RIVER‐PCI 2016.

Methods	Study design: parallel‐group trial Total study duration: about 3.5 years Duration of follow‐up: mean of 643 days Method of randomisation: interactive web‐based block randomisation system (block sizes of 10), with randomisation stratified by diabetes history (presence versus absence) and acute coronary syndrome presentation (acute versus non‐acute) Method of concealment of allocation: not described in enough detail, it is just mentioned that investigators and patients were masked to treatment allocation Blinding: double‐blind, referred to participants, clinicians ("masked to treatment allocation"), data collectors (independent Data Safety Monitoring Board, independent on‐site clinical monitors, independent angiographic core laboratory), outcome adjudicators (independent Clinical Endpoint Committee) and data analysts (independent statistical data analysis group, Duke Clinical Research Institute for quality of life and economic analyses) Power calculation: 85% power using a 2‐sided log‐rank test at the 5% significance level, with regard to the primary end point events Phases of the study: 1 (treatment phase, including a 7‐day run‐in period) Number of patients randomised: 2651 (1319/1332 for placebo/ranolazine group) Exclusions post‐randomisation: 32 exclusions in the placebo group (19 not treated, 3 scientific misconduct, 10 no qualifying PCI), 15 exclusions in the ranolazine group (7 not treated, 3 scientific misconduct, 5 no qualifying PCI). Additionally, for the quality of life sub‐study, there were 105 exclusions in the placebo group (97 questionnaires invalid, 8 questionnaires not done) and 110 exclusions in the ranolazine group (103 questionnaires invalid, 7 questionnaires not done) Withdrawals (and reasons): 463/1287 ‐ 529/1317 for placebo ‐ ranolazine group, reasons detailed in the appendix of the study report
Participants	Total number: 2651 Country of enrolment: 15 countries (Europe, Israel, Russia, USA) Setting/location: inpatient and outpatient Diagnostic criteria (stable angina pectoris): symptoms of stable angina Comorbidities: incomplete revascularisation (ICR) post‐PCI Age (mean): 63.3±10 / 63.3±10.4 years for placebo/ranolazine group Gender (male %): 80.3% / 79.7% for placebo/ranolazine group Inclusion criteria: Men and women aged ≥18 years History of chronic angina, defined as ≥ 2 episodes of angina pain or discomfort in the chest, jaw, shoulder, back, neck, or arm that is precipitated by exertion or emotional stress and relieved by rest or sublingual nitroglycerin, occurring on ≥ 2 separate days and ≥14 d before PCI (in the case of staged PCI procedures, a history of angina has to have occurred at least 14 days before the first PCI in the series) PCI for any indication (ACS or non‐ACS) Evidence of ICR post‐PCI. ICR is defined as the presence of ≥ 1 lesion with visually estimated ≥50% diameter stenosis in any coronary artery (including branch vessels) with reference vessel diameter ≥2.0 mm, whether in the target vessel or in a non‐target vessel. In the case of a participant post‐CABG, ICR is defined as the presence of ≥ 1 lesion with visually estimated ≥ 50% diameter stenosis in a non bypassed epicardial vessel ≥ 2.0 mm in diameter, or ≥1 visually estimated ≥ 50% diameter stenosis in a bypass graft supplying an otherwise non revascularised myocardial territory Clinically stable post‐PCI. Participants randomised in hospital on the day of planned discharge or in clinic are considered stable. Participants randomised in hospital before the day of planned discharge must meet all of the following criteria: CK‐MB < 3 times the upper limit of normal (ULN) ≥3 hours after PCI, or with evidence of decreasing CK‐MB (by at least 20% from the prior measurement) if ≥3 times the ULN, each as reported by local laboratory. If CK‐MB is not available, the participant must have evidence of normal or decreasing troponin levels (by at least 20% from the prior measurement) ≥3 hours after PCI, as reported by local laboratory Systolic blood pressure ≥ 90 mm Hg and not receiving pressors or inotropes No current requirement for an intra‐aortic balloon pump or any left ventricular assist device No current requirement for intravenous (IV) nitroglycerin Women of childbearing potential must have a negative pregnancy test result at screening (unless surgically sterile or postmenopausal) and must agree to use highly effective contraception methods from screening throughout the duration of study treatment and for 14 d