Abstract
Aim:
To examine the impact of CYP2C19 genotype on selective serotonin reuptake inhibitor (SSRI) prescribing patterns.
Patients & methods:
Observational cohort containing 507 unique individuals receiving an SSRI prescription with CYP2C19 genotype already in their electronic medical record. Genotype was distributed as follows: n = 360 (71%) had no loss of function alleles, 136 (26.8%) had one loss of function allele and 11 (2.2%) had two loss of function alleles.
Results & conclusion:
For poor metabolizers exposed to sertraline, citalopram or escitalopram, providers changed prescribing patterns in response to alerts in the electronic medical record by either changing the drug, changing the dose or monitoring serial EKGs longitudinally. For intermediate metabolizers exposed to sertraline, citalopram or escitalopram, no alert was needed (mean QTc = 440.338 ms [SD = 31.1273] for CYP2C19*1/*1, mean QTc = 440.371 ms [SD = 29.2706] for CYP2C19*1/*2; p = 0.995).
Keywords: : arrhythmia, citalopram, coronary artery disease, depression, escitalopram, pleiotropy, primary care, QT interval, QT prolongation, sertraline
CYP2C19 genotype is becoming routinely available in electronic medical records (EMRs). Healthcare providers are increasingly ordering this test in patients undergoing percutaneous coronary intervention (PCI), to guide choice of antiplatelet therapy [1]. Because a key principle in gene-based drug dosing has been that variation in a single pharmacogene can impact clinical outcome for many classes of drugs, decision support for CYP2C19 genotype must be deployed within a flexible framework capable of providing guidance for additional medications beyond antiplatelet agents [2]. For example, CYP2C19 metabolizes antifungals (e.g., voriconazole), proton pump inhibitors (e.g., omeprazole) and many other commonly prescribed drugs, including several types of antidepressants [3–5].
Antidepressants are among the most commonly prescribed drugs in the USA. As many as 10% of adults have a depressive disorder [6], and the prevalence of depression is even higher in patients with cardiovascular disease [7]. Uncontrolled depression is a poor prognostic indicator for patients with coronary artery disease (CAD) [8], and prescribing strategies are needed to optimize efficacy and decrease toxicity for antidepressants specifically in this population [9].
Serotonin selective reuptake inhibitors (SSRIs) are often prescribed to treat depression [10]. Three frequently prescribed SSRIs, sertraline, citalopram and escitalopram, are metabolized by CYP2C19, and CYP2C19 genotype can affect the efficacy of these drugs. The STAR*D trial quantified impact of CYP2C19 genotype on SSRI efficacy in a setting designed to approximate real-world conditions [11]. Within this treatment-based cohort, CYP2C19 genotype was associated with remission (p = 0.03) in patients adherent to therapy for the full duration of the study [12].
In addition to influencing efficacy, CYP2C19 genotype influences risk for toxicity in patients taking sertraline, citalopram and escitalopram [4]. Adverse drug reactions (ADRs) accompanying the use of SSRIs can include gastrointestinal upset (constipation, diarrhea, nausea and vomiting), sleep disorders (insomnia or somnolence), sexual dysfunction (abnormal ejaculation and reduced libido) and cardiac rhythm disturbances (QT prolongation and torsades de pointes) [13]. Gender-specific thresholds have been published for QTc (QT interval adjusted by RR interval, to correct for heart rate). While these thresholds can help clinicians identify QT prolongation (QTc >440 ms in men and QTc >460 ms in women), the strength of the correlation between QT interval and potentially lethal ventricular arrhythmias like torsades de pointes increases when QTc exceeds 500 ms [14].
