Background characteristics and postoperative outcomes of insufficient weight loss after laparoscopic sleeve gastrectomy in Japanese patients

Abstract Aim Laparoscopic sleeve gastrectomy (LSG) is becoming popular in Japan, but insufficient weight loss is often observed in patients after LSG. We investigated the effect of LSG on obesity‐related comorbidities and identified the background characteristics of Japanese patients with insufficient weight loss after LSG. Methods In this multi‐institutional retrospective study at 10 certified bariatric institutions, 322 Japanese patients who underwent LSG with a follow‐up period of more than 2 years were analyzed. Anthropometry, obesity‐related comorbidities and psychosocial background data were collected. Weight loss was expressed as 2‐year percent total weight loss (%TWL). Results Mean age, body weight, body mass index (BMI) and glycated hemoglobin were 46.9 years, 119.2 kg, 43.7 kg/m2 and 7.1%, respectively. Prevalence of mental disorders was 26.3%. Mean BMI declined to 30.3 kg/m2 at 2 years and %TWL was 29.9%. Improvements in the markers and prevalence of obesity‐related comorbidities were observed. Remission rates of diabetes, dyslipidemia and hypertension were 75.6%, 59.7% and 41.8%, respectively. %TWL at the respective cut‐off level of diabetes remission was 20.8%. Lower remission rates of diabetes in patients with %TWL <20%, and less calorie restriction and higher prevalence of mental disorders (46.9%) in patients with %TWL <15% were observed. Frequencies of %TWL <15% and <20% were 6.5% and 18.5%, respectively. Conclusion %TWL 20% was a candidate cut‐off point of insufficient weight loss for diabetes remission after LSG, and mental disorders might be relevant to intractable obesity in Japanese patients.


| INTRODUC TI ON
Obesity is a chronic health problem and has reached worldwide epidemic proportions. 1 Obesity is associated with an increased risk of obesity-related comorbidities such as type 2 diabetes (T2DM), dyslipidemia, hypertension and sleep apnea syndrome (SAS). In Japanese patients, T2DM tends to become severe at a relatively lower BMI. [2][3][4][5] Hence, obesity is also a critical health problem in Japan. Treatment of obesity aims to restore the imbalance between energy intake and energy expenditure. However, conventional approaches such as lifestyle modification, dietary control, increasing physical activity and pharmacotherapy are usually insufficient to achieve satisfactory weight loss.
Bariatric surgery has become a popular treatment option for obesity. Laparoscopic Roux-en-Y gastric bypass was the most widely used bariatric procedure and, recently, laparoscopic sleeve gastrectomy (LSG) as a single-stage procedure is becoming more popular worldwide. 6 In Japan, LSG is the only bariatric procedure which is covered by the national health insurance and is therefore the main procedure used. A Japanese multi-institutional survey reported that % total weight loss (%TWL) was 29% at 1 year and complete or partial diabetes remission rate was 85% after LSG. 7 The results are comparable to similar surveys in American and European countries. In contrast, insufficient weight loss is often observed in patients after undergoing LSG, but background characteristics of these patients and the effect of LSG on obesity-related comorbidities have not been fully investigated in Japan. The cause of obesity is multifactorial, including genetic, social, economic, educational, environmental and psychological factors.
Therefore, there is a possibility that these preoperative factors predict success or failure of postoperative weight loss. The number of patients who undergo bariatric surgery in Japan is still quite low compared with American, European and other Asian countries. We established a retrospective study group "Japanese Survey of Morbid and Treatment-Resistant Obesity" (J-SMART). The aim of the present study was to investigate the effects of LSG on obesity-related comorbidities and to identify the background characteristics of Japanese obese patients with insufficient weight loss after LSG.

| SUBJEC TS AND ME THODS
Japanese Survey of Morbid and Treatment-Resistant Obesity was supported by a grant for research on intractable diseases from the Ministry Correspondence Atsuhito Saiki, Center of Diabetes, Endocrine and Metabolism, Toho University Sakura Medical Center, 564-1, Shimoshizu, Sakura-City, Chiba 285-8741, Japan. Email: atsuhito156@sakura.med.toho-u. ac.jp

Funding information
Ministry of Health, Labour and Welfare of Japan, Grant/Award Number: H28-nanjiippan-014 Abstract Aim: Laparoscopic sleeve gastrectomy (LSG) is becoming popular in Japan, but insufficient weight loss is often observed in patients after LSG. We investigated the effect of LSG on obesity-related comorbidities and identified the background characteristics of Japanese patients with insufficient weight loss after LSG.

