Cost-effectiveness of pharmacogenetic testing to tailor smoking-cessation treatment

Article metrics


We evaluated the cost-effectiveness of a range of smoking cessation drug treatments, including varenicline, transdermal nicotine (TN), bupropion and the use of a genetic test to choose between TN and bupropion. We performed Monte Carlo simulation with sensitivity analysis, informing analyses with published estimates of model parameters and current prices for genetic testing and smoking-cessation therapy. The primary outcomes were discounted life-years (LY) and lifetime tobacco-cessation treatment costs. In the base case, varenicline treatment was optimal with an ICER, compared to bupropion, of $2985/LY saved. In sensitivity analyses, varenicline was in all cases (and bupropion in most cases) admissible; only under favorable assumptions was the genetically tailored approach competitive. Our data suggest that an untailored approach of treatment with either bupropion or varenicline is a cost-effective form of tobacco dependence treatment, but a tailored approach for selecting between TN and bupropion can be cost-effective under plausible assumptions.


Pharmacogenetic research holds promise for the future tailoring of pharmacotherapy to improve treatment outcomes, including treatment for psychiatric conditions and addictive behaviors.1, 2 For example, emerging data suggest that polymorphisms in central nervous system genes predict therapeutic response to antipsychotics,3 antidepressants,4 stimulant medications5 and pharmacotherapy for dependence on alcohol6 and tobacco.7 Genetic tests can also identify patients most likely to suffer adverse side effects of treatment, such as tardive dyskinesia8 and weight gain.9

As with any proposed treatment, evaluation of the cost-effectiveness of the pharmacogenetic approach should precede its translation to clinical practice.10, 11 In this paper, we use data from several clinical trials of tobacco dependence treatment to inform an evaluation of the cost-effectiveness of a pharmacogenetic treatment strategy. Tobacco dependence treatment is a subject ripe for economic analysis, as smoking is the greatest single preventable cause of illness and death in the United States, adding an estimated $85 billion to the U.S. health care costs in 2001.12, 13, 14 The topic is also of global significance, as the World Health Organisation estimates that there are currently 5 million tobacco-related deaths per year worldwide; the figure is expected to rise to 10 million by the early 2020s, with 70% in the developing world. Roughly half of long-term tobacco users die of tobacco-related diseases.15

Currently, there are three FDA-approved pharmacotherapies for the treatment of tobacco dependence: nicotine replacement therapies (NRTs), bupropion and varenicline. Analysis of large numbers of pharmacotherapy trials indicates that NRTs and bupropion increase the odds of successful quitting by a factor of 1.5–2.0 compared to placebo.16, 17, 18 The newest FDA-approved medication, varenicline, is superior to bupropion, increasing the odds of quitting by three to fourfold over placebo.18, 19, 20 Yet despite their proven efficacy relative to placebo, only one in four smokers are able to maintain long-term abstinence with any of these medications.21 Thus, the search for improved, cost-effective approaches to tobacco dependence treatment continues.

Inter-individual variation in successful quitting arises, in part, from inherited variation in genes for drug metabolism and drug targets.1, 22 For example, smokers who carry the minor allele (Asp40) of the mu opioid receptor (OPRM1) Asn40Asp polymorphism have almost a twofold greater odds of abstinence with 8 weeks of NRT than those carrying the common Asn40 variant.23 Similarly, smokers homozygous for the –141C Ins/Del Ins C allele in the dopamine receptor DRD2 gene have a 1.6-fold greater odds of abstinence with bupropion than those carrying at least one Del C allele, whereas smokers who have at least one copy of Del C have a 1.3-fold greater odds of abstinence on transdermal nicotine (TN) than Ins C homozygotes.7 Other reports have suggested that genetic variation in nicotine-metabolizing enzymes (for example, CYP2A6 and CYP2B6) might also predict response to NRT.24, 25, 26 Although pharmacogenetic data on varenicline are not yet available, emerging data from bupropion and NRT trials raise hopes that genetic tests may eventually help select more effective tobacco dependence treatments for individual smokers.

Despite their future promise, the cost-effectiveness of pharmacogenetic ‘test-and-treat’ approaches is largely unstudied. Indeed, a search of the Pharmacogenetics and Pharmacogenomics Knowledge Base web site ( identified only one article describing a cost-effectiveness analysis of a pharmacogenetically tailored treatment plan in any disease.27 Nevertheless, it is understood that the cost-effectiveness of a genetically tailored drug treatment plan will depend on a number of factors: the distribution of the relevant genetic variants, the costs of genetic testing and subsequent treatment and the effectiveness of the tailored therapy compared to a ‘one-size-fits-all’ approach.28, 29 In this article, we describe a model of tobacco dependence treatment that allows us to simulate cost and effectiveness outcomes to assess the value of pharmacogenetic testing using data on pharmacogenetic effects observed in tobacco dependence treatment trials.7 Estimates of pharmacogenetic effects in these trials are within the range of those observed for a variety of psychiatric medications.2 Therefore, this analysis may be considered as a case study with applicability to the broader field of pharmacogenetics in psychiatry.


