[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Table 1.  
Baseline Characteristics of Patients Experiencing Psychiatric Hospitalization
Baseline Characteristics of Patients Experiencing Psychiatric Hospitalization
Table 2.  
Psychiatric Hospitalization, Self-harm, and Nonpsychiatric Hospitalization During Drug Exposure vs Nonexposure Periods
Psychiatric Hospitalization, Self-harm, and Nonpsychiatric Hospitalization During Drug Exposure vs Nonexposure Periods
Table 3.  
Additional Analyses of Psychiatric Hospitalization, Self-harm, and Nonpsychiatric Hospitalization During Drug Exposure vs Nonexposure Periods
Additional Analyses of Psychiatric Hospitalization, Self-harm, and Nonpsychiatric Hospitalization During Drug Exposure vs Nonexposure Periods
1.
So  H-C, Chau  CK, Chiu  W-T,  et al.  When GWAS meets the Connectivity Map: drug repositioning for seven psychiatric disorders.  bioRxiv: the Preprint Server for Biology. https://www.biorxiv.org/content/early/2016/12/23/096503. Posted December 23, 2016. Accessed November 15, 2018.
2.
Lago  SG, Bahn  S.  Clinical trials and therapeutic rationale for drug repurposing in schizophrenia  [published online June 26, 2018].  ACS Chem Neurosci. doi:10.1021/acschemneuro.8b00205.Google Scholar
3.
Wettermark  B, Hammar  N, Fored  CM,  et al.  The new Swedish Prescribed Drug Register: opportunities for pharmacoepidemiological research and experience from the first six months [published correction appears in Pharmacoepidemiol Drug Saf. 2008;17(5):533].  Pharmacoepidemiol Drug Saf. 2007;16(7):726-735. doi:10.1002/pds.1294PubMedGoogle ScholarCrossref
4.
Shen  H, Li  R, Yan  R,  et al.  Adjunctive therapy with statins in schizophrenia patients: a meta-analysis and implications.  Psychiatry Res. 2018;262:84-93. doi:10.1016/j.psychres.2018.02.018PubMedGoogle ScholarCrossref
5.
Cipriani  A, Saunders  K, Attenburrow  MJ,  et al.  A systematic review of calcium channel antagonists in bipolar disorder and some considerations for their future development.  Mol Psychiatry. 2016;21(10):1324-1332. doi:10.1038/mp.2016.86PubMedGoogle ScholarCrossref
6.
Casamassima  F, Hay  AC, Benedetti  A, Lattanzi  L, Cassano  GB, Perlis  RH.  L-type calcium channels and psychiatric disorders: a brief review.  Am J Med Genet B Neuropsychiatr Genet. 2010;153B(8):1373-1390. doi:10.1002/ajmg.b.31122PubMedGoogle ScholarCrossref
7.
Ripke  S, O’Dushlaine  C, Chambert  K,  et al; Multicenter Genetic Studies of Schizophrenia Consortium; Psychosis Endophenotypes International Consortium; Wellcome Trust Case Control Consortium 2.  Genome-wide association analysis identifies 13 new risk loci for schizophrenia.  Nat Genet. 2013;45(10):1150-1159. doi:10.1038/ng.2742PubMedGoogle ScholarCrossref
8.
Krystal  JH, D’Souza  DC, Mathalon  D, Perry  E, Belger  A, Hoffman  R.  NMDA receptor antagonist effects, cortical glutamatergic function, and schizophrenia: toward a paradigm shift in medication development.  Psychopharmacology (Berl). 2003;169(3-4):215-233. doi:10.1007/s00213-003-1582-zPubMedGoogle ScholarCrossref
9.
Praharaj  SK, Jana  AK, Goyal  N, Sinha  VK.  Metformin for olanzapine-induced weight gain: a systematic review and meta-analysis.  Br J Clin Pharmacol. 2011;71(3):377-382. doi:10.1111/j.1365-2125.2010.03783.xPubMedGoogle ScholarCrossref
10.
Ying  MA, Maruschak  N, Mansur  R, Carvalho  AF, Cha  DS, McIntyre  RS.  Metformin: repurposing opportunities for cognitive and mood dysfunction.  CNS Neurol Disord Drug Targets. 2014;13(10):1836-1845. doi:10.2174/1871527313666141130205514PubMedGoogle ScholarCrossref
11.
Ramsay  EN, Roughead  EE, Ewald  B, Pratt  NL, Ryan  P.  A self-controlled case series to assess the effectiveness of beta blockers for heart failure in reducing hospitalisations in the elderly.  BMC Med Res Methodol. 2011;11:106. doi:10.1186/1471-2288-11-106PubMedGoogle ScholarCrossref
12.
Bauer-Staeb  C, Jörgensen  L, Lewis  G, Dalman  C, Osborn  DPJ, Hayes  JF.  Prevalence and risk factors for HIV, hepatitis B, and hepatitis C in people with severe mental illness: a total population study of Sweden.  Lancet Psychiatry. 2017;4(9):685-693. doi:10.1016/S2215-0366(17)30253-5PubMedGoogle ScholarCrossref
13.
Hollander  AC, Dal  H, Lewis  G, Magnusson  C, Kirkbride  JB, Dalman  C.  Refugee migration and risk of schizophrenia and other non-affective psychoses: cohort study of 1.3 million people in Sweden.  BMJ. 2016;352:i1030. doi:10.1136/bmj.i1030PubMedGoogle ScholarCrossref
14.
Ekholm  B, Ekholm  A, Adolfsson  R,  et al.  Evaluation of diagnostic procedures in Swedish patients with schizophrenia and related psychoses.  Nord J Psychiatry. 2005;59(6):457-464. doi:10.1080/08039480500360906PubMedGoogle ScholarCrossref
15.
Song  J, Sjölander  A, Joas  E,  et al.  Suicidal behavior during lithium and valproate treatment: a within-individual 8-year prospective study of 50 000 patients with bipolar disorder.  Am J Psychiatry. 2017;174(8):795-802. doi:10.1176/appi.ajp.2017.16050542PubMedGoogle ScholarCrossref
