Psychiatric symptom improvement from adjunctive statin prescribing in severe mental illness: three target trial emulation studies

BackgroundRandomized controlled trials (RCTs) of statins as adjunct therapy for severe mental illness (SMI) have produced mixed results. Specific statin-antipsychotic combinations might improve psychiatric symptoms through: 1) blood-brain barrier (BBB) penetrant statins being anti-inflammatory/neuroprotective, and/or 2) statins that inhibit p-glycoprotein enhancing the effects of antipsychotics with high p-glycoprotein affinity. AimTo investigate these mechanisms via three target trial emulation studies. MethodsWe identified patients with SMI (schizophrenia, bipolar disorder, other psychoses) prescribed antipsychotics/mood stabilisers and statins from 2000-2019 in English electronic health records. We defined three hypothetical RCTs and their observational analogues: 1) simvastatin (crosses BBB) vs. atorvastatin/pravastatin/rosuvastatin (non-penetrant); 2A) simvastatin/atorvastatin (p-glycoprotein inhibitors) vs. pravastatin in patients prescribed risperidone/olanzapine/aripiprazole (high p-glycoprotein affinity); 2B) risperidone/olanzapine/aripiprazole vs. quetiapine (low p-glycoprotein affinity) in patients prescribed simvastatin/atorvastatin. Primary outcome: 12-month psychiatric admissions. Secondary outcomes: self-harm, physical health, and accident/injury admissions. ResultsIn 72,096 patients prescribed statins and antipsychotics/mood stabilisers, we found no reduction in psychiatric admissions at 12 months in patients prescribed: 1) BBB-penetrant vs. non-penetrant statins (HR:1.07; 95%CI:0.88-1.31); 2A) antipsychotics with p-glycoprotein affinity and p-glycoprotein inhibiting statins vs. statins without inhibition (HR:0.77; 95%CI:0.28-2.15); 2B) p-glycoprotein inhibiting statins with antipsychotics having p-glycoprotein affinity vs. antipsychotics without affinity (HR:0.93; 95%CI:0.79-1.09). In 2B we observed reduced self-harm (HR:0.60; 95%CI:0.38-0.97) in per-protocol analysis and reduced psychiatric admissions in the other psychoses subgroup (HR:0.53; 95%CI:0.34-0.85). ConclusionsBBB permeability is unlikely to be the mechanism by which statins improve SMI symptoms. Further research is needed to understand statin-antipsychotic interactions, and whether interaction with p-glycoprotein is a plausible mechanism.