We simulated discounted survival and smoking-cessation costs for 40 000 hypothetical smokers treated with one of five possible regimens: no treatment, TN, bupropion, varenicline and a tailored approach using a genetic test to determine the choice between bupropion and TN. We repeated the analysis under a range of assumptions about quit rates, mortality and so on, as listed in Table 1. See the Methods section for details.

Table 1 Parameters used in the simulation model

With 40 000 replications per simulation, Monte Carlo standard errors of mean outcome under each set of parameters were less than 0.5% of the mean for both survival and costs. Nevertheless, comparisons of outcomes between regimens were, in some cases, not statistically significant relative to their simulation standard errors. When this occurred we considered the treatments to be equivalent.

In all but two conditions tested in the sensitivity analysis, the ordering of treatment plans on mean effectiveness is no treatment < TN < Bupropion < Tailored < Varenicline, which is consistent with the ranking of the treatment plans by overall quit rate. In the two exceptions, the least favorable case and the case of zero treatment-by-SNP (single-nucleotide polymorphism) interaction, bupropion is slightly, but not significantly, more effective than tailored therapy. ‘No treatment’ is associated with zero cessation costs and, therefore, is in all cases the least expensive regimen. Although a cycle of TN is less expensive than a cycle of bupropion, which is in turn less expensive than a cycle of either drug plus a genetic test, in all cases the average cost of TN exceeds that of bupropion, and in some cases it exceeds the cost of tailored therapy as well. This finding reflects the greater efficacy of bupropion and tailored therapy compared to TN, leading to potentially fewer repeat treatments. In every scenario, the long-term treatment cost of varenicline is the highest of all.

Table 2 shows the estimated residual mean life-years (LY) and smoking-cessation costs for the least favorable, base-case and most favorable treatment strategies. In the base-case analysis, bupropion is significantly more effective and less expensive than TN, which consequently is dominated. Tailored therapy is slightly, but not significantly, more effective than bupropion, but significantly more expensive, and therefore also dominated. Varenicline is significantly more expensive than bupropion but also significantly more effective. The incremental cost-effectiveness ratio (ICER) of bupropion compared to no treatment is estimated at $1557/LY, and the ICER of varenicline compared to bupropion at $2985/LY.

Table 2 Cost-effectiveness of the competing treatment strategies under the base-case assumptions and assumptions that are least and most favorable to tailored treatment

Under the conditions least favorable to tailored therapy, TN is equivalent in both cost and effectiveness to bupropion, which has an ICER of $6614/LY compared to no treatment. Bupropion/TN dominates tailored therapy, which is significantly more expensive but no more effective. Varenicline is both significantly more effective and more expensive than bupropion/TN, with an ICER of $6902/LY. So in this scenario, varenicline is the cost-effective treatment plan.

Under the assumptions most favorable to genetic tailoring, tailored therapy is significantly more effective and less expensive than TN, which is thus dominated. Bupropion is both less expensive and less effective than tailored therapy, but is dominated by a combination of no treatment and tailored therapy. The ICER of tailored therapy compared to no treatment is $386/LY. Varenicline is significantly more effective and expensive than tailored therapy, with an ICER of $1053/LY.

Table 3 presents the single-factor sensitivity analyses, displaying for each regimen the ordering of the treatments with regard to effectiveness, omitting dominated therapies and presenting ICERs for the remaining pairwise comparisons; the most effective admissible treatment appears in boldface. In every scenario, no therapy, bupropion and varenicline are admissible; TN is never admissible, and tailored therapy is admissible only in the case of a large treatment-by-genotype interaction. ICERs for varenicline compared to the next most effective admissible therapy range from $1091/LY to $5381/LY.

Table 3 One-way sensitivity analyses of cost-effectiveness of smoking-cessation treatments


Proof of cost-effectiveness is necessary for the translation of any new therapy into routine clinical practice. Using data from two randomized pharmacogenetic trials of tobacco dependence treatment, we have estimated the potential effect on treatment costs and long-term survival of using genetic testing to choose between TN and bupropion for individual smokers. Our approach differs from the previous cost-effectiveness analyses of tobacco dependence treatment30, 31, 32, 33, 34, 35, 36 in the following ways: (i) we consider pharmacogenetic tailoring of treatment; (ii) we allow treatment costs to recur during the smoker's lifetime, reflecting the relatively modest long-term effectiveness of these drugs in any single attempt21 and (iii) we include the newly approved drug varenicline in an untailored approach (pharmacogenetic data are not yet available for this drug).

Under all parameter combinations that we considered, non-tailored varenicline was the most effective (and expensive) admissible treatment. Bupropion was dominated only in the best-case scenario. Typically, the additional survival benefit of genetic testing was negligible whereas its additional cost, though modest, was detectable; only under favorable assumptions (in particular a large treatment-by-genotype interaction) was the pharmacogenetic approach potentially cost-effective. Yet our analyses do reveal that when conditions are right, a pharmacogenetic regimen can actually both increase survival and reduce costs compared to ‘one-size-fits-all’ drug treatment, simultaneously improving the cessation rate and reducing the number of repeated courses of therapy needed to effect a cure. Under the most favorable assumptions a tailored approach selecting between bupropion and TN is nearly as effective as varenicline, at a savings of 15% in lifetime smoking-cessation costs.

