Effect of on

Brian R. Matlaga,* Andrew D. Shore, Thomas Magnuson, Jeanne M. Clark, Roger Johns and Martin A. Makary

From the Departments of Urology, Surgery, Anesthesiology and Medicine, the Johns Hopkins University School of Medicine, and the Departments of Health Policy and Management, Johns Hopkins School of Public Health, Baltimore, Maryland

Purpose: Recent studies have demonstrated that mineral and electrolyte abnor- Abbreviations malities develop in patients who undergo . While it is known and Acronyms that these abnormalities are a risk factor for urolithiasis, the prevalence of stone BCBS ϭ Blue Cross Blue Shield disease after bariatric surgery is unknown. We evaluated the likelihood of being BMI ϭ body mass index diagnosed with or treated for an upper urinary tract calculus following Roux-en-Y ϭ gastric bypass surgery. JI jejunoileal Materials and Methods: We identified 4,639 patients who underwent Roux-en-Y RYGB ϭ Roux-en-Y gastric bypass gastric bypass surgery and a control group of 4,639 obese patients who did not have surgery in a national private insurance claims database in a 5-year period (2002 to Submitted for publication October 15, 2008. 2006). All patients had at least 3 years of continuous claims data. Our 2 primary Supported by The Hariri Family Foundation, and Mr. and Mrs. Chad and Nissa Richinson. outcomes were the diagnosis and the surgical treatment of a urinary calculus. The data set used in this study was originally Results: After Roux-en-Y gastric bypass surgery 7.65% (355 of 4,639) of patients created for a different research project on patterns were diagnosed with urolithiasis compared to 4.63% (215 of 4,639) of obese of care within selected BCBS plans. The Ͻ previous research project (but not the current study) patients in the control group (p 0.0001). Subjects in the treatment cohort more was funded by unrestricted research grants from commonly underwent shock wave (81 [1.75%] vs 19 [0.41%], Ethicon Endo-Surgery, Inc. (a Johnson & Johnson p Ͻ0.0001) and ureteroscopy (98 [2.11%] vs 27 [0.58%], p Ͻ0.0001). Logistic company); Pfizer, Inc and GlaxoSmithKline. The data and database development support and guidance regression demonstrated that Roux-en-Y gastric bypass surgery was a significant were provided by the BCBS Association, BCBS of predictor of being diagnosed with a urinary calculus (OR 1.71, CI 1.44–2.04) as Tennessee, BCBS of Hawaii, BCBS of Michigan, well as undergoing a surgical procedure (OR 3.65, CI 2.60–5.14). BCBS of North Carolina, Highmark, Inc. of Pennsyl- Conclusions: vania, Independence Blue Cross of Pennsylvania, Roux-en-Y gastric bypass surgery is associated with an increased Wellmark BCBS of Iowa and Wellmark BCBS of risk of kidney stone disease and kidney stone surgery in the postoperative period. South Dakota. Clinicians should be aware of this hazard and inform patients of this potential * Correspondence: James Buchanan Brady complication. Future studies are needed to evaluate preventive measures in the Urological Institute, The Johns Hopkins University School of Medicine, 600 North Wolfe St., Balti- high risk population. more, Maryland 21287 (telephone: 410-550-3506; FAX: 410-550-3341; e-mail: [email protected]). Key Words: kidney, calculi, bariatric surgery, epidemiology

OBESITY has become a leading public result in sustained and significant health epidemic with recent estimates weight loss in the majority of candidate that 30% of American adults are obese morbidly obese patients. In fact, the (BMI greater than 30).1 Even more number of bariatric procedures per- alarming is the trend that prevalence formed in the United States increased rates are increasing rapidly,1 creating a 5-fold from 1998 to 2002.3 deferred societal burden of type II dia- The earliest surgical therapy for betes, heart disease, hypertension, the treatment of obesity was the mal- pregnancy complications, sleep apnea absorptive JI bypass procedure. Al- and other health problems.2 Bariatric though JI bypass did induce rapid surgery is the only therapy to date to and lasting weight loss, increasing ex- our knowledge that has been shown to perience with the procedure led to the

