| BACKGROUND |
Nicotine replacement therapy (NRT) is frequently used as a component of smoking cessation strategies. It minimizes many of the physiological and psychomotor withdrawal symptoms usually experienced following smoking cessation and, therefore, may increase the likelihood of remaining abstinent (Gourlay 1990).
The first type of NRT to become widely available was chewing gum. The nicotine resin complex is presented in a buffered chewing gum base to enable the nicotine to be absorbed directly through the buccal mucosa, resulting in plasma concentrations which are approximately half that produced by smoking a cigarette (Russell 1976). Nicotine chewing gum is available either as a 2 mg or 4 mg preparation, and in many countries the lower dose is sold over-the-counter, without a prescription from a medical practitioner. Several factors limit the usefulness of nicotine chewing gum in some smokers, including oral and gastric side effects (Henningfield 1990), impaired absorption when taken with coffee or acidic beverages (Hughes 1986), inadequate dosing, and a risk that some smokers may transfer their dependency from cigarettes to the gum (Hughes 1986).
Other forms of NRT that aim to avoid some of the problems associated with nicotine gum have been developed, including transdermal nicotine patches, intranasal nicotine spray, and nicotine inhaler devices. Transdermal patches are now widely available. Nasal spray, inhaler, lozenges and tablets of nicotine are also licensed for use in a number of countries, and other formulations are being developed.
Transdermal patches are available in several different sizes, and deliver between 7 mg and 22 mg of nicotine over a 24-hour period, resulting in plasma levels similar to the trough levels seen in heavy smokers (Fiore 1992). Some brands of patch are designed to be worn for 24 hours, whilst others are intended to be worn for 16 hours, delivering a dose of 15mg over that period.
The introduction of transdermal patches was accompanied by strong marketing campaigns in a number of countries, targeted both at smokers and physicians, encouraging use of the patch as a "proven and effective" smoking cessation strategy (Saul 1993). This has caused much debate about the role of NRT in smoking cessation; including which group(s) of smokers should be offered NRT, which preparations should be used, in what dose regimen, and whether NRT is effective when used alone or only together with some form of additional support strategy.
More recently, the observation that nicotine patches and gum do not provide 100% nicotine replacement (Dale 1995; Hurt 1993) has led to interest in increasing the efficacy of nicotine replacement by raising patch doses (Jorenby 1995), or by combining different forms of NRT, for example, patches and gum (Kornitzer 1995; Puska 1995), or nasal spray with patches (Blondal 1999). In addition, there is growing interest in comparing NRT to newer pharmacotherapies, particularly the antidepressant bupropion.
This review assesses the effectiveness of the different forms of NRT when offered to smokers who have varying levels of dependency and motivation to quit, in a range of clinical settings, and with or without additional support programs.
| OBJECTIVES |
To determine the effectiveness of nicotine replacement therapy (NRT) (including gum, transdermal patch, intranasal spray and inhaled and oral preparations) in achieving long-term smoking cessation.
A second objective, added in 2001, is to determine the effectiveness of NRT in assisting long term reduction in the amount smoked by smokers who are unwilling or unable to quit.
We wished to test the following hypotheses:
1) The use of NRT is more effective than placebo or 'no NRT' intervention in promoting smoking cessation (Comparison 1);
2) 4mg nicotine gum is more effective than 2mg nicotine gum (Comparison 2), and fixed dose schedules are more effective than ad-lib use (Comparison 13);
3) The provision of high-intensity support, in addition to the use of NRT, is more effective in producing abstinence than addition of low-intensity support programs (Comparison 3);
4) The effectiveness of the nicotine patch is greater with longer duration of use (Comparison 4), with weaning rather than abrupt withdrawal (Comparison 5), and with 24 hour patches rather than with 16 hour patches (Comparison 6);
5) NRT is more effective when offered to smokers who are motivated to quit and will, therefore, be more effective in clinical settings that selectively recruit motivated smokers (Comparison 7);
6) Increasing the delivery of nicotine replacement by raising the dose of nicotine patch therapy (Comparison 8) or combining different forms of NRT (Comparison 9) is more effective than conventional dose monotherapy ;
7) NRT is effective in smokers who have relapsed after previous NRT use (Comparison 10).
8) NRT is more effective than the antidepressant bupropion for smoking cessation (Comparison 11).
9) NRT is more effective than placebo for achieving long term reduction in number of cigarettes smoked for people who cannot quit.
| CRITERIA FOR CONSIDERING STUDIES FOR THIS REVIEW |
Randomized controlled trials. Trials where allocation to treatment was by a quasi-randomized method were also included but where appropriate, sensitivity analysis was used to determine whether their inclusion altered the results.
Smokers of either gender were included irrespective of the setting from which they were recruited and/or their initial level of nicotine dependency. Studies which randomized therapists, rather than smokers, to offer NRT or a control were included providing that the specific aim of the study was to examine the effect of NRT on smoking cessation. Trials that randomized physicians or other therapists to receive an educational intervention, which included encouraging their patients to use NRT, were not included but are reviewed separately (Lancaster 2000).
Comparisons of nicotine replacement therapy (including nicotine chewing gum, transdermal nicotine patches, nicotine nasal spray, nicotine inhalers/inhalators and nicotine tablets) versus placebo or no nicotine replacement therapy control.
Trials of different doses of nicotine replacement therapy were also included.