Because CYP2C19 enzyme activity influences risk for QT prolongation with sertraline, citalopram and escitalopram, many EMRs now have the capability to deploy automated best practice advisories (BPAs) to guide physicians with SSRI dosing based on CYP2C19 genotype [2]. CYP2C19 poor metabolizers (patients with two loss of function alleles) have higher plasma concentrations of sertraline, citalopram and escitalopram, and gene-based dosing guidelines have been published for these drugs by the Clinical Pharmacogenetics Implementation Consortium (CPIC) [4]. The US FDA has also recommended changes to the product label for some SSRIs based on CYP2C19 genotype. While the prescription package insert for citalopram now states that 20 mg/day is the maximum recommended dose for this medication in patients with two loss of function CYP2C19 alleles, guidance is less clear for patients with one loss of function allele [15].
We therefore sought to describe the impact of CYP2C19 genotype on SSRI prescribing patterns in a large observational cohort where BPAs based on CPIC guidelines are available in the context of routine clinical care. We report examples of changes in physician behavior for patients with two loss of function alleles, and we quantify the relationship between genotype and QTc in patients with one loss of function allele.
Methods
The current study protocol was approved by the Sanford Research Center Institutional Review Board. Design was retrospective assessment of prescribing pattern in a large observational cohort. The setting was a multi-specialty group practice located in the upper Midwest. Sanford Health and Sanford Medical Center represent a fully integrated healthcare system with a longstanding comprehensive EMR. The majority of patient records within this system have been electronic for over a decade. Our EMR includes comprehensive lists of patient problems, a summary of each clinic encounter (including diagnoses and procedures), a variety of medication alerts and online access to over a decade of laboratory results. Since the majority of patients in the current study (92.3%) receive primary care through Sanford Health, chart fragmentation is minimal and this record is considered both longitudinal and comprehensive.
Study subjects
Patients 18 years or older who had been genotyped for CYP2C19 during the course of routine clinical care were eligible for study inclusion. This included patients genotyped pre-emptively in the context of primary care, and patients who had been genotyped reactively, to guide antiplatelet therapy within the context of PCI for CAD. On 1 January 2018, our EMR contained longitudinal data from ∼2000 patients genotyped for CYP2C19. A total of 507 of these patients have also had an SSRI prescribed at some point. Because gene-based dosing guidelines for CYP2C19 published by CPIC are limited to CYP2C19 *1, *2, *3, *17 [4], our analyses included only those variants.
Patient summary: demographics & covariates
Our EMR contains data from 1.8 million unique individuals. Our study cohort consisted of 507 unique individuals who had received an SSRI prescription with extant CYP2C19 genotype already available in their chart. Mean age of these study subjects was 64 years (64.03; 15.14 SD). Gender was evenly distributed, 48.5% males and 51.5% females. Based upon self-reported race, 96.3% of study subjects were White/Caucasian, 2.6% were American Indian or Alaskan Native and less than 1% were African–American or Hispanic/Latin–American (0.4 and 0.2% respectively). A very small percentage of the cohort (0.6%) declined to provide race or ethnic information. Most study subjects were married or partnered (57.8%) or widowed (13.2%). Average years of education was 14 years (SD = 2.1 years). Alcohol use was reported in 40.8% of study participants (with at least one 12-ounce beer, one 5-ounce glass of wine or 1.5 ounces of a distilled alcohol containing beverage per week set as criteria for alcohol use). History of smoking was reported in 42% of the cohort (n = 213), and 40.8% of subjects reported being active smokers (N = 207). Hypertension was common in study subjects, at 82.2% with at least one historical diagnosis in the EMR. Nearly half of study participants (44%; N = 223) met criteria for diabetes (at least one episode of a diabetes diagnosis listed within their EMR as indicated by diagnostic codes).