0022-5347/09/1816-2573/0 Vol. 181, 2573-2577, June 2009 ® THE JOURNAL OF UROLOGY Printed in U.S.A. www.jurology.com 2573 Copyright © 2009 by AMERICAN UROLOGICAL ASSOCIATION DOI:10.1016/j.juro.2009.02.029 2574 GASTRIC BYPASS SURGERY AND KIDNEY STONE DISEASE

recognition of serious postoperative sequelae such as a BMI greater than 35 and 3 continuous years of followup failure, osteomalacia, nephrolithiasis and renal data were selected and assigned an index date of 7 months failure.4 Nephrolithiasis and renal failure were in- after the first month of enrollment to match the median duced by over-absorption of dietary oxalate, which interval between first enrollment and surgery in the treat- resulted in severe .5 As a consequence ment group. We excluded from the treatment and control groups female subjects who were pregnant the year be- of these metabolic consequences the JI bypass was fore, the year of or the year following an obese BMI value. abandoned and in 1979 the Food and Drug Admin- We also excluded anyone with a preexisting renal disease 6 istration requested a moratorium on the procedure. diagnosis or who had undergone treatment of an upper The Roux-en-Y gastric bypass overcame many urinary tract calculus (64 and 62 patients, respectively). of the malabsorptive problems associated with JI The 2 groups were matched using the SurveySelect bypass, and is recognized as a safe and effective procedure (SAS®, 2003) on the basis of age (18 to 24, 25 to therapy.3,7 However, some reports have suggested 34, 45 to 54, 55 to 64, 65 to 74 and 75 to 84 years), gender, that patients undergoing certain types of bariatric diagnosis of diabetes or prescription for an antihypergly- surgery such as RYGB are at increased risk for cemic agent at any time before the date of surgery or index postoperative hyperoxaluria. It is not known date, and presence of a diagnosis of hypertension or a whether this hyperoxaluria will ultimately result claim for an antihypertensive medication within 6 months of the surgery/index date. One-to-one sampling with prob- in symptomatic renal calculi as it did in patients 8–10 ability proportional to size was used. undergoing JI bypass. To address this ques- The main dependent variables examined in this study tion we designed a study to measure the preva- were a stone removal procedure (shock wave lithotripsy, lence of kidney stone disease following modern ureteroscopy, nephrolithotomy) and a diagnosis of urinary RYGB surgery performed for bariatric indications. calculi (an exclusion criterion if preexisting). The codes used to define these procedures and conditions can be found in Appendix 2. In terms of analysis simple chi- MATERIALS AND METHODS square tests and logistic regression were performed. SAS® The data and in-kind database development support and version 9.13 was used for all analyses. guidance were provided by the Blue Cross/Blue Shield Association, BCBS of Tennessee, BCBS of Hawaii, BCBS RESULTS of Michigan, BCBS of North Carolina, Highmark, Inc. (of Pennsylvania), Independence Blue Cross (of Pennsylva- We identified 4,639 patients who had undergone nia), Wellmark BCBS of Iowa and Wellmark BCBS of RYGB surgery and met the criteria for the treatment South Dakota. All individuals with 1 of these 7 plans as group as well as a matched control group (table 1). primary insurer were eligible for inclusion in the data set. Females outnumbered males 5-to-1 and the majority The claims data used in this study were de-identified in (64%) of subjects were 35 to 54 years old. An upper accordance with the Health Insurance Portability and Ac- urinary tract calculus was diagnosed in 355 subjects countability Act of 1996 definition of a limited data set and (7.65%) in the bariatric surgery group compared to 215 were used in accordance with federal standards for pro- (4.63%) in the comparison cohort (p Ͻ0.0001). Mean tecting the confidentiality of the personal health informa- time from RYGB to diagnosis of an upper tract calcu- tion of the enrollee. The institutional review board of The lus was 558.65 days. In the subsample who experi- Johns Hopkins University found this analysis to be ex- empt from the requirement for review. The data set in- enced kidney stone events subjects who underwent cludes approximately 2.4 million insured lives during a RYGB were not more likely to have experienced mul- 5-year period (2002 to 2006) with information on enrollee tiple stone events than those in the control cohort (OR age, sex, enrollment dates and claims for reimbursement 1.03, 95% CI 0.736–1.437, p ϭ 0.871). for billable health care services. Included in these data are Of the bariatric surgery cohort 3.30% (153 of patient diagnoses as identified by ICD-9 codes and Diag- 4,639) underwent a urological procedure for an up- nosis Related Groups, and medical procedures classified per urinary tract calculus in the 3 years after bari- by CPT codes and ICD-9 procedure codes. atric surgery vs 0.93% (43 of 4,639) of the compari- Subjects undergoing RYGB surgery with enrollment son group (p Ͻ0.0001) (table 2). Ureteroscopy and for 3 years following the date of bariatric surgery were shock wave lithotripsy were used to a similar extent selected for the treatment group. The codes used to define whereas nephrolithotomy was relied on less com- these procedures are found in Appendix 1. There were 18 patients with cancers of the , , small monly. intestine or who were excluded from analysis as Table 1. Demographic information some of the relevant gastric procedure codes for cancer could be mistaken for bariatric operations. A comparison RYGB Group Control Group control group of obese patients meeting the criteria for bariatric surgery was drawn from a subset of the database Mean age (SE) 44.6 (0.14) 45.0 (0.15) % Male/female 19.5/80.5 19.5/80.5 with BMI data from the Health Risk Assessment, a med- Median yrs observation (SE) 4.6 (0.008) 4.1 (0.005) ical history document completed online or based on data Total person-yrs studied 21,362 19,031 collected from physician encounters. Control subjects with GASTRIC BYPASS SURGERY AND KIDNEY STONE DISEASE 2575

