 Advertisement | ||||||||
 | ||||||||
|
||||||||
 |
 |
 |
 |
 |
 |
|
||
Electronic letters to:
|
|
Electronic letters published:
![[Read eLetter]](/icons/misc/sectionDOWN.gif){kind=icon}
Response to Letter from Dr. Sestini
Re: Three of the authors respond to Piersante Sestini
Race and Smoking Cessation Therapies
Response to Letter “Meta-analysis Rooted in Expectations Not Science”
Response to Letter “Pharmacotherapy for Smoking Cessation”
Race and smoking cessation therapies
Three of the authors respond to Piersante Sestini
Two warnings and a correction
Comparison varenicline-bupropion
Meta-analysis rooted in expectations not science
Pharmacotherapy for smoking cessation
|
|
||||||||||||||
|
Mark J Eisenberg Jewish General Hospital/McGill University
Send letter to journal:
mark.eisenberg{at}mcgill.ca Mark J Eisenberg
|
We thank Dr. Sestini for his follow-up letter regarding our recently
published meta-analysis examining smoking cessation therapies (1).
Unfortunately, our Figure 7 contains 2 independent errors. The first, a
labeling error that occurred during copyediting, resulted in the 2nd and
3rd columns being labeled ‘Placebo’ and ‘Treatment’ rather than
‘Bupropion’ and ‘Varenicline’, respectively. The second, an analytical
error, resulted in varenicline being compared with placebo rather than
bupropion, our intended comparator. Consequently, the results of this
comparison in our meta-analysis are incorrect. We have included a
corrected version of this figure. Based on this revised analysis,
varenicline may increase the proportion of patients who are abstinent
compared with bupropion however, the credible interval is wide, preventing
strong conclusions from being drawn (odds ratio: 1.40, 95% CrI: 0.75,
2.66). All other analyses presented in the paper have been verified and
are correct. We deeply regret this error and thank Dr. Sestini for his
help in identifying it.
Mark J. Eisenberg, MD MPH, Kristian B. Filion MSc, Daniel Yavin BSc, Patrick Bélisle MSc, Salvatore Mottillo BSc, Lawrence Joseph PhD, André Gervais MD, Jennifer O’Loughlin PhD, Gilles Paradis MD MSc, Stéphane Rinfret MD MSc, Louise Pilote MD MPH PhD References: 1) Eisenberg MJ, Filion KB, Yavin D, Bélisle P, Mottillo S, Joseph L, Gervais A, O'Loughlin J, Paradis G, Rinfret S, Pilote L. Pharmacotherapies for smoking cessation: a meta-analysis of randomized controlled trials. CMAJ. 2008;179:135-44. Conflict of Interest:Mark Eisenberg is a member of the Varenicline Advisory Board of Pfizer Canada Inc. André Gervais has received speaker fees and consultant fees from Pfizer Canada Inc., as well as travel assistance from Pfizer Canada Inc. to attend a conference on the treatment of tobacco dependence. No competing interests declared by Kristian Filion, Daniel Yavin, Patrick Bélisle, Salvatore Mottillo, Lawrence Joseph, Jennifer O'Loughlin, Gilles Paradis, Stephane Rinfret or Louise Pilote. |
||||||||||||||
|
|
||||||||||||||
|
Piersante Sestini University of Siena, Italy
Send letter to journal:
sestini{at}unisi.it Piersante Sestini
|
I apologize for insisting, but changing the labels in figure 7 without also correcting the numbers would make it in my opinion even worst. The figures on the left column are in fact the same as the ones for placebo in figure 2, while they should be the ones from the right column (treatment with bupropion) of that figure.
For example: Data from Gonzales et al.
