Introduction
Nonalcoholic fatty liver disease (NAFLD) is a progressive condition affecting a large number of patients worldwide, with an estimated prevalence of 25% [
1]. Obesity is found to be an underlying condition for NAFLD frequently and can be considered as the hepatic manifestation of the metabolic syndrome [
2]. The rising incidence of NAFLD is closely linked to the rising incidence of obesity [
1,
3]. Prevalence of NAFLD in patients undergoing bariatric surgery has been described in up to 95% of cases [
4]. Weight loss improves NAFLD and reduces its likelihood of progression to nonalcoholic steatohepatitis (NASH) and cirrhosis [
5‐
10]. The positive impact of bariatric surgery on altitude of liver enzyme elevation has been assessed in a large prospective intervention study [
11]. Histologically, grade of NAFLD ameliorates after bariatric surgery, as measured by NAFLD activity score (NAS) [
6,
9,
12‐
14]. Even liver fibrosis can be resolved after bariatric surgery in some cases [
6,
9,
14‐
17]. However, there are also reports of acute liver failure or worsening of preexisting liver injury following bariatric surgery [
6,
18,
19].
As of today, the right surgical procedure as a treatment option for patients with NAFLD is subject to intense discussions. Some studies suggest sleeve gastrectomy (SG) and Roux-en-Y-gastric bypass (RYGB) to be of equal effectiveness in amelioration of NAFLD [
10,
12,
13], while others favor either RYGB [
6,
17,
20] or SG [
5,
16,
21]. Thus, the aim of this study was to compare postoperative recovery of liver function between RYGB and SG and to assess risk factors for potential deterioration of liver function capacity.
Discussion
Bariatric surgery presents an accepted treatment modality for NAFLD [
6,
11,
12,
17,
25,
26]. However, evidence on the modality of choice with regard to NAFLD (SG or RYGB) is sparse and a statement regarding the operative approach is lacking in the guidelines [
26,
27]. While some studies suggest superiority for one of the operative procedures, others advocate their equality [
12,
25]. A paramount concern about bariatric surgery in patients with chronic liver disease is potential worsening of preexisting liver conditions, underlining the importance of NAFLD assessment prior to surgery [
19,
26,
28,
29]. Aim of this study was to assess the impact of SG and RYGB on recovery or worsening of liver function in patients with obesity after bariatric surgery.
Different non-invasive screening tools for NAFLD are used in daily clinical routine. NAFLD practice guidelines by the American Association for the Study of Liver Diseases recommend the usage of the aspartate aminotransferase (AST) to platelet ratio (APRI), among others [
27]. APRI proved particularly useful for diagnosing NAFLD in patients with obesity [
30]. The non-invasive LiMAx test also reached a high sensitivity and specificity in detecting NAFLD in patients with obesity and has been developed for evaluation of liver function capacity in liver surgery [
30‐
32]. Alterations of the biliary flow as present in patients after RYGB do not seem to influence the postoperative liver function capacity measured by LiMAx test, as the test has been validated in patients undergoing major liver resection with biliary reconstruction [
33].
In our study, we found significantly improved mean liver function capacities as measured by both LiMAx test and APRI in patients who underwent SG but not RYGB (see Figs.
1 and
2). Our data concurs closely with Billeter et al., who found improved liver function following SG in comparison to RYGB in patients with elevated liver enzymes and type 2 diabetes mellitus (T2DM) [
21]. Although statistical significance was not reached, a prospective cohort study further described similarly improved liver regeneration in patients who underwent SG, but not RYGB [
34]. Contradictory to these results of enhanced function capacity in patients undergoing sleeve gastrectomy, a recently published meta-analysis found no significant difference in NAFLD remission between RYGB and SG as correlated through alanine transaminase, aspartate transaminase, NAFLD activity score, and NAFLD fibrosis score [
12].
While mean LiMAx scores improved significantly after SG and showed a positive trend after RYGB, the percentage of patients presenting with impaired liver function capacity remained similar succeeding both procedures (see Table
2). An explanation can be that while most patients’ liver function improves after surgery, other patients might suffer from impairment of liver function. Furthermore, even patients with improved liver function capacity after surgery might still suffer from impaired liver function due to severely impaired initial liver function capacity. This ultimately leads to the important question, which patients present with a constant or worsening liver function capacity after bariatric surgery.
