SPLENDOR – Bariatric Surgery and NASH

Association of Bariatric Surgery With Major Adverse Liver and Cardiovascular Outcomes in Patients With Biopsy-Proven Nonalcoholic Steatohepatitis.

Aminian A, Al-Kurd A, Wilson R, et al.

JAMA. 2021 Nov 23;326(20):2031-2042. [Full text]

SPLENDOR on Surfing the NASH Podcast

Summary by Srikar Reddy

Nonalcoholic steatohepatitis (NASH), which is typically associated with underlying obesity and metabolic disease, has become the leading cause of cirrhosis and hepatocellular carcinoma. Currently no drug therapies have shown to limit the adverse liver or cardiovascular outcomes associated with NASH. Bariatric surgery has emerged as a major treatment for reduction of excess body weight and improved outcomes for diabetes, hypertension, and dyslipidemia (see SOARD.com). Until this study, outcomes in NASH after bariatic surgery had not been assessed other than through small observation studies, which had shown improvement in some histological improvement. SPLENDOR (Surgical Procedures and Long-term Effectiveness in NASH Disease and Obesity Risk) was a significantly larger cohort study designed to evaluate the relationship between bariatric surgery and adverse liver and cardiovascular outcomes.

Patient population and Design

Liver biopsies from patients with fibrotic NASH but without cirrhosis were reviewed. The nonalcoholic fatty liver disease (NAFLD) activity score is defined as liver steatosis (grade 0 to 3), hepatocyte ballooning (grade 0 to 2), and lobular inflammation (grade 0 to 3). To qualify as NASH, at least one point for each of the three criteria of steatosis, ballooning, and lobular inflammation was required.

Patients ages 18 to 80 were included if they had obesity with a BMI≥30, a confirmed histological diagnosis of NASH, and presence of fibrosis on baseline liver biopsy (F1-F3 stages). Patient were excluded if the score was F4 (cirrhosis), had fibrosis from an etiology other than NASH, had a history of excessive alcohol use, hepatocellular carcinoma, organ transplant, HIV, were on dialysis, had severe heart failure (LVEF <20%), had a diagnosis of any type of cancer within 1 year of liver biopsy, or received total parenteral nutrition within 6 months of the biopsy.

Eligible patients were then split into groups that had underwent bariatric surgery (sleeve gastrectomy or Roux-en-Y) or were managed nonsurgically. There were 1158 patients in the study with obesity and biopsy-proven fibrotic NASH without cirrhosis. 650 patient underwent bariatric surgery and 508 patients were in the nonsurgical management group. Most (83%) of the bariatric surgery group underwent Roux-en-Y compared to sleeve gastrectomy (17%). Most patients were women (63.9%), and had a median age of 49.8 years, and median BMI pf 44.1.

The study was not a randomized clinical trial but was conducted as a retrospective cohort study that used propensity scores to account for differences. Propensity scoring sometimes tries to match patients between groups for balance but can ;ead to imprecision. This study used something called overlap weighting propensity scores that assigns weights to each patient that are proportional to the probability of that patient belonging to the opposite treatment group. With this method outliers have less of an impact than in traditional propensity matching.


At 10 years, the incidence of major adverse liver outcomes was 2.3% in bariatric surgery group compared to 9.6% in the nonsurgical group. The unadjusted absolute risk difference for major adverse liver outcomes was 7.3% compared to adjusted absolute risk difference was 12.4%. This adjustment takes into consideration baseline population differences between the bariatric and nonsurgical groups.

Major adverse cardiovascular outcomes at 10 years also favored the bariatric surgery group, with an 8.5% incidence in the bariatric surgery group compared to 15.7% in the nonsurgical group. The unadjusted absolute risk difference of major adverse cardiovascular outcomes was 7.2% compared to adjusted absolute risk difference of 13.9%.

The reduction in mean body weight at 10 years was 22.4% in bariatric surgery group compared to 4.6% in the nonsurgical group. Additionally, there was a statistically significant reduction in hemoglobin A1c of 1.6% in the bariatric surgery group.  

Of not, 9.5% of patients developed major adverse events within 30 days of bariatric surgery. These include postoperative sepsis (3.5%), gastrointestinal leaks (2.2%), pulmonary adverse events (2.2%), bleeding (1.7%), venous thromboembolism (1.4%), acute kidney injury (1.2%), small bowel obstruction (0.6%), and cardiac adverse events (0.5%). Within first year after bariatric surgery, 0.6% of patients passed away from surgical complications.


In this retrospective cohort study, bariatric surgery was associated with an improvement in major adverse liver and cardiovascular outcomes in patients with NASH and obesity. This is the first larger scale study showing a treatment modality connected to improvements in clinical endpoints for patients with biopsy-proven NASH. Moreover, this shows that bariatric surgery could at least be considered as a viable therapeutic option for NASH patients. The benefit may even be greater than reported. The adjusted absolute risk difference (adjusted for higher degree of comorbidities in the bariatric surgery population witha BMI of 45.7 compared to the non surgical BMI of 36) suggests even more favororable major adverse liver and cardiovascular outcomes in the bariatric surgery group.

It is important to note that bariatric surgery is associated with postsurgical complications. The incidence in this retrospective cohort study was 9.5%, with sepsis, gastrointestinal leaks, and pulmonary adverse events being the three most common. Patients should be selected with caution and postoperatively, patients should be closely monitored to minimize and mitigate any complications.

Limitations of this study include its observational, retrospective design. There may be the healthy user effect, in that patients who ultimately chose bariatric surgery were more motivated compared to their nonsurgical counterparts. Additionally, the non surgical group had no standard intervention, such as a targeted weight loss intervention. Finally, it should be noted that the larger majority (83%) of the bariatric surgery group underwent Roux-en-Y, but the sleeve gastrectomy is currently the most common bariatric surgical procedure done.

While there are multiple drugs in clinical development for NASH, none currently is approved. Few randomized trials have shown mild histological benefits with medications such as vitamin E, pioglitazone, obeticholic acid, liraglutide, and semaglutide. Weight loss remains the primary management for NASH and there are medications that have been effective for weight loss (i.e. liraglutide). Current estimates show that 20% of NASH patients will progress to cirrhosis during their lifetimes. This is especially a problem since the incidence and prevalence of NASH continues to grow substantially. Bariatric surgery, as the SPLENDOR study suggests, maybe an important tool to help prevent these major adverse liver outcomes as well as adverse cardiovascular outcomes.

F: Follow up10 years
R: RandomizationNO
I: Intention to treatNO
S: Similar at baselineNo
B: BlindingNO
E: Equal treatmentNO
S: Source (funding)Not discussed