Key Laboratory of Malignant Tumor Gene Regulation and Target Therapy of Guangdong Higher Education Institute, Sun Yat-sen University, Guangzhou, China
Division of Hematology and Oncology, Cedars-Sinai Medical Center, University of California at Los Angeles School of Medicine, Los Angeles, CA
General Surgery Department of Sun Yat-sen Memorial Hospital, Key Laboratory of Malignant Tumor Gene Regulation and Target Therapy of Guangdong Higher Education Institute, Sun Yat-sen University, Guangzhou, China
General Surgery Department of Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, China
General Surgery Department of Sun Yat-sen Memorial Hospital, Key Laboratory of Malignant Tumor Gene Regulation and Target Therapy of Guangdong Higher Education Institute, Sun Yat-sen University, Guangzhou, China
Potential conflict of interest: Nothing to report.
The optimal surgical strategy for treatment of patients with synchronous colorectal liver metastases (SCLRM) remains controversial. We conducted a systematic review and meta-analysis of all observational studies to define the safety and efficacy of simultaneous versus delayed resection of the colon and liver. A search for all major databases and relevant journals from inception to April 2012 without restriction on languages or regions was performed. Outcome measures were the primary parameters of postoperative survival, complication, and mortality, as well as other parameters of blood loss, operative time, and length of hospitalization. The test of heterogeneity was performed with the Q statistic. A total of 2,880 patients were included in the meta-analysis. Long-term oncological pooled estimates of overall survival (hazard ratio [HR]: 0.96; 95% confidence interval [CI]: 0.81-1.14; P = 0.64; I2 = 0) and recurrence-free survival (HR: 1.04; 95% CI: 0.76-1.43; P = 0.79; I2 = 53%) all showed similar outcomes for both simultaneous and delayed resections. A lower incidence of postoperative complication was attributed to the simultaneous group as opposed to that in the delayed group (modified relative ratio [RR] = 0.77; 95% CI: 0.67-0.89; P = 0.0002; I2 = 10%), whereas in terms of mortality within the postoperative 60 days no statistical difference was detected (RR = 1.12; 95% CI: 0.61-2.08; P = 0.71; I2 = 32%). Finally, selection criteria were recommended for SCRLM patients suitable for a simultaneous resection. Conclusion: Simultaneous resection is as efficient as a delayed procedure for long-term survival. There is evidence that in SCRLM patients simultaneous resection is an acceptable and safe option with carefully selected conditions. Due to the inherent limitations of the present study, future randomized controlled trials will be useful to confirm this conclusion. (HEPATOLOGY 2013;57:2346–2357)
Colorectal cancer (CRC) remains the second leading cause of cancer-related death in Western Europe and North America, and there are more than 940,000 new cases annually and nearly 500,000 deaths each year worldwide.1-3 Up to 50% of patients with CRC might have liver metastases during the course of the disease, and 15% to 20% have synchronous colorectal liver metastases (SCRLM) at the time of diagnosis, whereas an additional 20% to 25% develop metachronous hepatic tumors.4-7 The presence of liver metastases has an important influence on patient prognosis, and the median survival is 2.3 to 21.3 months for patients in whom the cancer is nonresected.3, 5, 8-10 Furthermore, liver resection has been accepted as the only treatment offering the chance for a cure and long-term survival, with 5-year survival rates of 25% to 60% and 10-year survival rates of 22% to 26% reported in the literature.11-19
However, optimal timing of liver surgery for synchronous metastases remains controversial and continues to evolve.20 Previously, most series reporting on the surgical management of SCRLM recommended a delayed approach with initial resection of the primary lesion followed by chemotherapy, then hepatectomy was typically offered only to patients who did not exhibit disease progression after prolonged courses of chemotherapy 2 to 3 months later.21, 22 Recently, the paradigm for the surgical management of SCRLM has begun to change. Due to advances in surgical technique of liver resection and enhancements in anesthesia and critical care, the safety and efficacy of simultaneous resection of colorectal and liver tumors have improved.23-27 In addition, it has been demonstrated that hepatotoxicity from contemporary chemotherapy regimens may damage liver parenchyma.28-30 Therefore, the present meta-analysis was designed to review and define the short- and long-term surgical outcomes following simultaneous and delayed resections for SCRLM patients.