Current Figure 7
The same apply for all the other studies in that figure
Regards, Conflict of Interest:None declared |
||||||||||||||
|
|
||||||||||||||
|
Kristian B. Filion Jewish General Hospital/McGill University
Send letter to journal:
kristian.filion{at}mail.mcgill.ca Kristian B. Filion
|
We thank Ms. Shivanthika and Dr. McLachlan for their letter regarding the importance of considering race when examining the efficacy of smoking cessation therapies. We agree that race, a marker for genetic variants, and ethnicity, a marker for social, geographical, and cultural factors, may alter the efficacy of smoking cessation pharmacotherapies. In our recent meta-analysis (1), we did not examine potential interactions between smoking cessation pharmacotherapies and race or ethnicity. We were limited to published, aggregate data and thus were unable to account for these variables at the patient-level. Furthermore, there were an insufficient number of trials per intervention to include these data in aggregate form as part of a meta-regression. Additional pharmacogenomic studies are required to identify genetic factors that may alter the efficacy of smoking cessation pharmacotherapies and to enable physicians to better prescribe individualized smoking cessation strategies. Kristian B. Filion MSc, Daniel Yavin BSc, Patrick Bélisle MSc, Salvatore Mottillo BSc, Lawrence Joseph PhD, André Gervais MD, Jennifer O’Loughlin PhD, Gilles Paradis MD MSc, Louise Pilote MD MPH PhD, Stéphane Rinfret MD MSc, Mark J. Eisenberg, MD MPH References: 1) Eisenberg MJ, Filion KB, Yavin D, Bélisle P, Mottillo S, Joseph L, Gervais A, O'Loughlin J, Paradis G, Rinfret S, Pilote L. Pharmacotherapies for smoking cessation: a meta-analysis of randomized controlled trials. CMAJ. 2008;179:135-44. Conflict of Interest:Mark Eisenberg is a member of the Varenicline Advisory Board of Pfizer Canada Inc. André Gervais has received speaker fees and consultant fees from Pfizer Canada Inc., as well as travel assistance from Pfizer Canada Inc. to attend a conference on the treatment of tobacco dependence. No competing interests declared by Kristian Filion, Daniel Yavin, Patrick Bélisle, Salvatore Mottillo, Lawrence Joseph, Jennifer O'Loughlin, Gilles Paradis, Stephane Rinfret or Louise Pilote. |
||||||||||||||
|
|
||||||||||||||
|
Kristian B. Filion Jewish General Hospital/McGill University
Send letter to journal:
kristian.filion{at}mail.mcgill.ca Kristian B. Filion
|
We thank Dr. Polito for his discussion of the importance in blinding in clinical trials. Blinding is undoubtedly a key component to the validity of inferences drawn from randomized controlled trials. For this reason, it is included in clinical trial quality assessment tools such as the Jadad scale (1). In our meta-analysis of randomized controlled trials of smoking cessation pharmacotherapies (2), we restricted our study to double-blind trials to only include trials of the highest quality. We agree with Dr. Polito that maintaining blinding may be difficult, particularly in the presence of withdrawal symptoms. However, although the importance of blinding is well established, assessing the integrity of blinding remains more controversial. In a recent analysis of randomized controlled trials that report tests for the success of blinding, Hróbjartsson and colleagues found that less than 2% of trials reported testing for blinding (3). Others have reported this proportion to be as high as 8% (4). One reason few trial reports include tests for blinding is that it is unclear what these tests actually measure. Some have argued that such tests are often biased (5). These tests often do not test blinding but rather test the participant’s or physician’s belief regarding the efficacy of the treatment (6). Thus, the fact that patients who are randomized to placebo and return to smoking correctly guess their allocated treatment may not be a reflection of the blinding itself but of their belief that placebo is not efficacious. Blinding plays an important role in maintaining the validity of clinical trial results. However, improved methods are needed to assess the success of this blinding. Until new methods are developed, incomplete blinding may be a potential limitation of most clinical trials. Kristian B. Filion MSc, Daniel Yavin BSc, Patrick Bélisle MSc, Salvatore Mottillo BSc, Lawrence Joseph PhD, André Gervais MD, Jennifer O’Loughlin PhD, Gilles Paradis MD MSc, Louise Pilote MD MPH PhD, Stéphane Rinfret MD MSc, Mark J. Eisenberg, MD MPH References: 1) Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, McQuay HJ. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996; 17:1-12. 