Our data suggest that the benefits of ‘test-and-treat’ may not justify its costs unless a constellation of conditions holds. First, the favorable genotype should be neither too rare nor too common, else the marginal success rate of the pharmacogenetic approach will barely exceed that of the treatment that is preferable for the more common genotype. Second, and for the same reason, the treatment-by-genotype interaction must be substantial, evidently greater than was observed in the clinical trials that informed our analysis. Third, the short-term outcome (such as 6-month quit rate) should be a sufficiently good surrogate for long-term outcome (that is, survival) so that discounting will not abrogate its benefit.

Although others have investigated the cost-effectiveness of pharmacologic smoking-cessation interventions,30, 31, 32, 33, 34, 35, 36, 37 we are the first to include a pharmacogenetic tailoring arm in such an evaluation. These previous studies have suggested that pharmacotherapy is cost-effective compared to no treatment or counseling alone, with ICERs in the range of $800–$1200 per quitter and $900–$11 200 per LY, depending on the regimen and the age and sex of the smoker. Our estimated ICER values lie in the middle and lower end of this range, even for the worst-case scenario. This is at least, in part, due to continuing declines in the prices of smoking-cessation pharmacotherapies.

Our analysis is subject to a number of limitations. First, although our model for smoking histories is intended to match the best available data,12, 16, 17, 18, 38, 39, 40 few studies have examined the long-term behavior of smokers attempting to quit,38, 39, 41 the long-term effectiveness of pharmacologic smoking-cessation aids42, 43, 44 and the effectiveness of these treatments when applied repeatedly after unsuccessful quit attempts.45, 46 Moreover, our estimated quit rates are taken from published clinical trials, where the efficacy of treatments is typically better than in actual clinical practice. Thus, our estimates of treatment effects may be to an extent inflated. Unless the inflation is very substantial, however, the ICERs are likely to continue to be well below the conventionally acceptable upper limit.

Second, we were unable to incorporate pharmacogenetic data on varenicline, the newest approved pharmacotherapy for smoking cessation, because no such data are yet available. Nevertheless, the odds ratios used in our example are within the ranges of those reported for pharmacogenetic studies of antidepressants, antipsychotics, addiction treatment and tobacco dependence treatment.1, 2, 6, 47 Thus, our data give some indication of the likely cost-effectiveness of pharmacogenetic approaches in a range of psychiatric diseases. A related concern is the use of data from smokers of European ancestry only. Multi-ethnic pharmacogenetic trials are rare due to concerns about bias from ethnic admixture, a problem that can be addressed only with larger studies on more diverse populations. Moreover, it is likely that in the future a panel of genetic markers, or even a whole-genome scan, will form the basis of treatment tailoring, potentially leading to a stronger benefit of tailored therapy.29, 48

Third, the effects of discounting across time and the modest survival benefits of increasing the 6-month quit rate lead to clustering of the mean cost and LY outcomes, and consequently our comparisons are sensitive to model parameters and assumed costs when varied in combination. In some cases, the differences between treatments were too small to distinguish statistically even with the large number of subjects we simulated, leading to ambiguity in the determination of which treatments are dominated. Nevertheless, under all but one tested assumption both bupropion and varenicline are cost-effective. Moreover, because the cost of genetic testing is modest compared to that of repeated drug treatment, there are likely to be many scenarios in which genetically tailored therapy will be optimal.

Fourth, we did not attempt to quantify other treatment costs such as expenses for travel to counseling sessions, lost wages and so on. Nor did we consider the potential savings to individuals who successfully quit and thereby avoid the expenses of tobacco addiction, such as the costs of the cigarettes themselves; work time lost to smoking and respiratory illness; expenses for treating smoking-related medical conditions; and medical and economic costs arising from exposure of others to second-hand smoke. Perhaps, the most important omission is the long-term effect of smoking cessation on the consumption of health care. On average, smokers consume more health resources than non-smokers at all stages of life, even well before the onset of tobacco-related chronic disease. But because non-smokers live longer, their lifetime health care costs may actually be higher.49 Modeling these differences is analytically complex and ideologically controversial. Finally, we excluded health-related quality of life from the quantification of effectiveness. Consideration of such effects might have increased the differences in clinical outcomes across strategies.

Despite its limitations, this study is, to our knowledge, the first to analyze the cost-effectiveness of pharmacogenetic strategies for tobacco dependence treatment. Some have cautioned that the value of genetically tailored therapies for tobacco addiction is over-promised.50 Our results support this view, but only in part. They confirm the findings of others that conventional drug treatment plans are cost-effective. They also suggest that one-size-fits-all drug treatment plans compare favorably to genetically tailored plans. But our data also reveal that the differences are small, and that a tailored treatment can provide worthwhile benefits if there is a truly favorable genotype (or panel of genetic variants) that is neither too common nor too rare, and sufficiently effective treatments exist for sub-populations defined by the genotype. Although independent validation of pharmacogenetic effects across different populations will be needed before the tailored approach can pass into routine clinical practice, our results support continued research aimed at identifying the genotypes that predict therapeutic response. Thus, our data should encourage researchers to continue to collect data on potential pharmacogenetic markers and on the costs of treatments to facilitate the translation of this approach to the clinical setting.


