Five-Year Outcomes After Laparoscopic Gastric Bypass and Laparoscopic Duodenal Switch in Patients With Body Mass Index of 50 to 60: A Randomized Clinical Trial

Importance  There is no consensus as to which bariatric procedure is preferred to reduce weight and improve health in patients with a body mass index higher than 50.

Objective  To compare 5-year outcomes after Roux-en-Y gastric bypass (gastric bypass) and biliopancreatic diversion with duodenal switch (duodenal switch).

Design, Setting, and Participants  Randomized clinical open-label trial at Oslo University Hospital, Oslo, Norway, and Sahlgrenska University Hospital, Gothenburg, Sweden. Participants were recruited between March 17, 2006, and August 20, 2007, and included 60 patients aged 20 to 50 years with a body mass index of 50 to 60. The current study provides the 5-year follow-up analyses by intent to treat, excluding one participant accepted for inclusion who declined being operated on prior to knowing to what group he was randomized.

Interventions  Laparoscopic gastric bypass and laparoscopic duodenal switch.

Main Outcomes and Measures  Body mass index and secondary outcomes including anthropometric measures, cardiometabolic risk factors, pulmonary function, vitamin status, gastrointestinal symptoms, health-related quality of life, and adverse events.

Results  Sixty patients were randomly assigned and operated on with gastric bypass (n = 31) and duodenal switch (n = 29). Fifty-five patients (92%) completed the study. Five years after surgery, the mean reductions in body mass index were 13.6 (95% CI, 11.0-16.1) and 22.1 (95% CI, 19.5-24.7) after gastric bypass and duodenal switch, respectively. The mean between-group difference was 8.5 (95% CI, 4.9-12.2; P < .001). Remission rates of type 2 diabetes mellitus and metabolic syndrome and changes in blood pressure and lung function were similar between groups. Reductions in total cholesterol, low-density lipoprotein cholesterol, triglycerides, and fasting glucose were significantly greater after duodenal switch compared with gastric bypass. Serum concentrations of vitamin A and 25-hydroxyvitamin D were significantly reduced after duodenal switch compared with gastric bypass. Duodenal switch was associated with more gastrointestinal adverse effects. Health-related quality of life was similar between groups. Patients with duodenal switch underwent more surgical procedures related to the initial procedure (13 [44.8%] vs 3 [9.7%] patients; P = .002) and had significantly more hospital admissions compared with patients with gastric bypass.

Conclusions and Relevance  In patients with a body mass index of 50 to 60, duodenal switch resulted in greater weight loss and greater improvements in low-density lipoprotein cholesterol, triglyceride, and glucose levels 5 years after surgery compared with gastric bypass while improvements in health-related quality of life were similar. However, duodenal switch was associated with more surgical, nutritional, and gastrointestinal adverse effects.

Trial Registration  clinicaltrials.gov Identifier: NCT00327912

Roux-en-Y gastric bypass (gastric bypass) and biliopancreatic diversion with duodenal switch (duodenal switch) are surgical procedures used in the treatment of severe obesity. In patients with a body mass index (BMI; calculated as weight in kilograms divided by height in meters squared) higher than 50, gastric bypass may be considered suboptimal owing to insufficient long-term weight loss and substantial weight regain.^1-7 However, there is no consensus on the preferred procedure. Although studies indicate a greater weight loss and greater improvements in glycemic control after duodenal switch, the procedure is less often performed, presumably owing to the surgical technical aspects and perceived risks of nutritional sequelae.^4,8-13

Long-term results from randomized clinical trials comparing gastric bypass with duodenal switch are lacking. Quiz Ref IDIn previous reports of this randomized clinical trial, we found that 2 years after surgery, duodenal switch led to greater weight loss and greater improvement in some blood lipids compared with gastric bypass. However, duodenal switch was associated with more nutritional complications and gastrointestinal adverse effects. Improvement in health-related quality of life (HRQL) was similar between the groups.^14-17 The current study provides the 5-year follow-up analyses of our trial. It extends our previous studies and addresses the durability of the outcomes from the procedures.