16.
Molero  Y, Lichtenstein  P, Zetterqvist  J, Gumpert  CH, Fazel  S.  Selective serotonin reuptake inhibitors and violent crime: a cohort study.  PLoS Med. 2015;12(9):e1001875. doi:10.1371/journal.pmed.1001875PubMedGoogle ScholarCrossref
17.
Allison  PD.  Fixed Effects Regression Models. Thousand Oaks, CA: SAGE Publications; 2009. Quantitative Applications in the Social Sciences Series.
18.
Rosendo  AB, Lima  LO, Dal-Pizzol  F, Almeida  S.  Lipid and C-reactive protein levels, cardiovascular disease risk factors and simvastatin treatment in Brazilian individuals.  Inflammation. 2010;33(4):244-250. doi:10.1007/s10753-009-9178-3PubMedGoogle ScholarCrossref
19.
Zhang  YY, Fan  YC, Wang  M, Wang  D, Li  XH.  Atorvastatin attenuates the production of IL-1β, IL-6, and TNF-α in the hippocampus of an amyloid β1-42–induced rat model of Alzheimer’s disease.  Clin Interv Aging. 2013;8:103-110. doi:10.2147/CIA.S40405PubMedGoogle Scholar
20.
Boulton  DW, DeVane  CL, Liston  HL, Markowitz  JS.  In vitro P-glycoprotein affinity for atypical and conventional antipsychotics.  Life Sci. 2002;71(2):163-169. doi:10.1016/S0024-3205(02)01680-6PubMedGoogle ScholarCrossref
21.
Tramontina  AC, Wartchow  KM, Rodrigues  L,  et al.  The neuroprotective effect of two statins: simvastatin and pravastatin on a streptozotocin-induced model of Alzheimer’s disease in rats.  J Neural Transm (Vienna). 2011;118(11):1641-1649. doi:10.1007/s00702-011-0680-zPubMedGoogle ScholarCrossref
22.
Wang  Q, Zengin  A, Ying  W,  et al.  Chronic treatment with simvastatin upregulates muscarinic M1/4 receptor binding in the rat brain.  Neuroscience. 2008;154(3):1100-1106. doi:10.1016/j.neuroscience.2008.04.026PubMedGoogle ScholarCrossref
23.
Wang  Q, Tang  XN, Wang  L,  et al.  Effects of high dose of simvastatin on levels of dopamine and its reuptake in prefrontal cortex and striatum among SD rats.  Neurosci Lett. 2006;408(3):189-193. doi:10.1016/j.neulet.2006.09.009PubMedGoogle ScholarCrossref
24.
Maron  DJ, Fazio  S, Linton  MF.  Current perspectives on statins.  Circulation. 2000;101(2):207-213. doi:10.1161/01.CIR.101.2.207PubMedGoogle ScholarCrossref
25.
Striessnig  J, Koschak  A, Sinnegger-Brauns  MJ,  et al.  Role of voltage-gated L-type Ca2+ channel isoforms for brain function.  Biochem Soc Trans. 2006;34(pt 5):903-909. doi:10.1042/BST0340903PubMedGoogle ScholarCrossref
26.
Agarwal  SM, Ahsan  Z, Lockwood  J,  et al.  S225: a systematic review and meta-analysis of pharmacological interventions for reduction or prevention of weight gain in schizophrenia.  Schizophren Bull. 2018;44(suppl 1):S413. doi:10.1016/j.biopsych.2018.02.1137Google Scholar
27.
Steiner  J, Bernstein  HG, Schiltz  K,  et al.  Immune system and glucose metabolism interaction in schizophrenia: a chicken-egg dilemma.  Prog Neuropsychopharmacol Biol Psychiatry. 2014;48:287-294. doi:10.1016/j.pnpbp.2012.09.016PubMedGoogle ScholarCrossref
28.
Cooper  SJ, Reynolds  GP, Barnes  T,  et al.  BAP guidelines on the management of weight gain, metabolic disturbances and cardiovascular risk associated with psychosis and antipsychotic drug treatment.  J Psychopharmacol. 2016;30(8):717-748. doi:10.1177/0269881116645254PubMedGoogle ScholarCrossref
29.
Osborn  DP, Hardoon  S, Omar  RZ,  et al.  Cardiovascular risk prediction models for people with severe mental illness: results from the Prediction and Management of Cardiovascular Risk in People With Severe Mental Illnesses (PRIMROSE) research program.  JAMA Psychiatry. 2015;72(2):143-151. doi:10.1001/jamapsychiatry.2014.2133PubMedGoogle ScholarCrossref
30.
Hippisley-Cox  J, Coupland  C, Brindle  P.  Development and validation of QRISK3 risk prediction algorithms to estimate future risk of cardiovascular disease: prospective cohort study.  BMJ. 2017;357:j2099. doi:10.1136/bmj.j2099PubMedGoogle ScholarCrossref
31.
Šimić  I, Reiner  Ž.  Adverse effects of statins: myths and reality.  Curr Pharm Des. 2015;21(9):1220-1226. doi:10.2174/1381612820666141013134447PubMedGoogle ScholarCrossref
32.
Eisenberg  MJ, Brox  A, Bestawros  AN.  Calcium channel blockers: an update.  Am J Med. 2004;116(1):35-43. doi:10.1016/j.amjmed.2003.08.027PubMedGoogle ScholarCrossref
33.
Kirpichnikov  D, McFarlane  SI, Sowers  JR.  Metformin: an update.  Ann Intern Med. 2002;137(1):25-33. doi:10.7326/0003-4819-137-1-200207020-00009PubMedGoogle ScholarCrossref
34.
Dalman  Ch, Broms  J, Cullberg  J, Allebeck  P.  Young cases of schizophrenia identified in a national inpatient register: are the diagnoses valid?  Soc Psychiatry Psychiatr Epidemiol. 2002;37(11):527-531. doi:10.1007/s00127-002-0582-3PubMedGoogle ScholarCrossref
35.
Jörgensen  L, Ahlbom  A, Allebeck  P, Dalman  C.  The Stockholm Non-Affective Psychoses Study (SNAPS): the importance of including out-patient data in incidence studies.  Acta Psychiatr Scand. 2010;121(5):389-392. doi:10.1111/j.1600-0447.2009.01500.xPubMedGoogle ScholarCrossref
Views 8,606
Citations 0
Original Investigation
January 9, 2019

