Serotonin is normally released by activated platelets,

Transcription

1 Effect of Antidepressants and Their Relative Affinity for the Serotonin Transporter on the Risk of Myocardial Infarction William H. Sauer, MD; Jesse A. Berlin, ScD; Stephen E. Kimmel, MD, MS Background Antidepressants, particularly selective serotonin reuptake inhibitors (s), attenuate platelet activation by depleting serotonin storage and may decrease risk of myocardial infarction (MI). These drugs differ in their affinity for the platelet serotonin transporter and therefore may vary in their effects on MI protection. Methods and Results A case-control study of first MI in patients aged 40 through 75 years was conducted among 36 hospitals in a 5-county area during a 3-year period. Case subjects were patients hospitalized with a first MI, and control subjects were randomly selected from the same geographic area. Detailed information regarding medication use and other clinical and demographic data were obtained by telephone interview. Among the 1080 cases and 4256 controls who participated, there were 223 users of antidepressants with high serotonin transporter affinity, all of which were s (paroxetine, fluoxetine, and sertraline). After adjustment with multivariable logistic regression for age, gender, race, education, physical activity, quantity of cigarettes smoked per day, body mass index, aspirin use, family history of MI, and history of diabetes, hypertension, or hypercholesterolemia, the odds ratio for MI among current users of antidepressants with high serotonin transporter affinity compared with nonusers was 0.59 (95% CI 0.39 to 0.91; P 0.02). Increasing serotonin transporter affinity was associated with reduced odds of MI among users of all s (P for trend 0.01) but not tricyclic (P 0.77) or atypical (P 0.70) antidepressants. There was no association detected between non- antidepressant use and MI. Conclusions Increasing serotonin transporter affinity correlates with greater MI protection with but not other antidepressant exposure. (Circulation. 2003;108:32-36.) Key Words: serotonin uptake inhibitors antidepressant agents epidemiology myocardial infarction drugs Serotonin is normally released by activated platelets, causing enhanced platelet aggregation, which may contribute to the pathogenesis of acute myocardial infarction (MI). 1,2 A genetic polymorphism of the serotonin transporter (the only mechanism for serotonin uptake into platelets) causes increased serotonin uptake and is associated with an increased risk of MI. 3 Antidepressants, particularly the selective serotonin reuptake inhibitors (s) with high affinity for the serotonin transporter, attenuate platelet activation by depleting serotonin storage 4 and have been shown to decrease platelet activity in patients with coronary artery disease. 5 The antiplatelet effects of s, most of which have high affinity for the serotonin transporter, have been implicated in an increased risk of gastrointestinal bleeding 6 and lower risk of MI among users of s. 7 Recently, van Walraven and colleagues 8 described an increased risk of upper gastrointestinal bleeding with increasing serotonin transporter affinity of antidepressants. We have previously described an association between use and reduced odds of MI. However, that study was limited in its ability to examine antidepressants other than s with nonselective serotonin uptake inhibition or the potential association between serotonin transporter affinity and MI protection because of the small number of antidepressant users included. 7 Therefore, the specific aim of this study was to examine the association between both antidepressant type and the extent of the serotonin uptake inhibition of an antidepressant with the risk of MI in a general population. Methods Source Population and Identification of Subjects Data were obtained from an ongoing case-control study of MI in the Philadelphia metropolitan area that is examining the effects of prescription and over-the-counter drug exposure on the risk of MI. Case subjects were patients aged 40 through 75 years with a first MI who were hospitalized between May 1998 and May 2001 at 1 of 36 Received February 10, 2003; revision received April 1, 2003; accepted April 4, From the Cardiovascular Division (W.H.S., S.E.K.), Department of Medicine, Center for Clinical Epidemiology and Biostatistics and Department of Biostatistics and Epidemiology (J.A.B., S.E.K.), and Center for Education and Research on Therapeutics (J.A.B.), University of Pennsylvania School of Medicine, Philadelphia, Pa. Dr Berlin has served as a consultant to GlaxoSmithKline in the past on projects unrelated to depression or drugs related to depression. Dr Kimmel has received research funding support from GlaxoSmithKline to study antidepressant-related heart disease and has received research funding support from Aventis, Pfizer, Pharmacia, and Novartis, unrelated to depression or drugs related to depression. Dr Kimmel has also served as a consultant to Bayer, unrelated to depression or drugs related to depression. Correspondence to Stephen E. Kimmel, MD, MS, University of Pennsylvania School of Medicine, Center for Clinical Epidemiology and Biostatistics, 717 Blockley Hall, 423 Guardian Dr, Philadelphia, PA skimmel@cceb.med.upenn.edu 2003 American Heart Association, Inc. Circulation is available at DOI: /01.CIR CD 32

