Surgery for Obesity and Related Diseases 7 (2011) 516 –525

Original article
Laparoscopic Roux-en-Y gastric bypass: 10-year follow-up
Kelvin Higa, M.D., F.A.C.S.a,b,c, Tienchin Ho, M.D., F.A.C.S.a,, Francisco Tercero, M.D.a,b,
Tahir Yunus, M.D.a,b, Keith B. Boone, M.D., F.A.C.S.a,b,c
a
Advanced Laparoscopic Surgery Associates, Fresno, California
b
University of California, San Francisco-Fresno, Fresno, California
c
Fresno Heart and Surgical Hospital, Fresno, California
Received November 9, 2009; accepted October 25, 2010

Abstract Background: The short-term benefits of bariatric surgery are well documented; however, few
reports with data beyond 10 years exist. Those that have been published have described only open
procedures. We present our 10-year follow-up results with laparoscopic Roux-en-Y gastric bypass
with hand-sewn gastrojejunal anastomosis in a group private practice.
Methods: We performed an institutional review board-approved retrospective review of a prospec-
tively maintained database, combined with office visits and telephone questionnaires, for patients
who underwent laparoscopic Roux-en-Y gastric bypass between February 1998 and April 1999.
Results: A total of 242 patients underwent surgery from February 1998 to April 1999. The office
follow-up rate was 33% at 2 years and 7% at 10 years. An additional 19% had telephone follow-up
at 10 years. The mean excess weight loss was 57% at 10 years. Of the 242 patients, 65 (33.2%) failed
to achieve an excess weight loss of ⬎50%; 86 (35%) had ⱖ1 complication during follow-up.
However, 83%, 87%, 67%, and 76% of patients with diabetes, hypertension, dyslipidemia, and
obstructive sleep apnea, respectively, experienced improvement or resolution. The internal hernia
rate was 16%, and the gastrojejunal stenosis rate was 4.9%. No surgery-related deaths occurred. Of
the 242 patients, 136 (51%) had nutritional testing at least once after postoperative year 1. Of these
136 patients, only 24 (18%) had remained nutritionally intact during follow-up.
Conclusion: The obstacles to follow-up have continued to impede the collection of accurate
long-term data. Of the 26% of patients with data, laparoscopic Roux-en-Y gastric bypass provided
sustainable weight loss and resolution of co-morbidities. However, nutritional deficiencies presented
sporadically over time and underscore the importance of routine testing. (Surg Obes Relat Dis 2011;
7:516 –525.) © 2011 American Society for Metabolic and Bariatric Surgery. All rights reserved.
Keywords: Adolescent; Adult; Anemia; Bariatric; Body mass index; Calcium; Co-morbidity; Complications; Diet; Folate;
Follow-up studies; Gastric bypass; Humans; Iron deficiency; Malabsorption syndromes; Nutrition; Obesity;
Complications; Surgery; Parathyroid hormone; Postoperative complications; Quality of life; Reoperation; Ret-
rospective studies; Serum albumin; Treatment outcome; Vitamin B12; Vitamin B6; Weight loss

Despite extensive positive short-term results and the

lack of an alternative medical therapy, critics of bariatric
surgery have cited the scarcity of long-term results and
unknown potential complications. The criticism is valid:
“Why are more long-term published results not available?”
Presented during the Plenary Session at the 28th American Society of Currently, the longest laparoscopic gastric bypass series
Metabolic and Bariatric Surgery Annual Meeting, June 2009. reported 5-year data [1].
*Reprint requests: Tienchin Ho, M.D., Medical Group, Inc., Advanced
Laparoscopic Surgery Associates, 6121 North Thesta Street, Fresno, CA
After 11 years and ⬎8000 patients, we thought it time to
93710. reflect on our experience and lessons learned regarding
E-mail: tienchinho@yahoo.com laparoscopic Roux-en-Y gastric bypass.
1550-7289/11/$ – see front matter © 2011 American Society for Metabolic and Bariatric Surgery. All rights reserved.
doi:10.1016/j.soard.2010.10.019
 K. Higa et al. / Surgery for Obesity and Related Diseases 7 (2011) 516 –525 517