The methods of literature search, inclusion and exclusion criteria, outcome measures, and methods of statistical analysis were defined in a protocol according to the Meta-analysis of Observational Studies in Epidemiology (MOOSE) recommendations for study reporting.31
Data Sources and Searches.
The primary sources of the reviewed studies through January 2012, without restrictions on languages or regions, were PubMed, Embase, Science Citation Index, and SpringerLink. We combined the database-specific search terms of simultaneous procedure and delayed procedure as well as truncated search terms using the wildcard (“*”) character for SCRLM patients. Additionally, the “related articles” function was also used to broaden the search, and the computer search was supplemented with manual searches for reference lists of all retrieved review articles, primary studies, and abstracts from meetings to identify other studies not found in the computer search. Authors of relevant abstracts were contacted to obtain any unpublished data (if available). When the results of a single study were reported in more than one publication, only the most recent and complete data were included.
All clinical studies in which a simultaneous strategy was compared with a delayed strategy in SCRLM were selected. Patients scheduled for a so-called “two-stage hepatectomy” procedure (two sequential hepatectomies for bilateral metastases unresectable by a single resection) were excluded from the meta-analysis. In addition, all of the studies included in the meta-analysis met the following criteria: (1) Liver metastasis as the first manifestation of M1 disease accompanied by no documented nonhepatic disseminated disease in preoperative imaging; intraoperative histologically proven colorectal carcinoma. (2) No prior history of liver-directed treatment such as hepatectomy, radiofrequency ablation, or other local modalities; no extrahepatic disease. (3) At least perioperative results or long-term results of survival and recurrence rates were assessed as outcome measures of the treatment effects. (4) They had been published or accepted for publication as full-length articles. (5) The study included at least 30 patients. Smaller studies were excluded because of poor reliability. The exclusion criteria were nonhuman studies, experimental trials, review articles, editorials, letters/case reports, and articles not reporting outcomes of interest.
Outcomes assessed were primary parameters of 1-, 3-, 5-year overall survival and recurrence-free survival, postoperative complications (those directly related to primary colorectal cancer resection [ileus, anastomotic leak, pelvic abscess, rectovaginal fistula], to hepatectomy [hepatic insufficiency/failure, subphrenic/perihepatic abscess, bile leakage, bleeding, etc.], to laparotomy [wound infection, intra-abdominal collection, pulmonary and cardiac diseases] and others), and postoperative mortality within 60 days; secondary parameters were blood loss during the operation, operative time, and length of hospitalization. In the delayed group, the parameters were the sum of the outcomes from the first primary CRC resection and the staged liver surgery. Accessory outcomes reported in some of the articles were also reviewed.
Data Extraction and Quality Assessment.
Two reviewers (Z.Y. and C.L.) independently considered the eligibility of potential titles and abstracts. When there was a disagreement about a study or a lack of information for an accurate assessment of eligibility, the study was carried to the full-text stage for evaluation. Data were extracted independently and in duplicate by another two reviewers (Y.C. and Y.B.); discrepancies were resolved by mutual discussion. We extracted the inclusion and exclusion criteria and the characteristics of each included study.
The quality of observational studies was assessed by modified criteria suggested by the Newcastle-Ottawa quality assessment tool.32 We also assessed the loss to follow-up and the ways in which missing data were handled for all studies.
Data Synthesis and Analysis.