2 Sauer et al Antidepressants and Myocardial Infarction 33 acute-care hospitals in a 5-county region that are part of the Delaware Valley Case-Control Network. Acute MI was defined with the criteria from the Minnesota Heart Survey 9 and verified in 86% of the cases that had sufficient information. Exclusion of those cases without complete information or charts available did not alter the results presented. Case subjects were excluded if they had a history of a prior MI, they did not have telephones or could not communicate on the telephone (eg, they did not speak English, were deaf, or were aphasic), they did not live in 1 of the 5 counties of the network, or they used multiple antidepressants at the time of enrollment. The participation rate among eligible cases was 60%. Control subjects were selected by random digit dialing and were not matched to case subjects except for meeting the same inclusion and exclusion criteria. The participation rate among eligible controls was 54%. Data Collection Exposure and covariate data were collected by a structured telephone interview for both cases and controls, who were unaware of the study hypothesis. To maximize the validity of exposure information, cases were interviewed only if they could be reached within 4 months of their MI. Similarly, controls were also interviewed only within 4 months of initially being identified. The index date was the date of MI for cases and the date of the telephone interview for controls. Detailed information was obtained regarding antidepressant use during the week before the index date, indication for antidepressant use, and other clinical characteristics. We did not measure depression or ask about symptoms related to depression except in asking antidepressant users for the indication for their medication. Categorization of Antidepressants Antidepressants were categorized by class (, tricyclic, and atypical) and by their affinity for the serotonin transporter with the dissociation constant (K d ) used as a continuous variable. 10 Although prior studies of gastrointestinal bleeding used 3 categories of serotonin transporter affinity (low, moderate, and high), 8 we found differences among classes of antidepressants on MI risk that made these groupings inappropriate without first stratifying on the basis of antidepressant class. In addition, there were no users of monoamine oxidase inhibitors or clomipramine (the only non- in the high-affinity group) in the study population. Statistical Analysis The odds ratio (OR) was used to estimate the relative risk of MI from use of antidepressants. The primary comparisons were between users of a particular class of antidepressant and nonusers of any antidepressant and by serotonin transporter affinity as a continuous variable within each antidepressant class. For these analyses, we used the log transformation of the dissociation constant. Higher values of this constant indicate lower affinity for the serotonin transporter (Table 1). Multivariable logistic regression analysis was used to control for possible confounding. The multivariable model included variables that are known risk factors for MI and any potential confounder that changed the unadjusted OR by more than 10% after adjustment. 11 The variables used in the final model were age, gender, race, education, a validated physical activity score, 12 aspirin use during the index week, number of cigarettes smoked per day, body mass index, and family history of coronary disease, diabetes mellitus, hypertension, and hypercholesterolemia. Other potential confounding variables tested (income, year and season of index week, vitamin use, lipid-lowering therapy, ACE inhibitor use, marital status, history of coronary disease, caffeine and alcohol consumption, type of insurance, -blocker use, and nonsteroidal anti-inflammatory drug use) did not affect any of the ORs and were therefore not included. The analyses examining tricyclic, atypical, and antidepressants used the same multivariable model described above. Statistical analyses were performed with the SPSS (version 10.0) software program, and statistical significance was defined as a 2-sided probability value less than TABLE 1. Antidepressants Categorized by Affinity for the Serotonin Transporter 10 Affinity for Serotonin Transporter High (K d 1 nmol) Paroxetine Clomipramine