ThesisPDF Available

Towards Longer Lives: The Ethics of Life Extension by Slowed Ageing

Authors:

Abstract and Figures

A significant reduction in calorie intake, known as calorie restriction (CR), has been shown to increase lifespan in a wide variety of animal subjects. If these results translated to humans, CR could substantially increase human lifespans by decelerating ageing. This possibility has led to an effort to develop calorie restriction mimetic drugs (CRMs) that mimic the effects of CR without the need to reduce calorie intake. This project examines the social and ethical implications of extending lifespans using CR and CRMs. The thesis is in three parts. Part I looks closer at the empirical questions about CR and CRMs, and in particular the issue of whether results from animal studies would translate to humans. I argue that although the evidence is far from conclusive, there are grounds to think that CR could slow ageing and extend lifespan in humans. Part II examines the implications of prolonging lifespan for individual welfare. I argue that historical and philosophical objections to life extension on the grounds of individual welfare are unsuccessful against CR. CR itself may have some undesirable effects, although these are due to the stringent diet and are unlikely to result from CRMs. Part III discusses the social impact of CRMs, assessing common ethical objections to life extension on the grounds of fairness and social welfare. I claim that it would be fair to distribute CRMs by public health services. Moreover, concerns about the demographic impact of longer lives can be mitigated. Indeed, a wide distribution of life extending technologies could improve social welfare. Overall, I claim that CR and CRMs are compatible with, and could further, values that are significant for individuals and societies.
Content may be subject to copyright.
PhD degree in Foundations of the Life Sciences and their Ethical
Consequences
European School of Molecular Medicine (SEMM) and
Department of Health Sciences, University of Milan
Settore disciplinare: FIL/02
Towards Longer Lives:
The Ethical Implications of Life Extension
by Calorie Restriction and Calorie
Restriction Mimetics
Christopher Wareham
IFOM-IEO Campus, Milan
Matricola n. R08431
Supervisors:
Prof. Giuseppe Testa
IFOM-IEO Campus, Milan
Dr Marco Giorgio
IFOM-IEO Campus, Milan
Prof John Harris
University of Manchester, Manchester
Anno accademico 2011-2012
CONTENTS
ABBREVIATIONS..............................................................................................................4
FIGURES .............................................................................................................................5
ABSTRACT .........................................................................................................................6
INTRODUCTION ...............................................................................................................7
i) Applied ethics ....................................................................................................................................9
ii) Applied ethics and ‘compatibilism’ ............................................................................................. 10
iii) A narrower focus .......................................................................................................................... 12
iv) Methodological limitations .......................................................................................................... 14
v) Conclusion ..................................................................................................................................... 15
PART I: EMPIRICAL QUESTIONS...............................................................................17
INTRODUCTION TO PART I........................................................................................17
1. WHAT ARE CR AND CRMS? ...................................................................................18
1.1 What is CR? ................................................................................................................................. 18
1.2 What are CRMs?.......................................................................................................................... 20
1.3 Conclusion.................................................................................................................................... 25
2. CR IN ANIMALS: LIFESPAN AND THE RATE OF AGEING .................................26
2.1 Lifespan measures of the rate of ageing ..................................................................................... 27
2.2 Rate of mortality measures.......................................................................................................... 29
2.3 Disease measures ......................................................................................................................... 31
2.4. Biomarkers of ageing and longevity.......................................................................................... 34
2.5 Conclusion.................................................................................................................................... 37
3. IMPLICATIONS FOR HUMANS: THE TRANSFER THESIS ...................................39
3.1 Lifespan predictions of the transfer thesis ................................................................................. 40
3.2 Doubts about the transfer thesis.................................................................................................. 43
3.3 Human studies of CR................................................................................................................... 54
3.4 Conclusion.................................................................................................................................... 65
CONCLUSION TO PART I ............................................................................................66
PART II: CR, CRMS AND INDIVIDUAL WELFARE ..................................................67
INTRODUCTION TO PART II ......................................................................................67
i) The significance of individual welfare.......................................................................................... 67
ii) What makes a life go better or worse? ......................................................................................... 68
iii) Distributions of welfare within a life .......................................................................................... 70
iv) Welfare and comparison .............................................................................................................. 71
v) Structure of arguments .................................................................................................................. 73
2
4. SUBSTANTIVE GOODS............................................................................................75
4.1 Health ............................................................................................................................................75
4.2 Procreation ....................................................................................................................................87
4.3 Self-development and flourishing ...............................................................................................92
4.4 Creativity and beauty....................................................................................................................96
4.5 Community ...................................................................................................................................99
4.6 Conclusion ..................................................................................................................................103
5. DESIRE FULFILMENT............................................................................................ 104
5.1 The badness of life......................................................................................................................106
5.2 Persistent desires.........................................................................................................................110
5.3 Changing desires.........................................................................................................................115
5.4 Desire satisfaction and receding deadlines ...............................................................................119
5.5 Conclusion ..................................................................................................................................122
6. MENTAL STATES................................................................................................... 123
6.1 Neutrality ....................................................................................................................................124
6.2 Suffering ......................................................................................................................................125
6.3 Declining happiness ...................................................................................................................129
6.4 Fear ..............................................................................................................................................131
6.5 Boredom ......................................................................................................................................133
6.6 Feeling old...................................................................................................................................138
6.7 CR, CRMs and Mental States ....................................................................................................144
6.8 Conclusion ..................................................................................................................................148
CONCLUSION TO PART II.........................................................................................149
PART III: CRMS AND SOCIAL VALUES................................................................... 151
INTRODUCTION TO PART III................................................................................... 151
i) What makes a society better or worse? ........................................................................................151
ii) The significance of social values ................................................................................................152
iii) Outline of arguments ..................................................................................................................154
7. TRANSLATION OF CRMS......................................................................................157
7.1 Over the counter .........................................................................................................................157
7.2 Regulatory approval ...................................................................................................................159
7.3 Provision by health services.......................................................................................................161
7.4 Conclusion ..................................................................................................................................166
8. FAIRNESS................................................................................................................ 169
8.1 The Fair Healthspan objection...................................................................................................170
8.2 Flawed responses: Laissez fair and banning .............................................................................171
8.3 Equal access through health services ........................................................................................175
8.4 Eliminating enhancement by unequal provision ......................................................................180
3
8.5 Conclusion.................................................................................................................................. 183
9. SOCIAL WELFARE .................................................................................................185
9.1 Social welfare, demography, and slowed ageing..................................................................... 185
9.2 Ageing societies ......................................................................................................................... 189
9.3 Overpopulation (at a time) ........................................................................................................ 203
9.4 Under-population (across time) ................................................................................................ 208
9.5 Conclusion.................................................................................................................................. 216
CONCLUSION TO PART III........................................................................................217
CONCLUSIONS..............................................................................................................219
i) Ethical conclusions....................................................................................................................... 219
ii) Conclusion on health policy ....................................................................................................... 220
iii) Directions for empirical research .............................................................................................. 221
iv) Conclusion on social policy ....................................................................................................... 223
v) Concluding remark ...................................................................................................................... 223
ACKNOWLEDGEMENTS.............................................................................................225
BIBLIOGRAPHY............................................................................................................226
4
ABBREVIATIONS
CR Calorie restriction
CRM Calorie restriction mimetics
2DG 2-deoxy-d-glucose
DHEA Dehydroepiandrosterone
DHEAS Dehydroepiandrosterone sulfate
CRS Caloric Restriction Society
CALERIE Comprehensive Assessment of Long-term Effects of Reducing Energy Intake
LET Life Extension Technology
h-LET Hypothetical Life Extension Technology
USPBC US Presidential Council on Bioethics
iPSC Induced Pluripotent Stem Cell
WHO World Health Organisation
QALY Quality Adjusted Life Year
NICE National Institute for Clinical Excellence
5
FIGURES
Figure 1. Lifespan predictions of the transfer thesis. Figure and data adapted from
Speakman and Hambly 2007............................................................................................41
Figure 2 Average and maximum lifespan in Okinawan, Japanese and US groups. (Willcox
et al 2007a)...................................................................................................................... 59
Figure 3 Age-related disease in Okinawans, Japanese and U.S groups. (Willcox et al
2007a). ............................................................................................................................63
Figure 4 Recent findings about potential health concerns.................................................86
Figure 5 Number of people at any age after 200 years of No life extension (left) and after
200 years of slowed ageing (right). ................................................................................189
Figure 6 Change in population size in CRM compared to Normal, given replacement rates
......................................................................................................................................206
6
ABSTRACT
A significant reduction in calorie intake, known as calorie restriction (CR), has been shown
to increase lifespan in a wide variety of animal subjects. If these results translated to
humans, CR could substantially increase human lifespans by decelerating ageing. This
possibility has led to an effort to develop calorie restriction mimetic drugs (CRMs) that
mimic the effects of CR without the need to reduce calorie intake. This project examines
the social and ethical implications of extending lifespans using CR and CRMs.
The thesis is in three parts. Part I looks closer at the empirical questions about CR and
CRMs, and in particular the issue of whether results from animal studies would translate to
humans. I argue that although the evidence is not conclusive, there are grounds to think
that CR could slow ageing and extend lifespan in humans.
Part II examines the implications of prolonging lifespan for individual welfare. I argue that
historical and philosophical objections to life extension on the grounds of individual
welfare are unsuccessful against CR. CR itself may have some undesirable effects,
although these are due to the stringent diet and are unlikely to result from CRMs.
Part III discusses the social impact of CRMs, assessing common ethical objections to life
extension on the grounds of fairness and social welfare. I claim that it would be fair for
public health services to distribute CRMs. Moreover, concerns about the demographic
impact of longer lives can be mitigated. Indeed, a wide distribution of life extending
technologies could improve social welfare.
Overall, I claim that CR and CRMs are compatible with, and could further, values that are
significant for individuals and societies.
7
INTRODUCTION
Ageing, widely regarded as a progressive decline that makes an organism less capable of
survival, is related to a host of maladies including cancers, cardiovascular diseases,
diabetes, and Alzheimer’s disease. As Harris has suggested, this means that there is a
strong chance that treating these and other diseases will have an impact on the rate of
ageing and thereby substantially prolong human lives (Harris 2004, 530). This raises the
prospect that life extension interventions may arrive largely unheralded.
Thus far, however, only one intervention—caloric restriction (CR)—has been shown to
consistently and substantially extend average and maximum lifespan in a wide variety of
organisms (Roth et al. 1995).1 As the name suggests, CR involves a significant reduction in
calorie intake. Remarkably this relatively simple intervention has been shown to prolong
the lives of rodents by as much as 60% (Speakman and Hambly 2007). If effective in
humans it would result in a substantial increase in average and maximum lifespan.2
Despite a growing number of CR practitioners, the caloric restriction regime is widely
thought to be too demanding for widespread human practice. As a result, biologists and
pharmaceutical companies are investigating a number of potential CR mimetics (CRMs)—
drugs that might replicate the effects of CR without the need to restrict calories.
Perhaps the most extensively investigated candidate CRMs are rapamycin, resveratrol and
metformin.3 Rapamycin is an immuno-suppressant commonly used in organ
transplantation. Metformin is a compound used in the treatment of diabetes. Resveratrol is
a polyphenol derived from grapes and found in small quantities in red wine. All of these
1 For an overview see Masoro 2005.
2 The extent of this increase is discussed in more detail in Chapter 3.
3 See Minor et al. 2010 for fuller discussion of recent research on candidate CRMs.
8
candidate CRMs have been trialled in humans. Indeed resveratrol-based products are
widely available without prescription in pharmacies, while metformin is used every day by
diabetes sufferers.
Research on CR and CRMs has garnered an increasing amount of attention in well-
respected scientific journals, in the media and, significantly, from the pharmaceutical
industry. Recently the pharmaceutical giant GlaxoSmithKline bought Sirtris, a
biotechnology company aimed at creating effective CRMs, for $720 million—an amount
indicative of the financial potential of such drugs.4 Despite this increased attention, and the
fact that life extension has been a human preoccupation for millennia, there is virtually no
discussion of the ethical implications of these accessible and potentially lifespan
augmenting interventions.
Clearly there is a large body of work, either literary or philosophical, that explores the
ethical implications of extending lifespan, and these concerns must be addressed. Yet there
is also a growing body of scientific literature yielding clearer ideas about the potential
benefits and drawbacks of life extension technologies. Research on CR and CRMs presents
the possibility to revisit the more speculative concerns about the impact of substantial life
extension on individual welfare with concrete empirical evidence in hand.
This thesis examines these potential ethical implications of prolonging life in light of
findings about CR and CRMs. I claim that, although longer lives raise important
challenges, research on CR and CRMs provides grounds to think that CRMs in particular
could contribute to a better and fairer society with healthier and more long-lived
individuals.
4 http://www.guardian.co.uk/business/2008/apr/23/glaxo.sirtris. Accessed 18 December 2012.
9
In the remainder of this introduction I discuss the role of empirical studies in this thesis and
in applied ethics generally. In doing so, I differentiate this project from other studies of life
extension in terms of its scope, and in terms of its focus on factual premises of moral
arguments. I also draw attention to a methodological tenet I refer to as ‘compatibilism’ and
discuss its strengths and weaknesses.
i) Applied ethics
Empirical studies play two significant roles in this thesis. First, they function as factual
premises in moral arguments; second, they allow a narrower focus that avoids unrealistic
and heavily hypothetical life extension scenarios.
The first role of empirical studies is as factual premises in ethical arguments. Below I
situate my methodology as a type of applied ethics. I explain the structure of arguments in
applied ethics, and the role of factual premises supplied by empirical studies.
The aims of applied ethics
Applied ethics attempts to arrive at justified conclusions about real ethical problems. Its
purpose is to inform decisions about what actions or decisions are good or right. The
claims that abortion is immoral,’ or that ‘murderers should receive the death penalty’ are
examples of claims in applied ethics.
This is a thesis in applied ethics. It aims to provide ethical discussion of, and guidance
about a potentially problematic technology. More narrowly, it can be regarded as a
bioethical study, since it applies ethical principles to an emerging biotechnology.
10
Argument in applied ethics
An argument in applied ethics must, of course, obey logical rules and has, at minimum, the
following components:
1. A moral premise which constitutes a moral conviction, principle or value
2. A factual premise
3. A practical conclusion that is logically entailed by the combination of the moral
and factual premises (Tännsjö 2011).
A simplified example is the following:
1. A person should do things that make her healthy
2. Eating vegetables makes a person healthy
3. So a person should eat vegetables
This simple structure means that there are three key ways that an argument in applied
ethics can contested. One can challenge the normative premise, the factual premise, or the
logical validity of the argument. Below I discuss how a focus on factual premises and
validity grounds a methodological approach I refer to as ‘compatibilism.’
ii) Applied ethics and ‘compatibilism’
My approach in this thesis is to accept the normative basis on which objections to life
extension are made. That is, I accept a range of normative moral premises, which can
include values and principles. Instead, I challenge the factual premises, and the validity, of
arguments about life extension.
I term this approach ‘compatibilism,’ because my aim is to determine the extent to which
life extension by CR and CRMs is compatible with the furtherance of a range of normative
11
principles. Even if we accept the substantive normative theories that underlie fears about
life extension, there may be reason to think that CRMs would be a good thing if these fears
are based on misapprehensions about the facts.
This approach is made workable by the narrower focus discussed below. Since I focus on a
particular life extending intervention about which much is known, it is possible to examine
the factual basis of concerns about life extension. I now discuss some of the strengths of,
and motivations for, compatibilism before examining its limitations.
Motivations for compatibilism
One motivation for assessing the compatibility between life extension and value claims is
to avoid pervasive disagreement at the level of values. If an intervention can be shown to
be compatible with various ethical principles, then it is possible to prevent some of this
disagreement. By taking moral premises for granted and instead challenging factual
premises, it may be possible to achieve an ‘overlapping consensus,’ (to use Rawls’s (1971)
term) on the morality of an intervention.
A further motivation for compatibilism is that if factual premises about an intervention
lead to an undesirable conclusion, it is sometimes possible to alter these premises. Facts
can sometimes be more malleable than values. Take, for example, the following argument:
1. It is immoral to cause an embryo to die.
2. Stem cell research causes embryos to die.
3. Thus stem cell research is immoral.
Much work in bioethics and applied ethics has focussed on disputing the morality of killing
embryos (premise 1). However, more recently, it has become possible to conduct stem cell
research that does not make use of embryonic stem cells. Induced pluripotent stem (iPS)
12
cells can be made from virtually any cell in the body.5 In effect the factual premise has
been altered, rendering false the above argument against the morality of stem cell research.
In a similar way, focusing on factual premises in arguments about life extension might help
draw attention to facts that are problematic, and which may similarly be subject to
biological manipulation. A relevant example of this, which emerges particularly in
Chapters 4 and 9, is the possibility of compressing or shortening morbidity. I highlight this
example in more detail the conclusion of this thesis.
