Study search and selection details

Our database search yielded 1007 records after duplicate studies were removed. The titles and abstracts of these records were screened for their relevance, and 35 articles were retained for further evaluation. Of these 35 articles, 26 articles were further excluded because of incorrect comparison (n = 8), mixed liver tumor (n = 8), noncomparative study (n = 4), lack of detailed data (n = 4), duplicate data (n = 1), or study presented as an abstract (n = 1). Thus, 9 studies were selected for the meta-analysis [15,16,17,18,19,20,21,22,23] (Fig. 1).

Fig. 1

Flow chart of study selection

Study characteristics

The included 9 studies comprising one RCT and 8 retrospective studies and involved a total of 3268 patients, of which 1703 patients underwent liver resection with IPM (Table 1). Five studies reported OS and DFS data and eight studies reported perioperative outcomes. The mean or median occlusion time varied among the studies and ranged from 19.5 to 50 min.

Table 1 Characteristics of the included studiesQuality assessment results

The risk of bias for studies reporting RCTs was low. The details of the quality assessment of these studies are shown in Supplementary Materials 2 and 3. Among the eight retrospective studies, seven had high quality and one had moderate quality (Table 1).

OS and DFS analysis

The HR values of OS and DFS were available for five studies (one RCT and four retrospective studies). Because there was no significant heterogeneity among the studies, the fixed-effects model was used. Because of the larger sample size, the two studies by Xia et al. and Famularo et al. carried the highest impact on the analysis of OS and DFS. The meta-analysis showed no significant difference in the OS and DFS between the two groups (OS: HR, 1.01; 95% CI, 0.85–1.20; p = 0.95; DFS: HR, 1.01; 95% CI, 0.88–1.17; p = 0.86) (Fig. 2). The results of OS and DFS of subgroup analysis of the retrospective study was consistent with the primary analysis (OS: HR, 1.07; 95% CI, 0.89–1.28; DFS: HR, 0.99; 95% CI, 0.85–1.16) (Supplementary Material 4).

Fig. 2

Forest plots for overall survival and disease-free survival. A Overall survival forest plot. B Disease-free survival forest plot

Sensitivity analysis was performed by excluding each study at a time and combining the HR values for the remaining included studies (Supplementary Material 5). Sensitivity analysis showed that excluding studies by Xia et al. and Famularo et al. changed the CIs considerably, but the results of OS and DFS were still stable. Publication bias was considered significant for both OS and DFS because of the presence of asymmetry in the funnel plots (Supplementary Material 6). The trim and fill method was used to evaluate the effect of publication bias on the results of OS and DFS (Supplementary Material 7). The OS and DFS results before and after trimming and filling were similar (OS: HR, 0.98; 95% CI, 0.83–1.16; p = 0.84; DFS: HR, 0.99; 95% CI, 0.86–1.13; p = 0.83) (Supplementary Material 8).

Recurrence rate

Four studies reported the tumor recurrence rate. Because of significant heterogeneity among the studies, the random-effects model was used. The meta-analysis showed no significant difference in the tumor recurrence rate between the two groups (RR, 1.22; 95% CI, 0.85–1.74; p = 0.28) (Table 2).

Table 2 Details of perioperative outcomes and tumor recurrenceOperation time

Operation time was reported in six studies. The random-effects model was used because of significant heterogeneity among the studies. The meta-analysis revealed no significant difference in operation time between both groups (SMD, 0.23; 95% CI, − 0.40 to 0.86; p = 0.47) (Table 2).

Blood loss

Seven studies reported blood loss during liver resection. As there was significant heterogeneity among the studies, the random-effects model was used. The meta-analysis showed no significant difference in the volume of blood loss between both groups (SMD, 0.02; 95% CI, − 0.30 to 0.33; p = 0.92) (Table 2). However, sensitivity analysis showed that the study by Fumularo et al. [18] had a great influence on the results including heterogeneity and significance (Supplementary Material 9). The meta-analysis after omitting this study showed that IPM significantly reduced blood loss (SMD, − 0.20; 95% CI, − 0.28 to − 0.12; p < 0.01) (Supplementary Material 10). Sensitivity analysis indicated that the result was stable (Supplementary Material 11).

Blood transfusion

Four studies reported blood transfusion. The random-effects model was used because of significant heterogeneity among the studies. The meta-analysis showed no significant difference in blood transfusion between both groups (RR, 1.02; 95% CI, 0.47–2.20; p = 0.97) (Table 2).

Total complications

Total complications were reported in five studies. Because of significant heterogeneity among the studies, the random-effects model was used. The meta-analysis revealed no significant difference in total complications between both groups (RR, 0.91; 95% CI, 0.70–1.20; p = 0.52) (Table 2).

Liver failure

Liver failure was reported in six studies. The fixed-effects model was used as there was no significant heterogeneity among the studies. The meta-analysis showed no significant difference in liver failure between both groups (RR, 0.73; 95% CI, 0.35–1.52; p = 0.41) (Table 2).

Pleural effusion

Four studies reported pleural effusion. The random-effects model was used because of significant heterogeneity among the studies. The meta-analysis showed no significant difference in pleural effusion between both groups (RR, 1.27; 95% CI, 0.61–1.64; p = 0.52) (Table 2).

Ascites

The occurrence of ascites was reported in three studies. Because of significant heterogeneity among the studies, the random-effects model was used. The meta-analysis showed no significant difference in the occurrence of ascites between both groups (RR, 0.85; 95% CI, 0.46–1.58; p = 0.61) (Table 2).

Hospital stay

The length of hospital stay was reported in four studies. The random-effects model was used as the studies exhibited significant heterogeneity. The meta-analysis showed no significant difference in the length of hospital stay between both groups (SMD, 0.01; 95% CI, − 0.28 to 0.31; p = 0.94) (Table 2).

Subgroup analysis for perioperative outcomes

Subgroup analysis based on the proportion of patients with Child A, the proportion of patients with liver cirrhosis, the proportion of patients receiving major liver resection, and the proportion of patients with multiple tumor resection to explore the sources of heterogeneity for perioperative outcomes.

Subgroup analysis based on the proportion of patients with Child A showed that different proportion of patients with Child A is one source of heterogeneity for operation time, blood loss, blood transfusion, pleural effusion, and hospital stay (Supplementary Material 12). For those studies with a proportion of patients with Child A > 90%, meta-analysis showed that IPM may prolong operation time (SMD, 0.16; 95% CI, 0.01 to 0.37).

Subgroup analysis based on the proportion of patients with liver cirrhosis showed that different proportion of patients with liver cirrhosis is one source of heterogeneity for operation time, blood loss, blood transfusion, total complication, and hospital stay (Supplementary Material 13).

Subgroup analysis based on the proportion of patients who received major liver resection showed that different proportions of patients who received major liver resection are one source of heterogeneity for operation time and blood loss (Supplementary Material 14). For those studies with a proportion of patients who received major liver resection > 60%, a meta-analysis showed that IPM may reduce blood loss (SMD, − 0.23; 95% CI, − 0.41 to − 0.05).

Subgroup analysis based on the proportion of patients who received major liver resection showed that different proportion of patients received major liver resection is one source of heterogeneity for operation time, blood loss, blood transfusion, total complication, pleural effusion, and hospital stay (Supplementary Material 15). For those studies with a proportion of patients with multiple tumors < 25%, meta-analysis showed that IPM may prolong operation time (SMD, 0.16; 95% CI, 0.01 to 0.37). For those studies with the proportion of patients with multiple tumors > 25%, meta-analysis showed that IPM may reduce blood loss (SMD, − 0.17; 95% CI, − 0.34 to − 0.00).