Sertraline Fluoxetine Moderate (K d 1 10 nmol) Citalopram Imipramine Fluvoxamine Amitriptyline Venlafaxine Low (K d 10 nmol) Desipramine Nortriptyline Protriptyline Amoxapine Doxepin Trimipramine Trazodone Nefazodone Maprotiline Bupropion Mirtazapine Results Type of Antidepressant K d indicates dissociation constant. The dissociation constants listed are derived from reference 10. Characteristics of Study Participants There were 404 antidepressant users (76 exposed cases and 328 exposed controls) identified among the 1080 cases and 4256 controls who participated. Compared with nonusers of antidepressants, high-affinity users were more likely to be female and had more risk factors for coronary disease, including hypertension, family history, and hypercholesterolemia (all P 0.05; Table 2). Interactions were tested between each potential confounding variable and use of antidepressants with low, moderate, or high serotonin transporter affinity; none were significant (P 0.10). Association of Antidepressant Serotonin Transporter Affinity and MI There was a significant interaction between drug class and serotonin transporter affinity on MI risk (adjusted P for interaction 0.04). That is, the effect of serotonin transporter affinity on MI risk varied depending on the class of antidepressant drug used. Higher serotonin transporter affinity was associated with a reduced odds of MI among users of s (adjusted P for trend 0.01; Figure 1) but not among users of tricyclic agents or atypical antidepressants (adjusted P for trend 0.77 for tricyclics and 0.70 for atypicals).

3 34 Circulation July 8, 2003 TABLE 2. Clinical Characteristics of Study Participants in the Control Group Serotonin Transporter Affinity Group High (n 190) Association Between Use and MI Table 3 displays the risk of MI by type of antidepressant. use was associated with a significantly reduced risk of MI compared with other antidepressant users (data not displayed in table; adjusted OR 0.51; 95% CI 0.29 to 0.90; P 0.02) but not when compared with nonusers of any antidepressant (adjusted OR 0.72; 95% CI 0.49 to 1.05; P 0.09). However, use of high-affinity s (paroxetine, fluoxetine, and sertraline) was associated with a significant reduction in risk of MI compared with use of other antidepressants (adjusted OR 0.38; 95% CI 0.21 to 0.70; P 0.01) Moderate (n 60) Low (n 78) No AD Use (n 3928) Age (mean SD), y BMI (mean SD), y Cigarettes/day (mean SD) Activity score (mean SD) Male gender, n (%) 55 (28.9) 13 (21.7) 23 (29.5) 1660 (42.2) Diabetes mellitus, n (%) 18 (9.5) 4 (6.7) 4 (5.1) 283 (7.2) Hypercholesterolemia, n (%) 64 (33.7) 23 (38.3) 24 (30.8) 996 (25.3) Hypertension, n (%) 71 (37.4) 18 (30.0) 29 (37.2) 1154 (29.4) Smoking status, n (%) Former 69 (36.3) 27 (45.0) 28 (35.9) 1258 (32.2) Current 45 (23.7) 9 (15.0) 22 (28.2) 844 (21.6) Never 76 (40.0) 24 (40.0) 28 (35.9) 1807 (46.2) Family history of CAD, n (%) 92 (48.4) 32 (53.3) 25 (32.1) 1489 (37.9) Married, n (%) 103 (55.1) 30 (50.0) 41 (52.6) 2452 (63.4) Aspirin use, n (%) 50 (26.3) 12 (20.0) 18 (23.1) 1014 (25.8) -Blocker use, n (%) 16 (8.4) 1 (1.7) 13 (16.7) 287 (7.3) History of CAD, n (%) 8 (4.2) 1 (1.7) 1 (1.7) 81 (2.1) Race, n (%) White 160 (85.6) 56 (93.3) 68 (87.2) 3152 (80.6) Black 18 (9.6) 3 (5.0) 6 (7.7) 628 (16.1) Other 9 (4.8) 1 (1.7) 4 (5.1) 132 (3.4) College education, n (%) 119 (63.0) 39 (65.0) 46 (59.0) 2912 (58.8) Vitamin use, n (%) 91 (47.9) 27 (45.0) 44 (56.4) 2002 (50.9) AD indicates antidepressant; BMI, body mass index (kilograms per meter squared); and CAD, coronary artery disease. and no antidepressant use (adjusted OR 0.59; 95% CI 0.39 to 0.91; P 0.02). Gender was the only potential confounder included in the multivariable model that weakened the association between high-affinity use and MI. Association Between Other Antidepressant Use and MI Users of low and moderate serotonin transporter affinity antidepressants did not have significantly altered odds ratios for MI compared with no antidepressant use. Use of tricyclic and atypical antidepressants, all of which were low or moderate serotonin transporter affinity antidepressants in the present study, was not associated with an altered risk of MI (Table 3). Odds of MI according to use of antidepressants. Reference group is users of moderate serotonin transporter affinity s (citalopram and fluvoxamine). K d indicates dissociation constant. Discussion Study Results This study demonstrates a significant reduction in risk of MI associated with current use of high-affinity s compared with both nonuse of antidepressants and use of other antidepressants. In addition, the extent of serotonin inhibition among s correlated with the degree of reduced odds of MI. Other antidepressants, regardless of serotonin transporter affinity, were not associated with an altered odds ratio of MI relative to those not using antidepressants.

