Roux-en-Y Gastric Bypass Vs. Adjustable Gastric Banding
Roux-en-Y Gastric Bypass Vs. Adjustable Gastric Banding
The follow-up flowchart of the cohort is summarized in Figure 1. Overall, 1452 patients had bariatric surgery at Lille University Hospital from December 1990 to June 2012 in our ongoing longitudinal study. Among them, RYGB or AGB was performed in 707 (48.7%) and 582 (40.1%) patients, respectively. Liver biopsy was performed at the time of surgery in 1236 of these patients (95.9%) who were enrolled in the present study, with interpretable results in 1201 (97.2%). Among the patients at risk, 1164 (95.2%) and 480 (80.1%) were seen for routine follow-up visits at 1 and 5 years, respectively. A repeated percutaneous biopsy sample was obtained in 578 patients at 1 year and in 413 patients at 5 years, representing 47.2% and 68.9%, respectively, of all patients at risk.
(Enlarge Image)
Figure 1.
Five-year flowchart of the Lille Bariatric Cohort. *Percentage of patients at risk.
Overall, 48 adjustable gastric bands had been removed before 5 years in patients operated before 2008 (12.3%) because of complications or insufficient weight loss. To avoid interpretation bias in patients with multiple operations, these patients were excluded from this analysis at the time of reoperation.
At baseline, the body mass index was 48.4 ± 7.6 kg/m in the overall cohort and 392 patients (32.6%) had T2DM. When analyzed with the mixed model (Table 1), all clinical and biological metabolic outcomes including body composition, diabetes, insulin resistance (HOMA-IR), fasting blood glucose, triglycerides, and high-density lipoprotein cholesterol significantly improved after surgery (time effect; P < 0.001).
To eliminate a selection bias specific to the patients who underwent liver biopsy, we compared all outcomes in patients with or without liver biopsies at each time point (see Supplemental Digital Content Table 1, available at http://links.lww.com/SLA/A681). This sensitivity analysis revealed significant differences only at 1 year, reflecting our recent policy to omit repeated liver biopsy in patients with mild or no liver disease at the time of operation.
On the basis of liver biopsies performed at the time surgery, NAFLD was present in 1016 patients (86%) and categorized as severe in 22% of cases. Mean NAS was 1.8 ± 1.4 at baseline and superior or equal to 3 in 27.7% of patients. Nine patients (AGB: n = 3; RYGB: n = 6) had cirrhosis, as defined by Kleiner et al.
As summarized in Table 1 , all biological and histological liver outcomes improved after surgery (time effect; P < 0.001), as did the proportions of patients with severe steatosis and/or with NAS of 3 or more (Fig. 2). As expected, the alanine aminotransferase/aspartate aminotransferase ratio was correlated with NAS at baseline (P < 0.001, r = 0.2); its mean value was inferior to 1 at baseline and reached 1.2 ± 0.4 at 5 years after surgery (P < 0.001 vs baseline).
(Enlarge Image)
Figure 2.
The proportion of patients with severe steatosis (A) and NAS of 3 or more (B) at baseline, 1 year, and 5 years after AGB and RYGB.
All these findings were confirmed in a sensitivity analysis restricted to the 315 patients who had all 3 consecutive biopsy samples available (see Supplemental Digital Content Table 2 , available at http://links.lww.com/SLA/A682). Among these patients with complete histological data, 12 patients had NAS of 5 or more at baseline. As illustrated in Figure 3, complete normalization of liver histology was observed at 5 years in 8 of them (66%). Thirteen patients [RYGB: n = 6; and AGB: n = 7] had bridging fibrosis (Kleiner fibrosis score ≥3) at the time of the operation (4.1%). At 5 years, fibrosis had regressed (Kl einer fibrosis score ≤2) in 6 of them, including RYGB (n = 3; 50%) versus AGB (n = 3; 43%) and disappeared in 2 patients [RYGB: n = 1; AGB: n = 1].
(Enlarge Image)
Figure 3.
