Evaluating the Use of Pitavastatin to Reduce the Risk of Cardiovascular Disease in HIV-Infected Adults

Cardiac ischemia or myocardial infarction component of the primary composite MACE outcome. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Deaths (without preceding event of interest) were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Cerebrovascular event (stroke or TIA) component of the primary composite MACE outcome. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Deaths (without preceding event of interest) were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Peripheral arterial ischemia component of the primary composite MACE outcome. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Deaths (without preceding event of interest) were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

CV death component of the primary composite MACE outcome. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Non-CV deaths and deaths from undetermined causes were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

CV or undetermined death component of the primary composite MACE outcome. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Non-CV deaths were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Cardiac cardiac catheterization or revascularization component of the primary composite MACE outcome. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Deaths (without preceding event of interest) were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Carotid or cerebrovascular revascularization component of the primary composite MACE outcome. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Deaths (without preceding event of interest) were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Peripheral arterial revascularization component of the primary composite MACE outcome. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Deaths (without preceding event of interest) were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

A composite outcome including MACE and death from any cause. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Death from any cause. The incidence rates were estimated based on time to event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

A composite of non-CV clinical diagnoses including: non-AIDS-defining cancers (excluding basal cell and squamous cell carcinomas of the skin), AIDS-defining events (based on Centers for Disease Control and Prevention [CDC] 2014 classification), end-stage renal disease, and end-stage liver disease. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Deaths (without preceding event of interest) were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Non-AIDS-defining cancer (excluding basal cell and squamous cell carcinomas of the skin) component of the composite non-CV clinical diagnoses outcome. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Deaths (without preceding event of interest) were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

AIDS-defining event component of the composite non-CV clinical diagnoses outcome. Events were captured based on the Centers for Disease Control and Prevention [CDC] 2014 classification. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Deaths (without preceding event of interest) were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

End-stage renal disease (defined as initiation of dialysis or renal transplantation) component of the composite non-CV clinical diagnoses outcome. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Deaths (without preceding event of interest) were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

End-stage liver disease (defined as cirrhosis or hepatic decompensation requiring hospitalization) component of the composite non-CV clinical diagnoses outcome. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated via cause-specific relative hazard (i.e. hazard ratio) of prescribed pitavastatin compared to placebo from Cox proportional hazards models, stratified by screening CD4 count and sex. Deaths (without preceding event of interest) were treated as competing events and censored. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Safety analysis outcome measure of non-fatal serious adverse event was defined by International Conference on Harmonisation (ICH) criteria. Fatal events were excluded as deaths were a secondary efficacy outcome (see outcome measure: incidence rate of death (all-cause)). The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated using incidence rate ratios from Poisson regression models (prescribed pitavastatin compared to placebo), adjusted for screening CD4 and sex. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Safety analysis outcome measure of diabetes was defined as new diagnosis of diabetes with initiation of anti-diabetic agent. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated using incidence rate ratios from Poisson regression models (prescribed pitavastatin compared to placebo), adjusted for screening CD4 and sex. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Safety analysis outcome measure of myalgia, muscle weakness or myopathy which were grade 3 or higher or treatment-limiting. Grade 3 or higher includes grade 3 and 4 events, where grade 3 refers to severe and grade 4 to life-threatening according to DAIDS AE Grading Table (version 2.1). The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated using incidence rate ratios from Poisson regression models (prescribed pitavastatin compared to placebo), adjusted for screening CD4 and sex. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Safety analysis outcome measure of rhabdomyolysis which was grade 3 or higher or treatment-limiting. Grade 3 or higher includes grade 3 and 4 events, where grade 3 refers to severe and grade 4 to life-threatening, according to DAIDS AE Grading Table (version 2.1). The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated using incidence rate ratios from Poisson regression models (prescribed pitavastatin compared to placebo). Due to small number of events, there was no adjustment for screening CD4 and sex. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Safety analysis outcome measure of Grade 3 or higher alanine transaminase (ALT). Grade 3 or higher includes grade 3 and 4 events, where grade 3 refers to severe and grade 4 to life-threatening, according to DAIDS AE Grading Table (version 2.1). The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated using incidence rate ratios from Poisson regression models (prescribed pitavastatin compared to placebo). Due to small number of events, there was no adjustment for screening CD4 and sex. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Safety analysis outcome measure of any AE. AE collection included events of grade ≥3, those that were serious (defined by International Conference on Harmonisation (ICH) criteria) or treatment-limiting, and targeted diagnosis of diabetes. Grade ≥3 includes events that were grade 3 (serious) or grade 4 (life-threatening) per DAIDS AE Grading Table (version 2.1). Fatal events were excluded as deaths were a secondary efficacy outcome (see outcome measure: incidence rate of death (all-cause)). The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated using incidence rate ratios from Poisson regression models (prescribed pitavastatin compared to placebo), adjusted for screening CD4 and sex. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

LDL-C level was derived as LDL-C calculated according to the Friedewald formula at triglycerides ≤400 mg/dL, and direct LDL-C at triglycerides >400 to <500 mg/dL. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Non-HDL cholesterol levels were calculated as total cholesterol minus HDL cholesterol. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Serious COVID-19 was defined as COVID-19 that resulted in hospitalization or death or was life-threatening as per the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use Guideline E2A definition. The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated using incidence rate ratios from Poisson regression models (prescribed pitavastatin compared to placebo), adjusted for GBD region to account for regional differences. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

COVID-19 was defined as COVID-19 clinical diagnosis or positive test result (SARS-CoV-2 PCR or rapid antigen tests). The incidence rates were estimated based on time to the first event using Poisson distribution, with follow-up time censored at last contact. The treatment effect was estimated using incidence rate ratios (prescribed pitavastatin compared to placebo) from Poisson regression models, adjusted for GBD region to account for regional differences. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

NCP was defined as plaque voxels with attenuation of <350. Change in NCP is expressed as absolute change from baseline (calculated as NCP volume at 2 years minus NCP volume at entry), based on quantitative read of the CT scan, whenever available. Participants without a quantitative read and no evidence of NCP based on the corresponding qualitative read were assigned a value of zero for the change. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Progression at Year 2 was defined as any progression/increase in NCP volume in participants with evidence of NCP at entry, or incident NCP in participants without evidence of NCP at entry. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Total plaque includes all plaque voxels (noncalcified + calcified). Change in total plaque volume is expressed as absolute change from baseline (calculated as volume at 2 years minus volume at entry). Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Level of inflammatory biomarker lipoprotein-associated phospholipase A2 (LpPLA2). Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Change in inflammatory biomarker LpPLA2 from baseline calculated as value at month 24 minus value at entry. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Level of inflammatory marker high-sensitivity C-reactive protein (HsCRP). Censored values below or above the assay limit were imputed. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Change in inflammatory biomarker hsCRP from baseline calculated as value at month 24 minus value at entry. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Level of immune biomarker soluble CD163. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).

Change in immune biomarker soluble CD163 from baseline calculated as value at month 24 minus value at entry. Treatment discontinuation was ignored, including the initiation of statin therapy as part of clinical care (intention to treat policy).