after the last dose of study drug Ability and willingness to comply with all study procedures during the course of the study Exclusion criteria: Any future planned revascularisation (including staged procedures) or possible planned revascularisation (e.g. planned stress test to assess the imminent need for additional revascularisation). Future planned stress tests for purposes of monitoring are permitted but strongly discouraged. Participants may be enrolled after the last PCI in the staged series or once a decision is made not to perform a follow‐up PCI, as long as randomisation occurs within 14 d from the last PCI. If a participant has had a stress test after PCI and before randomisation and no further intervention is planned, the participant may be enrolled within 14 days from the last PCI. Unrevascularised left main coronary artery lesion with diameter stenosis ≥ 50%. Participants with a history of CABG to the left coronary system will be considered to have a revascularised left main if at least 1 graft is patent. Major complication during or after the index PCI (in the case of staged PCI, the last in the series) including any of the following: TIMI major bleeding or any bleeding requiring blood transfusion of ≥ 2 units of red blood cells Coronary perforation requiring treatment Procedural complication requiring surgery (including CABG or peripheral vascular surgery) Stroke within 90 days before randomisation or any history of stroke with permanent major neurologic disability Cardiogenic shock within 90 d before randomisation (transient decreases in blood pressure without clinical sequelae are not considered to be cardiogenic shock) New York Heart Association class III or IV heart failure Severe renal insufficiency as defined by an estimated GFR 30 mL/min per 1.73 m² using the 4 variable modification of diet in renal disease equation (based on the last available measurement before randomisation, collected within 1 mo before the index PCI [or in the case of staged PCI, the last in the series]) Liver cirrhosis Use of class Ia, Ic, or class III anti‐arrhythmic agents, except for amiodarone Current treatment with strong inhibitors of CYP3A Current treatment with cytochrome P450 3A4 inducers or P‐gp inducers Participants taking > 20 mg simvastatin daily or > 40 mg lovastatin daily who cannot reduce the dose to 20 mg once daily for simvastatin or 40 mg once daily for lovastatin, or who cannot switch to another statin Participants taking > 1000 mg daily of metformin who cannot reduce the dose to a maximum total of 1000 mg daily (additional antidiabetic medications may be added as clinically indicated to allow participants to decrease their metformin dose and maintain glycaemic control) Previous treatment with ranolazine for > 7 consecutive days within 30 d before randomisation, or known hypersensitivity or intolerance to ranolazine or to any of the excipients Participation in another investigational drug or investigational device study within 30 d before randomisation (participation in registries is allowed) Women who are pregnant or breast‐feeding Non–coronary artery disease–related comorbid conditions (e.g. advanced malignancy, severe aortic stenosis), which are likely to result in death within 2 years of randomisation
Interventions	Number of intervention groups: 2 Concomitant medications: per the discretion of the investigator Excluded medications: those mentioned in exclusion criteria Placebo group Intervention: placebo Duration of intervention: mean (IQR) of 642 (575‐561) days Ranolazine group Intervention: ranolazine (type of formulation not specified) 1000 mg twice daily Duration of intervention: mean (IQR) of 644 (575‐757) days
Outcomes	Total number of outcomes: 1 According to study protocol: 14 (time from randomisation to the occurrence of ischemia‐driven revascularisation/hospitalisation, sudden cardiac death, cardiovascular death, myocardial infarction; all‐cause mortality; incidence of major adverse cardiovascular events: stroke, transient ischaemic attack, hospitalisation for heart failure; SAQ score, DASI score, symptoms by Rose Dyspnea Scale score (quality of life sub‐study);cumulative total US medical costs, health care costs and resource use, cost per life‐year added and cost per quality‐adjusted life‐year added (health economics sub study)) Reported: 16 (including incidence of ischemia‐driven revascularisation/hospitalisation, sudden cardiac death, cardiovascular death, myocardial infarction, missing time to event