Abstract
At standard doses used for schizophrenia or bipolar disorder, quetiapine has been associated with weight gain and increased levels of triglycerides, total cholesterol and low‐density lipoprotein (LDL) cholesterol, which are risk factors for cardiovascular morbidity and mortality. However, this drug is also commonly used off‐label at low doses for anxiolytic or hypnotic purposes, and its cardiovascular safety at these doses is unknown. We aimed to assess the risk of major adverse cardiovascular events with use of low‐dose quetiapine compared to use of Z‐drug hypnotics in a nationwide, active comparator‐controlled cohort study. The cohort included new users of either drugs in Denmark from 2003 to 2017, aged 18‐85 years, without history of ischemic stroke, myocardial infarction, cancer, and severe mental illness. The main outcome was the occurrence of major adverse cardiovascular events, defined as non‐fatal myocardial infarction or ischemic stroke, or death from cardiovascular causes. Selective serotonin reuptake inhibitors (SSRIs) were used as an alternative comparator in sensitivity analyses. Altogether, we compared 60,566 low‐dose quetiapine users with 454,567 Z‐drug users, followed for 890,198 person‐years in intent‐to‐treat analysis, and 330,334 person‐years in as‐treated analysis. In intention‐to‐treat analysis, low‐dose quetiapine was associated with an increased risk of major adverse cardiovascular events (adjusted hazard ratio, aHR=1.13, 95% CI: 1.02‐1.24, p=0.014) and cardiovascular death (aHR=1.26, 95% CI: 1.11‐1.43, p<0.001). In as‐treated analysis, continuous low‐dose quetiapine use was associated with increased risk of major adverse cardiovascular events (aHR=1.52, 95% CI: 1.35‐1.70, p<0.001), non‐fatal ischemic stroke (aHR=1.37, 95% CI: 1.13‐1.68, p=0.002) and cardiovascular death (aHR=1.90, 95% CI: 1.64‐2.19, p<0.001). The risk of major adverse cardiovascular events was greater in women (aHR=1.28, p=0.02) and those aged ≥65 years at initiation (aHR=1.24, p<0.001). Compared to SSRIs, low‐dose quetiapine use was associated with an increased risk of major adverse cardiovascular events (aHR=1.42, p<0.001), non‐fatal ischemic stroke (aHR=1.27, p=0.0028) and cardiovascular death (aHR=1.72, p<0.001). So, we conclude that the use of low‐dose quetiapine is associated with an increased risk of major adverse cardiovascular events, especially in women and the elderly. On the basis of these findings, we suggest that use of off‐label low‐dose quetiapine for sedative or hypnotic purposes should be discouraged.
Keywords: Low‐dose quetiapine, major adverse cardiovascular events, death from cardiovascular causes, off‐label use, anxiolytic or hypnotic use, cardiovascular safety
Quetiapine is a second‐generation antipsychotic labelled for use in schizophrenia, bipolar disorder, and as adjunctive treatment in major depression 1 . In addition to its antipsychotic effects, quetiapine has anxiolytic, sedative and hypnotic properties, due to its high affinity to serotonergic, histaminergic and muscarinic receptors 2 . These additional properties have led to considerable off‐label use of the drug, which has been documented across several countries3, 4, 5, 6. Quetiapine is now among the most frequently prescribed antipsychotics worldwide, with approximately 2 million users in the US alone7, 8. Evidence suggests that prescription by non‐psychiatrists contributes significantly to the increased off‐label, low‐dose use of the drug5, 9.
Antipsychotics, in general, have been associated with increased risk of cardiovascular morbidity and sudden cardiac death10, 11, 12. The increased risk of cardiovascular morbidity is driven by metabolic abnormalities, whereas the increased risk of sudden cardiac death is likely to stem from QT prolongation (increasing the risk of ventricular arrhythmias)13, 14.
While quetiapine has not been associated with clinically significant QT prolongation compared to other antipsychotics 15 , it has been found to induce weight gain and considerable increases in the levels of triglycerides, total cholesterol, and low‐density lipoprotein (LDL) cholesterol 16 , all of which are important risk factors for the development of cardiovascular disease morbidity 17 . In a study of 284,234 non‐elderly adults in the US, including 12,094 patients treated with quetiapine and 253,027 receiving antidepressants 18 , the use of quetiapine was found to be associated with an increased risk of stroke, hypertensive heart disease, and coronary artery disease.
Whether the increased risk of cardiovascular events observed with standard doses of quetiapine in schizophrenia is also present with the low doses mainly used for anxiety and insomnia is an important question, given the widespread off‐label use of the drug. Practical limitations with randomized controlled trials (e.g., small sample size and limited follow‐up duration) makes them less suitable to study long‐term adverse effects, such as cardiovascular morbidity or mortality. Furthermore, the exclusion of individuals with significant physical comorbidities from those trials limits the generalizability of their findings to the real‐world population 19 . For these reasons, observational studies are important to assess the long‐term cardiovascular safety of off‐label/low‐dose antipsychotic treatment in a representative population.
In this study, our aim was to assess the association between prescription of low‐dose quetiapine and major adverse cardiovascular events with an active comparator‐controlled design, using routinely collected health data from nationwide registers. We hypothesized that low‐dose quetiapine would be associated with an increased risk of major adverse cardiovascular events compared to Z‐drugs and to selective serotonin reuptake inhibitors (SSRIs).
