Purpose This meta-analysis compares open reduction and internal fixation with a plate (ORIF) versus nailing for humeral shaft fractures with regard to union, complications, general quality of life and shoulder/elbow function. Methods PubMed/Medline/Embase/CENTRAL/CINAHL was searched for observational studies and randomised clinical trials (RCT). Effect estimates were pooled across studies using random effects models. Results were presented as weighted odds ratio (OR) or risk difference (RD) with corresponding 95% confidence interval (95% CI). Subgroup analysis was performed stratified for study design (RCTs and observational studies). Results Eighteen observational studies (4906 patients) and ten RCT's (525 patients) were included. The pooled effect estimates of observational studies were similar to those obtained from RCT's. More patients treated with nailing required re-intervention (RD 2%; OR 2.0, 95% CI 1.0-3.8) with shoulder impingement being the most predominant indication (17%). Temporary radial nerve palsy secondary to operation occurred less frequently in the nailing group (RD 2%; OR 0.4, 95% CI 0.3-0.6). Notably, all but one of the radial nerve palsies resolved spontaneously in each groups. Nailing leads to a faster time to union (mean difference − 1.9 weeks, 95% CI − 2.9 to − 0.9), lower infection rate (RD 2%; OR 0.5, 95% CI 0.3-0.7) and shorter operation duration (mean difference − 26 min, 95% CI − 37 to − 14). No differences were found regarding nonunion, general quality of life, functional shoulder scores, and total upper extremity scores. Conclusion Nailing carries a lower risk of infection, postoperative radial nerve palsy, has a shorter operation duration and possibly a shorter time to union. Shoulder impingement requiring re-intervention, however, is an inherent disadvantage of nail fixation. Notably, absolute differences are small and almost all patients with radial nerve palsy recovered spontaneously. Satisfactory results can be achieved with both treatment modalities.