2) Eisenberg MJ, Filion KB, Yavin D, Bélisle P, Mottillo S, Joseph L, Gervais A, O'Loughlin J, Paradis G, Rinfret S, Pilote L. Pharmacotherapies for smoking cessation: a meta-analysis of randomized controlled trials. CMAJ. 2008;179:135-44. 3) Hróbjartsson A, Forfang E, Haahr MT, Als-Nielsen B, Brorson S. Blinded trials taken to the test: an analysis of randomized clinical trials that report tests for the success of blinding. Int J Epidemiol. 2007;36:654-63. 4) Fergusson D, Glass KC, Waring D, Shapiro S. Turning a blind eye: the success of blinding reported in a random sample of randomised, placebo controlled trials. BMJ. 2004;328:432. 5) Haynes RB, Sackett DL, Guyatt GH, Tugwell P. Clinical epidemiology, 3rd edition. How to do clinical practice research. Philadelphia: Lippincott Williams & Wilkins, 2005. 6) Sackett DL. Commentary: Measuring the success of blinding in RCTs: don't, must, can't or needn't? Int J Epidemiol. 2007;36:664-5. Conflict of Interest:Mark Eisenberg is a member of the Varenicline Advisory Board of Pfizer Canada Inc. André Gervais has received speaker fees and consultant fees from Pfizer Canada Inc., as well as travel assistance from Pfizer Canada Inc. to attend a conference on the treatment of tobacco dependence. No competing interests declared by Kristian Filion, Daniel Yavin, Patrick Bélisle, Salvatore Mottillo, Lawrence Joseph, Jennifer O'Loughlin, Gilles Paradis, Stephane Rinfret or Louise Pilote. |
||||||||||||||
|
|
||||||||||||||
|
Kristian B. Filion Jewish General Hospital/McGill University
Send letter to journal:
kristian.filion{at}mail.mcgill.ca Kristian B. Filion
|
We thank Dr. Brewster for her letter highlighting the importance of behavioral therapies for smoking cessation. The objective of our meta- analysis was to examine the efficacy of smoking cessation pharmacotherapies (1). We found that most pharmacotherapies approximately double the proportion of patients who are abstinent at 1-year. Consequently, our results support the use of pharmacotherapies as a first- line therapy for smoking cessation. However, as discussed by Dr. Brewster, behavioral interventions are also efficacious at promoting smoking abstinence. Recently completed meta-analyses suggest that behavioral interventions are associated with a relative increase in the proportion of patients who are abstinent between 1.3 and 2.0 (2-6). Practice guidelines now recommend the use of combination therapy, including both behavioral therapy and pharmacotherapy (7). Thus, both behavioral and pharmacotherapies should be used to effectively promote smoking cessation. Kristian B. Filion MSc, Daniel Yavin BSc, Patrick Bélisle MSc, Salvatore Mottillo BSc, Lawrence Joseph PhD, André Gervais MD, Jennifer O’Loughlin PhD, Gilles Paradis MD MSc, Louise Pilote MD MPH PhD, Stéphane Rinfret MD MSc, Mark J. Eisenberg, MD MPH References: 1) Eisenberg MJ, Filion KB, Yavin D, Bélisle P, Mottillo S, Joseph L, Gervais A, O'Loughlin J, Paradis G, Rinfret S, Pilote L. Pharmacotherapies for smoking cessation: a meta-analysis of randomized controlled trials. CMAJ. 2008;179:135-44. 2) Lancaster T, Stead LF. Individual behavioural counselling for smoking cessation. Cochrane Database Syst Rev 2005;(2):CD001292. 3) Rice VH, Stead LF. Nursing interventions for smoking cessation. Cochrane Database Syst Rev 2004;(1):CD001188. 4) Stead LF, Lancaster T. Group behaviour therapy programmes for smoking cessation. Cochrane Database Syst Rev 2005;(3):CD001007. 5) Stead LF, Lancaster T, Perera R. Telephone counselling for smoking cessation. Cochrane Database Syst Rev 2006;(1):CD002850. 6) Fiore MC. US public health service clinical practice guideline: treating tobacco use and dependence. Respir Care 2000; 45(10):1200-1262. Fiore MC, Jaén CR, Baker TB, et al. Treating Tobacco Use and Dependence: 2008 Update. Rockville, MD: US Dept of Health and Human Services; May 2008. http://www.surgeongeneral.gov/tobacco/treating_tobacco_use08.pdf. Accessed August 31st, 2008. Conflict of Interest:Mark Eisenberg is a member of the Varenicline Advisory Board of Pfizer Canada Inc. André Gervais has received speaker fees and consultant fees from Pfizer Canada Inc., as well as travel assistance from Pfizer Canada Inc. to attend a conference on the treatment of tobacco dependence. No competing interests declared by Kristian Filion, Daniel Yavin, Patrick Bélisle, Salvatore Mottillo, Lawrence Joseph, Jennifer O'Loughlin, Gilles Paradis, Stephane Rinfret or Louise Pilote. |
||||||||||||||
|
|
||||||||||||||
|
Craig S McLachlan UTS, Assistant Professor
Send letter to journal:
reperfusion{at}hotmail.com Craig S McLachlan
|
