Kidney stones are common after bariatric surgery

clinical investigation http://www.kidney-international.org & 2014 International Society of Nephrology Kidney stones are common after bariatric surgery John C. Lieske1,2, Ramila A. Mehta3, Dawn S. Milliner1, Andrew D. Rule1,4, Eric J. Bergstralh3 and Michael G. Sarr5 1 Division of Nephrology and Hypertension, Mayo Clinic, Rochester, Minnesota, USA; 2Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota, USA; 3Division of Biomedical Statistics and Informatics, Mayo Clinic, Rochester, Minnesota, USA; 4Division of Epidemiology, Mayo Clinic, Rochester, Minnesota, USA and 5Department of Surgery, Mayo Clinic, Rochester, Minnesota, USA Obesity, a risk factor for kidney stones and chronic kidney disease (CKD), is effectively treated with bariatric surgery. However, it is unclear whether surgery alters stone or CKD risk. To determine this we studied 762 Olmsted County, Minnesota residents who underwent bariatric surgery and matched them with equally obese control individuals who did not undergo surgery. The majority of bariatric patients underwent standard Roux-en-Y gastric bypass (RYGB; 78%), with the remainder having more malabsorptive procedures (very long limb RYGB or biliopancreatic diversion/duodenal switch; 14%) or restrictive procedures (laparoscopic banding or sleeve gastrectomy; 7%). The mean age was 45 years with 80% being female. The mean preoperative body mass index (BMI) was 46.7 kg/m2 for both cohorts. Rates of kidney stones were similar between surgery patients and controls at baseline, but new stone formation significantly increased in surgery patients (11.0%) compared with controls (4.3%) during 6.0 years of follow-up. After malabsorptive and standard surgery, the comorbidity-adjusted hazard ratio of incident stones was significantly increased to 4.15 and 2.13, respectively, but was not significantly changed for restrictive surgery. The risk of CKD significantly increased after the malabsorptive procedures (adjusted hazard ratio of 1.96). Thus, while RYGB and malabsorptive procedures are more effective for weight loss, both are associated with increased risk of stones, while malabsorptive procedures also increase CKD risk. Kidney International (2015) 87, 839–845; doi:10.1038/ki.2014.352; published online 29 October 2014 KEYWORDS: bariatric surgery; hyperoxaluria; nephrolithiasis; obesity Utilization of bariatric surgery continues to be high in the United States. Recent large, randomized trials confirm that patients have sustained weight loss, less mortality, and a decrease in obesity-related complications, such as diabetes, hypertension, and obstructive sleep apnea.1,2 Thus, the number of bariatric procedures performed annually in the United States has increased from 12,775 to a peak of 135,985 in 2004; rates have since plateaued. In 2008, B70% of bariatric procedures were Roux-en-Y gastric bypass (RYGB),3 the preferred procedure, because it is associated with acceptably low morbidity and improved absolute and sustained weight loss compared with restrictive procedures (mainly adjustable gastric banding). Recently, sleeve gastrectomy has been reported to have an efficacy between that of gastric banding and RYGB.3 Still RYGB is viewed as a more durable and effective procedure, especially in cases of severe obesity, and represented 56% of procedures in 2012.3 The number of existing persons in the United States with RYGB procedures performed between 1998 and 2008 can be estimated to be B830,000.4 We previously reported a high incidence of hyperoxaluria and kidney stones among patients who had undergone RYGB for obesity.5 Others have made similar observations in other patient cohorts.6–8 The risk of hyperoxaluria and perhaps kidney stones may be less with other forms of bariatric surgery.9–11 However, the risk for kidney stones and/or chronic kidney disease (CKD) with bariatric surgery remains unclear, because these studies were either not population-based or lacked