Smoking-cessation treatment plans

We used a simulation model to evaluate the costs and effectiveness of five alternative strategies for managing cigarette smokers who are attempting to quit: (i) a ‘no treatment’ plan, (ii) treatment of all individuals with TN, (iii) treatment of all individuals with bupropion, (iv) a genetically tailored plan that chooses between TN and bupropion based on the result of a genetic test that predicts therapy outcome7 and (v) treatment of all individuals with varenicline. In the ‘no treatment’ plan, individuals receive neither counseling nor drug therapy. In the TN regimen, individuals receive counseling and 8 weeks of TN therapy. In the bupropion regimen, individuals receive counseling and 10 weeks of bupropion (2 weeks before and 8 weeks after the target quit date). In the varenicline regimen, individuals receive counseling and 12 weeks of varenicline. In tailored therapy, one first tests for the –141C Ins/Del polymorphism in the DRD2 gene. Homozygotes for the Ins C allele (CC homozygotes) receive counseling plus 10 weeks of bupropion, as above; those who have at least one copy of Del C (NN homozygotes and CN heterozygotes) receive counseling plus 8 weeks of TN, as above. This plan is based on the findings of a recent pharmacogenetic analysis of data from two randomized trials, an open-label NRT study that included a TN arm (n=180 for the TN arm)51 and a placebo-controlled trial of bupropion (n=414).7

Simulation method

We estimate mean cost and survival using a Monte Carlo simulation over 40 000 hypothetical smokers. The simulation involves repeating the following steps for each smoker:

  1. 1)

    Generate the smoker's age and genotype. We generate age from a uniform distribution on the interval 20–60 years, and independently generate genotype (CC vs CN or NN) assuming a fixed prevalence for the CC genotype. (We combine the CN and NN genotypes because NN homozygotes are rare, representing only 1.7% of trial participants7).

  2. 2)

    Generate the individual's smoking history: we assume that all individuals are current smokers, and that each year, each smoker has a 50% probability of making a quit attempt,38 with a success rate that depends on the genotype and the treatment plan. A fraction of those who quit do so permanently, whereas the rest are subject to relapse with a fixed annual relapse hazard. The model reflects what is known about the experiences of smokers attempting to quit: (i) even smokers who are motivated to quit do not make an attempt every year,38, 39 and (ii) a fraction of smokers who quit initially succeed permanently,38, 39, 40 whereas (iii) others continue a lengthy cycle of temporarily successful quit attempts that may or may not end in permanent success.38, 39, 40

  3. 3)

    Compute the individual's mortality hazard pattern: using the sampled smoking history, we compute the individual's annual hazard of death from any cause in each future potential year of life. We use as baseline hazard the age-specific mortality hazard rate in the population. An individual who is smoking or has recently quit has an elevated hazard, as described below.

  4. 4)

    Compute the individual's expected residual LY from his future mortality hazard pattern. We use the sequence of future mortality hazards at the individual's current age to generate a distribution of ages at death for that individual; the distribution varies across individuals because of variability in their smoking histories. The individual's expected residual survival is the expectation of his (discounted) survival beyond his current age.

  5. 5)

    Compute the individual's expected total cessation costs: based on his future pattern of smoking and quitting, together with the assigned treatment plan, we calculate lifetime discounted smoking-cessation costs. The one-time cost of genetic testing, when it is a part of the treatment plan, accrues in the year of the first quit attempt. Costs of counseling and drug treatment accrue in each year in which the individual attempts to quit. We compute lifetime discounted treatment costs for each possible year of death and average against the individual's distribution of potential death times (from step 4 above). We follow individuals potentially to age 100, the greatest age for which the CDC publishes mortality hazards.

All computations were done in S-Plus Version 7.0 (Insightful Corporation, Seattle, WA, USA) on a Sun SPARC computer. Code is available from the first author.


We consider two outcomes: residual LY and lifetime smoking-cessation costs. Each simulated individual yields an expected discounted LY and lifetime cessation cost, given age, genotype and smoking history. Averaging across the 40 000 simulated individuals gives an overall mean LY and cost for the treatment plan under consideration.


For each set of assumptions we ranked the five plans from least to most effective, eliminating inadmissible plans (that is, those that are dominated by either an adjacent plan or a combination of adjacent plans). We then computed the ICER, or ratio of the difference in mean cost to difference in mean LY, for each adjacent pair of remaining strategies. Ranked in this way, the most effective treatment whose ICER relative to its next less effective neighbor lies below some pre-specified limit (conventionally $50 000–$100 000/LY) is considered the cost-effective plan.