4 Sauer et al Antidepressants and Myocardial Infarction 35 TABLE 3. Association Between Antidepressant Type and MI Cases, n Controls, n Unadjusted OR (95% CI) Multivariable OR (95% CI)* P No AD use Reference Reference use ( ) 0.72 ( ) 0.09 High serotonin transporter affinity ( ) 0.59 ( ) 0.02 AD use (all s) AD use ( ) 1.63 ( ) 0.19 AD use ( ) 1.28 ( ) 0.37 AD indicates antidepressant. *Adjusted for age, gender, race, education, activity score, smoking status, quantity of cigarettes smoked per day, aspirin use, body mass index, family history, diabetes mellitus, hypertension, and hypercholesterolemia. P value for multivariable OR. Potential Role of Serotonin Reuptake Inhibitors in Prevention of MI Serotonin is normally released when platelets are activated by proaggregatory factors, causing platelet effects, which may contribute to the pathogenesis of MI. 1,2 Recently, Fumeron et al 3 described the association between the LL genotype of the SLC6A4 polymorphism of the serotonin transporter gene, which results in increased platelet serotonin uptake, and higher risk of MI compared with other genotypes. Antidepressants that prevent the uptake of serotonin into platelets may thereby inhibit the only mechanism for platelet storage of serotonin. 4 All of the high-affinity s represented in the present study have been shown to reduce platelet aggregation. 5,13 15 We know of no studies that have suggested altered platelet activity attributed to the other antidepressants with low or moderate serotonin transporter affinity; however, one study demonstrated no effect of nortriptyline on platelet activity. 14 If the mechanism of reducing MI is related to the effect of antidepressants on platelet serotonin, and if antidepressants in the higher-affinity groups have a greater antiplatelet effect than lower-affinity ones, then these drugs might be expected to have the greatest effect on MI risk. The potential effects of high serotonin transporter affinity antidepressants on platelet biology therefore form the basis for the a priori hypothesis that the extent of serotonin reuptake inhibition correlates with reduction in serotonin-mediated platelet aggregation and thus reduced risk of MI in users of these agents. The results of the present study, as well as those described in a previous study examining gastrointestinal bleeding in users of antidepressants with moderate and high serotonin transporter affinity, 8 would suggest that the platelet inhibition caused by antidepressants is related to their affinity for the serotonin transporter. The reduced risk of MI associated with increasing serotonin transporter affinity was seen only among users of s and not other antidepressants. This may be due to a nonlinear effect on platelet inhibition by antidepressants with increasing affinity for the serotonin transporter. The lack of effect of increasing serotonin transporter affinity in the non- antidepressants may also reflect potentially harmful effects of these drugs, 16 negating any potential MI protective effects from platelet inhibition. Because there were no users of non- antidepressants in the highest-affinity group, we are unable to clarify the potential biological basis for this observation. Regardless of the reason for differences between s and non- antidepressants, the pharmacodynamic property of serotonin inhibition is a plausible biological explanation for the reduced risk of MI associated with s. 7,16 Prior Investigations Into the MI- Protection Relationship There have been several studies examining the potential for MI protection associated with use. In the SADHART study, a randomized trial investigating the safety and efficacy of sertraline in patients with post-mi depression, there was a nonsignificant reduction in MI risk in the treatment group. 17 Although the SADHART study was not powered to detect any meaningful reduction in MI risk, the results are consistent with what we observed in the present study. Cohen et al 16 used an administrative database to study the risk of MI for users of all antidepressants. In that study, tricyclic agents were associated with an increased risk of MI, and s were associated with a nonsignificant protective effect (OR 0.8). 