Abstract
Objective
Quetiapine use at standard doses has been associated with hyperglycemia and dyslipidemia. However, whether even frequently prescribed low‐dose quetiapine results in significant metabolic disturbances remains unclear. Thus, this study aimed to investigate the association between off‐label, low‐dose quetiapine and changes in glycosylated hemoglobin (HbA1c) levels/lipid parameters.
Methods
We identified new users of low‐dose quetiapine (≤50 mg tablets) in Denmark 2008–2018 with measurements of HbA1c, total cholesterol (TC), low‐density lipoprotein cholesterol (LDL‐C), high‐density lipoprotein cholesterol (HDL‐C), or fasting triglycerides (fTG) within 365 days before and after quetiapine initiation. Mixed‐effects linear regression models were used to estimate coefficients (β) with 95% confidence intervals (95%CIs) for change in cardiometabolic parameters after quetiapine initiation. Inverse probability weighting was used to mitigate selection bias. Higher doses of quetiapine (>50 mg) were included in sensitivity analyses.
Results
Among 106,711 eligible new low‐dose quetiapine users (median age = 45 years, females = 55%), low‐dose quetiapine initiation was associated with increased fTG (β = 1.049[95%CI:1.027–1.072]) and decreased HDL‐C (β = 0.982[0.978–0.986]). Although HbA1c did not change significantly and TC and LDL‐C even decreased considering all subjects, all three metabolic parameters increased significantly among individuals with normal pre‐quetiapine initiation levels. The adverse metabolic effect of quetiapine on HbA1c, TC, LDL‐C, and HDL‐C was dose‐dependent, which was not the case for fTG.
Conclusions
Low‐dose quetiapine was associated with a significant increase in fTG and decreases in HDL‐C in all subjects, as well as with significant increases in HbA1c, TC, and LDL‐C among those with normal baseline values. The risk of metabolic worsening with quetiapine was dose‐dependent, except for fTG.
Keywords: Antipsychotics, dyslipidemia, hyperglycemia, off‐label, quetiapine
Significant Outcomes
Overall, low‐dose quetiapine use (25–50 mg tablets) was only associated with increases in fasting triglyceride levels and decreases in HDL‐cholesterol.
Among individuals with normal HbA1c/lipid‐levels before quetiapine initiation, low‐dose quetiapine use was associated with increases in HbA1c, total cholesterol, and LDL‐cholesterol levels.
The risk of metabolic worsening with off‐label use of quetiapine seems to be dose‐dependent, especially for total cholesterol and LDL‐cholesterol levels.
Limitations
Blood test results before and after initiation of quetiapine were only available for ≤10% of the total population of low‐dose quetiapine users.
The probability of having measurements after initiation of quetiapine was higher with increasing pre‐initiation levels. Thus, selection bias might have attenuated the true effect of low‐dose quetiapine on risk factors for cardiometabolic disease.
1. INTRODUCTION
Quetiapine is the second‐generation antipsychotic with regulatory approval for treatment of schizophrenia, bipolar disorder, and major depression, and is commonly used off‐label in low doses for anxiolytic or hypnotic purposes. 1 , 2
Quetiapine use has been linked to increased levels of low‐density lipoprotein‐cholesterol (LDL‐C) and decreased high‐density lipoprotein‐cholesterol (HDL‐C) based on randomized controlled trial (RCT) data in schizophrenia 3 and various off‐label indications, for example, substance abuse disorders, generalized anxiety disorder (GAD), and obsessive–compulsive disorder. 4 However, patients included in RCTs receive much higher doses of quetiapine (150–800 mg/day) 3 , 4 than typically used off‐label for anxiety or insomnia. Furthermore, the use of quetiapine might increase blood glucose levels, although this effect remains unclear, as relevant RCTs are generally short and blood glucose changes take more time. 3
Observational studies found that both cholesterol and blood glucose levels increase after quetiapine initiation, 5 and that cholesterol‐lowering and antidiabetic medications are started more frequently after initiation of the second‐generation antipsychotics, including quetiapine. 6
Hyperglycemia and dyslipidemia that have been associated with quetiapine use at higher doses 3 , 7 are important risk factors for cardiovascular disease. 7 Recently, low‐dose quetiapine use has been associated with an increased risk of major adverse cardiovascular events, defined as nonfatal myocardial infarction, nonfatal ischemic stroke, or death from any cardiovascular causes. 8 However, it is currently not known whether low doses of quetiapine are also associated with an increased risk of hyperglycemia or dyslipidemia. As quetiapine is frequently used off‐label in low doses, 2 , 9 , 10 , 11 , 12 this question is of particular public health interest and concern.
1.1. Aims of the study
Our aim was to investigate whether low‐dose use of quetiapine is associated with alterations in glycosylated hemoglobin A1c (HbA1c), triglyceride, and cholesterol levels, combining prescription data with blood test results from nation‐wide health‐care registers.
2. MATERIALS AND METHODS
We conducted a nation‐wide cohort study to investigate the association between initiation of low‐dose treatment with quetiapine and the subsequent adverse changes in HbA1c, fasting triglycerides (fTG), total cholesterol (TC), LDL‐C, and HDL‐C. The use of pseudonymized data for this study were approved by the University of Southern Denmark (ref 10.350) and the Danish Health Data Authority (FSEID‐00004279). According to Danish legislation, no ethical approval or informed consent is necessary for a purely register‐based study as the present one. The study was reported following the recommendations in the REporting of studies Conducted using Observational Routinely collected health Data adaptation for non‐interventional PharmacoEpidemiological research (RECORD‐PE) guideline (checklist in Appendix 1). 13
2.1. Study population and data sources
The index populations were all new users of quetiapine in Denmark between 1 January 2008 and 31 December 2018. This population was identified using the Danish National Prescription Register (DNPrR), which holds information on all prescriptions filled at community pharmacies in Denmark since 1995. 14 The date of each individual's first prescription for quetiapine was defined as the index date. New users were defined as individuals without any prescriptions for quetiapine between its approval in Denmark in 2001 and the index date. This population was then linked to the Danish National Patient Register (DNPaR) 15 and the Danish Civil Registration System (DCRS); 16 using civil registration numbers to obtain information on psychiatric diagnoses, migrations, and vital status. DNPaR contains information on all in‐ and outpatient contacts to Danish hospitals since 1977 and 1995, respectively. 15 DCRS contains information on vital status and dates of birth, death, immigration, and emigration.