In some analyses we categorised the trials into two groups depending on the level of additional support provided (low or high). Low-intensity additional support was regarded as part of the provision of routine care. If the duration of time spent with the smoker (including assessment for the trial) exceeded 30 minutes at the initial consultation or the number of further assessment and reinforcement visits exceeded two, the level of additional support was categorized as high.
The review comprises a comparison of the effects of NRT versus control on smoking cessation, rather than withdrawal symptoms. Trials in which follow-up was of short duration (less than 6 months) were excluded.
In each study the strictest available criteria to define abstinence were used. For example, in studies where biochemical validation of cessation was available, only those participants who met the criteria for biochemically confirmed abstinence were regarded as being abstinent. Wherever possible a sustained cessation rate, rather than point prevalence, was used. In trials where patients were lost to follow-up they were regarded as being continuing smokers.
The review also includes trials which compared the effects of NRT versus placebo or other pharmacotherapies on achieving a sustained reduction in the number of cigarettes smoked amongst smokers not attempting to quit.
| SEARCH STRATEGY FOR IDENTIFICATION OF STUDIES |
The specialised register of the Tobacco Addiction group was searched for trials with any reference to the use of nicotine replacement therapy of any type in the title, abstract or keywords. The most recent issues of the databases included in the register as searched for the current update of this review are: Medline Express (SilverPlatter) 12/2001, Medline (PubMed) Jan-July 2002, PsycLit/PscINFO (SilverPlatter) 5/2002, Embase 12/2001 (SilverPlatter), Science Citation Index 6/2002 (Web of Science), Cochrane Controlled Trials Register (Cochrane Library) Issue 2, 2002. (Earlier versions of the review performed searches of additional databases; Cancerlit, Health Planning and Administration, Social Scisearch, Smoking & Health and Dissertation Abstracts. Since this search did not produce any additional trials these databases have not been used after December 1996. During preparation of the first version of this review letters were also sent to manufacturers of NRT preparations. Since this did not result in additional data the exercise was not repeated for subsequent updates.)
| METHODS OF THE REVIEW |
Data were extracted from the published reports by two individuals independently. Disagreements were resolved by discussion or referral to a third party. No attempt was made to blind any of these individuals either to the results of the primary studies or which treatment subjects received. Reports published only in non-English language journals were examined with the assistance of translators.
Smoking cessation rates in the intervention and control groups were identified from the published reports at 6 or 12 months. Since not all studies reported cessation rates at exactly these intervals, we allowed a window period of 6 weeks at each follow-up point. For trials without 12 month follow-up we have used 6 month data in all analyses.
In trials of smoking reduction, sustained reduction was defined as a (self reported) 50% reduction in the number of cigarettes smoked per day at the evaluation point compared with the baseline.
The Fagerstrom score (0 - 11) was used to classify the level of nicotine dependency. Smokers with a Fagerstrom score of 7 or more were classified as high nicotine dependence, those with a score less than 7 were classified as low nicotine dependence.
In trials where details of the methodology were unclear, or where results were not expressed in a form which allowed extraction of the necessary key data, we wrote to the investigators asking for the required information. Six of the eight investigators approached responded to our request, but only five were able to provide the data that we requested.
The statistical methods used have been previously described (Yusuf 1985). For each trial in the overview the number of expected events (E) in the experimental group was calculated, assuming that the intervention used had no effect. This calculation is based on the number of participants initially randomized, irrespective of whether they completed the study. The number of expected events was then subtracted from the number that were actually observed in the experimental group (O). Adding these separate differences (observed minus expected O-E) and their variances, allows the calculation of a statistic (and its variance) that is typical of the differences observed between experimental and control groups in the array of trials within each overview. This was used to test the null hypothesis and also to estimate how large, and hence how worthwhile, any differential effects were likely to be. For the latter purpose, the typical odds ratio (OR) and its confidence interval (CI) were calculated using a fixed-effects model (Peto method, Yusuf 1985). These methods meant that patients in one trial were never compared directly with patients in another. Instead, only patients within a trial were compared according to the intervention they received. This avoids differences between treatments, duration of treatment, follow-up and end-point definitions interfering with the estimate of effectiveness. Results have been expressed as an OR (NRT:control) for achieving abstinence from smoking at a given point in time together with the 95% CIs for this estimate.
We also analysed the odds of achieving abstinence with each type of NRT relative to control (either placebo or no intervention) in different clinical settings, and by the intensity of additional support offered. In all pooled analyses, tests for heterogeneity were performed using a Mantel-Haenszel approach (Cochran 1954).
In order to summarize the data from a clinical perspective we calculated the number of smokers who needed to be treated in order to produce 1 successful quitter at 12 months beyond that which would be achieved with the control intervention. This estimate was based on the inverse of the pooled typical event rate difference, calculated using the method described by Rothman (Rothman 1986).
| DESCRIPTION OF STUDIES |
The review includes 110 studies. Ninety-six included a placebo or non nicotine control arm and contribute to the primary analysis. In this group there were 51 trials of nicotine gum, 34 of transdermal nicotine patch, four of intranasal nicotine spray, four of inhaled nicotine and three of an oral tablet. Five trials compared combinations of two forms of nicotine therapy with only one form; patch with gum to patch alone (Kornitzer 1995); patch with gum to gum alone (Puska 1995); patch with nasal spray to patch alone (Blondal 1999); patch with inhaler to inhaler alone (Bohadana 2000) and patch with inhaler to either one alone (Tonnesen 2000). A factorial trial compared nicotine and bupropion (Zyban) (Jorenby 1999). One trial compared nicotine inhaler to placebo for smoking reduction (Bolliger 2000).