The reported odds ratio (OR) and mean difference (MD) with 95% confidence interval (CI) were used in the analysis (when the incidence of an outcome of interest is common in the study population [>10%], pooled OR was then corrected to express the result as a summary risk ratio [RR]33). The hazard ratio (HR) was used as a summary statistic for long-term outcomes (survival analysis) as described by Parmar et al.34 An HR of less than 1 represented a survival benefit favoring the simultaneous group. Medians were converted to means using the technique described by Hozo et al.35 The fixed-effect model was first used to pool the results, which assumes that all the studies share the same common (fixed or nonrandom) effect. The studies were weighted in the meta-analysis by the inverse variances of their effect estimates, that is, the validities of the included studies. Heterogeneity was considered not statistically significant when the Cochrane Q test P value was >0.1. A transformation of Q test, the I2 statistic (I2 = 100% × (Q−df)/Q), was used to assess the consistency of the effect sizes. In case of heterogeneity, meta-analysis was performed applying the random-effects model. In addition, an I2 value of less than 25% was defined to represent low heterogeneity, a value between 25 and 50% was defined as moderate heterogeneity, and a value of >50% was defined as high heterogeneity.36
Subgroup analyses, which considered more homogeneous studies, were performed to identify subsets of patients more likely to benefit from the treatment and to assess the efficacy of different studies. To determine the extent to which the combined risk estimate might be affected by individual studies, sensitivity analysis was performed by consecutively omitting every study from the meta-analysis (leave-one-out procedure). Funnel plots were used to screen for publication bias. Meta-analysis was conducted by Review Manager (RevMan) Meta-analysis software, v. 5.1.6. The 95% CIs were calculated as estimates of precision for OR. The statistical tests were two-sided, and P < 0.05 was considered statistically significant.
Detailed analytical methods can be found in the Supporting Algorithms for Data Combination in the Meta-analysis.
Characteristics of the Included Studies.
Table 1 lists the characteristics of the included studies and details of the enrolled participants. Figure 1 illustrates the study screening and selection process. A total of 2,880 patients (simultaneous resection 1,015, delayed resection 1,865) from 17 studies were included. Synchronous metastases were defined as liver metastases diagnosed before colorectal resection or at the time of surgery, and patients scheduled for a so-called “two-stage hepatectomy” procedure (two sequential hepatectomies for bilateral metastases unresectable by a single resection) were excluded from the meta-analysis. Most studies were from Western Europe and North America in single-centers analyzed retrospectively and the number of patients per study ranged from 36 to 610 (multicenter study for Reddy et al.).27, 39 Preoperative chemotherapy status was reported in five studies.27, 40, 42, 47, 49 Moreover, we observed that patients with restricted metastatic disease were more likely to undergo simultaneous resections, whereas extended and anatomical difficult resections were rather performed as staged procedures. Distributions of risk (Severity) characteristics for the included patients from each observational study are detailed in Supporting Table 1.
Table 1. Characteristics of Included Studies of Simultaneous Versus Delayed Hepatectomy in Resectable Synchronous Hepatic Metastases of Colorectal Cancer*
Abbreviations: NA, not available; SD, standard deviation; Retro, retrospective.
Patients scheduled for a so-called “two-stage hepatectomy” procedure (two sequential hepatectomies for bilateral metastases unresectable by a single resection) were excluded from the meta-analysis.
Chemotherapy was carried out before hepatectomy in 42 (44%) patients of both groups.
The agreement between two reviewers for study selection was 0.94 and for quality assessment of trials was 0.89. We evaluated the risk of bias in the 17 observational studies by modification of the Newcastle-Ottawa scale (Table 2).32 Detailed descriptions of follow-up were available in most studies. The outcomes may have been influenced by various selection biases for patient allocation, including age, major hepatectomy (more than three Couinaud segments), sex ratio, colorectal primary location, the distribution, total number, and the largest size of liver metastases, and preoperative chemotherapy (Supporting Fig. 1). More colectomy combined with minor hepatectomy was performed in the simultaneous group. In addition, No study described adequately the patient flow. Methods for handling missing data were not adequately described in most studies. As shown by Supporting Fig. 2, the funnel plots are symmetrical, similar to inverted funnels, which means little publication bias exists in this meta-analysis for primary measures.
Table 2. Risk of Bias in the Observational Studies Using Ottawa-Newcastle Rules and Other Criteria
Abbreviations: Preop, preoperative; ACA, available case analysis.
The reference group received staged operations. Patients scheduled for a so-called “two-stage hepatectomy” procedure (two sequential hepatectomies for bilateral metastases unresectable by a single resection) were excluded from the meta-analysis.
The physician decided the arm allocation, and no protocol was present.
The following matching criteria were used: age, sex, number (categorized as one, two or three, or more than three) and distribution (unilateral or bilateral) of CLMs at diagnosis. In addition, institutional results were compared with those of a multicenter cohort of surgically treated patients with CLMs (the international registry LiverMetSurvey: http://www.livermetsurvey.org).
