Is the report of widespread body pain associated with long-term
increased mortality? Data from the Mini-Finland Health Survey
G. J. Macfarlane, G. T. Jones, P. Knekt1, A. Aromaa1, J. McBeth2, M. Mikkelsson3and M. Heliovaara1
Objective. To determine whether an observation in a UK study, that persons with chronic widespread pain are at long-term increased risk of
cancer mortality, can be replicated in a different setting.
Methods. Subjects were participants aged ?30yrs in the Mini-Finland Health Survey conducted between 1979 and 1980. Information
collected included prevalent pains at different joints throughout the body, demographic, anthropometric, lifestyle and occupational factors.
During follow-up, until 1994, information on vital status and cause of death was obtained.
Results. 7182 persons participated (89.8%). The prevalence of widespread body pain (pain at four or more sites) was 20% in females and
12% in males, and during follow-up there were a total of 1647 deaths. The risk of death was not elevated amongst those with widespread pain
[relative risk (RR): 0.86; 95% confidence interval (CI): 0.74–1.00], and in particular, those with widespread pain were at a slightly lower risk of
several disease-specific causes of death and cancer death (RR: 0.64; 95% CI: 0.46–0.91).
Conclusions. This study of multiple pains has not confirmed a previous observation of an association between the reporting of widespread
pain and subsequent increased risk of cancer death. Differences in the definitions used or, more probably, the population studied, in
particular, a larger rural population with more multiple pains related to physical activity may account for the differences.
KEY WORDS: Widespread pain, Mortality, Cohort study, Cancer.
Chronic widespread body pain is the clinical hallmark of the
syndrome known to rheumatologists as fibromyalgia. Evidence
strongly suggests that fibromyalgia is not a distinct clinical entity
but one end of a spectrum of pain and tender points . Chronic
widespread pain has a population prevalence of 10–13% and its
onset is predicted by individual factors, such as psychological
distress, aspects of illness behaviour and by previous reports of
other somatic symptoms . Aspects of the social environment,
which have been particularly studied in a workplace setting,
also influence the likelihood of symptom onset . A variety
of abnormalities in biology have been investigated in patients
with fibromyalgia, although consistent evidence for many
symptoms and high levels of disability ; however, until recently
it was not believed that they resulted in premature mortality.
A population-based prospective cohort study from the UK,
however, demonstrated that persons reporting chronic widespread
pain were twice as likely to die over the next 9yrs (compared with
persons reporting no pain) and that this excess risk of death was
principally related to cancer deaths . These results, if true,
would have important implications for management. Preliminary
results from a further study conducted in the same geographical
area confirm some of the previous results . No other
studies examining this issue directly have reported, and the
available related indirect evidence provides only weak supporting
evidence. In addition, there is no clear biological mechanism
linking the report of pain with a long-term increase in disease-
is associatedwith persistent
It is therefore important, before placing too much emphasis on
this observation, to try to replicate these findings, and in doing so
it would be particularly advantageous to do so in a diverse setting
from the original observation. We, therefore, took advantage of
available information from the Mini-Finland Health Survey to
determine whether persons reporting multiple bodily pains
subsequently experience increased mortality either overall or
from specific causes of death.
The methods of the Mini-Finland Health Survey, conducted
between 1978 and 1980, have been described in detail pre-
viously . In brief, the study population was a stratified two-
stage cluster sample drawn from the Finnish population aged
?30yrs. In the first stage, areas of the country were selected which
were considered representative in terms of geographical area,
urbanization and employment. In the second stage, a sample of
inhabitants was drawn from the population register of each area.
Of the 8000 persons selected, all the data necessary for the current
study were obtained for 7182 (89.8%). The distributions of the
participants with respect to age, sex, marital status and level
of education corresponded closely with the whole Finnish
Information was collected in the study using interviews within a
mobile clinic. Subjects were asked, ‘Have you had a pain, ache or
tenderness on movement in one or more joints during the last
month? Can you identify these joints on the body manikin?’
Respondents were also specifically asked about neck and back
pain. The number of painful sites and joints were then summed.
