A total of 1219 of 1394 patients (87.4%) felt better or much better than before RYGB surgery, while 113 of 1394 (8.1%) felt worse than before surgery. Thirty-five patients did not answer this question.
eFigure 1. Flowchart of Patients in the Questionnaire Survey
eFigure 2. Questionnaire
eTable 1. Prevalence Ratios of Possible Predictors of Symptoms Leading to Hospital Admission After RYGB
eTable 2. Comparison of Self-Reported Data on Kidney Stones and Gallstones With the Gold Standard in the Danish National Patient Registry
eTable 3. Baseline Characteristics of the 2238 Patients Who Underwent RYGB Surgery, According to Response Status
eTable 4. Prevalence Ratios of Possible Time Since Surgery and Symptoms Leading to Contact With the Health Care System After RYGB
Gribsholt SB, Pedersen AM, Svensson E, Thomsen RW, Richelsen B. Prevalence of Self-reported Symptoms After Gastric Bypass Surgery for Obesity. JAMA Surg. 2016;151(6):504-511. doi:10.1001/jamasurg.2015.5110
Population-based studies on the prevalence of symptoms after Roux-en-Y gastric bypass (RYGB) surgery are sparse. Knowledge about possible predictors of these symptoms is important for prevention.
To examine patients’ overall well-being and the prevalence and predictors of medical, nutritional, and surgical symptoms after RYGB surgery, and their association with quality of life.
Design, Setting, and Participants
A survey was conducted from March 3 to July 31, 2014, among 2238 patients who underwent RYGB surgery between January 1, 2006, and December 31, 2011, in the Central Denmark Region. A comparison cohort of 89 individuals who were matched with patients according to sex and body mass index but who did not undergo RYGB surgery were surveyed as a point of reference. Data analysis was conducted from September 1, 2014, to June 25, 2015.
Main Outcomes and Measures
Prevalence and severity (based on contacts with health care system, ranging from no contact to hospitalization) of self-reported symptoms following RYGB surgery. Prevalence ratios (PRs) of symptoms associated with different predictors were computed. The association between number of symptoms and quality of life was investigated using the Spearman rank correlation coefficient.
Of 2238 patients undergoing RYGB surgery, 1429 (63.7%) responded to the survey. Among these patients, 1266 (88.6%) reported 1 or more symptoms a median of 4.7 years after RYGB surgery. Mean age at the time of the survey was 47.1 years (range, 26.9-68.0 years), and 286 were men (20.0%). A total of 1219 of 1394 patients (87.4%) reported that their well-being was improved after vs before RYGB surgery, while 113 (8.1%) reported reduced well-being. Symptoms after RYGB surgery were reported by 1266 patients (88.6%); 966 patients (67.6%) had been in contact with the health care system about their symptoms vs 31 [34.8%] of those in the comparison group, and 416 (29.1%) had been hospitalized vs 6 [6.7%] of those in the comparison group. The symptoms most commonly leading to health care contact after RYGB surgery were abdominal pain (489 [34.2%]), fatigue (488 [34.1%]), and anemia (396 [27.7%]). The risk of symptoms was higher among women (crude PR, 1.23; 95% CI, 1.11-1.37), among patients younger than 35 years (PR, 1.24; 95% CI, 1.13-1.36), among smokers (PR, 1.11; 95% CI, 1.02-1.20), among unemployed persons (PR, 1.15; 95% CI, 1.06-1.24), and in those with surgical symptoms before RYGB surgery (PR, 1.34; 95% CI, 1.25-1.43). Quality of life was inversely associated with the number of symptoms (r = –0.30; P < .001).
Conclusions and Relevance
Most patients reported improved well-being after RYGB surgery, but the prevalence of symptoms was high and nearly one-third of patients were hospitalized, 4- to 5-fold more than among the comparison group. Predictors of symptoms included young age, female sex, smoking, and experiencing symptoms before RYGB surgery. Development of weight loss procedures with fewer subsequent symptoms should be a high priority.