controls with similar obesity and comorbidities who did not undergo bariatric surgery. Thus, in the current study we used the resources of the Rochester Epidemiology Project12 to conduct a populationbased study to compare the incidence of stones in patients after bariatric surgery with that in comorbidity-matched obese controls. RESULTS Correspondence: John C. Lieske, Division of Nephrology and Hypertension, Mayo Clinic, 200 First Street South West, Rochester, Minnesota 55905, USA. E-mail: [email protected] Received 4 June 2014; revised 20 August 2014; accepted 4 September 2014; published online 29 October 2014 Kidney International (2015) 87, 839–845 There were 2683 patients with a history of bariatric surgery at Mayo Clinic during the study period. After excluding those without research authorization (n ¼ 63), Olmsted County residency (n ¼ 1832), or preoperative body mass index (BMI) greater than 35 kg/m2 (n ¼ 26), there were 762 bariatric 839 clinical investigation JC Lieske et al.: Kidney stones after bariatric surgery Table 1 | Types of bariatric operations 2000–2011 Malabsorptive procedure RYGB Restrictive Table 2 | Demographics of bariatric patients and controls Total number Number by year 2000–2003 2004–2007 2008–2011 105 591 56 7 43 36 26 127 273 191 0 15 41 0 1 6 Years of follow-up (mean (s.d.)) Age at surgery (mean (s.d.)) Sex (% female) BMI kg/m2 (mean (s.d.)) 6.2 (3.7) 6.2 (3.1) 3.9 (1.6) 3.1 (2.6) 44.1 (11.0) 44.9 (11.2) 42.7 (12.0) 52.3 (9.3) 70.5% 56.3 (8.4) 82.9% 45.3 (6.5) 78.6% 57.1% 43.6 (7.6) 45.4 (10.0) Abbreviations: BMI, body mass index; RYGB Roux-en-Y gastric bypass. surgery patients to be studied. There were 13,256 Olmsted County residents with a BMI 435 kg/m2 during the study period. After excluding those who had undergone bariatric surgery (n ¼ 699) and subjects who refused research authorization (n ¼ 63), 12,494 potential controls remained. With 1:1 matching, we were able to identify controls for 759 of the 762 bariatric surgery patients. Among the bariatric operations performed, most (n ¼ 591, 78%) were standard RYGB procedures (Table 1). The majority of standard RYGB operations before 2007 were open procedures (n ¼ 188), whereas laparoscopic procedures predominated after 2004 (n ¼ 404). When a greater amount of weight loss was deemed desirable, procedures that were typically more malabsorptive in nature were performed, including very, very long limb RYGB (n ¼ 55) or biliopancreatic diversion/duodenal switch (n ¼ 50). At our institution, a relatively small number of restrictive procedures, including laparoscopic banding (n ¼ 43) or laparoscopic sleeve gastrectomy (n ¼ 13), were completed during the years of the study. The mean (s.d.) age at the time of bariatric surgery was 44.7 (11.2) years; 80% were female, and the mean preoperative BMI was 46.7(7.9) kg/m2; because of matching, these parameters were similar in controls (Table 2). Baseline comorbidities, including hypertension, diabetes, osteoarthritis, and sleep apnea, were more common in bariatric surgery patients than in obese controls (Table 2). CKD at baseline was similar between the two groups (10.4% vs. 8.7%; P ¼ 0.26). Nephrolithiasis at baseline had similar frequency in bariatric surgery patients and controls (4.0% vs. 4.2%; P ¼ 0.70). In contrast, over a mean follow-up period of 6.0 (3.2) years, new (incident) stone events were found to be more common in bariatric surgery patients than in obese controls (11.1% vs. 4.3%; Po0.01, Table 2). Kaplan–Meier plots confirmed that kidney stone events increased among operated patients within the first 2 years, reaching B14% at 10 years compared with 7% of controls at that time point. Among the stones analyzed, calcium oxalate were the most common in bariatric patients before the procedure (73%) 840 Bariatric patients (n ¼ 759) Controls (n ¼ 759) P-valuea 44.7 (11.2) 80.6% 46.7 (7.9) 6.9 (3.4) 44.7 (11.2) 80.6% 46.7 (7.8) 7.0 (3.3) Matched Matched Matched 0.42 52.3% 27.1% 56.1% 56.8% 47.2% 