This randomized clinical trial was conducted at Oslo University Hospital, Oslo, Norway, and Sahlgrenska University Hospital, Gothenburg, Sweden. Eligibility criteria for study inclusion were a BMI of 50 to 60, age between 20 and 50 years, and previous nonsurgical weight reduction attempts that did not enable sustained weight loss. Exclusion criteria were previous bariatric or major abdominal surgery, malignancy, severe cardiopulmonary disease, systemic steroid treatment, severe psychiatric illness, and drug abuse.¹⁴ The 5-year follow-up was completed January 14, 2013. The regional ethics committees in Health Region South-East, Norway, and Gothenburg, Sweden, approved the study protocol. The full study protocol and list of participating sites and investigators can be found in the trial protocol in Supplement 1. All patients provided written informed consent.

Randomization was computer generated, with allocation concealment by sequentially numbered sealed envelopes and stratification for sex, age, BMI, and study center.¹⁴ The minimization method was used. Both procedures were performed laparoscopically. In gastric bypass, the gastric pouch was 25 mL or less, the alimentary limb was 150 cm, and the biliopancreatic limb was 50 cm. The 1-stage duodenal switch included a sleeve gastrectomy, a 200-cm alimentary limb, and a 100-cm common channel. Mesenteric defects were not closed in either procedure. A standard regimen of vitamin and mineral supplementation was prescribed (eTable 1 in Supplement 2).¹⁵

The primary end point of the trial was change in BMI at 2 years’ follow-up. All outcomes reported were predefined secondary outcomes including weight loss, cardiometabolic risk factors, lung function, vitamins, gastrointestinal symptoms, HRQL, and adverse events.

Data were collected from outpatient visits at baseline, 6 to 8 weeks; 6 months; and 1, 2, and 5 years after surgery. Each visit included a physical examination and fasting blood samples and a record of medical history, current medication, and adverse events. The patients completed study questionnaires at baseline and 1, 2, and 5 years after surgery.

Remission of type 2 diabetes mellitus was evaluated using the criteria from the American Diabetes Association, with complete remission defined as having a fasting glucose level less than 100 mg/dL (to convert to millimoles per liter, multiply by 0.0555) and partial remission defined as having a fasting glucose level of 100 to 125 mg/dL, both assuming at least 1 year without pharmacological antihyperglycemic treatment prior to evaluation.¹⁸ Metabolic syndrome was diagnosed using the criteria from the International Diabetes Federation.¹⁹ Glycated hemoglobin A_1c was measured with high-performance liquid chromatography (Tosoh G8; Tosoh Corporation; normal reference range, 4%-6%).

Spirometry was performed at baseline and after 5 years according to a standardized procedure with the patients in a sitting position.²⁰ The best of 3 acceptable measures of forced expiratory volume in 1 second and forced vital capacity were used.

A standard regimen of vitamin and mineral supplementation was prescribed (eTable 1 in Supplement 2). Vitamin and mineral status was measured at each study visit and supplements were individually adjusted after follow-up visits. Methods for vitamin analyses have been previously described in detail.¹⁵ Vitamin deficiencies were defined as concentrations less than the laboratory reference values.

To evaluate gastrointestinal adverse effects of the procedures, the patients completed the Gastrointestinal Symptom Rating Scale²¹ and a bowel-function questionnaire.²²

Health-related quality of life was evaluated with a generic questionnaire, the Short Form 36 (4-week recall; version 2.0),²³ and an obesity-specific questionnaire, the Obesity–related Problems Scale.²⁴ The Short Form 36 measures the health burden of chronic diseases and comprises 8 dimensions of physical and mental health, with scores ranging from 0 (poorest) to 100 (optimal). The Obesity-related Problems Scale evaluates the effect of obesity on psychosocial functioning, with scores ranging from 0 (optimal) to 100 (poorest).

At each study visit, patients were asked open-ended questions about any adverse events or complaints, including diagnostic or therapeutic procedures, and hospital admissions and patient medical records were reviewed. Perioperative adverse events have been reported previously.¹⁴ Before the outset of the study, surgical experience at the study centers included several hundred gastric bypass procedures and fewer than 20 duodenal switch procedures (15 in Norway and 18 in Sweden). In this study, we present adverse events from 30 days after surgery to 5 years’ follow-up. Complete data were available from 58 patients (96.7%; 55 from follow-up in the clinic, 2 from telephone interviews, and 1 patient who died who had his patient records assessed). Two patients who did not attend the 5-year follow-up had follow-up data available until 2 years after surgery.