Event summary: electronic extraction of longitudinal data
Longitudinal data were electronically extracted from the EMR for cardiovascular events (CAD and PCI) and for diagnostic codes linked to behavioral health (depression and anxiety). Approximately 25% of the cohort had been genotyped pre-emptively in the context of routine primary care, and the remainder were genotyped reactively (at the point of prescribing) in the coronary angiography suite. The study included more than 3 years of encounter data between 1 October 2014 and 1 January 2018. We used commonly captured EMR data, including ICD-9 codes, ICD-10 codes, CPT codes and laboratory test results for discrete variable abstraction (encounter diagnosis, hospital account diagnosis, hospital admission diagnosis; claims diagnosis/codes; as well as medical history, abstracted from the problem list). Coronary artery disease was identified using ICD-9 and ICD-10 diagnostic and procedural codes in the EMR. Greater than three quarters of the study population met criteria for a diagnosis of CAD (76.1%; N = 386), and greater than 60% of study participants (62.5%; N = 317) had undergone a PCI. Furthermore, 70.2% of study participants had at least one diagnostic code for depression (311.**, 296.2, 296.3), and 67.9% has at least one diagnostic code for anxiety (300.00, 300.01 and 300.02).
Validation & manual chart review
Documents for all 507 study subjects with CYP2C19 genotype and exposure to an SSRI were reviewed by a licensed practicing physician, and 50 of these records (10%) were reviewed in duplicate by two licensed practicing physicians for quality control. Clinical text documents and EKGs were manually reviewed for subject identifiers, and charts were reviewed longitudinally for SSRI prescribing history. QT interval was abstracted as QTc from time windows where subjects were observed to have been on one of three target SSRIs: sertraline, citalopram and escitalopram. Medication use was also confirmed at the time of most recent clinic visit. The reviewer also confirmed the presence or absence of electronically extracted covariates by reviewing clinical text documents dating back to 1997 (two decades). Types of documents encompassed a variety of clinical encounters (e.g., primary care office notes, routine physical exams and cardiology consultations). In order to manually reconstruct complete SSRI exposure histories, each physician reviewer recorded drug name (generic and/or trade name), drug dose and the time period over which the drug had been prescribed. In all cases, there was sufficient information embedded in the clinical text documents to identify specific dates when each SSRI had been started and stopped.
Statistical analysis
Association between CYP2C19 genotype and prescribing pattern for each SSRI was assessed using χ2 test or Fisher's exact test as appropriate. The mean QTc in carriers of CYP2C19 loss of function alleles was compared with the mean of patients with no loss of function allele for each of three target drugs (sertraline, citalopram, escitalopram) using a two sample t-test. Next, the effects of genotype and gender on QTc were examined through two-way ANOVA with post-hoc pairwise comparison using the TukeyHSD method. A p-value of less than 0.05 was considered significant for each of these analyses.
Results
On 1 January 2018, our EMR contained longitudinal data from approximately 2000 unique individual patients genotyped for CYP2C19; of these, 507 of patients had been prescribed an SSRI at some point. Table 1 contains a description of the cohort. As anticipated, most study subjects in this treatment-based cohort – defined by prior exposure to an SSRI – had either a diagnostic code for depression or anxiety.
Table 1. . Patient summary.
| Patient information and history | Reported as | Total† (n = 507) | UM/RM/EM (n = 360) | IM (n = 136) | PM (n = 11) |
|---|---|---|---|---|---|
| Age | Mean (SD) | 64.03 (15.14) | 64.540 (14.775) | 62.440 (15.536) | 67.270 (21.119) |
| BMI | Mean (SD) | 52.79 (12.72) | 52.569 (12.785) | 53.458 (12.374) | 51.631 (16.336) |
| Gender | Males n (%) | 246 (48.5%) | 172 (69.9) | 71 (28.9) | 3 (1.2) |
| Females n (%) | 261 (51.5%) | 188 (72.0) | 65 (24.9) | 8 (3.1) | |
| Diabetes | n (%) | 223 (44.0%) | 158 (70.9) | 59 (26.5) | 6 (2.7) |
| Hypertension | n (%) | 417 (82.2%) | 302 (72.4) | 105 (25.2) | 10 (2.4) |
| CAD | n (%) | 386 (76.1%) | 276 (71.5) | 101 (26.2) | 9 (2.3) |
| PCI | n (%) | 317 (62.5%) | 225 (71.0) | 85 (26.8) | 7 (2.2) |
| Angiogram | n (%) | 337 (66.5%) | 238 (70.6) | 90 (26.7) | 9 (2.7) |
| Alcohol use | n (%) | 207 (40.8%) | 147 (71.0) | 57 (27.5) | 3 (1.4) |
| Smoking history | n (%) | 213 (42.0%) | 161 (75.6) | 47 (22.1) | 5(2.3) |
| Depression | n (%) | 356 (70.2%) | 256 (71.9) | 92 (25.8) | 8 (2.2) |
| Anxiety | n (%) | 344 (67.9%) | 251 (73.0) | 85 (24.7) | 8 (2.3) |
†All percentages are within phenotype, except under column ‘Total’.
