Feasibility of a Supervised Inpatient Low-Calorie Diet Program for Massive Weight Loss Prior to RYGB in Superobese Patients
- First Online:
- Cite this article as:
- Huerta, S., Li, Z., Anthony, T. et al. OBES SURG (2010) 20: 173. doi:10.1007/s11695-009-0001-x
This study was undertaken to determine the feasibility of an inpatient low-calorie program for a substantial decrease of preoperative weight (>10 points in BMI) in superobese patients.
Five patients were hospitalized for an average of 11 weeks and were placed on a low-calorie liquid diet (<900 kcal/day) and an exercise program. Following a drop of ten points in BMI, they underwent a Roux-en-Y gastric bypass (RYGB). Hemoglobin A1c and lipid profiles were obtained at the beginning of the diet, prior to surgery and at the last follow-up appointment. Our results were compared to the National Surgical Quality Improvement Program (NSQIP) database, which included 1,046 bariatric operations performed at VA centers between October 1999 and August 2007.
All five patients were massively obese men (body mass index (BMI) = 64.3 ± 2.1 kg/m2; 54.7 ± 2.6 years old; four of five were white) with multiple comorbid conditions, which placed them in a substantially higher risk for bariatric surgery. Of the four diabetic patients, two were insulin dependent. There was an average decrease in BMI by 12.7 points (85.8 ± 6.0 lb) during the preoperative diet period (11 weeks). All patients underwent RYGB without complications. This cohort of patients further decreased their BMI by 10.6 points (88.4 ± 29.4 lb) following surgical intervention. The total combined preoperative and postoperative excess body weight loss was 89% (10.6-month average follow-up). Sleep apnea resolved following gastric bypass but did not improve during the preoperative weight loss period. Hypertension, osteoarthritis, and dyslipidemia all improved following surgical intervention. Hemoglobin A1c decreased by 1.9% during diet-induced weight loss with no further improvement being noted after surgery. The two insulin-dependent diabetic patients discontinued insulin therapy following surgery. The NSQIP database contained 77 patients with similar characteristics to our cohort of patients. The 30-day mortality for this cohort of patients was 3.9% with a complication rate of 33.8%.
Massive preoperative weight loss is possible to achieve with a liquid protein diet in superobese patients greatly facilitating gastric bypass surgery in an otherwise high-risk patient population.
KeywordsStaged operationsRYGBSuperobesityWeight reduction surgeryPostoperative complications
Forty-four percent of Veteran Administration (VA) patients are overweight, and one in four is obese . Weight reduction, even if moderate (≥10%), results in significant health benefits resulting from amelioration of obesity-related comorbid conditions . The benefit of achieving and maintaining weight loss in the VA population is substantial.
In morbidly obese patients (body mass index (BMI) ≥ 40 kg/m2), however, sustained weight loss induced by nonsurgical means is unrealistic. These patients may lose weight, but, without surgery, weight regain is inevitable. The benefits of weight loss surgery are proportional to body mass as are the risks of operative complications . Several studies have shown that the risk of bariatric surgical complications is proportional to body size. This is especially true for very large patients. Prior analysis of the VA experience with bariatric surgery demonstrated that superobesity was a modifiable risk factor . For this reason, we embarked on a protocol of preoperative weight loss for our patients. We required our patients to lose between 20 and 40 lb prior to surgical intervention concurrent to their preoperative psychological and preoperative evaluation. Using this strategy, our patients lost an average of 8% of their initial weight within 3 months prior to surgery. This modality has not affected operative morbidity or mortality . These findings are similar to preoperative weight loss prior to weight reduction surgery reported retrospectively  and prospectively  in patients undergoing laparoscopic Roux-en-Y gastric bypass (LRYGB). However, these strategies are limited in that they only accounted for a modest (≤8%) preoperative weight loss. The current study addresses the effects of massive preoperative weight loss in a high-risk population.
Most patients assessed for Roux-en Y Gastric Bypass (RYGB) at the Dallas VA Medical Center have three potential factors associated with increased morbidity and mortality: advanced age, male gender, and superobesity. Superobese patients (BMI ≥ 50 kg/m2) often present for evaluation for weight reduction surgery with a burden of comorbid conditions, which makes them at a prohibitively high risk for complications of surgical interventions.