For patients with morbid obesity, bariatric surgery, including Roux-en-Y gastric bypass (RYGB) surgery, is an effective treatment for weight loss and diseases associated with obesity.1,2 However, various medical, nutritional, and surgical symptoms requiring treatment may occur after RYGB surgery3- 5 and may impair patients’ quality of life (QoL). These symptoms have generally been classified as surgical or medical and nutritional depending on the primary treatment procedures. Quiz Ref IDCommon surgical symptoms after RYGB surgery include gallstones (7%-16% after 18-26 months’ follow-up)6,7 and kidney stones (≤8% after 19-42 months’ follow-up).8 Well-known medical and nutritional symptoms include dumping syndrome (12%-42% after 1-2 years),9- 11 hypoglycemia (0.2% of patients were hospitalized within 0-20 years),12 diarrhea (18% of patients with anal leakage of stool potentially provoked by diarrhea after 2 years),13 peripheral neuropathy (16% of patients, with unknown follow-up period),14 and anemia (23%-47% after 24-48 months).15- 17 Commonly observed symptoms in clinical practice, such as abdominal pain and fatigue, are sparsely described in the literature on RYGB surgery despite the importance of such symptoms for patients’ functional status and QoL.18,19 Thus, a large population-based study with long-term follow-up of patients who had undergone RYGB surgery, providing a more comprehensive examination of symptom prevalence, is warranted.
Knowledge of possible predictors is important to prevent symptoms related to RYGB surgery after the procedure. Most patients undergoing RYGB surgery are women; recent studies have pointed toward a higher risk of hospitalization after RYGB surgery for women than for men,20,21 whereas others have found no influence of sex on the rate of hospitalization after the procedure.22 Furthermore, advanced age has been associated with increased risk of hospitalization after surgery.20,21,23 Increased mortality and morbidity after other types of surgery has been observed for smokers,24 unmarried patients,25 and unemployed persons.26 Whether such demographic and lifestyle factors also predict symptoms after bariatric surgery, to our knowledge, have not yet been examined.
Previous studies have reported improved physical QoL but not mental QoL after bariatric surgery.27,28 The influence of medical and nutritional symptoms on QoL after RYGB surgery has, to our knowledge, not been examined previously. As surgical symptoms have been associated with reduced QoL,29 we hypothesized that patients with medical and nutritional symptoms following RYGB surgery have a poorer QoL than do patients with no symptoms.
We aimed to conduct a comprehensive study of the well-being and the severity and burden of common symptoms after RYGB surgery. Furthermore, we examined possible predictors of symptoms requiring health care contact after RYGB surgery, including demographic, clinical, socioeconomic, and lifestyle factors. Finally, we examined the association of medical, nutritional, and surgical symptoms with QoL after RYGB surgery.
The Central Denmark Region has a population of approximately 1.27 million people. All residents are registered in the Civil Registration System (CRS) and assigned a unique personal identification number (CRS number).30,31 The Danish National Board of Health provides tax-supported health care, including RYGB surgery. In Denmark, all patients must be seen by the general practitioner to be referred to secondary care (outpatient and inpatient clinics or hospitalization).32
We obtained CRS numbers for all 3121 patients undergoing RYGB surgery (International Classification of Diseases, 10th Revision, codes KJDF10 [gastric bypass] and KJDF11 [laparoscopic gastric bypass]) within Central Denmark between January 1, 2006, and December 31, 2011 (>99% of the operations were laparoscopic RYGB surgery). The CRS provided data on the vital status of all patients. Patients who were deceased (n = 32), had emigrated (n = 14), had unknown vital status (n = 1), had a CRS-registered restriction against researcher inquiries (n = 791), or were unknown at their registered address (n = 45) were excluded (eFigure 1 in the Supplement). The study questionnaire was mailed to 2238 eligible patients on March 3, 2014. Two reminders were sent to those who did not respond. On June 8, 2015, a comparison cohort of 89 individuals who did not undergo RYGB surgery and who were matched with patients according to sex and body mass index were recruited via the Internet; this cohort also responded to the questionnaire.
According to Danish legislation,33 questionnaire studies do not require a separate permission from the Danish Scientific Ethical Committee. Patients provided written consent to participate in the surveys.