These motivations point to ways in which the compatibilist use of empirical studies can
shed light on, and potentially resolve, ethical disputes even when moral premises are the
subject of deep and pervasive disagreement.
iii) A narrower focus
A second function of empirical studies is that they allows a narrower focus on a particular
type of life extending intervention. Below I explain what I mean by this by contrasting my
focus with broader ethical analyses of substantial life extension.
Broad approaches: life extension and other enhancements
At the broader end of the spectrum, prolonging life might be considered as one of many
different types of ‘enhancement’ technologies: interventions that raise individuals above
some level of welfare or health considered ‘normal’ (Daniels 2000). This type of broad
discussion focuses on general features of enhancement technologies and considers their
ethical status. Life extension would be judged alongside these.
5 See for example Cherry and Daley 2012.
13
This breadth of scope has the advantage of broader applicability, but often ignores more
fine-grained details. One example is the likelihood, discussed in Chapters 7 and 8, that
although potentially enhancements, CRMs will also be treatments, improving disease
conditions.
Narrower focus: substantial life extension itself
More narrowly, ethical theorists focus on a general category of life extending technologies
that could hypothetically exist. They attempt to identify ethical problems and sometimes
examine whether these problems can be resolved. Such evaluations are useful, since they
provide a rich source of potential problems an ethical gauntlet that a real life extending
technology must run. Considering the ethics of longer lives can also, arguably, clarify
questions about the meaning of life and why it’s good to be alive in the first place.
However, in relation to the implications of life extending technologies themselves, such
studies can be hampered by the lack of a factual basis. One can make use of ‘thought
experiments’ about anti-ageing drugs, but it is speculative that these thought experiments
will share any features with real technologies. As an example, in Chapter 9, I point out that
Singer’s argument against life extension fails to apply to CRMs, since it is relies on a
hypothetical life extension technology with little basis in fact.
Moreover, the lack of a factual basis often leads writers to focus on unlikely life extension
scenarios such as immortality. As Leigh Turner suggests, these discussions can
skew analysis by moving from legitimate concerns to a far more speculative, less
biologically grounded mode of deliberation. (Turner 2004, 220)
14
Narrower still: particular life extending interventions
At the narrower end of the scale, an ethical analysis of life extension examines the
implications of a particular intervention, or category of intervention, given empirical
studies. There are very few such analyses because no intervention is known with certainty
to substantially extend lifespan in humans.
Nonetheless the focus of this thesis is narrow in this sense. I examine the ethical
implications of life extension by CR and CRMs, in the light of empirical studies on CR.
This focus enables the making of claims that are more ‘biologically grounded,’ to use
Turner’s phrase, than any other claims about life extension that I have encountered.
However, it is important to note that my focus is not as narrow as it gets. If we knew with
certainty that an intervention extended lifespan in humans, we could conduct an even
narrower ethical inquiry – one without the need for well-grounded, but uncertain, claims.
This would obviously be more informative, both scientifically and ethically. However, as I
point out in Chapter 3, it is unlikely that we can have certainty that an intervention
substantially extends lifespan before it has done so.
iv) Methodological limitations
There are some aspects of the methodology I make use of in this thesis that could be
perceived as weaknesses. First, the factual premises I make use of may be shown to be
wrong in the long run. They are based on a hypothesis that is not known to be true that
the effects of CR will translate to humans and a hypothetical intervention – a drug that
will mimic these effects. Since the empirical claims are susceptible to falsification, some of
the ethical analysis of the effects of CR and CRMs may be rendered incorrect by current
and future research. Other accounts that rest less heavily on empirical studies are obviously
less susceptible to this outcome.
15
Second, genuine CRMs may have their own side-effects and ethical complications. It is not
possible to take these into account, since, as discussed in Part I, we do not know whether
there are any are genuine CRMs.
Despite these potential weaknesses, the claims advanced in this dissertation would have
broader applicability, even if some of the empirical claims turn out to be false or
incomplete. In the first place, the collection of arguments I discuss provides a framework
of values and potential problems that is useful for analysing any life extending
intervention.
Moreover, a key empirical premise I rely on is that of decelerated or slowed ageing, as
outlined in Chapter 2. Even if CR and CRMs do not slow ageing, it is not unlikely that
other interventions will. If so, many of the arguments will apply to these interventions.
v) Conclusion
This introduction has attempted to spell out the methodological role of empirical evidence
in this thesis. Two advantages are worth re-emphasising. First, the compatibilist focus on
acceptance of a variety of moral premises makes it possible to reach agreement about the
morality of an intervention despite disagreement about moral values. Second, a narrower
focus on a well-known intervention enables more fine-grained and less speculative claims
about life extension.
With these methodological points in place, I discuss the empirical evidence that will
inform the ethical argument in subsequent parts.
16
17
PART I: EMPIRICAL QUESTIONS
INTRODUCTION TO PART I
CR has been shown to increase average and maximum lifespan in multiple organisms
including yeast, flies, nematodes, mice, and dogs. This has resulted in a quest to investigate
and develop potential CRMs such as resveratrol, rapamycin, and metformin. This part
examines the empirical research on CR and CRMs.
Chapter 1 surveys empirical and terminological issues with defining CR and CRMs, and
gives examples of candidate CRMs. Chapter 2 discusses research in animals, and the
effects of CR on the rate of ageing. Chapter 3 explores the hypothesis, which I refer to as
the ‘transfer thesis,’ that the effects of CR observed in animals will translate to humans.
I claim that while it is difficult to be certain how CR and CRMs will affect humans in the
long-term, current findings from animal and human studies suggest that CR, and
interventions that mimic it, could slow ageing, leading to a substantially longer life. These
claims form the basis for the ethical analysis in subsequent chapters.
18
1. WHAT ARE CR AND CRMS?
In this chapter I discuss terminological differences in empirical literature that may cause
confusion. I also motivate a definition of CRMs, and discuss some interventions that are
regarded as potential CRMs.
1.1 What is CR?
Since McCay’s seminal studies on restricted food intake (McCay 1935), a number of
different experimental protocols have been referred to as calorie restriction. Conversely,
many of the very same protocols have been referred to by different names, including food
restriction, dietary restriction, energy restriction, and caloric restriction.
Although it is not uncommon to use these terms interchangeably (eg Redman et al 2007),
sometimes the use of a particular term has a purpose. For instance, the terms dietary or
food restriction are sometimes used operationally when attempting to isolate which aspects
of food restriction are responsible for the effects on longevity and ageing. For instance,
there have recently been attempts to restrict consumption of various dietary components,
such as proteins (Fontana et al 2008), fats (Sanz et al 2006), and molecules such as
methionine (Sun et al 2009), in an effort to isolate particular dietary inputs that contribute
to the effects of CR on longevity and ageing. The use of the terms food, or dietary
restriction in such cases expresses a suspension of judgement about any particular dietary
cause.
1.1.1 The term used in this thesis
Dietary restriction and food restriction are broader, encompassing the restriction of calories
and other dietary sources that may be responsible for the effects of reduced food intake.
19
This breadth of meaning means that the terms are somewhat inexact, and have been
referred to as ‘vague’ (Masoro 2006).
Calorie, or caloric, restriction is narrower, referring to the restriction of dietary energy.
There is general agreement that the restriction of energy, or calories, plays a large role in
the effects of a leaner diet (Masoro 2005). This means that it is somewhat justified to use
the term caloric restriction.
Even narrower terms are possible: as mentioned, protein restriction, and the restriction of
the amino acid methionine may contribute to the effects of a reduced diet. However, the
extent of the roles of protein and methionine are still disputed (ibid.). As a result I will use
the intermediate terms calorie, or caloric, restriction.
1.1.2 CR protocols
With this terminological query aside, I turn to the issue of what CR entails. Again this is a
vexed question. In its broadest formulation, CR involves a restriction of energy intake from
ad libitum levels, but without a reduction in essential nutrients. There are many difficulties
in constructing CR protocols and comparing their results.6 Here I mention two such
problems in order to clarify what CR and degrees of CR entail.
The first clarification concerns what is meant by an ad libitum or free-feeding diet.
Animals that are allowed to eat as much as they wish are often overweight. As a result it
has been suggested that some of the gains in longevity achieved by restricting calories may
be as a result of counteracting diseases related to overfeeding (Speakman and Mitchell
2011). To eliminate this possibility, it has become common to restrict the calorie intake of
6 See Speakman and Mitchell 2011 for a more detailed discussion of methodological difficulties in studies of
CR.
20
control mice to 10% of true ad libitum levels. This restriction serves as the comparandum
for greater reductions in calorie intake. Significantly, this may mean that, for example, a
50% reduction in calories relative to controls is actually a 60% reduction relative to true ad
libitum levels (Speakman and Hambly 2007). This has consequences for the degree of CR
that might be required in humans, and is thus important to mention.
The second clarification concerns interpretations of degrees or percentages of caloric
restriction. The extent of CR is expressed in two different ways. Sometimes a diet of 60%
that of controls - a reduction of 40% - is referred to as 60% CR, referring to the percentage
of an ad libitum diet that is consumed. However, the same degree of restriction is
sometimes called 40% CR, meaning that the organism’s diet is restricted by 40% from
normal levels. In this thesis I will use the latter formulation, so that 30% CR is 70% of the
diet of controls.
CR thus involves a specified restriction of caloric intake, without malnutrition. In Chapter
3, I will discuss factors that modulate the degree of life extension achieved by CR.
1.2 What are CRMs?
Broadly, CRMs are interventions that mimic the effects of caloric restriction. According to
Ingram and colleagues, the function of a CRM is
to trick the organism into a CR state and thereby activate the protective
mechanisms that are induced in CR. Moreover, the parallel objective should be to
minimize any reduction in actual caloric consumption. (Ingram et al. 2006, 99)
Ingram et al complain that in much of the literature, CRMs are construed too broadly as
including ‘any intervention that can evoke similar effects on aging, health, and lifespan to
those of CR’ (ibid., 98) This suggests there is a need to refine the conceptual scope of
21
CRMs. At the very least interventions that have radically different modes of action should
be excluded.
In order to clarify the aims of research into CRMs, Ingram et al propose that a CRM must
fulfil the following conditions:
(i) it mimics the metabolic, hormonal, and physiological effects of CR; (ii) it does
not significantly reduce long-term food intake; (iii) it activates stress response
pathways observed in CR and provides protection against a variety of stressors; and
(iv) it produces CR-like effects on longevity, reduction of age-related disease and
maintenance of function. (Ibid.)
However, it is possible that an intervention fulfils all the criteria, but lacks some inessential
hormonal, or metabolic effect. Indeed, presumably not all metabolic effects of restricting
calories could be replicated without actually restricting food intake. For instance, if food is
not restricted, then the metabolic activity of breaking food down to nutrients will not occur
to the same extent. This would be a difference in metabolism. So if we’re extremely strict
about the above definition, it’s not clear that anything could be a CRM without being CR
itself.
I will not focus extensively on definition of a CRM. Instead I will assert a working
definition of a CRM as an intervention that
i) does not involve long-term reduction of food intake and
ii) has identical effects to CR, barring only those effects that are solely
consequences of decreasing food intake.
If, for instance, a candidate CRM has exactly the same effects as CR, but doesn’t result in
weight loss or hunger, and if weight loss and hunger are solely consequences of having less
food, the candidate would successfully count as a CRM.
22
This narrow understanding of a CRM would have the effect of excluding interventions like
surgical reduction of the stomach and appetite suppressants, since both involve an actual
reduction of calorie intake. It would also inevitably exclude many of the current candidate
CRMs discussed below, since none is capable of perfectly mimicking CR. On the other
hand this conception of a CRM allows me to examine the ethical implications of CRMs
given the store of information that has been accrued across more than 80 years of research
on CR.
In the next sub-section I discuss three candidate CRMs, indicating the extent to which they
fulfil the criteria for a CRM suggested above.
1.2.1 Candidate CRMs
Numerous candidate CRMs are currently being investigated. These include 2-deoxy-d-
glucose (2DG), sirtuins such as resveratrol, rapamycin and rapalogs, and biguanides, such
as metformin, phenformin and buformin. Here I will discuss three that are regarded as most
promising: resveratrol, rapamycin, and metformin.7
Resveratrol is a polyphenol present in diverse dietary sources. It has been widely reported
in the media due to the fact that it is found in red wine and purportedly has ‘anti-ageing’
effects. Interestingly, the consumption of resveratrol-rich red wine has been proposed as an
explanation for the French Paradox: the phenomenon that France appears to have low
mortality rates, despite a traditionally high-fat diet and the prevalence of cigarette smoking
(Vidavalur et al 2006). The attraction of resveratrol is increased due to the fact that it is
generally regarded as safe for human consumption (Cottart et al 2010).
7 See Minor et al 2010 for fuller discussion of recent research on candidate CRMs.
23
Experiments on resveratrol suggest that, in sufficient quantities, it mimics many of the
effects of CR discussed in more detail later. In particular, in mice it has many of the same
effects in reducing the incidence of cardio-vascular disease and cancer (Barger et al 2008).
However, it has only been shown to be moderately successful in increasing maximum
lifespan. Only mice fed a very high fat diet have substantially higher maximum lifespan as
a result of resveratrol treatment (Baur et al 2006). Further, resveratrol is not easily
absorbed by the human digestive system, and the quantities needed to replicate the effects
of resveratrol in humans are thought to be large (Walle 2011).
Nonetheless, findings on resveratrol have been encouraging enough to generate massive
investment in developing a marketable version that would be effective in humans. As
mentioned in the introduction, Sirtris, a research-based company aimed at making a drug
that would have the same effects as resveratrol, was bought by the pharmaceutical
company GlaxoSmithKline for a staggering amount of money. Currently Sirtris is
conducting clinical trials on a number of resveratrol-derived drugs that may prove to be
effective CRMs.
Rapamycin has been shown to increase lifespan in a variety of genetically different rodents
(Harrison 2009; Miller 2011). Like CR it has anti-tumor effects, and is believed to slow
ageing. These effects have led to the development of rapalogs – drugs derived from
rapamycin such as everolimus – that are used in the treatment of cancer (Minor et al 2010).
However, rapamycin is also a potent immune-suppressant used in organ transplantation
(ibid.). Thus it may have negative effects on the immune system, although it is possible
that lower dosage may mimic the life effects of CR without introducing additional
deleterious consequences (Kaeberlein 2010). Indeed, some regard rapamycin as a viable
24
CRM (Blagosklonny 2010) while others suggest that it is the best candidate for the
development of CRMs (Sierra et al 2009).
Metformin is the most widely used drug for the treatment of type 2 diabetes. It is regarded
as a candidate CRM because it results in similar gene expression patterns (Minor et al.
2010, 2). Like metformin, CR is highly effective in the treatment of diabetes (Jonker et al.
2011). Metformin has also been shown to increase maximum lifespan in mice, though not
by as much as CR (Anisimov et al. 2011).
Unlike the above CRM candidates, metformin has been in general use for more than half a
century, having been introduced in the United Kingdom in 1958. It is clinically approved
and has been used by millions of humans. Significantly, this means it is possible to conduct
long-term mortality and morbidity studies in humans. Such studies have indicated that in
addition to its on-label effect of treating diabetes, metformin has cardio-protective and
anti-atherosclerotic effects (Scarpello 2003). Recent studies also suggest metformin may
also have a role in tumour suppression (Scarpello 2008). Again, these effects are similar to
those that occur as a result of CR.
It should be borne in mind that metformin may not meet the strict CRM criteria. In
particular, I am unaware of any controlled studies comparing the life expectancy of
metformin users to that of non-diabetics. Without these studies it is not clear that it
increases life expectancy in humans. Even so, metformin is of particular interest as a CRM
candidate due to its approved status and the fact that it is already widely used. This makes
it perhaps the most realistic target for health policy on CRMs, as discussed in the
conclusion of this thesis.
25
1.3 Conclusion
In this chapter I have discussed some terminological and empirical issues in research on
CR and CRMs. I also gave examples of candidate CRMs. Although none of these
candidates fully replicate the effects of CR, they are already of ethical interest due to their
CR-like effects and their widespread use in humans. Moreover, increased investment in
these and other compounds makes it highly plausible that research will lead to the
development of CRMs capable of reproducing the effects of CR discussed in Chapters 2
and 3.
26
2. CR IN ANIMALS: LIFESPAN AND THE
RATE OF AGEING
Animal studies suggest that CR extends lifespan by slowing ageing.8 One influential
definition of ageing is the following:
Aging (senescence) is defined as the deteriorative changes, during the adult period
of life, which underlie an increasing vulnerability to challenges, thereby decreasing
the ability of the organism to survive (Masoro 2006, 15).
In these terms, ageing is a progressive deterioration that makes an organism less capable of
survival. CR is widely regarded as slowing this process, thus allowing organisms to
survive for longer.
However, the claim that CR slows ageing is complicated by the fact that several different
measures of the rate of ageing are used in empirical studies. In this chapter I outline these
measures, and discuss the results that have been achieved in animal models. The discussion
below focuses primarily on mammals. Although CR has been shown to have life extending
effects in organisms such as yeast, nematodes and fruit flies, studies in mammals are more
likely to be of direct relevance to humans ageing.