for sudden cardiac death, cardiovascular death and myocardial infarction) Cardiovascular mortality Outcome definition: death from cardiac disease, stroke, pulmonary embolism (in the absence of conditions such as malignancy), peripheral artery disease or cardiovascular intervention/surgery Method and unit of measurement: absolute frequency Time points reported: overall study duration All‐cause mortality Outcome definition: number of deaths Method and unit of measurement: absolute frequency Time points reported: overall study duration Quality of life 1. Seattle Angina Questionnaire (SAQ) Outcome definition: includes 5 domains: angina frequency, angina stability, angina‐related treatment satisfaction, angina‐related physical functioning and QOL Upper and lower limits and whether a high or low score is good: each domain is scored separately from 0 to 100, with higher scores indicating better health status Method and unit of measurement: score for each domain Time points reported: 1, 6, 12 months 2. Duke Activity Satuts Index (DASI) Outcome definition:12‐item scale which focuses on physical activity ranging from self‐care to strenuous physical work; each activity is weighted by the metabolic output associated with its performance, and a final score weights the performed activities Upper and lower limits and whether a high or low score is good: score ranges from 0 (worst) to 58.2 (best) Method and unit of measurement: score Time points reported: 1, 6, 12 months 3. Mental Health Inventory‐5 (MHI‐5) Outcome definition: 5‐question scale derived from the 36‐item Short Form Health Survey (SF‐36) version 2.0 Upper and lower limits and whether a high or low score is good: not described Method and unit of measurement: score Time points reported: 1, 6, 12 months 4. European QOL Five Dimension Three‐Level Scale (EuroQOL‐5D‐3 L) Outcome definition: 5‐item instrument assessing specific domains of mobility, self‐care, usual activities, pain/discomfort, and anxiety/depression Upper and lower limits and whether a high or low score is good: not described Method and unit of measurement: score Time points reported: 1, 6, 12 months 5. Rose Dyspnea Scale (RDS) Outcome definition: not described Upper and lower limits and whether a high or low score is good: not described Method and unit of measurement: score Time points reported: 1, 6, 12 months Acute myocardial infarction incidence (fatal and non‐fatal) Outcome definition: episodes defined by symptoms suggestive of ischaemia/infarction in association with ECG, cardiac biomarker, or pathologic evidence of infarction Method and unit of measurement: absolute frequency Time points reported: overall study duration Need for revascularisation procedure Outcome definition: any PCI or CABG surgery occurring after randomisation for angina or angina equivalent symptoms, with or without documented ischaemia. PCI is defined as an attempt to cross a lesion with a wire with the intention of performing revascularisation Method and unit of measurement: absolute frequency Time points reported: overall study duration RESULTS Cardiovascular mortality Sample size: 2604 (intention‐to‐treat analysis) Missing participants: 47 (32/15 for placebo/ranolazine group) Summary data: 20/1287 ‐ 21/1317 for placebo ‐ ranolazine group Subgroup analyses: not performed All‐cause mortality Sample size: 823 (intention‐to‐treat analysis) Missing participants: 32 (22/10 for placebo/ranolazine group) Summary data: 36/1297 ‐ 42/1322 for placebo ‐ ranolazine group Subgroup analyses: not performed Quality of life 1182 1207 1. Seattle Angina Questionnaire (SAQ) Sample size: 1958 (980/978 for the placebo/ranolazine group) (intention‐to‐treat analysis) Missing participants: 202/229 for placebo/ranolazine group Summary data: baseline/12‐month (mean ± SD) score in the QOL domain: 49.5 ± 22.8 / 70.4 ± 22.2 for placebo group, 48.3 ± 22.3 / 70.3 ± 22.5 for ranolazine group Subgroup analyses: performed for the following variables: age, sex, indication for the qualifying PCI, baseline anti‐anginal use, diabetes mellitus, baseline angina 2. Duke Activity Satuts Index (DASI) Sample size: 1957 (980/977 for the placebo/ranolazine group) (intention‐to‐treat analysis) Missing participants: 202/230 for the placebo/ranolazine group Summary data: baseline/12‐month (mean±SD) score: 18.7±14.3 / 23.3±16.1 for placebo group, 18.7±14.9 / 22.5±15.8 for ranolazine group Subgroup analyses: performed for the following variables: age, sex, indication for the qualifying