Shivanthika M Jayalath BSc and Craig S McLachlan PhD MPH We read with interest the article by Eisenberg et al. [1], in which a meta analysis of published trials was performed to evaluate suitable pharmacotherapies for smoking cessation. The study showed positive results for future interventions for smoking cessation, validating the efficacy of antidepressant like drugs such as varenicline and bupropion [1]. However, it is also interesting that the study failed to take into account race that may in some circumstances diminish the impact of an antidepressant in smoking cessation. Indeed some populations such as Middle East and Africa, are somewhat ignored in the study [1]. According to Tsai et al. [2] and Aubin et al. [3], smoking cessation in Western trials include an ethnicity of < 3 % Asians. The failure to mention specific geographical populations is important due to evidence of pharmacogenetic, body weight and environmental differences among ethnicities - these factors may have direct effects on the outcome of smoking cessation therapies [2; 4]. According to Tsai et al. [3], nausea as a result of antidepressants, are remarkably higher in the Asian populations compared to that of Caucasians in the United States. This difference in nausea was believed to be a result of baseline body weight between the two race population groups. Chen et al. [5] also expresses the view that Asians develop side effects to antidepressants at lower doses compared to other ethnic groups. Asians are also known to metabolise drugs via CYP4502D6 more slowly and their relatively low body weights contribute toward more remarkable side effects when compared to non- Asians [5]. Most importantly the measured continuous smoking abstinence rate with varenicline in Japan is significantly lower than in the USA. In summary, it is important that race be considered when considering the efficacy of smoking pharmacological interventions such as varenicline. Analysis of other studies, should likely elicit the importance of stating the treated population(s), before confirming the efficacy of a pharmacotherapy for smoking cessation. References [1] Eisenberg, M.J., Filion, K.B., Yavin, D., Belisle, P., Mottillo, S., Joseph, L., Gervais, A., O’Loughlin, J., Paradis, G., Rinfret, S. & Pilote, L. 2008, ‘Pharmacotherapies for smoking cessation: a meta- analysis of randomized controlled trials’, CMAJ, vol 179, pp 135-144 [2] Tsai, S., Cho, H., Cheng, H., Kim, C., Hsueh, K., Billing, C.B. & Williams, K.E. 2007, ‘ A randomized placebo-controlled trial of varenicline a selective α4β2 nicotinic acetylcholine receptor partial agonist, as a new therapy for smoking cessation in Asian smokers’, Clinical Therapeutics, vol 29, pp 1027-1038 [3] Aubin, H.J., Bobak, A., Britton, J.R., Oncken, C., Billing, C.B., Gong, J., Williams, K.E. & Reeves K.R. 2008, ‘Varenicline versus transdermal nicotine patch for smoking cessation: results from a randomised open-label trial’, Thorax, vol 63, pp 717-724 [4]Nakamura, M., Oshima, A., Fujimoto, Y., Maruyama, N., Ishibashi, T. & Reeves, K.R. 2007, ‘Efficacy and tolerability of varenicline, and α4β2 nicotinic acetylcholine receptor partial agonist in a 12- week randomized placebo-controlled dose-response study with 40-week follow -up for smoking cessation in Japanese smokers’, Clinical Therapeutics, vol 29,pp 1040- 1056 [5] Chen, J., Barron, C., Lin, K. & Chung, H. 2002, ‘Prescribing medication for Asians with mental disorders, Western Journal of Medicine, vol 176, pp 271-275 Conflict of Interest:None declared |
||||||||||||||
|
|
||||||||||||||
|
Kristian B. Filion Divisions of Cardiology and Clinical Epidemiology, Sir Mortimer B. Davis Jewish General Hospital
Send letter to journal:
kristian.filion{at}mail.mcgill.ca Kristian B. Filion
|
We thank Dr. Sestini and colleagues for their letter regarding our recently published meta-analysis of smoking cessation pharmacotherapies (1). In their letter, they raise a number of important issues that we address below. First, Dr. Sestini and colleagues discuss our use of Bayesian techniques and, in doing so, touch upon several different issues: cumulative versus non-cumulative plots, Bayesian versus non-Bayesian meta- analyses methods, and random effects versus non-random effects (or hierarchical versus non-hierarchical) meta-analysis models. In any meta- analysis, three completely independent decisions can be made. One can decide whether to use a Bayesian or frequentist approach, a decision independent of whether one chooses to use simple pooling or a random effects (or hierarchical) model, and these first two decisions are again distinct from whether one decides to present a cumulative or non- cumulative forest plot. We will consider each of these decisions in turn. For modeling flexibility and many other reasons too numerous to discuss here (2), Bayesian methods are increasingly being used to analyse medical