Sample size calculation was estimated to detect a difference in BMI between groups as described previously.¹⁴ We fitted linear mixed models to all continuous outcomes with more than 2 measurements (anthropometric measures, cardiometabolic risk factors, vitamins, gastrointestinal symptoms, and HRQL). The effect of time was modeled as piecewise linear, with 1 knot at 1 year and 1 knot at 2 years after baseline. The fixed effects were treatment, time, and treatment × time interaction. The models included a random intercept and a random effect for time and an unstructured covariance matrix was used. Two variables (25-hydroxyvitamin D and the vitality domain of the Short Form 36) could not be modeled with the random effect for time and, for these variables, only a random intercept was used. We used χ² tests or the Fisher exact test to compare independent proportions. Ordered categorical outcomes in the bowel-function questionnaire were compared with the exact Mann-Whitney test. The 2-sampled t test for unequal variances (Welch U test) was used to compare the mean number of adverse events per patient. Weight regain and pulmonary function were compared within treatment groups with the paired-samples t test and between treatment groups with linear regression of 5-year measurement on treatment and nadir measurement (weight regain) or baseline measurement (lung function; analysis of covariance). Thiamine and glycated hemoglobin A_1c levels at 5 years were compared with the independent-samples t test. We used Stata 13.1 (StataCorp LP) and StatXact 9 (Cytel Inc) to perform the statistical analyses.

Of 60 included patients, 31 underwent gastric bypass and 29 underwent duodenal switch (30 patients in Norway and 30 patients in Sweden). Fifty-five patients (92%) attended the 5-year visit after a mean follow-up of 61 months (range, 54-73 months; Figure 1). Table 1 presents baseline patient characteristics.

After 5 years, the estimated reductions in BMI and weight were greater after duodenal switch compared with after gastric bypass, with a mean between-group difference of 8.5 kg (95% CI, 4.9 to 12.2) and 24.9 kg (95% CI, 13.6 to 36.2), respectively (P < .001 for both; Figure 2; Table 2). Quiz Ref IDTotal body weight loss was 26.4% (95% CI, 21.7 to 31.1) after gastric bypass and 40.3% (95% CI, 35.7 to 44.9) after duodenal switch at 5 years’ follow-up (P < .001). Mean weight regain from nadir at 1 or 2 years after the procedure to the end of the study was 9.9 kg (95% CI, 4.0 to 15.8) after gastric bypass and 8.7 kg (95% CI, 4.8 to 12.5) after duodenal switch; the mean between-group difference was 1.7 kg (95% CI, −6.6 to 9.9; P = .69). After 5 years, 15 patients with gastric bypass (55.6%) and 4 patients with duodenal switch (14.3%) had a BMI of more than 40 (P = .001).

Total cholesterol and low-density lipoprotein cholesterol levels decreased significantly only after duodenal switch (Table 2; eFigure 1 in Supplement 2). Triglyceride concentrations decreased significantly after both interventions, with greater reductions after duodenal switch (P = .01). High-density lipoprotein cholesterol concentrations increased significantly after both interventions, with larger increases after gastric bypass (P = .002).

Systolic blood pressure was significantly reduced after both procedures, with no difference between groups (P = .12; Table 2). Diastolic blood pressure was significantly reduced only after duodenal switch but with no difference between groups (P = .10).

Mean fasting plasma glucose concentrations decreased significantly after duodenal switch but not after gastric bypass, with greater reductions after duodenal switch (P = .03 between groups; Table 2; eFigure 1 in Supplement 2). Mean hemoglobin A_1c levels at 5 years’ follow-up were 5.6% (95% CI, 5.2 to 6.0) after gastric bypass and 4.8% (95% CI, 4.6 to 5.0) after duodenal switch (P = .001 between groups). None of the patients with type 2 diabetes mellitus at baseline (5 in the gastric bypass group and 6 in the duodenal switch group) used glucose-lowering medication after 5 years. In the gastric bypass group, 4 of 5 patients with type 2 diabetes mellitus preoperatively had complete remission of type 2 diabetes mellitus and 1 had partial remission. Complete remission was achieved in 6 of 6 patients after duodenal switch.