CAD: Coronary artery disease; PCI: Percutaneous coronary intervention; SD: Standard deviation.
Provider response to BPAs in poor metabolizers
According to guidelines published by CPIC [4], our EMR activates an automated BPA for any patient with two loss of function CYP2C19 alleles, when a provider attempts to prescribe sertraline, citalopram or escitalopram [2]. Within our study cohort of 507 individuals, the EMR was reviewed manually for every subject where a BPA had fired in the context of two loss of function CYP2C19 alleles (CYP2C19 poor metabolizer status). BPA events were extracted (with date stamps) electronically, then confirmed manually. Drug prescription dates were also extracted and confirmed longitudinally using the same approach. These two longitudinal datasets (BPA events, and SSRI prescribing events) were then merged to construct schematic timelines illustrating provider prescribing behavior in response to the BPAs. As shown in Figure 1, provider response to BPAs in these CYP2C19 poor metabolizers was found to fit into three distinct patterns: change drug, change dose or no change. In situations where no change in prescribing was implemented, patients were followed longitudinally to monitor risk/benefit ratio (Figure 1A).
Figure 1. . Provider response to Best Practice Advisories during sertraline prescription in CYP2C19 poor metabolizers.
Longitudinal data are shown for three representative examples, demonstrating three patterns of response. (A) No change: although QTc went up on sertraline, the provider did not change the prescription (but psychiatry was consulted to follow the patient and monitor the patient's risk/benefit ratio). (B) Change dose (decrease dose): The provider decreased sertraline dose, and since QTc stayed within normal range, the patient continues on a lower dose of sertraline (after documented discussion with provider). (C) Change drug (switch to an serotonin selective reuptake inhibitors not metabolized by CYP2C19): QTc goes up, BPA fires, provider stops offending agent (in this case, sertraline) the provider then starts a different serotonin selective reuptake inhibitors not metabolized by CYP2C19 (paroxetine per BPA).
BPA: Best practice advisories.
Net impact on prescribing pattern
Sertraline is the most commonly prescribed SSRI in our EMR. Among the 2000 unique individuals with CYP2C19 genotype available in our EMR, 159 have at some point received a sertraline prescription. Based upon temporal data, over half of these individuals showed a sertraline ‘start date’ that preceded their medication order date, indicating they had already been on the drug prior to the first prescription in our system (i.e., as a historical medication). Out of 159 study subjects in our cohort that ever received a prescription for sertraline, the prescription was initiated after the CYP2C19 genotype for 75 individuals. Within these 75, electronically extracted data indicate that sertraline was discontinued (for any reason) in 23 individuals, while the dose was later decreased in ten individuals and increased in five individuals.
The specific relationship between CYP2C19 genotype and current prescribing pattern for sertraline is illustrated in Figure 2. The relationship between genotype and prescription numbers are shown for all six available SSRIs (three CYP2C19 substrates, and three non-CYP2C19 substrates) as of January 2018.
Figure 2. . Current distribution of serotonin selective reuptake inhibitor prescription in 507 individuals based upon CYP2C19 genotype.
Histograms illustrate prescription frequency for three CYP2C19 substrates, and three non-CYP2C19 substrates.
