This systematic review evaluated the clinical effectiveness and safety of subcutaneous implantable cardioverter‐defibrillator (S‐ICD) in patients at an increased risk of sudden cardiac death and with an ICD indication for primary or secondary prevention. A systematic literature search was conducted in four databases (Medline via Ovid, Embase, the Cochrane Library, and HTA‐INAHTA). Randomized controlled trials (RCTs) and controlled observational studies with ≥100 S‐ICD patients and a low to moderate risk of bias were eligible for inclusion. The studies' quality and the available evidence's strength were assessed using the Cochrane risk of bias tool, the ROBINS‐I tool, and the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) approach. One RCT, a post hoc analysis of the RCT (
n
= 849) and four controlled observational studies (
n
= 7149) were included. The quality of the available evidence was graded as low to very low, except for the primary composite endpoint of the RCT, which was rated as moderate quality. After 4 years, the RCT showed that S‐ICD was non‐inferior to TV‐ICD regarding the composite endpoint of inappropriate shocks and device‐related complications (68 [15.1%] vs. 68 [15.7%], hazard ratio [HR] 0.99, 95% confidence interval [CI] [0.71, 1.39], non‐inferiority margin 1.45,
P
= 0.001). The RCT and two observational studies reported statistically significantly fewer lead complications in S‐ICD patients (after 4 years: 1.4% vs. 6.6%, HR 0.24, 95% CI [0.10, 0.54]; after 3 years: 0.3% vs. 2.3%,
P
= 0.03; and after 5 years: 0.8% vs. 11.5%,
P
= 0.03). Identified evidence about appropriate and inappropriate shocks was inconclusive: The RCT detected statistically significantly more appropriate shocks in patients with S‐ICD (83 [19.2%] vs. 57 [11.5%], HR 1.52, 95% CI [1.08, 2.12],
P
= 0.02), whereas one observational study showed a statistically significantly lower rate in the S‐ICD group (9.9%, 95% CI [7.0, 13.9], vs. 13.9%, 95% CI [10.8, 17.8],
P
= 0.003). Regarding inappropriate shocks, one observational study reported statistically significantly higher rates in the S‐ICD cohort (11.9% vs. 7.5%,
P
= 0.007), whereas the RCT and two other observational studies did not detect a statistically significant difference between the treatment groups (
P
> 0.05). None of the included studies showed a statistically significant difference in overall mortality and shock efficacy between patients with S‐ICD and TV‐ICD (
P
> 0.05). The available evidence is insufficient to show the superiority of S‐ICD compared with TV‐ICD, hindering the widespread use of the technology. Results of the recently completed ATLAS trial are to be awaited, and the anticipated role of the S‐ICD needs to be clearly formulated.