Longitudinal Retrospective Analysis of Prospectively Collected Data for Outcomes After Primary Sleeve Gastrectomy (N=20 000) in the Swedish National Quality Registry SOReg (Scandinavian Obesity Surgery Registry)

Few countries have a national quality registry for metabolic and bariatric surgery with a very high coverage and level of adherence. The national registry database in Sweden - SOReg -was founded in 2007 and since 2012 close to 100% of all metabolic and bariatric surgery in Sweden is captured, and all active clinics report to the registry [18]. Thus, SOReg data represents a unique opportunity to report the long-term outcomes of the sleeve gastrectomy on a national level and to correlate preoperative demographics, operating unit and procedural factors with weight loss outcomes, resolution of comorbidities and complications. It also provides us insight into the true prevalence of conversion of sleeve gastrectomy to another procedure and the reasons hereto.

Primary aim:

The primary aim of the SCAN study is to evaluate the surgical outcomes (weight loss, resolution of comorbidities and complications) of the sleeve gastrectomy in a national cohort with a long-term follow-up. Weight outcomes will be presented in weight strata.

The secondary aims are:

For the whole cohort:

• Description of anthropometrics prior to sleeve gastrectomy.
• Determining rate of self-paying vs publicly funded procedures.
• Description of procedural factors
• To evaluate possible correlations in preoperative demographics, operating unit and procedural factors with outcomes.
• Estimate the prevalence of conversion of sleeve gastrectomy to another procedure (RYGB, DS, SASI etc) at different follow-up times, as well as determining the indications that lead to conversion.
• Identifying the rate of hiatal hernia in the whole cohort (n=20435) as plotted against time (baseline= before the sleeve procedure, thereafter yearly).
• Identifying the rate of clinical dyspepsia in the whole cohort (n=20435) as plotted against time (baseline= before the sleeve procedure, thereafter yearly).
• Identifying the Incidence of initial suboptimal clinical response vs recurrent weight gain in the whole cohort versus the group of patients converted regardless of indication for conversion.
• Estimate eligibility for a second revisional procedure in the total cohort due to either suboptimal weight loss outcomes/recurrence of comorbidities or dyspepsia.
• Identifying potential predictors for incidence of clinical dyspepsia ("Time from sleeve operation to dyspepsia is the outcome") leading to conversion (use of PPI before sleeve, dyspepsia before sleeve, hiatal hernia before sleeve, age, sex, BMI, smoking status, and clinical operating unit (high-volume vs low-volume). Nb! Dyspepsia is noted yearly (column HQ: year one; IV: year two etc).

Study Design The study is a retrospective study on prospectively collected data from the national registry on bariatric surgery, SOReg, where data on metabolic and bariatric surgeries in Sweden are collected.

The cohort will be all primary sleeve operations in Sweden performed during the years 2012 to 2024, with a total of n=20.435 patients. TThe design will be register-based cohort study, with a maximum follow-up time up to 15 years.

Study outcomes and predictors in weight strata (5 BMI points each)

a Conversion: Prevalence of conversion of SG to another procedure: RYGB, SASI, Duodenal switch, or a two-stage planned procedure (first SG and then later another like the DS). The data from the excel file should be held up against the total number of sleeve procedures in the same time period to assess the true prevalence.

b Procedural factors: These include operative time, Length of Stay (LOS), early complications (leaks, bleeding, reoperation, non-operative intervention, readmission, cardiac events, pneumonia, thromboembolic complications, wound infection, sepsis, death), and late complications (bowel obstruction, marginal ulcers, abdominal pain, laparoscopy with no pathology, malnutrition, and malabsorption)

c Predictors for all complications en bloc, for mortality, for early vs late complications, and then for all complications individually. Potential variables to be evaluated: Height, weight, sex, BMI, medical comorbidities one-by-one, smoking status, operative time, prior thromboembolic complications, operating unit (high-volume vs low-volume as defined by the number of procedures the operating units perform annually; this information is not available from the excel file, but I suggest preliminarily that the number of complications are stratified according to the respective operating units)

e Predictors for resolution of symptoms: intestinal limb length (alimentary limb, biliopancreatic limb and common channel), crura plasty (posterior vs anterior), other concomitant reflux procedures, operating unit (high-volume vs low-volume).