Thus, in our study cohort of 507 individuals, only four CYP2C19 poor metabolizers are still on an SSRI known to be metabolized by CYP2C19 (only four on sertraline, 0 on citalopram and 0 on escitalopram). While the descriptive data illustrated in Figure 1 (representative timelines) clearly indicate that providers change their prescribing pattern in response to BPAs in CYP2C19 poor metabolizers (homozygous loss of function), the very small number of individuals with two loss of function alleles limits our statistical power to demonstrate a difference in prescribing patterns quantitatively in poor metabolizers. With respect to heterozygotes, we observed no difference in drug choice by genotype in patients with only one loss of function allele (p = 0.7335, Fisher's exact test). This is anticipated, because according to published guidelines [4], our EMR does not fire a BPA in heterozygotes for this class of drugs. However, given that STAR*D showed a difference in efficacy in heterozygotes [16], we opted to quantify the impact of CYP2C19 heterozygosity on toxicity as well (Part III, below).
QT interval in intermediate metabolizers
Last, to quantify the impact of CYP2C19 genotype on QT interval in heterozygotes, we manually abstracted all QT interval data (QTc) for study subjects with EKG data available specifically within the context of CYP2C19*1/*1 genotype and CYP2C19*1/*2 genotype. We limited our analysis to these two genotypes to maximize sample size while simultaneously avoiding misclassification bias introduced by other rare loss of function alleles (CYP2C19*3) as well as any gain of function alleles (CYP2C19*17). CYP2C19*2 is the most common loss of function allele for this gene in our study population, and other loss of function alleles beyond *2 may influence CYP2C19 enzyme activity to a different degree. Because misclassification can drive the odds ratio toward 1.0, a direct focused assessment of CYP2C19*1/*1 versus *1/*2 genotype retained phenotypic homogeneity among the heterozygotes, and it optimized sample size by utilizing the single most common detrimental allele. Results are summarized in Table 2.
Table 2. . There is no difference in mean QTc in intermediate metabolizers versus normal metabolizers.
| SSRI | CYP2C19 genotype | Mean | N | Standard deviation | p-values using t-test |
|---|---|---|---|---|---|
| Citalopram | *1/*1 | 440.558 | 52 | 33.4416 | 0.961 (two-tailed) |
| *1/*2 | 441.028 | 18 | 35.5844 | 0.519 (one-tailed) | |
| Total | 440.679 | 70 | 33.7434 | ||
| Escitalopram | *1/*1 | 441.875 | 28 | 28.1772 | 0.834 (two-tailed) |
| *1/*2 | 439.969 | 16 | 29.1379 | 0.417 (one-tailed) | |
| Total | 441.182 | 44 | 28.2057 | ||
| Sertraline | *1/*1 | 439.366 | 56 | 30.8100 | 0.906 (two-tailed) |
| *1/*2 | 440.146 | 24 | 25.1627 | 0.547 (one-tailed) | |
| Total | 439.600 | 80 | 29.0748 | ||
| Total | *1/*1 | 440.338 | 136 | 31.1273 | 0.995 (two-tailed) |
| *1/*2 | 440.371 | 58 | 29.2706 | 0.503 (one-tailed) | |
| Total | 440.348 | 194 | 30.5085 | ||
Current guidelines recommend action (i.e., change drug or change dose) if an SSRI known to be metabolized by CYP2C19 is prescribed in patients who arehomozygous for CYP2C19 loss of function alleles (poor metabolizers) [4]. They do not recommend action in patients who are heterozygous for loss of function (intermediate metabolizers). However, if heterozygotes were found to have a clinically significant increase in QT interval on these SSRIs, these patients may be at higher risk for a serious ADR (ventricular arrhythmia). We therefore tested the hypothesis that mean QTc on sertraline, citalopram and escitalopram may be increased in heterozygotes (CYP2C19*1/*2) when compared with patients with no loss of function alleles (CYP2C19*1/*1). As shown in Table 2, there is no difference in mean QTc between these two groups, indicating that BPAs are not needed in intermediate metabolizers.