From the population of new users of quetiapine, we excluded individuals:
with prescriptions for quetiapine in tablet strengths >50 mg on the index date, to focus on low‐dose use, specifically;
with prescriptions for other antipsychotics on the index date or in the 365 days preceding the index date, to isolate the potential effect of quetiapine from that of other antipsychotics;
with a history of schizophrenia, schizoaffective disorder, mania, or bipolar disorder, to focus on low‐dose off‐label use; and
with less than 365 days of register coverage prior to the index date, to ensure adequate exposure and covariate assessment.
As enforced by exclusion criterion (i), the study population consisted of individuals who filled prescriptions for 25 or 50 mg tablets of quetiapine. The exact daily dose prescribed or ingested could not be derived from our data sources, thus the tablet strength served as a proxy for the use of low daily doses of quetiapine. See Appendix 2 for codes used for exposure assessment and exclusion criteria.
2.1.1. Biochemical analysis results
For each eligible low‐dose quetiapine user, we created an observation period from 365 days before to 365 days after the index date. The observation period ended before 365 days had elapsed if the patient met one of the exclusion criteria.
We then used the nation‐wide Register of Laboratory Results for Research (RLRR) 17 , 18 to identify low‐dose quetiapine users who had >1 measurements within the observation period for ≥1 of the following laboratory test results: HbA1c, fTG, TC, LDL‐C, or HDL‐C.
The RLRR contains information on all blood samples analyzed at public hospital laboratories in Denmark from 2008 and onwards. 18 In Denmark, routine analyses from hospitals are almost exclusively analyzed at these laboratories. Furthermore, the majority of blood samples taken by general practitioners are also analyzed at these hospital laboratories, and thus included in this register. However, blood samples analyzed using certain point‐of‐care devices (e.g., point‐of‐care analyses of HbA1c in general practice) are not included in RLRR. HbA1c was chosen over plasma glucose, being independent of fasting status. Fasting measurements of triglycerides were chosen, as triglycerides are more susceptible to post‐prandial variation than cholesterols (fTG measurements were identified using specific NPU‐codes [Appendix 2]). 19
2.2. Covariates
For each new user of low‐dose quetiapine, we recorded sex and age at initiation. Using the DNPrR, we also assessed whether these individuals had filled prescriptions for antidiabetic drugs, lipid‐lowering drugs, or psychotropic drugs other than quetiapine within 365 days before the index date. Lastly, we used the DNPaR to assess whether these new quetiapine users had records of in‐ or outpatient hospital diagnoses for diabetes within 5 years prior to the index date. See Appendix 2 for codes used in the covariate definition.
2.3. Statistical analysis
We examined the association between initiation of low‐dose quetiapine and changes in HbA1c, triglyceride, and cholesterol levels using linear mixed‐effects regression models. Analyses were conducted using the Stata MP (release 17.0).
Models included an indicator variable of measurements being pre/post the index date as a fixed effect and a random effect for each individual to account for repeated observations of the same individual. Months relative to the index date was used as the underlying time scale. A graphical depiction of the analytical design is provided in Appendix 3.
Exploratory analyses confirmed our expectation of log‐transformed outcomes having normally distributed residuals. We therefore conducted the analysis based on this transformation (Appendix 4). This analysis method implies that the effect associated with the pre/post indicator variable can be interpreted as the relative change in the median outcome after initiation of quetiapine. Thus, coefficients larger than one imply that quetiapine initiation was associated with an increase in the target outcome.
Furthermore, exploratory analyses found that the probability of having measurements after the index date (and thus being included in the study) increased with higher levels prior to the index date (Appendix 5). This finding implies that persons with high pre‐initiation values were more likely to be included in the study. By preferentially including participants on the basis of high pre‐initiation values, estimates may become biased due to a regression‐toward‐the‐mean effect, which would confer a general liability to observe decreasing levels or a liability to mask a genuine upward trend. Therefore, individuals' observations were weighted by their inverse probability of having post‐initiation measurements, based on their average values of the outcome before the index date, in order to mitigate this potential bias of selective measurements. The inverse probability‐weights (IPWs) were estimated for each individual using logistic regression models incorporating sex, age at initiation, calendar year, and average levels of the respective metabolic outcome before the index date.
Lastly, analyses were based on all available measurements from 365 days before the index date to the index date, and all measurements beyond day 7 (day 30 for analyses of HbA1c) to the end of the observation period. Measurements between the index date and day 7(/30) were disregarded as the effect of quetiapine on outcomes likely requires some time to be evident. Furthermore, for obvious reasons, individuals without measurements beyond day 7/30 could not be included in the analyses.
2.3.1. Subgroup analyses
To assess the potential effect modification with certain patient characteristics, we conducted subgroup analyses stratified by:
sex,
age at initiation (0–17/18–39/40–64/≥65 years),
recent use of antidiabetic drugs (in analyses of HbA1c),
recent use of lipid‐lowering drugs (in analyses of triglyceride/cholesterol), and
average level of the parameter in the year before initiation of quetiapine (see Appendix 2 for cut‐offs used for each outcome).
Effect modification via these characteristics was tested using likelihood‐ratio tests comparing models with and without an interaction term including the respective characteristic.
2.3.2. Supplementary analyses
We conducted two supplementary analyses to test the impact of various factors on the risk of metabolic alterations following the initiation of quetiapine: First, we excluded individuals with any use of antidiabetic or lipid‐lowering drugs within the observation period (in analyses of HbA1c and cholesterol/triglyceride, respectively) to assess if the use of such medications would mask a potential association between use of low‐dose quetiapine and the outcomes of interest. Second, we changed the exposure definition to higher tablet strengths of quetiapine (first prescription for quetiapine on 100 or >100 mg tablets compared to ≤50 mg tablets) to assess if the association between quetiapine initiation and the outcomes was dependent on the assumed daily dose of quetiapine. Analyses were conducted using a mixed effect linear regression model with IPWs, similar to that in the main analysis, but stratified on the starting tablet strength (initiation of ≤50, 100 or >100 mg tablets). A potential dose–response trend was tested for each outcome using the starting dose category as an independent variable in the regression model.