Association of Hydroxylmethyl Glutaryl Coenzyme A Reductase Inhibitors, L-Type Calcium Channel Antagonists, and Biguanides With Rates of Psychiatric Hospitalization and Self-Harm in Individuals With Serious Mental Illness

Author Affiliations
  • 1Division of Psychiatry, University College London, London, United Kingdom
  • 2Department of Public Health Sciences, Epidemiology of Psychiatric Conditions, Substance Use, and Social Environment, Karolinska Institute, Stockholm, Sweden
  • 3Centre for Medicines Optimisation Research and Education, Research Department of Practice and Policy, School of Pharmacy, University College London, London, United Kingdom
  • 4Department of Pharmacology and Pharmacy, University of Hong Kong, Pokfulam, Hong Kong
JAMA Psychiatry. Published online January 9, 2019. doi:10.1001/jamapsychiatry.2018.3907
Key Points

Question  Are drugs in common use for physical health problems (hydroxylmethyl glutaryl coenzyme A reductase inhibitors, L-type calcium channel antagonists, and biguanides) associated with reduced rates of psychiatric hospitalization and self-harm in individuals with serious mental illness?

Findings  In this series of within-individual cohort studies of 142 691 patients with bipolar disorder, schizophrenia, or nonaffective psychosis, exposure to any of the study drugs was associated with reduced rates of psychiatric hospitalizaiton compared with unexposed periods. Self-harm was reduced in patients with bipolar disorder and schizophrenia during exposure to all study drugs and in patients with nonaffective psychosis taking L-type calcium channel antagonists.

Meaning  Hydroxylmethyl glutaryl coenzyme A reductase inhibitors, L-type calcium channel antagonists, and biguanides hold potential as repurposed agents in serious mental illness, and the central nervous system mechanism of action of these drugs requires further investigation.

Abstract

Importance  Drug repurposing is potentially cost-effective, low risk, and necessary in psychiatric drug development. The availability of large, routine data sets provides the opportunity to evaluate the potential for currently used medication to benefit people with serious mental illness (SMI).

Objective  To determine whether hydroxylmethyl glutaryl coenzyme A reductase inhibitors (HMG-CoA RIs), L-type calcium channel (LTCC) antagonists, and biguanides are associated with reduced psychiatric hospitalization and self-harm in individuals with SMI.

Design, Setting, and Participants  These within-individual cohort studies of patients with SMI compared rates of psychiatric hospitalization and self-harm during periods of exposure and nonexposure to the study drugs, with adjusting for a number of time-varying covariates. Participants included 142 691 individuals from the entire population of Sweden with a diagnosis of bipolar disorder (BPD), schizophrenia, or nonaffective psychosis (NAP) who were 15 years or older and who were treated with psychiatric medication from October 1, 2005, through December 31, 2016. Data were analyzed from April 1 through August 31, 2018.

Interventions  Treatment with HMG-CoA RIs, LTCC antagonists, or biguanides.

Main Outcomes and Measures  Psychiatric hospitalizations and self-harm admissions.

Results  Among the 142 691 eligible participants, the HMG-CoA RI exposure periods were associated with reduced rates of psychiatric hospitalization in BPD (adjusted hazard ratio [aHR], 0.86; 95% CI, 0.83-0.89; P < .001), schizophrenia (aHR, 0.75; 95% CI, 0.71-0.79; P < .001), and NAP (aHR, 0.80; 95% CI, 0.75-0.85; P < .001) and reduced self-harm rates in BPD (aHR, 0.76; 95% CI, 0.66-0.86; P < .001) and schizophrenia (aHR, 0.58; 95% CI, 0.45-0.74; P < .001). Exposure to LTCC antagonists was associated with reduced rates of psychiatric hospitalization and self-harm in subgroups with BPD (aHRs, 0.92 [95% CI, 0.88-0.96; P < .001] and 0.81 [95% CI, 0.68-0.95; P = .01], respectively), schizophrenia (aHRs, 0.80 [95% CI, 0.74-0.85; P < .001] and 0.30 [95% CI, 0.18-0.48; P < .001], respectively), and NAP (aHRs, 0.89 [95% CI, 0.83-0.96; P = .002] and 0.56 [95% CI, 0.42-0.74; P < .001], respectively). During biguanide exposure, psychiatric hospitalization rates were reduced in subgroups with BPD (aHR, 0.80; 95% CI, 0.77-0.84; P < .001), schizophrenia (aHR, 0.73; 95% CI, 0.69-0.77; P < .001), and NAP (aHR, 0.85; 95% CI, 0.79-0.92; P < .001), and self-harm was reduced in BPD (aHR, 0.73; 95% CI, 0.62-0.84; P < .001) and schizophrenia (aHR, 0.64; 95% CI, 0.48-0.85; P < .001).

Conclusions and Relevance  This study provides additional evidence that exposure to HMG-CoA RIs, LTCC antagonists, and biguanides might lead to improved outcomes for individuals with SMI. Given the well-known adverse event profiles of these agents, they should be further investigated as repurposed agents for psychiatric symptoms.

Introduction

Serious mental illnesses (SMIs), including bipolar disorder (BPD), schizophrenia, and nonaffective psychoses (NAP), are associated with high levels of morbidity and are challenging to treat. Many drugs have been identified as having potential for repurposing in these disorders.1,2 We cross-referenced these drugs with the most commonly prescribed drugs in the general population.3 We chose to investigate the following 3 classes of drugs for which we might expect clinical benefit: hydroxylmethyl glutaryl coenzyme A reductase inhibitors (HMG-CoA RIs; ie, statins), L-type calcium channel (LTCC) antagonists (eg, verapamil hydrochloride), and biguanides (eg, metformin hydrochloride).

A recent meta-analysis of statins as adjunctive therapy for schizophrenia included 6 placebo-controlled randomized clinical trials (RCTs) and found a reduction in the Positive and Negative Symptom Scale scores in patients receiving statins for 12 weeks; however, only 169 received active treatment, and the reduction fell below the threshold considered clinically meaningful.4 Theorized mechanisms for HMG-CoA RI psychiatric action are via anti-inflammatory effects or via increased absorption and central nervous system uptake of antipsychotics.4 A systematic review and meta-analysis of LTCC antagonists for the treatment of BPD5 included 2 RCTs of verapamil vs placebo and 4 of verapamil vs lithium for the treatment of mania. The investigators found no evidence of an effect. However, no trials of other LTCC antagonists and no studies in BPD depression or prophylaxis have been performed. Evidence of calcium dysregulation in BPD has long been available; this evidence is supported by more recent pharmacologic, genetic, and molecular findings.6 Similarly, calcium signaling is implicated in the etiology of schizophrenia, where LTCC antagonists are also a potential adjuvant treatment.7 However, very limited and inconsistent trial evidence for their use in this disorder exists.8 A trial of metformin has been undertaken to counter antipsychotic-related weight gain,9 but metformin is also hypothesized to improve cognitive and mood dysfunction symptoms via mitigation of metabolic disturbances.10 Therefore, although none of these drugs have been comprehensively investigated as repurposed agents to improve mental disorders, each has a theoretical basis for effectiveness.