The questionnaire included items on sex, age, height, weight, preoperative surgical symptoms, cohabiting status, educational level, employment status, tobacco smoking, surgical symptoms after RYGB surgery (abdominal pain [eg, leakage, herniation], kidney stones, and gallstones), and medical and nutritional symptoms after RYGB surgery (diarrhea, anemia, fatigue, neurologic symptoms, dumping, and hypoglycemia). Questions on hospitalizations owing to plastic surgery after RYGB surgery were not included. For each symptom that occurred after RYGB surgery, we asked patients if the symptom had led to any contact with the health care system, categorizing patients as follows: 1, no contact; 2, any contact (including contact with general practitioner, visit in outpatient specialty clinic, and hospital admission); 3, contact with secondary care only (visit in outpatient specialty clinic and hospital admission); and 4, hospital admission (eFigure 2 in the Supplement). Contacts with the health care system owing to routine follow-up visits after RYGB surgery were not included in this questionnaire. Presence of symptoms prior to RYGB surgery was also queried.
The Edinburgh Hypoglycemia Scoring System (18 questions)34 and the Dumping Symptom Rating Scale (9 questions)9 were included in the questionnaire. The 2 scales are Likert scales, both ranging from 1 to 7. For hypoglycemia and dumping (general and gastrointestinal symptoms after gastrectomy), the severity of the symptoms and health-seeking behavior were estimated from a scoring system, in contrast with the other symptoms. We defined mild hypoglycemia as a score of 36 to 71 (symptomatic to the patient), moderate hypoglycemia as a score of 72 to 89 (likely to lead to contact with primary care), and severe hypoglycemia as a score of 90 or more (likely requiring contact with secondary care). We defined mild dumping as a score of 18 to 35 (symptomatic to the patient), moderate dumping as a score of 36 to 44 (likely to lead to contact with primary care), and severe dumping as a score of 45 or more (likely to lead to contact with secondary care).
We assessed QoL using the Mental Composite score (MCS) (6 items) and the Physical Composite score (PCS) (6 items) of the 12-Item Short-Form Health Survey (SF-12), which has been validated in the general Danish population.35 The scores for all items are summed and transformed into a scale from 0 to 100 (0 indicates poorest possible health state; 100, best possible health state). A score below 50 indicates poor mental or physical health. The SF-12 scores were calculated using Health Outcome Scoring Software, version 4.5 (Optum Inc). We also included a question addressing current well-being compared with well-being before RYGB surgery.
Data analysis was conducted from September 1, 2014, to June 25, 2015. We tabulated characteristics of the cohort that underwent RYGB surgery and the comparison cohort and tabulated the prevalence and severity of symptoms. For the cohort that underwent RYGB surgery, we computed the prevalence and the unadjusted and adjusted prevalence ratios (PRs) of symptoms, comparing possible predictors of symptoms leading to any health care contact, using Poisson regression analysis. The variables predefined as possible predictors were sex (male or female), age (<35 years, 35-44 years, or ≥45 years), surgical symptoms before RYGB surgery (abdominal pain, kidney stones, or gallstones), cohabiting status (living with someone or living alone), educational level (≤9 years, 10-12 years, or >12 years), employment status (full time, part time, or unemployed), and tobacco smoking (yes or no). All PRs were adjusted for demographic (age and sex), clinical (preoperative BMI, percentage weight loss, and surgical symptoms before RYGB surgery), socioeconomic (cohabiting status, employment status, and educational level), and lifestyle (smoking) factors. We conducted stratified analyses of prevalence and PRs focusing on medical and nutritional symptoms alone (diarrhea, fatigue, anemia, neurologic symptoms, hypoglycemia, and dumping), surgical symptoms alone (abdominal pain, kidney stones, and gallstones), and symptoms requiring hospitalization. We presented graphically the overall well-being of patients who underwent RYGB surgery and calculated symptom prevalence according to this factor.
Patients who underwent RYGB surgery were grouped according to the number of symptoms leading to health care contact (0, 1, 2, 3, ≥4 symptoms). The median score and interquartile range (IQR) of the PCS and the MCS were calculated for each group as data were nonnormally distributed. We calculated the Spearman rank correlation coefficient (r) for the correlations between MCS and PCS and SF-12 scores.