Methods patient received a recent assessment by a physician, we

obtained the office notes.
The institutional review board approved our retrospec- Before surgery, the patients agreed to lifelong, annual
tive review of a prospectively maintained database for pa- follow-up. The decline in patients followed over time is
tients who underwent laparoscopic Roux-en-Y gastric by- presented in 2 ways. The first is a Kaplan-Meier curve.
pass from February 1998 to April 1999. The data points Starting with 100% of the patients, the numerator of this
included weight statistics, complications, resolution of co- fraction decreases if patients have no additional follow-up.
morbid conditions, and nutritional parameters. The denominator decreases as patients either die or reach
The database contained the preoperative weight and sta- the maximal number of postoperative days possible. Since
tus of co-morbid conditions, the postoperative weight, post- eventually, no patients are available for follow-up, the curve
operative status of co-morbid conditions, complications, decreases to 0. With the second method, follow-up is pre-
and nutritional laboratory results. We reviewed the office sented as the number of patients followed up divided by the
charts to find any missing information. In cases in which we total number of patients eligible for follow-up during each
could not locate our original office chart, we requested the postoperative year (POY). These 2 methods differ because
hospital operative notes and preoperative history and phys- some patients were seen 1 year and then again several years
ical examination results. For 53 patients, the information later. Such a patient is counted as a follow-up during both
was obtained solely through the database because the office years using the second method and as a “lost” patient in
and hospital charts were unavailable. intervening years. In the first method with a Kaplan-Meier
All patients underwent primary laparoscopic Roux-en-Y curve, however, this patient would only be lost to follow-up
gastric bypass by 1 of 2 surgeons (K.H. and K.B.). The at the most recent follow-up visit.
pouch size, orientation, and intestinal lengths have remained Weight loss was reported as the percentage of excess
consistent, but incremental changes have occurred in the weight loss (%EWL), the percentage of weight lost, post-
technique of internal hernia closure, which have decreased operative BMI, and percentage of excess BMI lost. For each
our overall morbidity. The present cohort represents our patient, the average of all the weights measured during each
early experience and, thus, our learning curve. POY was obtained. For each POY, the average weights of
Retrocolic, antegastric Roux-en-Y gastric bypass has all patients measured in that year were then averaged to
been previously described [2] and differed from other de- calculate the average %EWL, percentage of weight lost,
scriptions with respect to the orientation of the gastric pouch postoperative BMI, and percentage of excess BMI lost for
and construction of the gastrojejunal anastomosis. The that POY.
pouch was vertically constructed and based on the lesser Inadequate weight loss was defined using the criteria of
curve, as originally described by MacLean et al. [3], and not Halverson and Koehler [8] of ⬍50% %EWL or the criteria
the horizontal orientation of many current techniques [4 – 6]. of Biron et al. [9] of ⬍40 kg/m2 postoperative BMI for
A 100-cm Roux limb was used for patients with a preop- patients with a preoperative BMI of ⱖ50 kg/m2 and ⬍35
erative body mass index (BMI) ⬍50 kg/m2; a 150-cm limb kg/m2 postoperative BMI for patients with a preoperative
was used for those with a preoperative BMI ⬎50 kg/m2. BMI of ⬍50 kg/m2.
The gastrojejunostomy was hand sewn with absorbable su- At the time, no risk stratification criteria existed, so we
ture material to emulate the “open” procedures of the day created our own (Table 1) to assess the preoperative co-
[7], because we did not know whether small changes would morbid conditions and to grade subsequent changes as re-
alter the outcomes for the laparoscopic procedure. solved, improved, recurred, unchanged, or worse. The pre-
An automatic recall system was used to arrange the operative information used in our analysis was consistently
annual office or telephone follow-up. At the patient’s last assessed prospectively. However, the prospective recording
appointment, the patient either scheduled another appoint- of subsequent follow-up information was inconsistent.
ment or their name was placed on a recall list to be con- Therefore, the most recent information available for each
tacted later for their annual appointment. The office staff patient from the database, chart, or letter from the primary
made 3 attempts by mail or telephone, or both, to contact care physician regarding the patient’s condition status was
patients for missed appointments. As much as possible, we used in our analysis.
coordinated care with the primary care physician, especially The patients were advised to take a daily multivitamin
for patients who had moved out of our area or were inac- with iron after surgery. Beginning in 2000, we also recom-
cessible because of insurance. mended 1000 ␮g vitamin B12 and 1500 mg calcium citrate
Patients who were unavailable for office follow-up visits as daily supplementation. Actigall was prescribed for 6
were interviewed by telephone regarding their current months after surgery to help prevent gallstone formation.
weight, state of health, quality of life, co-morbid conditions, Annual nutritional surveys included serum albumin, vitamin
and signs and symptoms related to potential long-term com- B12, vitamin B6, calcium, folate, hemoglobin, and intact
plications. If patients indicated they had recent nutritional parathyroid hormone. The normal ranges are listed in Table
laboratory studies done, we obtained the results. If the 2. In addition, when symptoms suggested a problem, such as
518 K. Higa et al. / Surgery for Obesity and Related Diseases 7 (2011) 516 –525

Table 1
Stratification of co-morbid conditions
Co-morbidity Stratification of disease severity

Mild Moderate Severe

Osteoarthritis Occasional need for pain Daily use of pain medications Need for mobility support, such as cane/walker/
medications wheel chair
Diabetes Diet controlled Controlled with oral medications Requires insulin
Dyslipidemia Diet controlled Controlled with medications Uncontrolled despite medications
Hypertension Diet controlled Controlled with ⬍2 medications Controlled with ⱖ2 medications
Sleep apnea Mild symptoms only Diagnosed with sleep study but no CPAP Diagnosed with sleep study and need for CPAP
Cardiovascular disease Mild disease History of angina History of cardiac intervention
Asthma Occasional inhaler Daily inhaler Systemic steroids
GERD Occasional medications Controlled with daily acid blocking Uncontrolled despite daily prescription
medications medications
Urinary stress incontinence Occasional Need for wearing pad daily Continuous incontinence
Infertility Menstrual irregularities Controlled PCOS symptoms Uncontrolled PCOS symptoms
Varicose veins Mild discomfort Healed varicose ulcers Unhealed ulcers

CPAP ⫽ continuous positive airway pressure; GERD ⫽ gastroesophageal reflux disease; PCOS ⫽ polycystic ovary disease.