METHODS
Study design
We conducted a new‐user, active‐comparator cohort study based on data from Danish nationwide health care registers. We included initiators of benzodiazepine‐related drugs (Z‐drugs) as an active comparator, because this drug class is widely used for the treatment of insomnia 20 . In sensitivity analyses, we also used SSRIs as an active comparator, which are used to treat anxiety, the second off‐label indication for low‐dose quetiapine. Since SSRIs have been associated with a potential increase in cardiometabolic risk 10 , these sensitivity analyses were used to test the generalizability and robustness of the results of the primary analysis. Additionally, we used propensity score weighting methods to control for other potential confounders while utilizing the full cohort size.
Access to pseudonymized health care data was approved by the Danish Health Data Authority. No ethical committee approval is needed for purely register‐based studies according to Danish legislation. The study protocol was registered in the European Union Electronic Register of Post‐Authorization Studies (EUPAS‐38508), and data presentation followed the REporting of studies Conducted using Observational Routinely collected health Data statement for PharmacoEpidemiology (RECORD‐PE) 21 .
Data sources
Prescriptions of quetiapine, comparators and other medications were identified in the Danish National Prescription Register 22 , in which all prescriptions redeemed at community pharmacies are captured. In‐ and outpatient diagnoses were obtained from the Danish National Patient Register 23 for outcome and comorbidity assessment. Information on vital status and migration was collected from the Danish Civil Registration System 24 , and information on cause of death (for outcome assessment) was obtained from the Danish Cause of Death Register 25 . Further description of the registers is provided in the supplementary information.
Exposure
We identified all individuals who had filled prescriptions for quetiapine between January 1, 2003 and December 31, 2017 in the Danish National Prescription Register. The date of their first quetiapine dispensing was taken as the index date. From this population, we excluded individuals who: a) had filled prescriptions for the comparator or other antipsychotics within 365 days before the index date; b) had filled prescriptions for the comparator on the index date; c) had filled prescriptions for quetiapine in tablet strengths >50 mg on the index date; d) had a history of myocardial infarction, stroke, cancer, or severe mental illness (for definitions, see supplementary information); e) had not been continuously residing in Denmark for 365 days before the index date; and f) were <18 or >85 years old on the index date. Individuals in the comparator group were required to fulfill the same conditions.
Outcome measures
The main outcome was the occurrence of major adverse cardiovascular events, defined by first record of either non‐fatal myocardial infarction or non‐fatal ischemic stroke, or by death from cardiovascular causes. Secondary outcomes were each of the above cardiovascular events. The ICD codes used for outcome‐defining events are provided in the supplementary information.
Propensity score
We estimated each individual's propensity to fill prescriptions for low‐dose quetiapine using logistic regression, including age, sex, year of cohort entry, and the 100 most influential covariates, selected using a high‐dimensional propensity score algorithm 26 assessing all prescription fills and hospital diagnoses within 365 days before the index date (see supplementary information). Covariate balance was assessed using standardized mean differences (SMD), with SMD ≤0.1 indicating sufficient balance.
Intention‐to‐treat analysis
In intention‐to‐treat analysis, the study population was restricted to individuals with ≥1 additional prescription within 180 days of the index date, to minimize exposure misclassification. Follow‐up began at day 181 and lasted until individuals either experienced the outcome of interest, or died for non‐cardiovascular reasons, or were censored. Reasons for censoring were: filling of prescriptions for the other study drug, filling of >1 prescription for other antipsychotics, filling of prescriptions for quetiapine in tablet strengths >50 mg, receiving a diagnosis of a severe mental disorder, emigration, ≥10 years of follow‐up, or end of data availability, whichever came first.
To adjust for baseline confounding, we used fine stratification weights by trimming non‐overlapping regions of the propensity score distribution and then constructing ten propensity score strata where we weighted Z‐drug users (and SSRI users in the sensitivity analyses) according to the distribution of low‐dose quetiapine users. Hazard ratios (HRs) with 95% confidence intervals (CIs) were estimated using Cox regression models adjusted by the fine stratification weights. The proportion of cases attributable to low‐dose quetiapine use was calculated as (HR–1)/HR.
These analyses, as well as all the following ones, were conducted using Stata MP, release 16.1 (StataCorp, College Station, TX, USA).
As‐treated analysis
To assess the relationship between continuous treatment and outcomes, we conducted an as‐treated analysis where individuals were followed from the index date until they either experienced the outcome of interest, or died (from non‐cardiovascular reasons), ended their first treatment episode, or were censored. Reasons for censoring were similar to those used in the intent‐to‐treat analysis, except that the maximum follow‐up was confined to five years, as very few individuals remained on treatment beyond that point.
Treatment episodes were constructed by assigning a duration to each prescription corresponding to the number of tablets dispensed (assuming use of one tablet/day). To the duration of each prescription, we added a grace period of 120 days to account for irregular use. Gaps exceeding 120 days were considered as the end of the first treatment episode. These additional 120 days of observation were also added to the last prescription to capture events occurring shortly after (and potentially associated with) the treatment episode, and thus avoid immortal time bias 27 .
To adjust for baseline confounding, we used inverse probability of treatment weights and inverse probability of censoring weights estimated from baseline covariates (included in the high‐dimensional propensity score algorithm). The inverse probability of censoring weights was updated every 90 days and truncated at the 1st and 99th percentile. Pooled logistic regression was used with the product of inverse probability of treatment weights and inverse probability of censoring weights to estimate HRs, and 95% CIs were computed using robust variance estimators 28 .