In discussing different measures of the rate of ageing, I will point out some of their
limitations. However, my purpose here is not to privilege any particular measure or to
arrive at a clear cut operational account of the rate of ageing. Instead, my aim is to clarify
the claim that CR slows ageing. This is important since, as will become clear measures of
8 Slowed ageing is also referred to as retarded or decelerated senescence. Senescence more commonly refers
to cellular, rather than organismic senescence, so I will usually use the term ageing. See Campisi and D’Adda
di Fagagna 2007 for a review of the connection between cellular senescence and ageing.
27
ageing are subject to criticisms both about their usefulness and their relation to ageing
itself.
Examining claims of slowed ageing also gives a better understanding of the effects of CR
in non-human animals that lead to this attribution. In particular, I will indicate four
significant conclusions, each of which corresponds to a particular measure of ageing
applied to animal studies. Lifespan measures, rate of mortality measures, disease measures,
and biomarker measures of the rate of ageing all point to different respects in which CR
slows the rate of ageing in animals.
2.1 Lifespan measures of the rate of ageing
The idea that CR slows the rate of ageing was initially based on the observation that it
increases the average and maximum lifespan in populations of animals studied, and
reduces the age-relative likelihood of death. In particular, CR mice and rodents have been
known to live as much as 60 percent longer than controls.9
Studies on rhesus monkeys are in relatively early stage. Although it is too early to know
whether CR will increase average and maximum and average lifespan, results so far are
promising. As of 2009 only five of thirty-eight CR animals in one study had died of age-
related causes, compared to thirteen out of thirty-eight control animals (Colman et al
2009).10 This is in keeping with studies on rodents that suggest a lower age-relative
likelihood of death.
9 Fontana, Partridge and Longo 2010 provide an overview.
10 Intriguingly a very recent study has failed to replicate these results (Mattison et al. 2012). Results from this
research require further analysis and may provide an opportunity to explore the complex impact of dietary
component restriction on the rate of ageing as measured by lifespan.
28
2.1.1 Limitations of lifespan measures
Determining the rate of ageing on the basis of lifespan has at least one conceptual
shortcoming: decreased chance of survival is not all that is important about ageing. Ageing
is also associated with increased susceptibility to disease and physical decline. This is a
significant reason why we are interested in the rate of ageing in the first place, and lifespan
measures alone tell us very little about the prevalence of disease and deterioration.
Lifespan measures of the rate of ageing also present methodological and practical
difficulties. In particular, the age at which the oldest organisms die changes with the size of
the group studied. Perhaps unsurprisingly the oldest organisms in large samples tend to be
older than the oldest organisms in smaller samples (Speakman and Mitchell 2011). As a
result, it has become common to record maximum lifespan as the average age of a
percentage of the last survivors – often 10%.
A further practical problem is that lifespan measures of the rate of ageing are difficult to
apply in studies of longer-lived animals such as humans. For instance, the oldest recorded
human is Jeanne Calment, who lived to 122 years and 164 days. If we take Jeanne Calment
as our control, according to maximum lifespan measures we would only know if CR
slowed ageing in humans if the oldest members of the CR population surpassed this age (or
perhaps if the oldest 10% of the CR population lived longer than the oldest 10% of the
human population). Clearly it would be useful to find a shorter term method to determine
rates of ageing.
2.1.2. Effects of CR on lifespan
Although these limitations are important, lifespan measures of ageing yield a significant
finding:
29
1. CR increases average and maximum lifespan relative to control groups and
decreases the age-relative likelihood of death.
The extent to which CR increases lifespan varies in accordance with the degree of CR and
the time of life at which the intervention is begun. These factors are discussed in more
detail in S3.1.
2.2 Rate of mortality measures
In studies of CR the rate of ageing is often measured on the basis of the rate of mortality.
Slowed ageing is said to occur if, after an increase in the number of deaths in a group
which signifies the onset of age-related death – this number takes longer to increase than it
does in control groups.
For example, suppose a control group of mice lives to two years before individuals within
the group begin to die more frequently. Two months after this acceleration, all are dead.
Now suppose another group lives to two years and five months before individuals in the
group start dying more frequently. Four months after this acceleration, all are dead.
In this case, the period in which increased mortality occurs is extended by an additional
two months. Since, after an initial point of increased mortality, deaths occur across a
longer span of time than a control group, this is an example of slowed ageing on rate of
mortality measures (Merry 2005).
This slowed ageing is in contrast to delayed ageing. With delayed ageing only the onset of
increased mortality is delayed. After this, animals die at approximately the same rate. In
the example above, if the second group of animals lived until two years and five months
30
before the rate of death increased, and then lived a further two months after increased
mortality (the same as the control group), this would be an example of delayed ageing.
A further possibility is accelerated ageing. With accelerated ageing the period after the
increased onset of age-related mortality is compressed relative to controls. In the example
above, accelerated ageing is in evidence if the period between increased mortality and the
death of the last animals is less than two months.
Mortality studies on CR rodents point to slowed ageing according to rate of mortality
measures. The rate of mortality after the initial age-related increase is slowed down
(Weindruch 1996; Weindruch and Sohal 1997; Masoro 2005).
2.2.1 Limitations of rate of mortality measures
Rate of mortality measures of biological ageing are important tools. However, there are
obvious shortcomings. Here I will mention four. The first three are held in common with
lifespan measures: first, they tell us very little about the health effects of CR; second, they
do not tell us about the physiological processes that occur prior to the period of increased
mortality; third, they are again very difficult to apply in human experiments or in
experiments on other long-lived mammals.
The fourth problem is in direct contrast to lifespan measures. Rate of mortality measures
contradict the idea that length of life is an important factor in the rate of ageing. For
instance, on these measures it is possible that a group of mice that lived a shorter amount
of time than a control group would be regarded as having a slower rate of ageing than the
control group, or that a group that lived longer than a control group would be regarded as
having a faster or equivalent rate of ageing. Indeed the latter finding increased lifespan,
but rapid ageing was reported by George Sacher upon the administration of procaine
31
(Sacher 1977). In this case, despite decreased deaths at all ages, rate of mortality measures
yielded the conclusion that mice administered procaine aged at the same rate as control
mice
This type of finding can occur because the rate of ageing is determined by how quickly
animals die after reaching a ‘tipping point’ of increased group mortality. Nothing that
occurs before this tipping point, nor the length of the life as a whole, enters the equation
about the rate of ageing. This leads to the above conclusions in Sacher’s study, which
Masoro describes as ‘absurd’ (Masoro 2006, 16). Intuitively, there are processes occurring
before the tipping point that could be highly relevant to the rate of ageing. This is backed
by findings that suggest that age-related decline begins much earlier in life: perhaps earlier
than thirty in humans (Sehl and Yates 2001; Nakamura and Miyao 2003.)
2.2.2 Effects of CR on rate of mortality
Despite these limitations, studies that make use of rate of mortality measures lead to a
second sense in which CR slows ageing and a second independently significant conclusion:
2. Deaths in CR populations tend to occur across a longer span of time than those in
control animals.
2.3 Disease measures
Studies of CR have also concentrated on the occurrence of age-related diseases. Measuring
the occurrence of age-related diseases is a promising way to determine the rate at which an
organism is ageing. Age-correlated diseases might be regarded as examples of the
‘deteriorative changes’ that constitute ageing and decrease an organism’s chance of
survival. Indeed the above approaches have been criticised for their excessive focus on
32
mortality. Instead critics urge that morbidity, or disease, is ‘much more informative about
the aging process’ than mortality (Sundberg et al 2011).
In the mammals in which it has been applied, CR delays the occurrence of, and slows the
development of, age-related diseases (Masoro 2006).11 In particular, it has been shown to
have beneficial effects for the three main causes of death in rodents: it inhibits cancers, and
delays or prevents kidney diseases, and heart diseases. It also prevents diabetes, auto-
immune disorders, and respiratory diseases (Speakman and Mitchell 2011). Moreover, CR
was found to have positive effects in mouse models of debilitating human diseases,
including Alzheimer’s disease, Huntington’s disease, and Parkinson’s disease (Omodei and
Fontana 2011). Disease measures of the rate of ageing in rodents thus support the idea that
CR slows ageing.
Disease measures also have the advantage of taking less time for results than maximum
lifespan and rate of mortality measures. As a result studies are beginning to emerge on
longer-lived mammals, which are more likely to be of relevance to humans. Perhaps the
most striking observations in this respect again come from Colman and colleagues’
ongoing study on calorically restricted rhesus monkeys, begun in 1989.
In rhesus monkeys, as in humans, cardiovascular disease, diabetes and cancer are amongst
the most prevalent age-related diseases (Colman et al 2009). This is one factor that
supports the potential transferability of research on rhesus monkeys to humans.
As of 2009, the monkeys in Colman and colleagues’ study were 20 years of age around
half the maximum lifespan of rhesus monkeys. The study found that age-related disease is
significantly reduced in the CR population. Remarkably, sixteen of the thirty-eight control
11 See Pallavi, Giorgio and Pellici 2012 for a discussion of the impact of CR on cancer and healthy ageing.
33
animals were either diabetic or pre-diabetic, while none of the CR animals showed signs of
diabetes.
The incidence of cancer and cardio-vascular disease was reduced by 50% in CR monkeys
compared to controls. Eight control monkeys had developed neoplasms, compared to only
four CR monkeys. Four control monkeys had developed cardiovascular disease, compared
to two in CR monkeys. Overall, age-related disease occurred at about three times the rate
in control animals as in CR animals. Although the study is ongoing, disease measures of
the rate of ageing indicate that ageing is slowed down in the rhesus monkeys.
2.3.1 Limitations of disease measures
On the basis disease measures, it seems clear that ageing is significantly slowed down in a
wide variety of animals. As suggested, these measures have advantages over maximum
lifespan and rate of mortality measures. However, their relevance to the ageing process has
also been questioned. Some biologists distinguish between primary and secondary ageing.
Primary ageing, also referred to as intrinsic ageing, is regarded as
the inevitable, progressive decline in tissue structure and biological function that
occurs with advancing age, independently of disease or harmful lifestyle and
environmental factors. (Holloszy and Fontana 2007, 709)
Secondary, or extrinsic ageing is defined as
the deterioration in tissue structure and biological function that is secondary to
disease processes and harmful environmental factors. (Ibid.)
If this distinction is accepted, disease does not tell us about the underlying processes of
primary ageing, but instead gives a measure of the rate of secondary ageing.
34
The distinction between primary and secondary ageing has been justified on the grounds
that, although all animals age, not all animals experience age-related disease. Some birds,
for instance, appear to have in-built ‘biological clocks’ which result in them dying
‘catastrophically’ without morbidity (Ricklefs and Scheuerlein 2001). If correct, this
distinction implies that disease measures of the effects of CR may only provide evidence of
the likelihood of slowed secondary ageing, while telling us nothing about the underlying
intrinsic processes of ageing (Masoro 2006; Hayflick 2004).
However, the distinction between basic ageing processes and disease processes is not
universally accepted (Sprott 2010). In many animals, it is not always clear how diseases
that relate to secondary ageing can be distinguished from ‘progressive decline in tissue
structure and biological function.’
2.3.2 Effect of CR on diseases
Despite these controversies it is clear that ageing is slowed on disease measures of the rate
of ageing. From this I extract a third finding that does not rely on the controversial
question of whether disease measures are adequate measures of the rate of ageing per se:
3. CR postpones the onset of age-related diseases, and reduces the age-relative
likelihood of organisms having age-related diseases.
If this finding is applicable to humans, it is of obvious ethical importance. In later chapters
I will argue that it also has significance for health policy on CRMs.
2.4. Biomarkers of ageing and longevity
Biomarkers of ageing are physiological and behavioural indicators of degrees of ageing.
Identifying and testing candidate biomarkers of ageing is a methodologically difficult and
35
conceptually fraught endeavour (ibid.). Miller argues that biomarkers of ageing should be
traits that meet the following three conditions:
1. The biomarker should predict the outcome of a wide range of age-sensitive tests
in multiple physiological and behavioral domains, in an age-coherent way, and do
so better than chronological age;
2. It should predict remaining longevity at an age at which 90% of the population is
still alive, and do so for most of the specific illnesses that afflict the species under
study;
3. Its measurement should not alter life expectancy or the outcome of subsequent
tests of other age-sensitive tests. (in Butler et al 2004, 561)
Ingram and colleagues point out that an ideal of biomarker research is to identify features
that are transferable between animal models and humans (Ingram et al. 2001, 1025-6).
They suggest that
[I]f a candidate biomarker is a valid measure of the rate of aging, then the rate of
age-related change in the biomarker should be proportional to differences in
lifespan among related species. Thus, for example, the rate of change in a candidate
biomarker of aging in chimpanzees should be twice that of humans (60 vs 120
years maximum lifespan); in rhesus monkeys three times that of humans (40 vs 120
years maximum lifespan).
Numerous candidate biomarkers have been suggested. Appearance and risk of disease,
brain changes, levels of dehydroepiandrosterone (DHEA) and its sulfate (DHEAS), grip
strength, and the appearance of grey hair and wrinkles are some examples. Many, if not all
of the biomarkers listed above will fail to fully meet the ambitious standards outlined by
Miller and Ingram. Nonetheless, I will discuss these purported biomarkers and the results
that have been obtained when they are used in studies of caloric restriction.
36
CR and candidate biomarkers
As discussed, the prevalence and prominence of diseases in animals is sometimes taken as
an indicator of the rate of ageing. To the extent that such measures are reliable biomarkers,
they point to slowed ageing, as discussed in the previous section.
One characteristic of ageing in humans and primates is brain atrophy. In the above-
mentioned study of rhesus monkeys, investigators used brain imaging techniques to
examine the effects of CR relative to controls. They found that CR reduced age-associated
brain atrophy in several brain regions (Colman et al 2009). Thus if degree of brain atrophy
is regarded as a biomarker of ageing, it appears that CR slows ageing according to this
measure.
Levels of DHEA and DHEAS have been suggested as candidate biomarkers because levels
of these steroids decline with age (Lane et al.1997; Roth et al. 2002). In a study of CR
rhesus monkeys levels of DHEA were increased relative to controls, suggesting a slower
rate of ageing (Mattison et al 2003).
A relatively successful predictor of age is appearance, or perceived age. Another is
physical activity. On the basis of these more mundane biomarkers, humans are able to
predict the relative age of organisms – particularly those species whose ageing we are
familiar with – with a surprising degree of accuracy (Christensen 2009). Researchers often
report that the calorically restricted animals they work with appear much younger than
control groups (eg Gems 2011; Colman et al 2009). They are also likely to maintain
physical activity for a longer period of time. These easily recognisable biomarkers of
ageing suggest that ageing is slowed down in CR populations.
37
2.4.1 Limitations of biomarker measures
Once again, there are numerous problems with developing effective biomarkers. As is the
case with disease measures of ageing, there is no general agreement on which
physiological variables track age. Moreover, there are great difficulties associated with the
transferability of physiological traits between very different categories of organisms. It is
not obvious, for instance, that a successful biomarker of age in a mouse would have an
analogue in higher mammals like monkeys or humans.
In addition, it is possible that an intervention that slows ageing in some organ systems may
not do so in all systems. This means that even if accepted biomarkers suggest slowed
ageing, some organs relevant to mortality and morbidity may age at a different rate.
Organismic unity of ageing processes is by no means certain. This means that ‘batteries’ of
biomarkers might be necessary to provide robust predictions about degrees and speeds of
ageing (Sprott, 2010).
2.4.2 Effects of CR on biomarkers
Despite these controversies it appears that on the basis of the, albeit imperfect, biomarkers
in use, ageing is slowed in a wide variety of organisms. This leads to the fourth conclusion
derived from animal studies of ageing rates under CR:
4. CR delays the appearance and development of physiological and behavioural
traits –biomarkers – associated with ageing.
2.5 Conclusion
To summarise, measures of the rate of ageing in animals suggest that ageing is slowed in
four respects:
1. CR increases average and maximum lifespan relative to control groups and
decreases the age-relative risk of death.
38
2. Deaths in CR populations tend to occur across a longer span of time than those in
control animals.
3. CR postpones the onset of age-related diseases, and reduces the age-specific
likelihood of organisms having age-related diseases.
4. CR delays the appearance and development of physiological and behavioural traits
associated with ageing.
The following chapter examines debates about whether these effects would translate to
humans.
39
3. IMPLICATIONS FOR HUMANS: THE
TRANSFER THESIS
A crucial question for ageing research is whether effects observed in calorically restricted
animals will be replicated in humans. In what follows I will refer to the claim that these
results would translate to humans as the ‘transfer thesis’. This chapter examines the
predictions and viability of transfer thesis.
The lack of confidence about the effects of CR in humans is due in part to difficulties
designing studies in humans. There are obvious practical and methodological problems in
conducting sufficiently long-term experiments on free-living humans. This means it is
unlikely that fully satisfactory tests of the transfer thesis can be carried out.
Nonetheless, surrogate measures have been developed. Studies of calorically restricted
groups such as the participants in the Biosphere Two experiment, members of the caloric
restriction society, the CALERIE project, as well as calorically restricted Okinawans
provide a basis – albeit a methodologically limited one – for evaluating the transfer thesis.