anemia, we pursued focused testing. Identified deficiencies Figure 1 illustrates the decline in the number of patients
were corrected. All complications and nutritional results followed up over time using a Kaplan-Meier curve. Figure
identified were included in our analysis. 2 presents the number of patients followed up during each
As an additional outcomes assessment, we used the orig- POY. A total of 196 patients (81%) had follow-up beyond
inal Moorehead-Ardelt quality of life questionnaire [10]. the first POY. Sixteen patients (7%) were seen in the office
The questionnaire compared self-esteem, physical activity, within POY 10. An additional 49 patients (20%) were fol-
social life, work conditions, and sexual activity before and lowed up by telephone at 10 years. Although 6 patients were
after bariatric surgery. eligible to follow-up 11 years after surgery, only a single
Using the Bariatric Analysis and Reporting Outcome patient did so. This single weight measurement did not
System (BAROS) described by Oria and Moorehead [10], permit statistical conclusions and was excluded from our
we combined the data on weight loss, co-morbid conditions, analysis.
complications, and quality of life to calculate the BAROS
score. Weight loss
The average %EWL at 10 years after surgery was
Results 57.1% ⫾ 23.1% and showed a decreasing trend over time
(Fig. 2). The differences in the %EWL between the POYs
From February 1998 to April 1999, 242 patients under- were not statistically significant, except between POYs 1
went primary laparoscopic Roux-en-Y gastric bypass. Of
the 242 patients, 13 required conversion to open surgery. Table 3
Gender, age, and preoperative BMI were typical for most Preoperative demographic characteristics
reported series (Table 3). However, 65 of our patients (27%) Characteristic Patients (n) % of 242
had a preoperative BMI ⬎50 kg/m2 and 14 (6%) had a BMI
Gender
⬎60 kg/m2, which is unusual for most laparoscopic series at
Female 197 81
that time [11]. Male 45 19
Age (y)
Table 2 ⬍20 7 3
Definition of normal range for tested nutritional parameters 20–30 12 5
30–40 71 29
Parameter Normal Range 40–50 84 35
50–60 57 24
Male Female
60–70 8 3
Albumin (g/dL) 3.5–5.0 3.5–5.0 ⱖ70 3 1
Vitamin B12 (pmol/L) ⬎400 ⬎400 Preoperative BMI (kg/m2)
Vitamin B6 (ng/mL) 5–30 5–30 ⬍40 46 19
Calcium (mg/dL) 8.5–11 8.5–11 40–50 131 54
Folate (ng/mL) 4–20 4–20 50–60 51 21
Hemoglobin (g/dL) 14–17 12–16 ⱖ60 14 6
Intact parathyroid hormone (pg/mL) 10–65 10–65
BMI ⫽ body mass index.
 K. Higa et al. / Surgery for Obesity and Related Diseases 7 (2011) 516 –525 519

A total of 65 patients (33% of 196) had ⬍50% EWL at

their most recent follow-up visit. Of these, 48 (74%) had not
attained 50% EWL even initially and 17 (26%) had ⱖ50%
EWL initially but had then regained the weight. Of the 242
patients, 12 (5%) had ⬍50% EWL during the first POY but
continued to lose weight and had attained ⬎50% EWL at
their most recent follow-up visit. Using the criteria of Biron
et al. [9], 47 patients (24% of 196) had inadequate weight
loss; 8 patients underwent secondary procedures for weight
recidivism.
Of the 65 patients weighed during POY 10, 23 (35%) had
⬍50% EWL during POY 10. Of 17 patients with a preop-
erative BMI ⱖ50 kg/m2, 6 (34%) had a %EWL ⬍50% at
POY 10, and of 48 patients with a preoperative BMI ⬍50
kg/m2, 17 (35%) had a %EWL ⬍50% at POY 10. Using the
criteria of Biron et al. [9], 13 patients (20% of 65) had
inadequate weight loss.
Thirteen patients lost more than their ideal body weight.
Fig. 1. Percent Follow-up over Time. Of these 13, 6 were clinically and nutritionally intact, 2 had
marginal ulcers that resolved with treatment, 1 had symp-
tomatic cholelithiasis and an internal hernia that resolved
and 10 (P ⫽ .008). The %EWL was significantly better
with treatment, 2 died of suspected or confirmed malig-
(P ⬍ .001, analysis of variance) for the 177 patients with a
nancy, and 2 were lost to follow-up.
preoperative BMI ⬍50 kg/m2 than for the 65 patients with
a preoperative BMI of ⬎50 kg/m2 (Fig. 3). The %EWL
Co-morbid conditions
decreased for each increasing preoperative BMI decentile
(Table 4). No statistically significant difference (P ⫽ .619, The follow-up rate and most recent assessment of each
t test) was found in the %EWL for patients who were seen condition for all study patients and for the patients evaluated
in the office (53.8%) and those reporting by telephone during POY 10 are listed in Table 5. The percentage of
(57.2%). At 10 years, the average percentage of weight lost follow-up varied according to the co-morbid condition be-
was 29% ⫾ 11.3%, the average postoperative BMI was cause of the sporadic way in which patients returned to the
33 ⫾ 8.0 kg/m2, and the average percentage of excess BMI office. In addition, documentation of co-morbid conditions
lost was 66.4% ⫾ 27.9% (Table 4). was inconsistent. Of the 242 patients, only 29 (12%) had no

Fig. 2. Average %EWL with standard deviation. Each point is labelled with the number of patients weighed during each postoperative year. All 242 patients
were eligible for follow-up through postoperative year ten.
520 K. Higa et al. / Surgery for Obesity and Related Diseases 7 (2011) 516 –525