data, including meta-analyses. While we chose to run a Bayesian analysis, as Sestini et al. agree, similar numerical inferences can be obtained regardless of inferential paradigm chosen. This can be seen, for example, in the pooled confidence interval in their figure being very similar to our pooled credible interval, with our interval being slightly wider to reflect the extra between-study uncertainty we include in our hierarchical model and which non-hierarchical models omit. We chose to use a hierarchical model because simple pooling assumes that the results from every single study has an identical odds ratio (OR), which is unrealistic (for example, variations can be induced by the different methods employed by each study and varying population characteristics). Our hierarchical model more realistically assumes that each study has its own individual OR value, with the log(OR)s from the different studies not assumed to all lie on a single point, but to have a statistical distribution. One of the well-known advantages of hierarchical models is that, assuming a reasonable distribution has been used at the hierarchical level, one obtains improved estimates of each study OR in terms of overall Mean Square Error (MSE). Thus, whether frequentist or Bayesian, if one allows a random effects model to correctly "shrink" individual estimates, one improves both the individual level ORs and the overall meta-analytic estimate. Small studies “shrink” more than larger studies, hence the different estimates one can see in comparing our estimate, for example, of the small Muramoto study to theirs. The use of hierarchical models is quite separate from running a cumulative or non-cumulative meta-analysis, which again can be done from either a Bayesian or frequentist viewpoint. In this study, we did non- cumulative meta-analyses. Thus, our figures present not "combined" or "iterated" estimates, but rather individual estimates for each study, except for the final overall estimate at the bottom, which is by definition combined. Any differences between our figures and simple unpooled analysis from each study is that, while we too present results for each study alone, the estimates are taken from our hierarchical model. This results in individual study estimates that lead to smaller MSE than more simple analyses that ignore the random effects. One can certainly discuss whether one prefers random effects or non-random effects results to be reported, but one should not confuse random effect estimates with a cumulative meta-analysis, which we did not do here. Second, they have suggested that we have ‘double counted’ controls from trials with multiple treatment arms. Although many ‘plug-and-play’ meta-analyses programs are unable to properly account for the controls from trials with multiple treatment arms, we conducted our analyses using WinBUGS, a program that allow us to account for this issue. Consequently, we did not ‘double count’ any data and thus, in no way have we artificially inflated the number of controls. Each control group’s data were inputted and used only once, even in studies with multiple treatment groups. Finally, they have identified an error in our Figure 7 also discussed in an e-letter by Dr. Saad. Although we did indeed directly compare varenicline to bupropion, the columns in this figure are vaguely labeled as placebo and treatment. These column titles should read “Bupropion” and “Varenicline”, respectively. The text and figure legend correctly refer to this figure as a comparison of varenicline to bupropion. We regret this error and a correction to the labeling is forthcoming. Sincerely, Lawrence Joseph, PhD Department of Epidemiology, Biostatistics and Occupational Health, McGill University Kristian B. Filion, MSc Mark J. Eisenberg, MD MPH Divisions of Cardiology and Clinical Epidemiology, Sir Mortimer B. Davis Jewish General Hospital and McGill University Reference: 1) Eisenberg MJ, Filion KB, Yavin D, Bélisle P, Mottillo S, Joseph L, Gervais A, O'Loughlin J, Paradis G, Rinfret S, Pilote L. Pharmacotherapies for smoking cessation: a meta-analysis of randomized controlled trials. CMAJ. 2008;179:135-44. 2) Gelman A, Carlin JB, Stern HS, Rubin DB. Bayesian data analysis. Chapman & Hall/CRC, Boca Raton (Fl), 2004. |
||||||||||||||
|
|
||||||||||||||
|
Piersante Sestini, NONe Section of Respiratory Diseases, University of Siena
Send letter to journal:
sestini{at}unisi.it Piersante Sestini
|
Sir, we commend Eisenberg et al for their attempt to summarize the data of pharmacotherapies for smoking cessation using sophisticated statistical tools (1). However, the presentation of some of the data may mislead readers and some parts of the analysis are flawed.