With the exception of 12 gum trials and 13 patch trials, participants were followed for at least 12 months. Sixteen of the gum trials and six of the patch trials were conducted in a primary-care setting where smokers were usually recruited in response to a specific invitation from their doctor during a consultation. A further four gum trials were undertaken in workplace clinics, and one in a university clinic. Since participants in these trials were recruited in a similar way to primary-care, we aggregated them in analyses involving clinic setting. One patch trial conducted in a university clinic, one conducted in a worksite setting and one conducted in Veterans Affairs Medical Centers and recruiting patients with cardiac diseases (Joseph 1996) were also included in the primary care category. One trial in an antenatal clinic (Wisborg 2000) is kept in a separate category. Six of the gum trials, one of the nasal spray trials and one of the inhaler trials, were carried out in specialized smoking-cessation clinics to which participants had usually been referred. Seven trials (three gum and four patch) were undertaken with hospital in- or out-patients, who were usually recruited because they had a coexisting smoking-related illness. Recent interest has focused on use of nicotine replacement therapy obtained 'over the counter' rather than from a medical care provider. Three trials have compared patch to placebo in this type of setting (Davidson 1998; Hays 1999; Sonderskov 1997). One of these also allowed a comparison between purchased and free patch with minimal support (Hays 1999). Another trial compared patch with physician support to patch without support (Leischow 1999). The remaining gum, patch, inhaler and nasal spray trials were undertaken in participants from the community most of whom had volunteered in response to media advertisements, but who were treated in clinical settings. One of the patch trials was conducted in relapsed smokers (Gourlay 1995).
Of the 51 trials comparing nicotine gum with placebo or no gum, 37 used the 2mg dose, two used 4mg only, and seven used a variable or mixed dosage. In the other trials the dose was not stated. Two trials compared a fixed dosage regimen with an ad lib regimen (Killen 1990; Goldstein 1989). The duration of therapy ranged from 3 weeks to 12 months. Many of the trials included a variable period of dose tapering, but most encouraged participants to be gum free by 6 to 12 months. Five trials compared 2mg and 4mg gum (Garvey 2000; Herrera 1995; Hughes 1990; Kornitzer 1987; Tonnesen 1988).
Of the 34 trials comparing nicotine patch with placebo or no patch, 24 used the 24-hour preparations, nine used patches applied for 16 hours, and one trial (Daughton 1991) included comparison of groups wearing 16 hour or 24 hour patches or placebo. All but two (Cinciripini 1996; Wong 1999) of the patch trials compared active against placebo; one trial also included comparison against a no patch group (Buchkremer 1988). The minimum duration of therapy ranged from 6 weeks to 3 months, with a tapering period, if required, in 27 of the trials. Two trials directly compared two durations of therapy (CEASE 1999; Bolin 1999). Six trials compared a higher dose to a standard dose patch (Dale 1995; Hughes 1999; Jorenby 1999; CEASE 1999; Killen 1999; Paoletti 1996).
Data are available from four completed trials of intranasal nicotine spray (Sutherland 1992; Hjalmarson 1994; Schneider 1995; Blondal 1997), and for four trials of inhaled nicotine (Tonnesen 1993, Schneider 1996; Leischow 1996; Hjalmarson 1997). One trial of a nicotine inhaler was excluded as follow-up was for only 3 months (Glover 1992). Leischow refers to another study by different investigators which did not demonstrate any benefit of a nicotine inhaler. Three trials of nicotine sublingual tablets/lozenges are included, one of which has not yet been published in full (Glover 1999).Two used 2-mg sublingual tablets (Glover 1999, Wallstrom 2000). The third trial used a nicotine lozenge and stratified participants according to dependency level based on their time to first cigarette of the day (TTFC). The two groups are entered in the meta-analysis as separate trials. Smokers whose TTFC was >30 minutes were randomised to 2mg lozenges or placebo (Shiffman 2002A), whilst TTFC <30 minute smokers had higher dose 4mg lozenges or placebo (Shiffman 2002B).
One trial (Kornitzer 1987), conducted in a worksite setting, was confined to male smokers, two recruited only women (Pirie 1992, Wisborg 2000), . The remainder included smokers of both sexes. The range in the mean number of cigarettes smoked at entry into the trials, among the studies which provided this data, was 15.5 to 32.9.
In some analyses we categorised the trials into two groups depending on the level of additional support provided (low or high). Low-intensity additional support was regarded as part of the provision of routine care. If the duration of time spent with the smoker (including assessment for the trial) exceeded 30 minutes at the initial consultation or the number of further assessment and reinforcement visits exceeded two, the level of additional support was categorized as high. In some trials this level of support included group behaviour modification sessions. This sub group analysis is shown only for gum and patch studies, the number of trials of other types being too small.
One trial (Nebot 1992) randomized primary care physicians to offer nicotine gum or placebo to smokers. In all the other trials included the unit of randomization was the smoker.
One trial compared the nicotine inhalator to placebo for helping smokers to reduce the number of cigarettes smoked (Bolliger 2000). The primary outcome was a reduction in 50% or more from baseline after 24m. The number of quitters was also recorded.
| METHODOLOGICAL QUALITY |
Thirty-seven studies (33%) reported randomization procedures in sufficient detail to be rated A for their attempts to control selection bias. The majority of studies either did not report how randomization was performed or reported it in insufficient detail to determine whether a satisfactory attempt to control selection bias had been made (B). A small number of trials randomized to treatment according to day or week of clinic attendance (Page 1986; Richmond 1990; Russell 1983), birthdate (Fagerstrom 1984), or smokers' clinic group (McGovern 1992).
