Article

Reduction versus abrupt cessation in smokers who want quit

Primary Care Clinical Sciences, University of Birmingham, Edgbaston, Birmingham, West Midlands, UK, B15 2TT.
Cochrane database of systematic reviews (Online) (Impact Factor: 6.03). 03/2010; 3(3):CD008033. DOI: 10.1002/14651858.CD008033.pub2
Source: PubMed
ABSTRACT
The standard way to quit smoking is to smoke as normal until a quit day at which point the smoker stops using all cigarettes. Most smokers who try to quit end up relapsing, therefore there are a number of people who have tried to quit abruptly in the past without success, and are disillusioned with this approach. An alternative way to give up could be to reduce the amount of cigarettes smoked before going on to quit completely. There is evidence to suggest that reducing smoking before quitting would be popular with smokers. This means that offering this approach to quitting could encourage more smokers to give up, however before offering this approach it is important to ensure it is at least as successful as abrupt quitting. This is because given a choice smokers who would otherwise have quit abruptly may choose to reduce first instead. If reduction isn't as effective, smokers who choose that method will be at a disadvantage. The aim of this review was to compare quit rates in reduction to quit and abrupt quitting interventions to see if reducing to quit is at least as successful as abrupt quitting. Ten studies were found which compared reducing smoking before quitting with abrupt quitting. Pooled results found that neither reducing or abrupt quitting produced superior quit rates. This was true whether nicotine replacement therapy was used as part of the intervention or not, and whether participants were offered self-help materials or behavioural support. These results suggest that smokers should be given a choice of quitting methods, either reducing smoking before quitting or abrupt quitting, however, to inform the development of new interventions more research is needed into which method of reducing smoking is the most effective.

Full-text

Available from: Paul Aveyard
Reduction versus abrupt cessation in smokers who want to
quit (Review)
Lindson N, Aveyard P, Hughes JR
This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library
2010, Issue 3
http://www.thecochranelibrary.com
Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 1
T A B L E O F C O N T E N T S
1HEADER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2PLAIN LANGUAGE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Figure 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
13DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15AUTHORS’ CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
19CHARACTERISTICS OF STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
33DATA AND ANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Analysis 1.1. Comparison 1 Reduction to quit versus abrupt quitting, Outcome 1 Abstinence: Main analysis. . . . 33
Analysis 1.2. Comparison 1 Reduction to quit versus abrupt quitting, Outcome 2 Abstinence: Sub-group analysis split by
use of pharmacotherapy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Analysis 1.3. Comparison 1 Reduction to quit versus abrupt quitting, Outcome 3 Abstinence: Sub-group analysis split by
type of behavioural support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
36APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
39HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
39CONTRIBUTIONS OF AUTHORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
39DECLARATIONS OF INTEREST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
39SOURCES OF SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
39INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
iReduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 2
[Intervention Review]
Reduction versus abrupt cessation in smokers who want to
quit
Nicola Lindson
1
, Paul Aveyard
1
, John R Hughes
2
1
Primary Care Clinical Sciences, University of Birmingham, Birmingham, UK.
2
Dept of Psychiatry, University of Vermont, Burlington,
Vermont, USA
Contact address: Nicola Lindson, Primary Care Clinical Sciences, University of Birmingham, Edgbaston, Birmingham, West Midlands,
B15 2TT, UK.
nll839@bham.ac.uk.
Editorial group: Cochrane Tobacco Addiction Group.
Publication status and date: New, published in Issue 3, 2010.
Review content assessed as up-to-date: 6 December 2009.
Citation: Lindson N, Aveyard P, Hughes JR. Reduction versus abrupt cessation in smokers who want to quit. Cochrane Database of
Systematic Reviews 2010, Issue 3. Art. No.: CD008033. DOI: 10.1002/14651858.CD008033.pub2.
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
A B S T R A C T
Background
The standard way to stop smoking is to quit abruptly on a designated quit day. A number of smokers have tried unsuccessfully to quit
this way. Reducing smoking before quitting could be an alternative approach to cessation. Before this method is adopted it is important
to determine whether it is at l east as successful as abrupt quitting.
Objectives
1. To compare the success of reducing smoking to quit and abrupt quitting interventions. 2. To compare adverse events between arms
in studies that used pharmacotherapy to aid reduction.
Search strategy
We searched the Cochrane Tobacco Addiction Review Group specialised register, MEDLINE, EMBASE and PsycInfo for topic specific
terms combined with terms used to identify trials of tobacco addiction interventions. We also searche d reference l ists of relevant papers
and contacted authors of ongoing trials. Date of most recent search: November 2009.
Selection criteria
We included randomized controlled trials (RCTs) that recruited adults who wanted to quit smoking. Studies included at least one
condition which instructed participants to reduce their smoking and then quit and one condition which instructed participants to quit
abruptly.
Data collection and analysis
The outcome measure was abstinence from smoking after at least six months follow-up. We pooled the included trials using a Mantel-
Haenszel fixed-effect model. Trials were split for two sub-group analyses: pharmacotherapy vs no pharmacotherapy, self help therapy
vs behavioural support. Adverse events were summarised as a narrative. It was not possible to compare them quantitatively as there was
variation in the nature and depth of reporting across studies.
1Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 3
Main results
Te n studies were relevant for inclusion, with a total of 3760 participants included in the meta-analysis. Three of these studies used
pharmacotherapy as part of the interventions. Five studies included behavioural support in the intervention, four included self-help
therapy, and the remaining study had arms which included behavioural support and arms which included self-help therapy. Neither
reduction or abrupt quitting had superior abstinence rates when all the studies were combined in the main analysis (RR= 0.94, 95%
CI= 0.79 to 1.13), whether pharmacotherapy was used (RR= 0.87, 95% CI= 0.65 to 1.22), or not (RR= 0.97, 95% CI= 0.78 to
1.21), whe ther studies included behavioural support (RR= 0.87, 95% CI= 0.64 to 1.17) or self-help therapy (RR= 0.98, 95% CI=
0.78 to1.23). We were unable to draw conclusions about the difference in adverse events between interventions, however recent studies
suggest that pre-quit NRT does not increase adverse events.
Authors’ conclusions
Reducing cigarettes smoked before quit day and quitting abruptly, with no prior reduction, produced comparable quit rates, therefore
patients can be given the choice to quit in either of these ways. Reduction inter ventions can be carried out using self-help materials or
aided by behavioural support, and can be carried out with the aid of pre-quit NRT. Further research needs to investigate which method
of reduction before quitting is the most effective, and which categories of smokers benefit the most from each method, to inform future
policy and intervention development.
P L A I N L A N G U A G E S U M M A R Y
Comparing reducing smoking to quit with abrupt quitting.
The standard way to quit smoking is to smoke as normal until a quit day at which point the smoker stops using all cigarettes. Most
smokers who try to quit end up relapsing, therefore there are a number of people who have tried to quit abruptly in the past without
success, and are disillusioned with this approach. An alternative way to give up could be to reduce the amount of cigarettes smoked
before going on to quit completely. There is evidence to suggest that reducing smoking before quitting would be popular with smokers.
This means that offering this approach to quitting could encourage more smokers to give up, however before offering this approach it is
important to ensure it is at least as successful as abrupt quitting. This is because given a choice smokers who would otherwise have quit
abruptly may choose to reduce first instead. If reduction isnt as effective, smokers who choose that method will be at a disadvantage.
The aim of this review was to compare quit rates in reduction to quit and abrupt quitting interventions to see if reducing to quit is at
least as successful as abrupt quitting. Ten studies were found which compared reducing smoking before quitting with abrupt quitting.
Pooled results found that neither reducing or abrupt quitting produced superior quit rates. This was true whether nicotine replacement
therapy was used as part of the intervention or not, and whether participants were offered self-help materials or behavioural support.
These results suggest that smokers should be given a choice of quitting methods, either reducing smoking before quitting or abrupt
quitting, however, to inform the development of new interventions more research is needed into which method of reducing smoking
is the most effe ctive.
B A C K G R O U N D
Description of th e cond ition
Tobacco use is the largest preventable cause of death in the world,
as a risk factor for six of the eight leading causes of death. A
survey of a sample of 893 English smokers (Jarvis 2002) found
that most were disenchanted with smoking and, if they were given
their time again, they would not have started smoking in the first
place. Most of these smokers expected to be quit within a few
years, but historical data on quit rates suggest this is extremely
unlikely because most people who try to quit relapse quickly (
Hunt
1973
). The authors interpreted this as a delusion gap between
expectations and likely reality. Such a gap means we need to find
new ways to encourage smokers, most of whom have tried to quit
repeatedly, to keep on trying. Finding new ways to quit would be
helpful to this endeavour.
2Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 4
Description of th e intervention
The standard way to stop smoking is to quit abruptly. This means
that a person smokes as normal until a designated quit day, from
which point forward they try to abstain and avoid any smoking
whatsoever.
An alternative method is to quit gradually. Such gradual reduction
methods, when used as a means of achieving cessation, typically
have a quit day as in abrupt cessation. The key difference is that
smokers aim to reduce smoking prior to this day. Some researchers
have investigated the relative efficacy of different methods of re-
ducing smoking on the likelihood of achieving reduction and of
subsequently achieving abstinence, but this is not the focus of th is
review. All such different methods are pooled here.
How t he intervention might work
There are a number of ways that reducing the amount of cigarettes
smoked prior to total abstinence might help a smoker give up com-
pletely. The first of these is a principle of psychopathology, which
suggests that, as the dose of nicotine received by the individual
each day is reduced, drug dependence and therefore craving is re-
duced in response. Another is shaping’, a conditioning procedure,
whereby making successive approximations of the target behaviour
(gradually cutting down the number of cigarettes smoked) the de-
sired behaviour (abstinence) is eventually achieved. The third is
the cognitive psychology principle that completing a step toward a
goal (reducing smoking) increases self efficacy, which increases the
likelihood that the goal (abstinence) will be achieved. The fourth
is the classical and operant conditioning principle that reducing
the frequency of a behaviour decreases the association with envi-
ronmental cues, which in turn weakens the urge to partake in that
behaviour when those cues are present. Finally reducing provides
a goal which is more in-line with the smokers current behaviour
than complete abstinence and the refore appears more achievable.