Methods:
In this multi-institutional retrospective study at 10 certified bariatric institutions, 322 Japanese patients who underwent LSG with a follow-up period of more than 2 years were analyzed. Anthropometry, obesity-related comorbidities and psychosocial background data were collected. Weight loss was expressed as 2-year percent total weight loss (%TWL). in patients with %TWL <15% were observed. Frequencies of %TWL <15% and <20% were 6.5% and 18.5%, respectively. Conclusion: %TWL 20% was a candidate cut-off point of insufficient weight loss for diabetes remission after LSG, and mental disorders might be relevant to intractable obesity in Japanese patients.

K E Y W O R D S
diabetes remission, insufficient weight loss, Japanese, mental disorder, sleeve gastrectomy of Health, Labour and Welfare of Japan. This was a multicenter and retrospective database analysis. In this study, a total of 369 Japanese patients who underwent LSG at 10 bariatric institutions certified by Japanese Society for Treatment of Obesity between January 2011 and December 2014 with a follow-up period of more than 2 years were enrolled. Patients whose BMI was 30-34.9 kg/m 2 with at least one obesity-related comorbidity or more than 35 kg/m 2 at the first visit were included. Of these, 47 patients were excluded due to lack of clinical data at 2 years after LSG. Data from the remaining 322 patients were analyzed. Our surgical method for LSG generally follows the global standard techniques. 8 Average bougie size was 36.7 ± 3.2 Fr and the gastric resection started from 4-5 cm proximal to the pylorus.
The following preoperative data were collected from patient re- severe psychiatric illness such as depression, bipolar disorder, schizophrenia or binge-eating disorder. 9 The prevalence of mental disorders was the sum of mental retardation, developmental disorders, binge eating and/or other mental disorders. The Wechsler Adult Intelligence Scale (WAIS) III was carried out as the intelligence test within general practice. WAIS III was administered and scored by trained psychiatrist or psychologist. WAIS III was used not to eliminate patients, but to estimate their postsurgical outcomes such as lifestyle changes.
Childhood-onset obesity was defined as onset before 6 years of age and BMI above 95th percentile. Anthropometric measurements, VFA, SFA, blood pressure, and blood samples were also collected at 1, 2, 3, 4 and 5 years after LSG. Calorie intake and dietary composition were assessed by standardized interview by trained dieticians using a computerized database, based on the analysis of the semiquantitative food record of 3 consecutive days for each 2-week period. Physicians also retrieved these nutritional data. All participants were prescribed daily supplements of multivitamins and multiminerals. Participants were encouraged to regularly attend the bariatric surgery patient support group meetings.
Outcome of weight loss after LSG was evaluated as 2-year %TWL.
Patients were divided into five groups according to %TWL (≤14.9%, 15.0%-19.9%, 20.0%-24.9% 25.0%-29.9% and ≥30.0%). Remission of diabetes, dyslipidemia and hypertension was also evaluated at 2 years after LSG. Complete diabetes remission (CR) was defined as HbA1c <6.0% without using any diabetes medication. 10 Dyslipidemia remission was defined as total cholesterol (TC) level <220 mg/dL, triglyceride (TG) level <150 mg/dL and high-density lipoprotein cholesterol (HDL-C) level ≥40 mg/dL without using any medical treatment for dyslipidemia, based on the criteria of Japan Atherosclerosis Society Guidelines. 11 Hypertension remission was defined as systolic blood pressure (SBP) <130 mm Hg, diastolic blood pressure (DBP) <85 mm Hg and no medication for hypertension required, based on the normal range of the Japanese Society of Hypertension Guidelines for the Management of Hypertension. 12 Results were expressed as mean ± SD (SPSS 15.0; SPSS Inc., Chicago, IL, USA) in all statistical analyses. All parametric data were analyzed using Student's t test (paired and unpaired where appropriate). All non-parametric data were analyzed using the Mann-Whitney U test. Fisher's exact test was used to identify any significant difference between proportions and categorical variables. Comparisons of 2-year metabolic remission rates among %TWL groups were conducted using simple linear regression, one-way ANOVA and Tukey post-hoc test. Sensitivity and specificity with respect to diabetes remission were analyzed using a conventional receiver-operatingcharacteristic (ROC) curve. A two-sided P value of .05 was considered statistically significant. If a case had missing data for any of the variables, we simply excluded that case from the analysis.

