Spontaneous spinal cerebrospinal fluid (CSF) leaks are a major cause of spontaneous intracranial hypotension. Minimally invasive surgical (MIS) approaches have been developed to reduce perioperative morbidity while preserving repair durability, yet their comparative effectiveness remains incompletely defined. We performed a systematic review and meta-analysis of contemporary surgical series following PRISMA 2020 guidelines. Twenty studies met criteria for qualitative synthesis; nine contributed to the quantitative synthesis. Primary closure success was 95% (95% CI 88.8–97.8; I 2 = 0%), and recurrence or reoperation occurred in 10% (95% CI 5.9–15.8; I 2 = 0%). Subgroup analysis showed numerically similar rates between MIS and open repair in both closure (97% vs 92%; p > 0.05) and recurrence (10% vs 9%; p = 0.86), though formal non-inferiority testing was not appropriate because the available evidence consisted predominantly of nonrandomized retrospective cohorts without a prespecified non-inferiority margin or direct adjusted between-group comparisons. Complete symptom resolution was 71% in open cohorts (I 2 = 87%); MIS cohorts lacked consistent denominators for quantitative pooling. Some cohorts reported shorter length of stay and reduced postoperative pain in MIS groups, but these between-study comparisons were descriptive and confounded by case selection. However, the observational nature of the data, absence of randomization, and unmeasured selection bias—whereby more complex leaks may have been preferentially treated with open surgery—preclude definitive comparative conclusions. In selected cases, MIS dural repair may approach the success rates of open procedures with comparable complication profiles, but prospective studies controlling for selection bias are needed. • Systematic review/meta-analysis of surgery for spontaneous spinal CSF leaks. • Pooled primary closure was 95%, and pooled recurrence/reoperation was 10%. • MIS cohorts lacked extractable symptom-resolution denominators for pooling. • In selected cases, MIS outcomes approached open repair; selection bias remains.
Yevudza et al. (Sun,) studied this question.