We investigated whether people in the Swedish population with SMI had lower rates of psychiatric hospitalization and self-harm during periods when they were prescribed HMG-CoA RIs, LTCC antagonists, and biguanides. Owing to the risks of confounding by indication, we used a design that compared within-individual periods with and without medication exposure,11 which controls for time-fixed confounders. In view of the proposed mechanism of action of each drug on psychiatric symptoms, we hypothesized that effects would be similar across SMI subtypes. We also examined rates of nonpsychiatric hospitalization in these groups and hypothesized that, contrary to psychiatric effects, nonpsychiatric hospitalizations would only be reduced in the group exposed to biguanides because these agents have an acute physical effect, whereas HMG-CoA RIs and LTCC antagonists have a longer-term mechanism of action.

Methods
Study Population

We collected data from national registers of Sweden from January 1, 1973, until December 31, 2016. For the purpose of this study, the Total Population Register, Migration Register, Cause of Death Register, Prescribed Drug Register, and National Patient Registers were linked. The Prescribed Drug Register contains data on all prescriptions collected from July 1, 2005, onward (excluding over-the-counter medications and drugs used in hospitals).3 The National Patient Registers contain inpatient records from 1964 onward and outpatient records from 2001 onward. These registers contain sociodemographic and medical information on each resident of Sweden. Ethical approval for the study was obtained via the Regional Ethical Review Board in Stockholm, Sweden, which waived the need for informed consent.

Definition of SMI

We defined diagnoses of BPD, schizophrenia, and NAP in line with previous research using Swedish National Patient Registers.12,13 Individuals had at least 1 record of inpatient or outpatient contact with codes from the International Classification of Diseases, Ninth Revision (ICD-9) and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) consistent with the diagnosis (eTable 1 in the Supplement) and a period of drug treatment with an appropriate psychotropic medication (antipsychotic, lithium, or anticonvulsant class mood stabilizer) from October 1, 2005, through December 31, 2016. Validation studies suggest that using inpatient SMI diagnosis alone has high positive predictive power (0.94) but potentially misses suitable patients owing to type II error (negative predictive power, 0.23).14

Exposure to Study Medications

The main exposure was defined as treatment with HMG-CoA RIs, LTCC antagonists, or biguanides (eTable 2 in the Supplement) in the Prescribed Drug Register. In routine practice in Sweden, oral medications are rarely dispensed for longer than 3 months at a time. In line with previous studies, we defined a medication period as a sequence of at least 2 prescriptions, with no more than 3 months (93 days) between any 2 consecutive prescriptions.15,16 Therefore, individuals were defined as exposed to medication during the interval between 2 dispensed prescriptions unless the dispensed prescriptions occurred more than 3 months apart. To determine whether an individual was exposed or nonexposed to the medication initially, the follow-up start was set to October 1, 2005, because the coverage of the Prescribed Drug Register started on July 1, 2005.

Outcomes

Outcomes were psychiatric hospitalization (identified by ward admission codes and ICD-9 and ICD-10 codes for admission indication) and self-harm (with suicidal and undetermined intent) (eTable 1 in the Supplement). We also examined rates of nonpsychiatric hospitalization. Outcome data came from the National Patient Register.

Time-Varying Confounders

As a result of the within-individual design, time-fixed covariates (eg, sex, age at illness onset) cannot confound any association between the study drug and the outcome of interest. Time-varying confounders were included to capture the potential for changes in psychiatric symptoms over time, including age, year, number of previous outcome events (ie, hospitalizations, episodes of self-harm), exposure to psychiatric medication (ie, mood stabilizer or antipsychotic), and a count of exposed and nonexposed periods.

Statistical Analysis

Data were analyzed from April 1 through August 31, 2018. We performed within-individual analyses using stratified Cox proportional hazards regression. This method, by design, exclusively draws information from individuals who ever experienced the outcome of interest during follow-up.17 Therefore, a different cohort is identified and analyzed for each combination of drug exposure and outcome. We split follow-up time into consecutive periods. A new period started after a medication switch (ie, starting and stopping medication) or an outcome event (after the date of discharge). For the latter, we restarted the period at baseline (ie, the time since the last event). Drug treatments were defined as time-varying dichotomous exposures. We estimated adjusted hazard ratios (aHRs) and 95% CIs for differences in outcome event rate between medication exposure and nonexposure periods; 2-sided P values were calculated using Wald tests.

Additional Analyses

We tested for an interaction between HMG-CoA RI and antipsychotic medication exposure, because of the theory that HMG-CoA RIs may increase uptake of antipsychotics by the central nervous system.4 We then performed several sensitivity analyses. First, we examined differences in rates of the 3 outcomes in periods that patients with SMI were exposed or not exposed to thiazide diuretics. We chose this class of drug as a negative control; no evidence suggests that exposure to thiazide diuretics should improve psychiatric outcomes, so lower rates during exposed periods might suggest that stability of SMI symptoms is associated with a period of medication adherence rather than a direct effect of the medication on psychiatric symptoms (ie, reverse causality). Second, we adjusted for time with and without each of the other study drugs and physical health problems experienced by the individual (cardiovascular or cerebrovascular disease, type 2 diabetes, hypertension, and hyperlipidemia). Third, we excluded from the analysis individuals who died during the follow-up period because they may have a more rapidly varying risk of the outcomes of interest. Fourth, we excluded psychiatric admissions, which were prompted by admissions for self-harm (ie, a psychiatric admission in the 7 days after a self-harm admission), to examine whether self-harm and psychiatric admissions were independent. Finally, we extended the definition of the drug exposure period and defined it as constant if a new prescription was issued within 4 months (124 days) of the previous prescription and extended the exposure period to 2 months (62 days) after the last dispensed date to account for potential exposure misclassification in our primary analysis.

Results

These studies drew from the entire population of Sweden who ever received an inpatient or outpatient diagnosis of SMI and were treated with an antipsychotic or mood stabilizer medication from October 1, 2005, through December 31, 2016 (eFigure in the Supplement). This base population included 76 759 patients with BPD, 30 954 with schizophrenia, and 34 978 with NAP for a total of 142 691 patients.