3. RESULTS
We identified 106,711 eligible new users of quetiapine in the DNPR between January 1, 2008 and December 31, 2018. Among these new quetiapine users, the proportion of individuals with biochemical glucose or lipid metabolism measurements before and after their first prescription for low‐dose quetiapine ranged from 1.2% (n = 1300) for fTG to 9.3% (n = 9905) for total cholesterol (Appendix 6). The median number of quetiapine prescriptions in the analytical population was 2 (interquartile range:1–4, range: 1–47). Other characteristics of all eligible low‐dose quetiapine users and those included in the analyses are shown in Table 1. Those individuals that had measurements for the outcomes differed somewhat from the overall population of low‐dose quetiapine initiators: they had a higher median age (55–65 versus 45 years), and were more likely to have a diagnosis of diabetes and use antidiabetic/lipid‐lowering medications. The distribution of psychiatric diagnoses and use of other psychotropic medications among individuals with measurements of the targeted outcomes were largely similar to that of the overall population of low‐dose quetiapine users (Table 1).
TABLE 1.
Characteristics of eligible low‐dose quetiapine‐users with pre−/post measurements of outcomes
| All users | HbA1c | Total cholesterol | Triglycerides a | LDL‐C | HDL‐C | |
|---|---|---|---|---|---|---|
| Characteristic | (N = 106,711) | (N = 9420) | (N = 9905) | (N = 1300) | (N = 9220) | (N = 9524) |
| Sex, N (%) | ||||||
| Female | 58,760 (55) | 5455 (58) | 5540 (56) | 712 (55) | 5200 (56) | 5327 (56) |
| Median age, years (IQR) | 45 (29–63) | 56 (41–72) | 55 (41–70) | 55 (41–67) | 55 (41–71) | 55 (41–70) |
| Age groups, N (%) | ||||||
| 0–17 years | 3200 (3) | 96 (1) | 148 (1) | 17 (1) | 144 (2) | 146 (2) |
| 18–39 years | 39,890 (37) | 2047 (22) | 2135 (22) | 287 (22) | 1962 (21) | 2022 (21) |
| 40–64 years | 37,992 (36) | 3866 (41) | 4275 (43) | 608 (47) | 3978 (43) | 4152 (44) |
| ≥ 65 years | 25,629 (24) | 3411 (36) | 3347 (34) | 388 (30) | 3136 (34) | 3204 (34) |
| Year of cohort entry, N (%) | ||||||
| 2008–2010 | 18,277 (17) | 45 (<1) | 148 (1) | 50 (4) | 113 (1) | 126 (1) |
| 2011–2014 | 37,529 (35) | 1304 (14) | 1545 (16) | 382 (29) | 1316 (14) | 1406 (15) |
| 2015–2018 | 50,905 (48) | 8071 (86) | 8212 (83) | 868 (67) | 7791 (85) | 7992 (84) |
| History of mental disorders, N (%) | ||||||
| Dementia | 8957 (8) | 893 (9) | 798 (8) | 57 (4) | 731 (8) | 746 (8) |
| Major depression | 24,643 (23) | 2119 (22) | 2393 (24) | 352 (27) | 2213 (24) | 2311 (24) |
| Anxiety disorders | 28,142 (26) | 2660 (28) | 2909 (29) | 350 (27) | 2728 (30) | 2823 (30) |
| Personality disorders | 8371 (8) | 632 (7) | 671 (7) | 78 (6) | 630 (7) | 655 (7) |
| Other mental disorders | 27,853 (26) | 2349 (25) | 2544 (26) | 333 (26) | 2373 (26) | 2463 (26) |
| No history of mental disorders | 45,142 (42) | 3939 (42) | 4088 (41) | 567 (44) | 3794 (41) | 3910 (41) |
| History of somatic disorders, N (%) | ||||||
| Diabetes | 6847 (6) | 1846 (20) | 1572 (16) | 227 (17) | 1445 (16) | 1535 (16) |
| Alcohol‐related disorders | 11,590 (11) | 1100 (12) | 1201 (12) | 161 (12) | 1123 (12) | 1170 (12) |
| Recent use of psychotropic medications, N (%) | ||||||
| Antidepressants | 65,125 (61) | 5884 (62) | 6344 (64) | 933 (72) | 5890 (64) | 6097 (64) |
| Lithium | 792 (<1) | 55 (<1) | 78 (<1) | 10 (<1) | 75 (<1) | 75 (<1) |
| Anxiolytics | 23,182 (22) | 2398 (25) | 2636 (27) | 373 (29) | 2442 (26) | 2531 (27) |
| Hypnotics | 29,764 (28) | 3124 (33) | 3397 (34) | 429 (33) | 3160 (34) | 3271 (34) |
| Psychostimulants | 5637 (5) | 305 (3) | 356 (4) | 53 (4) | 317 (3) | 326 (3) |
| Recent use of somatic medications, N (%) | ||||||
| Antidiabetic medications | 5996 (6) | 1683 (18) | 1429 (14) | 210 (16) | 1316 (14) | 1397 (15) |
| Lipid‐lowering medications | 15,018 (14) | 2833 (30) | 3116 (31) | 488 (38) | 2904 (31) | 3035 (32) |
Abbreviations: HbA1c, glycosylated hemoglobin A1c; HDL‐C, high‐density lipoprotein cholesterol; IQR, interquartile range; LDL‐C, low‐density lipoprotein cholesterol; N, number.
Fasting measurements only.