Table 2. Summary of kidney stone procedures performed malabsorptive component of the surgery. As fat is mal- No. (%) absorbed fat soluble vitamins and calcium are sapon- p Value ified, leading to subsequent nutrient loss. Concomi- RYGB Group Control Group (chi-square test) tantly an increased oxalate load is delivered to the Shock wave lithotripsy 81 (1.75) 19 (0.41) Ͻ0.0001 colon as the calcium that would normally bind oxalate Ureteroscopy with or without 98 (2.11) 27 (0.58) Ͻ0.0001 in the intestinal lumen is lost with the malabsorbed lithotripsy Percutaneous nephrolithotomy 6 (0.13) 3 (0.06) 0.5076* fat. Oxalate, which cannot be metabolized by humans, Overall 153 (3.30) 43 (0.93) Ͻ0.0001 is cleared by the kidney, resulting in hyperoxaluria and calcium oxalate nephrolithiasis. Of patients * Fisher’s exact test. who underwent JI bypass the risk of stone disease as a result of hyperoxaluria was estimated to be When multivariate logistic regressions were per- approximately 21% at 5 years after surgery.5 formed to determine the strongest predictors of a sur- Recently published laboratory analyses of the long- gical procedure for an upper tract calculus, having term effects of RYGB on the risk of nephrolithiasis are undergone RYGB bariatric surgery was consistently concerning. Asplin and Coe reported their experience the most powerful predictor with odds ratios in the from a commercial laboratory specializing in the uri- range of 3.55 to 4.06 (table 3). RYGB bariatric surgery nary metabolic analyses of kidney stone formers.8 was a significant predictor of being diagnosed with a They found that stone formers who have undergone urinary calculus (OR 1.76, CI 1.49–2.07) as well as modern bariatric procedures are significantly hyperox- undergoing a surgical procedure (OR 3.55, CI 2.58– aluric to a greater extent than idiopathic calcium stone 4.89). A substantial increase in risk appeared to occur formers. In their laboratory analyses normal, non- after the age of 45 years and males were more likely stone formers have a mean urinary oxalate excretion than females to undergo a stone removal procedure. of 34.3 mg daily, calcium oxalate stone formers have a However, the results for the male gender did not mean urinary oxalate excretion of 37.3 mg daily and achieve statistical significance for those undergoing the bariatric population (patients who have undergone shock wave lithotripsy. Due to the small numbers of RYGB or gastric banding procedures) has a mean uri- cases we did not present the results for nephrolithot- nary oxalate of 78.4 mg daily. Importantly the urinary omy, although they were included in the category any calcium oxalate supersaturation of the modern bariat- urological procedure. Neither diabetes nor hyperten- ric surgery cases was not significantly different from a sion was associated with the likelihood of undergoing a cohort of historical cases treated with JI bypass. Sinha stone removal procedure, although hypertension was et al from the Mayo Clinic also evaluated a cohort of associated with the diagnosis of a urinary calculus patients treated exclusively with RYGB for bariatric independent of surgery status. All variables examined indications.9 In a cross-sectional study of patients who are included in table 3. underwent bariatric surgery they found that hyperox- aluria was present in 7 of 13 and calcium oxalate DISCUSSION supersaturation was significantly increased above the RYGB, the most commonly used modern bariatric pro- reference mean in 12 of 13. Duffey et al prospectively cedure, causes weight loss by the 2 mechanisms of 1) studied 24 patients who underwent bariatric sur- 10 surgically reducing the gastric reservoir and 2) bypass- gery. Using the patients as their own control the ing a length of .3 The small size of the authors found that there was a significant postopera- gastric reservoir limits the oral intake of calories by tive increase in urinary oxalate excretion. Urinary su- simple physical restriction, and the small bowel by- persaturation of calcium oxalate was similarly in- pass results in shorter intestinal transit time and re- creased. Taken together these data suggest that duced absorptive surface area. The lithogenic effects of patients undergoing bariatric surgery are at increased bariatric surgery likely stem from the bypass induced risk for the formation of calcium oxalate renal calculi.