Effectiveness of smoking-cessation treatments

In randomized tobacco dependence treatment trials, effective drug therapies such as TN and bupropion have yielded 6-month quit rates of 20–40%, although studies show considerable variation.16, 17, 21 We derive our data from the pharmacogenetic evidence for the role of the DRD2 –141C Ins/Del polymorphism as a modifier of bupropion and NRT treatment.7 In that study, CC homozygotes had a quit rate of 27% on bupropion and 19% on TN; those who bear at least one copy of Del C had quit rates of 17% on bupropion and 23% on TN. In sensitivity analyses, we assumed quit rates that reflect a stronger predictive value of the polymorphism (CC: 32% on bupropion and 17% on TN; CN or NN: 12% on bupropion and 25% on TN) or a weaker value (all individuals 22% on bupropion and 21% on TN). In other words, the base-case analysis uses the treatment-by-genotype interaction observed in the clinical trials, whereas the sensitivity analysis varies these rates to yield either double the interaction or no interaction. All comparisons assume that ‘no treatment’ gives an annual quit rate of 5%,12, 52, 53 and that varenicline gives an annual quit rate of 35%.18, 19, 20

Our model assumes that individuals who quit after treatment have a 60% chance that the quit will be permanent; those who do not quit permanently have an annual relapse hazard of 50%.12 In sensitivity analyses we varied the rate of permanent success after an initially successful quit attempt from 45% to 75%.

Prevalence of the favorable genotype

Clinical trial data7 suggest that roughly 84% of European-ancestry Americans are CC homozygotes. Thus, in our simulation we used 84% as the prevalence of this genotype, which we designate the ‘favorable’ genotype because it offers the best treatment success rate. In sensitivity analyses, we varied the prevalence of the favorable genotype between 94 and 74%.

Mortality hazard

We computed residual life expectancy from the recent U.S. census data on age-specific mortality hazards in the U.S. population.54 We used the data from U.S. whites only, in keeping with the existing pharmacogenetic evidence.7

Effect of smoking on mortality

Follow-up studies comparing smokers to non-smokers suggest that smoking roughly doubles the hazard of death from any cause.55, 56, 57, 58, 59 We assumed, moreover, that in smokers who quit, the hazard ratio for death declines from its maximal value of 2 to the non-smoker's level (of 1) linearly over a 20-year period,37 and that when a temporary quitter takes up smoking again, his relative hazard returns immediately to 2. In sensitivity analyses, we varied the maximal hazard ratio between 1.5 and 2.5.

Cost of smoking-cessation treatment

We considered costs in three categories: genetic testing, smoking-cessation counseling and pharmacotherapy. Genetic testing is a one-time cost in the year of the first quit attempt. Such tests typically involve a blood draw or buccal swab followed by analysis of extracted DNA to identify genetic markers predictive of response to treatment. Such a test might include a panel of SNPs or other polymorphisms in genes that code for drug targets or drug metabolism enzymes. Because the cost of individual SNP assays is modest, the main expense of such procedures is that of collecting and preparing the specimens. The current price to consumers of a genetic test that purports to identify best treatment for nicotine addiction, offered by the g-nostics group (, Oxford, UK), is £150. We therefore used $307.71, the current U.S. equivalent, as the commercial cost of a genetic test.

A typical pharmacologic smoking-cessation treatment plan involves a small number of counseling sessions combined with several weeks of drug treatment. At our institution, in 8 weeks of treatment participants receive 2.75 h of contact from counselors who are paid $30 per hour, together with minimal intake time from clerical staff. The total is no larger than $120, which we used as the cost for the counseling component of the regimen. This is comparable to the reimbursement for smoking-cessation therapy for Medicare patients approved by the Centers for Medicare and Medicaid Services, which amounts to $197.92 for eight sessions per year (four sessions per attempt, up to two attempts annually).60 A standard 8-week course of tapered TN therapy23 costs $123.96 by the lowest available price listed on the online pharmacy web site. The current lowest available price for 10 weeks of bupropion (150 mg bid) is $127.50, and the current lowest available price for 12 weeks of varenicline (1 mg bid) is $314.95.

A health insurance plan that covers tailored smoking cessation might involve a one-time benefit for genetic testing, together with an annual benefit for a single course of counseling and drug therapy. Thus in our hypothetical coverage plan, participants may receive one course of treatment per year. As indicated above, we assumed that individuals will avail themselves of the cessation benefit in roughly half the years in which they are smoking. We did not consider indirect costs that insurance would not typically cover, such as those due to time off from work, travel and other out-of-pocket expenses. Moreover, we did not consider cost offsets, such as the avoided costs of treating smoking-related conditions.


In the base case we used a discount rate of 3% and varied this rate in sensitivity analyses between 0% and 5%.61 We applied discounting to both costs and LY.

Sensitivity analysis

We conducted a series of one-way sensitivity analyses by varying each of five model parameters while holding the others fixed at their base-case values (Table 1). To identify a range of values of cost-effectiveness, we also ran analyses setting all parameters at values least and most favorable to pharmacogenetic treatment. The least favorable strategy fixed the relative mortality hazard at 1.5, the discount rate at 5%, the treatment-by-genotype interaction at 0, the probability of a permanent quit at 45% and the prevalence of the favorable genotype at 94%. The most favorable strategy fixed the relative mortality hazard at 2.5, the discount rate at 0%, the treatment-by-genotype interaction at double the base-case value, the probability of a permanent quit at 75% and the prevalence of the favorable genotype at 74%.


  1. 1

    Munafò MR, Shields AE, Berrettini WH, Patterson F, Lerman C . Pharmacogenetics and nicotine addiction treatment. Pharmacogenomics 2005; 6: 211–223.