Definitions of abstinence varied considerably with 28 of the trials reporting the primary long-term outcome abstinence measure as a point prevalence, 76 as a sustained measure, and five making no specific mention in the report as to which approach was used.
All but eleven of the trials used some form of validation of self-reported smoking cessation. Validation of abstinence was carried out by blinded methods (measurement of metabolites in body fluids) in 25 trials. Measurement of carbon monoxide in expired air was the most common form of validation used. However, the 'cut-off' level of carbon monoxide used to define abstinence varied from less than 4 to 11 parts per million. In one trial participants who smoked up to three cigarettes per week were still classified as abstinent (Abelin 1989).
The quality of bias control did not differ significantly between trials of different forms of NRT.
Three trials are included based on data available from abstracts (Glover 1999; Mori 1992; Nakamura 1990).
| RESULTS |
The five forms of NRT were all significantly more effective than placebo, or no NRT, in helping smokers achieve abstinence. The benefit from using NRT was evident throughout the 6-12 month period of follow-up despite the presence of a significant relapse rate with each type of preparation. Despite the range of variation in characteristics of trials included in this review, there was no statistical evidence of significant heterogeneity in any of the main pooled analyses. Only eight of the individual trials (Hall 1996; Hughes 1990; Campbell 1991; Harackiewicz 1988; Joseph 1996; Kornitzer 1995; Killen 1997 (Video); Niaura 1999) yielded a negative treatment effect with NRT at the end of follow-up, but in a further 49 trials the 95% CI for the odds of abstinence included unity. Many of these trials had small numbers of smokers, and hence, insufficient power to detect a modest treatment effect with reasonable certainty.
The percentage of smokers who were abstinent after 12 months (excluding trials with shorter follow-up; data not shown) was 18% (95% CI 17% to 19%) amongst smokers who had been allocated to receive nicotine gum and 14% (95% CI 13% to 15%) amongst those who had used transdermal patches. For intranasal spray, nicotine inhaler and sublingual tablet, the corresponding figures were 24% (20% to 28%), 17% (14% to 21%) and 20% (15% to 25%) respectively.
When the abstinence rates for all trials were pooled (Comparison 1), using the longest duration of follow-up available, 17% of smokers allocated to receive NRT had successfully quit compared with 10% in the control group. This represents a 72% increase in the odds of abstinence with the use of NRT (95% CI, 60% to 84%).
The pooled OR of abstinence for any form of NRT relative to control was 1.74 (95% CI 1.64 -1.86). For the different forms of NRT the OR ranged from 1.66 (95% CI 1.52 - 1.81) with nicotine gum to 2.27 (95% CI 1.61 - 3.20) with nicotine nasal spray. For transdermal patch, nicotine inhaler, and nicotine sublingual tablet the ORs for abstinence were 1.74 (95% CI 1.57 - 1.93), 2.08 (95% CI 1.43 - 3.04) and 2.08 (95% CI 1.63 - 2.65) respectively. Since the confidence intervals around these estimates of effect overlapped there was no evidence in this indirect comparison for a significant difference in the effectiveness of the five types of NRT. For trials of nicotine gum and transdermal patch, the ORs for not smoking were not affected by whether or not the control group was placebo or no NRT (data not shown). The OR for nicotine gum was also not affected by whether or not the trial which randomized the treating physician, rather than the smoker (Nebot 1992), was included (data not shown).
The pooled OR of abstinence in the trials which directly compared 4mg versus 2mg gum was 2.67 (95% CI 1.69 - 4.22, Comparison 2.2) in highly dependent smokers (Kornitzer 1987; Tonnesen 1988; Herrera 1995) . In low dependency or unselected smokers there was no evidence for an effect (Hughes 1990; Kornitzer 1987). Two trials compared a 'fixed'-dose regimen of 2mg nicotine gum against use of an ad-lib regimen (Goldstein 1989; Killen 1990). The fixed dose regimen had non significantly better quit rates (OR 1.29 95% CI 0.90 - 1.84, Comparison 13).
There was no evidence of a difference in clinical effectiveness for 16 hour compared to 24 hour patch, although there was significant heterogeneity in the results of the 8 trials which used a 16 hour patch (Chi-square 18.53, df 7) (Comparison 6). One trial directly compared the effect of only wearing the patch whilst awake (about 16 hours) versus wearing it continuously for 24 hours (Daughton 1991). The study found no significant difference in the self-reported OR of abstinence at 6 months follow-up but had low power (OR: 24hr patch versus 16hr patch: 0.62, 95% CI 0.26 - 1.47). In addition, use of the patch for up to 8 weeks was as effective as longer courses of treatment (Comparison 4). One large trial which compared a 28 to a 12 week course of treatment found no evidence of benefit from longer treatment (CEASE 1999). A smaller trial comparing a 3 week to a 12 week course also found no evidence for a difference (Bolin 1999). There was no difference in effect in trials where the dose was tapered, or weaned, compared to those where withdrawal was abrupt (Comparison 5). Similarly, in the two trials that directly compared weaning with abrupt withdrawal, no difference was found (Hilleman 1994; Stapleton 1995).