PCI, baseline anti‐anginal use, diabetes mellitus, baseline angina 3. Mental Health Inventory‐5 (MHI‐5) Sample size: 1954 (978/976 for the placebo/ranolazine group) (intention‐to‐treat analysis) Missing participants: 204/231 for placebo/ranolazine group Summary data: baseline/12‐month (mean±SD) score: 64.1 ± 19.6 / 70.4 ± 18.0 for placebo group, 64.9±19.2 / 70.3 ± 18.1 for ranolazine group Subgroup analyses: performed for the following variables: age, sex, indication for the qualifying PCI, baseline anti‐anginal use, diabetes mellitus, baseline angina 4. European QOL Five Dimension Three‐Level Scale (EuroQOL‐5D‐3 L) Sample size: 1934 (973/961 for the placebo/ranolazine group) (intention‐to‐treat analysis) Missing participants: 209/246 for placebo/ranolazine group Summary data: baseline/12‐month (mean±SD) score: 0.75 ± 0.23 / 0.78 ± 0.23 for placebo group, 0.75 ± 0.23 / 0.79 ± 0.22 for ranolazine group Subgroup analyses: performed for the following variables: age, sex, indication for the qualifying PCI, baseline anti‐anginal use, diabetes mellitus, baseline angina 5. Rose Dyspnea Scale (RDS) Sample size: 1939 (971/968 for the placebo/ranolazine group) (intention‐to‐treat analysis) Missing participants: 211/239 for placebo/ranolazine group Summary data: baseline/12‐month (mean ± SD) score: 1.7 ± 1.4 / 1.0 ± 1.3 for placebo group, 1.7±1.4 / 1.1 ± 1.3 for ranolazine group Subgroup analyses: performed for the following variables: age, sex, indication for the qualifying PCI, baseline anti‐anginal use, diabetes mellitus, baseline angina Acute myocardial infarction incidence (fatal and non‐fatal) Sample size: 2604 (intention‐to‐treat analysis) Missing participants: 47 (32/15 for placebo/ranolazine group) Summary data: 116/1287 ‐ 111/1317 for placebo ‐ ranolazine group Subgroup analyses: not performed Need for revascularisation procedure Sample size: 2604 (intention‐to‐treat analysis) Missing participants: 47 (32/15 for placebo/ranolazine group) Summary data: 200/1287 ‐ 201/1317 for placebo ‐ ranolazine group Subgroup analyses: performed for the following variables: sex, age, precedence, diabetes mellitus, indication for PCI, type of vessel disease, residual SYNTAX score, type of PCI device, total occlusion, previous CABG, baseline BNP and baseline LVEF Adverse events Dizziness, constipation, nausea, hypotension, vomiting and vertigo were reported more often in the ranolazine group than in the placebo group.
Notes	Relevant observations for the data provided before: none Source of funding: Gilead Sciences, Menarini Group Notable conflicts of interest: seven of the authors declare current of past financial relationships with Gilead Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random sequence. Random sequence generated by an interactive web‐based block randomisation system with block sizes of ten
Allocation concealment (selection bias)	Low risk	Described for participants and clinicians as "masked to treatment allocation". The use of a web‐based system for randomization (and allocation) can be considered sufficient to mantain blinded the study personnel until the momment of assignment.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Treatment phase is declared to be double‐blinded, but no description for participants and personnel is provided. However, the subjects idea about study arm was measured, and study data was collected by independent groups, then presumably both participants and investigators were blinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Data for events and quality of life/economic analyses were collected by independent outcome adjudication groups
Incomplete outcome data (attrition bias) All outcomes	Low risk	Exclusions and withdrawals are described, and reason are reported
Selective reporting (reporting bias)	High risk	There are two sub studies (quality of life and health economics) besides the main study for the RIVER‐PCI trial, and all the main endpoints considered in the protocol were reported (the health economics sub study has not yet been published). Of note, for the events under study, time to event occurrence was stated to be the endpoint rather than the rate/incidence of the event; however, results for the secondary endpoint events were reported only as rate/incidence, and no data about time to event occurrence was provided
Other bias	High risk	The study was supported by Gilead Sciences and Menarini Group. Seven of the authors declare current or past financial relationships with Gilead Sciences.