Statistical analyses were performed using STATA, version 13.0 (StataCorp LP). The study was approved by the Danish Data Protection Agency.
A total of 1429 patients undergoing RYGB surgery (63.7%) responded to the survey; 286 (20.0%) were male. Median time since RYGB surgery was 4.7 years. Mean age at the time of the survey was 47.1 years (range, 26.9-68.0 years); the current median body mass index (calculated as weight in kilograms divided by height in meters squared) was 29.7, corresponding to a median total weight loss of 35.7%. As seen in Table 1, the comparison cohort was matched on current body mass index (29.4) and sex (22.5% male), yet it was slightly younger (mean age, 37.7 years).
A total of 1266 patients (88.6%) who underwent RYGB surgery reported 1 or more symptoms (Table 2) vs 70 (78.7%) in the comparison cohort. Quiz Ref IDThe symptoms most commonly reported among patients who underwent RYGB surgery were fatigue (776 [54.3%]), dumping (749 [52.4%]), and abdominal pain (778 [54.4%]). Of those reporting fatigue, 512 (66.0%) also reported dumping and 533 (68.7%) reported abdominal pain. Of the patients reporting dumping, 494 (66.0%) also reported abdominal pain. Symptoms leading to any health care contact were reported by 966 patients (67.6%) in those undergoing RYGB surgery and 31 (34.8%) in the comparison cohort. The most frequent symptoms after RYGB surgery that led to any health care contact were fatigue (488 [34.1%]), abdominal pain (489 [34.2%]), and anemia (396 [27.7%]). We estimated that 130 patients (9.1%) were likely to have health care contact owing to dumping and 95 (6.6%) owing to hypoglycemia. Dumping and hypoglycemia alone accounted for only 24 (1.7%) of the total health care contacts.
Quiz Ref IDInpatient hospitalizations owing to any symptom after surgery were reported by 416 patients (29.1%) in those who underwent RYGB surgery vs 6 individuals (6.7%) in the comparison cohort (Table 2). A total of 376 of 476 hospitalizations (79.0%) after RYGB surgery were owing to surgical symptoms and 100 (21.0%) to medical and nutritional symptoms.
Quiz Ref IDWe found a higher prevalence of symptoms leading to health care contact among women (PR, 1.23; 95% CI, 1.11-1.37), patients with surgical symptoms before RYGB surgery (PR, 1.34; 95% CI, 1.25-1.43), unemployed patients (PR, 1.15; 95% CI, 1.06-1.24), smokers (PR, 1.11; 95% CI, 1.02-1.20), and patients younger than 35 years (PR, 1.24; 95% CI, 1.13-1.36) (Table 3). Mutually adjusting for demographic, clinical, socioeconomic, and lifestyle factors yielded almost unchanged PRs.
Stratification for medical and nutritional symptoms and surgical symptoms after RYGB surgery provided results that were consistent with the results of all symptoms combined, with some exceptions. More women had medical and nutritional symptoms (PR, 1.45; 95% CI, 1.24-1.69). For surgical symptoms, the sex difference was smaller (PR for women, 1.15; 95% CI, 0.98-1.34), but the association with surgical symptoms before RYGB surgery was strong (PR, 1.93; 95% CI, 1.74-2.14). Concerning analysis restricted to hospitalizations, results were similar except for higher PRs for patients who underwent RYGB surgery and experienced a large amount of postoperative weight loss (PR for weight loss ≥35%, 1.53; 95% CI, 1.29-1.82) and for those with surgical symptoms before RYGB surgery (PR, 1.75; 95% CI, 1.48-2.07) (eTable 1 in the Supplement).
In total, 1219 of 1394 (87.4%) patients felt better or much better and 113 (8.1%) reported reduced well-being after RYGB surgery (Figure). Those who felt worse had substantially higher prevalence of symptoms after RYGB surgery (102 of 113 patients [90.3%] vs 846 of 1281 [66.0%]; P < .001). There was no difference between the groups in the prevalence of surgical symptoms before RYGB surgery (27 of 113 patients [23.9%] vs 228 of 1281 [17.8%]; P = .11).