I begin this chapter outlining the predictions of the transfer thesis with respect to human
life expectancy. Thereafter I discuss a number of criticisms that have been made by of the
idea that CR will extend lifespan in humans. I argue that these criticisms are themselves
subject to doubts, and fail to undermine the viability of the transfer thesis. Finally, I
examine the transfer thesis on the basis of human studies. I make the case that the results of
such studies suggest that many of the physiological changes that occur in animals are
replicated in humans, indicating a likelihood of slowed ageing and increased average and
40
maximum lifespan. This gives considerable support to the idea that results from animal
studies on CRMs would translate to humans.
3.1 Lifespan predictions of the transfer thesis
Before assessing the plausibility of the transfer thesis, it is necessary to clarify its claims
with respect to humans. In particular it is necessary to provide estimates of longevity that it
predicts in humans, since the effects of CR on human lifespan are the main issue raised by
critics of the transfer thesis. Below I outline the factors that influence lifespan increases,
and suggest some estimates of longevity that will inform the ethical analysis in chapters to
follow.
3.1.1 Modulators of lifespan increase
Speakman and Hambly (2007) point out that two factors heavily influence the degree of
life extension achieved by CR animals. The first is the degree of caloric restriction; that is,
the reduction in food intake relative to controls. The second is the percentage of life
remaining at onset; that is how early or late in life the intervention is begun. The greatest
increases in lifespan occur when CR is initiated very early in life, and when calories are
severely restricted. Negligible increases occur when CR is initiated very late in life, or
when the degree of CR is very small.
With regard to the effects of different degrees of restriction, the greatest increases in
lifespan in rodents have occurred at a restriction of approximately 60% (Speakman and
Hambly 2007). That is, animals were restricted to 40% of the diet of control mice. These
have been reported to result in an increase in lifespan of 50%. Restrictions of 30%
increased average and maximum lifespan by about 20%.
41
Lifespan gains also differ according to the time of life at which CR is commenced. The
above gains were achieved in mice that commenced restriction at a very young age: before
they had been weaned. However, studies on rodents suggest that the longevity effect of CR
is reduced the later in life that CR is initiated (ibid.)
In humans it is unlikely that restriction would begin as early in life as it does in rodents.
This is due in part to ethical factors discussed in Chapter 4: such a strict regimen on
children appears likely to stunt children’s growth. Practising CR much later in life would
reduce the extent of life extension achieved. Extrapolations from mouse data suggest that
after the age of 50, very little additional life would be gained (see figure 1 below). This
means that later restrictors are less likely to achieve the substantial gains in lifespan that
might occur if CR was initiated earlier.
3.1.2 Lifespan predictions
If humans experience gains in lifespan proportionate to those in rodents, the lifespan of CR
practitioners would increase substantially. The table below details the life extension that
could be achieved given different degrees of restriction beginning at different ages,
assuming the truth of the transfer thesis.
Percentage
life at onset
Years
restricting
Years added by degree
of caloric restriction
Average
lifespan at
60%
restriction
Statistical
maximum
lifespan at
60%
restriction
Absolute
maximum
lifespan at
60%
restriction
60%
45%
30%
15%
20
62.4
22.4
16.8
11.2
5.6
100.4
131.66
157.48
30
54.6
18.8
14.1
9.4
4.7
96.8
127.72
153.26
40
46.8
13.2
9.9
6.6
3.3
91.2
120.69
145.05
50
39
9.6
7.2
4.8
2.4
87.6
116.26
139.94
60
31.2
5.6
4.2
2.8
1.4
83.6
110.73
133.14
70
23.4
0.22
0.75
0.5
0.25
78.22
101.44
120.61
80
15.6
0
0
0
0
78
101
120
Figure 1. Lifespan predictions of the transfer thesis. Figure and data adapted from Speakman and
Hambly 2007.
42
The numbers in the above table are adapted from Speakman and Hambly (2007), who
extrapolate lifespan data from rodents to humans. They assume that average lifespan is 78
years.
Note that estimates for 60% and 45% are extrapolated from their calculations, but are not
included in their text. Similarly, they do not provide estimates for maximum lifespan.
Here, maximum lifespan was estimated on the basis of that it is commonly reported that
CR increases maximum lifespan by the same proportion of years spent restricting as
average lifespan.
The current statistical maximum lifespan is taken to be 101. This figure (101 years) is
adapted from Willcox and colleagues, who give the maximum lifespan of US and Japanese
people as 101.1 and 101.3 respectively (Willcox et al. 2007). Although the absolute
maximum lifespan achieved is greater than this (just under 123 years), the lower number
statistical maximum lifespan is in keeping with the lifespan measures discussed earlier
(S2.1). That is, it takes the average longevity of a top percentile as the maximum lifespan.
The absolute maximum lifespan assumes a current maximum lifespan of 120 years.
On the basis of these figures I suppose that, if the transfer thesis is correct, CR and CRMs
would increase the absolute maximum lifespan to at most 160 years and average lifespan to
at most 100 years. Of course it is highly unlikely that humans would be able to achieve a
60% reduction in calories, so no-one is likely to reach the upper limit using CR. However,
retaining this upper limit for the purposes of ethical analysis is justified, since it may
become possible to achieve the effects of 60% CR using a CRM. These figures thus
represent good guesses as to the greatest extent of life extension that would be achieved
under differing periods and degrees of CR. All this assumes that the transfer thesis can be
justified, an issue to which I now turn.
43
In order to evaluate the transfer thesis I proceed as follows: first, I assess claims that results
in animals will not translate to humans; second, I examine the difficulties faced in studying
CR in human populations; finally, I discuss studies in humans and indicate the extent to
which they provide support for the transfer thesis.
3.2 Doubts about the transfer thesis
Broad questions remain unanswered about the extent to which effects achieved in animal
studies can be expected to occur in humans. There are significant genetic differences
between different strains of rodents, let alone between rodents and humans. Studies on
genetically closer rhesus monkeys mean that this chasm can be narrowed somewhat, but
differences remain.
Nonetheless the near ubiquitous inter and intra-species effects of CR on ageing and
longevity provide some reason to think that these effects would be retained in humans.
Below I discuss the claims of several commentators that dispute this inference and, indeed,
the extent to which CR effects are ubiquitous. Many of these arguments are based on
evolutionary theories, so it is worth briefly outlining a common evolutionary explanation
for the effects of caloric restriction.
The most widely accepted evolutionary account of why there is a CR effect is that it
evolved as a response to conditions of famine. In such conditions of reduced availability of
energy from food, having more offspring would strain already scarce food resources and
reduce survival. In times of famine, then, it may be advantageous to allocate resources
from reproduction to maintenance and repair of age-related damage. Doing so may allow
organisms to survive to reproduce in times of greater food resources. As such, the CR
44
effect may be a useful adaptation (Masoro and Austad 1996). In the following sections, I
discuss claims that the CR effect would not occur in humans.
3.2.1. Differing energy costs of reproduction
Shanley and Kirkwood, as well as Phelan and Rose claim that CR will not have as great a
longevity effect in humans as in rodents, due to the difference in the relative energy costs
of reproduction in rodents and humans (Shanley and Kirkwood 2006; Phelan and Rose
2005).
The trade-off hypothesis
These authors suggest that lifespans evolve as a result of strategies involving trade-offs
between increased longevity and increased reproduction. I refer to this as the trade-off
hypothesis. This hypothesis predicts that, other things being equal, organisms with higher
reproductive investment and greater fertility will live a shorter time, and organisms with
lower reproductive investment will live longer.
The trade-off hypothesis explains, for instance, the fact that the house mouse lives a short
time (less than 5 years) in which they reproduce intensively, while the little brown bat can
live a very long time (more than 33 years), but reproduces a single offspring per year. The
mouse invests heavily in rapid reproduction at a cost to longevity, while the bat invests
more in longevity and less in reproduction, and so produces offspring more slowly. Each of
these may be a viable evolutionary strategy and results in a different lifespan.
With regard to CR, this reproduction-longevity trade-off suggests that, in conditions of
famine, energy resources usually directed to maximising reproduction are instead
redirected to maintenance and repair. Again, this is because organisms that defer
procreation to more plentiful times have a reproductive and survival advantage. The
45
increased longevity that results from CR, therefore, is a result of reallocation of the energy
resource budget from reproduction to survival.
Lower reproductive investment, less longevity gain
Shanley and Kirkwood and Phelan and Rose note that humans have a much smaller
reproductive investment than rodents in relative terms. Rodents have more offspring than
humans and at a far greater relative energetic cost. Because humans invest less on
reproduction, redirecting resources from reproduction to maintenance would not have as
great a longevity effect in humans as it does in rodents. Procreation takes up much less of
human beings’ energy budget, so saving this fraction and spending it on maintenance
wouldn’t earn one much more lifespan.
Note that the general evolutionary explanation of the CR effect discussed above need not
make this prediction, since it does not postulate a trade-off between reproduction and
longevity. The general theory suggests that mechanisms that increase maintenance and
defer reproduction can be advantageous without predicting that doing so would reduce
overall fertility.
Criticisms of the trade-off hypothesis
The idea of evolutionary trade-offs between reproduction and longevity has been criticised
on the grounds that some relatively long-lived organisms have a great investment in
reproduction. Speakman cites the example of naked mole rats that live up to 28 years,
which is exceptional given their size, and which nonetheless produce one of the largest
litters of any mammal (Speakman 2011). Cases of this type, involving high levels of both
maintenance and reproduction, cast doubt on the broad applicability of the trade-off
hypothesis.
46
Of more direct relevance is the fact that a fundamental prediction of the hypothesis with
respect to CR is not borne out in experiments on CR animals. In particular, it predicts
animals that have endured CR will have increased lifespan and reduced fertility with
respect to controls. However, it has been found that, after emerging from a period of CR,
rodents have both increased fertility and increased lifespan (Selesniemi 2008). The fact
that both reproduction and lifespan are enhanced in animals that have been on CR suggests
that a trade-off between them does not explain the CR effect (Speakman and Mitchell
2011).
Conclusions on the trade-off hypothesis
The reproduction-longevity trade-off hypothesis predicts that CR will not extend lifespan
as much in humans, since less of our energy budget is directed towards reproduction.
However, the hypothesis is faced by problematic counter-examples like the naked mole rat.
Moreover, experiments on CR suggest that its effects do not result from a lifetime trade-off
between reproduction and longevity. Since this is so there are grounds to doubt the
prediction at issue: that CR would not extend lifespan in humans.
3.2.2. Migration as a famine response
Le Bourg claims that species, such as humans, that are capable of migrating and which are
relatively fast-moving should not be expected to have increased lifespan. He surmises that
rather than reducing reproduction in response to famine, many organisms such as flies,
birds capable of flight, and humans would simply migrate. In this way the costs to
reproduction could be avoided without the need for a life extension response. I refer to this
as the migration thesis.
47
Le Bourg makes several predictions. In particular he suggests that his hypothesis might be
falsified if flighted birds experience life extension in response to CR, and, conversely, if
non-flighted birds do not have extended lifespans in response to CR.
Preliminary evidence suggests that hens have increased lifespan (Holmes & Ottinger
2003), in keeping with Le Bourg’s hypothesis. Chickens would, as non-flighted birds,
potentially be unable to migrate in response to famine. Undergoing the CR effect would
thus potentially be useful. However, I am not aware of any studies of life extension under
CR in flighted birds.
Problems with the migration thesis
There are, however, other organisms that appear capable of migrating that do show a
response to CR. Fish, for example, were amongst the first organisms observed to exhibit
the life extending effects of CR (McCay et al 1929; Comfort 1963). However, it could be
argued that the fish in the cited experiments – the trout and the guppy – are less adapted to
emigration since they are freshwater species. It would be interesting to see if the effects of
CR are observed in fish that are more capable of escaping to better conditions.
In addition to this potential empirical problem, there are at least two theoretical problems
with Le Bourg’s hypothesis. First, it makes assumptions about the degree to which CR is,
or is not, an evolutionarily conserved response. If the effects of CR are conserved across
multiple lineages, then it is possible that the common ancestor was incapable of escaping
famine. If so, then by Le Bourg’s hypothesis, humans would have the same CR response as
rodents, in keeping with the transfer thesis.
The second problematic assumption is that early humans were capable of escaping a
famine. Le Bourg provides no information about the geographic reach of famines, or
48
humans’ ability to travel beyond that reach. It is far from obvious that humans could
escape famine at will. These problematic assumptions give Le Bourg’s evolutionary
hypothesis a rather ad hoc feel and diminish its status as a challenge to the transfer thesis.
3.2.3. CR as domestication artefact
It has also been suggested that effects of CR in animals are a result of domestication and
breeding adaptations to laboratory conditions, rather than a result of natural evolution (Le
Bourg, 2010). It is claimed that applying CR simply undermines these adaptive effects,
returning the organism to a wilder state with lower food intake, leaner bodies, more
activity and thus greater health. The view, which I refer to as the domestication objection,
is that CR works not by slowing ageing, but by undermining faster ageing that has been
bred in through years of adaptation to unhealthier conditions.
Again this makes sense within the evolutionary paradigm. It is thought that one common
response to surplus food availability is to divert food from maintenance to reproduction.
This may have happened in domesticated animals as they adapted to living with humans.
For a variety of reasons, laboratory animals may have been selected for increased fecundity
and appetite (Speakman and Mitchell 2011).
Limitations of the domestication objection
However, it is not clear that this is relevant to the transfer thesis. Humans in many parts of
the world have massively increased in numbers. This is due in part to the ability to avoid
food shortages. It has also led to decreases in physical activity in many parts of the world,
particularly in recent centuries. Compared to humans like the South African Khoisan and
the Australian aborigines, modern civilisations might be said to be more highly adapted to
conditions of plenty and less exercise.
49
Thus, adaptation to living in more plentiful circumstances is not necessarily a disanalogy
between rodents and modern humans. Even if CR is a laboratory artefact there’s no reason
to think we would experience less of a reduction in the rate of ageing than domesticated
organisms.
3.2.4. CR in the wild
Laboratory animals are kept in conditions that are different to those experienced by free-
living humans. As a result, some physiological features that would be detrimental ‘in the
wild’ might not impact on survival in cages. It is possible that these deleterious effects
might cancel out any beneficial effects of CR in terms of longevity, so we should not
regard CR effects as transferable.
For example, one effect of CR is that it results in lower bone mass. (Speakman and
Hambly 2007). Animals are kept in confined areas with little potential for accidents or
damaging activity, so lower bone mass is not detrimental. Humans, on the other hand, are
constantly at risk of damage due to injuries and accidents such as falls.
A further difference is that animals are kept in environments that are free from pathogens.
As such, if CR had a negative impact on immune systems these would not necessarily be
recognised. By contrast, humans are constantly exposed to pathogens that would threaten
our survival if immune functions were compromised. The condition of living in the wild
might counteract the survival benefits of CR.
Studies on bone quality and immune function
These threats to health may compromise individual welfare, and so are discussed in
Chapter 4 on the implications of CR and CRMs for the substantive good of health. There I
point to studies on rodents and non-human primates suggest that bone quality and immune
50
functioning are actually improved under CR. Thus there is no strong reason to think that
living in a relatively pathogen-ridden environment would counteract gains in average or
maximum lifespan.
3.2.5. Absolute, not proportionate lifespan increase
The previous arguments cast doubt on CR increasing lifespan by pointing to differences
between the environments, evolutionary needs and adaptations of model organisms and
humans. By contrast, De Grey suggests that a crucial environmental similarity undermines
the likelihood that CR will substantially extend lifespan in humans (De Grey 2005).
In particular, De Grey claims that the duration of famines is the same for all species. The
adaptation of deferred reproduction and slowed ageing would only need to last as long as
this duration. Thus we should expect all organisms to gain the same absolute number of
years. Any addition to human lifespan should be expected to be the same as, and not
proportionate to, that experienced in mice and other organisms. This would be enough to
secure survival of a famine in order to reproduce.
On the basis of this hypothesis, De Grey holds that the large increases in lifespan reported
in some rats will not be replicated in humans. Instead, humans should expect only the same
absolute number of years as that gained by other organisms. This would only be a
moderate maximum lifespan increase of at most two or three years.
De Grey suggests that empirical studies bear out his claims. He cites studies of six
organisms that he suggests provide empirical evidence for his hypothesis. Comparison
51
between nematodes, fruit flies, grasshoppers, mice, dogs and cows all appear to bear out
his thesis.12
Criticisms of De Grey
Both the evolutionary reasoning in de Grey’s argument, and the empirical support cited for
his hypothesis has been questioned (Rae 2006). With respect to the evolutionary claims, it
is far from clear that famine would not occur more than once in the reproductive life of
more long-lived organisms. If it did, slowed ageing might be beneficial more than once
and, by de Grey’s hypothesis, long-lived organisms might gain a greater reproductive
advantage by slowing ageing on more than one occasion. This would increase lifespan
more than the duration of a single famine.