At baseline, the metabolic syndrome was present in 20 patients with gastric bypass (64.5%) and 23 patients with duodenal switch (79.3%; P = .20 between groups). At 5 years’ follow-up, the corresponding numbers were 3 (11.1%) and 1 (3.6%; P = .28 between groups).

Pulmonary function expressed by the forced expiratory volume in 1 second and forced vital capacity improved significantly from baseline to 5 years’ follow-up with no differences between the groups (Table 2).

The use of dietary supplements is presented in eTable 1 in Supplement 2. Serum concentrations of vitamin A, 25-hydroxyvitamin D, and ionized calcium decreased significantly and parathyroid hormone increased significantly after duodenal switch compared with gastric bypass (eTable 2 in Supplement 2). The prevalence of deficiencies at 5 years’ follow-up was greater after duodenal switch compared with gastric bypass for all these biomarkers. For vitamin E adjusted for lipids, vitamin B₆, vitamin B₁₂, vitamin C, and folic acid, serum concentrations were either unchanged or increased compared with the baseline in both groups, with no between-group differences (eTable 2 in Supplement 2). The prevalence of anemia at 5 years was similar in the groups (eTable 2 in Supplement 2).

Patients in both groups reported more abdominal pain and indigestion during follow-up compared with the baseline, with no difference between groups (eFigure 2 in Supplement 2). The patients with duodenal switch reported significantly more diarrhea compared with baseline but the difference did not reach statistical significance (P = .07). The patients with duodenal switch reported increased gastroesophageal reflux symptoms compared with patients with gastric bypass (P = .002).

The mean number of daytime defecations at baseline and 5 years’ follow-up were 2.1 (95% CI, 1.6 to 2.6) and 3.0 (95% CI, 2.3 to 3.7) after duodenal switch, respectively, and 2.0 (95% CI, 1.5 to 2.4) and 1.8 (95% CI, 1.1 to 2.4) after gastric bypass (P = .03 between groups), respectively. Quiz Ref IDAt 5 years’ follow-up, 13 patients with duodenal switch (48.1%) reported loose stools vs 3 patients with gastric bypass (11.1%; P = .01). Social limitations owing to altered bowel function were reported by 17 patients (63.0%) after duodenal switch and 7 patients (25.9%) after gastric bypass (P = .02). Furthermore, 18 patients with duodenal switch (66.7%) reported negative implications in daily life owing to bowel symptoms compared with 10 patients with gastric bypass (37.0%; P = .03). There were no differences between the groups in regard to nighttime defecations, awareness of or leakage of stool, involuntary leakage of gas, ability to discriminate between stool and gas, or the reported use of antidiarrhea medications (data not shown).

Five of 8 dimensions of the Short Form 36 improved significantly after both procedures, with no difference in changes between groups (eFigure 3 in Supplement 2). Psychosocial function measured with the Obesity–related Problems Scale improved by 24.3% (95% CI, 15.5 to 33.2) after gastric bypass and 35.2% (95% CI, 26.4 to 44.0) after duodenal switch; the mean between-group difference was 10.9% (95% CI, −1.7 to 23.4; P = .09).

Overall, 1 or more adverse events occurred in 21 patients with gastric bypass (67.7%) and 23 patients with duodenal switch (79.3%; P = .31; Table 3). Nine patients with gastric bypass (29.0%) and 17 patients with duodenal switch (58.6%) were admitted to the hospital for any reason during the first 5 years after surgery (P = .02). Additional surgical interventions related to the initial procedure were performed in 3 patients with gastric bypass (9.7%) and 13 patients with duodenal switch (44.8%; P = .002; Table 3).

Four patients with duodenal switch (13.8%) required hospital admissions owing to protein-calorie malnutrition. One of these patients also had severe diarrhea and was subsequently operated on with elongation of the common channel 3 years after the initial procedure. Two additional patients with duodenal switch received elongation of the common channel owing to disabling diarrhea; 1 patient received this after 3 years and 1 patient received this after 5 years. Two of the patients with revisional surgery had improvements but 1 reported an unchanged condition with disabling diarrhea.

Five patients with duodenal switch (17.2%) and 1 patient with gastric bypass (3.2%) presented with anemia or iron deficiency that required iron injections or blood transfusions (P = .10).

