Four relevant studies were identified: two RCTs (n=64, range 25 to 39) and two CCTs (n=312, range 52 to 260). Both RCTs had power calculations, adequate recruitment and statistics, and good external validity, but no blinding, no intention-to-treat analysis. Both RCTs had 11% post randomisation exclusions. Only one RCT reported allocation concealment. There was potential bias in allocation to groups in both CCTs. Both CCTs had no blinding and there were some baseline differences between the trial groups. Both CCTs gave details of confounders, adverse effects and statistics used.
Meta-analysis of the RCTs gave a significant reduction in both primary length of stay and total length of stay for the enhanced recovery intervention compared to controls (WM -3.64 days, 95% CI -4.98 to -2.29, I2=0% and WM -3.75 days, 95% CI -5.11 to -2.40, I2=0%). Meta-analysis was not carried out for the CCTs, which only reported postoperative length of stay data.
Meta-analysis of the RCTS for morbidity gave no significant difference between the two groups. The pooled analysis of the CCTs showed a significant reduction in morbidity with enhanced recovery (RR 0.44, 95% CI 0.32 to 0.61, I2=0%). There was no significant difference in mortality found with the pooled analyses of either RCTs or CCTs for enhanced recovery compared to controls.
Only one RCT reported any 30-day readmissions and there was no significant difference between the control and intervention groups. The pooled analysis of the CCTs for 30-day readmissions found a significantly higher number of readmissions with enhanced recovery compared with controls (RR 1.73, 95% CI 1.00 to 3.01, I2=0%).