Pre- and postoperative liver biopsies performed in a 62 years-old woman with an NAS of 5 at baseline and complete remission 5 years after RYGB. Before surgery, liver biopsy samples revealed moderate pericentrolobular fibrosis (A) and severe steatohepatitis (D). At 1 year and 5 years, fibrosis (B, C) and steatohepatitis (E, F) drastically decreased. Bar: 50 μm; fibrosis: arrow; steatohepatitis: *.
At baseline, RYGB patients had a slightly but significantly higher mean body mass index (+3.0 kg/m, P < 0.001), more frequent metabolic syndrome (P = 0.005) and T2DM (P < 0.001), and a more severe NAFLD (P < 0.001), than AGB patients (Table 1). Weight loss was significantly higher at 1 year after RYGB than after AGB. Despite a partial weight regain observed beyond the first year, weight loss remained higher at 5 years in RYGB patients than in AGB patients (21.4% ± 12.7% after AGB vs 25.5% ± 11.8% after RYGB) (P < 0.001). When analyzed with the mixed model (Time × Intervention interaction), the postoperative improvement of all clinical and biological metabolic markers appeared superior and more rapid after RYGB than after AGB (P < 0.01). The postoperative impact of RYGBP was also superior on most biological liver parameters (P < 0.01) and on liver histological outcomes (P < 0.001). At 5 years, NAS and the amount of steatosis became significantly inferior in RYGB patients than in AGB patients (Table 1). The proportion of patients with advanced NAFLD (steatosis <60% and/or with NAS ≥3) was similar at baseline in both groups but became significantly inferior in RYGB patients compared with AGB patients after 1 and 5 years (Fig. 2).
To further explore the specific role of the type of operation type on NAFLD outcomes, we therefore performed a multivariate analysis (step-by-step linear multiple regression) adjusted on NAFLD baseline values, weight loss, and the presence of T2DM (Table 2). In this model, the distinct benefit observed in RYGB patients on steatosis and NAS was primarily explained by weight loss (60.1% and 57.2%, respectively). Nevertheless, T2DM and other weight independent mechanisms, specific to RYGB, contributed more marginally but yet significantly (P < 0.05) to improve NAFLD beyond weight loss.
Results
Follow-up
The follow-up flowchart of the cohort is summarized in Figure 1. Overall, 1452 patients had bariatric surgery at Lille University Hospital from December 1990 to June 2012 in our ongoing longitudinal study. Among them, RYGB or AGB was performed in 707 (48.7%) and 582 (40.1%) patients, respectively. Liver biopsy was performed at the time of surgery in 1236 of these patients (95.9%) who were enrolled in the present study, with interpretable results in 1201 (97.2%). Among the patients at risk, 1164 (95.2%) and 480 (80.1%) were seen for routine follow-up visits at 1 and 5 years, respectively. A repeated percutaneous biopsy sample was obtained in 578 patients at 1 year and in 413 patients at 5 years, representing 47.2% and 68.9%, respectively, of all patients at risk.
(Enlarge Image)
Figure 1.
Five-year flowchart of the Lille Bariatric Cohort. *Percentage of patients at risk.
Overall, 48 adjustable gastric bands had been removed before 5 years in patients operated before 2008 (12.3%) because of complications or insufficient weight loss. To avoid interpretation bias in patients with multiple operations, these patients were excluded from this analysis at the time of reoperation.
Metabolic Improvement After Surgery
At baseline, the body mass index was 48.4 ± 7.6 kg/m in the overall cohort and 392 patients (32.6%) had T2DM. When analyzed with the mixed model (Table 1), all clinical and biological metabolic outcomes including body composition, diabetes, insulin resistance (HOMA-IR), fasting blood glucose, triglycerides, and high-density lipoprotein cholesterol significantly improved after surgery (time effect; P < 0.001).
To eliminate a selection bias specific to the patients who underwent liver biopsy, we compared all outcomes in patients with or without liver biopsies at each time point (see Supplemental Digital Content Table 1, available at http://links.lww.com/SLA/A681). This sensitivity analysis revealed significant differences only at 1 year, reflecting our recent policy to omit repeated liver biopsy in patients with mild or no liver disease at the time of operation.