Fig. 3. Average percent excess weight loss over time for patients with preoperative BMI ⬍ 50 kg/m2 compared to patients with preoperative BMI ⬎⫽ 50
kg/m2

existing preoperative co-morbid condition. For 166 patients which 19 (48% of 39 patients) involved the mesocolon and
(69% of 242), no information could be obtained regarding 15 (38%) did not. One was due to a band adhesion, and 4
the co-morbid conditions postoperatively. patients did not have bowel within the potential hernia
space. Of the 242 patients, 70 (29%) had 71 complications
Complications
other than internal hernias.
Although no operative deaths occurred, 8 patients died; Early major morbidity included staple line failure and a
all the deaths were unrelated to gastric bypass (Table 6). Of thermal injury with perforation that resulted in gastrointes-
the 242 patients, 86 (35%) experienced 109 complications. tinal leakage. The staple line failures occurred along staple
Of the 65 patients who were evaluated during POY 10, 31 lines involved in the formation of the gastric pouch. No
(45%) had 31 complications (Table 7). The complications leaks occurred at the gastrojejunal anastomosis, which was
were predictably cumulative and included a high percentage always hand sewn. Gastrojejunal stenosis (4.9%) usually
of internal hernias, characteristic of our early experience. presented within the first postoperative month and was eas-
Most of the internal hernias involved the mesentery, of ily treated with endoscopic balloon dilation.

Table 4
Average %EWL stratified by preoperative BMI decentile and average %EWL, postoperative BMI, percentage of weight lost, and percentage of excess
BMI lost for all 242 patients
POY Average %EWL for each preoperative BMI (kg/m2) group For all 242 patients

⬍40 (n ⫽ 46) 40–50 (n ⫽ 131) 50–60 (n ⫽ 51) ⱖ60 (n ⫽ 14) %EWL %WL Postoperative BMI %EBMIL

1 85 67 59 44 67.8 ⫾ 19.0 34.0 ⫾ 7.5 30.9 ⫾ 6.6 78.3 ⫾ 24.6

2 82 65 62 52 66.9 ⫾ 20.5 33.3 ⫾ 9.1 30.4 ⫾ 6.2 78.0 ⫾ 25.7
3 73 63 61 62 64.4 ⫾ 22.1 32.7 ⫾ 11.0 31.0 ⫾ 6.3 74.4 ⫾ 26.3
4 70 69 49 57 65.2 ⫾ 23.1 32.5 ⫾ 10.7 31.0 ⫾ 7.2 76.1 ⫾ 28.9
5 81 69 53 — 68.6 ⫾ 19.6 33.8 ⫾ 8.3 29.9 ⫾ 5.6 79.8 ⫾ 25.0
6 68 64 56 — 64.4 ⫾ 23.2 31.9 ⫾ 9.7 30.8 ⫾ 5.7 75.0 ⫾ 29.5
7 82 57 70 — 62.8 ⫾ 26.4 30.6 ⫾ 11.4 31.2 ⫾ 7.4 73.9 ⫾ 33.9
8 66 54 51 — 56.1 ⫾ 27.1 28.5 ⫾ 12.5 33.7 ⫾ 7.5 64.1 ⫾ 32.6
9 57 57 66 — 58.6 ⫾ 21.3 30.5 ⫾ 10.7 33.2 ⫾ 8.2 67.0 ⫾ 24.9
10 66 55 57 37 57.1 ⫾ 23.1 28.8 ⫾ 11.3 33.2 ⫾ 8.0 66.4 ⫾ 27.9

%EWL ⫽ percentage of excess weight loss; BMI ⫽ body mass index; %WL ⫽ percentage of weight loss; %EBMIL ⫽ percentage of excess BMI lost.
Data presented as numbers of patients or mean ⫾ standard deviation.
 K. Higa et al. / Surgery for Obesity and Related Diseases 7 (2011) 516 –525 521

Table 5
Outcomes of co-morbid conditions for 242 study patients and 51 patients evaluated during postoperative year 10
Co-morbid condition Patients (n) % of 242 242 Study patients 51 Patients with 10-y follow-up

Follow-up (%) Resolved or improved (%) Follow-up (%) Resolved or improved (%)

Osteoarthritis 110 45 35 84 100 78

Diabetes 45 19 27 83 75 67
Dyslipidemia 6 2 100 67 100 80
Hypertension 108 45 36 87 100 86
Infertility 5 2 40 50 100 100
Obstructive sleep apnea 45 19 47 76 95 79
Asthma 23 10 30 100 100 100
Gastroesophageal reflux disease 121 50 36 89 94 90
Urinary stress incontinence 35 14 46 69 92 55
Varicose veins 21 9 29 100 63 100