| RE SULTS
The 322 patients analyzed in the present study comprised 144 men and 178 women with a mean age of 46.9 years. At the first visit, mean body weight, BMI and HbA1c were 119.2 kg, 43.7 kg/m 2 and 7.1%, respectively ( Table 1). The follow-up rates, which were expressed as the ratio of expected number to actual number, during the study period at 1, 2, 3, 4 and 5 years after LSG were 87.3%, 87.3%, 81.3%, 77.3% and 73.5%, respectively. There was no intraoperative and postoperative mortality. Prevalence of T2DM, dyslipidemia and hypertension before surgery was 63.2%, 77.1% and 77.1%, respectively (data not shown).

| Weight loss
Changes in body weight and BMI at various time points are shown in Table 1. After LSG, rapid weight loss was observed during the initial first year and the weight stabilized thereafter.   1 year and 29.9% at 2 years. Figure 1 shows a histogram of %TWL at 2 years after LSG. Frequencies of %TWL <15% and <20% at 2 years, which were indicators of insufficient weight loss, were 6.5% and 18.5%, respectively (data not shown). Changes in VFA and SFA are shown in Table 1, and both decreased significantly in all postoperative periods compared to the first visit.

| Changes in metabolic parameters and relationship with %TWL
After LSG, mean HbA1c and FBG decreased rapidly to 5.9% and 96.5 mg/ dL, respectively, at 1 year, and to 5.7% and 97.8 mg/dL at 2 years (  Figure 2B. At the cut-off level (20.8%), sensitivity was 86.2% and specificity was 52.7%. Remission rate of dyslipidemia also decreased in low %TWL groups. Remission rate in %TWL ≤14.9% group was 17.7% and was significantly lower than that in %TWL ≥30.0% group (77.3%). The remission rate of hypertension was lower than the other metabolic remissions and did not differ significantly among %TWL groups.

| Other obesity-related comorbidities
Changes in markers of liver and renal functions, and uric acid from the first visit to various postoperative periods are shown in Table 1.
Liver function markers and serum uric acid decreased significantly in all postoperative periods, and renal markers decreased significantly at

| Psychosocial background, eating behavior and physical activity
Parameters for psychosocial background, eating behavior and physical activity at first visit in patients stratified by %TWL are shown in  Figure 3A). In patients with %TWL <15%, the frequency was 46.9% and significantly higher than in those with %TWL ≥15% ( Figure 3B).
Similarly, the prevalence of mental retardation and developmental disorders in patients with %TWL <15% tended to be higher than in those with %TWL ≥15% (P = .0817) (data not shown). Furthermore, the prevalence of mental disorders in patients with %TWL ≥45% was also significantly higher than in those with %TWL <45% ( Figure 3C).