  2. 2

    Kirchheiner J, Nickchen K, Bauer M, Wong ML, Licinio J, Roots I et al. Pharmacogenetics of antidepressants and antipsychotics: the contribution of allelic variations to the phenotype of drug response. Mol Psychiatry 2004; 9: 442–473.

  3. 3

    Schafer M, Rujescu D, Giegling I, Guntermann A, Erfurth A, Bondy B et al. Association of short-term response to haloperidol treatment with a polymorphism in the dopamine D(2) receptor gene. Am J Psychiatry 2001; 158: 802–804.

  4. 4

    Peters EJ, Slager SL, McGrath PJ, Knowles JA, Hamilton SP . Investigation of serotonin-related genes in antidepressant response. Mol Psychiatry 2004; 9: 879–889.

  5. 5

    Polanczyk G, Zeni C, Genro JP, Guimaraes AP, Roman T, Hutz MH et al. Association of the adrenergic alpha2A receptor gene with methylphenidate improvement of inattentive symptoms in children and adolescents with attention-deficit/hyperactivity disorder. Arch Gen Psychiatry 2007; 64: 218–224.

  6. 6

    Oslin DW, Berrettini W, Kranzler HR, Pettinati H, Gelernter J, Volpicelli JR et al. A functional polymorphism of the mu-opioid receptor gene is associated with naltrexone response in alcohol-dependent patients. Neuropsychopharmacology 2003; 28: 1546–1552.

  7. 7

    Lerman C, Jepson C, Wileyto EP, Epstein LH, Rukstalis M, Patterson F et al. Role of functional genetic variation in the dopamine D2 receptor (DRD2) in response to bupropion and nicotine replacement therapy for tobacco dependence: results of two randomized clinical trials. Neuropsychopharmacology 2006; 31: 231–242.

  8. 8

    Lerer B, Segman RH, Fangerau H, Daly AK, Basile VS, Cavallaro R et al. Pharmacogenetics of tardive dyskinesia: combined analysis of 780 patients supports association with dopamine D3 receptor gene Ser9Gly polymorphism. Neuropsychopharmacology 2002; 27: 105–119.

  9. 9

    Reynolds GP, Zhang Z, Zhang X . Polymorphism of the promoter region of the serotonin 5-HT(2C) receptor gene and clozapine-induced weight gain. Am J Psychiatry 2003; 160: 677–679.

  10. 10

    Corrigan OP . Pharmacogenetics, ethical issues: review of the Nuffield Council on Bioethics Report. J Med Ethics 2005; 31: 144–148.

  11. 11

    Shields A, Lerman C, Sullivan P . Translating emerging research on the genetics of smoking into clinical practice: ethical and social considerations. Nicotine Tob Res 2004; 6: 675–688.

  12. 12

    US Department of Health and Human Services. The Health Benefits of Smoking Cessation, USDHHS, PHS, CDC, Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health; DHHS Publication No. (CDC) 90-8416 1990.

  13. 13

    American Legacy Foundation. Saving Lives, Saving Money: Why States Should Invest in a Tobacco-Free Future. American Legacy Foundation: Washington, DC, 2002.

  14. 14

    US Department of Health and Human Services. Healthy People 2010, 2nd edn. USDHHS: Washington, DC, 2000.

  15. 15

    Esson KM, Leeder SR . The Millennium Development Goals and Tobacco Control: An Opportunity for Global Partnership. World Health Organization: Geneva, Switzerland, 2004.

  16. 16

    Hughes J, Stead L, Lancaster T . Antidepressants for smoking cessation. Cochrane Database Syst Rev 2007; 1: CD000031.

  17. 17

    Etter JF, Stapleton JA . Nicotine replacement therapy for long-term smoking cessation: a meta-analysis. Tob Control 2006; 15: 280–285.

  18. 18

    Wu P, Wilson K, Dimoulas P, Mills EJ . Effectiveness of smoking cessation therapies: a systematic review and meta-analysis. BMC Public Health 2006; 6: 300.

  19. 19

    Nides M, Oncken C, Gonzales D, Rennard S, Watsky EJ, Anziano R et al. Smoking cessation with varenicline, a selective alpha4beta2 nicotinic receptor partial agonist: results from a 7-week, randomized, placebo- and bupropion-controlled trial with 1-year follow-up. Arch Intern Med 2006; 166: 1561–1568.

  20. 20

    Gonzales D, Rennard SI, Nides M, Oncken C, Azoulay S, Billing CB et al. Varenicline, an alpha4beta2 nicotinic acetylcholine receptor partial agonist, vs sustained-release bupropion and placebo for smoking cessation: a randomized controlled trial. JAMA 2006; 296: 47–55.

  21. 21

    Schnoll RA, Lerman C . Current and emerging pharmacotherapies for treating tobacco dependence. Expert Opin Emerg Drugs 2006; 11: 429–444.

  22. 22

    Rutter JL . Symbiotic relationship of pharmacogenetics and drugs of abuse. AAPS J 2006; 8: E174–E184.

  23. 23

    Lerman C, Wileyto EP, Patterson F, Rukstalis M, Audrain-McGovern J, Restine S et al. The functional mu opioid receptor (OPRM1) Asn40Asp variant predicts short-term response to nicotine replacement therapy in a clinical trial. Pharmacogenomics J 2004; 4: 184–192.