The authors have used Bayesian statistics and readers may not be aware that what is presented in figures 2, 3, 4 and 5 are not the actual odds ratios of the single studies, but an iterative posterior estimate of the combined odds ratio after the addition of each study to the preceding value (2). This figure can be used to show how the combined OR converges to its final value, but not to describe the group of included studies, as it results in a considerable smoothing of the differences, and hence a false impression of homogeneity of results among studies. To illustrate this point, we include a figure of a frequentist meta-analysis performed using the same data provided in figure 2. The estimate of the combined odds ratio is almost identical to the corresponding Bayesian model, but the heterogeneity of results among studies is much clearer when the actual ORs are presented. We think that individual ORs should be presented,regardless how the combined OR is calculated.
A second concern is the inclusion of different arms of the same study as if they were separate studies, rather then combining them. This has the effect of artificially inflating the number of controls. This is particularly problematic in the meta-analysis on varenicline (figure 5), where 6 studies are analysed as if they were 13, and 941 of 2156 controls (44%) are duplicates.
Finally, the meta-analysis in figure 7, presented as a comparison between varenicline and bupropion is in fact a comparison between varenicline and placebo. Using bupropion data from figure 2 we estimated the actual combined OR as 1.4 with 95% CI 1.1-1.8 in favour of varenicline; still significant but less impressive. Hence the last sentence of the results and the second paragraph of page 141 are wrong. If we are correct, this will require correction in the journal.
Sincerely, Piersante Sestini MD Letizia Mori MD, Magda Kouri MD Lou Irving MD*
Smoking Cessation Centre. University Hospital Santa Maria alle Scotte, University of Siena, Italy and * Royal Melbourne Hospital, Melbourne, Australia
Conflict of Interest:None declared |
||||||||||||||
|
|
||||||||||||||
|
Emile Saad
Send letter to journal:
emile_saad01{at}yahoo.ca Emile Saad
|
Ragarding figure 7 in the article, I think the comparison is varenciline versus bupropion and not varenicline versus placebo. Am I right ? Conflict of Interest:None declared The authors respond:
We thank Dr. Saad for his interest in our article. As discussed in his letter, Figure 7 regrettably contains an error. The second and third columns of this figure should be labeled as “Bupropion” and “Varenicline”, respectively. As described in the text and the figure legend, this figure involves the direct comparison of the effect of varenicline and bupropion for smoking cessation. Sincerely, Mark J. Eisenberg, MD MPH Kristian B. Filion, MSc |
||||||||||||||
|
|
||||||||||||||
|
John R. Polito Editor WhyQuit.com
Send letter to journal:
john{at}whyquit.com John R. Polito
|
I submit that the validity of the authors' entire meta-analysis is grounded in a false premise, that researchers were somehow able to hide the onset of full-blown nicotine withdrawal from placebo group members, participants whose prior quitting history taught them exactly how it felt, and a way to hide from active group members withdrawal syndrome reduction. It has been four full years since Mooney's "Blind Spot" study told researchers that NRT studies were generally not blind as claimed.1 How hard would it have been in the Jorenby varenicline study to question still fresh memories and make an early determination of assignment beliefs of the roughly 83% of placebo group members who were smoking at two weeks?2 Why no blinding integrity assessments in any varenicline study to date? When Mooney's "Blind Spot" study is coupled with the authors' meta-analysis finding that pharmacology nearly always prevails in clinical trials, and then combined with the fact that pharmacology has failed to prevail over those quitting without it in nearly all real-world use surveys to date,3 it strongly suggests that chemical dependency