However, the standard assumption of smoking cessation treatment
is that cessation begins on a quit day and that cutting down prior to
quitting is not advised. This is based on nicotine addiction theory,
which claims that the user has impaired control over their drug
use, and that it should therefore be difficult for them to control
their usage in any way, e.g. by reducing. It also assumes that with
reduction each remaining cigarette will become more rewarding
and harder to give up, and that the smoker will suffer a loss of mo-
tivation before attaining total abstinence. However medication to
reduce withdrawal, such as nicotine replacement therapy (NRT),
could be used to counteract this effect, and has successfully been
used to do so in smokers who are not yet ready to quit (
McRobbie
2006
; Wang 2008). Wang et al conducted an assessment report
which examined the effectiveness and cost- effectiveness of NRT
alongside ’cut down to quit’ (CDTQ) smoking in people who were
either unwilling or unable to quit (
Wang 2008). The approach was
found to be both effective, and cost effective, although abstinence
rates were not as high as those documented in abrupt quitting
regimes. Studies which utilise the CDTQ approach, and therefore
people who are unable or unwilling to quit, have already been in-
cluded in the Cochrane review of harm reduction (Stead 2007).
Our review will focus only on those smokers who want to quit.
Surveys have been carried out across England and Wales (
West
2001) and the UK, US, Canada and Australia (Cheong 2007),
investigating the success of quit attempts when smokers reduce
cigarettes smoked with an aim to quitting completely. Both of
these observational studies found that abrupt quitting was almost
twice as successful as quitting gradually in those sampled. How-
ever participants in these studies were from the general population
and hadnt used a particular service or intervention. They could
potentially have used a wide range of gradual quitting techniques,
ranging from no str ucture, no reduction goals and no set quit day
to highly structured, with set reduction goals and a target quit day
to work toward, which may have influenced success rates.
Although British (
NICE 2008) and American (Fiore 2008) na-
tional guidelines for smoking cessation services do not recommend
reducing smoking before quitting, both conclude that further re-
search is needed into whether it could be used as a successful in-
tervention to help those who have tried unsuccessfully to quit in
the past. The US Medicines Regulator, and some other pharma-
ceutical regulators, have not approved the use of NRT for smokers
who wish to cut down the amount they smoke without wanting
to quit. However, the Medicines and Healthcare Regulatory Au-
thority (MHRA) in the UK and other medicines regulators have
licensed the use of NRT for this purpose. UK guidelines suggest
that, until further evidence is available, this strategy should only
be used in properl y designed and conducted research studies. New
Zealand’s smoking cessation guidelines (
NZ MoH 2007) men-
tion cutting down cigarettes smoked, but as a strategy that should
only be implemented in those unwilling to quit. Cutting down
is therefore a strategy that is either not recommended by national
guidelines or is only recommended for smokers not ready to stop.
Why it is important to do this review
Without help, most peopl e who try to stop smoking relapse within
one week and only 4% sustain abstinence for one year (
Hughes
2004
). The UK is the only country with a truly nationwide net-
work of smoking cessation clinics, although a growing number of
countries are developing a variety of free or subsidised services to
help smokers to quit. A lthough these clinics, substantially increase
rates of abstinence, most people who try to stop smoking will fail
to do so. For example, the evaluation of the UK National Health
Service (NHS) specialist stop smoking services showed that 15%
of patients achieved abstinence for a whole year (
Ferguson 2005).
Thus, while treatment substantially improves the number who
achieve abstinence, a return to smoking is the norm f or the ma-
jority, whatever method of stopping is used. Consequently there
is a cadre of patients who have been through treatment services
3Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 5
a number of times. Smoking cessation services currently recom-
mend abrupt cessation for all quit attempts (first or repeated),
whereas alternative methods might be more successful or at least
give renewed hope and encourage cessation in those who have
given this up as impossible. Gradual cessation could offer a new
way to quit for those who have failed previously, but it can only
be recommended by therapists if it is an e ffective strategy for ces-
sation.
There is evidence to suggest that some people f eel that reducing
the number of cigarettes they smoke is an important first step to-
wards quitting completely. In the English Smoking Toolkit Study,
40% of quit attempts included cutting down first (
West 2006).
A survey of respondents to an advertisement for people interested
in cessation found that 66% planned to stop by cutting down
gradually, while 13% planned to stop abruptly. In a survey of peo-
ple responding to an advertisement for those planning to reduce
smoking, 57% planned to reduce and then stop (Hughes 2006).
A sur vey of US daily smokers showed that in their most recent
quit attempt 35% tried to stop gradually while 65% tried to stop
abruptly (
Hughes 2007). Those who chose gradual cessation were
as motivated to stop and as confident of success as those who used
abrupt cessation. A random sample of US smokers showed that
nearly half of smokers planning to quit would choose reduction
over abrupt cessation (
Shiffman 2007). There was little interest
among th ese smokers in reduction as an end in itself, only as a
means to abstinence. Reducing the number of cigarettes smoked as
a means to giving up smoking may prove to be a popular approach,
and may draw people into treatment services. Given behavioural
support and pharmacotherapy increase the likelihood of achieving
abstinence (
Lancaster 2005; Stead 2005; Stead 2008) this would
have public health benefits.
O B J E C T I V E S
To compare the success of smoking cessation interventions
that instruct the smoker to reduce the amount they smoke before
quitting with interventions that instruct the smoker to stop
smoking abruptly.
To compare adverse events by arm, stratified by whethe r
they use pharmacotherapy.
M E T H O D S
Criteria for considering studies for this review
Types of studies
Randomized controlled trials. We included a trial where alloca-
tion to treatment arms was cluster randomized, and carried out a
sensitivity analysis to adjust for this clustering. To meet the second
objective we examined adverse events only in those trials which
had a reduction arm utilising pre-quit pharmacotherapy and an
abrupt quitting arm th at did not utilise pre-quit pharmacotherapy.
Types of participants
Cigarette smokers of any age who intended to stop smoking soon.
Participants demonstrated their commitment to quitting by en-
rolling in a smoking cessation programme. Trials that enrolled
smokers who did not intend to quit soon were excluded, as they are
covered by the Cochrane review of harm reduction (
Stead 2007).
Types of interventions
We compared any instruction to participants to reduce the amount
of cigarettes smoked before quitting, with any instruction to stop
smoking abruptly without prior reduction. We did not include
trials with arms where participants spontaneously reduced before
quitting without being advised to do so, versus arms whe re par-
ticipants stopped abruptly.
Interventions included anything from no behavioural support to
extensive behavioural support, but studies were excluded if be-
havioural support differed substantially in type or duration be-
tween arms. Beh avioural support pre- and post-quit could vary
between the reduction and abrupt quit arms as long as overall
contact was roughly equal. Trials could also include concomitant
pharmacotherapy to support cessation, as long as it was equivalent
in all trial arms after cessation. Pharmacotherapy used prior to quit
day could vary as a necessary component of the intervention i.e.
to support smoking reduction.
Types of outcome measures
Primary outcomes
The primar y outcome was abstinence from smoking at least six
months after the quit day. We excluded trials with a follow up of
less than six months.
In trials with more than one measure of abstinence, we preferred
the measure with the strictest criteria. We used pr ol onged or con-
tinuous abstinence over point prevalence abstinence, and preferred
biochemically validated abstinence, such as by exhaled carbon
monoxide, over self-report.
Secondary outcomes
The se condary outcome was the type and number of adverse events
recorded.
4Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 6
Search methods for identification of stud ies
We searched the Cochrane Tobacco Addiction Review Group spe-
cialised register, which has been developed from electronic searches
of MEDLINE, EMBASE and PsycINFO, together with hand-
searching of specialist journals, conference proceedings and refer-
ence lists of previous trials and overviews. We also searched MED-
LINE, EMBASE and PsycINFO for possible trials to include in
the review, searched the reference lists of relevant trials, and where
necessary contacted the authors of ongoing trials.
We searched MEDLINE (Ovid, 1966 to 5th November 2009),
EMBASE (Ovid, 1980 to 2009 week 44) and PsycINFO (Ovid,
1967 to 23rd November 2009) using the following topic-specific
terms:
cold turkey.mp
(schedul* adj3 smok*).mp
(cut* down or cut-down).mp
(({Gradual* or abrupt*}) adj3 (reduction or reduce* or quit*
or stop* or abstin* or abstain* or cessat*)).mp
fading.mp
taper*.mp
(controlled adj smoking).mp
[mp=title, original title, abstract, name of substance word, subject
heading word]
We combined these with the terms used for the regular searches of
MEDLINE, E MBASE and PsycINFO to identify trials of tobacco
addiction interventions for the Tobacco Addiction Review Group
specialised register. Full strategies are shown in the Appendices.
We also searched the specialised register in November 2009 using
the foll owing ter ms: Cold turkey or schedul* or Cut* down or cut-
down or Gradual* or abrupt* or fading or reduction or reduce* or
taper* or controlled smoking.
Data c ollection and analysis
Selection of studies
One author checked the titles and abstracts of studies generated
by the search strategy for relevance, and obtained copies of papers
reporting relevant trials. Two authors then independently assessed
the reduced trials l ist for inclusion in the review. Any disagree-
ments were resolved through discussion with the remaining review
author. We based eligibility decisions on the following questions:
1. Is the study described as randomized or quasi- randomized?
2. Were the participants cigarette smokers who wanted to quit?
3. Did the study include at least two groups, i.e. one group
advised to reduce their smoking before quitting and one advised
to quit abruptly on quit day?
4. If the intervention includes behavioural support with or
without pharmacotherapy, is overall contact for behavioural
support and post-quit pharmacotherapy similar between both
groups?
5. Is the intervention an instruction to reduce the number of
cigarettes smoked, rather than an instruction to reduce harm,
e.g. smoking cigarettes with lower levels of nicotine?
6. Does the study report smoking abstinence at least six
month after the quit date?
If the answer to any of the above questions was ’No then the trial
was not included in the review.