HMG-CoA Reductase Inhibitors

Of the patients described above, 6176 with BPD (3493 female [56.6%] and 2683 male [43.4%]; mean [SD] age, 55.3 [12.3] years), 2705 with schizophrenia (1084 female [40.1%] and 1621 male [59.9%]; mean [SD] age, 49.6 [12.6] years), and 2958 with NAP (1603 female [54.2%] and 1355 male [45.8%]; mean [SD] age, 57.8 [15.0] years) were also prescribed HMG-CoA RIs and experienced psychiatric hospitalization (Table 1). All subgroups of SMI had reduced rates of psychiatric hospitalization during HMG-CoA RI exposure after adjusting for age, year, number of previous hospitalizations, psychiatric medication exposure, and number of exposure periods (aHR for BPD, 0.86 [95% CI, 0.83-0.89]; aHR for schizophrenia, 0.75 [95% CI, 0.71-0.79]; and aHR for NAP, 0.80 [95% CI, 0.75-0.85]; P < .001 for all). Individuals with BPD (aHR, 0.76; 95% CI, 0.66-0.86; P < .001) or schizophrenia diagnoses (aHR, 0.58; 95% CI, 0.45-0.74; P < .001) had lower rates of self-harm during HMG-CoA RI exposure periods compared with nonexposure periods (Table 2; eTable 3 in the Supplement). Exposure status for HMG-CoA RIs was not associated with nonpsychiatric hospitalization, apart from in the NAP subgroup (aHR, 1.04; 95% CI, 1.01-1.08; P = .02) (Table 2; eTable 4 in the Supplement). We found no evidence of an interaction between HMG-CoA RI and antipsychotic exposure. Adjustment for additional time-varying confounders, such as LTCC antagonist and biguanide exposure, cardiovascular or cerebrovascular disease, type 2 diabetes, hypertension, and hyperlipidemia, did not affect the results for psychiatric hospitalization or self-harm outcomes, but nonpsychiatric hospitalization rates were significantly reduced during HMG-CoA RI exposure periods (aHR, 0.73; 95% CI, 0.72-0.74; P < .001) (Table 3). Excluding patients who died during follow-up did not change results for rates of psychiatric hospitalization or self-harm, but HMG-CoA RI exposure periods became associated with increased rates of nonpsychiatric hospitalization (aHR, 1.08; 95% CI, 1.06-1.11; P < .001). After excluding psychiatric hospitalizations prompted by self-harm, the association between HMG-CoA RI exposure periods and psychiatric hospitalization remained similar (aHR, 0.82; 95% CI, 0.80-0.84; P < .001). Redefining exposure period as constant for less than 4 months between prescriptions and extending to 2 months after the last prescription date did not change the findings.

LTCC Antagonists

Of those meeting inclusion criteria and prescribed LTCC antagonists, 4636 with BPD (2676 female [57.7%] and 1960 male [42.3%]; mean [SD] age, 56.1 [13.5] years), 1581 with schizophrenia (701 female [44.3%] and 880 male [55.7%]; mean [SD] age, 54.2 [12.1] years), and 2337 with NAP (1406 female [60.2%] and 931 male [39.8%]; mean [SD] age, 60.9 [15.4] years) experienced psychiatric hospitalization (Table 1). During periods of LTCC antagonist exposure, all groups had reduced psychiatric admissions (aHR for BPD, 0.92 [95% CI, 0.88-0.96; P < .001]; aHR for schizophrenia, 0.80 [95% CI, 0.74-0.85; P < .001]; and aHR for NAP, 0.89 [95% CI, 0.83-0.96; P = .002]) and reduced self-harm events (aHR for BPD, 0.81 [95% CI, 0.68-0.95; P = .01]; aHR for schizophrenia, 0.30 [95% CI, 0.18-0.48; P < .001]; aHR for NAP, 0.56 [95% CI, 0.42-0.74; P < .001]) (Table 2 and eTable 3 in the Supplement). Exposure to LTCC antagonists was not associated with the nonpsychiatric hospitalization rate (eTable 4 in the Supplement). Additional adjustment for the other study drugs and physical health problems did not affect the psychiatric hospitalization or self-harm results, but nonpsychiatric admissions were significantly reduced during LTCC antagonist exposure (aHR, 0.95; 95% CI, 0.93-0.97; P < .001) (Table 3). Excluding patients who died suggested an increased rate of nonpsychiatric admission during LTCC antagonist exposure (aHR, 1.09; 95% CI, 1.06-1.12; P < .001). Excluding psychiatric admissions prompted by self-harm did not alter the association (aHR for all SMI, 0.86; 95% CI, 0.83-0.88; P < .001). Redefining the drug exposure period did not alter our findings.

Biguanides

In total, 7349 patients were exposed to metformin and had psychiatric hospitalizations during follow-up, of whom 3493 had BPD (1998 female [57.2%] and 1495 male [42.8%]; mean [SD] age, 50.5 [14.4] years), 2294 had schizophrenia (974 female [42.4%] and 1320 male [57.5%]; mean [SD] age, 47.2 [13.2] years), and 1562 had NAP (839 female [53.7%] and 723 male [46.3%]; mean [SD] age, 50.5 [16.4] years) (Table 1). All subgroups had similarly reduced psychiatric hospitalization rates during metformin exposed compared with unexposed periods (aHR for BPD, 0.80 [95% CI, 0.77-0.84; P < .001]; aHR for schizophrenia, 0.73 [95% CI, 0.69-0.77; P < .001]; and aHR for NAP, 0.85 [95% CI, 0.79-0.92; P < .001]). Self-harm was reduced in the BPD (aHR, 0.73; 95% CI, 0.62-0.84; P < .001) and schizophrenia (aHR, 0.64; 95% CI, 0.48-0.85; P < .001) subgroups (Table 2 and eTable 3 in the Supplement). Across all diagnostic groups, metformin exposure was associated with reduced rates of nonpsychiatric hospitalization (aHR for BPD, 0.84 [95% CI, 0.81-0.87]; aHR for schizophrenia, 0.84 [95% CI, 0.80-0.89]; and aHR for NAP, 0.85 [95% CI, 0.80-0.90]; P < .001 for all) (eTable 4 in the Supplement). Additional adjustment for the other study drugs and physical health problems, exclusion of those who died during follow-up, and redefining the exposure period did not affect the results (Table 3), nor did exclusion of psychiatric admissions in the 7 days after self-harm (aHR, 0.79; 95% CI, 0.77-0.81; P < .001).

Thiazide Diuretics

No evidence suggested that thiazide diuretic exposure was associated with psychiatric hospitalization (aHR, 0.99; 95% CI, 0.93-1.05; P = .64) or self-harm (aHR, 0.94; 95% CI, 0.75-1.19; P = .61). However, exposure was associated with a reduced rate of nonpsychiatric admissions (aHR, 0.92; 95% CI, 0.88-0.96; P < .001) (Table 3 and eTable 5 in the Supplement).