  24. 24

    Lee AM, Jepson C, Hoffmann E, Epstein L, Hawk LW, Lerman C et al. CYP2B6 genotype alters abstinence rates in a bupropion smoking cessation trial. Biol Psychiatry 2007; 62: 635–641.

  25. 25

    Lerman C, Tyndale R, Patterson F, Wileyto EP, Shields PG, Pinto A et al. Nicotine metabolite ratio predicts efficacy of transdermal nicotine for smoking cessation. Clin Pharmacol Ther 2006; 79: 600–608.

  26. 26

    Malaiyandi V, Lerman C, Benowitz NL, Jepson C, Patterson F, Tyndale RF . Impact of CYP2A6 genotype on pretreatment smoking behaviour and nicotine levels from and usage of nicotine replacement therapy. Mol Psychiatry 2006; 11: 400–409.

  27. 27

    Hughes DA, Vilar FJ, Ward CC, Alfirevic A, Park BK, Pirmohamed M . Cost-effectiveness analysis of HLA B*5701 genotyping in preventing abacavir hypersensitivity. Pharmacogenetics 2004; 14: 335–342.

  28. 28

    Flowers CR, Veenstra D . The role of cost-effectiveness analysis in the era of pharmacogenomics. Pharmacoeconomics 2004; 22: 481–493.

  29. 29

    Roden DM, Altman RB, Benowitz NL, Flockhart DA, Giacomini KM, Johnson JA et al. Pharmacogenomics: challenges and opportunities. Ann Intern Med 2006; 145: 749–757.

  30. 30

    Fiscella K, Franks P . Cost-effectiveness of the transdermal nicotine patch as an adjunct to physicians’ smoking cessation counseling. JAMA 1996; 275: 1247–1251.

  31. 31

    Wasley MA, McNagny SE, Phillips VL, Ahluwalia JS . The cost-effectiveness of the nicotine transdermal patch for smoking cessation. Prev Med 1997; 26: 264–270.

  32. 32

    Croghan IT, Offord KP, Evans RW, Schmidt S, Gomez-Dahl LC, Schroeder DR et al. Cost-effectiveness of treating nicotine dependence: the Mayo Clinic experience. Mayo Clin Proc 1997; 72: 917–924.

  33. 33

    Curry SJ, Grothaus LC, McAfee T, Pabiniak C . Use and cost effectiveness of smoking-cessation services under four insurance plans in a health maintenance organization. N Engl J Med 1998; 339: 673–679.

  34. 34

    Stapleton JA, Lowin A, Russell MA . Prescription of transdermal nicotine patches for smoking cessation in general practice: evaluation of cost-effectiveness. Lancet 1999; 354: 210–215.

  35. 35

    Woolacott NF, Jones L, Forbes CA, Mather LC, Sowden AJ, Song FJ et al. The clinical effectiveness and cost-effectiveness of bupropion and nicotine replacement therapy for smoking cessation: a systematic review and economic evaluation. Health Technol Assess 2002; 6: 1–245.

  36. 36

    Song F, Raftery J, Aveyard P, Hyde C, Barton P, Woolacott N . Cost-effectiveness of pharmacological interventions for smoking cessation: a literature review and a decision analytic analysis. Med Decis Making 2002; 22S5: S26–S37.

  37. 37

    Cornuz J, Pinget C, Gilbert A, Paccaud F . Cost-effectiveness analysis of the first-line therapies for nicotine dependence. Eur J Clin Pharmacol 2003; 59: 201–206.

  38. 38

    Centers for Disease Control and Prevention. Cigarette smoking among adults—United States 2000. MMWR 2002; 51: 642–645.

  39. 39

    Hymowitz N, Cummings KM, Hyland A, Lynn WR, Pechacek TF, Hartwell TDK . Predictors of smoking cessation in a cohort of adult smokers followed for five years. Tob Control 1997; 6: S57–S62.

  40. 40

    West R, Gilsenan A, Coste F, Zhou X, Brouard R, Nonnemaker J et al. The ATTEMPT cohort: a multi-national longitudinal study of predictors, patterns and consequences of smoking cessation; introduction and evaluation of internet recruitment and data collection methods [erratum appears in Addiction 2006 Oct;101:1528]. Addiction 2006; 101: 1352–1361.

  41. 41

    Hyland A, Li Q, Bauer JE, Giovino GA, Steger C . Predictors of cessation in a cohort of current and former smokers followed over 13 years. Nicotine Tob Res 2004; 6S3: S363–S369.

  42. 42

    Killen JD, Fortmann SP, Murphy Jr GM, Hayward C, Arredondo C, Cromp D et al. Extended treatment with bupropion SR for cigarette smoking cessation. J Consult Clin Psychol 2006; 74: 286–294.

  43. 43

    Jamerson BD, Nides M, Jorensby DE, Donahue R, Garrett P, Johnston JA et al. Late-term smoking cessation despite initial failure: an evaluation of bupropion sustained release, nicotine patch, combination therapy, and placebo. Clin Ther 2001; 23: 744–752.

  44. 44

    Hall SM, Humfleet GL, Reus VI, Muñoz RF, Cullen J . Extended nortriptyline and psychological treatment for cigarette smoking. Am J Psychiatry 2004; 161: 2100–2107.