Data extraction and management
For each included trial one author e xtracted the data and another
author checked the m. The only included paper published in Span-
ish was translated into English (
Roales-Nieto 1992). We extracted
the following information for inclusion in the
Characteristics of
included studies
table:
Methods:
The design of the trial, for example randomized or quasi-
randomized.
Country and setting
Method by which participants were selected
The definition of a smoker
Duration of the study
Time to follow up(s)
Participants:
The number of participants randomized to each
intervention group
Demographics of participants (age, gender, ethnicity)
The average number of cigarettes pe r day, and number of
past quit attempts
Average Fagerstrom Test of Nicotine Dependence (FTND)
or equivalent score
Particular preference for abrupt or gradual cessation
Interventions:
The method of rapid reduction intervention used
The method of abrupt quitting intervention used
Whether pharmacotherapy was used as part of the
intervention, and if so details of use
Details of any behavioural support provided
Duration of reduction period
Who delivered the intervention?
5Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 7
Outcomes:
Did the tr ial examine whether the reduction arm reduced as
instructed, and that the abrupt arm did not reduce?
Outcomes measured
The strictest definition of abstinence used
Whether abstinence was biochemically verified, and if so,
how
whether enough data are available for an intention-to-treat
analysis
the proportion of quitters in each intervention arm
the number of adverse events in each arm
Amount of reduction in cigarettes per day in each arm (self
report and/or chemical biomarkers)
Additional outcome results
Drop-out rates
Information about withdrawals
Further information about adverse events
Missing data in both arms.
Assessment of risk of bias in included studies
Risk of bias for each trial was assessed within the domains of se -
quence generation, allocation concealment and incomplete out-
come data, using the risk of bias table, as outlined in the Cochrane
Collaboration Handbook (
Handbook 2008).
Measures of treatment effect
We compared quit rates between the abrupt cessation and reduc-
tion groups, calculated on an intention-to-treat (ITT) basis, in-
cluding all participants originally randomized to a trial arm. Any
participants lost to follow up were treated as relapsed, excluding
any deaths. We used relative risk as the summary statistic in any
meta-analyses, using the Mantel-Haenszel fixed-effect model for
pooling results, checking for no significant heterogeneity. We also
compared the number of adverse events between arms, however no
meta-analysis was carried out for this outcome as data was sparse
and not consistently measured across studies.
Assessment of heterogeneity
Any inconsistency across study results was identified and assessed
by examining forest plots for poor overlap of confidence intervals,
and by examining the I-squared statistic.
Subgroup analysis and investigation of heterogeneity
We conducted sub-group analyses comparing trials which used
pharmacotherapy as part of the interventions with those that did
not. We also grouped interventions by whether or not the instruc-
tion on how to quit smoking was given alongside behavioural sup-
port or by self-help methods.
Sensitivity analysis
We tested study design by investigating the sensitivity of the main
effect, when adjusting for the only cluster randomized trial eligible
for inclusion in the meta-analysis and when excluding the studies
where non-validated self-report data was used for the meta-analy-
sis,
R E S U L T S
Description of st udies
See:
Characteristics of included studies; Characteristics of excluded
studies
; Characteristics of ongoing studies.
See Characteristics of included studies, Characteristics of excluded
studies
, Characteristics of ongoing studies.
Results of the search
The searche s of the Cochrane Specialised Register, MEDLINE,
EMBASE and PsycINFO resulted in 543 unduplicated references.
Additionally one of the authors of this review has just completed
a study comparing reduction to abrupt quitting and has written a
study report (
Hughes 2009), which cited two further studies pos-
sibly relevant for inclusion. These 543 references were screened for
eligibility based on their titles and abstracts, resulting in a reduced
total of 30 studies. These studies were then independently assessed
by two authors for eligibility, based on the questions specified
above. We found 10 studies which were relevant for inclusion in
the review based on these criteria; seven of these took place within
the United States of America (
Flaxman 1978; Cummings 1988;
Curry 1988; Cinciripini 1995; Jerome 1999; Rile y 2005; Hughes
2009
), th e remaining three were situated in Austria (Gunther
1992
), Switzerland (Etter 2009) and Spain (Roales-Nieto 1992).
We also discovered three ongoing studies (
Riley 2001; Cinciripini
2006; Lindson 2009) which, when completed, may also be rele-
vant for inclusion. The authors of eight studies (
Cummings 1988;
Curry 1988; Jerome 1999; Riley 2001; Riley 2005; Roales-Nieto
1992
; Cinciripini 2006; Etter 2009) provided additional informa-
tion when contacted.
Included studies
Characteristics of participants
The 10 included studies all recruited adult cigarette smokers with
an aim to quit. Seven studies recruited participants from the com-
munity using advertisements (
Flaxman 1978; Cummings 1988;
Curry 1988; Cinciripini 1995; Riley 2005; Etter 2009; Hughes
6Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 8
2009). One study recruited work-sites to take part and then re-
cruited their employees by posting advertisements and internal
memos (
Jerome 1999). Another recruited students usingadvertise-
ments at a university (
Roales-Nieto 1992), and another recruited
patients consulting a hospital based smoking counselling service
(
Gunther 1992).
In one study these participants were then randomized in clusters
(work-sites) to study arm (
Jerome 1999), however for all other
included studies participants were individually randomized. In the
eight studies where participant gender was repor ted participants
were on average evenly split between males and females, and the
average reported age of participants (averaged across seven studies)
was 42.8 years. Eight studies reported average baseline cigarettes
per day in all participants, and this ranged from 23 to 28 cigarettes
per day, with an average of 25.4.
Sample sizes
The total sample size across the 10 included studies ranged from
23 to 1895, with a mean sample size of 487. However not all con-
ditions in all of the studies were used in the meta-analysis. When
only the conditions relevant to this review were taken into account,
sample sizes ranged from 14 to 1277, with a mean of 376. In five
of the included studies all conditions randomized were relevant
to the current review and were therefore included in the meta-
analysis, however five of the studies randomized participants to
interventions which were not relevant.
Cummings 1988, Jerome
1999
and Hughes 2009 all included a control condition, which
did not provide specific advice on how to quit, but provided in-
formation about the health implications of smoking, praise for
quitting, and material emphasizing the importance of a general
program of phy sical health (including quitting smoking) respec-
tively. Flaxman 1978 included an immediate quit condition where
participants were asked to quit the day after enrolling in the study
and received substantially less behavioural support then the other
conditions.
Roales-Nieto 1992 included two conditions where the
participants goal was to reduce their smoking and control it rather
than to reduce and quit completel y. All of these conditions were
deemed not relevant to this review and were excluded from any
meta-analyses.
Interventions
All of the included studies had at least one group of participants
who were instructed to reduce the amount they smoked before th ey
quit, and at least one group instructed to quit smoking abruptly.
In four of the studies, participants were advised on either abrupt
or gradual cessation by self-help manuals or a handheld computer
programme (Cummings 1988, Jerome 1999, Riley 2005, Etter
2009
). Participants in another five studies were given face-to-face
(Flaxman 1978, Gunther 1992, Roales-Nieto 1992; Cinciripini
1995
) or telephone based (Hughes 2009) behavioural support as a
means to assist either reduction or abrupt cessation. In the remain-
ing study one reduction arm and one abrupt arm consisted of self-
help therapy, and participants in the other reduction and abrupt
arms were provided with behavioural support (
Curry 1988). The
behavioural support varied in terms of the overall length of time
for which support was provided, the length of support sessions,
number of support sessions, whether these were provided to indi-
viduals or groups, and who provided the support, however they all
included pre-quit sessions where participants were taught strate-
gies to help them avoid smoking when tempted, such as strate-
gies to maximise self-control, and post-quit sessions focusing on
relapse prevention. Most of the se lf-help interventions consisted
of information booklets, some of which provided the participants
with written activities. However the reduction interventions in
Jerome 1999 and Riley 2005 gave participants the LifeSign hand-
held computer (PICS Inc); LifeSign structures a gradual reduction
schedule, prompts users to smoke and allows them to record each
cigarette they smoke. In the
Jerome 1999 study this computer was
provided, with a 48 page manual, which consisted of instructions
on how to use the computer and information about behaviour
modification strategies and relapse prevention. In the
Riley 2005
study participants only received brief instructions on how to use
the device and no further information. This was designed as a min-
imal contact intervention, which matched the minimal instruc-
tions provided to the abrupt quitting intervention group mem-
bers, who received a cale nder log to record their smoking.
The abrupt quitting method advised for participants did not vary
much across the ten studies. Participants were either given a quit
date or asked to choose one themselves, and then asked to smoke
as normal and quit abruptly on this date, with no prior cutting
down. Quit dates ranged from zero to five weeks fol lowing base-
line assessment. The smoking reduction interventions were more
varied across studies as follows, h owever all reduction methods
culminated in a quit day:
Cummings 1988 gave participants unspecific advice on how
to quit; they were simply advised to reduce the amount smoked
over two weeks before quitting. Suggestions were provided on
how they could reduce, such as setting daily goals, switching
brands, changing habits and delaying the first cigarette; but
ultimately it was left to participants to choose by how much to
reduce and which, if any, strategies to use to achieve this.
Three studies asked participants to reduce cigarettes per day
by a certain quota over a set time interval without providing
participants with any particular strategy to do so.
Etter 2009
asked participants to reduce th eir smoking to 50% of baseline
over four weeks and then quit completely.
Gunther 1992 asked
participants to reduce their smoking by five to ten cigarettes per
week, depending on how much they were smoking at baseline,
over five weeks until they were not smoking at all.
Roales-Nieto
1992
instructed participants to reduce by 25% of baseline in
week one, 50% in week two, 75% in week three and to quit
completely in week four.
7Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 9
In the Cinciripini 1995 study two groups of participants
were asked to reduce their smoking; one of the groups reduced
smoking by a set quota but did not use a specific technique to
achieve this, as in the three studies above. Participants cut down
to 66% of their baseline smoking rate in the first week of
reduction, to 33% of baseline in the second week, and to 22% of
baseline in the third week, until they reached two to four
cigarettes per day. The second reduction group reduced by the
same quota of cigarettes, but this was structured. Each week the
advised smoking rate was divided by the number of hours in the
participants waking day to calculate an inter-cigarette interval.
