Summary OBJECTIVES The optimal revascularization strategy for patients with ischaemic left ventricular systolic dysfunction (iLVSD) remains controversial. We aimed to compare percutaneous coronary intervention (PCI), coronary artery bypass grafting (CABG) and medical therapy (MT) in a network meta-analysis. METHODS All randomized controlled trials and observational studies comparing any combination of PCI, CABG and MT in patients with iLVSD were analysed in a frequentist network meta-analysis (generic inverse variance method). Primary outcome was mortality at longest available follow-up. Secondary outcomes were cardiac death, stroke, myocardial infarction (MI) and repeat revascularization (RR). RESULTS Twenty-three studies were included (n = 23 633; 4 randomized controlled trials). Compared to CABG, PCI was associated with higher mortality [incidence rate ratio (IRR) 1.32, 95% confidence interval (CI) 1.13–1.53], cardiac death (IRR 1.65, 95% CI 1.18–2.33), MI (IRR 2.18, 95% CI 1.70–2.80) and RR (IRR 3.75, 95% CI 2.89–4.85). Compared to CABG, MT was associated with higher mortality (IRR 1.52, 95% CI 1.26–1.84), cardiac death (IRR 3.83, 95% CI 2.12–6.91), MI (IRR 3.22, 95% CI 1.52–6.79) and RR (IRR 3.37, 95% CI 1.67–6.79). Compared to MT, PCI was associated with lower cardiac death (IRR 0.43, 95% CI 0.24–0.78). CABG ranked as the best revascularization strategy for mortality, cardiac death, MI and RR; MT ranked as the strategy associated with the lowest incidence of stroke. Left ventricular ejection fraction, year of study, use of drug-eluting stents did not affect relative treatment effects. CONCLUSIONS CABG appears to be the best therapy for iLVSD, although mainly based on observational data. Definitive randomized controlled trials comparing CABG and PCI in iLVSD are required. PROSPERO registration ID 132414.