Abstract
OBJECTIVE: The efficacy and safety of minimally invasive nephrolithotomy (MPCNL) versus retrograde intrarenal surgery (RIRS) was assessed for lower calyceal (LC) stones. METHODS: Our team conducted a systematic literature search up to December, 2022, using PUBMED, EMBASE and the Cochrane Library. The study was registered in PROSPERO, CRD 42021247197. Randomized controlled trials evaluating the efficacy and safety of MPCNL versus RIRS for LC stones were collected. Heterogeneity among the studies was assessed using the χ 2 test based on the Q and I2 tests. Pooled effect sizes were calculated using a fixed model if I2 is less than 50%; otherwise, a random-effects model was chosen. The primary outcomes were the 3-month stone-free rate (3SFR) and total complications, while the secondary outcomes were the operating time, hospital stay, haemoglobin reduction, bleeding, postoperative fever and complications with the Clavien-Dindo system. A subgroup analysis of 10-20 mm LC stones was also designed. RESULTS: A total of 7 peer-reviewed trials comprising 711 patients were identified. No statistical differences were observed in the heterogeneity results of the 3SFR or total complications ( P >0.1, I2 < 50%). Compared with RIRS, MPCNL had an unfavourable safety profile, resulting in total complications [odds ratio (OR): 1.87 (95% CI: 1.05, 3.33); P =0.03], haemoglobin reduction [OR: 0.81 (95% CI: 0.15, 1.47); P =0.02] and complications with Grade I [OR: 5.52 (95% CI: 1.34, 22.83); P =0.02] but an improved efficacy and 3SFR [OR: 2.43 (95% CI: 1.48, 3.97); P =0.0004]. As for the 10-20 mm LC stones, compared with RIRS, MPCNL also had an unfavourable safety profile, resulting in total complications [OR: 2.47 (95% CI: 1.20, 5.07); P =0.01], complications with Grade I [OR: 4.97 (95% CI: 0.99, 25.01); P =0.05] and an increased hospital stay [OR: 2.46 (95% CI: 2.26, 2.66); P =0.00001] but an improved efficacy and 3SFR {OR: 3.10 (95% CI: 1.61, 5.99); P =0.0008]. The efficacy effect of MPCNL and safety effect of RIRS were nearly equal for both stones sized less than 20 mm (number needed to treat = 17, number needed to harm = 20) and stones sized 10-20 mm (number needed to treat = 20, number needed to harm = 13). No statistical difference was found between the MPCNL and RIRS groups for the rest of outcomes. CONCLUSION: Both MPCNL and RIRS are safe and effective management methods. Moreover, compared with RIRS, MPCNL had an unfavourable safety profile but improved efficacy for LC stones of ≤20 mm or 10-20 mm, and the differences were statistically significant. The relative profit of efficacy of MPCNL was similar to the relative profit of safety of RIRS.