Late complications included cholelithiasis (6.2%), de- An exception to this was folate deficiencies that appeared to
spite routine prescription of Actigall for 6 months after develop only after POY 7 (Table 9).
surgery, and marginal ulcers (4.5%), primarily seen in pa- Anemia, albumin, vitamin B6, calcium, and folate defi-
tients who used tobacco and/or nonsteroidal anti-inflamma- ciencies were statistically significantly more prevalent for
tory drugs. Seven patients (2.9%) developed substance patients with a preoperative BMI ⬎50 kg/m2 (P ⬍ .05).
abuse. Five of these patients disclosed 10 years after surgery Although vitamin B12 and intact parathyroid hormone ab-
that the dependency had been present before surgery. We do normalities were also more prevalent in patients with a
not knowingly offer surgery to patients with active sub- preoperative BMI ⬎50 kg/m2, the differences were not
stance abuse. statistically significant (P ⬎ .05; Table 10).
Nutritional outcomes
On average, for each POY, only 12% of patients ob- Table 7
tained nutritional testing. Of those tested, 66% were found Complications of 242 study patients and 65 patients evaluated during
to have a deficiency. Each year, we lost 5% to future POY 10
nutritional follow-up. Of the 242 patients, 136 (51%) had Morbidity 242 Study 65 Patients
nutritional testing at least once after POY 1 (Table 8). Only patients, n (%) evaluated during
24 (18%) were nutritionally intact throughout their fol- POY 10, n (%)
low-up period. For the 23 patients tested during POY 10, 12 Early
(52%) were found to have a deficiency. No statistically Incomplete division of stomach 4 (1.7)
significant difference was found in the occurrence of nutri- Staple line failure 2 (.8)
Thermal injury with perforation 1 (.4)
tional deficiencies between those patients who were seen in Marginal ulcer perforation 1 (.4)
the office and those reporting by telephone (P ⬎ .167 for all Bleeding, observation 1 (.4) 1 (1.5)
parameters). Deep venous thrombosis 1 (.4)
The abnormalities in every parameter tested were evenly Stenosis, gastrojejunostomy 12 (5.0) 4 (6.2)
distributed through the POYs. That is, patients many years Stenosis, mesocolon 1 (.4)
Fever 1 (.4)
after surgery seemed to have the same risk of developing Readmission 5 (2.1) 1 (1.5)
deficiencies as those who had undergone surgery recently. Hypoglycemia 1 (.4) 1 (1.5)
Central pontine myelinolysis 1 (.4)
Subtotal 31 (12.8) 7 (10.8)
Table 6 Late
Mortality Internal hernia 39 (16.1)
Marginal ulcer 11 (4.5) 5 (7.7)
Cause of death Postoperative interval
Gastrogastric fistula 1 (.4) 1 (1.5)
Asthma 5 mo Biliary requiring cholecystectomy 17 (7.0) 12 (18.5)
Car accident 1 y Alcohol dependency 6 (2.5) 5 (7.7)
Shot by police 4 y Other substance abuse 1 (.4)
Staph sinus infection 5 y Hernia, trocar 3 (1.2) 1 (1.5)
Malignancy 7 y Subtotal 78 (32.2) 24 (36.9)
Bacterial pneumonia 8 y Total 109 (45.0) 31 (47.7)
Ovarian cancer 9 y
Myocardial infarction 10 y POY ⫽ postoperative year.
Data in parentheses are percentages.
522 K. Higa et al. / Surgery for Obesity and Related Diseases 7 (2011) 516 –525

Table 8 data from Wittgrove and Clark [1] suggested that laparo-
Cumulative incidence of nutritional deficiencies between POY 1 and 10 scopic procedures will provide the same weight loss and
for 136 study patients tested
amelioration of co-morbidities as open procedures. Cur-
Parameter Patients with % of 136 rently, although the published data describes a certain
deficiency (n) patients tested
amount of weight regain [18,20], how much is usual and
Vitamin B12 81 60 when it occurs is unclear. Thus, comparisons of differing
Hemoglobin 71 52 bariatric procedures and approaches will require ever-longer
Albumin 46 34
Intact parathyroid hormone 36 26 periods of follow-up.
Vitamin B6 23 17 Our average %EWL was 57.1% ⫾ 33.9% at 10 years.
Calcium 18 13 We found no statistically significant difference in the
Folate 2 1 %EWL for each POY compared with POY 1 until we
POY ⫽ postoperative year. reached POY 10. Whether significant weight regain will
occur after POY 10 remains to be seen. Table 13 shows
our patients’ weight loss compared with that of other
Moorehead quality of life questionnaire long-term studies of the open gastric bypass.
The Moorehead quality of life questionnaire completed Using Halverson and Koehler’s definition, approxi-
the BAROS scoring system in 51 patients (21%) (Table 11). mately 1 of 3 patients had inadequate weight loss in our
Our average BAROS score was 3.9 ⫾ 2.6 (Table 12). series. Using the definition from Biron et al. [9], approxi-
Fourteen percent of the 51 patients were scored as failures. mately 1 of 4 patients had inadequate weight loss. Combin-
ing information from the other outcome measures in a
BAROS score reduced the number of patients with poor
Discussion outcomes to 14% at 10 years.
Although the current bariatric procedures have been Similar to those of other studies, we observed a greater
available for nearly 40 years, there are few reports of the %EWL for patients with a preoperative BMI ⬍50 kg/m2
long-term (⬎10 years) results [12,13]. O’Brien et al. [14] than for patients with a preoperative BMI ⬎50 kg/m2. The
reported medium-term (3–10-year) weight loss data in a question remains as to how to advise our “super-obese”
2006 systematic review. Evaluations of gastric bypass in- patients. Altering the limb length might improve the weight
cluded reports by Jones [15] of 71 patients at 10 years, loss but may also increase the risk of nutritional deficiencies
Pories et al. [16] of 158 patients at 10 years, Sugerman et al. [21]. Brolin et al. [22] showed better weight loss with
[17] of 135 patients at 10 –12 years, Sjostrom et al. [18] of long-limb bypass through 3 years of follow-up; however,
34 patients at 10 years, and Christou et al. [19] of 161 Christou et al. [19] did not at 10 years of follow-up. Our
patients at 12 years. retrospective data showed that patients who underwent
The laparoscopic era in bariatric surgery brought new 150-cm limb bypass (preoperative BMI ⱖ50 kg/m2) did
ideas, resources, and patient interest that rapidly increased have more deficiencies than those with the 100-cm limb
the number of procedures performed and sharply reduced bypass (preoperative BMI ⬍50 kg/m2).
short-term morbidity. Differences in the size and shape of Our results were similar to the meta-analysis by Buch-
the gastric pouch and anastomotic size raised doubt regard- wald et al. [23] with respect to the resolution or improve-
ing the efficacy of the laparoscopic approach. The 5-year ment of diabetes, hypertension, hyperlipidemia, and ob-