Subgroup and sensitivity analyses
To assess potential differences in risk between subgroups, we conducted analyses stratified on sex (male/female), age group (</≥65 years), history of ischemic heart disease (yes/no), and history of diabetes (yes/no). Furthermore, we conducted three sensitivity analyses: a) including any tablet strength of quetiapine in the exposure definition, to assess the potential difference in risk with dose; b) using SSRIs as an alternative comparator that targets individuals suffering from anxiety instead of insomnia; and c) excluding individuals with in‐/outpatient contacts for major depression, which might increase the risk of cardiovascular events.
Case‐control analysis
To investigate whether cumulative dose of quetiapine (as low‐dose treatment) was associated with the outcomes, we additionally conducted a case‐control analysis nested among quetiapine users.
For each case exposed to quetiapine, we identified the pool of quetiapine users of same sex and birth year who did not have the outcomes and randomly selected 20 such controls among them, or as many as were available if there were <20 controls. The controls were given an index date identical to their matched case. We estimated odds ratios (ORs) with 95% CIs for the association between cumulative quetiapine dose and major adverse cardiovascular events. The cumulative dose of quetiapine was assessed between first prescription and censoring (similar to the intent‐to‐treat analysis), and analyzed using predefined cumulative total dose strata (2,501‐5,000, 5,001‐10,000, 10,001‐25,000, 25,001‐50,000, >50,000 mg). Individuals with a cumulative dose of ≤2,500 mg were used as reference group for the analyses, as this dose corresponds to 100 tablets of 25 mg quetiapine (the smallest marketed package of quetiapine in Denmark). Trends in the association between outcomes and total cumulative quetiapine dose were tested using pooled logistic regression with dose strata as independent variable.
RESULTS
A total of 515,133 patients were included in the cohort (58% females; median age: 49 years, interquartile range, IQR: 36‐63). Of these, 60,566 were users of low‐dose quetiapine, and 454,567 of Z‐drugs. Risk factors for cardiovascular disease and use of preventive medications did not differ significantly between the groups (see Table 1).
Table 1.
Baseline characteristics of low‐dose quetiapine and Z‐drug users
| Intention‐to‐treat population | SMD | As‐treated population | SMD | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Low‐dose quetiapine (N=22,849) | Z‐drugs (N=131,623) | Total (N=154,472) | Before FSW | After FSW | Low‐dose quetiapine (N=60,566) | Z‐drugs (N=454,567) | Total (N=515,133) | Before IPTW | After IPTW | |
| Sex | ||||||||||
| Female, N (%) | 12,387 (54) | 76,244 (58) | 88,631 (57) | 0.1 | <0.1 | 32,347 (53) | 266,113 (59) | 298,460 (58) | 0.1 | <0.1 |
| Age, N (%) | ||||||||||
| Median (IQR) | 43 (29‐59) | 54 (42‐68) | 53 (40‐67) | 0.5 | <0.1 | 40 (27‐54) | 50 (38‐64) | 49 (36‐63) | 0.4 | 0.2 |
| 18‐44 years, N (%) | 12,021 (53) | 40,256 (31) | 52,277 (34) | 0.5 | <0.1 | 35,518 (59) | 177,099 (39) | 212,617 (41) | 0.4 | 0.1 |
| 45‐64 years, N (%) | 6,274 (27) | 49,890 (38) | 56,164 (36) | 0.2 | 0.1 | 16,184 (27) | 170,825 (38) | 187,009 (36) | 0.2 | 0.1 |
| 65‐85 years, N (%) | 4,554 (20) | 41,477 (32) | 46,031 (30) | 0.3 | <0.1 | 8,864 (15) | 106,643 (23) | 115,507 (22) | 0.2 | 0.2 |
| Year of cohort entry, N (%) | ||||||||||
| 2003‐2005 | 681 (3) | 40,683 (31) | 41,364 (27) | 0.8 | 0.1 | 1,306 (2) | 124,303 (27) | 125,609 (24) | 0.8 | 0.2 |
| 2006‐2008 | 2,193 (10) | 32,334 (25) | 34,527 (22) | 0.4 | <0.1 | 5,047 (8) | 110,128 (24) | 115,175 (22) | 0.4 | 0.1 |
| 2009‐2011 | 4,047 (18) | 23,679 (18) | 27,726 (18) | <0.1 | <0.1 | 9,881 (16) | 85,913 (19) | 95,794 (19) | 0.1 | 0.1 |
| 2012‐2014 | 7,579 (33) | 19,615 (15) | 27,194 (18) | 0.4 | 0.1 | 20,168 (33) | 73,492 (16) | 93,660 (18) | 0.4 | 0.1 |
| 2015‐2017 | 8,349 (37) | 15,312 (12) | 23,661 (15) | 0.6 | 0.1 | 24,164 (40) | 60,731 (13) | 84,895 (16) | 0.6 | 0.1 |
| Comorbidities, N (%) | ||||||||||
| Ischemic heart disease | 958 (4) | 7,629 (6) | 8,587 (6) | 0.1 | <0.1 | 2,126 (4) | 20,508 (5) | 22,634 (4) | 0.1 | 0.1 |
| Heart failure | 244 (1) | 3,196 (2) | 3,440 (2) | 0.1 | <0.1 | 565 (<1) | 7,847 (2) | 8,412 (2) | 0.1 | <0.1 |
| Peripheral vascular disease | 435 (2) | 4,556 (3) | 4,991 (3) | 0.1 | <0.1 | 1,025 (2) | 11,605 (3) | 12,630 (2) | 0.1 | <0.1 |