Patients with BMI of 50 to 60 who were operated on with either gastric bypass or duodenal switch both demonstrated substantial and sustained weight loss 5 years after surgery in this randomized clinical trial. Quiz Ref IDPatients with duodenal switch had greater weight loss compared with patients with gastric bypass, extending to our findings at 2 years’ follow-up and confirming the differences in weight loss in previous nonrandomized studies.^3,4,25,26 In contrast to our findings at the 2-year follow-up assessment, we observed greater reductions at 5 years’ follow-up in fasting serum triglyceride and glucose levels after duodenal switch compared with gastric bypass.

Surgical treatment of patients with BMI higher than 50 is challenging because while some bariatric procedures lead to unsatisfactory long-term weight loss, other procedures lead to both greater weight loss and more unwanted adverse effects. Previous systematic reviews^27,28 and nonrandomized studies^4,29,30 have found greater improvements of type 2 diabetes mellitus, dyslipidemia, and hypertension after duodenal switch compared with gastric bypass. A systematic review and meta-analysis of gastric bypass and duodenal switch reported remission of type 2 diabetes mellitus in 88% of patients after duodenal switch and 76% of patients after gastric bypass.³¹ In the present study, the remission rates for type 2 diabetes mellitus and the metabolic syndrome were similar in both groups but the number of patients with type 2 diabetes mellitus was limited. Nevertheless, we observed significantly lower mean values for fasting plasma glucose and hemoglobin A_1c levels after duodenal switch. This could be clinically relevant because fasting plasma glucose level is associated with future incidence of type 2 diabetes mellitus, even with glucose within the normal range.^32-34

Changes in HRQL were similar between the surgical groups, primarily showing improvements in domains for physical health and psychosocial functioning. Changes in the Short Form 36 scores were in accordance with findings in uncontrolled long-term studies.^35,36 A previous study found a close association between changes in HRQL and weight loss and weight regain 10 years after surgery.³⁷ In contrast, a systematic review found no association between weight loss and overall HRQL but a modest association between weight loss and physical health.³⁸ The changes in HRQL in our study were comparable between groups despite considerably greater weight loss after duodenal switch.

There were more adverse events after duodenal switch compared with after gastric bypass, consistent with observations reported by others.³⁹ Surgery to correct malnutrition or severe diarrhea was required in 10% of patients after duodenal switch. Other long-term studies have reported revision rates ranging from 0.6% to 5.7%.^40-43 The mesenterial defects were not closed in our study, possibly influencing the incidence of abdominal pain and internal hernias.

The reduction in levels of vitamin A and 25-hydroxyvitamin D after duodenal switch was consistent with our findings 1 and 2 years after surgery.^15,16 A high prevalence of deficiencies in fat-soluble vitamins after duodenal switch has also been observed in noncontrolled studies.^44-46 In contrast, Marceau et al⁴² reported increased levels of long-term 25-hydroxyvitamin D after duodenal switch in patients prescribed high doses of fat-soluble vitamins. Additional supplementation should be prescribed after duodenal switch using substantially higher doses of fat-soluble vitamins. Abnormalities in calcium metabolism after duodenal switch underscore the need of further studies on the consequences on bone health. Thus, monitoring nutritional parameters is particularly essential after duodenal switch.

Quiz Ref IDStrengths of this trial included the randomized design, standardized surgery, and a high follow-up rate. We included patients with BMI of 50 to 60 because the choice of procedure in this patient group was considered particularly challenging. Limitations included a small sample size that provided limited statistical power for assessing some of the secondary end points. Both centers had considerable experience in performing gastric bypass but less experience with duodenal switch. Therefore, we focused on the long-term adverse events (>30 days after surgery) in an attempt to reduce the importance of the potential technical aspect of less surgical experience with duodenal switch. The follow-up regimen was similar after both procedures, making it unlikely that any differences in outcomes were explained by differences in postoperative care. However, as demonstrated, patients with duodenal switch were more prone to nutritional deficiencies and could thus benefit from closer follow-up.

The results may not generalize to patients with BMI outside the range of 50 to 60. A systematic review of gastric bypass and duodenal switch found that differences in BMI reductions increased with higher preoperative BMI, indicating that patients with higher BMI ranges might benefit more from duodenal switch.³¹

Patients with BMI higher than 50 account for more than 25% of the bariatric population in the United States.^47,48 Given the complex health issues associated with very high BMIs, this group represents a particular challenge to the health care system. In our study, 56% of patients with gastric bypass had a BMI higher than 40 at 5 years after surgery, which may represent less efficiency in treating weight-bearing health problems with gastric bypass compared with duodenal switch.