Participants were then able to smoke only in the first five
minutes of each interval, and any missed cigarettes could not be
accumulated for later use. Both groups quit in the week
following the third week of reduction, and were combined for
the purposes of our meta-analysis.
Jerome 1999 and Riley 2005 also used inter-cigarette
intervals to reduce smoking to nil. They implemented this using
a handheld computer called LifeSign, which developed a
smoking reduction schedule, lasting between 10 to 28 days,
depending on each individuals baseline smoking rate and
progress th r ough the programme. The mach ine beeped and put a
reminder on its screen to prompt participants to smoke.
Hughes 2009 advised participants to reduce their smoking
by 25% of baseline in week one, 50% in week two and 75% in
week three, before quitting completely. They were also provided
with four structured ways to do this, which they could choose
between. The first was scheduled reduction where participants
were advised to gradually increase the time between cigarettes
(the inter-cigarette interval). The second asked participants to
rate each cigarette of the day in terms of how difficult it would
be to give up and then eliminate each in turn starting with the
most difficult first. The third was the same as the second but
participants started with the easiest first. The fourth involved the
participant increasingly de laying the time from waking to the
first cigarette of the day. Abstinence results did not appear to
differ across the methods and so the data was pooled.
Flaxman 1978 differed from the previous approaches as
participants were not asked to reduce by a certain quota of
cigarettes, but to identify situations that caused them to smoke.
They were then asked to rate these situations in terms of how
difficult it would be to abstain from smoking and then to
eliminate smoking in one situation every three days, starting
with the easiest situation and proceeding to the most difficult. In
one reduction group participants continued this until they were
not smoking at all and in the other they reduced until they were
smoking in 50% of their baseline smoking situations and then
quit abruptly. These two reduction groups were combined into
an overall reduction group in our me ta-analysis.
One study gave very limited information as to how
reduction took place (
Curry 1988); the method was described as
cigarette tapering and a gradual acquisition of coping skills. The
author confirmed that this was a reduction method relevant for
inclusion in this review, however no further detail could be
provided.
Pharmacotherapy
Three of the studies included in this review gave participants phar-
macotherapy as a part of their interventions. In all cases this was
in the form of nicotine replacement therapy (NRT); one study
used gum (
Etter 2009), another lozenges (Hughes 2009) and the
third nasal spray (
Riley 2005). In the reduction arm of each study
participants used the NRT both pre- and post- quit, and in the
abrupt quitting arm post-quit only. In the pre-quit period
Etter
2009 advised participants to use at least 10 piece s of 4mg nico-
tine gum per day,
Hughes 2009 requested that participants replace
each cigarette missed with a 2mg or 4 mg lozenge (4mg for those
who smoked within 30 minutes of waking and 2mg for others).
Riley 2005 signalled when participants should use the nasal spray
using the same LifeSign handheld computer as was used to signal
smoking. The appropriate nasal spray dosage was determined for
each individual user depending on their recorded baseline smok-
ing rate.
Outcomes
Nine of the 10 studies reported smoking abstinence as an outcome
at either six month follow-up (
Flaxman 1978; Cummings 1988;
Hughes 2009), 12 month follow-up (Gunther 1992; Curry 1988;
Etter 2009) or both (Cinciripini 1995; Jerome 1999; Riley 2005).
The remaining study (
Roales-Nieto 1992) reported cigarettes per
day over seven days at six month, nine month and 12 month fol-
low-ups for individual participants; it was possible to calculate ab-
stinence rates from this information. Where abstinence was mea-
sured at six and 12 month follow-ups the 12 month r ates were
used in the meta-analysis. In three studies smoking abstinence
was reported as point prevalence (
Roales-Nieto 1992; Jerome
1999
; Riley 2005), and in six studies as prolonged/continuous
(
Cummings 1988; Curry 1988; Gunther 1992; Cinciripini 1995;
Etter 2009; Hughes 2009). Flaxman 1978 did not report how ab-
stinence was defined. Abstinence was verified in eight of th e in-
cluded studies, by either expired carbon monoxide (
Jerome 1999;
Riley 2005; Etter 2009; Hughes 2009), saliva cotinine (Cinciripini
1995
; Etter 2009), saliva thiocyanate (Curry 1988), or asking a
relative or friend to confirm the participant had stopped smok-
ing (
Cummings 1988; Roales-Nieto 1992). However verified data
were not used for one of the studies (
Cummings 1988), as there
were pr oblems with the naming of a friend or relative to verify par-
ticipants self report. If participants did not name a person to ver-
ify their self-report, or if their self-report contradicted with their
friend/relatives then they were classed as smoking, however 20%
of those claiming abstinence did not provide a friend/relative. Par-
ticipants who lived alone were four times more likely not to name a
8Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 10
person for verification than th ose who lived with others. All of the
study reports either reported ITT analysis or provided sufficient
information to allow us to calculate this, apart from
Cummings
1988
where the author provided this information when contacted.
Only two studies (
Etter 2009, Hughes 2009) reported information
about adverse events. Further information was obtained from these
authors and some limited information was also obtained from the
authors of
Riley 2005. Reporting was not consistent across studies
and so it was not possible to carry out a meta-analysis, therefore
these data are synthesised qualitatively.
Excluded studies
Studies which were identified as potentially relevant but later ex-
cluded are listed, with reasons for exclusion, in the
Characteristics
of excluded studies
table. The primary reasons for exclusion f ell
into one of three categories: 1) The goal of the intervention was to
reduce smoking and control it, rather than quit (
Hatsukami 1988;
Bolliger 2000b), 2) the main outcome was smoking rates, and it
was not possible to calculate abstinence rates from the data pre-
sented or to get these from the authors (
Marston 1971), 3) both
of the trial arms quit in the same way (
Bernard 1972; Glasgow
1989; Cinciripini 1994; Herrara 1995; Daughton 1998; Jerome
& Fiero 1999;
Rose 1998; Bolliger 2000a; Schuurmans 2004;
Rose 2006; Rezaishiraz 2007; Bullen 2008; Rose 2009 ; Shiffman
2009
). Five of the excluded studies examined pre-treatment with
NRT vs placebo prior to the quit date, and did not instruct smok-
ers to reduce pre-quit. Three of these reported that participants
spontaneously reduced whilst using the NRT (
Rose 1998; Rose
2006
; Rose 2009), two of which found that participants who re-
duced their smoking the most were more likely to achieve absti-
nence (
Rose 2006; Rose 2009). However, as none of the studies
instructed subjects to reduce their smoking during the pre-cessa-
tion ph ase of the treatment, this success cannot be attributed to
an instruction to reduce and so the studies were excluded.
Risk of bias in included studies
We extracted information from each study to assess the risk of
biased randomization, whether allocation concealment took place,
and whether incomplete outcome data was addressed. This was
assessed as either likely to cause bias (No), unlikely to cause bias
(Yes) or unclear, if insufficient information was present to make a
judgment (
Figure 1).
9Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 11
Figure 1. Methodological quality summary: review authors’ judgements about each methodological quality
item for each included study.
10Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 12
Randomization sequence generation. Five studies reported
adequate information on sequence generation to be classified as
having minimal chance of bias in this regard. Five of the studies
(
Flaxman 1978; Roales-Nieto 1992; Cinciripini 1995; Jerome
1999
; Riley 2005) did not describe the method of randomization
used, and so were classified as unclear in this category. All of the
studies randomized individual participants, apart from
Jerome
1999
who randomized work-sites to trial arms. Trials that
randomize clusters to treatment arms can be given a higher
weighting than they should if data on individuals are entered in
the meta-analysis. This is because the analysis assumes there is no
connection between individuals in the same group in the
likelihood of them stopping smoking successfully. However
when we carried out an analysis adjusting for the clustering in
Jerome 1999, although the study weighting decreased from 22%
to 17% the main result was not sensitive to the adjustment (Risk
ratio (RR)= 0.93, 95% CI= 0.78 to 1.12). This adjustment used
an intra-class correlation of 0.0105 (as recommended by
Martinson 1999 for an outcome of percentage quit in the work
place) and an average number of people per group of 18.3
(design effect = 1.18).
Allocation concealment. When rated in terms of allocation
concealment from clinicians enrolling participants into studies,
four studies (Cummings 1988; Jerome 1999; Riley 2005; Etter
2009
) were rated as unlikely to cause bias, as all interventions
consisted of self-help therapy and there was either no or minimal
contact with investigators/enrolling clinicians. Consequently,
participants enrolment in the studies could not depend on
knowledge of the allocation sequence as there was no clinician
deciding on whether to enrol or which treatment to give. Hughes
2009
was also rated as unlikely to cause bias in this category as a
statistician generated a concealed allocation sequence. The five
remaining studies did not report on allocation concealment and
were therefore classed as unclear.
Incomplete outcome data. In the category of incomplete
outcome data six studies were classed as unlikely to cause bias, as
participant attrition was reported as similar in all trial arms. The
four remaining studies were classed as unclear; three of the
studies (
Flaxman 1978; Cinciripini 1995; Riley 2005 ) did not
provide any information about participant attrition or missing
data, and the abstinence rates table in the
Cummings 1988
report appeared to leave 18 participants unaccounted for. Due to
the length of time since the study had been completed the author
could not confirm why this was th e case, but did provide further
information so that an intention-to-treat analysis could be
carried out, in which the missing participants were classified as
not abstinent. Participants attrition in general was similar across
arms, however the study was classified as unclear as we didnt
know the allocation of the missing participants and whether this
was similar across arms.
Two of the included studies (
Flaxman 1978; Cinciripini 1995)
were rated as unclear for all th ree of the above bias categories
and another two were rated as unclear for two (
Roales-Nieto
1992; Riley 2005). We carried out a sensitivity analysis to establish
whether the main result was sensitive to th e exclusion of th ese four
studies and found that it was not (RR= 0.94, 95% CI= 0.76 to
1.16).
Other potential sources of bias were failure to verify smoking absti-
nence by biochemical means and whether participants conformed
to their allocated intervention.
Biochemical verification. Studies that did not validate self-
reports of abstinence (
Flaxman 1978; Gunther 1992), or where
validation was potentially flawed, and th erefore not used in this
review (
Cummings 1988) could potentially over-estimate
abstinence. However we would not expect this to differ between
arms, and a sensitivity analysis confirmed that the main findings
were not se nsitive to the exclusion of studies where abstinence
was not validated (RR= 0.91, 95% CI= 0.74 to 1.12).