In the simultaneous resection group, liver metastases were identified before surgery in 37 patients and intraoperatively in three patients. In the delayed resection group, liver metastases were recognized before colorectal resection in 169 patients, during resection of the primary tumor in eight patients, and within 12 months after resection of the primary in two patients.
Simultaneous liver resections were generally not performed in patients with contraindications to additional major surgery.
Reviewed at monthly intervals for the first 3 months and then at 6-month intervals thereafter
Distribution of liver metastases (unilobar, bilobar), total number and the largest size of liver metastases
Long-Term Oncological Outcomes Were Similar for Both Simultaneous and Delayed Resections.
To evaluate the long-term oncological outcomes of simultaneous and delayed hepatic resections for treating SCRLM, HRs of overall survival and recurrence-free survival were calculated and combined in the present study using the data extracted from Kaplan-Meier curves (Fig. 2). Supporting Fig. 3 displays the constructed version of overall survival Kaplan-Meier graphs based on data inputted to an HR Calculations Spreadsheet, 10 studies with a total of 1,190 patients were included, and the postoperative duration for overall survival analysis ranged from 36 months to 168 months. The final pooled estimate of overall survival showed similar outcomes for both simultaneous and delayed resections (HR: 0.96; 95% CI: 0.81-1.14; P = 0.64; I2 = 0). When considering the effects of tumor recurrence on postoperative survival, analysis for the 486 patients from the four studies also did not detect a significant difference for the two surgical treatment strategies, and the final pooled HR of recurrence-free survival was found to be 1.04 (95% CI: 0.76-1.43; P = 0.79; I2 = 53%), with the follow-up time ranging from 18 months to 120 months (Supporting Fig. 4).
Simultaneous Hepatectomy Is a Relatively Safe Surgical Strategy in Resectable SCRLM Under Selected Conditions.
Meta-analyses for the efficacy (postoperative survival) and safety (postoperative complication and mortality) of the two hepatic resection strategies were the primary parameters in the current study. As mentioned above, simultaneous resection seemed endowed with a comparable long-term surgical oncological efficacy to delayed resection, whereas for safety considerations, a summary parameter of the two strategies implied a lower incidence of postoperative complication in the simultaneous group than that in the delayed group (modified RR = 0.77; 95% CI: 0.67-0.89; P = 0.0002; I2 = 10%) as shown by Fig. 3. In terms of postoperative mortality, significant difference was not observed based on the data included (RR = 1.12; 95% CI: 0.61-2.08; P = 0.71; I2 = 32%). Additionally, preoperative patient conditions in the simultaneous resection group were less severe, which were thought unavoidable in these observational studies due to the lack of randomized controlled trials (RCTs) so far. Nevertheless, distributions for various postoperative complications have been detailed and categorized in the present study (Supporting Table 2; Supporting Figs. 5, 6), and a conclusion could be drawn that simultaneous resection is safe for patients of SCRLM under some selected conditions. Thus, interpreting the limitations of selection bias associated with this present meta-analysis could be viewed as another interesting outcome, that is, to establish suitable selection criteria for patients who would really benefit from simultaneous resections (see below).
Six studies reported blood loss during operation (Supporting Fig. 7); the pooled estimate showed simultaneous hepatectomy was 181.19 mL significantly less than the delayed resection (95% CI: −357.41, −4.96; P = 0.04; I2 = 97%). As for operative time and hospital stay, the simultaneous strategy also had a significantly lower summary results compared to delayed strategy, with the pooled estimates of −46.97 min (95% CI: −94.50, 0.56; P = 0.05; I2 = 97%) and −4.64 day (95% CI: −6.38 to −2.90; P < 0.01; I2 = 96%), respectively.
Subgroup and Sensitivity Analyses.