We defined widespread pain as pain present in at least four sites,
since examination of the data distribution showed that this
resulted in a prevalence of widespread pain similar to other
prevalence studies, and since it had face validity with respect to the
American College of Rheumatology (ACR) definition of wide-
spread pain in their criteria for fibromyalgia . Pain reported in
between one and three sites was defined as regional pain.
Information was also collected on age, sex, years of education,
present or most recent occupation, current alcohol consumption
and tobacco habits. Height and weight were measured allowing
body mass index (BMI) to be derived.
1 of 3
Aberdeen Pain Research Collaboration (Epidemiology Group), University of
Aberdeen, Scotland, UK,
2Arthritis Research Campaign Epidemiology Unit, Division of Epidemiology and
Health Sciences, University of Manchester, England, UK and
Rehabilitation, The Rheumatism Foundation Hospital, Heinola, Finland.
1National Public Health Institute, Helsinki, Finland,
Submitted 19 May 2006; revised version accepted 31 October 2006.
Correspondence to: Prof. Gary J. Macfarlane, Epidemiology Group, Department
of Public Health, School of Medicine, Polwarth Building, Forresterhill, Aberdeen,
Scotland, AB25 2ZD, UK. E-mail: firstname.lastname@example.org
Rheumatology 2006; 1 of 3 doi:10.1093/rheumatology/kel403
? The Author 2006. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: email@example.com
Rheumatology Advance Access published December 22, 2006
by guest on June 3, 2013
Information was collected on physical and mental work stress.
The presence of physically stressful features was recorded in the
present or in the previous occupation of longest duration as
dichotomies (no¼0, yes¼1) (lifting or carrying heavy objects;
stooped, twisted or otherwise awkward work posture; shaking of
the whole body or use of vibrating equipment; a constantly
repeated series of movements; and work paced by a machine)
and summed to form an index (0–5). Similarly mental stress
features (uninteresting or monotonous work, a hurried or
tight work schedule, and worry about making mistakes)
were recorded at three levels of severity (none¼0, mild¼1,
severe¼2) and summed to form an index of mental stress at work
During follow-up, the vital status of the participating subjects
was determined by obtaining information from the Central
Statistical Office of Finland. Principal causes of death were
coded according to the 8th revision of the International Standard
Classification of Disease, Injuries and Causes of Death (ISD-8).
Follow-up for this analysis was continued until 31 December
1994, equivalent to a follow-up between 14yrs and 16yrs, and
providing a total of 88870 person-years.
mortality was analysed using Cox proportional hazard models,
with adjustment for potential confounding variables. Hazard
ratios were derived from the model (i.e. the ratio of probability
of death in the pain groups compared with those with no
pain) and are expressed as relative risks (RR) with 95%
confidence intervals (CIs). Assumptions of proportionality were
relationshipbetweenpain reportingand subsequent
The socio-demographic characteristics and reported lifestyle of the
participants is shown in Table 1: 46% were male, most had
elementary level of education (68%), and most were never
smokers (56%). In terms of their jobs, 6% were never employed
while for the remainder their current or longest-serving jobs
(if currently not employed) were evenly split between professional
(22%), agricultural (23%), industrial (25%) and the service
sector (23%). An intermediate or high physical work stress
was more common (45%) than an intermediate or high
mental work stress (34%) (Table 1). Older age, low number of
years of education, obesity and a high mental or physical
workload were all significantly associated with the reporting of
Males were more likely to report no pain (39 vs 30%) and less
likely to report widespread pain (four or more body sites: 12
vs 20%). Regional pain exhibited a similar prevalence (pain at 1–3
body sites: 49 vs 50%). During the follow-up period there were a
total of 856 and 791 deaths amongst male and female participants,
respectively. The risk of death was unrelated to the number of
pain sites reported on the recruitment screening survey in either
males or females. Neither for overall mortality nor for any of the
disease-related deaths was there an excess risk of death in those
reporting regional or widespread pain. Indeed for cardiovascular,
respiratory and cancer deaths, and all other disease-related deaths
(combined), there was reduced risk of death for those reporting
widespread pain which for cancer death was statistically
significant (Table 2).