Table 3. Multivariate logistic regression of predictors of the diagnosis and treatment of a urinary calculus

Diagnosis of Urinary Calculus Shock Wave Lithotripsy Ureteroscopy Any Urological Procedure OR (95% CI) p Value OR (95% CI) p Value OR (95% CI) p Value OR (95% CI) p Value

RYGB 1.76 (1.49–2.07) Ͻ0.0001 4.06 (2.57–6.43) Ͻ0.0001 4.01 (2.62–6.13) Ͻ0.0001 3.55 (2.58–4.89) Ͻ0.0001 Age 45ϩ yrs 1.33 (1.11–1.59) 0.0019 1.54 (1.02–2.33) 0.0407 1.69 (1.15–2.49) 0.0081 1.68 (1.24–2.28) 0.0009 Male 1.82 (1.52–2.17) Ͻ0.0001 1.39 (0.91–2.14) 0.1274 2.30 (1.59–3.32) Ͻ0.0001 1.79 (1.33–2.4) 0.0001 Diabetes 1.19 (0.98–1.44) 0.0761 1.26 (0.81–1.96) 0.3135 0.91 (0.60–1.40) 0.6774 1.08 (0.78–1.50) 0.6337 Hypertension 1.33 (1.11–1.59) 0.0019 0.88 (0.58–1.32) 0.5248 0.97 (0.67–1.42) 0.8894 0.94 (0.70–1.27) 0.6773 2576 GASTRIC BYPASS SURGERY AND KIDNEY STONE DISEASE