| Hypertension | 4,135 (18) | 34,596 (26) | 38,731 (25) | 0.2 | <0.1 | 9,297 (15) | 94,343 (21) | 103,640 (20) | 0.1 | 0.1 |
| COPD | 2,859 (13) | 18,304 (14) | 21,163 (14) | <0.1 | <0.1 | 7,102 (12) | 53,355 (12) | 60,457 (12) | <0.1 | <0.1 |
| Diabetes | 1,326 (6) | 8,808 (7) | 10,134 (7) | <0.1 | <0.1 | 3,213 (5) | 24,475 (5) | 27,688 (5) | <0.1 | 0.1 |
| Alcohol‐related disorders | 6,621 (29) | 14,293 (11) | 20,914 (14) | 0.5 | 0.1 | 17,956 (30) | 54,648 (12) | 72,604 (14) | 0.4 | 0.1 |
| Obesity | 1,828 (8) | 5,605 (4) | 7,433 (5) | 0.2 | <0.1 | 4,901 (8) | 18,711 (4) | 23,612 (5) | 0.2 | <0.1 |
| Major depression | 4,155 (18) | 3,764 (3) | 7,919 (5) | 0.5 | 0.1 | 10,278 (17) | 10,795 (2) | 21,073 (4) | 0.5 | <0.1 |
| Recent use of medications, N (%) | ||||||||||
| Acetylsalicylic acid | 1,797 (8) | 15,834 (12) | 17,631 (11) | 0.1 | <0.1 | 3,660 (6) | 40,661 (9) | 44,321 (9) | 0.1 | 0.1 |
| Statins | 2,464 (11) | 18,517 (14) | 20,981 (14) | 0.1 | <0.1 | 5,556 (9) | 52,116 (11) | 57,672 (11) | 0.1 | <0.1 |
SMD – standardized mean difference, FSW – fine stratification weighting, IPTW – inverse probability of treatment weighting, IQR – interquartile range, COPD – chronic obstructive pulmonary disease
The intention‐to‐treat population (>1 prescription required) included 22,849 low‐dose quetiapine users and 131,623 Z‐drug users. The total follow‐up was 890,198 person‐years, with a median of 2.6 years (IQR: 1.2‐4.7) for low‐dose quetiapine users and 7.0 years (IQR: 3.2‐9.5) for Z‐drug users. In this population, 59% of low‐dose quetiapine users and 55% of Z‐drug users had ≥5 prescriptions (see also supplementary information).
In the as‐treated population, the total follow‐up was 330,334 person‐years, with a median follow‐up of 7.2 months (IQR: 7.2‐11.8) for low‐dose quetiapine users and 4.6 months (IQR: 4.3‐5.3) for Z‐drug users. In this population, 50% of low‐dose quetiapine users and 29% of Z‐drug users had ≥2 prescriptions (see also supplementary information).
In the intent‐to‐treat analysis, there were 877 major adverse cardiovascular events among low‐dose quetiapine users and 11,464 among Z‐drugs users. After adjusting for baseline covariates, the risk for major adverse cardiovascular events was significantly higher with use of low‐dose quetiapine (adjusted HR, aHR=1.13, 95% CI: 1.02‐1.24, p=0.014; attributable proportion of cases, APC=11%, 95% CI: 2‐19%) (see Table 2).
Table 2.
Risk of major adverse cardiovascular events and secondary outcomes with use of low‐dose quetiapine (QUE) compared to use of Z‐drugs (ZDR)
| N. patients | N. events | Follow‐up | Hazard ratio (95% CI) | p | |
|---|---|---|---|---|---|
| QUE/ZDR | QUE/ZDR | QUE/ZDR | |||
| Major adverse cardiovascular events | |||||
| Intention‐to‐treat analysis (adjusted) | 22,827/131,582 | 877/11,464 | 73/817 | 1.13 (1.02‐1.24) | 0.014 |
| As‐treated analysis (adjusted) | 60,564/454,552 | 850/5,513 | 59/272 | 1.11 (1.00‐1.24) | 0.046 |
| As‐treated analysis (fully adjusted) | 60,564/454,552 | 850/5,513 | 59/272 | 1.52 (1.35‐1.70) | <0.001 |
| Non‐fatal myocardial infarction | |||||
| Intention‐to‐treat analysis (adjusted) | 22,828/131,588 | 138/2,895 | 74/829 | 0.91 (0.73‐1.14) | 0.42 |
| As‐treated analysis (adjusted) | 60,564/454,552 | 109/1,307 | 59/273 | 0.69 (0.52‐0.90) | 0.007 |
| As‐treated analysis (fully adjusted) | 60,564/454,552 | 109/1,307 | 59/273 | 0.91 (0.69‐1.21) | 0.53 |
| Non‐fatal ischemic stroke | |||||
| Intention‐to‐treat analysis (adjusted) | 22,827/131,586 | 267/4,378 | 74/825 | 0.98 (0.83‐1.15) | 0.81 |
| As‐treated analysis (adjusted) | 60,564/454,552 | 256/1,920 | 59/273 | 1.01 (0.83‐1.21) | 0.95 |
| As‐treated analysis (fully adjusted) | 60,564/454,552 | 256/1,920 | 59/273 | 1.37 (1.13‐1.68) | 0.002 |
| Death from cardiovascular causes | |||||
| Intention‐to‐treat analysis (adjusted) | 22,828/131,593 | 565/6,262 | 74/837 | 1.26 (1.11‐1.43) | <0.001 |
| As‐treated analysis (adjusted) | 60,564/454,552 | 558/2,903 | 59/274 | 1.37 (1.20‐1.56) | <0.001 |
| As‐treated analysis (fully adjusted) | 60,564/454,552 | 558/2,903 | 59/274 | 1.90 (1.64‐2.19) | <0.001 |
Follow‐up in 1,000 person‐years. The intention‐to‐treat analysis is adjusted for baseline confounding by fine stratification weights. The as‐treated analysis is adjusted for baseline confounding by inverse probability of treatment weights, or fully adjusted by inverse probability of treatment weights and informative censoring by inverse probability of censoring weights.