3.1. Changes in HbA1c
Overall, low‐dose quetiapine initiation was not associated with significant HbA1c level increases (n = 9420; β = 0.999; and 95%CI = 0.997–1.002; Table 2). However, significant HbA1c increases were observed among those individuals with normal HbA1c before low‐dose quetiapine initiation, compared to those with prediabetic or diabetic HbA1c levels before initiation (β = 1.006 vs. β = 0.999/0.971; p < 0.001; Table 3). An estimated median ratio of 1.006 should be interpreted as a 0.6% higher median HbA1c level among patients after quetiapine initiation. Significant HbA1c increases after low‐dose quetiapine initiation were also observed among individuals aged 40–64 years at the index date compared to those <40/≥65 years at low‐dose quetiapine initiation (β = 1.006 vs. β = 0.983–0.998; p < 0.001; Table 3). Similarly, changes in HbA1c were higher among individuals without versus with recent antidiabetic medication use (p < 0.001; Table 3). Lastly, the effect of low‐dose quetiapine on HbA1c did not differ between men and women (p = 0.45; Table 3).
TABLE 2.
Change in metabolic parameters after initiation of treatment with low‐dose quetiapine
| Outcome | Individuals/samples, N | Coefficient (95%CI) | 95%PI for baseline value |
|---|---|---|---|
| HbA1c | 9420/33740 | 0.999 (0.997 to 1.002) | 29–83 |
| Total cholesterol | 9905/33678 | 0.993 (0.989 to 0.996) | 3.0–7.2 |
| Triglycerides | 1300/4195 | 1.049 (1.027 to 1.072) | 0.6–7.0 |
| LDL‐cholesterol | 9220/30940 | 0.984 (0.977 to 0.990) | 1.1–4.8 |
| HDL‐cholesterol | 9524/32111 | 0.982 (0.978 to 0.986) | 0.8–2.6 |
Abbreviations: CI, confidence interval; HbA1c, glycosylated hemoglobin A1c; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein; N, number; PI, prediction interval depicting the 2.5% and 97.5% percentile of average levels before the index date.
TABLE 3.
Subgroup analyses of change in metabolic parameters after initiation of treatment with low‐dose quetiapine
| Subgroup | Number individuals/samples | Coefficient (95%CI) | 95%PI baseline values | Test for interaction a p‐value |
|---|---|---|---|---|
| HbA1c | ||||
| Sex | ||||
| Female | 5455/18925 | 0.998 (0.995 to 1.001) | 29–81 | |
| Male | 3965/14815 | 1.001 (0.997 to 1.006) | 29–86 | 0.45 |
| Age group | ||||
| 0–17 years | 96/311 | 0.983 (0.944 to 1.024) | 19–85 | |
| 18–39 years | 2047/6236 | 0.998 (0.993 to 1.002) | 28–87 | |
| 40–64 years | 3866/13861 | 1.006 (1.002 to 1.010) | 29–85 | |
| ≥ 65 years | 3411/13332 | 0.993 (0.988 to 0.998) | 31–79 | <0.001 |
| Mean HbA1c before baseline b | ||||
| Normal | 6971/21813 | 1.006 (1.003 to 1.008) | 28–41 | |
| Prediabetic | 1127/4562 | 0.999 (0.991 to 1.007) | 42–48 | |
| Diabetic | 1322/7365 | 0.971 (0.960 to 0.981) | 48–105 | <0.001 |
| Recent use of antidiabetic medications c | ||||
| Yes | 1683/9325 | 0.992 (0.983 to 1.002) | 36–101 | |
| No | 7737/24415 | 1.001 (0.999 to 1.004) | 28–48 | <0.001 |
| Total cholesterol | ||||
| Sex | ||||
| Female | 5540/18591 | 0.998 (0.993 to 1.002) | 3–7 | |
| Male | 4365/15087 | 0.986 (0.981 to 0.992) | 3–7 | <0.001 |
| Age group | ||||
| 0–17 years | 148/430 | 1.020 (0.994 to 1.046) | 3–6 | |
| 18–39 years | 2135/6381 | 1.007 (1.000 to 1.014) | 3–7 | |
| 40–64 years | 4275/14724 | 0.994 (0.988 to 0.999) | 3–8 | |
| ≥ 65 years | 3347/12143 | 0.979 (0.973 to 0.986) | 3–7 | <0.001 |
| Increased levels before initiation | ||||
| Yes | 4164/13977 | 0.944 (0.939 to 0.949) | 3–5 | |
| No | 5741/19701 | 1.026 (1.022 to 1.031) | 5–8 | <0.001 |
| Recent use of lipid‐lowering drugs c | ||||
| Yes | 3116/12500 | 0.981 (0.973 to 0.989) | 3–7 | |
| No | 6789/21178 | 0.999 (0.995 to 1.002) | 3–7 | <0.001 |
| Fasting triglycerides | ||||
| Sex | ||||
| Female | 712/2246 | 1.043 (1.015 to 1.071) | 1–4 | |
| Male | 588/1949 | 1.057 (1.019 to 1.095) | 1–14 | 0.75 |
| Age group | ||||
| 0–17 years | 17/54 | 0.953 (0.810 to 1.120) | 1–3 | |
| 18–39 years | 287/847 | 1.066 (1.008 to 1.127) | 1–22 | |
| 40–64 years | 608/1978 | 1.054 (1.022 to 1.087) | 1–7 | |
| ≥ 65 years | 388/1316 | 1.033 (0.999 to 1.067) | 1–4 | 0.72 |
| Increased levels before initiation | ||||
| Yes | 327/1124 | 0.861 (0.826 to 0.898) | 2–22 | |
| No | 973/3071 | 1.115 (1.088 to 1.143) | 1–2 | <0.001 |
| Recent use of lipid‐lowering drugs c | ||||
| Yes | 488/1776 | 1.011 (0.978 to 1.045) | 1–14 | |
| No | 812/2419 | 1.074 (1.043 to 1.106) | 1–5 | 0.0021 |
| LDL‐cholesterol | ||||
| Sex | ||||
| Female | 5200/17240 | 0.989 (0.981 to 0.997) | 1–5 | |
| Male | 4020/13700 | 0.977 (0.967 to 0.987) | 1–5 | 0.042 |
| Age group | ||||
| 0–17 years | 144/416 | 0.990 (0.941 to 1.042) | 1–4 | |
| 18–39 years | 1962/5755 | 0.999 (0.988 to 1.011) | 1–5 | |
| 40–64 years | 3978/13479 | 0.982 (0.973 to 0.992) | 1 to 5 | |
| ≥ 65 years | 3136/11290 | 0.974 (0.963 to 0.986) | 1–5 | 0.047 |
| Increased levels before initiation | ||||
| Yes | 3314/10765 | 0.902 (0.894 to 0.911) | 3–5 | |
| No | 5906/20175 | 1.030 (1.022 to 1.038) | 1–3 | <0.001 |
| Recent use of lipid‐lowering drugs c | ||||
| Yes | 2904/11436 | 0.966 (0.953 to 0.979) | 1–5 | |
| No | 6316/19504 | 0.992 (0.986 to 0.999) | 1–5 | <0.001 |
| HDL‐cholesterol | ||||
| Sex | ||||
| Female | 5327/17692 | 0.991 (0.986 to 0.996) | 1–3 | |
| Male | 4197/14419 | 0.972 (0.965 to 0.978) | 1–2 | <0.001 |
| Age group | ||||
| 0–17 years | 146/423 | 1.008 (0.978 to 1.039) | 1–2 | |
| 18–39 years | 2022/5952 | 0.995 (0.986 to 1.003) | 1–2 | |
| 40–64 years | 4152/14194 | 0.990 (0.984 to 0.997) | 1–3 | |