f Weight loss outcomes weight in kg, body mass index (BMI), percent of total weight loss (%TWL), percent excess weight loss (%EWL), and delta BMI (ΔBMI). The above mentioned are calculated as follows:

• %TWL = (Initial Weight - Postop Weight)/ (Initial Weight x 100)
• %EWL = (Initial Weight - (Postop Weight) / (Initial Weight - Ideal Weight)
• ΔBMI = (Initial BMI - Postop BMI) (Ideal weight is defined as the weight in kg that corresponds to a BMI of 25 kg/m2) g Medical comorbidities Diabetes mellitus (DM), Hypertension (HT), Dyslipidaemia (DL), and Obstructive Sleep Apnoea (OSA) i Weight loss outcomes (BMI, kg, EWL and TWL) should be assessed for Group 1 as a whole and also when dichotomized according to the following two groups: I) Recurrent weight gain (previously called poor responders) II) Initial suboptimal clinical response (previously called non-responders). j Eligibility here is defined as having a BMI ≥ 35 kg/m2 or BMI ≥ 30 kg/m2 along with a recurrence/presence of obesity-associated medical comorbidities (DM, HT, DL, OSA, Arthritis).

Predictors of resolution of comorbidities Variables to be considered potential predictors: preoperative weight and BMI, sex, age, smoking status, distance from angle of Hiss, distance to pylorus, numbers of staplers used, summated length of staplers, operative time.

**Complications: Peri-operative complications: Injury to sleen, bowel perforation, other complication and bleeding (100-499, 500-1500, and >1500 ml).

Complications within 6 weeks: Leakage, Bleeding Abcess/deep infection, Wound dehiscence, Other WOund complication, bowel obstruction/paralysis, Port related complication, Stricture, Ulcer, Cardio-vascular complication, DVT/PE, Pulmonary complication, Unrinary tract infection, Other complication, and Severity of complications according to Clavien Dindo classification.

Complications until 1y:

Complications can then be analysed yearly according to the data available in the spread sheet.

Reasons for conversion to another bariatric procedure and relief of problems:

Group 1: Symptoms: Clinical GERD, dyspepsia, dysphagia, stricture or obstruction, nausea or vomiting, staple line leak, hiatal hernia, intrathoracic migration of the sleeve or anything else that can be classified as complications after a sleeve gastrectomy.

Group 2: Suboptimal weight loss: Initial suboptimal clinical response and recurrent weight gain, as well as recurrence of obesity-associated medical comorbidities or persistent obesity-associated medical comorbidities.

Statistical analyses

Summary tables (descriptive statistics and frequency tables) will be provided for all baseline variables, outcome variables, and safety variables, as appropriate. Continuous variables will be summarized with descriptive statistics (n, mean, standard deviation, range, and median). Frequency counts and percentage of subjects within each category will be provided for categorical data.

For the primary aim, time to first ever conversion will be evaluated with the Kaplan-Meier estimates and Cox proportional hazards models. Time to event will be calculated as the time difference between the index operation with SG to another procedure. Persons without another procedure will be censored at the the time point when the register was complete. From the Kaplan-Meier and Cox-models, cumulative incidence of proportion converting will be evaluated since the time of primary SG operation.

For persons with symptoms leading to conversion, resolution of these symptoms will be evaluated at x months/years since conversion. Time-to-event or logistic regression models.

For persons with suboptimal weight loss as reason for conversion, weight loss outcomes will be evaluated.

Proportions with resolution of co-morbidities and incidence of complications will be analysed with time-to-event/logistic regression model.