In the as‐treated analysis, after adjusting for baseline confounding, the risk of major adverse cardiovascular events was significantly higher with use of low‐dose quetiapine (aHR=1.11, 95% CI: 1.00‐1.24, p=0.046). With additional adjustment for informative censoring (using inverse probability of censoring weights), continuous use of low‐dose quetiapine was significantly associated with major adverse cardiovascular events compared to continuous use of Z‐drugs (aHR=1.52, 95% CI: 1.35‐1.70, p<0.001; APC=34%, 95% CI: 26‐41%) (see Table 2).
Analysis of individual major adverse cardiovascular events showed that the association was mainly driven by an increased risk of cardiovascular death (intent‐to‐treat analysis: aHR=1.26, 95% CI: 1.11‐1.43, p<0.001; as‐treated analysis: aHR=1.90, 95% CI: 1.64‐2.19, p<0.001). Use of low‐dose quetiapine was not associated with increased risk of non‐fatal myocardial infarction in either intent‐to‐treat analysis (aHR=0.91, 95% CI: 0.73‐1.14, p=0.42) or as‐treated analysis (aHR=0.91, 95% CI: 0.69‐1.21, p=0.53). An association between use of low‐dose quetiapine and non‐fatal ischemic stroke was only present in the as‐treated analysis (aHR=1.37, 95% CI: 1.13‐1.68, p=0.002) (see Table 2).
The cumulative incidence of major adverse cardiovascular events and secondary outcomes is shown in Figure 1. For major adverse cardiovascular events and cardiovascular death, the difference between groups became evident beyond 3‐4 years of follow‐up.
Figure 1.
Cumulative incidence of major adverse cardiovascular events with use of low‐dose quetiapine compared to use of Z‐drugs (intention‐to‐treat analysis using fine stratification weights). A: major adverse cardiovascular events, B: non‐fatal myocardial infarction, C: non‐fatal ischemic stroke, D: death from cardiovascular causes
In subgroup analyses of the intent‐to‐treat population, use of low‐dose quetiapine had a stronger association with major adverse cardiovascular events among females than males (aHR=1.28, 95% CI: 1.11‐1.48 vs. 1.02, 95% CI: 0.90‐1.16, p=0.02). Age ≥65 years at initiation of low‐dose quetiapine treatment was also more strongly associated with major adverse cardiovascular events (aHR=1.24, 95% CI: 1.10‐1.40) vs. initiation at age <65 years (aHR=0.88, 95% CI: 0.75‐1.03, p<0.001). A history of ischemic heart disease or diabetes was not significantly associated with an increased risk of major adverse cardiovascular events (p=0.67 and p=0.42, respectively) (see Table 3). None of these subgroups was significantly related to an increased risk of non‐fatal myocardial infarction or ischemic stroke, when comparing low‐dose quetiapine with Z‐drug use. Female sex and age ≥65 years at treatment initiation were associated with increased risk of cardiovascular death, when comparing low‐dose quetiapine with Z‐drug use (p=0.0086 and p<0.001, respectively) (see supplementary information).
Table 3.
Subgroup analysis of major adverse cardiovascular events with use of low‐dose quetiapine compared to use of Z‐drugs in the intention‐to‐treat population
| Low‐dose quetiapine | Z‐drugs | |||||||
|---|---|---|---|---|---|---|---|---|
| N. patients (%) | N. events | Rate (per 1,000 person‐years) | N. patients (%) | N. events | Rate (per 1,000 person‐years) | Adjusted hazard ratio (95% CI) | p | |
| Sex | ||||||||
| Female, N (%) | 12,375 (54) | 424 | 10.3 | 76,226 (58) | 5,509 | 11.3 | 1.28 (1.11‐1.48) | |
| Male, N (%) | 10,452 (46) | 453 | 13.9 | 55,356 (42) | 5,955 | 18.2 | 1.02 (0.90‐1.16) | 0.02 |
| Age group, N (%) | ||||||||
| <65 years | 18,276 (80) | 268 | 4.3 | 90,130 (68) | 3,629 | 6.1 | 0.88 (0.75‐1.03) | |
| ≥65 years | 4,551 (20) | 609 | 51.6 | 41,452 (329) | 7,835 | 34.6 | 1.24 (1.10‐1.40) | <0.001 |
| History of ischemic heart disease, N (%) | ||||||||
| No | 21,870 (96) | 743 | 10.5 | 123,959 (94) | 9,745 | 12.6 | 1.13 (1.02‐1.25) | |
| Yes | 957 (4) | 134 | 51.5 | 7,623 (6) | 1,719 | 42.4 | 1.19 (0.93‐1.53) | 0.67 |
| History of diabetes, N (%) | ||||||||
| No | 21,504 (94) | 760 | 10.9 | 122,781 (93) | 9,773 | 12.7 | 1.16 (1.05‐1.29) | |
| Yes | 1,323 (6) | 117 | 31.5 | 8,801 (7) | 1,691 | 37.6 | 1.04 (0.81‐1.34) | 0.42 |
| Overall estimate | 22,827 (100) | 877 | 11.9 | 131,582 (100) | 11,464 | 14.0 | 1.13 (1.02‐1.24) | |
The hazard ratio is adjusted for baseline confounding by fine stratification weights
In sensitivity analyses of the intent‐to‐treat population, low‐dose quetiapine use compared with use of SSRIs was associated with an increased risk of major adverse cardiovascular events (aHR=1.42, p<0.001), non‐fatal ischemic stroke (aHR=1.27, p=0.0028) and cardiovascular death (aHR=1.72, p<0.001), but not of non‐fatal myocardial infarction (aHR=0.86, p=0.23) (see also supplementary information).