| ≥ 65 years | 3204/11542 | 0.960 (0.953 to 0.968) | 1–3 | <0.001 |
| Decreased levels before initiation | ||||
| Yes | 1407/5126 | 1.076 (1.065 to 1.087) | 1–1 | |
| No | 8117/26985 | 0.965 (0.961 to 0.969) | 1–3 | <0.001 |
| Recent use of lipid‐lowering drugs c | ||||
| Yes | 3035/12058 | 0.984 (0.977 to 0.991) | 1–2 | |
| No | 6489/20053 | 0.981 (0.976 to 0.986) | 1–3 | 0.34 |
Abbreviations: CI, confidence interval; HbA1c, glycosylated hemoglobin A1c; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein; N, number; PI, prediction interval (depicting the 2.5% and 97.5% percentile of average levels before the index date.).
Likelihood‐ratio test.
Individuals were classified as having normal (<42 mmol/mol), prediabetic (42‐47 mmol/mol), or diabetic (≥48 mmol/mol) HbA1c‐levels, based on their average HbA1c‐level within a year before the index date.
Recent use of antidiabetic/lipid‐lowering medications was defined as having prescriptions fills for such medications (ATC: A10/C10) within 1 year of the index date. See Appendix 1 for relevant codes.
3.2. Changes in triglyceride and cholesterol
Overall, initiation of low‐dose quetiapine was associated with decreased levels of both TC (n = 9905; β = 0.993; 95%CI = 0.989–0.996), LDL‐C (n = 9220; β = 0.984; and 95%CI = 0.977–0.990), and HDL‐C (n = 9524; β = 0.982; and 95%CI = 0.978–0.986); as well as increased levels of fTG (n = 1300; β = 1.049; and 95%CI = 1.027–1.072; Table 3). However, the level of TC/LDL‐C increased significantly after initiation of low‐dose quetiapine among those with normal levels prior to initiation of low‐dose quetiapine (Table 3). Sex, age at initiation, and the recent use of lipid‐lowering drugs modified the risk of in−/decreases in lipid levels after low‐dose quetiapine initiation. However, no single characteristic was associated with significant in‐ or decreases in lipid levels, except that triglycerides increased significantly among those individuals without recent use of lipid‐lowering medications compared to those who had filled prescriptions for this class of medications (Table 3).
3.3. Supplementary analyses
Excluding individuals with any use of lipid‐lowering medications during the observation period found that such medications likely attenuated the impact of low‐dose quetiapine on TC, LDL‐C, and fTG (Appendix 7). A similar effect was not seen for the use of antidiabetic medications on HbA1c, or for lipid‐lowering medication use on HDL‐C.
The impact of quetiapine initiation on HbA1c increased with the quetiapine dose, although numbers of individuals available for the analyses were limited and all confidence intervals overlapped (p = 0.031; Figure 1). Similarly, higher quetiapine doses were associated with greater impact on TC, LDL‐C, and HDL‐C (p = 0.0071, p < 0.001, and p = 0.004, respectively; Figure 1). Analyses indicated that higher quetiapine doses were associated with numerically larger increases in fTG, but the number of individuals exposed to 100/>100 mg tablets who had measurements of fTG were limited, resulting in wide CIs (p = 0.41; Figure 1).
FIGURE 1.
Change in metabolic parameters after initiation of treatment with quetiapine in various tablet strengths. Abbreviations: CI: confidence interval; Coef: Coefficient; HbA1c: Glycosylated hemoglobin A1c; HDL: High‐density lipoprotein; LDL Low‐density lipoprotein; N, number. †Tablet strength on the index date. ‡Mixed‐effects linear regression adjusted using inverse probability weights estimated from the probability of having measurements of the outcome after initiation of quetiapine based on age, sex, calendar year, and the average level of the outcome in the year before the index date. §P‐value from mixed‐effects linear regression model including tablet strength as continuous variable.
4. DISCUSSION
In this nation‐wide, register‐based cohort study of low‐dose quetiapine initiators in Denmark who were prescribed 25–50 mg tablets that were likely prescribed off‐label for insomnia or anxiety, we found that the use of low‐dose quetiapine was associated with significant increases in triglycerides and with decreases in HDL‐C levels among all subjects. Furthermore, we found that the use of low‐dose quetiapine was also associated with statistically significant increases in HbA1c, TC, and LDL‐C among individuals with normal levels of these metabolic parameters before low‐dose quetiapine initiation.
Our findings of increased triglycerides and decreased HDL‐C are consistent with meta‐analyses of RCTs involving quetiapine. 3 , 4 However, these meta‐analyses included either individuals with schizophrenia or GAD, who generally received higher doses of quetiapine than typically used off‐label (i.e., 400‐800 mg/day in RCTs in schizophrenia and 50–300 mg/day in RCTs in GAD). 3 , 4 The association between low‐dose use of quetiapine and adverse metabolic changes has also been explored in several smaller cohort studies where quetiapine use was associated with increases in fasting blood glucose, 20 LDL‐C, 5 TC, 5 and triglycerides. 20 However, the generalizability of these observational studies to the average off‐label use of quetiapine has been limited by the fact that these studies only included individuals with psychiatric diagnoses and that the mean doses herein were higher than what is typically used off‐label for symptoms of anxiety or insomnia. Thus, the present study adds to the evidence of safety concerns with off‐label use of quetiapine by specifically assessing the safety with use of low doses (≤50 mg tablets).