Based on the SF-12, the median MCS was 41.8 (IQR, 34.8-45.9) for patients who underwent RYGB surgery and 39.7 (IQR, 31.7-44.7) in the comparison cohort. The median PCS was 53.0 (IQR, 46.1-55.4) and 53.5 (IQR, 49.4-55.2), respectively. For patients who underwent RYGB surgery and experienced no symptoms, the median MCS was 44.1 and the median PCS was 54.2 compared with a median MCS of 33.6 and a median PCS of 45.0 for those with 4 or more symptoms. The Spearman rank correlation coefficients between number of symptoms and MCS and PCS were ρ = –0.30 and ρ = –0.23, respectively (Table 4).
Quiz Ref IDOur study showed that 87.4% of the patients reported that they felt better after RYGB surgery. Nevertheless, 88.6% of patients reported 1 or more symptoms (mild to severe) a median of 4.7 years after RYGB surgery. In the comparison cohort, 78.7% of the participants reported symptoms, indicating that mild symptoms are a part of “normal” physiological features in patients with obesity. Approximately two-thirds of patients who underwent RYGB surgery had health care contact owing to symptoms and 29.1% had been hospitalized, 79.0% owing to surgical symptoms and 21.0% owing to medical symptoms, a higher percentage than previously reported.21 The patients who underwent RYGB surgery were hospitalized 4- to 5-fold more than the comparison cohort.
Predictors of symptoms included younger age, female sex, smoking, and unemployment. Finally, a higher number of symptoms was associated with lower health-related QoL and reduced well-being after RYGB surgery.
Most patients reported a better or much better well-being after surgery despite a high prevalence of symptoms. Those who reported worse well-being were characterized by a high prevalence of symptoms. Obesity is associated with reduced QoL, possibly owing to stigmatization, reduced mobility, and obesity-related diseases.27 Satisfaction with weight loss after RYGB surgery may outweigh some dissatisfaction with symptoms after surgery. Previous studies have focused on symptoms leading to hospitalization.20,22,23 Morgan et al20 found that 18% of patients were hospitalized after bariatric surgery (mean follow-up, 41 months) compared with 29.1% in our study. However, they did not report on anemia, neurologic complications, gallstones, kidney stones, fatigue, and diarrhea, which may explain this difference. Our study also revealed a high prevalence of symptoms requiring contact with a general practitioner, figures that, to our knowledge, have not been reported previously. Our results are consistent with the prevalences reported in studies regarding severe dumping, anemia, kidney stones, gallstones, and neurologic symptoms.6- 11,14- 17,36- 39 Dumping may, however, be an integrated part of the effect of RYGB surgery in inducing weight loss. Because hypoglycemia consists of unspecific symptoms, patients with this condition might be overlooked or misdiagnosed. Prevalence estimates to date have been based on hospital admission data although many patients with hypoglycemia are not hospitalized.12
In corroboration with the findings by Telem et al21 and Morgan et al,20 we identified a higher prevalence of symptoms among women. The discrepancy with Stroh et al,22 who found no difference between sexes after adjustments, may be explained by anemia not being included in that study. In an additional analysis excluding anemia, we found that both the unadjusted and adjusted results were unchanged.
In contrast to 3 previous studies,20,21,23 we found that younger age was associated with a higher prevalence of symptoms. However, the results of those studies are not directly comparable. Those studies included different symptoms,20,21,23 examined 30-day complication rates,23 or only examined hospitalizations.20,21 After restricting our study to care received in the hospital, the association with younger age persisted (eTable 1 in the Supplement). We hypothesize that age may be an important factor in the conception of symptoms (ie, younger patients may accept fewer symptoms than older patients before contacting a physician).
We found that both cohorts had a reduced health-related QoL vs that of the comparison cohort, which is consistent with the results reported by Raoof et al.40 We only measured the QoL after surgery and therefore were not able to identify changes in QoL. In those undergoing RYGB surgery, 955 of 1314 patients (72.7%) had an MCS below the cutoff for depression risk; in the comparison cohort, 66 of 85 individuals (77.6%) had an MCS below the cutoff for depression risk.41 The negative effect of symptoms that occur after RYGB surgery on the patients’ reported QoL suggests that screening for depression might be beneficial among patients undergoing RYGB surgery, especially among patients with several symptoms.