Table 9
Percentage of patients tested found to have deficiency in each postoperative year
POY Albumin (%) Vitamin B12 (%) Vitamin B6 (%) Calcium (%) Folate (%) Hemoglobin (%) Parathyroid hormone (%)

0 26.9 68.4 11.1 0 0 34.2 NA

1 30.8 81.1 9.7 12.5 0 38.1 NA
2 29.2 70.0 22.2 23.5 0 37.9 75.0
3 13.6 70.8 20.0 12.0 0 44.8 50.0
4 38.5 53.8 25.0 21.9 0 45.2 22.2
5 18.5 60.7 18.2 15.4 0 35.7 29.2
6 23.8 60.0 5.0 8.0 0 27.3 40.9
7 25.0 75.0 0 0 0 33.3 16.7
8 22.2 62.5 0 12.5 14.3 30.0 75.0
9 27.8 52.6 22.2 16.7 6.3 36.4 44.4
10 22.2 58.3 16.7 14.3 8.3 25.0 60.0
Average 27.9 71.3 15.0 13.7 2.9 38.8 41.3

POY ⫽ postoperative year.

 K. Higa et al. / Surgery for Obesity and Related Diseases 7 (2011) 516 –525 523

Table 10
Comparison of nutritional deficiencies between patients with preoperative BMI ⬍50 kg/m2 and preoperative BMI ⱖ50 kg/m2
Parameter Preoperative BMI ⬍50 kg/m2 Preoperative BMI ⱖ50 kg/m2 P value

Total Deficient % Total Deficient %

Intact parathyroid hormone 114 43 38 33 13 39 ⬎.5

Vitamin B12 189 118 62 55 40 73 .1596
Calcium 177 19 11 62 13 21 .0417
Folate 177 0 0 51 3 6 .0081
Vitamin B6 140 15 11 45 15 33 ⬍.001
Albumin 294 42 14 89 30 34 ⬍.001
Hemoglobin 355 83 23 112 51 46 ⬍.001
Total 1,446 320 22 447 165 37 ⬍.001

P ⬍ .05 significant.

structive sleep apnea. We were unable to obtain follow-up underscores the need for long-term routine laboratory test-
information for the 7 patients with cardiovascular disease. ing and vitamin supplementation. In our study, 82% of those
We found the follow-up of co-morbid conditions to be patients tested were deficient at some point during follow-
difficult to score consistently and reproducibly using up. Similarly, Gasteyger et al. [24] found that by 2 years
BAROS, especially when multiple conditions had conflict- after surgery, 98% of patients tested required supplements.
ing outcomes. Most active receptor-mediated absorption of iron, folate,
Our complications were cumulative and represented our calcium, and protein occurs in the proximal small bowel.
early experience—the beginning of our learning curve. De- Shortening the section of proximal small bowel exposed to
spite, this, the occurrence of major perioperative complica- the secretions required for absorption lowers the absorptive
tions was stable, including a gastrojejunal stenosis rate of capacity for these micronutrients. Acid pH-sensitive intrin-
nearly 5%. Rare with open bypass, the high incidence of sic factor-mediated vitamin B12 absorption is affected by
internal hernias required modifications in our mesenteric fundal exclusion with creation of the gastric pouch. Vitamin
closure technique. The incidence decreased to ⬍1% with B12, however, is absorbed in the ileum. Therefore, length-
the adoption and use of continuous, nonabsorbable suture ening the distance between the gastrojejunostomy and the
fixation. Biliary tract disease was included as a complica- jejunojejunostomy (alimentary limb) should not change vi-
tion; however, its prevalence in the general population prob- tamin B12 absorption. We observed that patients with an
ably surpasses that associated with surgical weight loss. alimentary limb of 150 cm did have increased rates of
Other complications, such as marginal ulceration and sub- anemia and folate, calcium, and albumin deficiency com-
stance addiction, were probably underreported and contrib- pared with those with an alimentary limb of 100 cm. Vita-
uted to the long-term, rather than the short-term, complica- min B12 and intact parathyroid hormone abnormalities were
tion rate. Complications such as wound infections and not significantly different statistically between the groups.
incisional hernias common in the open experience were rare Vitamin B6 is absorbed by passive diffusion along the
in our laparoscopic series. entire length of the small bowel. This suggests that patients
Tracking nutritional deficiencies is a continuing chal- with a 150-cm alimentary limb should have rates of vitamin
lenge because of the sporadic manner in which they occur. B6 deficiency similar to those with a 100-cm alimentary
Although an unknown number of patients declined to follow limb, as was documented by Brolin’s randomized study
the recommendation for a daily multivitamin, our findings [21]. However, this was not the case with our patients. Our
were probably not unlike the experience of other bariatric patients were not randomized, and other differences be-
practices. The high overall incidence of nutritional deficien- tween the 2 groups of patients related to preoperative BMI
cies despite the recommendation for a daily multivitamin
Table 12
Complete BAROS scores for 51 patients
Table 11
BAROS score Patients, n (%)
Moorehead quality of life questionnaire
Failure 7 (14)
Score Patients (n) % of 51 patients
Fair 9 (18)
Greatly diminished ⫺3 to ⫺2.1 0 0 Good 14 (27)
Diminished ⫺2 to ⫺1.1 5 10 Very good 15 (29)
Minimal to no change ⫺1 to ⫹1 4 8 Excellent 6 (12)
Improved ⫹1.1 to ⫹2 28 55
Greatly improved ⫹2.1 to ⫹3 14 27 BAROS ⫽ Bariatric Analysis and Reporting Outcome System
Data in parentheses are percentages.
524 K. Higa et al. / Surgery for Obesity and Related Diseases 7 (2011) 516 –525