| Characteristics of patients with %TWL <15%
As mentioned above, patients with %TWL <15% had a high prevalence of mental disorders. Therefore, to clarify the characteristics F I G U R E 1 Histogram of percent total weight loss (%TWL) at 2 y after laparoscopic sleeve gastrectomy. Patients were divided into eight groups according to %TWL (every 5% of TWL). Black dotted line indicates mean %TWL (29.9 ± 11.0%) of patients with %TWL <15%, further investigation was carried out ( Figure 4). Patients with %TWL <15% not only had a smaller decrease in body weight compared to patients with %TWL ≥15%, they also showed significant weight regain from 1 year to 2 years after LSG (P = .0002). Patients with %TWL <15% had significantly higher daily calorie intake at 2 years after LSG than those with %TWL ≥15% (P < .0001) and showed a significant increase in calorie intake from 1 year to 2 years after surgery (P = .0486). and there were 10 institutions that carried out more than 10 bariatric/metabolic surgery cases in 2015. Therefore, the sample size in the present study could correspond to that nationwide. LSG has been established as a safe and effective procedure worldwide. [13][14][15] Seki et al 16  Uric acid (mg/dL) 6.3 ± 1.6 5.7 ± 1.5 6.8 ± 1.7 6.5 ± 1.6 6.2 ± 1.6 6.2 ± 1.6
As a definition of weight loss failure after bariatric surgery, Reinhold's classification 17 of BMI >35 or % excess weight loss (EWL) <50% is widely used and, using this definition, the rate of weight loss failure was reported to be 20.4%. 18 However, the definition of weight loss failure using %TWL has not been established.
van de Laar et al 19 reported that ≥25% TWL was a useful criterion with sensitivity and specificity of 90%, but that ≥15% and ≥20% TWL were inadequate criteria with specificity <30% for weight loss success with reference to the baseline BMI-independent weight loss percentile chart based on retrospective data after bariatric surgery. In the present study, diabetes remission rates were considerably lower in patients with %TWL <20% compared with higher %TWL. The ROC curve of %TWL at the optimal cut-off level of diabetes CR was 20.8%. There is a possibility that %TWL 20% is a candidate cut-off point of insufficient weight loss for diabetes remission after LSG. Furthermore, a lower remission rate of dyslipidemia was shown in patients with %TWL <15%.
This is the first report that comprehensively investigated obesity-related comorbidities other than metabolic disorders in bariatric patients in Japan. In the systematic review conducted by Buchwald However, the relationship between %TWL and improvement in heart failure, menstrual disorder, joint disorders and SAS was not detected in the present study as a result of a lack of available data.
Further prospective investigation is required to elucidate how much weight obese patients need to lose to improve their comorbidities.
The present study is also the first nationwide survey in Japan Note: One-way ANOVA detected no statistically significant differences in these parameters among the %TWL groups. Abbreviations: AHI, apnea hypopnea index; SAS, sleep apnea syndrome; %TWL, percent total weight loss.

| CON CLUS IONS
This is the first nationwide survey in Japan to clarify the relationship between insufficient weight loss after LSG, metabolic remission and psychosocial background. Our study showed that mean %TWL was 29.9%, and that weight loss resulted in significant improvement in obesity-related comorbidities. %TWL 20% might be an optimal cut-off point for diabetes remission in Japanese obese patients. Furthermore, patients with %TWL <15% might have increased appetite and mental disorders. Patients with insufficient weight loss may have treatment resistance that involves certain physical and psychological factors. Further investigation in collaboration with pediatricians and psychiatrists is required to clarify the mechanism of insufficient weight loss.

D I SCLOS U R E
Funding: J-SMART was supported by a grant for research on intrac- F I G U R E 3 Prevalence of mental disorders stratified by percent total weight loss (%TWL). (A) Patients were divided into eight groups according to %TWL (every 5% of TWL). Relationship between frequency of mental disorders and %TWL was a U-shaped curve, and the trend was significant using Fisher's exact test (P = .0046). (B) In patients with %TWL <15%, the frequency was 46.9% and significantly higher than in those with %TWL ≥15% (P = .0101). (C) Frequency in patients with %TWL ≥45% was also significantly higher than in those with %TWL <45% F I G U R E 4 Changes in (A) body weight and (B) daily calorie intake in patients with percent total weight loss (%TWL) <15% compared with %TWL ≥15%. In patients with %TWL <15%, significant weight regain was observed from 1 y after laparoscopic sleeve gastrectomy (LSG) (mean body weight was 99.8 kg at 1 y and 106.2 kg at 2 y, P = .0002). In patients with %TWL <15%, daily calorie intake at 2 y after LSG was significantly higher compared to %TWL ≥15% (2012.4 vs 1339.2 kcal/d, P < .0001), and a significant (P = .0486) increase in daily calorie intake from 1 y (1666.8 kcal/d) to 2 y after LSG (2012.4 kcal/d) was observed