The finding of significant increases in all metabolic parameters among individuals with non‐abnormal levels prior to quetiapine initiation suggests that low doses of quetiapine are associated with metabolic disturbances in individuals with non‐abnormal levels, whereas the impact of such low doses appears to be more limited to TG and HDL‐C when levels are already disturbed, which creates a possible ceiling and/or regression to the mean effect. We found the relative increase in metabolic parameters in individuals with non‐abnormal levels to range from 0.6% for HbA1c to 11.5% for fTG within the one‐year observation period. Such changes might not represent clinically relevant changes in most patients. However, for some patients, the use of low‐dose quetiapine might result in a shift toward abnormal levels (e.g., prediabetes), especially with long‐term treatment or in patients with additional risk factors.
Lastly, the analyses of change in metabolic parameters with higher doses of quetiapine than 50 mg suggest that these adverse metabolic changes are likely connected to the dose of quetiapine. Changes were more pronounced with increasing doses above 50 mg for all metabolic outcomes, except for HbA1c, where dose‐proportionality was modest. This finding is consistent with a recent cohort study from Switzerland where quetiapine doses ≥150 mg/day were associated with larger changes in triglycerides, TC, LDL‐C, and HDL‐C than with doses <150 mg/day. 5
Two secondary, but important findings from this study are that (i) <10% of the eligible off‐label, low‐dose quetiapine users in the study period had records of assessed HbA1c and/or lipid parameters in the RLRR, and (ii) that the probability of having such measurements after initiation of off‐label, low‐dose quetiapine treatment depended on the pre‐initiation level, that is, that prescribers were more likely to monitor those that have already increased HbA1c, triglyceride, or cholesterol levels.
4.1. Strengths and limitations
The present study has a number of strengths and limitations that should be acknowledged when interpreting the results.
A major strength is the high number of individuals included in the study, especially, in comparison with previous observational studies of the association between low‐dose/off‐label use of quetiapine and the risk of metabolic alterations. 5 , 20 Furthermore, the study was based on routinely collected biochemical analysis results from the entire population of eligible off‐label, low‐dose quetiapine users in Denmark, increasing the generalizability of the results. Lastly, the data analytic strategy incorporated inverse probability weights to attenuate the impact of selective testing among low‐dose quetiapine users.
Limitations of the present study include first that only a minor proportion of low‐dose quetiapine users had measurements recorded in the RLRR, which creates a risk of ascertainment bias. However, we believe that the influence of this bias is limited, as the clinical characteristics of the included low‐dose quetiapine‐users did not differ substantially from those of the overall cohort of eligible low‐dose quetiapine‐users. Second, we were not able adjust for the influence of other potentially relevant factors (e.g., smoking, alcohol consumption, and exercise) on alterations in metabolic parameters during follow‐up, as these parameters are not contained in the registers available for analysis. Third, previous research on the use of low‐dose quetiapine in Denmark found that a large proportion of low‐dose quetiapine users will only fill one prescription for quetiapine. 21 This fact means that some individuals will not be exposed during the entire observation period. However, this limitation would only bias the results toward the null hypotheses, representing a conservative bias that only strengthens the results. Fourth, the assessment of low‐dose quetiapine use was based on tablet strength, which is a crude proxy for the dose used. However, exclusion/censoring criteria were included based on filling higher tablet strengths to avoid individuals using substantially higher doses of quetiapine biasing the results. Fifth, measurements for HbA1c (or other measurements for blood glucose) might be conducted during visits with the general practitioner, using point‐of‐care devices, which means that they are not recorded in the RLRR and were thus not available for analyses. Sixth, we did not have data to assess the degree of durability of the adverse metabolic effects of low‐dose quetiapine after its discontinuation. Future studies should investigate this relevant aspect and identify high‐risk subgroups in which the cardiometabolic burden persists beyond quetiapine use, as well as whether duration of quetiapine use is associated with reversibility/durability of the adverse metabolic effects. Finally, we did not include a control group in our analyses. As we were interested in the effect of low‐dose quetiapine and used a within‐subject design, a control group would be of limited utility. Theoretically, however, the observed trends in biochemical markers of cardiovascular risks could be a spontaneous drift by age or by calendar year. With the time frame used in our study, such drift is likely to be quite small. Moreover, we used higher dose quetiapine as a quasi‐control and found that the adverse metabolic effects only strengthened, supporting the validity and relevance of the findings.
4.2. Implications for low‐dose, off‐label use of quetiapine
Continuous monitoring of risk factors for cardiometabolic disease, including blood glucose and lipid parameters, has become standard‐of‐care when treating severe mental illness as schizophrenia and bipolar disorder. 22 Monitoring of risk factors when prescribing antipsychotics to individuals with psychotic disorders is mandatory in Denmark. 23 As our study finds that even the use of low doses of quetiapine is associated with worsening in several important risk factors for cardiovascular disease, especially in individuals with “normal” levels prior to initiation, we suggest that metabolic monitoring should be considered regardless of indication, dose, or duration of treatment with quetiapine. Such monitoring would ensure that alternatives to quetiapine could be sought, or interventions to reduce the quetiapine's adverse impact on metabolic parameters could be initiated. 24 However, further studies on the development of metabolic alterations should investigate the relation with other risk factors and hereby identify certain groups at high risk of metabolic alterations where metabolic monitoring might be considered.