Table 13
Comparison of weight loss findings from long-term studies of Roux-en-Y gastric bypass
Investigator Patients (n) Follow-up (y) Patients eligible Patients at %EWL Weight lost (%) Postoperative
for follow-up (n) follow-up, n (%) BMI (kg/m2)

Jones [15] 352 10 71 36 (51) 62 NR 30

Pories et al. [16] 608 10 NR 158 (NR) 55 NR 35
Sugerman et al. [17] 1025 10–12 361 135 (37) 52 28 36
Sjostrom et al. [18] NR 10 NR 34 (NR) NR 34 NR
Christou et al. [19] 272 11.6–12.3 272 161 (59) 68 NR 38
Present study 242 10 242 65 (27) 57 29 33

%EWL ⫽ percentage of excess weight loss; NR ⫽ not reported.

Data in parentheses are percentages.

could also have contributed to the differences in deficien- their parents, who are less likely to move. Many patients are
cies. embarrassed by weight regain and might not want to be
The prevalence of vitamin deficiencies is highly variable perceived as a “failure.” Others desire more weight loss or
in the published data [25,26], likely reflecting differences in have become unhappy with weight recidivism.
the definitions of deficiency, as well the multifactorial na- Given our volume and declining reimbursement, we will
ture of this problem. Dietary choices, exposure to daylight, likely be unable to meet the American Society of Metabolic
and even unintentional supplement intake, such as antacids and Bariatric Surgery Surgical Review Corporation require-
or fortified cereal, defy reporting. In addition, symptoms ment of a 75% long-term follow-up rate without committing
associated with mild micronutrient deficiencies are often significant additional resources specific to data manage-
vague and make assessment of the clinical significance ment. The volume of patients presenting for follow-up will
difficult. Anemia, however, can be used as an objective require a completely different strategy than what we are
reflection of nutritional sufficiency. In our study, more than doing currently.
one half of the patients tested had anemia between POYs 1
and 10. Including this large number of nutritional deficien-
cies in the BAROS scoring system still resulted in a lower, Conclusion
14% failure rate than the 25% failure rate defined using the
weight criteria alone. Laparoscopic Roux-en-Y gastric bypass for morbid obe-
With only a 26% follow-up rate at 10 years, the obvious sity can deliver sustainable medium-term weight loss and
question is what happened to the other 74% of patients. No control of obesity-related co-morbidities safely. Follow-up
assumptions could be made regarding capturing a greater is essential to identify and treat nutritional deficiencies but
number of patients who had a complication or, in contrast, is not associated with better weight loss. Although our goal
who did not return because they were experiencing difficul- has been to improve the health and quality of life of our
ties. Between the patients who were followed up routinely patients, measurements of success remain nebulous, defined
and those who were not, no differences were found in the by arbitrary criteria with imprecise metrics.
long-term weight loss or presence of nutritional deficien-
cies. Shen et al. [27] have shown that follow-up plays a
significant role in weight loss for laparoscopic adjustable Acknowledgment
gastric banding but not for Roux-en-Y gastric bypass. The authors thank Pleshette Fitzgerald, Rebecca Galvan,
Harper et al. [28] found that patients who automatically Mooney Keonoupheth, and Kimberly Zurita for helping to
returned for their annual follow-up visit did significantly find patients and charts, Ronna Mallios M.S., for statistical
better than those who did not (76% versus 65% excess body analysis, and Abdelrahman Nimeri, M.D., F.A.C.S., and
weight loss; P ⬍ .003). Both studies were based on 1-year Alice Jackson, N.P.-C., C.B.N., for their contributions to
data, and extrapolation to long-term performance is ques- this report.
tionable.
Obtaining follow-up data and statistics contributed sig-
nificantly to our office overhead expenses. Our policy has
Disclosures
always been to follow up patients lifelong; however, our
statistics indicated several barriers to obtaining this ideal. K. Higa, consultant for Endogastric Solutions, past con-
Patients often do not perceive a need for routine follow-up, sultant for Ethicon Endo-Surgery and W. L. Gore & Asso-
especially in the absence of symptoms. Their insurance ciates, Inc., T. Ho, none; F. Tercero, none; T. Yunus, none;
status and residence often change, as does their primary K. B. Boone, consultant for Ethicon Endo-Surgery and
physician. Often, the only way to track a patient is through W. L. Gore & Associates, Inc.
 K. Higa et al. / Surgery for Obesity and Related Diseases 7 (2011) 516 –525 525