Six trials have compared a high patch dose to standard dose (Comparison 8) . Three used 24 hour patches and compared 42/44mg doses to standard 21/22mg doses (Dale 1995; Hughes 1999; Jorenby 1995). Three used 16 hour patches and compared a 25mg high dose to 15mg standard dose (CEASE 1999; Killen 1999; Paoletti 1996). Two studies (Hughes 1999; Killen 1999) specifically recruited heavy smokers and one selected smokers with baseline cotinine levels of over 250ng/ml (Paoletti 1996). Pooling all six studies gives an OR of 1.21 (95% CI 1.03 - 1.42) suggesting that there may be a small benefit from higher doses.
Combinations of nicotine therapy:
In the two trials which compared a combination of patches and gum with gum or patch alone, early increases in abstinence rates in the more intensively treated group were not sustained at one year follow-up (Kornitzer 1995; Puska 1995). A trial comparing nasal spray and patch with patch alone found a significant increase in sustained abstinence at one year with the combined therapy (Blondal 1999). One trial combining patch with inhaler also showed a non significant increase in cessation compared to inhaler alone (Bohadana 2000). A trial combining patch and inhaler had nonsignificantly lower quit rates from the combination than with either of the forms alone (Tonnesen 2000). Pooling all five trials suggests an overall benefit (OR 1.55, 95% CI 1.17 -2.05), but statistical heterogeneity approaches significance (p=0.094), and the trials are clinically heterogeneous in the combinations and comparison therapies used.
Clinical Settings:
The pooled OR of not smoking at 6 to 12 months when NRT is offered to smokers attending smoking-cessation clinics did not differ significantly from the OR amongst those recruited from the community as volunteers, or those who were recruited opportunistically through primary-care (Comparison 7). However, since the absolute abstinence rate was higher in community volunteers and smoking-cessation clinics, the percentage of smokers helped to quit by using NRT was higher in these settings than in primary-care or hospital patients.
Smokers recruited as hospital in-patients, or through out-patient clinics, have a lower increase in quitting using gum than smokers seen in other clinical settings, (OR 1.13, 95% CI 0.84 - 1.51). The results using transdermal patches in hospitals, based on three trials, are more consistent with the results seen in other settings (OR 1.74, 95% CI 1.19 - 2.54). In the single trial of a nicotine patch for women trying to quit during pregnancy no benefit of the patch was detected (OR 1.05. 95% CI 0.54 - 2.18).
Increasingly, various forms of NRT are available without a medical prescription and can be purchased in pharmacies or other shops. Three placebo controlled trials of nicotine patch have assessed effectiveness in an 'Over the Counter' setting with minimum levels of support. The effectiveness of the patch was similar to that in other settings. One trial (Leischow 1999) comparing patch with minimal physician support and patch with no support in a simulated OTC setting did not detect a significant difference. Continuous abstinence rates at one year were very low in both conditions (Comparison 12).
Intensity of additional support:
The absolute probability of not smoking at 6 to 12 months was greater in trials which provided high-intensity additional support, rather than low-intensity, particularly with nicotine gum (Comparison 3). Although the pooled OR of abstinence was greater in the trials of gum or patch in which smokers only received low-intensity additional support, the confidence intervals overlapped.
Only two small studies, both in primary care, directly compared the effect of providing high versus low-intensity follow-up to participants receiving nicotine gum (Fagerstrom 1984; Marshall 1985). The pooled results favour intensive follow-up but the result was not statistically significant (OR intensive follow-up:minimal follow-up: 1.30, 95% CI, 0.75 - 2.28, data not shown). In the one patch trial which compared minimal counselling with two forms of more intensive counselling in patients receiving one of two nicotine doses, the intensive intervention did not lead to improved outcomes (Jorenby 1995).
Relapsed smokers:
Although many of the trials reported here did not specifically exclude subjects who had previously tried and failed to quit with NRT, only one trial has specifically looked at the effectiveness of NRT (patch) in smokers who had relapsed after previous patch use (Gourlay 1995). Although this study did not find any difference between active treatment and placebo for continuous abstinence, they did find a small increase in quitters in the patch group using their predetermined endpoint of abstinence in the 28 days before assessment. The absolute quit rates were low (Comparison 11).
Cost of therapy:
One study comparing the effectiveness of free and purchased patch found no significant difference in quit rates between the two conditions (Hays 1999) (Comparison 12).
Comparison with bupropion:
Nicotine patch and placebo tablet was significantly less effective than bupropion and placebo patch in one study (Jorenby 1999). The combination of bupropion and nicotine patch was significantly more effective than placebo alone or patch alone, but not significantly different from bupropion alone (Comparison 11).
Harm reduction:
In the single trial of the use of nicotine inhaler for smoking reduction, the odds of smokers reducing their consumption to <50% of their baseline level was significantly improved (OR 3.59, 95% CI 1.58 - 8.31) two years later.
Adverse Effects:
No attempt was made in this overview to synthesize quantitatively the incidence of the various side effects reported with the different NRT preparations. This was because of the extensive variation in reporting the nature, timing and duration of symptoms. However, the major side effects usually reported with nicotine gum, including hiccups, gastrointestinal disturbances, jaw pain, and orodental problems, are not seen with transdermal patch (Fiore 1992; Palmer 1992). The only side effect which appears to interfere with use of the patch is skin sensitivity and irritation; this may affect up to 54% of patch users, but it is usually mild and rarely leads to withdrawal of patch use (Fiore 1992). The major side effects reported with the nicotine inhaler and nasal spray are related to local irritation at the site of administration (mouth and nose respectively). For example, symptoms such as throat irritation, coughing, and oral burning were reported significantly more frequently with subjects allocated to the nicotine inhaler than to placebo control (Schneider 1996), however none of the experiences were reported as severe. With the nasal spray, nasal irritation and runny nose are the most commonly reported side effects. Nicotine sublingual tablets have been reported to cause hiccups, burning and smarting sensation in the mouth, sore throat, coughing, dry lips and mouth ulcers (Wallstrom 1999). A review of adverse effects based on 35 trials with over 9,000 participants did not find evidence of excess adverse cardiovascular events assigned to nicotine patch, and the total number of such events was low (Greenland 1998).
