This assessment updates and expands on two previous technology assessments that evaluated implantable cardioverter defibrillators (ICDs) for arrhythmias and cardiac resynchronisation therapy (CRT) for heart failure (HF). To assess the clinical effectiveness and cost-effectiveness of ICDs in addition to optimal pharmacological therapy (OPT) for people at increased risk of sudden cardiac death (SCD) as a result of ventricular arrhythmias despite receiving OPT; to assess CRT with or without a defibrillator (CRT-D or CRT-P) in addition to OPT for people with HF as a result of left ventricular systolic dysfunction (LVSD) and cardiac dyssynchrony despite receiving OPT; and to assess CRT-D in addition to OPT for people with both conditions. Electronic resources including MEDLINE, EMBASE and The Cochrane Library were searched from inception to November 2012. Additional studies were sought from reference lists, clinical experts and manufacturers' submissions to the National Institute for Health and Care Excellence. Inclusion criteria were applied by two reviewers independently. Data extraction and quality assessment were undertaken by one reviewer and checked by a second. Data were synthesised through narrative review and meta-analyses. For the three populations above, randomised controlled trials (RCTs) comparing (1) ICD with standard therapy, (2) CRT-P or CRT-D with each other or with OPT and (3) CRT-D with OPT, CRT-P or ICD were eligible. Outcomes included mortality, adverse events and quality of life. A previously developed Markov model was adapted to estimate the cost-effectiveness of OPT, ICDs, CRT-P and CRT-D in the three populations by simulating disease progression calculated at 4-weekly cycles over a lifetime horizon. A total of 4556 references were identified, of which 26 RCTs were included in the review: 13 compared ICD with medical therapy, four compared CRT-P/CRT-D with OPT and nine compared CRT-D with ICD. ICDs reduced all-cause mortality in people at increased risk of SCD, defined in trials as those with previous ventricular arrhythmias/cardiac arrest, myocardial infarction (MI) > 3 weeks previously, non-ischaemic cardiomyopathy (depending on data included) or ischaemic/non-ischaemic HF and left ventricular ejection fraction ≤ 35%. There was no benefit in people scheduled for coronary artery bypass graft. A reduction in SCD but not all-cause mortality was found in people with recent MI. Incremental cost-effectiveness ratios (ICERs) ranged from £14,231 per quality-adjusted life-year (QALY) to £29,756 per QALY for the scenarios modelled. CRT-P and CRT-D reduced mortality and HF hospitalisations, and improved other outcomes, in people with HF as a result of LVSD and cardiac dyssynchrony when compared with OPT. The rate of SCD was lower with CRT-D than with CRT-P but other outcomes were similar. CRT-P and CRT-D compared with OPT produced ICERs of £27,584 per QALY and £27,899 per QALY respectively. The ICER for CRT-D compared with CRT-P was £28,420 per QALY. In people with both conditions, CRT-D reduced the risk of all-cause mortality and HF hospitalisation, and improved other outcomes, compared with ICDs. Complications were more common with CRT-D. Initial management with OPT alone was most cost-effective (ICER £2824 per QALY compared with ICD) when health-related quality of life was kept constant over time. Costs and QALYs for CRT-D and CRT-P were similar. The ICER for CRT-D compared with ICD was £27,195 per QALY and that for CRT-D compared with OPT was £35,193 per QALY. Limitations of the model include the structural assumptions made about disease progression and treatment provision, the extrapolation of trial survival estimates over time and the assumptions made around parameter values when evidence was not available for specific patient groups. In people at risk of SCD as a result of ventricular arrhythmias and in those with HF as a result of LVSD and cardiac dyssynchrony, the interventions modelled produced ICERs of < £30,000 per QALY gained. In people with both conditions, the ICER for CRT-D compared with ICD, but not CRT-D compared with OPT, was < £30,000 per QALY, and the costs and QALYs for CRT-D and CRT-P were similar. A RCT comparing CRT-D and CRT-P in people with HF as a result of LVSD and cardiac dyssynchrony is required, for both those with and those without an ICD indication. A RCT is also needed into the benefits of ICD in non-ischaemic cardiomyopathy in the absence of dyssynchrony. This study is registered as PROSPERO number CRD42012002062. The National Institute for Health Research Health Technology Assessment programme.