treatment pharmacology research may be the only known area where blinding is impossible. Ignored common sense dictates that in regard to no other health condition, other than chemical dependency, will placebo group assignment guarantee that within 24 hours of cessation, participants will feel vastly worse than when randomized. Their hope of a 50/50 chance of weeks or months of free withdrawal syndrome reduction medication was demolished by onset of the same rising tide of anxieties, anger, dysphoria, concentration difficulty and sleep fragmentation felt during prior unaided attempts. Mooney warned researchers in 2004 that "the validity of NRT clinical trial results could be questioned" if future studies fail to assess the integrity of study blindness. This unheaded warning has equal application to every study evaluating any cessation chemical enhancing dopamine stimulation, or punishing study participation by placebo group assignment and withdrawal syndrome onset. How bad was it? The blinding analysis in Dar's 2005 study found that 3.3 times as many placebo group members correctly guessed that they had received placebo (54.5%) as guessed that they had received nicotine (16.4%).4 Although the Dar study focused on smoking reduction, Tonnesen's 1993 nicotine inhaler study produced strikingly similar findings in that 3.8 times as many in the placebo group correctly guessed placebo (58%) as guessed nicotine (15%). Among inhaler users, Tonnesen found that 3.5 times more correctly guessed inhaler (46%) as guessed placebo (13%).5 As the authors show, never in cessation history have so many different cessation products promised to double success. Yet the pharmacology cessation era, with mandatory U.S. Guideline use recommendations since June 2000, has heralded a screeching halt to decline in smoking rates.6 In January we watched the Ontario Medical Association revise and adopt a "Smoking Medications" position paper, with revisions that actually removed physician discretions.7 Although varenicline excitement should provide a brief cessation bump, Canadian health policymakers must someday confront the reality that toying with dopamine pathway stimulating chemicals is not more effective than teaching those hooked on nicotine how to quickly and more comfortably adapt to natural stimulation. One would think that the horrific life and death implications of even the remote possibility of having subjected smokers to two decades of expectations driven efficacy findings would have motivated scores of cessation researchers to want to prove or disprove the scientific foundation of their earlier work. Frankly, it has not happened. Is it fair to ask why? John R. Polito 1. Mooney M, et al, The blind spot in the nicotine replacement therapy literature: Assessment of the doubleblind in clinical trials [PubMed abstract], Addict Behav., 2004 June;29(4):673-84 2. Jorenby DE, et al, Efficacy of varenicline, an alpha4beta2 nicotinic acetylcholine receptor partial agonist, vs placebo or sustained-release bupropion for smoking cessation: a randomized controlled trial [free full-text], JAMA, 2006 Jul 5;296(1):56-63 3. UK NHS, Statistics on NHS Stop Smoking Services in England, April to December 2007 [full text - see Table 6], April 16, 2008; also see 2006 Unpublished U.S. National Cancer Institute Survey of 8,200 quitters, as reported in the Wall Street Journal, Page A1, February 8, 2007; also see Doran CM, et al, Smoking status of Australian general practice patients and their attempts to quit [PubMed Abstract], Addict Behav., 2006 May;31(5):758-66; also see, Ferguson J, The English smoking treatment services: one-year outcomes, Addiction [PubMed abstract], 2005 Apr;100 Suppl 2:59-69 [see Table 6]; also see, Alberg AJ, et al, Nicotine replacement therapy use among a cohort of smokers [PubMed abstract], J Addict Dis., 2005;24(1):101-13; also see, SmokeFree London, Tobacco In London, Facts and Issues, [online PDF report, see Figure 14], November 26, 2003; also see Boyle RG, et al, Does insurance coverage for drug therapy affect smoking cessation? [free full-text], Health Aff (Millwood), 2002 Nov-Dec;21(6):162-8; also see, Pierce JP, et al, Impact of Over-the-Counter Sales on Effectiveness of Pharmaceutical Aids for Smoking Cessation [free full text], JAMA, September 11, 2002;288:1260-1264. 