Discussion
CYP2C19 genotype is becoming routinely available within EMRs. We now report the impact of extant CYP2C19 genotype (genotype already in the chart for another reason) on provider prescribing behavior for one of the most commonly prescribed class of drugs in the US, the SSRIs. This observational cohort study is unique because it describes the impact of automated decision support (BPAs based on CYP2C19 genotype) in the context of routine clinical care. Based on our findings, providers clearly change the drug (or reduce the dose and monitor) in response to BPAs fired in CYP2C19 poor metabolizers exposed to sertraline, citalopram and escitalopram. This is consistent with recommendations published by CPIC [4], as well as recent product label changes recommended by the FDA. For example, the citalopram product label now specifically warns against prescribing more than 20 mg daily in CYP2C19 poor metabolizers [15].
While the actionability of this drug–gene relationship is unequivocal in poor metabolizers, neither CPIC nor FDA currently recommend any intervention in intermediate metabolizers. We therefore compared mean QTc in patients with one loss of function CYP2C19 allele to mean QTc in patients with no loss of function alleles, for SSRIs known to be metabolized by CYP2C19 (sertraline, citalopram, escitalopram). It is reassuring to observe no difference between CYP2C19*1/*1 and CYP2C19*1/*2. Thus, the current approach to deploy BPAs in only poor metabolizers seems prudent for this drug–gene relationship.
One limitation of this study that must be noted is an inadequate sample size for quantifying the rate of adverse clinical events related to QT prolongation in poor metabolizers (torsades de pointes, other ventricular arrhythmias and death). Due to the relatively low frequency of the poor metabolizer genotype, and the rarity of clinical ADR events like torsades de pointes, larger observational studies aggregating data across multiple institutions will be needed to quantify the impact of this approach on event reduction for ventricular arrhythmias. Efforts are underway to address this limitation [17].
There is also great need to define the impact of this drug–gene relationship on the efficacy of SSRIs, particularly in patients with cardiovascular disease. Patients with CAD have a higher frequency of depression, and medication nonadherence in the context of failed antidepressant therapy can lead to poor outcome for CAD as well. In a recent systematic review of antidepressant efficacy in patients with CAD, current prescribing strategies are only ‘moderately efficacious’ [18]. Yet uncontrolled depression is an independent risk factor for CAD [19,20]. After controlling for covariates such as smoking, each additional mental health diagnosis further increases the odds of CAD by 44%, and patients with depression alone have a 60% higher odds of developing CAD [21]. In the future, the impact of CYP2C19 genotype on efficacy for SSRIs needs rigorous assessment specifically within patients that have CAD.
Summary points.
Variation in a single pharmacogene can impact clinical outcome for many classes of drugs.
CYP2C19 genotype influences risk of toxicity for sertraline, citalopram and escitalopram.
Providers changed prescribing patterns in response to poor metabolizer alerts in the electronic medical record by either changing the drug, changing the dose or monitoring serial EKGs.
Mean QTc was not different for intermediate metabolizers compared with normal metabolizers.
This observational cohort study is unique because it describes the impact of automated decision support (CYP2C19 genotype-based prescribing for serotonin selective reuptake inhibitors) in the context of routine clinical care.
Acknowledgements
The authors would like to thank E Dalseide (University of South Dakota) for assistance with the construction of Figures 1 and 2.
Footnotes
Financial & competing interests disclosure
This work was funded by the NIH (U01HG007253) and through a generous gift from T Denny Sanford that created Imagenetics (merging Internal Medicine and Genetics). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
No writing assistance was utilized in the production of this manuscript.
Ethical disclosure
In first section of methods, the authors state that they have obtained appropriate institutional review board approval for all human experimental investigations. No informed consent used as study was retrospective chart review.