Moreover, many evolutionary adaptations can be triggered and maintained in the absence
of the environmental factor to which they evolved as a response. For instance, metformin,
a potential CRM, works in part by triggering a natural response to excessive glucose. Thus
it is possible that a biological mechanism triggered by famine would continue to work in
the absence of an actual famine. There’s no clear reason that these mechanisms would
cease to work after a nominal amount of time.
In addition to these theoretical difficulties, the use of empirical studies cited in De Grey’s
paper has been heavily criticised on several grounds (Speakman 2011; Rae 2006). There
are two central problems. First, the organisms were not restricted to the same degree. As
mentioned earlier, the degree of life extension is highly dependent on the extent to which
organisms are calorically restricted. If they are not restricted to the same degree then
comparisons between extents of life extension are unlikely to be informative. If one
12 Note that the nematode studies he cites do not achieve close to the same absolute lifespan. He argues that
this may be because the maximum lifespan in nematodes has not yet been achieved.
52
organism is restricted by 30% and another is restricted by 60%, then of course the transfer
thesis will not predict the same proportional increase in both. The comparisons in De
Grey’s study are thus inappropriate tests of whether lifespan increases will be
proportionate or absolute.
The second, and related problem is that there is little reason to think that the degree of
lifespan extension exhibited in the chosen studies is the maximum lifespan the organisms
could achieve under CR. Again, this is because the degrees of CR in each study are very
different. The mouse study cited is of a 53% calorie reduction, while the study in dogs is
only 25%. The response in cows was to a 60% reduction in calories, but interspersed with
ad libitum feeding. Given that the degree of life extension is sensitive to the consistency
and degree of CR, these studies provide little reason to think that the maximum lifespan
was achieved in the relevant organisms. The claim that these organisms attain the
maximum lifespan increase achievable by any organism is thus highly doubtful.
In contrast, comparative studies of species using the same degree of CR appear to directly
contradict De Grey’s claims. They appear instead to support the idea that
a given degree of CR imposed on an animal of a given species leads to a similar
extension of [lifespan] expressed as a proportion of the species maximum [lifespan]
(Rae 2006, 95)
That is, comparative studies appear to support the transferability of CR’s effects on
longevity between species tested.
3.2.6. Biological limits to lifespan
Carnes, Olshansky, and Grahn claim that there is evidence for ‘biological warrantee
periods’ that are upper limits of maximum and average lifespan (Carnes, Olshansky, and
Grahn 2003). Other animals seldom reach these limits because their survival is highly
53
dependent on extrinsic factors – predation, infectious and parasitic diseases, and the
availability of food.
Humans, however, have experienced radical increases in average lifespan across the 20th
century due to reductions in these extrinsic causes of mortality. Elimination of many of
these causes has revealed an underlying ‘intrinsic’ life expectancy that is determined by the
rate of ageing. Based on patterns of age-related mortality, the authors argue that the
biological limit for human lifespan has been reached in many parts of the world. We
should not, thus, expect life expectancy to increase much above 85 years of age.
A clarification: lifespan limits and the rate of ageing
It is important to clarify this claim. The authors’ contention that there are biological limits
to lifespan can be misunderstood as a denial of inability to increase lifespan through
medical and biological interventions. However, they explicitly and repeatedly deny this
implication (ibid.; Carnes & Olshansky 2007). Their claim is instead directed against
demographic models that make long-term predictions that simply extrapolate from the life
expectancy gains of the 20th century (eg, Vaupel and Gowen 1986). Doing so ignores
evidence about an existing intrinsic limit that they argue is likely to prevent, or at least
slow down gains in life expectancy.
However, Carnes and colleagues do not deny that the rate of intrinsic ageing can be
modulated. On the contrary, they observe that although ‘a repetition of the large and rapid
gains in life expectancy observed during the 20th century is extremely unlikely’, such
gains could be achieved if we had the ‘ability to slow the rate of aging’ (Carnes, Olshanky
& Grahn 2003, 43. My italics.)
54
Here they explicitly acknowledge that slowing ageing would increase biological lifespan.
As I indicated in Chapter 2, studies indicate that CR slows the rate ageing on a variety of
measures. Thus there is no reason to think that Carnes’ and colleagues ideas of a biological
limit rules out the possibility that CR, or a drug that mimicked it, would result in
considerable gains in average and maximum lifespan.
3.2.7 The status of doubts about the transfer thesis
CR is effective in increasing lifespan in a wide variety of organisms. However, the
proposals above could cast doubt on the idea that the effects of CR would be transferable
to humans. I have argued that these proposals are themselves subject to theoretical and
empirical criticisms, or to clarifications that diminish their efficacy as objections to the
transfer thesis.
This is not to make a strong claim that CR would have equivalent effects, or that the above
proposals are false. Other objections may arise, or the proposals discussed may be
vindicated by further studies. However, the difficulties with these challenges mean that the
transfer thesis remains plausible.
3.3 Human studies of CR
In the previous chapter I discussed the effects of CR in animal studies, and their
implications for slowed ageing and life extension. In this chapter I have examined claims
that cast doubt on the likelihood that these effects will be replicated in humans. I argued
that they fail to provide convincing evidence that the CR effect will not transfer to humans.
In this section I discuss more direct attempts to test the transfer thesis: studies of CR
humans.
55
First I discuss some difficulties with conducting such tests, before providing examples of
studies in humans. Though they are by no means conclusive in favour of the transfer thesis,
these studies indicate that many of the effects of CR on lifespan and the rate of ageing
occur in humans. In combination with the discussion of criticisms of the transfer thesis
discussed above, the studies below indicate that CR and interventions that mimic it are
worthy of the ethical investigation that follows in later chapters.
3.3.1. Obstacles to empirical studies in humans
As Holloszy and Fontana suggest, ‘the only way to be sure that CR ‘works’ in humans is to
conduct studies in people’ (Holloszy and Fontana 2007). However, there are obvious
practical and methodological difficulties in conducting sufficiently long-term experiments
on free-living humans.
The costs of lifelong experiments would be prohibitive. It would also be necessary to
control for genetic and environmental factors that affect ageing. All this would have to be
done whilst accounting for the range of free human activities that might affect the
experiment. Moreover, waiting a significant amount of time for results would significantly
delay evaluation of interventions with potentially profound social and ethical importance.
They would also be of no help to the thousands of people currently practising CR or
making use of candidate CRMs, or to those considering the use of these interventions.
It is thus unlikely that fully satisfactory long-term studies will ever be conducted in
humans. Nonetheless there have been attempts to conduct research on humans that that
does not require impractical lifelong trials. These make use of the measures of ageing
mentioned earlier: lifespan measures, rate of mortality measures, disease measures, and
biomarkers of ageing and longevity. Even though these methods are unlikely to lead to
56
certainty about veracity of the transfer thesis, they mean that we are able to improve our
guess about whether CR ‘works’ in humans.
3.3.2. Examples of studies in humans
Several studies in humans have been undertaken. Below I briefly describe the four most
commonly cited studies of calorically restricted humans: Biosphere Two volunteers,
members of the Caloric Restriction Society (CRS), participants in the CALERIE study, and
the Okinawan cohort.13 Thereafter, I summarise the results of these studies in terms of the
relevant measures of the rate of ageing, and the corresponding claims about CR. Thus far
results suggest that the transfer thesis is a viable hypothesis.
Biosphere Two
Biosphere Two was designed to conduct experiments in a closed ecosystem. In 1991 eight
volunteers were sealed in the biosphere for two years to study the complex interactions
between living processes. One unforeseen consequence of the experiment was that food
supplies in the biosphere dropped well below the expected level. As a result, the
participants were forced to restrict dietary intake (Walford et al. 2002). Studies of the
occupants thus give an indication of the immediate adaptations to a CR diet.
The Caloric Restriction Society
The Caloric Restriction Society (CRS) was founded in 1994 by Roy Walford, one of the
participants in the Biosphere Two project. Members of the CRS restrict calories for health
and longevity purposes. It has become common for practitioners of a CR diet to volunteer
for research, providing data about the short and long-term effects of caloric restriction.
13 These are discussed by among others, Holloszy and Fontana 2007, Redman and Ravussin 2011, and Roth
and Polotsky 2012.
57
CALERIE
The Comprehensive Assessment of Long-term Effects of Reducing Intake of Energy
(CALERIE) is the first randomised controlled trial of caloric restriction
(http://calerie.pbrc.edu/). The study examines the effects of 25% caloric restriction (75% of
weight maintenance requirements) in non-obese women and men between the ages of 25
and 45 years. The trial consists of about 150 calorically restricting participants and is
ongoing.
Okinawan studies
The Okinawans – inhabitants of the Japanese Okinawa Prefecture – have one of the highest
life expectancies in the world, as well as a disproportionate number of centenarians. The
Okinawa Centenarian Study (http://okicent.org) has investigated more than 600 Okinawans
in order to discover the reasons for their exceptional longevity. Significantly it was found
that a particularly long-lived cohort of Okinawans had a diet that contained substantially
fewer calories, leading to the suggestion that their increased longevity may be as a result of
CR (Willcox et al. 2007a). This means that Okinawans are perhaps the closest we will get
to a lifelong study of CR in humans.
The above studies provide a testing ground for the transfer thesis. Again, the transfer thesis
holds that effects observed in animals will be replicated in humans. If the transfer thesis is
true:
1. CR humans will have higher average and maximum lifespan relative to non-CR
populations, and will have a reduced age-relative risk of mortality.
2. Deaths in populations of CR humans will occur across a longer span of time than
those in non-CR humans.
58
3. In CR humans the onset of age-related diseases will occur later than in non-CR
groups, and at any age they will be less likely to have an age-related disease.
4. In CR humans, the appearance and development of other physiological and
behavioural traits associated with ageing will be delayed.
Below I discuss the extent to which these claims are justified by human studies.
3.3.3. Lifespan measures of the rate of ageing
As mentioned, it is difficult to conduct scientifically rigorous lifelong studies in humans.
However, studies of Okinawans provide the closest available surrogate. These provide
some reason to think that the first claim of the transfer thesis will be met.
Willcox and colleagues (2007a) report that the cohort of Okinawans had a high average
lifespan of 83.8 years, compared to 82.3 in the rest of Japan and 78.9 in the USA.
Similarly, the maximum lifespan of Okinawans was 104.9 years compared to 101.1 in the
rest of Japan and 101.3 in the USA. Significantly, Okinawans had higher average and
maximum lifespan relative to Japan, the world’s longest lived population.
59
Figure 2 Average and maximum lifespan in Okinawan, Japanese and US groups. (Willcox et al 2007a).
It is important to note that, in keeping with the maximum lifespan measures outlined
earlier, Willcox and colleagues did not take the oldest old person in these groups as the
maximum lifespan: the oldest old Okinawan person was 114 in 2009,14 while the oldest
American was Sarah Knauss (119 years) and the oldest Japanese person was Tane Ikai
(116 years).15 Instead the average of the top 1% of old people is taken as the maximum
lifespan.
Recall, however, that lifespan measures of the rate of ageing a top percentage as
identifying the maximum in order to cancel out some of the effects of sample size. As
mentioned earlier, there’s a higher chance that the oldest member of the larger sample will
be older than the oldest member in a smaller sample. Thus it should not be surprising that
14 http://www.japanupdate.com/?id=9793. Last accessed 19 December 2012. I was unable to provide
independent verification for this claim, or discover a recent update. This may be because, as the article
claims, the super-centenarian prefers not to be named.
15 http://en.wikipedia.org/wiki/Supercentenarian. Accessed 19 December 2012.
60
the oldest member of a relatively small group – the Okinawans – is younger than the oldest
member of much larger groups – Japan and America.
Limitations
A further factor to bear in mind is that studies of Okinawans don’t account for many of the
limitations to human studies mentioned earlier. In particular, it is difficult to exclude the
range of genetic and environmental factors other than CR that may have contributed to
their impressive longevity. Without this, it is not certain that their diet was the central
contributor to increased lifespan. However, Willcox and colleagues note that since the
westernisation (increased meat, fat and bread) and Japanisation (more polished white rice)
of the Okinawan diet, the ‘Okinawan mortality advantage has all but disappeared’ (Willcox
et al. 2007a, 436). This suggests that calorie intake plays a major role in the increased
longevity of older Okinawan groups.
It is also apparent that the increase in lifespan achieved by Okinawans is nowhere near the
60% that has been achieved in animal studies. However, it should be recognised that the
degree of CR practiced by Okinawans was estimated to be about 10.9% fewer calories than
is recommended to sustain body weight (Willcox et al. 2007a). This is much less than the
60% CR that resulted in drastic life extension in rodents. Moreover the Okinawans studied
are only believed to have been under CR for approximately half their adult lives. It may be
that reductions greater than this and/or for a greater proportion of life, or CRMs that
mimicked such reductions, would result in a higher average and maximum lifespan than
that achieved by the Okinawans.
The Okinawans are the closest that we have to a life-long CR study on humans. As a result
they are the only humans on which it is possible to conduct studies that make use of
lifespan measures of ageing. Use of these measures imply that CR in humans slows ageing
61
in the first respect above: Okinawans have a higher average and maximum lifespan and
reduced age-relative risk of death. Thus available evidence suggests the first aspect of the
transfer thesis is borne out.
3.3.4. Rate of mortality measures
Rate of mortality measures equate the rate of ageing with the rate at which organisms die
after reaching a tipping point of increased deaths. Again the application of these metrics
would require a life-long study in a group such as the Okinawans. Unfortunately, however,
I have been unable to find a study that applies rate of mortality measures to the Okinawans.
Nonetheless, the results displayed in figure 2 suggest that Okinawans have a slower rate of
ageing relative to mainland Japanese. This is suggested by the widening gap between the
two plots that begins between the ages 65 and 75. However, to the naked eye, the gap
between the Okinawan and USA plots does not widen, perhaps implying delayed, rather
than decelerated ageing. Interpreting the rate of mortality is thus inconclusive. Indeed,
attempting to estimate rate of mortality on the basis of the above lifespan plots may be
methodologically unsound. Until a rate of mortality analysis is available in humans, it is
necessary to make recourse to studies in animals that indicate slowed ageing.
3.3.5. Disease measures
Disease measures of ageing record the incidences of age-related disease. Recall that in
rodents and rhesus monkeys it was found that CR reduced the incidence of a wide range of
age-related diseases such as cancer, cardiovascular disease, and diabetes relative to
controls.
Human studies are in keeping with these findings. During the period of CR, calorically
restricted occupants of Biosphere Two exhibited improvements in a number of risk factors
62
for age-related disease. They showed, amongst other beneficial changes, improved levels
of cholesterol, blood pressure and low density lipo-proteins, all of which are indicators of
the likelihood of cardiovascular disease. They also showed levels of glucose and insulin
that indicated a lower risk of diabetes (Walford, 2002).
Similar changes were also evidenced in members of the CRS (Meyer et al 2006), leading
Holloszy and Fontana and Omodei and Fontana to contend that they have a decreased risk
of atherosclerosis and type 2 diabetes (Holloszy and Fontana 2007; Omodei and Fontana
2011). Similar observations have been made in the CALERIE study. Given acknowledged
indicators of age-related diseases, CRS members and subjects in CALERIE appear to have
a lower risk of age-related health problems (Stein et al 2012).
Significantly, the CALERIE study also found that CR reduces oxidative damage in
humans. Oxidative damage has been implicated as a major cause of ageing and age-related
disease. It is, for instance, thought to contribute to cancer, heart failure, atherosclerosis,
Parkinson’s disease, and Alzheimer’s disease (Redman and Ravussin 2011).
Although the findings indicate physiological changes that point strongly to the conclusion
that CR would reduce the incidence of disease, none of these studies is life-long. This
means that there are no advanced cases of disease in either the control group or the CR
group. Despite these positive indications, therefore, it has not been shown conclusively that
CR delays the onset of age-related diseases, or reduces the age-relative risk of age-relative
diseases in these groups.
The Okinawans again provide the best evidence concerning disease measures of ageing in
humans. Willcox and colleagues report that, in keeping with animal studies,
63
[c]oronary heart disease, and forms of cancer, such as lymphoma, and cancer of the
prostate, breast, and colon are remarkably low in age-matched Okinawans versus
other Japanese and Americans. (Willcox et al. 2007a, 445)
Figure 3 Age-related disease in Okinawans, Japanese and U.S groups. (Willcox et al 2007a).
This means that at any age Okinawans display a lower risk of dying from an age-related
disease than Japanese or Americans (see figure 3 above). Moreover, in Okinawans the
onset of the period in which age-related disease is more likely is delayed.
Shorter term studies of CR strongly suggest that adaptations occur that will delay the onset
of age-related diseases and reduce their likelihood at any age. Studies on Okinawans
provide further evidence for this claim. Thus there is a growing body of evidence that the
third prediction of the transfer thesis will be fulfilled: in CR humans, as in CR animals, the
onset of age-related diseases will occur later than in non-CR groups, and at any age they
will be less likely to have an age-related disease.
64
3.3.6. Biomarkers of ageing and longevity
The fourth way to measure the rate of ageing is to determine physiological traits that
deteriorate progressively with age. Again, there are very few, if any, completely reliable
inter-species biomarkers. As mentioned earlier, indicators of disease risk are sometimes
regarded as biomarkers of ageing. To the extent that they are, the disease measures
discussed above indicate that ageing is slowed in humans, replicating findings in animals.