There has been concern about the safety of NRT in smokers with cardiac disease (TNWG 1994). A trial of nicotine patch (Joseph 1996) which recruited smokers aged over 45 with at least one diagnosis of cardiovascular disease found no evidence that serious adverse events were more common in smokers in the nicotine patch group. Events related to cardiovascular disease such as an increase in angina severity occurred in approximately 16% of patients, but did not differ according to whether or not patients were receiving NRT.
| DISCUSSION |
This overview provides reliable evidence from over 35,600 smokers that offering NRT to dependent smokers is more effective in helping them to stop smoking than when NRT is not offered or if placebo is used. This applies to all forms of NRT and is independent of any variations in methodology or design characteristics of trials included in the overview.
Comparisons between the relative effectiveness of the different forms of NRT can only be made indirectly. The increased odds of not smoking at 6 to 12 months follow-up were greatest with the intranasal nicotine spray and nicotine inhaler; however there are still only a small number of trials involving these delivery systems and confidence intervals are wide. In making indirect comparisons it should be noted that most of the trials included in the comparison of nicotine gum versus placebo used 2 mg gum. The pooled OR of abstinence in the trials which directly compared 4 mg versus 2 mg gum in highly dependent smokers found a significant benefit in favour of 4 mg gum (2.18, 95% CI 1.48 - 3.17). There have been no direct comparisons of the relative effectiveness between 4 mg gum and nicotine patch. In one study directly comparing inhaler to patch the patch was nonsignificantly more effective (Tonnesen 2000). One study in which smokers were randomized to nicotine gum, patch, spray or inhaler found no significant differences in abstinence rates after 12 weeks (Hajek 1999).
There is some evidence that using combinations of NRT products is better than one product alone. Updated US clinical practice guidelines (Fiore 2000) recommend the use of nicotine patch with another form of NRT taken ad libitum as a second-line therapy for patients unable to quit on a single type of NRT or bupropion. However the strength of evidence was recognised as less than optimal due to the clinical heterogeneity of the studies in the meta-analysis. Whilst two further trials have been published since then, there is still a lack of evidence that combinations produce significant additional benefits. As a recent review (Sweeney 2001) points out, it is not yet clear whether any benefit of combination therapy is due to the sensory effects provided by multiple types of delivery systems, to the higher percentage of nicotine substitution achieved, or some combination of these and other factors. The review also notes that not enough is known for NRT products to be appropriately labelled so that nonexperts can be guided in the most safe and effective use of combinations of products.
All forms of NRT were associated with a high relapse rate in the first 3 months. Minimizing this relapse is important if long-term smoking cessation rates are to be substantially improved. There is suggestive evidence (Gourlay 1995) that repeated use of NRT in patients who have relapsed after an initial course may produce further quitters, though the absolute effect is small.
Clinical Setting:
The two factors which appear to be the major determinants of the effectiveness of NRT are the setting in which it is offered, and the smoker's level of dependency on nicotine. Both of these factors have been recognized in previous reviews (Gourlay 1990; Lam 1987). The nature and flexibility of the dosage regimen appears to be a far less important determinant of the effectiveness of NRT.
Nicotine gum and transdermal patches were more effective when offered to volunteer smokers recruited from the community or those attending specialized clinics than if offered to smokers in primary-care. These findings are likely to be partly explained by the high motivation to quit among many of the smokers in the community who volunteer for trials in response to media advertisements and, similarly, among those participants who are recruited as a result of their attendance at specialized smoking-cessation clinics. The latter group also have access to trained therapists who specialise in assisting smokers to quit. However, given the limited number of specialized smoking-cessation clinics, access will be restricted to a small proportion of smokers wanting help to quit.
In contrast, most of the smokers recruited into trials conducted in primary-care settings were unselected, and hence, may be less motivated to quit. In addition, the treating physician or practice nurse had frequently received little training in smoking cessation skills. As a result, compliance with NRT among smokers treated with in primary-care is reported to be lower than in other settings (Lam 1987). There has been some debate about the amount of evidence for efficacy of NRT when obtained over the counter without advice or support from a health care professional (Walsh 2000, Hughes 2001, Walsh 2001). The small number of placebo controlled trials in OTC settings, support the conclusion that the relative effect of NRT is similar, although quit rates in both control and intervention groups have been low.
The poor result seen with use of nicotine gum in hospital-based patients was disappointing. Smokers recruited in this setting frequently had coexisting smoking-related diseases which, hopefully, would have acted as an added incentive to quit. In addition, their level of dependency on nicotine was generally high. However, it appears to be difficult to change behaviour in this group.