4. Dar R, et al, Assigned versus perceived placebo effects in nicotine replacement therapy for smoking reduction in Swiss smokers [PubMed abstract], J Consult Clin Psychol., 2005 Apr;73(2):350-3. 5. Tønnesen P, et al, A double-blind trial of a nicotine inhaler for smoking cessation [PubMed abstract - blinding data provided by author], JAMA, 1993 Mar 10;269(10):1268-71. 6. CDC, Tobacco Use Among Adults - United States, Oct. 27, 2005, MMWR; 55(42);1145-1148 7. Ontario Medical Association, Rethinking stop-smoking medications: treatment myths and medical realities, [full text PDF]. Ont Med Rev 2008 Jan;75(1):22-34. Conflict of Interest:Editor of WhyQuit, an abrupt nicotine cessation forum, compensated by South Carolina for presenting 63 abrupt nicotine cessation prison programs during 2007-08 |
||||||||||||||
|
|
||||||||||||||
|
Joan M Brewster Dalla Lana School of Public Health, University of Toronto
Send letter to journal:
joan_brewster{at}camh.net Joan M Brewster
|
The meta-analysis by Eisenberg and colleagues(1) shows that pharmacotherapies are more effective than placebo in RCTs for smoking cessation, although the authors note that the overall cessation rate is disappointing. The trials differed in level of behavioural support in addition to the drug tested, but the authors required that behavioural support be the same for drug and placebo groups in each trial reviewed. A difficulty with trials comparing pharmacotherapy with placebo for smoking cessation is that participants expecting a drug may detect that they have received a placebo. More important, a placebo is not an accepted smoking cessation therapy. To support the recommendation that smoking cessation pharmacotherapies should be widely recommended and distributed, as advocated by Ebbert & Hays in an accompanying commentary(2), this approach needs to be shown to be more effective in the long term than good behavioural therapy, not only more effective than placebo. The 2008 update of the US Clinical Practice Guideline(3) concludes in its review that brief treatment and counselling are effective approaches to smoking cessation (as are medications). In addition, there is evidence that about three quarters of long-term former smokers quit smoking with no help at all(4,5). Taking these considerations together with the possibility of harmful side effects with pharmacotherapy, caution is needed in recommending to professionals that pharmacotherapy should be the first approach to smoking cessation. And smokers should not be given the erroneous impression that they cannot quit without the assistance of a drug. 1 Eisenberg MJ, Filion KB, Yavin D et al., Pharmacotherapies for smoking cessation: a meta-analysis of randomized controlled trials. CMAJ 2008;179(2):135-44. 2 Ebbert JO, Hays T. The missing link in tobacco control. CMAJ 2008;179(2):123-4. 3 Fiore MC et al. Clinical Practice Guideline: Treating tobacco use and dependence: 2008 update. 2008: US Department of Health and Human Services, Public Health Service. http://www.surgeongeneral.gov/tobacco/treating_tobacco_use08.pdf Accessed July 28, 2008 4 Brewster JM, Victor JC, Ashley MJ. Views of Ontarians about health professionals’ smoking cessation advice. Can J Pub Health 2007;98(5):395- 9. 5 Canadian Tobacco Use Monitoring Survey (CTUMS). Smoking in Canada: an overview 2002. http://www.hc-sc.gc.ca/hl-vs/alt_formats/hecs- sesc/pdf/tobac-tabac/research-recherche/stat/_ctums-esutc_fs- if/2002_overview-survol-eng.pdf Accessed July 28, 2008. Conflict of Interest:None declared |
||||||||||||||
| HOME | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | HELP | SEARCH |
|
|||||
|
Copyright 1995-2010, Canadian Medical Association. All rights reserved. ISSN 1488-2329
(e) 0820-3946 (p)
All editorial matter in CMAJ represents the opinions of the authors and not necessarily those of the Canadian Medical Association. |
|||||
eLetters: 