Efforts to ensure appropriate monitoring of risk factors for cardiometabolic disease should be directed toward all types of health‐care providers, given the increasing number of users and prescriptions for quetiapine issued in general practice and other medical specialities. 11 , 25 , 26 While, this study is concerned with quetiapine, similar recommendations should likely be made for other antipsychotics when used off‐label for anxiolytic or hypnotic purposes (e.g. chlorprothixene 27 ), as the antihistaminergic properties responsible the anxiolytic and hypnotic effects of antipsychotics, used in low doses, are likely to be involved in the pathophysiology of metabolic disturbances. 7 , 28 , 29
In conclusion, this nation‐wide cohort study indicates that even low doses of quetiapine can adversely affect blood levels of glucose as measured with HbA1c, triglycerides, and TC, LDL‐C as well as HDL‐C, all of which are important risk factors for the long‐term risk of cardiovascular disease. Therefore, the need for and alternatives of off‐label‐prescribed low‐dose quetiapine should be carefully evaluated, and appropriate monitoring of cardiovascular disease risk factors should be implemented when quetiapine is prescribed, regardless of indication, dosage, or duration of treatment.
AUTHOR CONTRIBUTIONS
All authors participated in the design of the study. Mikkel Højlund conducted the statistical analyses. Mikkel Højlund and Jesper Hallas wrote the initial draft. All authors participated in the interpretation of data and critical revision of the manuscript. All authors approved the final draft and the decision to submit for publication.
CONFLICT OF INTEREST
Mikkel Højlund has received honoraria for consultancy/lecturing from the Lundbeck Foundation and Otsuka Pharmaceuticals Inc. without relation to the present work. Henrik Støvring reports participation in research funded by LEO Pharma A/S (no personal fees). He has received personal consulting fees from Bristol‐Myers‐Squibb, Novartis, Roche, Merck and Pfizer outside the submitted work. He has received teaching fees from Atrium. Kjeld Andersen declares no conflict of interest. CUC has been a consultant and/or advisor to or has received honoraria from: AbbVie, Acadia, Alkermes, Allergan, Angelini, Aristo, Boehringer‐Ingelheim, Cardio Diagnostics, Cerevel, CNX Therapeutics, Compass Pathways, Darnitsa, Gedeon Richter, Hikma, Holmusk, IntraCellular Therapies, Janssen/J&J, Karuna, LB Pharma, Lundbeck, MedAvante‐ProPhase, MedInCell, Merck, Mindpax, Mitsubishi Tanabe Pharma, Mylan, Neurocrine, Newron, Noven, Otsuka, Pharmabrain, PPD Biotech, Recordati, Relmada, Reviva, Rovi, Seqirus, SK Life Science, Sunovion, Sun Pharma, Supernus, Takeda, Teva, and Viatris. He provided expert testimony for Janssen and Otsuka. He served on a Data Safety Monitoring Board for Lundbeck, Relmada, Reviva, Rovi, Supernus, and Teva. He has received grant support from Janssen and Takeda. He received royalties from UpToDate and is also a stock option holder of Cardio Diagnostics, Mindpax, LB Pharma and Quantic. Jesper Hallas reports participation in research projects funded by Alcon, Almirall, Astellas, Astra‐Zeneca, Boehringer‐Ingelheim, Novo Nordisk, Servier and LEO Pharma, all regulator‐mandated phase IV‐studies, all with funds paid to the institution where he was employed (no personal fees) and with no relation to the work reported in this paper. In addition, Jesper Hallas has received honoraria for teaching epidemiology at the Danish Association of Pharmaceutical Manufacturers.
PEER REVIEW
The peer review history for this article is available at https://publons.com/publon/10.1111/acps.13515.
Supporting information
APPENDIX S1: Supporting Information
ACKNOWLEDGMENTS
This work was supported by the Research Fund of Mental Health Services in the Region of Southern Denmark (grant A2957).
Højlund M, Støvring H, Andersen K, Correll CU, Hallas J. Impact of low‐dose quetiapine‐use on glycosylated hemoglobin, triglyceride and cholesterol levels. Acta Psychiatr Scand. 2023;147(1):105‐116. doi: 10.1111/acps.13515
Funding information Mental Health Services in the Region of Southern Denmark
DATA AVAILABILITY STATEMENT
The access to nation‐wide health care registers for this project was approved by the Danish Health Data Authority. For privacy reasons, the underlying data cannot be shared, but can be obtained through application to the Danish Health Data Authority (https://sundhedsdatastyrelsen.dk/da/english). Statistical code can be obtained through request to the authors.
REFERENCES
- 1. Coe HV, Hong IS. Safety of low doses of quetiapine when used for insomnia. Ann Pharmacother. 2012;46(5):718‐722. doi: 10.1345/aph.1Q697 [DOI] [PubMed] [Google Scholar]
- 2. Gjerden P, Bramness JG, Tvete IF, Slørdal L. The antipsychotic agent quetiapine is increasingly not used as such: dispensed prescriptions in Norway 2004‐2015. Eur J Clin Pharmacol. 2017;73(9):1173‐1179. doi: 10.1007/s00228-017-2281-8 [DOI] [PubMed] [Google Scholar]
- 3. 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(1):64‐77. doi: 10.1016/S2215-0366(19)30416-X [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Stogios N, Smith E, Bowden S, et al. Metabolic adverse effects of off‐label use of second‐generation antipsychotics in the adult population: a systematic review and meta‐analysis. Neuropsychopharmacology. 2022;47:664‐672. doi: 10.1038/s41386-021-01163-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Dubath C, Piras M, Gholam M, et al. Effect of quetiapine, from low to high dose, on weight and metabolic traits: results from a prospective cohort study. Pharmacopsychiatry. 2021;13:279‐286. doi: 10.1055/a-1525-2820 [DOI] [PubMed] [Google Scholar]
- 6. Skrede S, Tvete IF, Tanum L, Steen VM, Bramness JG. Incident users of antipsychotic agents and future use of cholesterol‐lowering drugs: an observational pharmacoepidemiologic study. J Clin Psychiatry. 2015;76(1):e111‐e116. doi: 10.4088/JCP.14m08996 [DOI] [PubMed] [Google Scholar]
