Abstract Background Low-volume high-intensity interval training (HIT) appears to be an efficient and practical way to develop physical fitness. Objective Our objective was to estimate meta-analysed mean effects of HIT on aerobic power (maximum oxygen consumption [VO_2max] in an incremental test) and sprint fitness (peak and mean power in a 30-s Wingate test). Data Sources Five databases (PubMed, MEDLINE, Scopus, BIOSIS and Web of Science) were searched for original research articles published up to January 2014. Search terms included ‘high intensity’, ‘HIT’, ‘sprint’, ‘fitness’ and ‘VO_2max’. Study Selection Inclusion criteria were fitness assessed pre- and post-training; training period ≥2 weeks; repetition duration 30–60 s; work/rest ratio <1.0; exercise intensity described as maximal or near maximal; adult subjects aged >18 years. Data Extraction The final data set consisted of 55 estimates from 32 trials for VO_2max, 23 estimates from 16 trials for peak sprint power, and 19 estimates from 12 trials for mean sprint power. Effects on fitness were analysed as percentages via log transformation. Standard errors calculated from exact p values (where reported) or imputed from errors of measurement provided appropriate weightings. Fixed effects in the meta-regression model included type of study (controlled, uncontrolled), subject characteristics (sex, training status, baseline fitness) and training parameters (number of training sessions, repetition duration, work/rest ratio). Probabilistic magnitude-based inferences for meta-analysed effects were based on standardized thresholds for small, moderate and large changes (0.2, 0.6 and 1.2, respectively) derived from between-subject standard deviations (SDs) for baseline fitness. Results A mean low-volume HIT protocol (13 training sessions, 0.16 work/rest ratio) in a controlled trial produced a likely moderate improvement in the VO_2max of active non-athletic males (6.2 %; 90 % confidence limits ±3.1 %), when compared with control. There were possibly moderate improvements in the VO_2max of sedentary males (10.0 %; ±5.1 %) and active non-athletic females (3.6 %; ±4.3 %) and a likely small increase for sedentary females (7.3 %; ±4.8 %). The effect on the VO_2max of athletic males was unclear (2.7 %; ±4.6 %). A possibly moderate additional increase was likely for subjects with a 10 mL·kg⁻¹·min⁻¹ lower baseline VO_2max (3.8 %; ±2.5 %), whereas the modifying effects of sex and difference in exercise dose were unclear. The comparison of HIT with traditional endurance training was unclear (−1.6 %; ±4.3 %). Unexplained variation between studies was 2.0 % (SD). Meta-analysed effects of HIT on Wingate peak and mean power were unclear. Conclusions Low-volume HIT produces moderate improvements in the aerobic power of active non-athletic and sedentary subjects. More studies are needed to resolve the unclear modifying effects of sex and HIT dose on aerobic power and the unclear effects on sprint fitness.