Until our present work it has been uncertain that patients who undergo bariatric surgery are at whether laboratory detected hyperoxaluria in patients significantly increased risk for stone disease and even treated with bariatric surgery would result in an in- oxalate nephropathy.14 Therefore, these patients creased incidence of kidney stone disease. In part this should be counseled on maneuvers to attenuate kidney issue could not be resolved due to the lack of a large stone risk factors including fluid intake to maintain a scale, comparative data set such as the one we have urine output greater than 2 l daily, sufficient calcium presented. However, what data have been previously intake, as well as restricted sodium, animal protein published are suggestive of such a correlation. Enci- and oxalate intake. Should a patient treated with bari- nosa et al reviewed the MarketScan Commercial atric surgery ultimately be diagnosed with nephroli- Claims and Encounter Database and found that thiasis, a metabolic evaluation would be prudent as among the common diagnoses for postoperative condi- these patients are at increased risk for metabolic ab- tions 180 days following bariatric surgery, calculus of normalities, particularly hyperoxaluria. urinary tract occurred with a frequency of 3.6%.7 Un- Our present study has several limitations which fortunately the MarketScan data set has only limited merit mention. In general these limitations are inher- patient demographic information and does not track ent to the use of an administrative claims database. patient BMI, precluding a comparison analysis such as Erroneous and incomplete coding (ie failing to list we were able to perform. codes for all diagnoses relevant to a given admission) The first case series commenting on stone disease may limit our ability to capture all outcomes. How- following modern bariatric surgery was presented in ever, we would expect that incomplete coding and cod- 2005 by Nelson et al.11 They identified 23 patients in ing errors would affect both groups equally. BMI data whom hyperoxaluria and calcium oxalate stone dis- were not available for the cohort of patients undergo- ease developed following RYGB. Of these patients ox- ing RYGB. The BMI threshold of 35 used to define the alate nephropathy and renal failure had developed in control cohort was selected to adequately capture the 2. Durrani et al subsequently reported a single insti- morbidly obese. It should be noted that although this tution, retrospective analysis of stone disease in pa- value does represent a morbidly obese BMI, it is pos- tients undergoing bariatric surgery.12 Of 972 patients sible that certain bariatric surgical centers rely on a identified 8.8% had a preoperative history of upper higher threshold for bariatric surgery. tract calculi and de novo stones developed in 3.2% in the postoperative period. Of the known stone form- CONCLUSIONS ers 31% were diagnosed with a recurrent stone in Modern bariatric surgery is associated with an in- the postoperative period. Overall stone prevalence creased risk of kidney stone disease in the postopera- was reported to be increased by approximately tive period. It is likely that hyperoxaluria, which has 70% in the bariatric surgery cohort compared with been reported to be prevalent following these surgical expected rates derived from the National Health procedures, is one of the inciting factors in stone for- and Nutrition Examination Survey III. mation. As obesity is currently an important national This report represents the largest analysis to our health issue and bariatric surgery to treat it is increas- knowledge of the risk of kidney stone disease following ing at a rapid rate, further studies are urgently re- modern bariatric surgery. As one might expect based quired to better define the etiology of hyperoxaluria as on the previously reported laboratory studies of pa- well as other stone risk factors in this population. tients who had undergone bariatric surgery, we did Thus, future efforts can be devoted to altering risk find an increased incidence of stone disease in this factors for this disorder, thereby attenuating kidney population relative to our control population of obese stone formation following bariatric surgery. However, patients who did not undergo bariatric surgery. How- until the pathophysiology of stone formation in this ever, one should not necessarily interpret our findings population is better understood it is incumbent upon of an increased risk of stone disease following bariatric physicians to maintain a high level of suspicion for surgery as an indictment of this surgical intervention. underlying urinary metabolic abnormalities in the The long-term health benefits of bariatric surgery are stone former after undergoing bariatric surgery. well characterized and significant.13 Rather our work should serve as a cautionary note to those physicians who treat patients undergoing bariatric surgery. The ACKNOWLEDGMENTS epidemiological findings presented here confirm previ- Eric Bass and Jonathan Weiner provided study sup- ous laboratory studies and case series, and suggest port. GASTRIC BYPASS SURGERY AND KIDNEY STONE DISEASE 2577

APPENDIX 1 Codes Used to Define Bariatric Surgical Procedures

CPT Code Description

43644 , surgical, gastric restrictive procedure; with gastric bypass and Roux-en-Y (roux limb 150 cm or less) 43645 Laparoscopy, surgical, gastric restrictive procedure; with gastric bypass and small intestine reconstruction to limit absorption 43846 Gastric restrictive procedure, with gastric bypass for morbid obesity; with short limb (150 cm or less) Roux-en-Y gastroenterostomy 43847 Gastric restrictive procedure, with gastric bypass for morbid obesity; with small bowel reconstruction to limit absorption