Since some of the symptoms (eg, diarrhea and fatigue) are common in the general population, we focused on symptoms leading to health care contact as indicators of symptom severity. All Danish patients undergoing RYGB surgery are offered 2 years of postoperative follow-up outpatient visits and then annual follow-up visits in primary care; there is a risk that patients will be unable to remember whether the visit was related to symptoms or simply a part of the postoperative follow-up program, which could lead to inaccurate reports on symptom-related visits. This risk does not apply for symptoms leading to hospitalization; these data were verified by data from the Danish National Patient Registry42 on admissions owing to kidney stones and gallstones. We found satisfactory consistency between self-reported and registry-based data as described by a sensitivity and specificity greater than 0.80 (eTable 2 in the Supplement).43 We assumed that these diagnoses were accurately remembered by the patients and were coded correctly.
The main strengths of our study are its large size, population-based sample, relatively long time frame, and examination of a broad range of symptoms and demographic factors. The size of our comparison cohort was rather small; however, the matching of the comparison cohort for sex and body mass index was good and we believe our comparisons reflect the prevalence of symptoms in overweight individuals . Another concern is that the questionnaire relied on self-reported symptoms elicited a long time after surgery. We did not ask about time relation between surgery and symptoms since we considered that to be a risk of misclassification and that recall bias would be large. The time elapsed since surgery also poses a risk of misclassification in answers concerning preoperative symptoms and weight. Furthermore, we had to construct cutoff points for hypoglycemia and dumping syndrome based on clinical knowledge, and not all scales have been validated exactly in the way we used them.9,34 The cutoff points we defined for dumping produced prevalence estimates consistent with those in the existing literature. Although our questions were comprehensive, focusing on the most common symptoms after RYGB surgery, we may have missed some symptoms. These often rare symptoms are, however, not expected to change the outcome.
Only 63.7% of patients completed the questionnaire. Nonrespondents tended to be younger, but the age distribution differed only by a few percentage points, not comprising a risk of bias (eTable 3 in the Supplement). More nonrespondents than respondents underwent RYGB surgery during 2010-2011, but there were no significant differences in PRs at time since surgery (eTable 4 in the Supplement). More of the nonrespondents were male; since male sex was associated with a lower risk of symptoms, we would expect an even larger sex difference than what we observed. It is possible that the respondents had symptoms that they were interested in addressing. At the same time, respondents satisfied with the results of their operation may have wished to report their satisfaction.
The vast majority of patients reported better or much better well-being after RYGB surgery. However, the amount of symptoms and health care contacts after RYGB surgery is rather high and represents a significant burden from an individual perspective as well as for society. Nearly one-third of patients had been admitted to a hospital owing to these symptoms, mainly surgical symptoms. Focus on the QoL among patients with many symptoms may be required since such patients are at risk of depression. Development of new weight loss treatments with less risk of subsequent symptoms should be a high priority.
Accepted for Publication: October 1, 2015.
Corresponding Author: Sigrid Bjerge Gribsholt, MD, Department of Endocrinology and Internal Medicine, Aarhus University Hospital, Tage-Hansens Gade 2-4, 8000 Aarhus C, Denmark (firstname.lastname@example.org).
Published Online: January 6, 2016. doi:10.1001/jamasurg.2015.5110.
Author Contributions: Dr Gribsholt had full access to all 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: Gribsholt, Richelsen.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Gribsholt, Pedersen, Svensson.
Critical revision of the manuscript for important intellectual content: Svensson, Thomsen, Richelsen.
Statistical analysis: Gribsholt, Pedersen.
Obtained funding: Richelsen.
Administrative, technical, or material support: Richelsen.
Study supervision: Svensson, Thomsen, Richelsen.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by the NovoNordisk Foundation, The A.P. Møller Foundation, and the Research Council of Central Denmark Region.
Role of the Funder/Sponsor: The funding sources 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: Henrik Toft Sørensen, DMSc, PhD, Department of Clinical Epidemiology, Aarhus University Hospital, shared his ideas concerning the design of study, supported the accomplishment of the study, and revised the manuscript critically. He was not compensated for his contribution.