This study was undertaken to determine the feasibility of a supervised inpatient, low-calorie, liquid, protein diet in order to attain a substantial degree of weight loss (a decrease in BMI of ≥10 kg/m2) in massively obese patients with a BMI ≥ 60 kg/m2 prior to RYGB as a risk-reduction strategy. The aims of this study were to: (1) determine if massively obese VA patients were able to adhere to a supervised low-calorie diet for an average of ≥10 weeks, (2) determine if a decrease in BMI of ≥10 kg/m2 would decrease perioperative and postoperative risk for these massively obese individuals, and (3) establish if this strategy had an effect on long-term weight loss and improvement of comorbid conditions.
Six patients were identified from the Dallas VA bariatric surgery consult database. These patients were originally rejected for weight reduction surgery because they were considered to have an excessive risk of adverse events from bariatric surgery because of their size. These patients were invited to participate in an inpatient low-calorie liquid protein diet program. Patients with a BMI of ≥60 or ≥55 kg/m2 with insulin-requiring diabetes were invited to participate. This study protocol was approved by the Institutional Review Board of the Dallas VA Medical Center.
All patients were hospitalized for an average of 10 weeks prior to surgery and were placed on a low-calorie liquid diet (<900 kcal) consisting of Glucerna® three times a day with additional supplements of three packages of Beneprotein powder three times a day. Patients were instructed to combine three packets of protein supplement per can of Glucerna®. This combination provided 891 kcal/day and 83.7 g per protein per day. All participants’ were also provided with multivitamin supplementation daily. The goal of the diet was for patients to decrease their weight by ten points of body mass index or by 100 lb.
All patients underwent psychological assessment by a bariatric psychologist specialist prior to admission to the program. All prospective patients participated in a thorough psychological interview and were given a number of psychological tests, which included: a questionnaire on Weight and Eating Patterns-Revised (QWEP-R) which is intended to identify the presence or absence of bulimia and eating disorder not otherwise specified (NOS; binge eating disorder). Each patient was also given the Minnesota Multiphasic Personality Inventory-2 (MMPI-2) in order to detect any impairing psychopathology that may have existed.
All patients were enrolled in water-based exercise activities one to twice per week for at least 30 min throughout the study protocol.
Patients with diabetes were placed on an insulin sliding scale and were continued on decreased doses of their oral hypoglycemic medications with adjustments based on the fasting glucose levels throughout the study protocol.
Weight was recorded once a week. Electrolytes were measured every week for the first 3 weeks of the study and every other week thereafter. All patients with diabetes had a fasting glucose measurement in the morning and prior to their meal replacements. Patients with diabetes had hemoglobin A1c measured at the beginning of the study, prior to surgery and 3 months after surgery.
Following the ten-point reduction in BMI, the patients had an RYGB performed by one surgeon (EHL). All patients undergoing surgical intervention had an open retrocolic undivided Roux-en-Y gastric bypass with a 90–150-cm Roux limb and a 20-cm biliopancreatic limb. The alimentary limb was passed through a transverse mesocolic defect in a retrocolic, antegastric fashion. A 20-cm3 gastric pouch was created by stapling the stomach in continuity using two firings of an Ethicon 30-mm heavy-wire stapler. The mesenteric defect was closed, and the Roux-en-Y limb was secured to the mesentery. The gastrojejunostomy was created in an interrupted single-layer fashion over a nasogastric tube. The jejunojejunostomy was created in a two-layer hand-sewn fashion with running 3-0 polydioxane (PDS®; ETHICON, Livingston, UK) suture and an interrupted 3-0 silk .
Assessment of Comorbid Conditions
Improvement of obstructive sleep apnea (OSA) was determined by the discontinued used of continuous positive airway pressure (CPAP). Osteoarthritis (OA) and gastroesophageal reflux disease (GERD) were assessed subjectively and by the decreased use of medications for the management of these conditions. An improvement in hypertension was assessed by the decreased use of medications needed for the control of this disease. Dyslipidemia was assessed by both the decreased use of medications and changes in lipid profiles from before the study, prior to surgical intervention, and at the last visit postoperatively. Improvement of diabetes was determined by measurement of hemoglobin A1c prior to the start of the diet, before surgery, and during the last postoperative visit. Medication requirements for diabetes control were used to assess improvements in diabetes management.
All data were abstracted from the VA’s Computerized Medical Record System. Statistical analysis was performed by using Student’s t test for continuous variables, and Chi-squared or Fisher’s exact test was employed for categorical data where appropriate. All values are presented as means ±SE, and statistical significance is assumed when p ≤ 0.05.