Author's contributions
N Petry substantially contributed to the conception or design of the work, interpretation of data, drafting and revising the work, final approval and agreement to be accountable for work. R Lupu substantially contributed to the conception or design of the work, acquisition and analysis/interpretation of data, drafting and revising the work, final approval and agreement to be accountable for work. A Gohar substantially contributed to the acquisition and analysis of data for the work, drafting and revising the work, final approval and agreement to be accountable for work. EA Larson substantially contributed to the conception or design of the work, interpretation of data, drafting and revising the work, final approval and agreement to be accountable for work. C Peterson substantially contributed to the conception and acquisition of data for the work, drafting and revising the work, final approval and agreement to be accountable for work. V Williams substantially contributed to the conception and acquisition of data for the work, drafting and revising the work, final approval and agreement to be accountable for work. J Zhao substantially contributed to the conception or design of the work, acquisition and analysis/interpretation of data, drafting and revising the work, final approval and agreement to be accountable for work. RA Wilke substantially contributed to the conception or design of the work, interpretation of data, drafting and revising the work, final approval and agreement to be accountable for work. LJ Hines substantially contributed to the conception or design of the work, interpretation of data, drafting and revising the work, final approval and agreement to be accountable for work.
References
Papers of special note have been highlighted as: • of interest; •• of considerable interest
- 1.Cavallari LH, Denny JC, Lee CR, et al. Prospective Clinical implementation of CYP2C19-genotype guided antiplatelet therapy after PCI: a multi-site investigation of MACE outcomes in a real-world setting. Circulation. 2016;134:e711–e712. [Google Scholar]; • In patients with CPY2C19 loss of function, cardiovascular outcomes can be improved when clinical genotype is made available and alternative therapy is prescribed early after percutaneous coronary intervention (PCI).
- 2.Oberg V, Hines L, Differding J, Fisher M, Wilke RA. Navigating pleiotropy in precision medicine – pharmacogenes from trauma to behavioral health. Pharmacogenomics. 2016;17(5):499–505. doi: 10.2217/pgs.16.6. [DOI] [PubMed] [Google Scholar]; • As a pharmacogene, CYP2C19 can have pleiotropic clinical impact, influencing outcomes for both antiplatelet agents and antidepressants.
- 3.Swen JJ, Nijenhuis M, de Boer A, et al. Pharmacogenetics: from bench to byte – an update of guidelines. Clin. Pharmacol. Ther. 2011;89(5):662–673. doi: 10.1038/clpt.2011.34. [DOI] [PubMed] [Google Scholar]
- 4.Hicks JK, Bishop JR, Sangkuhl K, et al. Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline for CYP2D6 and CYP2C19 genotypes and dosing of selective serotonin reuptake inhibitors. Clin. Pharmacol. Ther. 2015;98:127–134. doi: 10.1002/cpt.147. [DOI] [PMC free article] [PubMed] [Google Scholar]; • Dosing recommendations have been published pertaining to drug efficacy and safety for fluvoxamine, paroxetine, citalopram, escitalopram and sertraline based on CYP2D6 and/or CYP2C19 genotype.
- 5.Hicks JK, Swen JJ, Thorn CF, et al. Clinical Pharmacogenetics Implementation Consortium guideline for CYP2D6 and CYP2C19 genotypes and dosing of tricyclic antidepressants. Clin. Pharmacol. Ther. 2013;93:402–408. doi: 10.1038/clpt.2013.2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Taylor WD Clinical practice. Depression in the elderly. N. Engl. J. Med. 2014;371(13):1228–1236. doi: 10.1056/NEJMcp1402180. [DOI] [PubMed] [Google Scholar]
- 7.Kohlmann K, Gierk B, Murray AM, Scholl A, Lehmann M, Lowe B. Base rates of depressive symptoms in patients with coronary heart disease: an individual symptom analysis. PLoS ONE. 2016;11(5):e0156167. doi: 10.1371/journal.pone.0156167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lichtman JH, Froelicher ES, Blumenthal JA, et al. Depression as a risk factor for poor prognosis among patients with acute coronary syndrome: systematic review and recommendations. Circulation. 2014;129(12):1350–1369. doi: 10.1161/CIR.0000000000000019. [DOI] [PubMed] [Google Scholar]; • Systematic review identifying depression as risk factor for all-cause and cardiovascular mortality after acute coronary syndrome.