References [16] Pories W, Swanson M, MacDonald K. Who would have thought it?
An operation proves to be the most effective therapy for adult-onset
[1] Wittgrove AC, Clark GW. Laparoscopic gastric bypass, Roux-en-Y— diabetes mellitus. Ann Surg 1995;222:339 –50.
500 patients: technique and results, with 3– 60-month follow-up. [17] Sugerman HJ, Wolfe LG, Sica DA, Clore JN. Diabetes and hyper-
Obes Surg 2000;10:233–9. tension in severe obesity and effects of gastric bypass-induced weight
[2] Higa K, Boone K, Ho T. Laparoscopic Roux-en-Y gastric bypass for loss. Ann Surg 2003;237:751– 8.
morbid obesity: technique and preliminary results of our first 400 [18] Sjostrom L, Lindroos A, Peltonen M, et al. Lifestyle, diabetes, and
patients. Arch Surg 2000;135:1029 –34. cardiovascular risk factors 10 years after bariatric surgery. N Engl
[3] MacLean LD, Rhode BM, Sampalis JS, Forse RA. Results of the J Med 2004;351:2683–93.
surgical treatment of obesity. Am J Surg 1993;165:155– 62. [19] Christou NV, Look D, Maclean LD. Weight gain after short- and
[4] Wittgrove AC, Clark GW, Schubert KR. Laparoscopic gastric bypass, long-limb gastric bypass in patients followed for longer than 10 years.
Roux-en-Y: technique and results in 75 patients with 3–30-months Ann Surg 2006;244:734 – 40.
follow-up. Obes Surg 1996;6:500 – 4. [20] Lewis CE, Jacobs DR Jr, McCreath H, et al. Weight gain contin-
[5] Schauer PR, Ikramuddin S, Hamad G, et al. Laparoscopic gastric ues in the 1990s: 10-year trends in weight and overweight from
bypass surgery: current technique. J Laparoendosc Adv Surg Tech A the CARDIA study: Coronary Artery Risk Development in Young
2003;13:229 –39. Adults. Am J Epidemiol 2000;151:1172– 81.
[6] de la Torre RA, Scott JS. Laparoscopic Roux-en-Y gastric bypass: a [21] Brolin RE. Long limb Roux-en-Y gastric bypass revisited. Surg Clin
totally intra-abdominal approach—technique and preliminary report. North Am 2005;85:807–17.
Obes Surg 1999;9:492– 8. [22] Brolin RE, Kenler HA, Gorman JG, Cody RP. Long-limb gastric
[7] MacLean LD, Rhode BM, Nohr CW. Late outcome of isolated gastric bypass in the super-obese: a prospective randomized study. Ann Surg
bypass. Ann Surg 2000;231:524 – 8. 1992;215:387–95.
[8] Halverson JD, Koehler RE. Gastric bypass: analysis of weight loss [23] Buchwald H, Avidor Y, Braunwald E, et al. Bariatric surgery: a
and factors determining success. Surgery 1981;90:446 –55.
systematic review and meta-analysis. JAMA 2004;292:1724 –37.
[9] Biron S, Hould FS, Lebel S, et al. Twenty years of biliopancreatic
[24] Gasteyger C, Suter M, Gaillard RC, Giusti V. Nutritional deficiencies
diversion: what is the goal of the surgery? Obes Surg 2004;14:160 – 4.
after Roux-en-Y gastric bypass for morbid obesity often cannot be
[10] Oria HE, Moorehead MK. Bariatric Analysis and Reporting Outcome
prevented by standard multivitamin supplementation. Am J Clin Nutr
System (BAROS). Obes Surg 1998;8:487–99.
2008;87:1128 –33.
[11] Nguyen NT, Ho HS, Palmer LS, Wolfe BM. Laparoscopic Roux-
en-Y gastric bypass for super/super obesity. Obes Surg 1999;9: [25] Marinella MA. Anemia following Roux-en-Y surgery for morbid
403– 6. obesity: a review. South Med J 2008;101:1024 –31.
[12] Scopinaro N, Adami GF, Marinari GM, et al. Biliopancreatic diver- [26] Bloomberg RD, Fleishman A, Nalle J, Herron DM, Kini S. Nutri-
sion. World J Surg 1998;22:936 – 46. tional deficiencies following bariatric surgery: what have we learned?
[13] White S, Brooks E, Jurikova L, Stubbs RS. Long-term outcomes after Obes Surg 2005;15:145–54.
gastric bypass. Obes Surg 2005;15:155– 63. [27] Shen R, Dugay G, Rajaram K, Cabrera I, Siegel N, Ren CJ. Impact of
[14] O’Brien PE, McPhail T, Chaston TB, Dixon JB. Systematic review of patient follow-up on weight loss after bariatric surgery. Obes Surg
medium-term weight loss after bariatric operations. Obes Surg 2006; 2004;14:514 –9.
16:1032– 40. [28] Harper J, Madan AK, Ternovits CA, Tichansky DS. What happens to
[15] Jones K. Experience with the Roux-en-Y gastric bypass, and com- patients who do not follow-up after bariatric surgery? Am Surg
mentary on current trends. Obes Surg 2000;10:183–5. 2007;73:181– 4.