One trial of nicotine patch in pregnant women is now included in the review. Women still smoking after their first trimester were recruited, and they were followed up until one year post partum. No significant benefit of treatment was detected, although the confidence intervals do not exclude the possibility of benefit. Quit rates 1 year after delivery were 15% in the patch and 14% in the placebo group. Using quit rates at the final prenatal follow-up did not alter the conclusions, with rates of 28% versus 25%. Possible explanations for the lack of relative benefit may have been low compliance with patch use, and the intensive cessation counselling offered to all participants. A second trial of the patch in pregnancy (Kapur 2001) is not included here since follow up was only to end of treatment at 12 weeks. In this trial 0/13 in the placebo group quit compared to 4/17 (24%) in the active treatment group. Enrolment was ended early in this study because of a possible adverse event in the placebo arm. A recent study measuring nicotine metabolism in smokers during their pregnancy and postpartum has suggested that nicotine is metabolised more quickly by pregnant women and that this may affect the dose of NRT required. (Dempsey 2002). More studies are needed to establish whether or not NRT does aid quitting in pregnancy and what effects there are on birth outcomes (Benowitz 2000)
Intensity Of Additional Support:
In previous versions of this overview there seemed to be a clear trend towards a lower OR for abstinence (NRT versus control) in trials which included high intensity support programmes than in those with low intensity support. NRT had a relatively greater effect when given with minimal support even though the absolute increase in abstinence rates was larger when combined with high intensity support. The trend remains, but the overlap in the confidence intervals are such that this could have arisen by chance.
It is important that smokers do not misinterpret these results by believing that NRT offers an easy option 'medical cure' for the far more complex problem of addictive behaviour. Almost all the trials in this review included some form of additional support together with the use of NRT as part of the intervention. In the case of trials in an 'over the counter' settings the adjunctive support was limited. The absolute probability of abstinence for an individual is still low, irrespective of what support strategies are used and whether or not they include use of NRT. Many smokers will therefore need to have multiple attempts to quit using a variety of strategies before they finally succeed. Falsely raising the expectations of smokers who purchase these products 'over-the counter' without at least providing minimal support and an adequate explanation of the limitations of using NRT may be counterproductive in the long term.
Dependency On Nicotine:
The benefit of using nicotine gum in smokers with high levels of dependency on nicotine has been previously recognized (Gourlay 1990; Lam 1987; Tang 1994). In such patients, the 4mg gum is significantly more effective than the lower dose.
Direct comparison with non-nicotine pharmacotherapies:
There is evidence from one large study that bupropion is more effective than nicotine patch. A combination of NRT and bupropion was not found to be significantly more effective than bupropion alone.
Harm reduction:
Using NRT to help smokers who cannot quit could help them reduce the number of cigarettes smoked and therefore decrease the harmful effects of smoking. It is not clear what reduction in consumption is needed for a clinically useful health benefit. Taking a >50% self reported reduction confirmed by some reduction in carbon monoxide levels as a cut off, one trial has shown a benefit from nicotine inhaler.
Methodological Limitations:
There are two possible methodological limitations of this overview which need to be borne in mind: use of tabulated data predominantly derived from published reports (Stewart 1993) and publication bias (Simes 1986). We tried to partly address any shortcomings from having limited our analysis to tabulated data by approaching investigators, where necessary, to obtain additional unpublished data or to clarify areas of uncertainty. Although steps were taken to minimize publication bias by writing to the manufacturers of NRT products when this review was first prepared , the response was poor and we have not repeated this exercise. It is therefore possible that there are some unpublished trials, with less favourable results, that we have not identified despite our systematic efforts to do so. Indeed, a recent statistical analysis (Egger 1997; Egger personal communication) suggests that this is the case. Using a regression method to assess the symmetry of funnel plots, they showed evidence of asymmetry (and hence possible publication bias) for both nicotine chewing gum and transdermal patches. For the nicotine inhaler we are aware of one unpublished trial with a non significant result. The practical effect of this is that the magnitude of the effectiveness of nicotine replacement may be smaller than our estimates suggest.
| REVIEWER'S CONCLUSIONS |
1. All of the commercially available forms of NRT (nicotine gum, transdermal patch, nicotine nasal spray, nicotine inhaler and nicotine sublingual tablet) are effective as part of a strategy to promote smoking cessation. They increase long term quit rates approximately 1.5 to 2 fold regardless of setting. Use of NRT should be preferentially directed to smokers who are motivated to quit (as demonstrated by their initiative to request assistance) and have high levels of nicotine dependency. There is little evidence about the role of NRT for individuals smoking less than 10-15 cigarettes/day.
2. The choice of which form to use should reflect patient needs, tolerability, and cost considerations. Patches are likely to be easier to use than gum or nasal spray in primary care settings.
3. Eight weeks of patch therapy is as effective as longer courses and there is no evidence that tapered therapy is better than abrupt withdrawal. Wearing the patch only during waking hours (16 hours/day) is as effective as wearing it for 24 hours/day.
4. If gum is used, it may be offered on a fixed dose or ad lib basis. For highly dependent smokers, or those who have failed with 2mg gum, 4mg gum should be offered.
5. There is some evidence of a small benefit from combining the nicotine patch with a form allowing ad lib dosing compared to use of a single form. Use of combination therapy may be considered for patients who have been unable to quit using a single type of NRT.
6. There is borderline evidence that there is a small benefit from use of the nicotine patch at doses higher than 22mg/24 hours, or 15mg/16 hours compared to the standard dose patch. Use of these may be considered for heavy smokers (>= 30/day) or for patients relapsing because of persistent craving and withdrawal symptoms on standard dose therapy (Hughes 1995).
7. The effectiveness of NRT appears to be largely independent of the intensity of additional support provided to the smoker. Provision of more intense levels of support, although beneficial in facilitating the likelihood of quitting, is not essential to the success of NRT.
8. There is minimal evidence that a repeated course of NRT in patients who have relapsed after recent use of nicotine patches will result in a small additional probability of quitting.