Including all tablet strengths of quetiapine in the exposure definition did not result in increased risk of either major adverse cardiovascular events (aHR: 1.00 vs. 1.13) or cardiovascular death (aHR: 1.07 vs. 1.26). The same was observed when excluding individuals with in‐/outpatient contacts for major depression (aHR: 1.27 vs. 1.13 for major cardiovascular events; 1.52 vs. 1.26 for cardiovascular death) (see also supplementary information).
In the case‐control analysis, increasing cumulative doses of quetiapine (as low‐dose treatment) was not significantly associated with increased risk of major adverse cardiovascular events (p=0.21), while it was associated with a significantly increased risk of cardiovascular death with cumulative doses ≥50,000 mg (OR=1.32, 95% CI: 1.09‐1.60, p=0.014) (see supplementary information).
DISCUSSION
In this large cohort study, we found an increased risk of major adverse cardiovascular events with low‐dose quetiapine, one of the most frequent uses of any individual antipsychotic medication, compared to use of Z‐drugs. This increased risk was mainly driven by an increased risk of cardiovascular death, while we only found an increased risk of non‐fatal ischemic stroke with continuous treatment, and no increase in the risk of non‐fatal myocardial infarction. The association between use of low‐dose quetiapine and major adverse cardiovascular events or cardiovascular death was robust, as it was confirmed when analyzing continuous low‐dose quetiapine treatment and when using SSRIs as an alternative comparator.
The increased risk (in intent‐to‐treat analysis) of major adverse cardiovascular events and cardiovascular death, but not of non‐fatal myocardial infarction or ischemic stroke, might seem surprising, given quetiapine's association with dyslipidemia 16 . Actually, the number of non‐fatal myocardial infarction or ischemic stroke events during the follow‐up period was low in each group. To ensure internal validity, we used strict censoring criteria in both intent‐to‐treat and as‐treated analyses (e.g., filling >1 prescription for other antipsychotics), which might have limited the follow‐up duration to capture a potential difference between groups in the number of cardiovascular events. On the other hand, cardiovascular death was assessed based on information from death certificates, which includes both the primary cause of death and any underlying causes of death. Additionally, the outcome definition incorporated a wide selection of cardiovascular diagnoses. Therefore, the increased risk of cardiovascular death might reflect a higher degree of cardiovascular morbidity (e.g., from ischemic heart disease or heart failure) among low‐dose quetiapine initiators, potentially caused by dyslipidemia, than is captured by the relatively well‐defined conditions of myocardial infarction and ischemic stroke.
The absence of a clear relationship between the risk of cardiovascular outcomes and cumulative quetiapine dose can have different interpretations. First, the observed risk of cardiovascular outcomes might be due to residual confounding by risk factors associated with off‐label quetiapine use (e.g., mental illness, smoking, unhealthy lifestyle). However, we adjusted analyses for 100 potentially relevant confounders, making this less likely. Second, even the use of low cumulative doses adopted as reference category in the case‐control analysis (≤2,500 mg) might be sufficient to increase the risk of major adverse cardiovascular events, either due to increased lipid levels or other adverse cardiometabolic mechanisms. The latter interpretation is supported by data from randomized controlled trials, where even short‐term quetiapine exposure significantly increased lipid levels 16 .
Z‐drugs users, especially long‐term users, might have more physical comorbidities than the average low‐dose quetiapine user. Therefore, we conducted a sensitivity analysis with SSRI users as the comparator group. SSRI users were not chosen as the primary comparator, as this drug class can also cause weight gain and possibly increase cardiovascular morbidity/mortality 10 , and simultaneously has platelet‐inhibiting properties, which might result in a decreased risk of thrombotic events 29 , although this has not been demonstrated with certainty 30 . The increased risk of major adverse cardiovascular events, non‐fatal ischemic stroke, and cardiovascular death in low‐dose quetiapine users vs. SSRI users further supports the results of the main analyses. The small differences between results from the main and sensitivity analyses incorporating all tablet strengths of quetiapine (25‐400 mg) might suggest that the association between quetiapine dose and cardiovascular outcomes is largely independent of daily dose.
Despite the common and increasing use of low‐dose quetiapine as an anxiolytic or hypnotic3, 4, 5, 6, 31, no study has investigated its long‐term cardiovascular safety in a nationwide setting with long follow‐up as the present analysis. Cardiometabolic risk factors with use of low‐dose quetiapine have been investigated in smaller cohort studies32, 33 and in one nationwide database study 34 , but the evidence of cardiovascular risk/safety has been insufficient35, 36. The only nationwide database study 34 found increased risk of cardiovascular death with 6‐12 months of cumulative exposure, but pooled low‐dose quetiapine with low‐dose olanzapine, providing no specific evidence for the risks associated with quetiapine.