Discussion

As far as we are aware, this study is the first to suggest that periods of exposure to HMG-CoA RIs, LTCC antagonists, and biguanides are associated with lower rates of psychiatric admission and self-harm in patients with SMI. Each of these drugs has a theoretical basis for effectively reducing psychiatric symptoms.

We found that periods of HMG-CoA RI exposure were associated with reduced psychiatric hospitalization in all subgroups of SMI and with reduced self-harm in BPD and schizophrenia. Several actions of HMG-CoA RIs could explain our findings. Statins are anti-inflammatory, with effects on levels of interleukin-1β, interleukin-6, tumor necrosis factor, and C-reactive protein.18,19 Extensive evidence suggests that systemic and neuroinflammatory processes are involved in the pathophysiology of psychiatric disorders.4 Alternatively (or synergistically), HMG-CoA RIs may interact with other substrates for P-glycoprotein (including antipsychotic medications). We found no evidence of interaction between periods of statin and antipsychotic exposure in our study. However, to assess this mechanism, we would need to specifically examine drugs with a high affinity for P-glycoprotein, such as quetiapine fumarate or risperidone.20 The low number of events in those exposed to specific antipsychotics prevented robust analysis. In addition, animal models have found that statins up-regulate N-methyl-d-aspartate receptors21 and increase muscarinic receptor binding,22 potentially producing neuroprotective effects, and increase dopamine receptor levels centrally,23 potentially causing antipsychotic effects. Any of these mechanisms may reflect the relatively acute effect of statins on rates of psychiatric hospitalization and self-harm that we observe. In contrast, the mechanism for treatment of hypercholesterolemia is slower, involving inhibition of HMG-CoA reductase, a resultant reduction in hepatocyte cholesterol level, and an increase of hepatic low-density lipoprotein receptor levels, which clear circulating low-density lipoprotein and its precursors.24

Exposure to LTCC antagonists was associated with reduced rates of psychiatric hospitalization and self-harm. The LTCCs have been implicated in the physiopathology of a range of mental health conditions, including mood disorders, addiction, dementia, sleep disorders,6 and psychotic illness.7 Inhibition of LTCCs has cardiodepressant and vasodilatory actions, but isoforms of LTCCs are widely distributed beyond the cardiovascular system, including neurons.25 These isoforms appear to control emotional behavior in animal models and are associated with antidepressant and anxiolytic effects.25

Periods of metformin exposure were associated with reduced psychiatric and nonpsychiatric hospitalization across all SMI subgroups. Self-harm was reduced in individuals with BPD and schizophrenia during biguanide exposure. Metformin has been widely investigated in psychiatry because of its potential to attenuate antipsychotic-induced weight gain, including in patients without diabetes.26 Unfortunately, these trials did not routinely report changes in psychiatric symptoms. Metformin may potentially improve symptoms in SMI because of underlying disturbed cerebral use of glucose, particularly in brain areas linked to cognitive impairments in schizophrenia.27

Hypercholesterolemia, hypertension, diabetes, and prediabetes are more common in people with SMI, and evidence suggests that they are not as rigorously treated as in the general population.28 Our findings notwithstanding, more effort should be made to prescribe these drugs when individuals with SMI fulfill established clinical criteria for their use. In addition, use of these medications within clinical guidelines is already broader than in the general population. For example, a statin will be indicated in any man older than 60 years with SMI, but no other risk factors for cardiovascular disease, according to risk calculators,29,30 and adjunctive metformin therapy is recommended to manage antipsychotic-associated weight gain.28 All 3 classes have relatively good adverse event profiles.31-33 Ideally, further RCTs of these drugs will be conducted that are appropriately powered, with meaningful clinical end points.

Strengths and Limitations

Compared with possible RCTs, our large, population-based longitudinal sample avoids selection bias because it potentially includes all individuals. In RCTs, recruiting representative samples, such as actively suicidal participants or those who are acutely unwell, is more challenging. Our study is therefore potentially more generalizable and representative. However, to be included in the study, individuals had to be prescribed a study drug, suggesting poor physical health, and experience at least 1 outcome, suggesting poor mental health. We therefore potentially included the most unwell individuals. In addition, most individuals were older than 50 years at the start of follow-up, so survival bias is possible. Whether our results generalize to the wider population with SMI remains unclear, but the group studied would usually be considered most difficult to treat. The comparison of exposed and unexposed periods within the same individual automatically controls for time-invariant confounders, reducing the likelihood of confounding by indication. We then additionally adjusted for potentially important time-varying confounders. However, as with other observational designs, we cannot rule out residual and unmeasured time-varying confounding. In particular, confounding may be linked to varying severity of SMI symptoms. We attempted to investigate whether periods of symptom stability were associated with adherence to any prescribed medication and a reduced rate of psychiatric hospitalization and self-harm by examining nonpsychiatric hospitalization rates and by studying a drug class that should have no effect on psychiatric outcomes. Nonpsychiatric hospitalization was not associated with HMG-CoA RI or LTCC antagonist exposure periods, suggesting no global effect on health service use. The effects seem specific to psychiatric outcomes. Biguanide exposure was probably associated with reduced nonpsychiatric hospitalization because this class of medication has more acute effects. Thiazide diuretics, a group of medications that have no theorized basis for improving psychiatric symptoms, were not associated with rates of psychiatric hospitalization or self-harm rate but were associated with reduced rates of nonpsychiatric hospitalization. This finding suggests that medication adherence in and of itself is not a proxy for mental state stability, which would be reflected in reduced rates of psychiatric hospitalization and self-harm.

Exposure misclassification is a potential risk; in our primary analysis, patients were considered unexposed after the date of their last prescription. This classification potentially underestimates exposure time but would bias estimates toward the null. In the sensitivity analysis with increased exposure time, effect estimates were comparable. Unfortunately, the Prescribed Drug Register does not include information on dose and days supply. Similarly, we lack direct information on treatment adherence and defined patients collecting 2 or more prescriptions as adherent. Nonadherence would mean our reported associations represent underestimates of the true effect. Misclassification of SMI subgroup is possible, but given the inclusion criteria for this study, we were unlikely to include patients who do not have some form of SMI.14,34,35 Self-harm that did not result in hospital attendance is not captured by our study.

Conclusions

If substantiated, this study has considerable implications for clinical practice and drug development. The study drugs—HMG-CoA RIs, LTCC antagonists, and biguanides—are globally licensed, commonly used, cheap, and relatively safe medications. They are therefore ideal candidates for repurposing. Understanding their mode of action on the central nervous system may facilitate better understanding of the pathophysiology of SMI and offer opportunities for innovative pharmacotherapy development.