22.8% 8.8% 30.4% 0.05 0.05 o0.001 o0.001 CKD events Prevalence at baseline New (incident) 10.4% 7.9% 8.7% 9.6% 0.26 0.24 Nephrolithiasis events Prevalence at baseline New (incident) stones 4.3% 11.1% 4.0% 4.3% 0.70 o0.01 Other Demographics Age, years (mean, s.d.) Female sex BMI, kg/m2 (mean, s.d.) Years of follow-up (mean, s.d.) Hypertension Diabetes Arthritis Sleep apnea Abbreviations: BMI, body mass index; CKD, chronic kidney disease. a P-value from w2 test (nominal factors), rank sum test (continuous factors), and logrank test (time to incident CKD and stones). Table 3 | Available stone composition of bariatric and obese patients, before and after the incident date Bariatric cohort Hydroxyapatite Calcium oxalate Struvite Uric acid Not available Prevalent Incident 3 21 2 3 2 63 1 1 (10%) (73%) (7%) (10%) 4 Obese cohort Prevalent and incident (3%) (94%) (1.5%) (1.5%) 17 7 (31%) 15 (65%) 1 (4%) 0 (0%) 40 Values expressed as total number and % of those available. and in obese control patients (65% before or after the incident date combined; Table 3). However, calcium oxalate stones were even more common after bariatric surgery, representing 94% of those analyzed (P ¼ 0.02 for comparison of the stone distribution between the post-bariatric group and obese stone formers, using Fisher’s exact test). Stone risk varied by type of procedure, being highest in patients who had undergone a malabsorptive procedure, intermediate in those who had undergone standard RYGB, and lowest in those who had undergone restrictive procedures, which were similar to the rates seen in obese controls (Figure 1). Baseline diabetes, osteoarthritis, sleep apnea, and being a bariatric surgery case were all significant risk factors for incident stones (Table 4). In multivariable models, only diabetes, RYGB, and malabsorptive procedures remained statistically significant risk factors. Patients with a history of a prior stone (prevalent) at the time of bariatric surgery were more likely to develop a stone after surgery compared with those without a prior stone history (42% vs. 14% at 10 years; hazard ratio ¼ 4.1, Po0.001). However, the risk of prevalent obese patients forming a second stone was slightly higher (52% at Kidney International (2015) 87, 839–845 clinical investigation 50 50 Malabsorptive RYGB Control Restrictive 40 New onset chronic kidney disease (%) New onset nephrolithiasis (%) JC Lieske et al.: Kidney stones after bariatric surgery 30 20 10 40 30 20 10 0 No. at risk Malabsorptive RYGB Control Restrictive 0 2 4 6 8 Years after surgery/index date 101 563 729 55 77 502 673 54 58 408 572 36 47 293 416 18 0 10 36 208 282 0 27 94 135 0 Malabsorptive Control RYGB Restrictive No. at risk Malabsorptive Control RYGB Restrictive 0 2 4 6 8 Years after surgery/index date 93 693 530 51 78 644 484 50 61 548 400 32 49 392 296 15 10 34 255 213 0 24 124 99 0 Figure 1 | Risk of new-onset nephrolithiasis after bariatric surgery. The risk of incident stones was greater after Roux-en-Y gastric bypass (RYGB) or malabsorptive bariatric procedures, compared with that in matched obese controls (Po0.001 overall). Patients with restrictive procedures were not at increased risk. Figure 2 | Risk of new-onset chronic kidney disease (CKD) after bariatric surgery. The risk of incident CKD was greater after malabsorptive bariatric procedures compared with that in matched obese controls (P ¼ 0.004 overall). Patients were not at increased CKD risk after Roux-en-Y gastric bypass (RYGB) or restrictive procedures. Table 4 | Univariable and multivariable models of hazard ratios for kidney stones Table 5 | Univariable and multivariable models of hazard ratios for CKD Univariable Risk factor HR Age at surgery (per 10 years) Male sex Hypertension Diabetes Arthritis Sleep apnea 1.04 0.88–1.23 1.11 0.89 0.51 0.46 0.67 95% CI Multivariable P-value 0.65 0.70–1.76 0.64 0.61–1.28 0.52 0.35–0.75 0.0005 0.31–0.65 o0.001 0.47–0.97 0.034 Type of bariatric surgery Control (referent) 1.00 NA NA