The present analysis has several strengths. Besides the large number of individuals and the possibility of long‐term follow‐up in nationwide registers, we applied several design features to further strengthen our confidence in the findings. First, we adopted a new‐user, active‐comparator design to limit the impact of prior exposure and confounding from mental illness/distress. Second, we attempted to minimize the impact of additional confounding by using a high‐dimensional propensity score drawing on all prescriptions and hospital contacts for the population. Third, we used strict censoring criteria to investigate the risk with low‐dose quetiapine specifically, and adjusted for informative censoring in as‐treated analysis. Lastly, the positive predictive value in the registers is considered high and well‐defined in time 23 .
Limitations with the present study must also be acknowledged. First, no ideal comparator to off‐label, low‐dose quetiapine exists, as other antipsychotics have at least some risk of cardiometabolic adverse events and are not used for the same indication as quetiapine, especially outside psychiatry. Z‐drugs were chosen as the primary comparator, as they are a common alternative to low‐dose quetiapine use as hypnotics 20 . However, Z‐drug users are not an ideal comparator, likely having more comorbidities (especially in long‐term users) than the average low‐dose quetiapine user. As this difference would bias analyses towards less difference between groups, SSRIs, an alternative to low‐dose quetiapine use as an anxiolytic, were included as comparator in sensitivity analyses.
Second, long‐term follow‐up is needed to sufficiently assess cardiovascular safety, but longer follow‐up increases the potential influence of other factors on outcomes. Confining the intent‐to‐treat analysis to individuals with ≥1 additional prescription within 180 days of the first quetiapine prescription, and limiting the maximum follow‐up to 10 years were done to minimize the impact of other factors on the observed risk. However, this analytic choice may have led to underestimation of the real number of cardiovascular events in this cohort attributable to low‐dose quetiapine use. Third, despite adjusting the analyses for 100 potentially relevant confounders, residual confounding cannot be excluded. However, results were generally consistent and robust across multiple approaches. Finally, since this is a non‐randomized database study, associations cannot prove causation.
In conclusion, our findings indicate that low‐dose quetiapine use is associated with increased risk of major adverse cardiovascular events, especially cardiovascular death. The risk increases with continuous treatment and in vulnerable populations, including females and the elderly. On the basis of these findings, we suggest that off‐label low‐dose quetiapine use for sedative or hypnotic purposes should be discouraged.
ACKNOWLEDGEMENTS
This study was supported by the Research Fund of the Mental Health Services in the Region of Southern Denmark (grant no. A2957). The funding source had no role in the design or conduct of the study, data analysis, interpretation, writing of the paper, or the decision to submit the manuscript. The authors want to thank L.C. Lund (University of Southern Denmark) for valuable comments on study design. C.U. Correll and J. Hallas are joint senior authors of the paper. Supplementary information on this study can be found at https://osf.io/658vq/.
REFERENCES
- 1. European Medicines Agency . Summary of product characteristics: Seroquel, Seroquel XR and associated names (quetiapine). https://www.ema.europa.eu.
- 2. Correll CU. From receptor pharmacology to improved outcomes: individualising the selection, dosing, and switching of antipsychotics. Eur Psychiatry 2010;25(Suppl. 2):S12‐21. [DOI] [PubMed] [Google Scholar]
- 3. Alexander GC, Gallagher SA, Mascola A et al. Increasing off‐label use of antipsychotic medications in the United States, 1995‐2008. Pharmacoepidemiol Drug Saf 2011;20:177‐84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Duncan D, Cooke L, Symonds C et al. Quetiapine use in adults in the community: a population‐based study in Alberta, Canada. BMJ Open 2016;6:e010861. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Højlund M, Andersen JH, Andersen K et al. Use of antipsychotics in Denmark 1997‐2018: a nation‐wide drug utilisation study with focus on off‐label use and associated diagnoses. Epidemiol Psychiatr Sci 2021;30:e28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Marston L, Nazareth I, Petersen I et al. Prescribing of antipsychotics in UK primary care: a cohort study. BMJ Open 2014;4:e006135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Hálfdánarson Ó, Zoëga H, Aagaard L et al. International trends in antipsychotic use: a study in 16 countries, 2005‐2014. Eur Neuropsychopharmacol 2017;27:1064‐76. [DOI] [PubMed] [Google Scholar]
- 8. ClinCalc.com. Quetiapine – drug usage statistics. https://clincalc.com.