9. NRT does not lead to an increased risk of adverse cardiovascular events in smoker with a history of cardiovascular disease.
10. Nicotine patch was less effective than bupropion in one trial. However any decision about which pharmacotherapies to use should take into account potential adverse effects as well as benefits.
11. Finally, marketing claims by manufacturers of NRT products should reflect these points and avoid the possible misunderstanding by health professionals and members of the public that any of these products alone offer a magical 'cure' for the smoking habit.
Further research is required in several areas:
1. Direct comparisons between the various forms of NRT and between different doses and durations of treatment.
2. Use of combinations of different forms of NRT.
3. Direct comparisons between NRT and other products such as bupropion.
| ACKNOWLEDGEMENTS |
Mark Lodge assisted in the preparation of the initial version of this review. Ruth Ashenden provided technical support in updating information. Drs. Tjeder-Burton, Campbell, Hjalmarson, Fagerstrom, Mori, Glover, Hughes, Fortmann, Killen and Varady cooperated with our requests for clarification of previously reported data. The Imperial Cancer Research Fund Library Services assisted in obtaining articles. Z. Ilic, and L. Silagy assisted with translation of foreign-language reports. P. Yudkin provided statistical advice. Marc Mooney provided copies of two papers we had not been able to obtain. Rafael Perera assisted with data extraction.
| POTENTIAL CONFLICT OF INTEREST |
CS has received funds for consultancy work undertaken (at various times) on behalf of Pharmacia and Upjohn, Marion Merrell Dow, Glaxo Wellcome and SmithKline Beecham. G. Fowler and D. Mant were involved in a trial of transdermal nicotine (ICRF 1994).
| NOTES |
Prof Chris Silagy died in December 2001. In recognition of his major contribution to the review he will remain listed as first author of this and future updates. The contact author for the review is now Lindsay Stead.
| TABLES |
|
|
|
| REFERENCES |
| COMMENTS AND CRITICISMS |
The comment states that NRT is not more effective than abrupt cessation. We summarise the supporting arguments and our response to each below:
Author's Reply:1. Pierce & Gilpin (Pierce JP, Gilpin EA. Impact of over-the-counter sales on effectiveness of pharmaceutical aids for smoking cessation. JAMA 2002;288:1260-4) found no difference in long-term cessation rates between those who did and who did not use NRT.
This point is addressed by in a letter commenting on the study (Stead LF et al. Effectiveness of over-the-counter nicotine replacement therapy. JAMA 2002;288:3109-10). The main limitation of their study is that the comparison between groups of people who chose or did not chose to use NRT, These two groups probably differ in many respects related to their chance of successful quitting, and it is impossible to adjust for these possible confounders. Therefore the conclusions of the study are stronger than the evidence justifies.
The criticism authors also cite the Minnesota insurance review (Boyle RG et al. Does insurance coverage for drug therapy affect smoking cessation? Health Affairs 2002 Nov-Dec;21:162-8) but it does not seem to give further support to the point made. The main finding of Boyle et al was that introducing an insurance benefit did not increase use of NRT.
2. In the real-world those relying exclusively upon NRT are relapsing and dying at pre-NRT rates.
This is an assertion which is not supported by evidence.
3. NRT study instruction is designed and sequenced in order to foster device transfer. In fact the placebo group must be deprived of critical abrupt cessation instructional tips because if given and followed many could have a negative impact upon the active group.
The review does not make the assertion or implication attributed to it. In the studies involving behavioural support as well as active versus placebo NRT, both active and placebo groups are typically given instructions designed to maximise their chances of success. In these circumstances NRT if anything shows a larger advantage over placebo than it does in minimal support settings. If it is being asserted that placebo groups are being deprived of progressive cigarette weaning or some form of lapse management strategy, there is no evidence to suggest that this is approach is effective.
4. The duration of abstinence for NRT groups should begin from the time they stop using NRT.
In response to this it should be noted that it is cigarettes which are causing the harm to health and the aim is to help
people stop smoking. Secondly, studies that have followed up smokers long term show that the medication genuinely improves long-term cessation rates and does not simply set the relapse clock back by the time period when nicotine replacement is being used.
5. There are clinic programmes achieving success rates at least as good as those using NRT.
It is necessary to make direct comparisons ensuring that the same criteria are applied to both groups to be able to draw conclusions.
Finally it must be noted that the Cochrane review shows that NRT is estimated to help some 7% smokers to stop long term who would not have stopped had they used a similar approach but without NRT. This effect is small but given the health benefits from stopping smoking it is a highly cost-effective life-preserving medication. That is not to say that other interventions, including a different kind of behavioural intervention that was incompatible with NRT could not get better results. However, it is not enough just to assert the possibility; with so many lives at stake it would be imperative to demonstrate the effectiveness of such approaches.
Contributors:Comment by John R. Polito. Response by Tim Lancaster & Lindsay Stead on behalf of review authors. Criticism editor Robert West.
| GRAPHS |
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| COVER SHEET |
|
| SOURCES OF SUPPORT |
| SYNOPSIS |
All forms of nicotine replacement therapy can help people quit smoking, almost doubling long term success rates
Nicotine replacement therapy (NRT) aims to reduce withdrawal symptoms associated with stopping smoking by replacing nicotine in the blood. NRT is available as chewing gum, patches for the skin, nose spray, inhalers, and tablets. The review of trials found that all these forms of NRT made it more likely that a person's attempt to quit smoking would succeed. There is no evidence that one form of NRT is better than any other. NRT works with or without additional counselling.
| Index Terms |
Mesh check words: Human