This population-based prospective study from Finland has not
supported an initial observation in a UK study that widespread
body pain is associated with an increased risk of death and,
in particular, cancer death. There are a number of methodological
issues which are relevant to consider:
(i) Unlike the original observation, the study was unable to
define widespread pain according to the definition used in
the ACR criteria for fibromyalgia . Instead the study used
a count of painful joints of the body. It is possible therefore
that the relationship with increased mortality is only with
widespread pain rather than multiple joint pains. Against
this, however, is the fact that the most common areas for
pain to be reported on the body manikins is in the axial
skeleton and major joints (i.e. shoulder, knee and hip) ,
that the ACR criteria are most commonly satisfied by
multiple regional pains rather than pain over the whole body
and that when an alternative definition of widespread pain is
applied in population studies (Manchester criteria) which
requires the pain to be more truly widespread, prevalence is
considerably lower [12, 13].
(ii) Information is not available about the duration of pain
reported at multiple sites. However, data from other studies
has shown that in 80–90% of persons reporting widespread
body pain, the pain has been present for >3 months .
Therefore, how can we reconcile this observation of no relation-
ship between multiple joint pain reports and subsequent mortality
experience with our previous study showing a strong relationship
between widespread pain, death from cancer and deaths from
suicides and accidents? Firstly it may be that the original
observation is wrong and that although strong, the relationship
was observed by chance. This may be the case, although initial
results from a second prospective study in the UK have confirmed
the initial results and shown a doubling of risk of both
TABLE 1. Characteristics of participants and their relationship to the report of
CharacteristicsSubjects n (%)RR (95% CI)
Educational level (yrs)
Physical work stress
Mental work stress
Alcohol consumption (g/week)
Not at all
Body mass index (kg/m2)
2 of 3G. J. Macfarlane et al.
by guest on June 3, 2013
cardiovascular and cancer death over the 8yrs after pain reports Download full-text
. Interestingly, this second UK study was also not able to use
the ACR definition of widespread pain, but instead had a count of
painful joint areas in the body, more similar to the definition used
in the current study—thus suggesting that the different results are
not due to different definitions of widespread pain. Alternatively,
it may be that we are studying different types of populations and
whose symptoms have a diverse aetiology. The populations in the
two studies showing a strong relationship between pain and cancer
death have both been from the Greater Manchester area of the
UK. This is a densely populated conurbation, with employment in
either industrial or service sectors, and including areas of social
deprivation. There have been many studies of chronic widespread
pain conducted in this geographical area—and they have shown
that the reports of widespread pain are associated with a
psychological (rather than mechanical) aetiology and social
factors. In contrast, a much greater proportion of the population
in the Mini-Finland Health Survey lived in rural settings, the
majority finished education at the elementary level and roughly
equal proportions of persons were employed across professional,
agricultural, industrial and service industries. The development
from a poor agricultural society to an industrialized one and,
later on, to the post-industrial stage has been more rapid and
drastic than elsewhere. In Finland in 1980, consequently, a
substantial proportion of people had been exposed to very heavy
manual labour, a risk factor for widespread pain. A previous
report from this study on the prevalence and associated features of
fibromyalgia (which has widespread pain as one of its primary
symptoms) demonstrated that fibromyalgia was significantly more
common in those employed in the agriculture sector, was strongly
related to high physical work stress, but not with level of mental
work stress . A particularly strong relationship between
physical work stress and widespread pain is shown in the current
analysis. It could, therefore, be that the multiple pains reported in
this study are more likely to be as a result of mechanical load and
injury than is the case in urban populations. If so, the conclusion
would be that if there is an increased risk of cancer (and perhaps
cardiovascular) death it is not linked to the reporting of multiple
pains per se, but to the chronicity of symptoms and the commonly
associated comorbid features. This hypothesis will need to be
tested in future studies.