Several other purported biomarkers of ageing have been studied in the human CR groups
discussed. These include body temperature, levels of DHEA and DHEAS, physical
activity, grip strength and appearance. As discussed below, these tend to indicate slowed
ageing on biomarker measures.
Reduced body temperature is a feature that occurs in CR animals and long-lived humans
and so may play a significant role in slowed ageing, or provide an indication that similar
effects are occurring.16 In keeping with this indicator, subjects in the CALERIE
experiments exhibited the reduced body temperature seen in animals and long-lived
humans (Redman and Ravussin 2011).
Levels of DHEA decline progressively with age in rhesus monkeys and also in humans. As
a result DHEA is regarded as a candidate biomarker (Lane et al.1997; Roth et al. 2002). In
CALERIE participants, no change was observed in levels of DHEA after 6 months under
CR, perhaps due to the short duration of the trial. I was unable to find studies about DHEA
levels in CRS members, many of whom would have been restricting calories for
substantially longer. In the Okinawan cohort, however, DHEA levels were observed to be
substantially higher than age-matched controls, suggesting that CR slows the rate of ageing
16 Given the earlier definition of a biomarker as a physiological trait that deteriorates with age, it is not clear
that body temperature should be included in this category. Body temperature does not change predictably
with age, so this may disqualify it as a biomarker of ageing on the strict definition discussed in S2.4.
65
as measured by DHEA. Subsequent publications from the CALERIE study, as well as
investigation of DHEA levels in CRS members could provide more information about
whether CR slows ageing on the basis of this candidate biomarker.
Other biomarkers provide further support for the idea that CR slows ageing in humans.
Okinawans appear physically younger, are more active, and have stronger grip strength
than age-matched controls (Willcox et al 2007b). To the extent that available biomarkers
can be regarded accurate measures of the rate of ageing, and to the extent that they have
been made use of in studies of CR humans, they tend to provide evidence for the fourth
aspect of the transfer thesis: in CR humans, the appearance and development of other
physiological and behavioural traits associated with ageing are delayed.
3.4 Conclusion
In the above I have attempted to make apparent the limitations of existing research of CR
humans. To the extent that studies are available, however, human studies appear to bear
out the transfer thesis. Lifespan measures, rate of mortality measures, disease measures,
and biomarker measures of the rate of ageing suggest many of the effects observed in
animals may be replicated.
66
CONCLUSION TO PART I
It is impossible, on the basis of the available studies, to be fully confident that CR and
CRMs would result in the substantial lifespan gains detailed in S3.1. However, it is worth
noting that this lack of certainty in advance is likely with any genuine life extending
intervention. Without methodologically rigorous, extremely long-term tests it is impossible
to know that an intervention will extend average and maximum lifespan.
In the absence of such long-term evidence we must make decisions on the evidence that is
available. The above research on CR provides a strong case for the claim that the effects of
CR and drugs that mimicked it would translate to humans, extending lifespan and slowing
ageing. This premise, and in particular the four central claims of the transfer thesis, provide
the empirical basis for the ethical analysis in chapters to follow.
67
PART II: CR, CRMS AND INDIVIDUAL
WELFARE
INTRODUCTION TO PART II 17
In the previous chapter, I discussed the transfer thesis the hypothesis that the effects of
CR and CRMs would be replicated in humans. In this chapter I evaluate whether, assuming
the truth of the transfer thesis, extending lifespan by using CR or CRMs would, other
things equal, be good or bad for a person.
This question can be posed using different terms: would the intervention be in or against a
person’s interest? Would it harm or benefit her? Would it increase or decrease the value of
her life, or make her life go better or worse? Would it be prudent for her to use the
intervention, or not? Would it have positive or negative impact on her well-being or
welfare? It may be possible to distinguish these questions by appealing to differences in
everyday uses between terms such as ‘welfare’ and ‘well-being,’ though I have doubts that
doing so is useful. Instead, I will treat these questions as equivalent.
i) The significance of individual welfare
Theories of individual welfare provide accounts of what makes a person’s life go well and
badly for them. If smoking is bad for me, or harms me, theories of welfare should be able
to designate the features of smoking that make my life go worse. Similarly, if exercise is
good for me, or benefits me, a theory of welfare should be able to account for what it is
about exercise that makes my life go better.
17 The arguments of this Part have benefited from useful discussion at the 2009 Philosophical Society of
South Africa (PSSA) Conference, University of Fort Hare, South Africa, the 2012 Well-being in
Contemporary Society (WICS) 2012’ conference in Twente, Netherlands, as well as the ‘2012 Symposium
on Enhancing Human Experience via Emerging Technologies,’ in Laval, France. Sections of this Part are
forthcoming in Philosophical Papers and the International Journal of Design and Innovation Research.
Thanks to the reviewers for their comments and suggestions.
68
Individual well-being is significant for prudential and moral individual choices, as well as
for policy choices. Prudential choices are guided by considerations about whether a
particular action will increase or lower one’s welfare – whether it would be prudent or wise
to make a certain choice. Amongst other things, moral choices may require that in pursuing
her own welfare a person’s choices don’t undermine the welfare of others. Some moral
systems suggest that we are also obligated to improve the welfare of others.
In biomedical policy, beneficence and non-maleficence are regarded as key virtues of
political institutions.18 Other things being equal, institutions should aim to benefit people,
increasing the welfare of citizens. Moreover, it is thought that political institutions have an
even stronger duty of non-maleficence: to avoid harming, or decreasing the welfare, of
individuals. Ethical policy choices may also require that there is a fair distribution of
welfare, or opportunities for welfare across society and between individuals.
Well-being may thus play a significant role in ethical individual and policy decisions. As a
result, a useful analysis of life extension by CR or CRMs should take into account their
effects on well-being.
ii) What makes a life go better or worse?
In this Part, I examine the likely impact of CR and CRMs on the basis of ethical ideas
about what values make someone’s life go better or worse for her. These accounts of
welfare are typically divided into substantive good, desire fulfilment, and mental state
accounts.19
18 See, for example, Beauchamps and Childress 2001.
19 For example, see Parfit 1984, Griffin 1989, Sumner 1996, Scanlon 1998, Crisp 2006, Keller 2004.
Substantive good accounts are often referred to as ‘objective list’ accounts. Here I prefer Scanlon’s term,
since the term ‘list’ may invoke an arbitrary string of goods.
69
Substantive good accounts delineate objectively valuable factors required in order for a
person’s life to go best. For instance, autonomy, self-development and friendship are often
considered to be prerequisites for a high degree of well-being.20 These goods are seen as
intrinsically valuable. They increase welfare whether or not they are wanted by the person
and whether or not they have any other good consequences.
Desire satisfactionist accounts hold that a person’s welfare depends on whether her desires
are fulfilled.21 If a person desires that her lover is faithful to her, and the lover is not, then
she is harmed and her welfare is lowered. On desire satisfactionist accounts, this is so even
if she gains in other (un- or less desired) respects, such as having more freedom, or
pleasure. The only thing of value for a person is that her desires are satisfied.
Mental state theories hold that a person’s life goes well or badly to the extent that she has
good or bad mental states. On these theories, only certain experiences have intrinsic value.
For instance, hedonist mental state theories hold that a person’s life goes better if she has
pleasurable experiences and worse if she has painful experiences.22
All the above theories have been subjected to criticism. My intention here is not to prefer
any particular theory of welfare. Instead I use the theories as a structural device to consider
the possible impact of CR-related life extension on individual welfare. That is, the
accounts above provide a way to structure arguments and situate objections to life
extension in terms of prudential values.
20 See, for instance, Nussbaum 2000.
21 For example, Bernard Williams 1973.
22 Examples of hedonist theories are those of Feldman 1991, and Bradley 2004.
70
Moreover, considering a variety of claims about welfare is in keeping with the
compatibilist methodology outlined in the introduction. Since I address objections on the
basis of several accounts of welfare, it is possible that broader agreement may be achieved.
Note that there is obviously some overlap between the prescriptions of these accounts. For
instance, no plausible account of value for a person would deny that a degree of happiness
is important. They will, however, differ on the justification. Some hedonist accounts, such
as those of Mill, Bentham, and Epicurus hold that happiness is the only thing that has
intrinsic value. Substantive goods theories may hold that happiness is one good among
many, or that it is instrumentally valuable in that it contributes to some other intrinsic
good. Desire satisfaction theories will hold that happiness is good because people want
happiness. Thus, although the accounts of individual benefit can be distinguished, they will
often overlap in their recommendations.
iii) Distributions of welfare within a life
The theories of welfare above designate goods and bads that make a person’s life go better
or worse. However, the value of a life may depend in part on the distribution of these
goods within a life (Benatar 2006, 61-64). In particular, individual welfare may involve an
interplay between the total, or cumulative goods within a life, as well as the order in which
goods occur. I will explain in more detail.
The simplest view of welfare is that a life goes better or worse to the extent that it contains
more or less of a particular good. If pleasure is a good, hedonists claim, then a life is better
if it contains more pleasure and worse if it contains less. All that matters is the total good
accumulated in a life.23
23 For such a view see Bradley 2004.
71
In addition to cumulative good, David Velleman has argued that the order in which these
goods are achieved, and the structure of a life, also play a role in evaluations of a life
(Velleman 1991). For instance, it is usually thought that, other things being equal, a life
that goes from worse to better is an improvement over a life that goes from better to worse.
Suppose we have to choose between a life of declining happiness and a life of increasing
happiness. Velleman holds that if the total good in these lives is the same, most people
would choose the second option. This suggests a further contributor to the value of life. In
addition to the total good achieved, the structure of the life, and the order in which goods
are achieved may be relevant.
iv) Welfare and comparison
In asking whether CRMs make a life go better or worse, it is important to specify what the
life is better or worse than. That is, it is necessary to indicate the basis or bases of
comparison. We think that smoking is bad for a person because it leads to conditions like
cancer and cardiovascular disease. If a person smokes she is more likely to have one of
these conditions.
Smoking may thus be regarded as bad because it leads from a better to a worse state. From
the state of being healthy, one becomes unhealthy. In this case smoking can be said to be
bad on the basis of a comparison between a better and a worse state.
Life-death comparisons
This type of comparison is difficult to make in the case of life extending interventions.
Assuming that people cease to exist when they die, life extension postpones going from a
state of life to a state of non-existence. However, there appears to be a conceptual problem
in saying that life is better for a person than non-existence. Silverstein has argued that it is
incoherent to assign any value, even zero to times when a person does not exist (Silverstein
72
1980 & 2000). Since there is no person at such times there can be no value for her.24 This
type of life-death comparison may thus be inappropriate to make. As a result, it is
surprisingly difficult to argue that life extension is good for you because it postpones going
from a good state (life) to a bad state (death).
Life-life comparisons
In order to avoid this type of life-death comparison, it has become common to invoke a
conceptually coherent life-life comparison when considering whether an event is harmful
or beneficial.25 On this view an event is good for you if it makes your life better than it
would have been had that event not occurred. Ben Bradley formalises this idea as a general
view about harms as follows:
[t]he overall value for a person x of an actually occurring or obtaining event or state
e = the value of x’s actual life minus what the value of x’s life would have been had
e not occurred or obtained. (Bradley 2007, 115)26
On this view, then, taking a CRM is good for a person if it makes her life better than it
would have been if she had not taken the CRM. Doing so is bad for her if it makes her life
worse than it would have been had she not taken the CRM. Suppose, for instance, I take a
CRM and, as a result, my life has a value of 10. If I did not take the CRM my life would
have had a value of 13. The value of taking a CRM is -3: it is bad for me, since I would
have been better off if I had not taken it. On this view something can be bad for me even if
it does not result in an absolutely bad state. What matters is that I am comparatively worse
off than I would have been.
24 See my ‘Deprivation and the See-saw of Death’ 2009 for a detailed discussion of this problem. See also
Parfit 1984, 175.
25 See Nagel 1979, McMahan 2002, Feldman 1991, Bradley 2004.
26 See McMahan 2002, and Broome 2004, 11 for variants of this claim.
73
Life-life comparisons and other judgments about welfare
This way of conceiving of harm and benefit avoids the problematic life-death comparison
that results from holding that harm involves going from a better to a worse state. However,
it also makes sense of some of our other intuitive judgements about benefit and harm, good
and bad. For instance, say Maria has a job that she likes, but she applies for another
position that would make her better off. Unbeknownst to her, Paolo, a competitor for the
position, tells a lie about Maria and, as a result, she does not get the job. Maria continues in
her position, at the same level of well-being she was before. Yet it is justified to think she
was harmed by Paolo, despite the fact that her level of welfare has not changed. She is
harmed because she is worse off than she would have been if Paolo had not lied.
The life-life comparison is effective in a range of cases. It is thus appropriate to use it to
determine the prudential value of CR and CRMs. On life-life comparisons, objections to
life extension may attempt to show that prolonging one’s life would a) add no value to a
person’s life, so that the intervention would not be worthwhile, or b) would make a
person’s life worse than it would have been had it not been prolonged.
This latter (b) can occur in two ways: first, a person’s life may be made worse because
CRMs result in an absolutely bad condition that subtracts from the actual value of one’s
life. This might be the case if, for instance, a CRM resulted in pain so severe that life was
not worth living. Alternatively life extension can be bad for me, even if an extended life is
good: if taking a CRM adds less good to life, and so is worse, than not taking the
intervention, making use of a CRM is harmful.
v) Structure of arguments
Having delineated relevant axiological conditions for evaluating interventions in terms of
their prudential value, I proceed as follows: In each chapter that follows, I outline the
74
intuitive positive argument for extending lifespan in terms of a relevant value. Thereafter I
discuss general objections to life extension that have been made on the basis of each value
or theory of individual welfare. I determine the extent to which these objections are
applicable to CRMs. Where they apply, I assess how damaging they are to the case for life
extension by CR and CRMs. Finally, I discuss ethical concerns that have been raised about
this particular type of intervention, and which are unlikely to have been considered in
relation to other life-extending interventions. The overall claim of this part is that, despite
some reservations, CRMs are likely to be beneficial.
75
4. SUBSTANTIVE GOODS
As mentioned above, substantive good accounts designate goods whose presence or
absence make a person’s life go better or worse. Some examples include ‘moral goodness,
rational activity, the development of one’s abilities, having children and being a good
parent, knowledge, and the awareness of true beauty’ (Parfit 1984, 499). Martha Nussbaum
has a list of central ‘capabilities,’ which are necessary conditions for a good life. These
include, inter alia, life, bodily health, affiliation, and play (Nussbaum 2000).27
There is obviously significant dispute about what goods should belong on the list. In this
chapter, I restrict the discussion to those substantive goods that have been discussed in
relation to life extension: health, procreation, self-development and flourishing, and the
value of community.28 I argue that, contra the objections, CR and CRMs may improve
lives on substantive good accounts.
4.1 Health
Health is often regarded as having special value. As early as 380 BC Plato argued that
health is desirable in itself and not merely for its consequences (Penner 2003, 312). More
recently, Daniels has linked the value of health to its impact on opportunity. He writes:
[d]isease and disability, by impairing normal functioning, restrict the range of
opportunities open to individuals. Health care thus makes a distinct but limited
contribution to the protection of equality of opportunity. (Daniels 2001, 2)
However one justifies the value of health, it clearly has key significance to individual
welfare. A reasonable standard of physical health is a precondition for having a good life.
27 See also Nussbaum and Sen 1993.
28 Note that pleasurable mental states and the satisfaction of desires would form part of most plausible
accounts of substantive goods. Thus for substantive good theorists, the sections that follow may be seen as
subsets of objective list theory.
76
Partly for this reason, many concerns about life extension centre on potential negative side-
effects that may impact on health related well-being.
Assuming a person had a sufficiently high standard of health, an intervention that
postponed death could increase the value of life. By delaying death, a life extending
technology (LET) could increase the amount of time spent in a condition that adds to the
health-related value of a life. If so, a person’s life would have a greater value if she took
the intervention than if she did not. On a life-life comparison, the use of the LET would be
good for a person.
The objections that follow challenge this idea. The Struldbrug objection makes the point
that there are some age-related health conditions that would detract from the value of life.
Life extension technologies may prolong these conditions, making an extended life worse
than a normal one. I argue that, although this may be true, it is not the case with CR and
CRMs.
4.1.1 The Struldbrug objection
In a study of community attitudes to life extension, Partridge et al established that one of
the major concerns people have about life extension is that gains in length of life come at
the expense of losses in quality of life in old age (Partridge et al. 2009). Individuals are
concerned that new life extending technologies will have what I will call the Struldbrug
effect of prolonging and worsening age-related health problems.
The Struldbrugs in Jonathan Swift’s Gulliver’s Travels are an often used example of
unhealthy life extension. Swift’s Gulliver writes of the Struldbrugs:
77
Besides the usual Deformities in extreme old age, they acquired an additional
Ghastliness in Proportion to their Number of Years, which is not to be described.