APPENDIX 2 Codes Used to Define Urological Procedures and Conditions

Indicator Type of Code Description

Urological Procedures Shock wave lithotripsy CPT code 50590 Lithotripsy, extracorporeal shock wave S0400 Global fee for extracorporeal shock wave lithotripsy treatment of kidney stone(s) ICD-9 procedure code 98.5 Extracorporeal shock wave lithotripsy 98.51 Extracorporeal shock wave lithotripsy of the kidney, ureter and/or bladder Ureteroscopy CPT code 52352 Cystourethroscopy, with ureteroscopy and/or pyeloscopy; with removal or manipulation of calculus 52353 Cystourethroscopy, with ureteroscopy and/or pyeloscopy; with lithotripsy Nephrolithotomy CPT code 50060 Nephrolithotomy; removal of calculus (percutaneous and open) 50065 Nephrolithotomy; secondary surgical operation for calculus 50070 Nephrolithotomy; complicated by congenital kidney abnormality 50075 Nephrolithotomy; removal of large staghorn calculus filling renal pelvis and calices 50080 Percutaneous nephrostolithotomy or pyelostolithotomy, up to 2 cm 50081 Percutaneous nephrostolithotomy or pyelostolithotomy, over 2 cm ICD-9 procedure code 55.03 Nephrostomy Renal Disease Indicators Urinary calculi Diagnosis Related Group 323 Urinary stones with complication or comorbidity and/or extracorporeal shock wave lithotripsy 324 Urinary stones without complication or comorbidity ICD-9 diagnosis code 274.11 Uric acid nephrolithiasis 592 Calculus of kidney and ureter 592.0 Calculus of kidney—nephrolithiasis not otherwise specified 592.1 Calculus of ureter 592.9 Urinary calculus, unspecified

REFERENCES 1. Flegal KM, Carroll MD, Ogden CL and Johnson tinuing potential need for surgical reversal. Arch 11. Nelson WK, Houghton SG, Milliner DS, Lieske JC CL: Prevalence and trends in obesity among US Surg 1995; 130: 318. and Sarr MG: Enteric hyperoxaluria, nephrolithi- adults, 1999-2000. JAMA 2002; 288: 1723. asis, and oxalate nephropathy: potentially serious 7. Encinosa WE, Bernard DM, Chen CC and Steiner and unappreciated complications of Roux-en-Y 2. Roth J, Qiang X, Marban SL, Redelt H and Lowell CA: Healthcare utilization and outcomes after gastric bypass. Surg Obes Relat Dis 2005; 1: 481. BC: The obesity pandemic: where have we been bariatric surgery. Med Care 2006; 44: 706. and where are we going? Obes Res 2004; 12: 12. Durrani O, Morrisroe S, Jackman S and Averch T: 88S. 8. Asplin JR and Coe FL: Hyperoxaluria in kidney Analysis of stone disease in morbidly obese pa- 3. Santry HP, Gillen DL and Lauderdale DS: Trends stone formers treated with modern bariatric sur- tients undergoing gastric bypass surgery. J En- in bariatric surgical procedures. JAMA 2005; 294: gery. J Urol 2007; 177: 565. dourol 2006; 20: 749. 1909. 9. Sinha MK, Collazo-Clavell ML, Rule A, Milliner 13. Buchwald H, Avidor Y, Braunwald E, Jensen MD, 4. Kirkpatrick JR: Jejunoileal bypass. A legacy of DS, Nelson W, Sarr MG et al: Hyperoxaluric Pories W, Fahrbach K et al: Bariatric surgery: a late complications. Arch Surg 1987; 122: 610. nephrolithiasis is a complication of Roux-en-Y systematic review and meta-analysis. JAMA gastric bypass surgery. Kidney Int 2007; 72: 100. 2004; 292: 1724. 5. Mole DR, Tomson CR, Mortensen N and Winearls CG: Renal complications of jejuno-ileal bypass for obesity. QJM 2001; 94: 69. 10. Duffey BG, Pedro RN, Makhlouf A, Kriedberg C, 14. Nasr SH, D’Agati VD, Said SM, Stokes MB, Lar- Stessman M, Hinck B et al: Roux-en-Y gastric goza MV, Radhakrishnan J et al: Oxalate ne- 6. Requarth JA, Burchard KW, Colacchio TA, Stukel bypass is associated with early increased risk phropathy complicating Roux-en-Y gastric bypass: TA, Mott LA, Greenberg ER et al: Long-term factors for development of calcium oxalate neph- an underrecognized cause of irreversible renal morbidity following jejunoileal bypass. The con- rolithiasis. J Am Coll Surg 2008; 206: 1145. failure. Clin J Am Soc Nephrol 2008; 3: 1676.