In conclusion, the data on whether there is a link between the
experience of widespread body pain and premature mortality is
contradictory. This study (of multiple joint pains) has found no
relationship. This suggests that, if there is a link, it is not a simple
biological link triggered by the experience of multiple pains.
We, therefore, should remain cautious about whether a link exists
but studies currently underway will allow us to determine whether
the link is specific to pain that is truly ‘widespread’ (with its
common comorbidities) and/or whether behavioural and lifestyle
changes consequent upon the experience of chronic symptoms
may explain the relationship. Understanding any potential link is
important for the many patients who present to primary and
secondary care with widespread body pain symptoms.
The authors have declared no conflicts of interest.
1 Macfarlane GJ. Fibromyalgia, Chronic widespread pain. In: Crombie IK, Croft PR,
Linton SJ, LeResche L, von Korff M, eds. Epidemiology of pain. International
Association for the Study of Pain, IASP Press, 1999.
2 ClauwDJ, Crofford LJ. Chronic widespread pain
we know, and what we need to know. Best Pract Res Clin Rheumatol
3 Bennett R. Fibromyalgia: present to future. Curr Rheumatol Rep 2005;7:371–6.
4 Papageorgiou AC, Silman AJ, Macfarlane GJ. Chronic widespread pain in the
population: a seven year follow-up study. Ann Rheum Dis 2002;61:1071–4.
5 Macfarlane GJ, McBeth J, Silman AJ. Widespread body pain and mortality:
prospective population-based study. Brit Med J 2001;323:1–5.
6 Macfarlane GJ, Symmons DPM, Silman AJ et al. Widespread pain is associated with
a long-term increased risk of cancer death. Arthritis Rheum 2005;52(Suppl):S77.
7 Ma ¨kela ¨ M, Heliovaara M, Sievers K, Knekt P, Maatela J, Aromaa A. Musculoskeletal
disorders as determinants of disability in Finns aged 30 years or more. J Clin
8 Aromaa A, Helio ¨vaara M, Impivaara O, Knekt P, Maatela J. The execution of the
mini-Finland health survey. Part 1: Helsinki, Turku: Social Insurance Institution 1989
(In Finnish with English summary).
9 Wolfe F, Smythe HA, Yunus MB et al. The American College of Rheumatology, 1990
Criteria for the Classification of Fibromyalgia. Report of the Multicenter Criteria
Committee. Arthritis Rheum 1990;33:160–72.
10 Ma ¨kela ¨ M, Helio ¨vaara M, Sievers K, Impivaara O, Knekt P, Aromaa A. Prevalence,
determinants and consequences of chronic neck pain in Finland. Am J Epidemiol
11 Macfarlane GJ, Jones GT, McBeth J. Epidemiology of pain. In: McMahon SB,
Koltzenburg M, eds. Wall and Melzack’s textbook of pain, 5th edn. (Chapter 76)
Elsevier Churchill: Livingstone, 2005.
12 Hunt IM, Silman AJ, Benjamin S, McBeth J, Macfarlane GJ. The prevalence and
associated features of chronic widespread pain in the community using the
‘Manchester’ definition of chronic widespread pain. Rheumatology 1999;38:275–9.
13 Macfarlane GJ, Croft PR, Schollum J, Silman AJ. Widespread pain: is an improved
classification possible? J Rheumatol 1996;23:1628–32.
14 Ma ¨kela ¨ M, Helio ¨vaara M. Prevalence of primary fibromyalgia in the Finnish
population. Brit Med J 1991;303:216–9.
TABLE 2. Pain report and subsequent specific cause of death
RR (95% CI)No. of deaths
0 1–3 4–10
Cause of death/pain sites No pain Regional painWidespreadN
Other disease-related deaths
All analyses adjusted for age group, gender, education, physical work stress, mental work stress, alcohol consumption, tobacco smoking and BMI.
Rheumatology key messages
? This study has not confirmed that persons with widespread body
pain are at an increased risk of death (or, in particular, cancer).
? The most likely explanation for this disagreement between studies
is either differences in populations studied or in the definition of
widespread pain used.
Widespread pain and mortality3 of 3
by guest on June 3, 2013