(Swift 1826, 89)
These pitiful characters age normally, or at least consistently, but continue to decline
physically long after the normal’ life span. As a result they are tragic and twisted beings
with a life of ever-decreasing welfare. This kind of prolonged physical decline is, arguably,
increasingly evident in the developed world where lifespan has increased due to medical
advances. This has contributed to a rising number of cases of age-associated disorders such
as Alzheimer’s disease, diabetes, and cancer (Olshansky et al. 2006). In general, getting
older is associated with worsening health. Moreover, interventions such as life support
machines can lengthen lifespan without increasing health-related quality of life, resulting
in lives that many regard as not worth living.
If life extension extends and exacerbates age-related health decline, the contribution it
makes to welfare will at some point cease to be positive and the intervention would detract
from the health-related value of a person’s life. The Struldbrug problem thus captures an
important fear that people have about life extension: health decline may make old age so
bad that having the additional life is not worthwhile.
A distinction: decreased health minimum and prolonged health decline
To deal with the above fears, it is necessary to distinguish two separate concerns that are
present in the case of the Struldbrugs. The first is that an extended lifespan will result in a
decreased minimum of health. The second concern is prolonged health decline that life
extension would stretch out the period of declining health. In the case of the Struldbrugs
both concerns are in evidence. The decline in health continues below a minimum, or
tolerable level. This decline is also prolonged; in the case of the Struldbrugs it is extended
indefinitely without ever ‘bottoming out’ or being cut short by death.
78
In response to the Struldbrug objection to life extension, I argue below that CR and CRMs
would not result in a decreased health minimum. However, it does appear likely to prolong
a period of worse health by decelerating ageing. Although a decreased health minimum
would be harmful for individuals, I claim that the prolonged health decline that would
result from CRMs may add to the overall health-related value of life. As a result, I argue,
the Struldbrug effect is not something that should worry us in the case of CRMs.
Decreased health minimum
As discussed in Chapter 3, several studies of the health effects of CR and CRMs on
humans have been and are currently being conducted. The CALERIE project is one of the
most recent of these. The study involves the observation of calorically restricting humans,
the measurement of health status, and assessment of the biomarkers of ageing.
Thus far the results of the CALERIE study suggest that, far from resulting in a Struldbrug-
like scenario, caloric restriction improves the health of subjects, in keeping with rodent and
primate studies (Redman and Ravussin 2011). Similar results have been found in a recent
clinical trial of a resveratrol-based CRM compound (Kennedy et al. 2010). These shorter
term results are in keeping with the transfer thesis outlined in Chapter 3. If the transfer
thesis holds true in the longer term, age-related diseases will not be suffered to a worse
degree. Instead, since ageing is slowed age-related diseases will be deferred to a later point
in life.
Thus, studies on CR and CRMs strongly suggest that life extension by CR derived methods
will not to result in the decreased minimum health scenario exemplified by the Struldbrugs.
Other healthcare tools such as life support machines may extend life below a minimally
tolerable level of health. Existing studies suggest that CRMs will not.
79
This should go some way to alleviating the fear that about life extension by CRMs would
worsen health in advanced old age. However, the concern about prolonged health decline
remains. Although CR and CRMs won’t result in a lower minimum of health related
welfare, they might result in the decline becoming extended across a longer period. That is,
although health will not decrease below normal levels as a result of CR, the period in
which health decline occurs might be longer.
Prolonged health decline
Would life extension by CR and CRMs result in prolonged health decline? Some advocates
of anti-ageing research claim that anti-ageing technologies may instead result in the
compression of morbidity (Olshansky et al. 2006, 35). That is, they suggest that rather than
extending the existing period of age-related decline, CR shortens it. Lifespan might be
extended through a prolongation of what has been termed ‘healthspan’ – the period of
relative health and vitality.
As discussed in the previous chapter, research on animal models points to slowed ageing.
As a result, CRMs would most likely delay the onset of diseases related to ageing,
extending the period of relatively good health before old age. However, slowed ageing also
means that the period of decreased health, while postponed, could also be extended. In
contrast to the compressed morbidity model, a time of decreased health is not shortened,
only put off and perhaps lengthened. If CRMs result in decelerated ageing, as appears
likely, there will indeed be a longer period in which age-related diseases will occur.
Compressed morbidity thus does not seem likely to be a consequence of CR. This may
strike some as unfortunate, since many people would like to avoid a period of illness at the
end of life as much as possible. In a moment I will argue that prolonged morbidity is not,
80
in fact, a bad thing in terms of individual health. However, for those who regard the
compression of morbidity as a goal it may be interesting to point out that there is some
evidence that at the level of population, a reduction of late life illness appears to be
occurring in the absence of CR (Fries 1983). Compressed morbidity appears unlikely to be
achieved by CR itself. However other interventions and lifestyle changes may enable its
achievement with or without the intervention at issue in this thesis.
While studies suggest that CRMs won’t result in decreased minimum health, it appears
they would result in a prolonged period of health decline. This is an aspect that tends to be
given little attention in academic literature on CR and CRMs, and it is important for
individuals that make use of existing and future CRMs to take this into account.
Benefits of prolonged health decline
Many potential users may be deterred from using a CRM by the thought of a prolonged
health decline. However, prolonging health decline isn’t necessarily harmful. Indeed,
interventions that decelerate ageing and lead to prolonged health decline could contribute
beneficially to the overall health-related value of life. This is so for three reasons,
discussed below.
First, decelerating ageing means that healthy period before old age would also be extended.
One still has more healthy life before health starts to decline. Even if a lengthened period
of health decline was regarded as undesirable, some might regard it as a worthy trade-off
for an earlier increase in the number of healthy years. The second and third points below
go further than this and suggest that this extended period of worse health is actually
desirable if it is achieved by CRMs. In this case, the trade-off need not take place.
81
The second reason why prolonged health decline would be beneficial is that studies on CR
suggest that the number of diseases suffered will not increase (Gems 2011, 110). Although
the length of the time in which a person is more likely to get sick is longer, at any point
during extended old age she is less likely to have a particular age-related disease. So
despite the fact that one would be more susceptible to age-related diseases for longer, the
frequency of diseases appears to be lower. In this sense prolonged health decline could be
regarded as an improvement over normal ageing.
The third reason why prolonging health decline could be a desirable effect of using CRMs
is that years spent in imperfect health can benefit one. Even if older people experience a
longer period of increased susceptibility to disease this doesn’t mean that their lives are
bad for them or are getting worse on the basis of the value of health. The extra years are
simply not as good as they could be. As Christine Overall suggests,
long-living people who do have an illness or disability are not thereby prevented
from leading rich, full lives; to the extent that they are dependent, their dependence
should not be interpreted as evidence that increased longevity is bad. (Overall 2003,
188)
Years spent with a disease don’t subtract from the overall health value of a time, unless
they have a negative health value.29 It simply means that less of this good is added to that
life.
Another way of putting this third point is that older people generally have good lives,
despite declining health and vigour. If a person’s life is worth living through most of her
old age, an extended old age would still add to the overall or cumulative goodness of her
29 Note that some measures of health-related quality of life, like the quality adjusted life year (QALY) used
by the UK’s National Institute for Clinical Excellent (NICE), controversially assign a health value of zero to
death. This means that only years spent in a state that is worse than death would detract from the value of
life. Years spent in a state better than death would increase the health-related value of life. See Phillips and
Thompson 1998.
82
life, even it would add less good than healthier years would. In comparison to a shorter life
without CRMs, then, using the intervention would be beneficial.
These three points provide grounds to think that, far from reducing the quality of a
person’s life, prolonging the period of health decline might increase it. The possibility that
CRMs would result in prolonged health decline is no objection to extending life by these
means.
4.1.2 Particular effects of CR on health-related welfare
Above I argued that life extension by CRMs will not result in a Struldbrug situation of
indefinitely declining welfare, and that extension of the period of health should not be seen
as reducing health related welfare. However, while the slowed ageing that results from CR
would not detract from individual health, CRMs may have particular effects with a more
problematic impact on well-being. For instance Dirks and Leeuwenburgh caution that CR
may result in immune system problems, loss of strength and stamina, decreased blood
pressure, bone thinning, osteoporosis, and reduced cognitive performance (Dirks and
Leeuwenburgh 2006). Below I discuss the extent to which these problems will apply.
Immune functioning
Results from empirical studies indicate that many of the above concerns are unlikely to
affect humans when calorie restriction is practiced without malnutrition.30 Omodei and
Fontana point out that there is insufficient evidence about the long-term effects of calorie
restriction on immune function (Omodei and Fontana 2011). However, they cite evidence
of improved immune functioning in the short term.
30 See figure 4 below.
83
Strength and stamina
With regard to strength and stamina, there appear to be conflicting findings. The
CALERIE study found that, although subjects experienced muscle loss, levels of vitality
remained the same, while physical functioning increased (Redman et al. 2011).
Nonetheless, given reduced calorie intake, it should be expected that severe levels of CR
would be incompatible with a high degree of physical activity.
CRMs, as defined in Chapter 1, on the other hand, would have the life extending effects of
CR without restrictions in energy intake from food. Although further empirical
investigation is required, reduced strength and stamina appear to be the result of reduced
food intake, and so are unlikely to be negatively affected. With similar calorie intake there
is little reason to think CRMs will contribute to an energy deficit, in the way that CR
would.
Bone strength
Bone thickness and strength are significant considerations, since reductions in these
contribute to fractures, particularly in the elderly and in malnourished groups. Villareal and
colleagues report that, despite a loss in bone density, there is no degradation in bone
quality (Villareal et al 2010). Since it is lower bone quality that is more strongly associated
with fractures in humans, CR practitioners appear to be at no greater risk.
Moreover, even if CR results in decreased bone strength, there is little reason to think that
CRMs would do so. The most influential explanation for bone loss is decreased mechanical
stress due to decreased weight (ibid.). If this explanation is correct, CRMs would not result
in decreased bone density, since they do not involve a reduction of food intake and
concomitant decreases in weight. Thus there is no reason to think that CRMs would result
in decreased bone density.
84
Hypotension
A further concern expressed by Dirks and Leeuwenburgh is that CR will result in
hypotension, or decreased blood pressure. Although blood pressure levels were
significantly decreased in human CR subjects, this was not beyond the normal range
(Redman 2008). Furthermore, lower blood pressure is associated with reductions in
cardiovascular disease, and is only considered problematic in the presence of other
symptoms.
Cognitive performance
The CALERIE study also assessed levels of cognitive performance and found that these
were not negatively affected (ibid.). CR has also been show to improve cognitive function
in the elderly (Witte 2009). Further, as discussed in Chapter 2, brain ageing appears to be
slowed in CR model organisms. In mice, resveratrol improves later life cognitive
performance relative to control mice (Oomen et al 2009). Thus if mouse and primate
experiments are transferable to humans, it should be expected that full cognitive function
should be retained longer than is normally the case.31
Thus the particular concerns mentioned by Dirks and Leeuwenburgh appear unlikely to be
problematic in the case of CRMs. Moreover, as part of its investigation of health concerns
about calorie restriction and CRMs, the CALERIE study conducted health related quality
of life assessment. Such assessments are, for instance, used in health policy to determine
whether, and the extent to which an intervention will benefit a person in terms of health.
On the basis of this health metric, the researchers concluded that there is little evidence that
31 This should also offset Glannon’s fear that life extension will slow bodily ageing, but not brain ageing,
which is a version of the Struldbrug concern discussed above (Glannon 2009).
85
CR will have unintended consequences, and that there is evidence that the salutary effects
observed in animals appear to transfer to humans (Redman et al 2008, 647).
Stunted growth
One health concern derived from animal studies that is not mentioned by Dirks and
Leeuwenburgh is that growth may be stunted if the intervention is used when young
(Anderson and Weindruch 2012). How should we take this fact? Stunted growth may not
be a bad thing in itself. Indeed Liao and colleagues have controversially argued that
reducing people’s size may be help to reduce carbon emissions (Liao, Sandberg, and
Roache 2012).
However, being smaller may have negative effects for individual welfare. In men in
particular, shorter size is associated with disadvantages in education, relationships and
employment, and many other aspects relevant to welfare (Christensen et al. 2007). These
possibilities may counsel against making use of the interventions at a young age, even if
earlier use would extend lifespan more. At the very least it is important to point out that
using the intervention when still growing involves a trade-off between the goods of life
extension discussed in this section, and the possibility of size-associated drawbacks.32
For the remainder of this thesis I will assume that a person would start the intervention
after she is fully grown. It is important to flag that this is a value-laden assumption. That is,
I make this assumption because I think it is plausible that, given the fact of stunted growth,
and the values that retarded growth would conflict with, it would be reasonable to postpone
taking a CRM.
32 There is also a further difficult question about who decides whether stunted growth is an acceptable price
to pay for a longer life. At the ages at which CR and CRMs could reduce growth, parents are likely to be
largely responsible for health decisions regarding the user’s welfare. It may thus raise difficult ethical issues
about parental responsibility. These are well beyond the scope of this thesis.
86
Potential health concerns
Recent evidence
Immune problems
Immune decline slowed down, but long-term effects
unknown.
Loss of strength and stamina
Muscle loss, but vitality, physical functioning unaffected.
Lower blood pressure
Yes, but not a problem.
Bone thinning, osteoporosis
No decline in bone quality.
Cognitive performance
No reduction. Brain ageing appears to be slowed in
rhesus monkeys. Evidence for extended preservation of
cognitive function in resveratrol fed mice.
Stunted growth
Likely if used early in life
Figure 4 Recent findings about potential health concerns
4.1.3 Conclusions on health
As it stands, studies of CR in general, and results from the CALERIE project in particular
seem positive in terms of the effects of CR on health-related well-being. First, the worry
about indefinite decline in health due to extended ageing seems unwarranted, since CR
extends the biological ageing process rather than causing it to continue beyond its ‘normal’
limit. Second, the concern that the period of decline at the end of life will be extended is
also mitigated. Although an extended decline may occur, a person is likely to have a
satisfactory level of health for longer.
Finally, the available studies indicate, albeit cautiously, that life extension by calorie
restriction will not have negative side effects for health. On the contrary, as discussed in
Chapters 2 and 3, CRMs may delay or prevent the onset of age-associated health
conditions such as cardiovascular diseases, Alzheimer’s disease, diabetes and cancers. In
terms of the substantive good of health, then, CR and CRMs may have a significant role in
enhancing individual welfare.
87
4.2 Procreation
It is likely that procreation will feature on many lists of substantive goods.33 Martha
Nussbaum, for instance lists reproductive health, and reproductive choice as significant
capabilities for a good life (Nussbaum 2000). Similarly, Kass regards procreation as the
‘eternal renewal of human possibility’ (Kass 2004, 318).
Even if procreation is not given the central significance that Kass and Nussbaum accord it,
many people do regard it as an important good. Some regard their own or others’ lives as
less complete if they don’t give rise to other lives. For these, an accompanying decrease in
reproductive ability would at least make prolong life less attractive.
In what follows I discuss two ways that a life extension might interfere with procreation.
The first potential conflict between life extension and the value of procreation is raised by
Kass. He argues that the desire to extend one’s lifespan is attitudinally at odds with
reproduction. The second problem is that the life extending intervention itself might reduce
fertility. I discuss each of these in turn. I claim that desiring to extend lifespan is
compatible with procreation. Moreover, fears that CR and CRMs would decrease fertility
can be mitigated.
4.2.1 Life extension and attitudes to procreation
Kass has argued that the pursuit of life extension for self-interested reasons displaces
procreation as a value in people’s lives. Thus, if procreation is a substantive good one
that is good whether or not people want it and if displacing the value of procreation
33 Some may associate the good of procreation with the good of health. Here I have kept them separate, since
it seems plausible that one can be considered healthy without the ability to procreate. Moreover, as discussed
below, there are attitudinal concerns about life extension and procreation that are separable from concerns
about health.
88
prevents or reduces the likelihood of people having children, then desiring life extension
might make people’s lives worse.34
Kass implies that the individual pursuit of increased longevity entails a self-interested
attitude which is inimical to reproduction. He claims that
simply to covet a prolonged life span for ourselves is both a sign and a cause of our
failure to open ourselves to procreation … one cannot pursue agelessness for
oneself and remain faithful to the spirit and meaning of perpetuation. (Kass 2004,
317)
In other words, wanting to live much longer signals a rejection of the deeper importance of
procreation.
There is a weak but important claim here, which is that the immoderate, single-minded
focus on selfishly pursuing extended lifespan – ‘coveting’ a much longer life ‘for oneself’
may threaten other values, like procreation, that may be significant for individuals and
society. Mythology and popular culture are replete with examples of this. To take an
obvious one, the vampire legend promises immortal life at the cost, inter alia of losing the
ability to reproduce. This warning that a self-interested obsession with living longer will
detract from other values is surely correct.
However, it might also be possible to extract a stronger claim from the above quote: that
wanting to live longer is incompatible with valuing perpetuation, such that if we want
longer life, then we are necessarily being unfaithful to the value of procreation. This strong
claim is implausible. It cannot, for instance, be the case that wanting more happiness – and
more happy years for ourselves entails that a person rejects the value of procreation.