Data were also extracted from the NSQIP database on all patients who underwent RYGB in 12 VA centers in the USA, excluding our institution. Data were obtained between October 1999 and September 2007. The NSQIP cohort was approximately, but not precisely matched. Patients were selected from the NSQIP cohort by having a BMI > 60. NSQIP patients had both open and laparoscopic RYGB. NSQIP patients were followed for 30 days postoperatively.
Between June 2006 and July 2007, six patients were hospitalized for an inpatient low-calorie diet program. Of the six patients originally invited to participate in the study, one was discharged from the program because of dietary noncompliance. He was a 60-year-old, 473-lb obese man with a BMI of 63 kg/m2 and a past medical history significant for diabetes, obstructive sleep apnea, and knee OA. He had not lost weight after 3 weeks of the inpatient program, resulting in his exclusion from the study. His current weight (11-month follow-up) was 485 lb, and his BMI is 64 kg/m2. The remaining patients (n = 5) were able to complete the study and underwent surgical intervention between October 2006 and November 2007. All demographic and statistical analyses were based on these five patients.
Age (years old)
54.7 ± 2.6
51.3 ± 0.9
468.8 ± 16.8
437.9 ± 7.3
64.3 ± 2.1
66.0 ± 0.8
3.2 ± 0.2
All patients received the QWEP-R which was intended to identify the presence or absence of bulimia and eating disorder NOS (binge eating disorder). None of the patients were identified as bulimic; two were identified as meeting criteria for binge eating disorder. Each patient was also given the MMPI-2 in order to detect any impairing psychopathology that may have existed. MMPI-2 results indicated that all but one patient were free of impairing psychopathology. The other was identified as meeting the criteria for PTSD by clinical history, although he was currently asymptomatic; he was also diagnosed with major depressive disorder, recurrent, in partial remission. He was followed by a team of mental health providers and, due to their ongoing involvement, was deemed to be adequately managed psychiatrically.
Weights and BMIs before diet, prior to surgery and postoperatively
N = 5
Prestudy weight and BMI
486.8 ± 16.8
64.3 ± 2.1
Weight at surgery and BMI
401.0 ± 15.5
51.6 ± 2.1
Postoperative weight and BMI
312.6 ± 15.0
41.0 ± 2.0
Weight loss during study and BMI
85.8 ± 6.0
12.9 ± 1.8
Study time (days)
98.4 ± 11.7
Postoperative follow-up (days)
220.4 ± 74.4
An average of a decrease in BMI of 12.7 points was achieved by all patients exclusively by dietary intervention. Following surgical intervention, an additional decrease by 10.6 BMI units was observed compared to the BMI prior to surgical intervention. Since the start of the study, all patients have been able to accomplish a decrease in BMI by 23.3 units.
Operative Time and LOS
There were no intraoperative complications. The intraoperative time was 133.8 ± 19.2 min (range 83 to 201 min). All patients were discharged from the hospital 3 days following surgical intervention. No postoperative complication occurred.
Assessment of comorbid conditions
Obstructive Sleep Apnea
All five patients required CPAP prior to surgical intervention. None of the patients were able to discontinue the use of CPAP during dietary intervention. However, all five patients discontinued its use within 4 weeks postoperatively.
Medications of all patients prior to the start of the diet and following surgical intervention with an average follow-up of 220 days
Medications on admission
Felodipine 10 mg QD
Glyburide 1.25/Metformin 250 mg BID
Glyburide 7.5 mg BID
Insulin 70/30 70 U QD
Insulin 70/30 80 U QD
Lisinopril 50 mg QD
Simvastatin 40 mg QHS
Rosiglitazone 8 mg QD
Metformin 850 mg TID
Doxazosin 2 mg QD
Metoprolol 50 mg QD
Citalopram 10 mg QHS
Metformin 1,000 mg BID
Doxazosin 8 mg QD
Lisinopril 10 mg QD
Torsemide 20 mg QD
Ibuprofen 800 mg PRN
Gemfibrozil 600 mg BID
Lisinopril 10 mg QD
Simvastatin 10 mg QD
Gemfibrozil 600 mg QD
Fosinopril 20 mg QQ
Metoprolol 50 mg BID
Omeprazole 20 mg QD
Diltizem 360 mg BID
Simvastatin 40 mg QD
Lisinopril 20 mg QD
Etodolac 200 mg PRN
Furosemide 40 mg QD
Omeprazole 20 mg QD
Etodolac 200 mg PRN
Hydrochlorothiazide 12.5 mg QD
Simvastatin 40 mg QHS
Glyburide 2.5 mg BID
Metformin 5,000 mg BID
Simvastatin 10 mg QD
Omeprazole 20 mg QD
Simvastatin 40 mg QD
Lisinopril 10 mg QD
Metoprolol 25 mg BID
Etodolac 200 mg PRN
Hydrochlorothiazide 12.5 mg QD
Lisinopril 20 mg QD
Gastroesophageal Reflux Disease
Subjectively, both subjects affected by GERD had an improvement of their symptoms. One had discontinued the use of omeprazole at his follow-up visit (Table 3).