- 7. De Hert M, Detraux J, van Winkel R, Yu W, Correll CU. Metabolic and cardiovascular adverse effects associated with antipsychotic drugs. Nat Rev Endocrinol. 2012;8(2):114‐126. doi: 10.1038/nrendo.2011.156 [DOI] [PubMed] [Google Scholar]
- 8. Højlund M, Andersen K, Ernst MT, Correll CU, Hallas J. Use of low‐dose quetiapine increases the risk of major adverse cardiovascular events: results from a nationwide active comparator‐controlled cohort study. World Psychiatry ; 2022;21(3):444‐451. doi: 10.1002/wps.210 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Højlund M, Andersen JH, Andersen K, Correll CU, Hallas J. 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: 10.1017/S2045796021000159 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Marston L, Nazareth I, Petersen I, Walters K, Osborn DPJ. Prescribing of antipsychotics in UK primary care: a cohort study. BMJ Open. 2014;4(12):e006135. doi: 10.1136/bmjopen-2014-006135 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Duncan D, Cooke L, Symonds C, Gardner D, Pringsheim T. Quetiapine use in adults in the community: a population‐based study in Alberta, Canada. BMJ Open. 2016;6(3):e010861. doi: 10.1136/bmjopen-2015-010861 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Alexander GC, Gallagher SA, Mascola A, Moloney RM, Stafford RS. Increasing off‐label use of antipsychotic medications in the United States, 1995‐2008. Pharmacoepidemiol Drug Saf. 2011;20(2):177‐184. doi: 10.1002/pds.2082 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. 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. Published online November 14, 2018:k3532. doi: 10.1136/bmj.k3532 [DOI] [PMC free article] [PubMed]
- 14. Pottegård A, Schmidt SAJ, Wallach‐Kildemoes H, Sørensen HT, Hallas J, Schmidt M. Data resource profile: the Danish National Prescription Registry. Int J Epidemiol. 2017;46(3):798‐798f. doi: 10.1093/ije/dyw213 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J Public Health. 2011;39(7 Suppl):30‐33. doi: 10.1177/1403494811401482 [DOI] [PubMed] [Google Scholar]
- 16. Pedersen CB. The Danish civil registration system. Scand J Public Health. 2011;39(7 Suppl):22‐25. doi: 10.1177/1403494810387965 [DOI] [PubMed] [Google Scholar]
- 17. Authority Danish Health Data. Danish National Laboratory Databank [Den Nationale Labdatabank]. Accessed June 15, 2018. https://sundhedsdatastyrelsen.dk/da/registre-og-services/om-de-nationale-sundhedsregistre/doedsaarsager-og-biologisk-materiale/laboratoriedatabasen
- 18. Arendt JFH, Hansen AT, Ladefoged SA, Sørensen HT, Pedersen L, Adelborg K. Existing data sources in clinical epidemiology: laboratory information system databases in Denmark. Clin Epidemiol. 2020;12:469‐475. doi: 10.2147/CLEP.S245060 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Ridker PM. Fasting versus nonfasting triglycerides and the prediction of cardiovascular risk: do we need to revisit the Oral triglyceride tolerance test? Clin Chem. 2008;54(1):11‐13. doi: 10.1373/clinchem.2007.097907 [DOI] [PubMed] [Google Scholar]
- 20. Carr CN, Lopchuk S, Beckman ME, Baugh TB. Evaluation of the use of low‐dose quetiapine and the risk of metabolic consequences: a retrospective review. Ment Health Clin. 2016;6:308‐313. doi: 10.9740/mhc.2016.11.308 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Højlund M, Lund LC, Andersen K, Correll CU, Hallas J. Association of low‐Dose Quetiapine and Diabetes. JAMA Netw Open. 2021;4(5):e213209. doi: 10.1001/jamanetworkopen.2021.3209 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. De Hert M, Vancampfort D, Correll CU, et al. Guidelines for screening and monitoring of cardiometabolic risk in schizophrenia: systematic evaluation. Br J Psychiatry J Ment Sci. 2011;199(2):99‐105. doi: 10.1192/bjp.bp.110.084665 [DOI] [PubMed] [Google Scholar]
- 23. Sundheds‐ og Ældreministeriet . Vejledning Om Behandling Med Antipsykotiske Lægemidler Til Personer over 18 År Med Psykotiske Lidelser . Vol VEJ nr 9276 af 06/05/2014.; 2014. Accessed June 1, 2022. https://www.retsinformation.dk/eli/retsinfo/2014/9276
- 24. Stahl SM, Sy S, Maguire GA. How and when to treat the most common adverse effects of antipsychotics: expert review from research to clinical practice. Acta Psychiatr. Scand. 2021;143(2):172‐180. doi: 10.1111/acps.13266 [DOI] [PubMed] [Google Scholar]
- 25. Højlund M, Rasmussen L, Olesen M, Munk‐Olsen T, Pottegård A. Who prescribes quetiapine in Denmark? Br J Clin Pharmacol. 2022;10:4224‐4229. doi: 10.1111/bcp.15388 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Lee TC, Desforges P, Murray J, Saleh RR, McDonald EG. Off‐label use of quetiapine in medical inpatients and Postdischarge. JAMA Intern Med. 2016;176(9):1390‐1391. doi: 10.1001/jamainternmed.2016.3309 [DOI] [PubMed] [Google Scholar]
- 27. Højlund M, Wagner CB, Wesselhoeft R, Andersen K, Fink‐Jensen A, Hallas J. Use of chlorprothixene and the risk of diabetes and major adverse cardiovascular events: a nation‐wide cohort study. Basic Clin Pharmacol Toxicol. 2022;130(4):501‐512. doi: 10.1111/bcpt.13711 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Stahl SM, Mignon L, Meyer JM. Which comes first: atypical antipsychotic treatment or cardiometabolic risk? Acta Psychiatr Scand. 2009;119(3):171‐179. doi: 10.1111/j.1600-0447.2008.01334.x [DOI] [PubMed] [Google Scholar]
- 29. Correll CU, Detraux J, De Lepeleire J, De Hert M. Effects of antipsychotics, antidepressants and mood stabilizers on risk for physical diseases in people with schizophrenia, depression and bipolar disorder. World Psychiatry. 2015;14(2):119‐136. doi: 10.1002/wps.20204 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
APPENDIX S1: Supporting Information
Data Availability Statement
The access to nation‐wide health care registers for this project was approved by the Danish Health Data Authority. For privacy reasons, the underlying data cannot be shared, but can be obtained through application to the Danish Health Data Authority (https://sundhedsdatastyrelsen.dk/da/english). Statistical code can be obtained through request to the authors.