Improvement of NAFLD After Surgery
On the basis of liver biopsies performed at the time surgery, NAFLD was present in 1016 patients (86%) and categorized as severe in 22% of cases. Mean NAS was 1.8 ± 1.4 at baseline and superior or equal to 3 in 27.7% of patients. Nine patients (AGB: n = 3; RYGB: n = 6) had cirrhosis, as defined by Kleiner et al.
As summarized in Table 1 , all biological and histological liver outcomes improved after surgery (time effect; P < 0.001), as did the proportions of patients with severe steatosis and/or with NAS of 3 or more (Fig. 2). As expected, the alanine aminotransferase/aspartate aminotransferase ratio was correlated with NAS at baseline (P < 0.001, r = 0.2); its mean value was inferior to 1 at baseline and reached 1.2 ± 0.4 at 5 years after surgery (P < 0.001 vs baseline).
(Enlarge Image)
Figure 2.
The proportion of patients with severe steatosis (A) and NAS of 3 or more (B) at baseline, 1 year, and 5 years after AGB and RYGB.
All these findings were confirmed in a sensitivity analysis restricted to the 315 patients who had all 3 consecutive biopsy samples available (see Supplemental Digital Content Table 2 , available at http://links.lww.com/SLA/A682). Among these patients with complete histological data, 12 patients had NAS of 5 or more at baseline. As illustrated in Figure 3, complete normalization of liver histology was observed at 5 years in 8 of them (66%). Thirteen patients [RYGB: n = 6; and AGB: n = 7] had bridging fibrosis (Kleiner fibrosis score ≥3) at the time of the operation (4.1%). At 5 years, fibrosis had regressed (Kl einer fibrosis score ≤2) in 6 of them, including RYGB (n = 3; 50%) versus AGB (n = 3; 43%) and disappeared in 2 patients [RYGB: n = 1; AGB: n = 1].
(Enlarge Image)
Figure 3.
Pre- and postoperative liver biopsies performed in a 62 years-old woman with an NAS of 5 at baseline and complete remission 5 years after RYGB. Before surgery, liver biopsy samples revealed moderate pericentrolobular fibrosis (A) and severe steatohepatitis (D). At 1 year and 5 years, fibrosis (B, C) and steatohepatitis (E, F) drastically decreased. Bar: 50 μm; fibrosis: arrow; steatohepatitis: *.
Superior Impact of RYGB
At baseline, RYGB patients had a slightly but significantly higher mean body mass index (+3.0 kg/m, P < 0.001), more frequent metabolic syndrome (P = 0.005) and T2DM (P < 0.001), and a more severe NAFLD (P < 0.001), than AGB patients (Table 1). Weight loss was significantly higher at 1 year after RYGB than after AGB. Despite a partial weight regain observed beyond the first year, weight loss remained higher at 5 years in RYGB patients than in AGB patients (21.4% ± 12.7% after AGB vs 25.5% ± 11.8% after RYGB) (P < 0.001). When analyzed with the mixed model (Time × Intervention interaction), the postoperative improvement of all clinical and biological metabolic markers appeared superior and more rapid after RYGB than after AGB (P < 0.01). The postoperative impact of RYGBP was also superior on most biological liver parameters (P < 0.01) and on liver histological outcomes (P < 0.001). At 5 years, NAS and the amount of steatosis became significantly inferior in RYGB patients than in AGB patients (Table 1). The proportion of patients with advanced NAFLD (steatosis <60% and/or with NAS ≥3) was similar at baseline in both groups but became significantly inferior in RYGB patients compared with AGB patients after 1 and 5 years (Fig. 2).
Multivariate Analysis
To further explore the specific role of the type of operation type on NAFLD outcomes, we therefore performed a multivariate analysis (step-by-step linear multiple regression) adjusted on NAFLD baseline values, weight loss, and the presence of T2DM (Table 2). In this model, the distinct benefit observed in RYGB patients on steatosis and NAS was primarily explained by weight loss (60.1% and 57.2%, respectively). Nevertheless, T2DM and other weight independent mechanisms, specific to RYGB, contributed more marginally but yet significantly (P < 0.05) to improve NAFLD beyond weight loss.