Back to top
Article Information

Accepted for Publication: October 5, 2018.

Published Online: January 9, 2019. doi:10.1001/jamapsychiatry.2018.3907

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Hayes JF et al. JAMA Psychiatry.

Corresponding Author: Joseph F. Hayes, PhD, Division of Psychiatry, University College London, 149 Tottenham Court Rd, 6th Floor Maple House, London W1T 7NF, United Kingdom (joseph.hayes@ucl.ac.uk).

Author Contributions: Drs Hayes and Dalman had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Hayes, Lundin, Osborn.

Acquisition, analysis, or interpretation of data: Hayes, Wicks, Lewis, Wong, Osborn, Dalman.

Drafting of the manuscript: Hayes.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Hayes, Wicks.

Obtained funding: Hayes.

Administrative, technical, or material support: Hayes, Lundin, Dalman.

Supervision: Hayes, Lundin, Lewis, Wong, Osborn, Dalman.

Conflict of Interest Disclosures: Dr Wong reported receiving research funding from Bristol-Myers Squibb, Pfizer, and Janssen Pharmaceutica outside the submitted work. No other disclosures were reported.

Funding/Support: This study was supported by grant 211085/Z/18/Z from the Wellcome Trust (Dr Hayes), the University College London Hospitals National Institute for Health Research Biomedical Research Centre (Drs Hayes, Lewis, and Osborn), and grant 523-2010-1052 from the Swedish Research Council (Dr Dalman).