16 However, the study was limited by its inability to adjust for confounding comorbidities, including depression. Furthermore, the small number of exposed cases substantially impeded its ability to detect a meaningful effect of use on MI risk. In a recent study of s that included antidepressants with both high and moderate serotonin transporter affinity, there was no detectable altered MI risk observed in users of these agents compared with users of other antidepressants. 18 However, that study used a definition of exposure that would include some subjects not currently taking an and included an atypical agent as an, both of which could have biased the results toward no effect on MI risk. In addition, all subjects with risk factors for cardiac disease were excluded, thereby limiting the generalizability of the results. We 7 previously described an association between use and reduced odds of MI in a population of smokers. That study had insufficient numbers of users of other antidepressants to perform the analyses used in the present study. The results of the present study are consistent with the previous observation because 99% of the group of our prior study consisted of high-affinity s, and subsequent analyses excluding the very small number of moderate-affinity users in the prior study did not substantially alter the results of that study. Potential Limitations Nonparticipation could bias the results of the present study if nonparticipant cases were more likely to use high-affinity s

5 36 Circulation July 8, 2003 for treatment of depression than were participants. This bias is unlikely because use of other antidepressants, which would be expected to have a similar participation rate, is not associated with reduced odds of MI. In addition, the significant association between increasing serotonin transporter affinity and reduced risk of MI among users would be unaffected by differential participation of antidepressant users. Alternatively, if nonparticipant control subjects were less likely to use a high-affinity, there could be a false association with these agents and reduced MI risk. However, the prevalence of use of these agents in the control group was similar to what is described on a national level 19 and therefore cannot explain these results. Differential recall of antidepressant use could create a false association if control subjects were more likely than case subjects to accurately report high-affinity exposure. However, high-affinity use was associated with reduced odds of MI compared with users of other antidepressants, who would have the same issues of differential recall, which makes this unlikely. Similarly, the association of serotonin transporter affinity and MI risk among s users should not be subject to recall bias. Uncontrolled confounding is another potential limitation of the present study. If high-affinity users are a low-risk group because of other unmeasured characteristics, there would be a false association between high-affinity use and lower risk for MI. However, high-affinity users had more traditional risk factors for MI, except that they included a higher proportion of women than did the nonuser group, and adjustment for all known potential risk factors, including variables related to access to medical care, strengthened rather than weakened the inverse association between high-affinity use and MI, with the exception of gender. Confounding by indication (ie, if high-affinity s were used for a condition that is itself associated with lower MI risk) could not explain these results because the primary indication for high-affinity s (depression), if it had any effect, should increase the risk of MI. 20,21 This would bias the results toward a falsely harmful effect of high-affinity s. Because we did not have information on the severity of depression, we could not adjust for that potential confounder. However, depression severity is likely to be nondifferential among the high-affinity users, given equivalent efficacy in this class, 22 and could not explain the results. Finally, if users were more likely to experience sudden cardiac death, then a false association with MI protection could be seen, because cases only included nonfatal MI. However, unlike the tricyclic agents, s are not arrhythmogenic and did not change ECG parameters in randomized trials. 