There is growing evidence that major changes in liver function and volume occur within the first 6 months after surgery [
19,
20]. In our study, generalized linear model revealed significant influences for T2DM, preoperative weight, preoperative LiMAx, APRI, and GGT on liver function capacity after 6 months in patients after RYGB (see Table
3). In contrast, 6 month liver function capacity for SG patients seemed only affected by GGT, showing a slight, albeit, positive impact. %EBMIL was not associated with a decreased liver function capacity after 12 months neither in the SG nor in the RYGB group but revealed a trend towards a decreased liver function capacity in both groups after 6 months (see Table
4). Additionally, %EBMIL revealed a negative correlation with percentage change in LiMAx value after 6 months. Taken together, these findings indicate that rapid weight loss in the first 6 months after surgery might be associated with an initial worsening of liver function. Supporting this, a recent randomized controlled trial by Kalinowski et al. found a deterioration in liver function 1 month after surgery that was resolved after 12 months [
19]. In this study, patients with NAFLD were at a higher risk of postoperative impaired liver function after RYGB than patients after SG. Likewise, Nickel et al. described an increased APRI shortly (1 month) after bariatric surgery [
20]. However, they also found significant improvement of APRI 12 months after surgery with no difference between SG and RYGB [
20]. Correspondingly, Yeo et al. report on a positive correlation of weight loss and amelioration of NAFLD after 12 months but not after 6 months. Mean BMI prior to operation in that study was 42 kg/m
2 and hence much lower than in ours (mean BMI 53 kg/m
2). One factor influencing early postoperative liver function capacity in patients after bariatric surgery is the development of portomesenteric vein thrombosis [
35]. This rare but severe condition appears to evolve more frequently after SG than after RYGB, most likely due to locally diminished blood flow after dissection of the greater curvature [
36,
37].
Our findings furthermore suggest an influence of T2DM on regeneration of liver function after metabolic surgery. This is of high clinical importance due to the high prevalence of diabetes in patients with obesity. In patients who underwent RYGB, presence of T2DM was associated with a decrease in LiMAx at 6 months after surgery (see Table
3). For SG, patients with T2DM showed no significant decrease in LiMAx after 6 months. Our data indicate a slight advantage of SG in patients with T2DM with regard to early postoperative liver function capacity. In another matched pair study, SG has also been described superior to RYGB in terms of ameliorating liver function in patients with T2DM [
21].
Other factors influencing postoperative liver function capacity include APRI, GGT, and preoperative LiMAx. In particular for RYGB, we found an association of higher GGT levels with lower postoperative liver function capacity values after 6 months. Surprisingly, in SG, GGT levels showed a positive impact on liver function capacity after 6 months (see Table
3).
These findings have some clinical implications. Liver function shows a significant improvement after SG but not after RYGB. In terms of preexisting liver diseases, SG is preferred in patients with manifest cirrhosis due to a lower complication profile by some authors [
5,
29,
38‐
40]. Furthermore, SG has been described as being more cost-effective than RYGB in patients with NASH cirrhosis [
41]. While RYGB has recently been proposed as superior in EWL for patients that suffer from super obesity [
42], two-stage RYGB with an initial sleeve gastrectomy followed by RYGB yields comparable results [
43]. As higher weight was associated with a decreased liver function capacity after 6 months in patients with RYGB but not SG, our data encourage choosing SG as a first step with possible step-up to RYGB after initial weight loss in patients with high preoperative weight. To sum up, in patients with high preoperative weight and underlying preexisting liver disease, SG can be considered less harmful than RYGB. Furthermore, our findings reveal risk factors for impaired liver function capacity and indicate which patients might benefit from a closer monitoring of liver function postoperatively.
There are some limitations to be mentioned. One limitation is the high number of missing data for LiMAx after 6 and 12 months which might be due to the fact that LiMAx test is not available everywhere. However, we could obtain APRI values from almost all patients and could show a similar tendency. The results were not correlated to histological findings of liver biopsies. Furthermore, shear wave elastography and magnetic resonance elastography have recently been gaining importance in the diagnosis of fibrosis in NAFLD and might be a valuable option for preoperative evaluation of liver function [
44,
45]. Another limitation is the relatively small number of patients included and the lack of randomization. Notwithstanding these limitations, as guidelines for the selection of the operative procedure in NAFLD are lacking, this study provides encouraging results towards patient selection and monitoring of patients at risk for worsening of liver function after bariatric surgery.
Conclusion
In patients undergoing bariatric surgery, liver function improved significantly 12 months after sleeve gastrectomy, but not after Roux-en-Y-gastric bypass. We found a negative correlation between weight loss and amelioration of liver function. Factors associated with a decreased liver function capacity included type 2 diabetes mellitus, high preoperative weight, impaired APRI, elevated GGT, and male sex. Therefore, patients with these constellations should receive monitoring of liver function after the operation.
Data indicated as n (percent) or mean (SEM). Abbreviations: BMI, body mass index; LiMAx, liver function capacity test LiMAx; SG, sleeve gastrectomy; RYGB, Roux-en-Y-gastric bypass; SEM, standard error of the mean; APRI, aspartate aminotransferase (AST) to platelet ratio; T2DM, type 2 diabetes mellitus; * statistically significant using chi-square test
Data indicated as n (percent) or mean (SEM). Abbreviations: BMI, body mass index; LiMAx, liver function capacity test LiMAx; SG, sleeve gastrectomy; RYGB, Roux-en-Y-gastric bypass; SEM, standard error of the mean; APRI, aspartate aminotransferase (AST) to platelet ratio; * statistically significant using two-sided t-test
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.