All five patients were within the normotensive rage prior to, during, and after the study on their respective medications (Table 3). There were no alterations in the medications of any of the patients during the diet program. Two subjects had discontinued the used of all antihypertensive medications at the last follow-up, and three had decrease the number of medications required for blood pressure control.
There was a 1.9% decrease in the levels of hemoglobin A1c during the diet period (p ≤ 0.05; Fig. 3). No significant decrease has occurred from the time of the RYGB to the last follow-up. A total decrease by 2.6% occurred from the beginning of the study to the last follow-up (p ≤ 0.01; Fig. 2).
Of four diabetic patients, two were insulin dependent prior to the beginning of the study. However, they discontinued using insulin at the last follow-up visit. One subject is only on 500 mg of daily metformin, and the second was on no diabetic medications whatsoever during his last follow-up visit. Of the two subjects requiring oral hypoglycemic medications, one completely discontinued their use, and the second decreased the use of glyburide from 7.5 mg twice a day to 2.5 mg twice daily.
NSQIP results were obtained for all assessed bariatric cases performed in 12 VA medical centers which approved bariatric surgery, excluding Dallas. Data were obtained for operations performed between Oct. 1, 1999 and September 30, 2007. There were 1,046 bariatric procedures, of which, 77 patients had BMI ≥ 60. Of these, 67 were male and ten were female. The mean (±SE) body weight was 437.9 ± 7.3 lb with a mean BMI of 66.0 ± 0.8. The mean age for this group was 51.3 ± 0.9 years old. The 30-day mortality for this cohort was 3.9% with a complication rate of 33.8%. The mean hospital stay was 8.8 ± 0.9 days with the median being 7.0 days.
We have previously shown that morbidly obese VA patients have a substantial amelioration or elimination of obesity-related comorbid conditions following weight reduction surgery . In spite of a higher prevalence of preoperative comorbid conditions, VA patients have a similar rate of complications when compared with other non-VA bariatric cohorts . We have also shown that a small reduction in weight ≥ 8.0% was associated with a decrease in operative time in patients submitting to an RYGB compared to patients who do not lose weight preoperatively. However, the decrease in long-term weight loss or effect on comorbid conditions following a much more substantial amount of preoperative weight loss (>15% of initial body weight, BMI decrease by ten points, or >100 lb) remains unclear.
A substantial number of VA patients present to the bariatric surgery clinic with a burden of obesity-related comorbid conditions and with such a high BMI that might make these patients at a prohibitive risk for weight reduction surgery. We, therefore, initiated a pilot study to determine if massively obese patients (BMI ≥ 60 or ≥ 55 kg/m2 with insulin-dependent diabetes) would be able to lose sufficient weight (≥10-kg/m2 decrease in BMI) in a supervised inpatient diet program to sufficiently reduce the risk of complications to restore their eligibility for surgical intervention. Our secondary aims were to assess long-term weight loss and the effect on comorbid conditions.
We demonstrated that very large patients could successfully lose substantial amounts of preoperative weight on a low-calorie diet. In an average of 3.3 months, our patients lost over 80 lb, which corresponded to 44.2% excess body weight exclusively from dietary intervention. Non-VA patients undergoing open RYGB typically lose and average of 88 lb [9, 10] at 1 year following surgical intervention. We were able to achieve the same degree of weight loss with massively obese patients with an inpatient low-calorie diet. Our previous observations with our bariatric VA patients who did not lose weight preoperatively showed a weight loss of 109 ± 7 lb at 18-month follow-up, which corresponded to an excess body weight loss of 67% following surgical intervention . In the present study, we demonstrated a much greater weight loss in less than a year totaling 174.2 ± 26.9 lb (mean follow-up 7.3 months), corresponding to an excess body weight loss of 89%. For the massively obese, long-term weight loss was more substantial when patients completed a rigorous preoperative weight loss program.