SRNT
2002
concludes that population based studies with limited face-
to face contact, and where data collection is optimally by mail,
telephone, or on the Internet are unlikely to benefit from
biochemical verification. Population studies have much higher
biochemical verification refusal rates than clinic based studies, if
all participants who refused were classed as smoking then this
would be likely to overestimate smoking rates. In reality the
extent that self-reports inflate abstinence rates is small and rarely
differs across conditions. Also, in studies where there is very littl e
contact with an investigator or therapist this reduces demand
characteristics, meaning there is little incentive to lie.
Adherence to method of quitting allocated. Six of the 10
studies assessed whether participants followed the instructions
they had been given on how to quit i.e. to reduce or quit abruptly
without prior reduction. Three of these studies (
Roales-Nieto
1992
; Etter 2009; Hughes 2009) found that participants
followed instructions; the participants in the reduction group
reduced before quitting and the participants in the abrupt group
quit abruptly with no prior reduction.
Cinciripini 1995 found
that the reduction group complied well with their instructions
but the abrupt group also reduced by seven to eight cigarettes per
day less than baseline before quitting. However the reduction
group smoked significantly fewer cigarettes than the abrupt
group before quit day. The two remaining studies to report on
adherence to the intervention allocation found that participants
did not abide by intervention instructions. In
Flaxman 1978 the
group which reduced until they were not smoking at all reduced
by a mean of 6 cigarettes per day, and the group who reduced to
50% of baseline then quit reduced by a mean of 3.5 cigarettes
per day. However the abrupt quit group also reduced by an
average of 3.4 cigarettes per day before th ey quit, meaning there
11Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 13
was little difference between reduction in the partial reduction
group and the abrupt quit group.
Cummings 1988 asked
participants after quit day whether they had quit abruptly. 39%
of participants in the abrupt group quit abruptly and 40% of the
reduction group also quit abruptly, therefore there appeared to
be l ittle difference between the arms in the methods of quitting
that were actually used. As is the case with all IT T analyses, it is
only ever possible to examine the effect of allocation to a quitting
method, not the effectiveness of actually following it.
Effects of interventions
Abstinence Outcome
The meta-analysis included 10 trials with a total of 3760 partic-
ipants. There was evidence that reduction produced similar quit
rates to abrupt cessation and that any difference in effectiveness
was small. The overall rate ratio for abstinence for reduction ver-
sus abrupt cessation was 0.94, 95% CI= 0.79 to 1.13 (Figure 2).
There was low heterogeneity (I² = 14%), suggesting that the e ffect
of reduction relative to abrupt cessation did not differ across tri-
als. For all studies confidence intervals spanned one, indicating no
study achieved statistically significant superiority of either gradual
or abrupt cessation. We have not reporte d pooled quit rates be-
cause studies varied on a number of factors, such as definition of
abstinence (point prevalence or prolonged), length of abstinence
(6 months or 12 months), whether or not behavioural support was
provided , and whether pharmacotherapy was provided, meaning
that average rates would not be useful.
Figure 2. Reduction to quit versus abrupt quitting. Outcome: abstinence
The effect of gradual versus abrupt cessation in participants
using pharmacotherapy
The studies were split into two sub-groups to assess whether the ef-
fect of gradual cessation depended on whether people used smok-
ing cessation pharmacotherapy or not. One sub-group included
studies that didnt use any pharmacotherapy as part of the inter-
ventions (
Flaxman 1978; Cummings 1988; Curry 1988; Gunther
1992
; Roales-Nieto 1992; Cinciripini 1995; Jerome 1999). The
other sub-group included the remaining studies (
Riley 2005; Etter
2009
; Hughes 2009), which utilised nicotine replacement therapy
pre- and post-quit in the reduction interventions and post-quit in
the abrupt interventions. There was no evidence of the superiority
of either gradual or abrupt cessation whether pharmacotherapy
(NRT) was used (RR= 0.89, 95% CI= 0.65 to 1.22), or not (RR=
0.97, 95% CI= 0.78 to 1.21), and neither was there evidence that
pharmacotherapy modified the effect of reduction versus abrupt
cessation (Analysis 1.2).
The effect of the type of behavioural support utilized
We also conducted a sub-group analysis by the type of th er-
apy provided. Some of the included studies used self-help ther-
apy (Cummings 1988; Jerome 1999; Riley 2005; Etter 2009),
and some behavioural support (
Flaxman 1978; Gunther 1992;
Roales-Nieto 1992; Cinciripini 1995; Hughes 2009). Curry 1988
included study arms that were self-help and others that were be-
havioural, so these were split accordingly for the sake of this analy-
sis. Again the risk estimates were similar whether the instruction in
how to quit and support for achieving this was given by self-help
(RR= 0.98 95% CI= 0.78 to 1.23) or by beh avioral support (RR=
12Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 14
0.87 95% CI= 0.64 to 1.17), and neither reduction nor abrupt
quitting resulted in superior quit rates in either case (
Analysis 1.3).
Adverse Events Outcomes
The secondary objective of this review was to compare adverse
events between arms, however no attempt has been made to do this
quantitatively as there was a a lot of variation in nature and depth of
reporting. The seven studies that did not utilise pharmacotherapy
did not report information about adverse events. Of the three stud-
ies using pharmacotherapy,
Riley 2005 reported no information
on adverse events in the study report, but the author kindly sup-
plied further information for this review.
Etter 2009 and Hughes
2009
also provided additional information as well as data reported
in the publications.
Etter 2009 and Riley 2005 reported that no
participants in these studies experienced serious adverse events.
Etter 2009 also provided data obtained in response to the question:
“If you experienced undesirable effects due to the nicotine gum,
please describe them (open e nded question), asked two months
after target quit day. Overall the most commonly reported symp-
toms were mouth pain/dry mouth/throat burns, hiccups, stomach
pain/heartburn- the most common side-effects from oral NRT.
Nine of the total symptoms reported occurred more frequently
in the reduction groups (mouth pain/dry mouth/throat burns,
hiccups, stomach pain/heartburn, pain/cramp in jaws, mouth ul-
cers, headache, eructation, heart palpitations, cough) four in the
abrupt groups (nausea, bad taste, insomnia, vomiting) and three
were reported as frequently in both groups (malaise, constipation,
diarrhoea).
Hughes 2009 reported that the incidence of adverse
events rated severe was small and similar across conditions. 3% of
participants randomized to the reduction to quit group reported
severe adverse events and 5% of the abrupt quit group; the inci-
dence of discontinuation was 1% for both groups.
D I S C U S S I O N
The 10 studies included in this review compared interventions
that instructed participants to quit smoking gradually by reducing
the amount they smoke with interventions that instructed par-
ticipants to quit smoking abr uptly without prior reduction. The
results provide evidence that reduction to quit provides similar
quit rates to abrupt quitting with no evidence that one method is
significantly superior to the other in adults trying to quit smoking.
This applies whether therapy is self-help or includes behavioural
support and wheth er the quit attempt uses NRT or not. The sim-
ilarity of the result in the NRT sub-group and the non NRT sub-
group suggests that the success of the reduction interventions rel-
ative to the abrupt quit interventions is not due to the use of pre-
quit NRT. We were unable to combine data on absolute quit rates
as studies varied on a number of factors expected to influence quit
rates, for example length of follow-up, and so can not provide
meaningful estimates of average quit rates as a result of reduction
to quit and abrupt quit interventions.
We were unable to combine statistically the adverse events data,
and therefore could not determine whether adverse events differed
significantly between the intervention groups that reduced and
used NRT pre- and post-quit, and the intervention groups whe re
participants quit abruptly and used NRT post-quit. However a
recent review conducted a meta-analysis (
Moore 2009) of seven
placebo controlled RCTs, which used NRT to assist reduction to
stop smoking and found that there were no significant differences
in deaths (odds ratio (OR)= 1.00, 95% CI= 0.25 to 4.02), seri-
ous adverse events (OR= 1.16, 95% CI= 0.79 to 1.50), and dis-
continuation due to adverse events (OR= 1.25, 95% CI= 0.64
to 2.51), between the placebo and NRT interventions. The only
adverse event that was more common in the NRT interventions
was nausea (OR= 1.69, 95% CI= 1.21 to 2.36), which is a com-
mon side effect of nicotine replacement therapy. Taken with othe r
safety data on concurrent smoking and use of NRT (
Fagerstrom
2002), there appears to be no reason to recommend against the
practice of gradual reduction assisted by NRT. At least one trial
shows that among smokers trying to quit smoking by gradual re-
duction, using NRT is more effective than use of placebo in sup-
porting abstinence (Shiffman 2008), and a Cochrane Review of
smoking harm reduction (
Stead 2007) found that people who did
not originally want to quit smoking were more likely to be absti-
nent from cigarettes at long-term follow-up when NRT was used
as an aid to reduction than when a placebo was used (OR= 1.90,
95% CI= 1.46 to 2.47). On this basis, if reduction is to be used
as a means of quitting, use of NRT or other pharmacotherapy
appears desirable. NRT is licensed for use in this way in the UK
and Australia, however the US Medicines Regulator, along with
other pharmaceutical regulators have not yet licensed NRT for
this purpose.
An important limitation of any meta-analysis is that methods vary
across studies and the underlying assumption that the meta-anal-
ysis is trying to estimate a single true rate ratio might not hold.
In this instance patient populations, outcome definitions, provi-
sion of pharmacotherapy and the behavioural support provided
varied across the included trials. Despite this, the measure of het-
erogeneity was low suggesting that heterogeneity of these elements
did not translate into heterogeneity of effectiveness of reduction.