34 Kass may not view procreation as, in the first instance, an aspect of individual welfare. It appears instead to
have some metaphysical importance. This is, however one way of interpreting the argument that is relevant in
the current context.
89
Valuing procreation does not require completely forgoing the pursuit of other aspects of
one’s own welfare.
Kass is right to draw attention to the fact that single-minded desire for longer life can
damage and destabilise other values. The value of life is not determined by its length alone.
However, it would be misguided to suggest that a wanting a better, longer life for oneself
must come at the expense of procreation.
4.2.2 CR, CRMs and reproduction
The second concern about the relation between CRMs and procreation concerns features
particular to the intervention itself. A fear that CR and CRMs will impact on reproduction
is given credence by evolutionary theories of ageing discussed in Chapter 3, and by studies
on a variety of organisms. Nalam and colleagues suggest two central evolutionary reasons
that support the idea that CR will reduce reproduction in mammals:
First, if there is a food shortage, then it would be advantageous for reproduction to
be temporarily halted because this would result in conservation of food for existing
parents and offspring until the food resources have been replenished. Second,
gestation and lactation are energetically costly, so if there is not enough food to
support these processes, then mother, child and future offspring would be lost.
(Nalam, Pletcher, and Matzuk 2008)
Restricting reproduction in times of decreased food intake is advantageous because it gives
adults greater access to food, and reduces energetically costly child-rearing. This
evolutionary claim is borne out by a plethora of animal studies demonstrating that
reproduction is decreased due to decreased fertility under conditions of calorie restriction.35
35 See Speakman and Mitchell 2011 for a review of the effects of CR on reproduction.
90
The possibility that CR and CRMs will decrease reproduction in humans may thus seem to
argue against their use as life extending techniques. If procreation makes a person’s life go
better, then one’s welfare appears likely to be reduced by CR. Below I discuss five
considerations that mitigate this problem. The first is that one can wait until one has
reproduced before taking the intervention. The second is that it is possible to interrupt the
intervention in order to reproduce. The third is that, empirically, fertility and CR are not as
clearly incompatible as the evolutionary argument suggests. The fourth consideration is
that CRMs are less likely to result in the fertility-compromising effect of the stringent CR
diet. The fifth is that advances in technology mean that levels of fecundity no longer
dictate one’s ability to procreate.
Mitigating factors
If CR and CRMs prevent reproduction, perhaps one could wait until after having children
to commence caloric restriction. If the transfer thesis is correct this means that less lifespan
would be gained. If, for instance, one reproduced then, at age 39, commenced the use of an
extremely potent CRM, capable of mimicking the life extending effects of a 60%
reproduction, one could expect to live to 87 years old.36 This is less than if the intervention
commenced at a younger age. Nonetheless it is still a significant gain in life expectancy,
and mitigates concerns about the effects of CRMs on procreation.
A second possibility is that one might interrupt the intervention when one wished to have
children. In this way perhaps one could keep lifespan gains earned before reproducing. As
mentioned in Chapter 3, in a study on calorically restricted female rats, it was found that
fertility increased on return to a normal diet, increasing lifespan and fecundity relative to
controls (Selesniemi 2008). If this result is replicated in humans, women could enjoy both
fertility and extended life.
36 See figure 1.
91
A third consideration that may ameliorate concern about the effects of CR and CRMs is
that it has been demonstrated that some genetically modified mice are able to retain the life
extending effects of CR whilst retaining normal fertility (Partridge, Gems, and Withers
2005). This raises the possibility that the life extending effects of CR and its effects on
fertility are separable. If so, it seems possible that an intervention might be derived that
mimics the effects of CR without undermining procreation.
A further factor is that the fertility inhibiting effects of CR are in part due to lower body
mass. At a low body mass, females become incapable of menstruation. If low body mass is
responsible for declines in fertility, as seems reasonable, then a CRM might not reduce
fertility, since a CRM would not require reduced food intake. Some support for this is
found in the fact that the most widely used candidate CRM, metformin, can be used before,
during, and after pregnancy.37
The final mitigating factor is that procreation is less tightly linked to a natural ability to
reproduce. People who are less fecund, or infertile can procreate using artificial means.
Social factors and technological interventions play an increasingly important role in
procreation. Thus, decreasing fecundity does not necessarily defeat the value of
procreation.
The above considerations do not entirely remove the concern that CRMs will reduce
reproductive ability. More studies in humans are required to determine whether a trade-off
between reproductive potential and the likelihood of extended lifespan is avoidable.
37 It is also thought that metformin increases fertility in women with polycystic ovary syndrome. See Morin-
Papunen et al. 2012
92
Nevertheless, given the points above, it is possible to tentatively conclude that the
reproduction-life extension trade-off may not be required.
4.3 Self-development and flourishing
The notion of self-development has a strong grounding in the three main families of
secular western moral thought. It is apparent in the Aristotle’s virtue ethical notion of
flourishing, Mill’s utilitarian interpretation of the value of liberty, as well as the Kantian
idea of autonomy, or self-rule, as deployed in deontology. As such it is unsurprising that
self-development is a key value on many substantive good theories. Here I interpret self-
development as entailing personal, moral and intellectual growth, since it is on this ground
that life extension has been attacked.
On a simple understanding of flourishing as personal growth, it would seem likely that an
extended life would provide more opportunities to improve one’s personal and moral
skills. Given a longer life, it seems plausible that one would gain a greater degree of
wisdom, and be able to develop one’s skills and interests. However, two objections have
been raised against life extension on the grounds that it will fail to contribute to, and
perhaps even impede self-development.
The first objection suggests that wanting life extension impedes self-development by
causing us to have a damaging focus on hanging onto life. The second concern is based on
the idea that personal growth requires progress through certain natural phases. Kass and
others have argued that life extension might disrupt these phases, and so impede the natural
cycle of development required for a good life.
93
4.3.1 Life extension as a diversion from living well
Kass argues that the pursuit of ‘life-extension will deflect us from realising more fully the
aspirations to which our lives naturally point, from living well rather than staying alive’
(Kass 2003, 25). This objection is similar to the attitudinal fears about procreation
discussed earlier. My response here is the same. Kass is perhaps right to draw attention to
the fact that single-minded desire for longer life can come at the expense of living well.
However, if, as I argue in this Part, living longer is itself a route to living better in terms of
some of the values discussed in this chapter, then there is no conflict been living well and
having a longer life. One wants to live longer precisely because it is a way of improving
one’s life and achieving flourishing.
4.3.2 CRMs, personal growth and the life cycle
In the introduction to this Part, I indicated that some claims about individual welfare hold
that the structure of a life and the order in which goods are achieved are significant for
welfare. ‘Life cycle traditionalism’ is one such view.38 For life cycle traditionalists,
personal growth requires that a life progresses naturally through certain phases. Kass puts
this idea as follows,
the ‘lived time’ of our natural lives has a trajectory and a shape, its meaning
derived in part from the fact that we live as links in the chain of generations. For
this reason, our flourishing as individuals might depend, in large measure, on the
goodness of the natural human life cycle, roughly three multiples of a generation: a
time of coming of age; a time of flourishing, ruling and replacing of self; and a time
of savoring and understanding. (Kass 2003, 26)
38 This term is used by Gems 2011.
94
A life with this structure is paradigmatic of the good life. A life extension technology that
disrupted this natural cycle might thus reduce one’s chances, or, more strongly, be entirely
incompatible with a life of maturity and growth.39
Kass’s argument is deeply rooted in the Natural Law tradition, or ethical naturalism. This
brand of ethical thinking has been heavily criticised in normative ethical theory, not least
because it appears to commit the ‘naturalistic fallacy.’40 That is, it has been criticised for
jumping from the claim that something is natural to the claim that it is good, without
intervening premises that give reasons why we should think that what is natural is good. In
the current context, Kass’s argument might be criticised for fallaciously moving from the
claim that humans in fact have a natural life cycle to the claim that the life cycle is a good,
and interference with it is bad.
Although the ethical naturalist basis of Kass’s argument is controversial, I will not criticise
it here. Instead, in keeping with the methodology outlined in the introduction to this thesis,
I will pursue a compatibilist approach. That is, I argue below that decelerated senescence is
compatible with self-development that is bound to life cycles. This is because CRMs
would introduce few changes to ‘shape’ of life, and would not prevent the passage through
any of the phases that life cycle traditionalists view as valuable.
CR, CRMs and the life cycle
CR and CRMs slow ageing. This means that the that transition through life’s phases would
be secured. We would still be children, and then adults, and then elderly. A CRM user
would simply spend a longer period of chronological time in each phase. If the passage
through the phases is all that matters to self-development, the response to life cycle
39 Note that procreation is an important aspect of the life cycle as conceived by life cycle traditionalists.
Earlier I argued that the use of CRMs is compatible with procreation, so that CRMs shouldn’t pose a problem
to life cycles in terms of this value.
40 See for instance Horrobin 2006.
95
traditionalists is relatively easy. There would be no conflict with self-development because
we pass through all the relevant phases of personal growth.
However, life cycle traditionalists may also to be concerned about the shape of a life. This
might be conceived of as requiring that the phases of life remain in proportion to one
another. If we accept this proportionality requirement CR and CRMs might be
problematic. This is because the intervention would only slow ageing from the point at
which the intervention was begun. Phases of growth before this would not be extended.
This is a concern given the earlier assumption that the use of CRMs should only
commence after a person is fully grown, in order to avoid stunted growth (S4.1.2). If the
intervention was begun after a person was full-grown, then later phases would potentially
be disproportionately long with respect to the growth phase.
A further way in which the phases of the life cycle could be made disproportionate is if the
intervention was stopped earlier. If, for instance, one begun the intervention at adulthood
and discontinued the use of a CRM after the middle phase of development the time of
‘ruling and replacing of self’ then the middle phase might be disproportionately long in
relation to both the later and the earlier phase.
It is worth pointing out a further concern that does not apply. It might be thought that the
last phase - the time of ‘savouring and understanding’ - could be made disproportionately
long relative to the earlier phases. However, as argued in Chapter 1, the degree of lifespan
achieved by CRMs is likely to diminish the later the intervention is used. If, for instance,
the intervention was begun at 55, there would be almost no increase in lifespan.41 Increases
in lifespan achieved by use of CRMs in the last phase are likely to be negligible. The last
phase will not, therefore, be longer relative to both other phases.
41 See figure 1.
96
The above considerations mean that there are two ways in which CRMs might make the
phases discussed by Kass disproportionate. First, the middle period of ‘flourishing, ruling,
and replacing of self’ and the final period of ‘savouring and understanding’ might be
comparatively longer than the first period of ‘coming of age.’ Second, the middle period
might be made longer than both the initial phase and the final phase.
Would life cycle traditionalists find the relative shortness of earlier and /or later phases
problematic in terms of self-development? That is, would they insist on strict
proportionality with respect to these phases? The proportionality requirement seems very
difficult to argue for. In the first place, it’s far from clear that these phases are as distinct as
Kass suggests. Moreover, to the extent that it is possible to distinguish between these
periods, there’s not much reason to think that they are in proportion in the normal case.
Intuitively people spend very different amounts of time in each phase. If so, the fact that
CRMs alter the proportionality of phases wouldn’t represent a substantial departure from
what occurs in a normal life.
Life cycle traditionalism at its most plausible should thus be thought of as holding that
what is important for a complete life is that one experiences these phases of development
and maturation; not that they are proportionate. Since CRMs do not impede this transition
to full maturity, a CRM user is at least as likely as anyone else to achieve a life of
flourishing and personal growth. Indeed, in a longer life a person might achieve greater
development of her personal skills, and moral and intellectual characteristics.
4.4 Creativity and beauty
The value of beauty has historical significance, having been linked with other
‘transcendental’ values, such as ‘goodness’ and ‘truth’ (Blackburn 2010). More recently,
97
Derek Parfit includes ‘awareness of true beauty’ as an example of something that might be
considered objectively valuable (Parfit 1984, 499). Two separate problems with life
extension have been raised in relation to the values of creativity and beauty. The first holds
that a sense of mortality underlies the creation and awareness of beauty. The second
highlights potential effects of CR on the human body. I claim that the first argument fails
because technologies that decelerate ageing do not prevent death. The success of the
second argument depends on how highly one values one’s own beauty, and may require a
trade-off with other values if one is to commence CR. However, CRMs would avoid the
second objection altogether.
4.4.1 Mortality as a prerequisite for beauty
Kass has suggested that mortality is a condition for the creation of beauty:
Perhaps … only a mortal being, aware of his mortality and the transience and
vulnerability of all natural things, is moved to make beautiful artifacts, objects that
will last, objects whose order will be immune to decay as their maker is not. (Kass
2001, 21)
Kass goes further and implies that, in addition to mortality being a prerequisite for the
creation of beautiful objects, it may also be necessary for their appreciation:
Could the beauty of flowers depend on the fact that they will soon wither? Does the
beauty of spring warblers depend upon the fall drabness that precedes and follows?
What about the fading, late afternoon winter light or the spreading sunset? Is the
beautiful necessarily fleeting, a peak that cannot be sustained? (Ibid.)
A number of responses are possible to Kass’s considerations. First, it is possible to
challenge the idea that impermanence is the source of beauty. It is dubious, for instance
that flowers are beautiful just because they will die. Second, even if we accept that
impermanence is the sole source of beauty, human death is not the only type of
98
impermanence. Perhaps we would still be inspired to create beautiful objects by the
impermanence of spring warblers and late afternoon winter light.
Though these are possible responses to Kass’s claim, I will not argue for them strongly
here. It is unnecessary to do so since Kass’s argument only targets immortality
interventions that would prevent death altogether. CRMs have no such power and would at
most postpone death. As such, they do not remove this potential inspiration for the creation
and appreciation of beauty. Thus, if it were raised against the intervention at issue, the
argument would fail altogether.
4.4.2 CR and physical beauty
One positive argument in favour of life extension on the basis of the value of beauty is
Hackler’s suggestion that life extending interventions would have the effect of ‘prolonging
the period of attractiveness and desire’ (Hackler 2004, 192).
However, the effects of CR introduce a problem with the possibility raised by Hackler. In
particular, some practitioners of CR complain that the diet worsens their self-perceived
aesthetic qualities and makes them less attractive to others.42 In particular, they are much
thinner and less muscular, failing to conform to existing beauty ideals. Thus, rather than
prolonging a period of perceived attractiveness, CR may conflict with a person’s aesthetic
beauty.
Note that this dip in self-perceived attractiveness is not universal. Michael Rae, a biologist
who practices CR suggests he prefers being much thinner.43 Nonetheless, it is a potential
side-effect of a severe CR diet that should be taken into account. Practitioners are
42 http://www.crsociety.org/resources/risks. Last accessed 20 December 2012.
43 http://www.macleans.ca/science/health/article.jsp?content=20070115_139289_139289. Last accessed 20
December 2012.
99
confronted with the possibility that efforts to remain biologically younger will reduce their
attractiveness. Consequently, in the case of CR itself, this negative effect should be
weighed against the positive effects on health and other values outlined in this Part. For
some, a reduction of attractiveness may be a sufficient reason not to practice a rigorous diet
with perhaps uncertain payoffs.
This difficulty is a further motivation for the development of CRMs. Evidence from studies
on mice treated with resveratrol suggests that CRMs could increase healthy lifespan even
in overweight mice (Baur et al 2006). So a CRM need not result in the emaciated physique
that CR practitioners complain of. As a result, there’s reason to think that CRMs would not
compromise physical beauty in the way that CR would.
Furthermore, in keeping with Hackler’s suggestion, CRMs would potentially prolong the
time which a person is typically regarded as physically attractive. Since ageing appears to
be slowed throughout the organism, phases usually regarded as being times of increased
attractiveness will be extended. In this way an effective CRM could contribute to a
person’s perceived aesthetic value, and also allow more time to experience other things of
beauty.
4.5 Community
The importance of community as a self-interested value can be derived from the following
quote:
selves tend to be defined or constituted by various communal attachments (e.g., ties
to the family or to a religious tradition) so close to us that they can only be set aside
at great cost, if at all… [W]e also need to sustain and promote the social
attachments crucial to our sense of well-being and respect, many of which have
been involuntarily picked up during the course of our upbringing. (Bell 2012)
100
Understanding community as a prudential value in this way implies that individual welfare
is constituted in part by others. Valuing community prudentially means that an insistence
on a definite separation between one’s own good and the good of (at least some) others
becomes untenable. It is at least partly for this reason that communitarians criticise the
atomistic pursuit of one’s own good that they associate with modern liberalism.44 If it is
true that our own welfare is intertwined with the welfare of others, then the pursuit of our
own good necessitates a concern for the well-being of others.
In this section I address the fear that life extension will result in negative changes to the
nature and fabric of human relationships. Before I do so, however, I should point to two
additional community related issues that are dealt with in subsequent sections. The first is
that extending lifespan means that users of CRMs are more likely to be lonely at the end of
their lives, and more likely to experience the loss of loved ones. The second concern is that
individuals will feel that they are a burden on their community. Although these problems
could equally be discuss