A 5–7% preoperative weight loss prior to bariatric surgery has been reported for patients undergoing LRYGB [6, 7]. One study demonstrated that preoperative weight loss resulted in greater excess body weight loss 1 year postoperatively. A 5% or greater preoperative weight loss was also associated with a decrease in intraoperative time by 36 min. In a prior analysis, we had found that an 8% preoperative weight loss resulted in a 22-min decreased operative time when compared to a cohort of patients who did not lose weight (161 vs. 183 min). In the present study, the operations were completed in an average of 134 min, which was 44 min less than patients who experienced modest preoperative weight loss in our prior study. The operative times in our current study were an hour less than for patients who did not lose weight preoperatively in our previous study. Of note, all the operations were performed by the same surgeon (EHL).
There were no (perioperative) postoperative complications in our cohort of patients. All of our patients were discharged on postoperative day 3. Data obtained from the NSQIP database demonstrated a mortality of 3.9% and a morbidity of 33.8% and a hospital length of stay (LOS) of 8.8 ± 0.9 days. Although our cohort of patients is smaller compared to the 77 subjects extracted from the NSQIP database, our results suggest a substantial and drastic improvement in morbidity and mortality in superobese patients subjected to a massive weight reduction diet program and a much reduced hospital LOS.
A meta-analysis of 147 bariatric surgery studies demonstrated an improvement of hypertension in 89% of patients and dyslipidemia in 32% submitting to weight reduction surgery . Our results were similar to these observations. In our small cohort of patients, two subjects were able to discontinue their gemfibrozil following RYGB. There were no statistically significant differences in the lipid profiles of our patients before or after treatment. All of our patients had decreased need for antihypertensive medications. All patients had resolution of their sleep apnea.
Meta-analyses investigating the effect of weight reduction surgery on diabetes have documented an improvement in 64–100% of patients . Since Pories’ seminal observation that bariatric surgery dramatically improves diabetes , there has been mounting interest in the effects of obesity surgery on diabetes. Many studies have suggested a direct effect of surgical intervention on insulin metabolism [11–25]. However, a 7% reduction in body weight by diet and exercise reduced the incidence of diabetes by 58% in patients with impaired glucose tolerance tests , suggesting that weight loss itself rather than a metabolically induced rearrangement results from weight reduction surgery. Because weight reduction by dietary intervention is not sustained for a significant period of time in the morbidly obese patient, the effects of diet on diabetes are not comparable to the weight reduction achieved by bariatric surgery. In our pilot study, we found that patients with diabetes were able to achieve a 1.9% decrease in hemoglobin A1c by dietary intervention alone. There was no further improvement in the hemoglobin A1c levels following RYGB, suggesting that reduced food ingestion causing weight loss is substantially responsible for the improvement in diabetes. These findings support the notion that lipotoxicity resulting from overnutrition causes diabetes .
Our cohort of five patients had similar characteristics as did the National VA system for patients with BMI ≥ 60 kg/m2 with respect to weight, age, and the predominance of male gender (Table 1). In contrast to our study, the NSQIP database demonstrated a substantially high mortality rate for superobese VA patients. Additionally, one third of these patients experienced complications related to weight reduction surgery. Although the small number of patients we treated in our dietary program precludes definitive comparison to the larger cohort of patients, the significant decrease in operative times, the decreased hospital LOS, and the absence of (perioperative) postoperative complications in these five patients suggest that massive preoperative weight loss results in a substantially less challenging operation than would be expected for patients with a BMI ≥ 60 kg/m2.
Our results suggest that bariatric surgery consistently ameliorates or eliminates obesity-related comorbid conditions. Massive preoperative weight loss is possible in very large patients by implementation a low-calorie diet. This strategy allows decreases in operative time and hospital LOS and is associated with substantial long-term weight loss. This modality also permits patients previously considered ineligible for bariatric surgery to proceed with weight reduction surgery. More emphasis should be placed on increasing bariatric surgery programs in Veteran Administration hospitals with substantial preoperative weight loss requirements.
This pilot study undertook a great deal of effort by a number of individuals who in addition to their routine busy activities took time to assist with the program. The authors specially acknowledge the participation and help of: Elizabeth Hayden, MS/RD; David Hales, MD; Theresa Hale, Ph.D.; Daniel Villarreal, PA; Emily Armour-Serold, PA; Wanda Frey, RN; Catherine Gordon, MD; Rhonda Harmon, MD; Ali Siddiqui, MD; and KC Murphy.