One of the studies also varied because it used cluster randomiza-
tion, but sensitivity analysis suggested that allowing for this or not
had little influence on the result of the meta-analysis. Four of the
studies included in the meta-analysis (
Flaxman 1978; Cummings
1988
; Curry 1988; Cinciripini 1995) had more than one inter-
vention th at qualified as reduction and/or abrupt quitting, and
we combined these to create one reduction arm and one abrupt
quit arm per study. We considered entering th e data for each trial
arm separately to see if this would give us any more detailed in-
formation about the relative success of different reduction meth-
13Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 15
ods, however the methods used differed in each study (scheduled,
non scheduled, group behavioural support, individual behavioural
support, reduction to zero cigarettes before quit, reduction to 50%
of baseline before quit etc), so they could not be pooled for a sub-
group analysis, and therefore would be no more informative than
the original studies.There is however some evidence that struc-
tured methods of reduction are more effective than simple advice
to cut down without following specific methods (
Levinson 1971;
Cinciripini 1995).
Two of the 10 included studies (
Etter 2009; Hughes 2009) were
assessed as unlikely to cause bias for all three categories assessed.
These studies were the most recent of th e 10 studies, which may
suggest that their increased reporting, relative to the other e ight
studies, is due to awareness of the revised CONSORT reporting
guidelines (Moher 2001), which were published in 2001, and ad-
vise reporting methods of sequence generation and allocation con-
cealment, and a flow diagram illustrating the flow of participants
through the study. Seven of the studies rated in at least one bias
category as ’No or ’Unclear were published before 2001 and the
remaining study was not written up for publication. Therefore lack
of reporting may be for these reasons rather than because bias is
present. This may also e xpl ain why the reporting of adverse events
was only present in the most recent studies. The main results of the
two most recent studies do not differ much from the main results
of the eight olde r studies, therefore there is no evidence that stud-
ies reporting better randomization procedures produced different
results. Many of the older studies did not propose a hypothesis
that favoured eithe r a reduction or an abrupt quitting intervention
(
Flaxman 1978; Curry 1988; Gunther 1992; Roales-Nieto 1992;
Cinciripini 1995) so in the cases where allocation concealment
was not reported, and so may not have occurred, there is no reason
why allocation would have been carried out to favour any partic-
ular arm.
Whilst assessing studies for eligibility there were two studies (
Curry
1988
; Jerome 1999) where uncertainty arose about whether the
intervention methods were abrupt or reduction to quit.
Curry
1988
reported that one method of quitting used in the study was
“cold turkey” and that the other was tapering and nicotine fad-
ing”. There is no f urther detail given on these methods so we
contacted the authors who confirmed that one of these methods
was an abrupt quit method and the other was a reduction to quit
method, which met our inclusion criteria.
Jerome 1999 consisted
of a study arm where participants reduced and then quit using a
handheld computer, and an arm where participants were provided
with an American Lung Association self-help booklet called “Free-
dom From Smoking For You and Your Family”. The study report
did not specify whether this booklet advised an abrupt quitting
method or a reduction to quit method and the authors and the
American Lung Association were unable to provide additional in-
formation. However
Davis 1992 includes a table comparing the
content of three self-h elp guides including this one, which re-
ported that the topic of cutting down smoking is not covered. We
therefore believe that including these studies in the review is ap-
propriate. The failure of studies to clarify methods used to achieve
abstinence does raise the possibility that studies could have been
missed because authors described them in terms we did not expect.
We f ol lowed up included studies reference lists to check for other
studies that came up in our search and we found no other studies.
Nevertheless, we could have failed to include all extant studies but
there is no reason why publication bias or failure to find less clearly
described or l ess pr ominent studies would be expected to bias the
results towards reduction or abrupt cessation methods.
Surveys carried out in the general population (
West 2001; Cheong
2007
) have found that gradual quitting isnt as effective as abrupt
quitting, however these differ from the RCTs included in this
meta-analysis in ways that may explain the difference in outcomes.
The participants quitting gradually in the RCTs (whether support
was behavioural or self-help) were all pr ovided with some instruc-
tions as to how to quit, which included setting quotas of cigarettes
to reduce by, and setting time intervals at which participants could
smoke. All of the included studies also appeared to require partic-
ipants to set a target quit day providing them with a goal to work
toward. However the participants included in the observational
studies will have quit using a number of methods of gradual reduc-
tion, and it is likely that these will vary in their levels of success.
The UK and US national guidelines do not recommend cutting-
down before quitting and therefore services such as the UK NHS
Stop Smoking Services (NHS SSS ) only offer abrupt quitting as a
cessation method. This means that those participants who chose
gradual cessation were less likely to have benefited from any kind of
support whilst quitting (behavioural or self-help materials), which
in the case of the NHS SSS has been found to increase quit rates
by up to four times. Therefore quitters choosing gradual reduction
are automatically put at a disadvantage. A person who quits with-
out support is also more likely to use an unstructured method,
with no reduction goals, no particular method of reducing, and
no target quit day.
Cinciripini 1995 found that those participants
that quit using unstructured reduction were less successful than
those who used a more structured method. Two previous meta-
analyses (
Law 1995; USPHS 2008) have looked at nicotine fading
as a smoking cessation intervention. These, however, differ from
the current analysis, because as well as including studies where
participants were asked to reduce nicotine intake by reducing the
number of cigarettes they smoked, they also included studies where
participants were asked to use graduated filters to remove progres-
sively more nicotine from inhaled smoke, and studies where par-
ticipants changed brands to cigarettes of successively lower nico-
tine yield. We ch ose not to combine all of these approaches in the
current analysis as there is reason to believe that these methods
do not all work by the same mechanisms. For example, one of
the ways reducing cigarettes smoked may work is by weakening
links between environmental cues (e.g. socialising) and smoking
a cigarette. This wouldnt be applicable to using nicotine filters as
14Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 16
the person is still smoking in all the same situations and therefore
still associates smoking with the same environmental cues. One
of the reviews (
Law 1995) compared the gradual quitting inter-
ventions with sudden or abrupt cessation as we have done in this
case, however the second (
USPHS 2008) compared nicotine fad-
ing with untreated control conditions and therefore the relative
effectiveness of reducing nicotine intake and abrupt quitting was
not reported.
USPHS 2008 found that the re was no effect of us-
ing nicotine fading techniques when compared to no treatment,
however
Law 1995 found that gradual cessation was 5% (95%
CI= -2% to 11%) more effective than abrupt quitting, although
this difference was not significant (p>0.10). Therefore as in this
analysis neither abrupt quitting or reducing to quit provided su-
perior quit rates.
The result of this analysis suggests that public health messages on
cessation and cessation services supporting individuals who smoke
could advocate or offer reduction as a way to quit for people who
intend to quit soon. They can be confident that if people choose
to quit by reducing before stopping entirely, this would not put
them at a disadvantage compared with those who choose to smoke
as normal and then quit abruptly. Reduction to quit might help
those who h ave tried to quit a number of times without success
and are disillusioned with the abrupt quit method. Having a new
way to quit could give renewed hope, especially as many smokers
see reduction as an intuitive first step toward stopping smoking
completely. Offering reduction to quit may also appeal to those
who would otherwise not have sought behavioural support and
pharmacotherapy because the y want to pursue gradual ce ssation
and this is not currently supported. This would then enhance the
proportion of the population that make assisted quit attempts and
boost population ce ssation rates . Without help, around 4% of
people who try to stop smoking sustain abstinence for one year
(
Hughes 2004), but with the aid of cessation treatment in the
UK, around 15% of quitters abstain for a year (
Ferguson 2005).
This increase in success when behavioural support is provided
suggests that we should be trying to encourage as many people as
possible into cessation services. O ur sub-group analysis, however,
suggests that reduction is as successful as abrupt quitting whether
the intervention consists of behavioural support or is sel f-help.
Therefore this result could also benefit people who want to quit
smoking on their own without behavioural support. If people who
smoke are aware of an additional effective quitting method then
this could also encourage more of them to quit who want to do so
independently.
Reduction versus abrupt quitting risk r atios vary across the studies
included in this me ta-analysis (from
Flaxman 1978: RR= 0.50,
95% CI= 0.25 to 1.01 to
Roales-Nieto 1992: RR= 5.00, 95% CI=
0.28 to 88.53). We would expect the effect of the abrupt interven-
tions to be constant across studies as abrupt quit instructions did
not vary, therefore there may be a difference in the success rates
of different reduction methods. This is supported by the fact that
gradual reduction has been found to be less successful than abrupt
quitting in observational studies, but as successful in RCTs. The
studies included in this review used a number of different meth-
ods, including scheduled reduction, non-scheduled reduction, re-
ducing to zero cigarettes before quitting, reducing to 50% of base-
line before quitting. There are conceivably many more ways that
people could reduce before going on to quit completely. Trials
that have been carried out so far to compare different reduction
methods are small and often participants aim to reduce rather than
to quit completely. There has been no attempt to combine all of
these studies into a review so that conclusions can be drawn and
applied to policy devel opment. Therefore furth er research needs
to be carried out to investigate the methods of gradual reduction
that smokers in the general population are using, and whether
they are using any type of support alongside, to see whether this
accounts for the difference in results between observational studies
and RCTs. Further work is needed to identify the most effective
reduction methods in those wanting to quit. Ideally this would be
a review which amalgamates existing evidence and identifies litera-
ture gaps, leading to large-scale RCTs that directly compare differ-
ent methods. In turn, this could inform policy and service devel-
opment as to the most successful reduction to quit method(s). If
there are marginal differences in the effects of different reduction
methods then quitters could choose from a number of options.
However, if there are methods shown to be significantly less ef-
fective, quitters should not be advised to use them as this might
disadvantage them. It may prove useful to establish whether dif-
ferent quitting methods benefit different groups of smokers, e.g.
a particular method may benefit a highly addicted smoker more
than a less addicted one. If so, then a person could use a quitting
method tailored to their individual profile, to produce the optimal
likelihood of abstinence.
In summary, we found no big differences in effectiveness between
advising people who smoke to quit abruptly or advising the m to
reduce cigarette consumption prior to quit day i.e. gradual quit-
ting. These results apply to gradual quitting methods that all em-
ployed a definite quit day and it is not clear whether tell ing people
to cut down and quit when they are ready would achieve the same
results. Given these findings, it seems reasonable to offer smokers
a choice of whether to cut down in preparation f or quitting or to
continue to smoke as normal and quit abruptly.
