Efficacy and safety of Shenmai injection and Shenfu injection on postoperative cognitive dysfunction: a systematic review and meta-analysis

Why clinicians should know about it

• Picked for Neurology (clinical) (top studies of the week, 19 July 2026).
• Picked for Pharmacology (medical) (paper of the day, 16 July 2026).

Abstract

BACKGROUND: Postoperative cognitive dysfunction (POCD) is a common neurological complication after surgery, associated with increased adverse events and mortality. While studies suggest that Shenmai injection (SMI) or Shenfu injection (SFI) may be associated with reduce POCD, rigorous systematic reviews specifically on these two botanical drug injections are lacking. METHODS: We comprehensively searched eight databases (PubMed, Embase, Cochrane Library, Web of Science, CNKI, Wanfang, VIP, CBM) from inception to May 2025 for RCTs on SMI and SFI for POCD. Data were analyzed using RevMan 5.4. Primary outcomes were Mini-Mental State Examination (MMSE) score and POCD incidence. Secondary outcomes included serum S100β protein concentration and postoperative consciousness recovery time. Continuous outcomes used mean difference (MD) or standardized mean difference (SMD) with 95% CI; dichotomous outcomes used relative risk (RR) with 95% CI. This study was registered in INPLASY (registration number: 202470131). RESULTS: A total of 552 records were identified, and after deduplication and screening, sixteen RCTs (N = 1,199 patients) were finally included. Both SMI and SFI were associated with statistically significant higher postoperative MMSE scores at multiple time points (all P < 0.01) and were associated with lower POCD incidence at 3 days (SMI RR 0.38; SFI RR 0.52) and 7 days (SMI RR 0.26; SFI RR 0.47). However, all included trials were at high risk of bias (lack of allocation concealment and blinding). SFI was associated with lower serum S100β at 24 h (SMD = -0.40); the effect of SMI could not be evaluated due to missing data. Both injections were associated with shorter consciousness recovery time (MD = -5.67 min, 95% CI: -6.46 to -4.87, P < 0.001). Subgroup analyses showed a trend toward reduced POCD incidence across doses. Most studies did not report detailed preparation composition or chemical characterization. Only three of the 16 included RCTs reported adverse events, and therefore no reliable safety conclusions can be drawn from the current RCT evidence. Sensitivity analysis confirmed robustness. GRADE evidence quality was low for all outcomes. CONCLUSION: Low-certainty evidence suggests a possible association of SMI or SFI with improved cognitive outcomes, but causality is unproven. These preliminary findings require confirmation in rigorous RCTs with proper blinding and standardized measures. SYSTEMATIC REVIEW REGISTRATION: https://inplasy.com/inplasy-2024-7-0131, identifier 202470131.

Abstract as published, via PubMed.