Subgroup analyses were performed to evaluate whether the pooled estimates of long-term oncological outcomes were different according to different follow-up times (Table 3). The 1-, 3-, and 5-year pooled HRs of overall survival for simultaneous and delayed resections were found to be 0.95 (95% CI: 0.72-1.25; P = 0.70; I2 = 0%), 0.96 (95% CI 0.80-1.15; P = 0.67; I2 = 0%), and 0.97 (95% CI 0.81-1.16; P = 0.76; I2 = 0%). Similarly, as for 1-, 3- and 5-year recurrence-free survivals, no significant difference was detected from the meta-analysis either, and the pooled HRs between the two procedures were 1.15 (95% CI: 0.84-1.58; P = 0.37), 0.98 (95% CI: 0.74-1.29; P = 0.86), and 0.94 (95% CI: 0.72-1.24; P = 0.68), with nil heterogeneity. (Forest plots in Supporting Figs. 8, 9). Moreover, the results in sensitivity analyses by a leave-one-out procedure were all consistent with the above outcomes, indicating the strong robustness of the current study.
Table 3. Survival Hazard Ratios (HR) for the 1-, 3-, 5-Year Overall Survivals (OS) and Recurrence-Free Survivals (RFS)
No. of Studies
No. of Patients
HR (95% CI)
0.95 (0.72, 1.25)
0.96 (0.80, 1.15)
0.97 (0.81, 1.16)
1.15 (0.84, 1.58)
0.98 (0.74, 1.29)
0.94 (0.72, 1.24)
Establishment of Selection Criteria for Patients Who Are Suitable for a Simultaneous Resection.
Based on the included studies and current published prognostic models (Supporting Tables 3-5; Supporting Figs. 10-12), several factors were considered as the selection criteria for simultaneous liver resection directed against delayed resection: liver resection no more than three segments, colon resection (especially the right-sided colectomy), age less than 70 years old, and exclusion of coexisting severe conditions. These factors were exclusive to the simultaneous resection group. Future large and well-designed RCTs may be conducted under these selection criteria to confirm our conclusion.
For more detailed comments, see Supporting Mini-Systematic Review and Meta-Analysis on the Establishment of Selection Criteria for Patients Who are Suitable for a Simultaneous Resection.
In the present study we did not find a significant difference with regard to long-term outcomes of both overall survival and recurrence-free survival. Further, from the subgroup analyses of postoperative 1-year, 3-year, and 5-year survival data, the pooled results were also similar between the two groups. Thus, strictly speaking, simultaneous resection was as efficient as a delayed procedure for the long-term oncological outcomes. Many prognostic parameters, such as the distribution of liver metastases (unilobar, bilobar), total number and the largest size of metastases, stage of the primary colorectal tumor, and neoadjuvant chemotherapy have been reported to influence the survival after hepatectomy in SCRLM. A “time-test” approach was advocated by some authors to evaluate the biological behavior of the neoplasm, to treat potentially occult disease, and to avoid operation in patients with rapidly progressing tumors.51 Furthermore, neoadjuvant chemotherapy can be administered before delayed hepatectomy for synchronous liver metastases. However, no difference in survival was encountered between the two hepatectomy strategies either in the present analysis or other series,22, 26, 40 and no clear benefit from a time-test approach was defined.
In the past decades, the strategy of the delayed hepatectomy approach gained popularity and has been established as the standard surgical practice. Some authors hold the view that simultaneous resections may increase the rate of postoperative complications, particularly the risk of insufficiency of the colorectal anastomosis by the additional burden of a simultaneous major hepatectomy.52, 53 However, the fact that delayed resections require two separate operations and the negligible morbidity and mortality in modern hepatectomy demonstrated by the accumulating evidence have prompted some surgeons to attempt simultaneous resections of primary tumors and liver metastases. When the specific laparotomy complications were evaluated, it was increased in delayed group in Martin et al.'s study.41 Reddy et al. found that simultaneous resection strategy would increase the complications compared to liver surgery alone is not surprising, given the evidence from large series that simultaneous extrahepatic procedures increase morbidity after partial hepatectomy.27, 54 Other large studies38, 49, 55 have also shown that simultaneous resections were not associated with elevated hepatic or colon complications compared to delayed resections. This systematic review and meta-analysis also indicated that SCRLM patients who underwent only one procedure in selected conditions in which both safety and effectiveness are enhanced by the simultaneous resection strategy is acceptable.
It was expected that delayed hepatectomy would have a longer duration of procedure and hospital stay as well as more blood loss during operation. These findings were also confirmed in the current analysis. Blood loss has been shown to have a deleterious impact on both short- and long-term outcomes of the operation and is considered one of the important selective factors of hepatectomy strategy for SCRLM. Furthermore, hepatectomy is associated with a median blood loss of 450 to 1,500 mL and perioperative transfusion for major blood loss is also associated with an adverse prognosis.56, 57 Thus, in patients who have already had a large volume of blood loss during colectomy, a delayed hepatectomy approach should be considered instead.
Neoadjuvant chemotherapy prior to hepatectomy may facilitate the resectability of the liver lesions and treat occult metastasis, but it may also lead to hepatic parenchyma damage in patients with resectable colorectal liver metastases. Thus, the oncologic benefit of neoadjuvant chemotherapy in patients who may be suitable for a curative hepatectomy is still controversial. Five included studies reported the use of preoperative chemotherapy before resections. In the de Hass et al., Luo et al., and Reddy et al. studies,27, 47, 49, simultaneous patients were less often treated with chemotherapy before hepatic resection. This may explain the higher recurrence rate with the simultaneous resection strategy found in the de Haas et al. study.47 However, of note was the observation that the dropout in the delayed resection patients with progressive intrahepatic and/or extrahepatic disease after resection of the primary colorectal tumor may have selected a residual group with a more favorable prognosis, which may be the reason why preoperative chemotherapy was not an independent predictor of recurrence in the de Haas et al. study. In addition, the role of adjuvant treatment postliver resection should be viewed in the context of prior treatment, surgical preference, and individual patient characteristics. Current evidences have suggested that perioperative and regional therapies both showed an increase in recurrence-free survival in patients with resectable colorectal liver metastases. Nordlinger et al.58 concluded that perioperative chemotherapy with FOLFOX4 is compatible with major liver surgery and reduces the risk of events of progression-free survival in eligible and resected patients. Still, optimal regimens and sequencing of chemotherapies are unclear, and it is difficult to conduct RCTs examining the role of adjuvant chemotherapy due to the rapidly changing chemotherapies.
Lastly, as demonstrated by the main meta-analysis of timing of hepatectomy for patients with synchronous liver metastases, long-term outcomes of overall survival and recurrence-free survival were similar between the simultaneous and delayed groups; accompanied by selection bias, short-term outcome of postoperative morbidity was less detected in the simultaneous group. Therefore, safety was more often considered when establishing the selection criteria for simultaneous resection. Furthermore, it was true that patients in the simultaneous group had less severe disease compared with those in the delayed group because of the nature of the included study types, but this was not the main concern in the present study. The most interesting result of the current study would be the establishment of selection criteria for patients who could really be suitable for simultaneous resection, besides simply comparing the safety and efficacy of the two hepatic strategies.
The main limitation of this review is that only retrospective studies were available, which makes it difficult to acquire enough data for meaningful results. In most observational studies patients who underwent simultaneous hepatectomy or delayed hepatectomy was decided by the surgeon, and the allocation was based on surgeon preferences according to intraoperative findings and their own experience, which tended to increase the risk of bias for the results. In addition, universal definitions of postoperative complications were not available among the included studies. It is common sense that the best method to establish the selection criteria would be based on the individual patient data analysis (IPD meta-analysis); however, this is not always available, and the diverse reporting forms from the included studies could reduce the reliability of the conclusion.
In summary, this systematic review and meta-analysis was conducted at an appropriate time because enough data has accumulated for inspection by meta-analytical methods when a simultaneous resection strategy for SCRLM patients is used more commonly. From the present pooled estimates, simultaneous resection is as efficient as a delayed procedure for long-term survival. There is evidence implying that in selected SCRLM patients less than 70 years old, without severe coexistent disease and undergoing colectomy combined with minor liver resection, simultaneous resection is a safe procedure. Future RCTs are awaited to confirm this conclusion.
The authors thank Prof. Yuan-Tao Hao, Department of Medical Statistics, Sun Yat-sen University, Guangdong, China, for statistical advice; Yan Jia, Medical Library of North Campus, Sun Yat-sen University, Guangdong, China, for literature search.
Author Contributions: Z.Y., C.L., Y.C., Y.B., and C.S.: data collection and elaboration, statistical analysis; Z.Y., C.L., Y.C., and R.Y.: writing of the article; D.Y. and J.W.: supervision of the study.