A U T H O R S C O N C L U S I O N S
Implications for practice
Patients can be given a choice to quit smoking either by
reducing cigarettes smoked before quitting or by quitting
abruptly with no prior reduction.
Reduction to quit can be implemented via self-help therapy
or with the aid of behavioural support.
NRT can be used to aid pre-quit reduction.
15Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 17
Implications for research
Further research should focus on methods of reduction that
smokers in the general population use to quit and whether they
utilise behavioural or self-help support alongside these.
A review of the existing literature on methods of smoking
reduction is needed, and RCTs developed to determine which
methods of reduction are the most effective.
Research is needed to try and establish people who may
benefit most from the abr upt and gradual approach to quitting
smoking, in order to tailor smoking cessation to individuals.
A C K N O W L E D G E M E N T S
We would like to thank Lindsay Stead for her help with devel-
oping the search strategy and carrying out database searches, as
well as offering general editorial support along with Kate Cahill.
Statistical support was given by Rafael Perera and the following
authors provided us with extra information about their studies:
Albert Jerome, Al Behar, William Riley, Michael Cummings, Car-
los Jaen, Sue Curry, Jean Francois Ettter, Jesús Gil Roales-Nieto,
Jan Blalock and Paul Cinciripini.
R E F E R E N C E S
References to studies included in this review
Cinciripini 1995 {published data only}
Cinciripini PM, Lapitsky L, Seay S, Wallfisch A, Kitchens K, Van
Vunakis H. The effects of smoking schedules on ces sation outcome:
Can we improve on common methods of gradual and abrupt
nicotine withdrawal?. Journal of Consulting and Clinical Psychology
1994;63:388–99.
Cummings 1988 {published and unpublished data}
Cummings KM, Emont SL, Jaén C, Sciandra R. Format a nd
quitting instructions as fac tors influencing the impact of a self-
administered quit smoking program. Health Educ ation Quarterly
1988;15:199–216.
Curry 1988 {published data only}
Curry SJ, Marlatt GA, Gordon J, Baer JS. A comparison of
alternative theoretical approaches to smoking cessation and relapse.
Health Psychology 1988;7:545–56.
Etter 2009 {published data only}
Etter J, Cornuz J, Hugelet P, Perneger T. Pre-cessation treatment
with nicotine gums for smoking cessation: a randomized trial
[PA12-1]. Society for Research on Nicotine and Tobacco 14th
Annual Meeting; Portland, Oregon. 2008.
Etter J, Huguelet P, Perneger TV, Cornuz J. Nicotine gum
treatment before smoking cessation: A randomized trial. Archives of
Inter nal Medicine 2009;169:1028–34.
Flaxman 1978 {published and unpublished data}
Flaxman J. Gradual and abrupt quitting strategies and a successful
self-control technique for smoking modification. PhD Dissertation,
Northwestern University 1974.
Flaxman J. Quitting smoking now or later: gradual, abrupt,
immediate, and delayed quitting. Behavior Therapy 1978;9:260–70.
Gunther 1992 {published data only}
Günther V, Gritsch S, Meise U. Smoking cessation - gradual or
sudden stopping?. Drug and Alcohol Dependence 1992;29:231–36.
Hughes 2009 {unpublished data only}
Hughes JR, Soloman LJ, Livingston AE, Callas PW, Peters EN. [A
randomized controlled trial of gradual cessation (aided by NRT) vs.
abrupt cessation of smoking]. provided by Hughes 2009.
Jerome 1999 {unpublished data only}
Jerome A, Behar A, Boyd NR. [Computerized scheduled gradual
reduction for smoking cessation: A randomized work–site outcome
study with 12–month follow–up]. provided by Jerome 1999.
Riley 2005 {published and unpublished data}
Riley W. [Final Report: Computerized scheduling of nicotine nasal
spray]. provided by Riley 2005.
Riley WT, Behar A, Shields R, Kelleher K, Vasiliadis M. [Effects of
a scheduled gradual crossover of cigarettes and nicotine nasal spray
[POS3–043]]. Society for Research on Nicotine and Tobacco 11th
Annual Meeting; Prague, Czech Republic. 2005.
Roales-Nieto 1992 {published data only}
Roales-Nieto JG, Fernández Parra A. Efficiency of a self-control
program for the treatment of smoking behaviour: Different effects
in two strategies, one with an immediate quitting and other with a
gradual quitting [Eficacia de un Programa de Autocontrol Para el
Tratamiento del Tabaquismo. Efectos Diferenciales de dos
Estrategias de Retirada y Reduccion]. Análisis y Modificación de
Conducta 1992;18(59):329–44.
References to studies excluded from this review
Bernard 1972 {published data only}
Bernard H S, Efran JS. Eliminating versus reducing smoking using
pocket timers. Behavior, Research & Therapy 1972;10:399–401.
Bolliger 2000a {published data only}
Bolliger CT. Practical experiences in smoking reduction and
cessation. Addiction 2000;95(S1):S19–S24.
Bolliger 2000b {published data only}
Bolliger CT. Practical experiences in smoking reduction and
cessation. Addiction 2000;95(S1):S19–S24.
Bullen 2008 {published data only}
Bullen C, Howe C, Lin RB, et al.Does pre-quitting nicotine
replacement therapy improve quitting success compared with usual
treatment? Findings from the PQNIQ trial. Society for Research
on Nicotine and Tobacco 14th Annual Meeting Februar y 26-March
1, Portland, Oregon. Rapid Communications Abstracts. 2008.
Cinciripini 1994 {published data only}
Cinciripini PM, Lap itsky LG, Wallfisch A, Mace R, Nezami E.
An evaluation of a multicomponent treatment program involving
16Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 18
scheduled smoking and relapse prevention procedures: initial
findings. Addi ctive B ehaviors 1994;19:13–22.
Daughton 1998 {published data only}
Daughton D, Susman J, Sitorius M, Belenky S, Millatmal T,
Nowak R, Patil K, Rennard SI. Transd ermal Nicotine Therapy and
Primary Care. Archives of Family Medicine 1998;7:425–430.
Glasgow 1989 {published data only}
Glasgow RE, Morray K, Lichtenstein E. Controlled s moking
versus abstinence a s a treatment goal: the hopes and fears may be
unfounded. Behavior Therapy 1989;20:77–91.
Hatsukami 1988 {published data only}
Hatsukami DK, Dahlgren L, Zimmerman R, Hughes JR.
Symptoms of tobacco withdrawal from total cigarette cessation
versus partial cigarette reduction. Psychopharm acology 1988;94:
242–247.
Herrara 1995 {published data only}
Herrara N, Franco R, Herrara L, Partidas A, Rolando R,
Fagerström KO. Nicotine Gum, 2 and 4 mg, for Nicotine
Dependence. Chest 1995;108:447–451.
Jerome & Fiero 1999 {unpublished data only}
Jerome A, Fiero PL. Computerized scheduling of nicotine gum use.
provided by Riley 1999.
Jerome A, Fiero PL, Behar A, Riley WT. Using a computerized,
portable device to schedule nicotine gum for smoking cessation.
provided by Jerome 1999.
Marston 1971 {published data only}
Marston AR, McFall RM. Comparison of behavior modification
approaches to s moking reduction. Journal of Consulting and
Clinical Psychology 1971;36(2):153–162.
Rezaishiraz 2007 {published data only}
Rezaishiraz H , Hyland A, Mahoney MC, O’Connor RJ,
Cummings KM. Treating smokers before the quit date: Can
nicotine patches and denicotinized cigarettes reduce cravings?.
Nicoti ne & Toba cco Research 2007;9(11):1139–1146.
Rose 1998 {published data only}
Rose JE, Behm FM, Westman EC. Nicotine-mecamylamine
treatment for smoking cessation: the role of pre-cessation therapy.
Experimental and Clinical Psychopharmacology 1998;6:331–43.
Rose 2006 {published data only}
Rose JE, Behm FM, Westman EC, Kukovich P. Precessation
treatment with nicotine skin patch facilitates smoking cessation.
Nicoti ne & Toba cco Research 2006;8:89–101.
Rose 2009 {published data only}
Rose JE, Herskovic JE, Behm FM, Westman EC. Pre-cessation
treatment with nicotine patch significantly inc reases abstinence
rates relative to conventional treatment. Nicotine & Tobacco
Research 2009;11:1067–75.
Schuurmans 2004 {published data only}
Schuurmans MM, van Diacon AHBX, Bolliger CT. Effect of pre-
treatment with nicotine patch on withdrawal symptoms and
abstinence rates in smokers subsequently quitting with the nicotine
patch: a randomized controlled trial. Addiction 2004;99:634–40.
Shiffman 2009 {published data only}
Shiffman S , Ferguson SG, Strahs KR. Quitting by gradual smoking
reduction using nicotine gum: A randomized controlled trial.
American Journal of Preventive Medicine 2009;36:96–104.
References to ongoing studies
Cinciripini 2006 {published data only (unpublished sought but not
used)}
Cinciripini PM, Lam C, Blalock JA, Robinson J, Wetter DE, Baile
W. Does scheduled reduced smoking have a place among smoking
cessation treatments? [SYM 10C]. Society for Research on Nicotine
and Tobacco 12th Annual Meeting; Orlando, Florida. 2006.
Lindson 2009 {published and unpublished data}
Lindson N, Aveyard P, Ingram JT, Inglis J, Beach J, West R, Susan
M. Rapid reduction versus abrupt quitting for smokers who want
to stop soon: a randomised controlled non-inferiority trial. Trials
2009;10(69):1–14.
Riley 2001 {published and unpublished data}
Riley W, Jerome A, Behar A, Weil J. Computerized scheduled
gradual reduction of smoking prior to nicotine patch use: Evidence
of add itive efficacy [PO3 28]. Society for Research on Nicotine and
Tobacco 7th Annual Meeting, Seattle, Washington. 2001.
Additional references
Cheong 2007
Cheong Y, Yong H, Borland R. Does how you quit affect success? A
comparison between abrupt and gradual methods using data from
the international tobacco control policy evaluation study. Nicotine
& Tobacco Research 2007;9(8):801–810.
Davis 1992
Davis SW, Cummings KM, Rimer BK Sciandra R, Stone JC. The
Impact of Tailored Self-Help Smoking Cessation Guides on Young
Mothers. Health Education & Behavior 1992;19(4):495–504.