The benefits of duodenal switch must be balanced with the risks of adverse events and impact on daily life (gastrointestinal adverse effects) as well as with the higher use of health care resources compared with after gastric bypass. Patient commitment to recommended follow-up regimens might improve outcome after duodenal switch. Unfortunately, there are no validated presurgical methods to identify patients’ adherence to follow-up regimens.

We recommend that duodenal switch be used with caution owing to a higher rate of additional surgical procedures and risk of nutritional complications. In patients in whom duodenal switch is considered, a 2-stage approach could be an alternative strategy. Following a sleeve gastrectomy, a second surgical procedure could complete the duodenal switch in patients adhering to follow-up regimens if the patients experience unsatisfactory weight loss and continue to have comorbidities that might benefit from further weight loss.^49-51 However, it remains to be determined whether a 2-staged duodenal switch will improve patient outcomes.

In our study, duodenal switch resulted in greater weight loss and greater improvements in blood lipids and glucose compared with gastric bypass across 5 years in patients with baseline BMIs of 50 to 60. Improvements in HRQL were comparable. However, duodenal switch was associated with more long-term surgical and nutritional complications and more gastrointestinal adverse effects compared with gastric bypass.

Corresponding Author: Torsten Olbers, MD, PhD, Department of Gastrosurgical Research and Education, Institute of Clinical Sciences, Sahlgrenska Academy at the University of Gothenburg, Sahlgrenska University Hospital, PO Box 100, SE-405 30 Gothenburg, Sweden (torsten.olbers@gu.se).

Accepted for Publication: August 19, 2014.

Published Online: February 4, 2015. doi:10.1001/jamasurg.2014.3579.

Author Contributions: Dr Risstad had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: All authors.

Acquisition, analysis, or interpretation of data: Risstad, Søvik, Aasheim, Fagerland, Fagevik Olsén, Kristinsson, le Roux, Bøhmer, Birkeland, Mala, Olbers.

Drafting of the manuscript: Risstad.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Risstad, Fagerland.

Obtained funding: Risstad, Aasheim, Kristinsson, Bøhmer, Mala, Olbers.

Administrative, technical, or material support: Risstad, Søvik, Engström, Aasheim, Fagevik Olsén, Bøhmer, Birkeland, Mala.

Study supervision: Kristinsson, le Roux, Mala, Olbers.

Conflict of Interest Disclosures: Dr Aasheim has received grants from the University of Oslo and South Eastern Norway Regional Health Authority for biochemical analysis of vitamin concentrations. Drs Kristinsson and Mala receive travel reimbursements from Johnson and Johnson and Covidien. Dr le Roux is funded by grant 12/YI/B2480 from the Science Foundation Ireland. Dr Birkeland has served as a consultant for and has received lecture fees or travel reimbursements from Novo Nordisk, Eli Lilly, Sanofi, Merck, Boehringer Ingelheim, Bristol-Myers Squibb, Novartis, and Pfizer. Dr Olbers serves on the advisory board and receives lecture fees and travel reimbursements from Johnson and Johnson. No other disclosures were reported.

Funding/Support: The present study was supported by a research fellowship from the South Eastern Norway Regional Health Authority (Drs Risstad, Søvik, and Aasheim), Oslo University Hospital, and a postdoctoral research grant from Sahlgrenska University Hospital (Dr Olbers). The use of the Short Form 36 questionnaires was permitted by the payment of license fees (license number QM017638).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We thank Hans Lönroth, MD, PhD, and Almantas Maleckas, MD, PhD, Sahlgrenska University Hospital, and Carl Fredrik Schou, MD, Oslo University Hospital, for providing surgical expertise; Kari Julien, MLT, Hormone Laboratory, Oslo University Hospital, for help with hormone analysis; Anne Hove, MLT, and Merete Berge Pettersen, MLT, Nutritional Laboratory, Biochemical Department, Oslo University Hospital, for help with vitamin analyses; and Niclas Björnfot, RN, Sahlgrenska University Hospital, for data collection. None of these individuals received financial compensation for their contributions.