  45. 45

    Han ES, Foulds J, Steinberg B, Gandhi KK, West B, Richardson DL et al. Characteristics and smoking cessation outcomes of patients returning for repeat tobacco dependence treatment. Int J Clin Pract 2006; 60: 1068–1074.

  46. 46

    Partin MR, An LC, Nelson DB, Nugent S, Snyder A, Fu SS et al. Randomized trial of an intervention to facilitate recycling for relapsed smokers. Am J Prev Med 2006; 31: 293–299.

  47. 47

    Cinciripini P, Wetter D, Tomlinson G, Tsoh J, De Moor C, Cinciripini L et al. The effects of the DRD2 polymorphism on smoking cessation and negative affect: evidence for a pharmacogenetic effect on mood. Nicotine Tob Res 2004; 6: 229–239.

  48. 48

    Liu QR, Drgon T, Walther D, Johnson C, Poleskaya O, Hess J et al. Pooled association genome scanning: validation and use to identify addiction vulnerability loci in two samples. Proc Natl Acad Sci USA 2005; 102: 11864–11869.

  49. 49

    Barendregt JJ, Bonneux L, van der Maas PJ . The health care costs of smoking. N Engl J Med 1997; 337: 1052–1057.

  50. 50

    Carlsten C, Burke W . Potential for genetics to promote public health: genetics research on smoking suggests caution about expectations. JAMA 2006; 296: 2480–2482.

  51. 51

    Lerman C, Kaufmann V, Rukstalis M, Patterson F, Perkins K, Audrain-McGovern J et al. Individualizing nicotine replacement therapy for the treatment of tobacco dependence: a randomized trial. Ann Intern Med 2004; 140: 426–433.

  52. 52

    Fiore MC, Novotny TE, Pierce JP, Giovino GA, Hatzaiandreu EJ, Newcomb PA et al. Methods used to quit smoking in the United States. Do cessation programs help? JAMA 1990; 263: 2760–2765.

  53. 53

    Hughes JR, Keely J, Naud S . Shape of the relapse curve and long-term abstinence among untreated smokers. Addiction 2004; 99: 29–38.

  54. 54

    Arias E . United States life tables, 2002. Nat Vital Stat Rep 2004; 53: 1–38.

  55. 55

    Doll R, Peto R, Boreham J, Sutherland I . Mortality in relation to smoking: 50 year's observations on male British doctors. BMJ 2004; 328: 1519–1527.

  56. 56

    Foody JM, Cole CR, Blackstone EH, Lauer MS . A propensity analysis of cigarette smoking and mortality with consideration of the effects of alcohol. Am J Cardiol 2001; 87: 706–711.

  57. 57

    Gotfredsen NS, Holst C, Prescott E, Vestbo J, Osler M . Smoking reduction, smoking cessation, and mortality: a 16-year follow-up of 19, 732 men and women from the Copenhagen centre for prospective population studies. Am J Epidemiol 2002; 156: 994–1001.

  58. 58

    McCarron P, Davey Smith G, Okasha M, McEwen J . Smoking in adolescence and young adulthood and mortality in later life: prospective observational study. J Epidemiol Community Health 2001; 55: 334–335.

  59. 59

    Qiao Q, Tervahauta M, Nissinen A, Tuomilehto J . Mortality from all causes and from coronary heart disease related to smoking and changes in smoking during a 35-year follow-up of middle-aged Finnish men. Eur Heart J 2000; 21: 1621–1626.

  60. 60

    US Department of Health and Human Services. Publication 100–04 Center for Medicare & Medicaid Manual System, Transmittal 562. USDHHS: Washington, DC, 2005.

  61. 61

    Gold MR, Siegel JE, Russell LB, Weinstein MC (eds) Cost-Effectiveness in Health and Medicine. Oxford University Press: New York, 1996.

Download references


We acknowledge the contributions of the following investigators to the pharmacogenetic studies on which this analysis was based: Wade H Berrettini, Susan Crystal-Mansour, Leonard Epstein, Larry Hawk, Vyga Kaufmann, David Main, Ray Niaura, Angela Pinto, Peter Shields, Susan Ware and E Paul Wileyto. We also thank Steven Siegel for helpful comments on an earlier draft.

Sources of funding: This study was supported by Grants P50 CA084718 from the National Cancer Institute and National Institutes on Drug Abuse and R01 CA063562 from the National Cancer Institute, which had no role in its design, conduct or reporting.

Author information

Correspondence to D F Heitjan.

Additional information

Duality of interest and financial disclosures

Dr Lerman has served as a consultant to Glaxo Smith-Kline, who provided bupropion and placebo for the studies described. She has also served as a consultant for Pfizer and has received funding for a project unrelated to the data presented in this paper. The other authors have no conflicts of interest to disclose.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Heitjan, D., Asch, D., Ray, R. et al. Cost-effectiveness of pharmacogenetic testing to tailor smoking-cessation treatment. Pharmacogenomics J 8, 391–399 (2008) doi:10.1038/sj.tpj.6500492

Download citation


  • addiction
  • genetic testing
  • Monte Carlo simulation
  • tobacco dependence

Further reading