- 9.Pizzi C, Wilhelmina A, Rutjes S, et al. Meta-analysis of selective serotonin reuptake inhibitors in patients with depression and coronary heart disease. Am. J. Cardiol. 2011;107(7):972–979. doi: 10.1016/j.amjcard.2010.11.017. [DOI] [PubMed] [Google Scholar]
- 10.Teply RM, Packard KA, White ND, Hilleman DE, DiNicolantonio JJ. Treatment of depression in patients with concomitant cardiac disease. Prog. Cardiovasc. Dis. 2016;58:514–528. doi: 10.1016/j.pcad.2015.11.003. [DOI] [PubMed] [Google Scholar]
- 11.Lin E, Chen PS. Pharmacogenomics with antidepressants in the STAR*D study. Pharmacogenomics. 2008;9(7):935–946. doi: 10.2217/14622416.9.7.935. [DOI] [PubMed] [Google Scholar]
- 12.Sinyor M, Schaffer A, Levitt A. The sequenced treatment alternatives to relieve depression (STAR*D) trial: a review. Can. J. Psychiatry. 2010;55(3):126–135. doi: 10.1177/070674371005500303. [DOI] [PubMed] [Google Scholar]
- 13.Westenberg HGM, Sandner C. Tolerability and safety of fluvoxamine and other antidepressants. Int. J. Clin. Pract. 2006;60(4):482–491. doi: 10.1111/j.1368-5031.2006.00865.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Bednar MM, Harrigan EP, Anziano RJ, Camm AJ, Ruskin JN. The QT interval. Prog. Cardiovasc. Dis. 2001;43(5 Suppl. 1):1–45. doi: 10.1053/pcad.2001.21469. [DOI] [PubMed] [Google Scholar]
- 15.US FDA. FDA Drug Safety Communication: revised recommendations for Celexa (citalopram hydrobromide) related to a potential risk of abnormal heart rhythms with high doses. 2012. www.fda.gov/Drugs/DrugSafety/ucm297391.htm
- 16.Mrazek DA, Biernacka JM, O'Kane DJ, et al. CYP2C19 variation and citalopram response. Pharmacogenet. Genomics. 2011;21(1):1–9. doi: 10.1097/fpc.0b013e328340bc5a. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Cavallari LH, Beitelshees AL, Blake KV, et al. The IGNITE pharmacogenetics working group: an opportunity for building evidence with pharmacogenetic implementation in a real-world setting. Clin. Transl. Sci. 2017;10(3):143–146. doi: 10.1111/cts.12456. [DOI] [PMC free article] [PubMed] [Google Scholar]; • Collaborative working group examining outcomes for CYP2C19 genotype-guided antiplatelet therapy after percutaneous coronary intervention across institutions.
- 18.Nieuwsma JA, Williams JW, Namdari N, et al. Diagnostic accuracy of screening tests and treatment for post-acute coronary syndrome depression. A systematic review. Ann. Intern. Med. 2017;167:725–735. doi: 10.7326/M17-1811. [DOI] [PubMed] [Google Scholar]
- 19.Khawaja IS, Westermeyer JJ, Gajwani P, Feinstein RE. Depression and coronary artery disease: the association, mechanisms, and therapeutic implications. Psychiatry (Edgmont) 2009;6(1):38–51. [PMC free article] [PubMed] [Google Scholar]
- 20.Nemeroff CB, Goldschmidt-Clermont PJ. Heartache and heartbreak – the link between depression and cardiovascular disease. Nat. Rev. Cardiol. 2012;9(9):526–39. doi: 10.1038/nrcardio.2012.91. [DOI] [PubMed] [Google Scholar]
- 21.Gerber MR, King MW, Iverson KM, Pineles SL, Haskell SG. Association between mental health burden and coronary artery disease in US women veterans over 45: a national cross-sectional study. J. Womens Health (Larchmt) 2018;27(3):238–244. doi: 10.1089/jwh.2017.6328. [DOI] [PubMed] [Google Scholar]; • Women military veterans over age 45 with depression had 60% higher odds of coronary artery disease.