14,17 Conclusions Use of s with high affinity for the serotonin transporter is associated with reduced odds of MI, and the extent of serotonin inhibition among s correlates with the degree of reduction in MI risk. Although adequate treatment of depression itself could translate into a reduced risk of MI, the present results suggest that the relative reduction in platelet serotonin uptake caused by s is the reason for the risk reduction. Further investigation is needed to confirm these findings, to evaluate the effect of varying levels of antidepressant serotonin transporter affinity on platelet inhibition, and to evaluate further potential differences among antidepressants regarding cardiovascular risk. Acknowledgments This work was supported by a grant from the National Institutes of Health (R01HL57312). A supplement for research still ongoing was provided by Searle Pharmaceuticals, Inc (now Pharmacia). References 1. Li N, Wallen NH, Ladjevardi M, et al. Effects of serotonin on platelet activation in whole blood. Blood Coagul Fibrin. 1997;8: Skop BP, Brown TM. Potential vascular and bleeding complications of treatment with selective serotonin reuptake inhibitors. Psychosomatics. 1996; 37: Fumeron F, Betoulle D, Nicaud V, et al. Serotonin transporter gene polymorphism and myocardial infarction: Etude Cas-Temoins de l Infarctus du Myocarde (ECTIM). Circulation. 2002;105: Hergovich N, Aigner M, Eichler HG, et al. Paroxetine decreases platelet serotonin storage and platelet function in human beings. Clin Pharmacol Ther. 2000;68: Serebruany VL, O Connor CM, Gurbel PA. Effect of selective serotonin reuptake inhibitors on platelets in patients with coronary artery disease. Am J Cardiol. 2001;87: de Abajo FJ, Rodriguez LA, Montero D. Association between selective serotonin reuptake inhibitors and upper gastrointestinal bleeding: population based case-control study. BMJ. 1999;319: Sauer WH, Berlin JA, Kimmel SE. Selective serotonin reuptake inhibitors and myocardial infarction. Circulation. 2001;104: van Walraven C, Mamdani MM, Wells PS, et al. Inhibition of serotonin reuptake by antidepressants and upper gastrointestinal bleeding in elderly patients: retrospective cohort study. BMJ. 2001;323: Mascioli SR, Jacobs DRJ, Kottke TE. Diagnostic criteria for hospitalized acute myocardial infarction: the Minnesota experience. Int J Epidemiol. 1989;18: Tatsumi M, Groshan K, Blakely RD, et al. Pharmacological profile of antidepressants and related compounds at human monoamine transporters. Eur J Pharmacol. 1997;340: Mickey RM, Greenland S. The impact of confounder selection criteria on effect estimation. Am J Epidemiol. 1989;129: Baecke JA, Burema J, Frijters JE. A short questionnaire for the measurement of habitual physical activity in epidemiological studies. Am J Clin Nutr. 1982;36: Markovitz JH, Shuster JL, Chitwood WS, et al. Platelet activation in depression and effects of sertraline treatment: an open-label study. Am J Psychiatr. 2000;157: Roose SP, Laghrissi-Thode F, Kennedy JS, et al. Comparison of paroxetine and nortriptyline in depressed patients with ischemic heart disease. JAMA. 1998;279: Menys VC, Smith CC, Lewins P, et al. Platelet 5-hydroxytryptamine is decreased in a preliminary group of depressed patients receiving the 5-hydroxytryptamine re-uptake inhibiting drug fluoxetine. Clin Sci. 1996;91: Cohen HW, Gibson G, Alderman MH. Excess risk of myocardial infarction in patients treated with antidepressant medications: association with use of tricyclic agents. Am J Med. 2000;108: Glassman AH, O Connor CM, Califf RM. Sertraline treatment of major depression in patients with acute MI or unstable angina. JAMA. 2002;288: Meier CR, Schlienger RG, Jick H. Use of selective serotonin reuptake inhibitors and risk of developing first-time acute myocardial infarction. Br J Clin Pharmacol. 2001;52: Kaufman DW, Kelly JP, Rosenberg L, et al. Recent patterns of medication use in the ambulatory adult population of the United States: the Slone survey. JAMA. 2002;287: Ariyo A, Haan M, Tangen C, et al. Depressive symptoms and risks of coronary heart disease and mortality in elderly Americans. Circulation. 2000; 102: Pratt LA, Ford DE, Crum RM, et al. Depression, psychotropic medication, and risk of myocardial infarction: prospective data from the Baltimore ECA follow-up. Circulation. 1996;94: Kroenke K, West SL, Swindle R, et al. Similar effectiveness of paroxetine, fluoxetine, and sertraline in primary care: a randomized trial. JAMA. 2001; 286: