Background
Multiple and large liver cancers remain a major challenge in liver surgery, although hepatectomy has become the most effective treatment [
1]. A very common reason is the limited future liver remnant (FLR), one of the determining factors leading to postoperative liver failure (PLF), which restricts the application of this method. Commonly, a volume of over 20% FLR must remain to avoid PLF in a normal liver [
2]. Conventional two-stage hepatectomy (TSH), including portal vein embolization (PVE) or portal vein ligation (PVL) accompanied by subsequent hepatectomy, represents one solution to this dilemma [
3]. Regrettably, the shortcomings of a long time interval and low regeneration efficiency seriously limit its application [
4]. In 2012, a new surgical approach named “associating liver partition and portal vein ligation for staged hepatectomy (ALPPS)”, characterized by great liver regeneration efficiency, was proposed by Schnitzbauer et al. [
5]. However, standard methods for patient selection and the long-term outcomes of this new approach remain controversial.
Several systematic reviews have examined the advantages and disadvantages of ALPPS, especially the ability to promote an increased FLR. However, this method continues to provoke heated debate because of its high mortality and unclear feasibility compared with other technologies, such as conventional TSH. Moreover, patient outcomes of ALPPS also remain uncertain [
6‐
8]. In the current analysis, we systematically searched for published studies comparing ALPPS and TSH to evaluate the liver regeneration efficiency, safety, and complication rates of these two methods.
Methods
The prospective agreement on study objectives, literature search methods, inclusion and exclusion criteria, outcome measurements, and statistical analysis methods were selected according to the Preferred Items for Systematic Reviews and Meta-Analysis (PRISMA) (Additional file
1: Appendix S1) and Meta-Analysis of Observational Studies in Epidemiology (MOOSE) guidelines (Additional file
2: Table S1).
There is currently no agreement on the definition of TSH. For the purpose of statistical analysis, the TSH group included patients who received PVE, PVL, or both.
Search strategy and selection criteria
A search of the PubMed, Embase, and Cochrane Library databases was performed to identify all studies comparing ALPPS and TSH. The following terms were searched: “liver partition,” “liver transection,” “portal vein occlusion,” “PVO,” “portal vein embolization,” “PVE,” “portal vein ligation,” “PVL,” “associating liver partition and portal vein ligation for staged hepatectomy,”, “ALPPS,” “staged hepatectomy,” and “two-stage hepatectomy.” The publication time was from January 1, 2009, to September 30, 2017, and the language was restricted to English.
All comparative studies comparing ALPPS with TSH in primary or secondary liver tumor patients who received surgery for staged hepatectomy were included. These studies included at least one quantitative outcome of interest, such as regeneration efficiency, time interval of the two stages, completion rate for second-stage operations, tumor deterioration, insufficient regeneration, R0 resection ratio, liver failure, bile leak, 90-day mortality, and 1-year disease-free survival. The excluded studies were those with irrelevant topics, case reports, non-comparative studies, review articles, letters, incomplete multiple published reports, and conference abstracts.
Data extraction and outcomes of interest
Two authors independently extracted and summarized the following data: patient characteristics, research designs, inclusion and exclusion criteria, and reported outcomes. Any disagreements were jointly resolved by the authors.
Quality assessment and statistical analysis
The studies ultimately included were classified according to the “2011 Levels of Evidence for Common Harms (Treatment Harms)” (Centre for Evidence Based Medicine, Oxford, UK). These levels of evidence can be described as follows: systematic reviews of randomized trials, systematic reviews of nested case-control studies and n-of-1 trials (level 1); individual randomized trials or (exceptionally) observational studies with a dramatic effect (level 2); non-randomized controlled cohort/follow-up studies, provided there are sufficient numbers to rule out a common harm (level 3); case-series, case-control studies, or historically controlled studies (level 4); and mechanism-based reasoning (level 5) [
9].
The weighted mean difference (WMD) and OR were used to compare continuous and dichotomous variables, respectively. The OR and 95% confidence interval (CI) were calculated for binary data.
P values < 0.05 were considered to indicate statistical significance. When the mean and variance were not reported, they were calculated using the median and range through a formula reported by Hozo SP [
10]. Heterogeneity between studies was assessed using the chi-squared test and
I
2. A
P value < 0.10 was used to indicate heterogeneity. Fixed-effect models were used for cumulative analyses when there was a lack of heterogeneity; otherwise, random-effect models were used.
The quality of all included studies was assessed using the Newcastle-Ottawa Scale (NOS) system, and patient selection, study comparability, and outcomes were evaluated. Scores ranging from 0 to 9 were calculated for each included study, and studies achieving a score of 6 or more were considered to be of good quality. Statistical analysis was performed using Review Manager Version 5.2 software (Cochrane Collaboration, Oxford, UK) (Additional file
3: Table S2).
Discussion
This systematic review and meta-analysis included 719 patients and aimed to compare the regeneration efficiency, safety, and complication rates of ALPPS and TSH. The absolute value of FLR regeneration in ALPPS was significantly higher than that in TSH, and the interval of the two stages in ALPPS was clearly shorter than that in TSH. In addition, ALPPS was associated with a higher completion rate, a lower probability of tumor progression during the stage interval, and a lower insufficient regeneration rate; these findings are similar to those of previous studies [
21‐
23]. However, complications, especially bile fistulas, were much more frequent in ALPPS, likely because of the liver splitting required in this procedure. Although ALPPS was associated with a lower rate of 1-year disease-free survival, there was no significant difference in the 90-day mortality rate between these two methods.
Given the safety of operation and possibility of PLF, a surgeon’s decision regarding a secondary ALPPS surgery is often affected by various factors. Two common factors are residual liver volume and function. In general, once the volume of liver regeneration reaches 20% of the normal liver and 30% of the cirrhotic liver, secondary surgery can be safely performed [
13‐
16,
18‐
20]. The rate of FLR regeneration differs during the regeneration progress. For instance, Correa reported that for 10 patients who did not accept resection but received PVE, the FLR regeneration rate was different at various periods within the 1-year follow-up, and the liver continued to grow even 1 year after PVE [
24]. However, the FLR regeneration volume alone still cannot reflect effective liver regeneration since patients usually have different regeneration rates after the operation. In our study, the absolute value of FLR regeneration was higher in ALPPS than in TSH, similar to the findings of previous studies. However, there was no uniform calculation of kinetic growth rate (KGR), and we did not perform a unified analysis of these included studies to better assess liver regeneration.
Furthermore, the pathological state of the liver also influences the potential for regeneration. Commonly, insufficient regeneration is accompanied by chronic liver disease. Chia et al. reported that patients with hepatitis B virus (HBV)-related cirrhosis require a longer time for regeneration before stage 2 [
25]. Furthermore, neoadjuvant chemotherapy and chemotherapy have been reported to inhibit liver regeneration to some extent.
What is the best way to evaluate liver regeneration? In 2013, Shindoh et al. proposed the concept of KGR to try to resolve this issue [
26]. However, KGR may reflect only the increased volume, not the increased function of the liver. Moreover, there is no uniform calculation for KGR, and therefore, this index remains controversial. Additionally, several studies have shown that many cytokines expressed during the liver regeneration process following ALPPS are associated with regeneration [
27‐
29]. Portal hemodynamics and immature hepatocytes during the process were both considered influencing factors [
15,
21,
30,
31]. Given all these factors, the most effective method to evaluate the function of the FLR remains unclear and further work is necessary to address the timing and selection of optimal candidates for staged hepatectomy.
Additionally, the interval time in TSH was longer, and tumor progression was more common [
32]. It is possible that the long time interval of TSH caused a higher risk of tumor progression. Only 4 patients were diagnosed with tumor progression in the ALPPS group in this study. Moreover, Fukami reported that Ki-67, a marker of cell proliferation, was increased to 80% in tumor cells after the second stage of ALPPS, and this increase was higher than that identified in the first stage in patients with colon cancers and synchronous multiple liver metastases [
33]. However, there is no consensus regarding whether the rapid regeneration response in ALPPS can aggravate the progression of the primary tumor. In addition, some researchers have reported that the incidence of postoperative tumor recurrence ranges from 14 to 87%, which is higher than that of TSH. Our study also indicated a similar conclusion. Furthermore, a higher R0 ratio could be seen in the ALPPS group; this result may be related to the entire splitting of stage 1. Additionally, there is no benefit to extending the interval time before stage 2, as a decreased KGR may occur at the seventh day after stage 1 [
34]. At the same time, considering that performing the stage 2 operation without sufficient preoperative assessment of the liver regeneration may easily lead to liver failure, it is necessary to establish an evaluation standard for assessing liver function after stage 1.
The mortality and complication rate are believed to be higher following ALPPS than following TSH. Our results also showed this trend although there was no significant difference between the two groups. Commonly increased portal blood, the most distinctive characteristic differentiating ALPPS from TSH, would lead to portal hyperperfusion which can conversely cause liver failure. Allard et al.’s research has shown a positive correlation between the portal vein pressure (PVP) and 90-day mortality. A multivariate analysis also indicated that PVP after hepatectomy was an independent predictor for PLF [
35]. Thus, the hemodynamic change in the portal vein after hepatectomy is closely related to patient prognosis. The increased portal vein flow in ALPPS can block the circulation from two parts of the liver and result in centralized hepatic blood flow, which ultimately leads to portal hyperperfusion of the FLR [
36]. However, with the hepatic artery buffering effect, the arterial flow in the non-ligated portion decreases to partly adapt to the increased portal vein flow [
37]. Vicente et al. investigated the pathological specimens of small-size syndrome patients and found that sinusoidal dilatation of the tissue was directly related to liver failure [
38].
There are many limitations of this meta-analysis. First, most of the included studies presented a low level of evidence, ranging from level 3 to level 4. Second, there was significant heterogeneity in some of the results because of the differences across studies in institutions, numbers of patients, types of primary tumors, and other factors. Third, there was obvious publication bias among some of the included studies.