Role of the Funder/Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

References
1.
So  H-C, Chau  CK, Chiu  W-T,  et al.  When GWAS meets the Connectivity Map: drug repositioning for seven psychiatric disorders.  bioRxiv: the Preprint Server for Biology. https://www.biorxiv.org/content/early/2016/12/23/096503. Posted December 23, 2016. Accessed November 15, 2018.
2.
Lago  SG, Bahn  S.  Clinical trials and therapeutic rationale for drug repurposing in schizophrenia  [published online June 26, 2018].  ACS Chem Neurosci. doi:10.1021/acschemneuro.8b00205.Google Scholar
3.
Wettermark  B, Hammar  N, Fored  CM,  et al.  The new Swedish Prescribed Drug Register: opportunities for pharmacoepidemiological research and experience from the first six months [published correction appears in Pharmacoepidemiol Drug Saf. 2008;17(5):533].  Pharmacoepidemiol Drug Saf. 2007;16(7):726-735. doi:10.1002/pds.1294PubMedGoogle ScholarCrossref
4.
Shen  H, Li  R, Yan  R,  et al.  Adjunctive therapy with statins in schizophrenia patients: a meta-analysis and implications.  Psychiatry Res. 2018;262:84-93. doi:10.1016/j.psychres.2018.02.018PubMedGoogle ScholarCrossref
5.
Cipriani  A, Saunders  K, Attenburrow  MJ,  et al.  A systematic review of calcium channel antagonists in bipolar disorder and some considerations for their future development.  Mol Psychiatry. 2016;21(10):1324-1332. doi:10.1038/mp.2016.86PubMedGoogle ScholarCrossref
6.
Casamassima  F, Hay  AC, Benedetti  A, Lattanzi  L, Cassano  GB, Perlis  RH.  L-type calcium channels and psychiatric disorders: a brief review.  Am J Med Genet B Neuropsychiatr Genet. 2010;153B(8):1373-1390. doi:10.1002/ajmg.b.31122PubMedGoogle ScholarCrossref
7.
Ripke  S, O’Dushlaine  C, Chambert  K,  et al; Multicenter Genetic Studies of Schizophrenia Consortium; Psychosis Endophenotypes International Consortium; Wellcome Trust Case Control Consortium 2.  Genome-wide association analysis identifies 13 new risk loci for schizophrenia.  Nat Genet. 2013;45(10):1150-1159. doi:10.1038/ng.2742PubMedGoogle ScholarCrossref
8.
Krystal  JH, D’Souza  DC, Mathalon  D, Perry  E, Belger  A, Hoffman  R.  NMDA receptor antagonist effects, cortical glutamatergic function, and schizophrenia: toward a paradigm shift in medication development.  Psychopharmacology (Berl). 2003;169(3-4):215-233. doi:10.1007/s00213-003-1582-zPubMedGoogle ScholarCrossref
9.
Praharaj  SK, Jana  AK, Goyal  N, Sinha  VK.  Metformin for olanzapine-induced weight gain: a systematic review and meta-analysis.  Br J Clin Pharmacol. 2011;71(3):377-382. doi:10.1111/j.1365-2125.2010.03783.xPubMedGoogle ScholarCrossref
10.
Ying  MA, Maruschak  N, Mansur  R, Carvalho  AF, Cha  DS, McIntyre  RS.  Metformin: repurposing opportunities for cognitive and mood dysfunction.  CNS Neurol Disord Drug Targets. 2014;13(10):1836-1845. doi:10.2174/1871527313666141130205514PubMedGoogle ScholarCrossref
11.
Ramsay  EN, Roughead  EE, Ewald  B, Pratt  NL, Ryan  P.  A self-controlled case series to assess the effectiveness of beta blockers for heart failure in reducing hospitalisations in the elderly.  BMC Med Res Methodol. 2011;11:106. doi:10.1186/1471-2288-11-106PubMedGoogle ScholarCrossref
12.
Bauer-Staeb  C, Jörgensen  L, Lewis  G, Dalman  C, Osborn  DPJ, Hayes  JF.  Prevalence and risk factors for HIV, hepatitis B, and hepatitis C in people with severe mental illness: a total population study of Sweden.  Lancet Psychiatry. 2017;4(9):685-693. doi:10.1016/S2215-0366(17)30253-5PubMedGoogle ScholarCrossref
13.
Hollander  AC, Dal  H, Lewis  G, Magnusson  C, Kirkbride  JB, Dalman  C.  Refugee migration and risk of schizophrenia and other non-affective psychoses: cohort study of 1.3 million people in Sweden.  BMJ. 2016;352:i1030. doi:10.1136/bmj.i1030PubMedGoogle ScholarCrossref
14.
Ekholm  B, Ekholm  A, Adolfsson  R,  et al.  Evaluation of diagnostic procedures in Swedish patients with schizophrenia and related psychoses.  Nord J Psychiatry. 2005;59(6):457-464. doi:10.1080/08039480500360906PubMedGoogle ScholarCrossref
15.
Song  J, Sjölander  A, Joas  E,  et al.  Suicidal behavior during lithium and valproate treatment: a within-individual 8-year prospective study of 50 000 patients with bipolar disorder.  Am J Psychiatry. 2017;174(8):795-802. doi:10.1176/appi.ajp.2017.16050542PubMedGoogle ScholarCrossref
16.
Molero  Y, Lichtenstein  P, Zetterqvist  J, Gumpert  CH, Fazel  S.  Selective serotonin reuptake inhibitors and violent crime: a cohort study.  PLoS Med. 2015;12(9):e1001875. doi:10.1371/journal.pmed.1001875PubMedGoogle ScholarCrossref
17.
Allison  PD.  Fixed Effects Regression Models. Thousand Oaks, CA: SAGE Publications; 2009. Quantitative Applications in the Social Sciences Series.
18.
Rosendo  AB, Lima  LO, Dal-Pizzol  F, Almeida  S.  Lipid and C-reactive protein levels, cardiovascular disease risk factors and simvastatin treatment in Brazilian individuals.  Inflammation. 2010;33(4):244-250. doi:10.1007/s10753-009-9178-3PubMedGoogle ScholarCrossref
19.
Zhang  YY, Fan  YC, Wang  M, Wang  D, Li  XH.  Atorvastatin attenuates the production of IL-1β, IL-6, and TNF-α in the hippocampus of an amyloid β1-42–induced rat model of Alzheimer’s disease.  Clin Interv Aging. 2013;8:103-110. doi:10.2147/CIA.S40405PubMedGoogle Scholar
20.
Boulton  DW, DeVane  CL, Liston  HL, Markowitz  JS.  In vitro P-glycoprotein affinity for atypical and conventional antipsychotics.  Life Sci. 2002;71(2):163-169. doi:10.1016/S0024-3205(02)01680-6PubMedGoogle ScholarCrossref
21.
Tramontina  AC, Wartchow  KM, Rodrigues  L,  et al.  The neuroprotective effect of two statins: simvastatin and pravastatin on a streptozotocin-induced model of Alzheimer’s disease in rats.  J Neural Transm (Vienna). 2011;118(11):1641-1649. doi:10.1007/s00702-011-0680-zPubMedGoogle ScholarCrossref
22.
Wang  Q, Zengin  A, Ying  W,  et al.  Chronic treatment with simvastatin upregulates muscarinic M1/4 receptor binding in the rat brain.  Neuroscience. 2008;154(3):1100-1106. doi:10.1016/j.neuroscience.2008.04.026PubMedGoogle ScholarCrossref
23.
Wang  Q, Tang  XN, Wang  L,  et al.  Effects of high dose of simvastatin on levels of dopamine and its reuptake in prefrontal cortex and striatum among SD rats.  Neurosci Lett. 2006;408(3):189-193. doi:10.1016/j.neulet.2006.09.009PubMedGoogle ScholarCrossref
24.
Maron  DJ, Fazio  S, Linton  MF.  Current perspectives on statins.  Circulation. 2000;101(2):207-213. doi:10.1161/01.CIR.101.2.207PubMedGoogle ScholarCrossref
25.
Striessnig  J, Koschak  A, Sinnegger-Brauns  MJ,  et al.  Role of voltage-gated L-type Ca2+ channel isoforms for brain function.  Biochem Soc Trans. 2006;34(pt 5):903-909. doi:10.1042/BST0340903PubMedGoogle ScholarCrossref
26.
Agarwal  SM, Ahsan  Z, Lockwood  J,  et al.  S225: a systematic review and meta-analysis of pharmacological interventions for reduction or prevention of weight gain in schizophrenia.  Schizophren Bull. 2018;44(suppl 1):S413. doi:10.1016/j.biopsych.2018.02.1137Google Scholar
27.
Steiner  J, Bernstein  HG, Schiltz  K,  et al.  Immune system and glucose metabolism interaction in schizophrenia: a chicken-egg dilemma.  Prog Neuropsychopharmacol Biol Psychiatry. 2014;48:287-294. doi:10.1016/j.pnpbp.2012.09.016PubMedGoogle ScholarCrossref
28.
Cooper  SJ, Reynolds  GP, Barnes  T,  et al.  BAP guidelines on the management of weight gain, metabolic disturbances and cardiovascular risk associated with psychosis and antipsychotic drug treatment.  J Psychopharmacol. 2016;30(8):717-748. doi:10.1177/0269881116645254PubMedGoogle ScholarCrossref
29.
Osborn  DP, Hardoon  S, Omar  RZ,  et al.  Cardiovascular risk prediction models for people with severe mental illness: results from the Prediction and Management of Cardiovascular Risk in People With Severe Mental Illnesses (PRIMROSE) research program.  JAMA Psychiatry. 2015;72(2):143-151. doi:10.1001/jamapsychiatry.2014.2133PubMedGoogle ScholarCrossref
30.
Hippisley-Cox  J, Coupland  C, Brindle  P.  Development and validation of QRISK3 risk prediction algorithms to estimate future risk of cardiovascular disease: prospective cohort study.  BMJ. 2017;357:j2099. doi:10.1136/bmj.j2099PubMedGoogle ScholarCrossref
31.
Šimić  I, Reiner  Ž.  Adverse effects of statins: myths and reality.  Curr Pharm Des. 2015;21(9):1220-1226. doi:10.2174/1381612820666141013134447PubMedGoogle ScholarCrossref
32.
Eisenberg  MJ, Brox  A, Bestawros  AN.  Calcium channel blockers: an update.  Am J Med. 2004;116(1):35-43. doi:10.1016/j.amjmed.2003.08.027PubMedGoogle ScholarCrossref
33.
Kirpichnikov  D, McFarlane  SI, Sowers  JR.  Metformin: an update.  Ann Intern Med. 2002;137(1):25-33. doi:10.7326/0003-4819-137-1-200207020-00009PubMedGoogle ScholarCrossref
34.
Dalman  Ch, Broms  J, Cullberg  J, Allebeck  P.  Young cases of schizophrenia identified in a national inpatient register: are the diagnoses valid?  Soc Psychiatry Psychiatr Epidemiol. 2002;37(11):527-531. doi:10.1007/s00127-002-0582-3PubMedGoogle ScholarCrossref
35.
Jörgensen  L, Ahlbom  A, Allebeck  P, Dalman  C.  The Stockholm Non-Affective Psychoses Study (SNAPS): the importance of including out-patient data in incidence studies.  Acta Psychiatr Scand. 2010;121(5):389-392. doi:10.1111/j.1600-0447.2009.01500.xPubMedGoogle ScholarCrossref
×