- 9. Olfson M, King M, Schoenbaum M. Antipsychotic treatment of adults in the United States. J Clin Psychiatry 2015;76:1346‐53. [DOI] [PubMed] [Google Scholar]
- 10. Correll CU, Detraux J, De Lepeleire J et al. Effects of antipsychotics, antidepressants and mood stabilizers on risk for physical diseases in people with schizophrenia, depression and bipolar disorder. World Psychiatry 2015;14:119‐36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Vancampfort D, Stubbs B, Mitchell AJ et al. Risk of metabolic syndrome and its components in people with schizophrenia and related psychotic disorders, bipolar disorder and major depressive disorder: a systematic review and meta‐analysis. World Psychiatry 2015;14:339‐47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Vancampfort D, Correll CU, Galling B et al. Diabetes mellitus in people with schizophrenia, bipolar disorder and major depressive disorder: a systematic review and large scale meta‐analysis. World Psychiatry 2016;15:166‐74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. De Hert M, Detraux J, van Winkel R et al. Metabolic and cardiovascular adverse effects associated with antipsychotic drugs. Nat Rev Endocrinol 2012;8:114‐26. [DOI] [PubMed] [Google Scholar]
- 14. Solmi M, Murru A, Pacchiarotti I et al. Safety, tolerability, and risks associated with first‐ and second‐generation antipsychotics: a state‐of‐the‐art clinical review. Ther Clin Risk Manag 2017;13:757‐77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Huhn M, Nikolakopoulou A, Schneider‐Thoma J et al. Comparative efficacy and tolerability of 32 oral antipsychotics for the acute treatment of adults with multi‐episode schizophrenia: a systematic review and network meta‐analysis. Lancet 2019;394:939‐51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Pillinger T, McCutcheon RA, Vano L et al. Comparative effects of 18 antipsychotics on metabolic function in patients with schizophrenia, predictors of metabolic dysregulation, and association with psychopathology: a systematic review and network meta‐analysis. Lancet Psychiatry 2020;7:64‐77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Grundy SM. Metabolic syndrome: connecting and reconciling cardiovascular and diabetes worlds. J Am Coll Cardiol 2006;47:1093‐100. [DOI] [PubMed] [Google Scholar]
- 18. Correll CU, Joffe BI, Rosen LM et al. Cardiovascular and cerebrovascular risk factors and events associated with second‐generation antipsychotic compared to antidepressant use in a non‐elderly adult sample: results from a claims‐based inception cohort study. World Psychiatry 2015;14:56‐63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Franklin JM, Schneeweiss S. When and how can real world data analyses substitute for randomized controlled trials? Clin Pharmacol Ther 2017;102:924‐33. [DOI] [PubMed] [Google Scholar]
- 20. Scharner V, Hasieber L, Sönnichsen A et al. Efficacy and safety of Z‐substances in the management of insomnia in older adults: a systematic review for the development of recommendations to reduce potentially inappropriate prescribing. BMC Geriatr 2022;22:87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Langan SM, Schmidt SA, Wing K et al. The reporting of studies conducted using observational routinely collected health data statement for pharmacoepidemiology (RECORD‐PE). BMJ 2018;363:k3532. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Pottegård A, Schmidt SAJ, Wallach‐Kildemoes H et al. Data resource profile: the Danish National Prescription Registry. Int J Epidemiol 2017;46:798‐798f. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Schmidt M, Schmidt SAJ, Sandegaard JL et al. The Danish National Patient Registry: a review of content, data quality, and research potential. Clin Epidemiol 2015;7:449‐90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Schmidt M, Pedersen L, Sørensen HT. The Danish Civil Registration System as a tool in epidemiology. Eur J Epidemiol 2014;29:541‐9. [DOI] [PubMed] [Google Scholar]
- 25. Helweg‐Larsen K. The Danish Register of Causes of Death. Scand J Public Health 2011;39(Suppl. 7):26‐9. [DOI] [PubMed] [Google Scholar]
- 26. Schneeweiss S, Rassen JA, Glynn RJ et al. High‐dimensional propensity score adjustment in studies of treatment effects using health care claims data. Epidemiology 2009;20:512‐22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Suissa S. Immortal time bias in pharmaco‐epidemiology. Am J Epidemiol 2008;167:492‐9. [DOI] [PubMed] [Google Scholar]
- 28. Williamson T, Ravani P. Marginal structural models in clinical research: when and how to use them? Nephrol Dial Transplant 2017;32(Suppl. 2):ii84‐90. [DOI] [PubMed] [Google Scholar]
- 29. de Abajo FJ. Effects of selective serotonin reuptake inhibitors on platelet function. Drugs Aging 2011;28:345‐67. [DOI] [PubMed] [Google Scholar]
- 30. Maslej MM, Bolker BM, Russell MJ et al. The mortality and myocardial effects of antidepressants are moderated by preexisting cardiovascular disease: a meta‐analysis. Psychother Psychosom 2017;86:268‐82. [DOI] [PubMed] [Google Scholar]
- 31. Gjerden P, Bramness JG, Tvete IF et al. The antipsychotic agent quetiapine is increasingly not used as such: dispensed prescriptions in Norway 2004‐2015. Eur J Clin Pharmacol 2017;73:1173‐9. [DOI] [PubMed] [Google Scholar]
- 32. Cates ME, Jackson CW, Feldman JM et al. Metabolic consequences of using low‐dose quetiapine for insomnia in psychiatric patients. Community Ment Health J 2009;45:251‐4. [DOI] [PubMed] [Google Scholar]
- 33. Carr CN, Lopchuk S, Beckman ME et al. Evaluation of the use of low‐dose quetiapine and the risk of metabolic consequences: a retrospective review. Ment Health Clin 2016;6:308‐13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Berge J, Abri P, Andell P et al. Associations between off‐label low‐dose olanzapine or quetiapine and cardiometabolic mortality. J Psychiatr Res 2022;149:352‐8. [DOI] [PubMed] [Google Scholar]
- 35. Anderson SL, Vande Griend JP. Quetiapine for insomnia: a review of the literature. Am J Health Syst Pharm 2014;71:394‐402. [DOI] [PubMed] [Google Scholar]
- 36. Modesto‐Lowe V, Harabasz AK, Walker SA. Quetiapine for primary insomnia: consider the risks. Cleve Clin J Med 2021;88:286‐94. [DOI] [PubMed] [Google Scholar]