Fagerstrom 2002
Fagerstrom KO, Hughes JR. Nicotine concentrations with
concurrent use of cigarettes and nicotine replacement: A review.
Nicoti ne & Toba cco Research 2002;4(Suppl 2):S73–S79.
Ferguson 2005
Ferguson J, Bauld L, Chesterman J, Judge K. The English smoking
treatment services: one year outcomes. Addiction 2005;100:59–69.
Fiore 2008
Fiore MC, Jaen CR, Baker TB, Bailey WC, Benowitz NL, Curry
SJ, et al.Treating tobacco use and dependence: 2008 update. http:/
/www.ahrq.gov/path/tobacco.htm [Accessed 11 February 2009].
Rockvill, M D: Department of Health and Human Services, 2008.
Handbook 2008
Higgins JPT, Green S (eds). Cochrane Handbook for Systematic
Reviews of Interventions. 5.0.1. The Cochrane Collaboration,
2008.
Hughes 2004
Hughes JR, Keely J, Naud S. Shape of the relapse curve and long
term abstinence among untreated smokers. Addiction 2004;99(1):
29–38.
17Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 19
Hughes 2006
Hughes JR, Callas PW, Peters E N. Interest in gradual cessation.
Nicoti ne & Toba cco Research 2006;9(6):671–5.
Hughes 2007
Hughes JR. Smokers who choose to quit gradually versus abruptly.
Addiction 2007;102(8):1326–7.
Hunt 1973
Hunt WA, Matarazzo JD. Three years later: Recent developments
in the experimental modification of smoking behavior. Journal of
Abnormal Psychology 1973;81(2):107–114.
Jarvis 2002
Jarvis MJ, McIntyre D, Bates C. Effectiveness of smoking cessation
initiatives. BMJ 2002;324:608.
Lancaster 2005
Lancaster T, Stead LF. Individual behavioural counselling for
smoking cessation. Cochrane Database of Systematic Reviews 2005,
Issue 2.
Law 1995
Law M, Tang JL. An analysis of the effectiveness of interventions
intended to help people stop smoking. Archives of Internal Medicine
1995;155(18):1933–1941.
Levinson 1971
Levinnson BL, Shapiro D, Schwartz GE, Tursky B. Smoking
Elimination by Gradual Reduction. Behavior Therapy 1971;2:
477–487.
Martinson 1999
Martinson BC, Murray DM, Jeffrey RW, Hennrikus DJ. Intraclass
Correlation for Measures from a Worksite Health Promotion Study:
Estimates , Correlates, and Applications.. American Journal of
Health Promotion 1999;13(6):347–357.
McRobbie 2006
McRobbie H, Whittaker R, Bullen C. Using nicotine replacement
therapy to assist in reducing cigarette cons umption before quitting:
another strategy for smoking cessation?. Di sease Management and
Health Outcomes 2006;4(6):335–40.
Moher 2001
Moher D, Schulz KF, Altman DG, for the CONSORT group. The
CONSORT statement: revised recommendations for improving
the quality of reports of parallel group randomized trials. JAMA
2001;285:1987–1991.
Moore 2009
Moore D, Aveyard P, Connock M , Wang D, Fry-Smith A, Barton P.
Effectiveness and safety of nicotine replacement therapy assisted
reduction to stop smoking: systematic review and meta-analysis..
BMJ 2009;338:867–880.
NICE 2008
National Institute for Health and Clinical Excellence. Smoking
cessation services in primary care, pharmacies, local authorities and
workplaces, particularly for manual working groups, pregnant
women and hard to reach communities. Public health guidance 10.
http://www.nice.org.uk/nic emedia/pdf/PH010guidance.pdf
[Accessed 20 February 2009] 2008.
NZ MoH 2007
Ministry of Health. New Zealand smoking cessation guidelines.
http://www.moh.govt.nz/moh.nsf/pagesmh/6663/$File/nz-
smoking-cessation-guidelines-v2-aug07.pdf [Accessed 11 February
2009]. Wellington: Ministry of Health, 2007.
Shiffman 2007
Shiffman S , Hughes JR, Ferguson SG, Pillitteri JL, Gitchell JG,
Burton SL. Smokers interest in using nicotine replacement to aid
smoking reduction. Nicotine & Tobacco Research 2007;9(11):
1177–82.
Shiffman 2008
Shiffman S , Ferguson SG. Nicotine p a tch therapy prior to quitting
smoking: a meta-analysis. Addiction 2008;103:557-563.
SRNT 2002
SRNT Subcommittee on Biochemical Verification. Biochemical
verification of tobacco use and cessation. Nicotine & Tobacco
Research 2002;4(2):149–159.
Stead 2005
Stead LF, Lancaster T. Group behaviour therapy programmes for
smoking cessation. Cochrane Database of Systematic Reviews 2005,
Issue 2.
Stead 2007
Stead LF, Lancaster T. Interventions to reduce harm from
continued tobacco use. Cochrane Da t abase of Systematic Reviews
2007, Issue 3. [DOI: 10.1002/14651858.CD005231.pub2]
Stead 2008
Stead LF, Perera R, Bullen C, Mant D, Lancaster T. Nicotine
replacement therapy for smoking cessation. Cochrane Database of
Systematic Reviews 2008, Issue 1.
USPHS 2008
Treating Tobacco Use and Dependence: 2008 Update.. Rockville
(MD): U.S. Department of Health and Human Services, Public
Health Service May 2008.
Wang 2008
Wang D, Connock M, Barton P, Fry-Smith A, Aveyard P, Moore D.
’Cut down to quit with nicotine replacement therapies in smoking
cessation: a systematic review of effectiveness and economic
analysis.. Health Technology Assessment 2008;12(2):1–135.
West 2001
West R, McEwen A, Bolling K, Owen L. Smoking cessation and
smoking patterns in the general population: A 1-year follow-up.
Addiction 2001;96(6):891–902.
West 2006
West 2006. Smoking and smoking cessation in England: 2006.
http://66.102.1.104/scholar?hl=en&lr=&client=firefox-a&q=cache:
UDuwOCSGZssJ:www.aspsilverbackwebsites.co.uk/
smokinginengland/Ref/pa-
per4.pdf+%22smoking+and+smoking+cessation+in+england%22+2006+west
2006.
References to other published versions of this review
Lindson 2009
Lindson N, Aveyard P, Hughes JR. Reduction versus abrupt
cessation in smokers who want to quit. Cochrane Database of
Systematic Reviews 2009, Issue 4. [Art No: CD008033; DOI:
10.1002/14651858]
Indicates the major publication for the study
18Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 20
C H A R A C T E R I S T I C S O F S T U D I E S
Characteristics of in cluded st udies [ordered by study ID]
Cinciripini 1995
Methods Country: USA
Recruitment: participants recruited from the community, method not stated.
Participants 128 smokers randomized to 4 groups, with at least 3 years smoking history, consumption
of 15+ cigarettes per day (CPD), no current cessation treatment, psychiatric disorder
or uncontrolled systemic illness. 58% F, av. age 45, av. CPD 24, av. 4 previous quit
attempts.
Interventions 1. Scheduled reduced: inter-cigarette interval progressively lengthened, until quit day at
week 5.
2. Non-scheduled reduced: CPD reduced using same quota as scheduled group but
participants were free to choose when they smoked their cigarettes, until quit day at week
5.
3.Scheduled non-reduced: participants instructed to smoke at regular time intervals but
the time intervals were not progressively reduced to quit day at week 5.
4. Non-scheduled, non-reduced: No manipulation of inter-cigarette interval or cigarette
frequency, until quit day at week 5.
Pharmacotherapy: No pharmacotherapy
Type of support: Two-hour weekly group meetings; cognitive behavioural intervention
weeks 2-5; relapse prevention weeks 5-9.
Outcomes Abstinence: Prolonged abstinence (defined as smoking on fewer than 5 days between
assessments) at treatment end (week 9) & 1, 6 and 12 month post-treatment. (PP at quit
week (week 5) also reported).
Validation: CO <6ppm at quit week, cotinine <14 mg/ml at treatment e nd & 1,6 and
12 month follow-up
Other outcomes: CPD, coping behaviour, withdrawal score, tension & fatigue mood
states, urge frequency, self-efficacy.
Notes Groups 1 and 2 combined to create reduction group and Groups 3 & 4 combined to
create abrupt quitting group.
Risk of bias
Item Authors’ judgement Description
Adequate sequence generation? Unclear Randomized, method not stated.
Allocation concealment? Unclear No information given.
Incomplete outcome data addressed?
All outcomes
Unclear Information on attrition/exclusions not
given. Those with missing data were
counted as non-abstainers.
19Reduction versus abrupt cessation in smokers who want to quit (Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Page 21
Cummings 1988
Methods Country: USA
Recruitment: calle rs responding to advertisement of stop smoking hotline, who accepted
a free stop smoking booklet.
Participants 1895 randomized to 4 experimental groups and 1 control group. 18+ year old current
smokers. 65% F, av. age 42, av. CPD 28, av. 3 previous quit attempts.
Interventions 1. Booklet instructing smokers to gradually reduced cigarettes smoked before quitting.
Day by day structured guide.
2. Booklet instructing smokers to gradually reduced cigarettes smoked before quitting.
No day by day instructions.
3. Booklet instructing smokers to quit abr uptl y. Day by day guide.
4. Booklet instructing smokers to quit abr uptl y. No day by day instructions.
5.Control: booklet providing information on the heal th hazards of smoking and the
nature of tobacco addiction, but did not give specific advice on how to stop smoking.
Pharmacotherapy: No pharmacotherapy.
Type of support: Self-help booklet.
Outcomes Abstinence: Continuous between 1 month & 6 month follow-up. (1 week PP at 1 & 6
months post-enrolment, 1 month prolonged at 6 months post-enrolment also reported)
.
Validation: surrogate interview conducted with family member or friend.
Other outcomes: report of cessation attempt, cigarettes per day, percentage of booklet
read, booklet e valuation, actions taken in preparation for quit and after quit, method of
quitting.
Notes Groups 1 and 2 combined to create reduction group and Groups