RYGB 2.49 1.63–3.81 o0.001 Malabsorptive 5.23 3.02–9.05 o0.001 Restrictive 0.50 0.07–3.69 0.50 HR 95% CI Univariable P-value 0.99 0.82–1.20 0.93 1.00 1.11 0.55 0.77 1.00 0.99 0.62 0.004 0.25 0.99 0.62–1.61 0.73–1.69 0.37–0.82 0.50–1.2 0.66–1.50 1.00 NA NA 2.13 1.30–3.49 0.003 4.15 2.16–8.00 o0.001 0.46 0.06–3.45 0.45 95% CI Multivariable Risk factor HR Age at surgery (per 10 years) Male sex Hypertension Diabetes Arthritis Sleep apnea 1.44 1.23–1.70 o0.001 1.23 1.02–1.48 1.86 0.47 0.28 0.98 0.72 1.40 0.70 0.34 1.05 1.10 P-value 1.27–2.74 0.0014 0.33–0.66 o0.001 0.20–0.39 o0.001 0.68–1.40 0.91 0.52–1.02 0.06 Type of bariatric surgery Control (referent) 1.00 NA RYGB 0.71 0.48–1.05 Malabsorptive 1.91 1.14–3.20 Restrictive 0.61 0.15–2.49 NA 0.09 0.01 0.49 HR 95% CI P-value 0.03 0.93–2.10 0.11 0.47–1.04 0.08 0.24–0.49 o0.001 0.67–1.65 0.83 0.75–1.61 0.63 1.00 NA 0.70 0.44–1.12 1.96 1.06–3.64 0.73 0.18–3.04 NA 0.14 0.03 0.67 Abbreviations: CI, confidence interval; HR, hazard ratio; NA, not applicable; RYGB, Roux-en-Y gastric bypass. Abbreviations: CI, confidence interval; CKD, chronic kidney disease; HR, hazard ratio; NA, not applicable; RYGB, Roux-en-Y gastric bypass. 10 years). Thus, this observation likely reflects the known tendency for stone event risk to increase as the number of prior events increases,13 and does not suggest that bariatric surgery disproportionately augments stone risk among those with past stone events. Overall, bariatric surgery was not a risk factor for developing CKD (hazard ratio ¼ 0.95, confidence interval 0.67–1.35). However, when evaluated by type of bariatric procedure, those patients with malabsorptive procedures were at increased risk (Figure 2). In multivariable modeling that controlled for diabetes and hypertension, the hazard ratio remained increased for those undergoing the malabsorptive procedure (Table 5). Urinary supersaturation profiles were available after bariatric surgery for 55 patients with follow-up stones and for 248 without follow-up stones, as well as for 20 obese controls with follow-up stones (Table 6). For the bariatric surgery group, data are presented broken down as o8 months after surgery (first follow-up visits) and 48 months after surgery (later visits). Urine oxalate excretion increased with time after surgery (o8 vs. 48 months; Po0.001) and was most prominent in the post-bariatric patients who developed stones 48 months after surgery (0.70 (0.37) (stone) vs. 0.47 (0.34) mmol/day (no stone); Po0.001). Urine citrate was also lower in this group (448 (340) (stone) vs. 610 (417) mg/day (no stone); Po0.001), resulting in higher overall CaOx supersaturations (2.12 (0.92) (stone) vs. 1.50 (0.95) delta Gibbs (no stone); Po0.001). Urine volumes were appreciably lower in the o8 months’ group regardless of stone status, resulting in higher supersaturations despite lower oxalate excretion at this time period. Increasing urine Kidney International (2015) 87, 839–845 841 clinical investigation JC Lieske et al.: Kidney stones after bariatric surgery Table 6 | Twenty-four hours urine chemistries by stone status Post-bariatric surgery o8 Months Time after surgery/index date Number w/labs Oxalate, mmol Calcium, mg Citrate, mg Uric acid, mg pH Sodium, mEq Chloride, mEq Potassium, mEq Phosphorous, mg Magnesium, mg Sulfate, mmol Creatinine, mg Volume, ml CaOx SS, DG CaP (apatite) SS, DG CaP (brushite) SS, DG Uric acid SS, DG 48 Months No stone Stone No stone Stone 136 0.35 (0.18) 152 (90) 690 (397) 434 (153) 5.9 (0.5) 122 (62) 115 (61) 41 (22) 629 (293) 89 (39) 11.4 (7.3) 1228 (383) 1380 (572) 2.02 (0.77) 2.98 (2.03)  0.73 (1.42) 0.90 (2.73) 13 0.29 (0.20) 163 (73) 739 (485) 477 (190) 6.0 (0.6) 127 (81) 121 (74) 35 (19) 568 (335) 88 (44) 9.9 (5.5) 1308 (573) 1094 (141) 2.36 (0.44) 3.80 (2.12)  0.12 (1.18) 1.32 (2.86) 112 0.47 (0.34)@ 138 (99) 610 (417) 466 (176) 6.0 (0.5) 165 (88)@ 155 (92)@ 51 (25)@ 823 (390)@ 134 (62)@ 14.7 (7.0)@ 1269 (449) 1970 (1039)@ 1.50 (0.95)@ 2.69 (2.02)  1.11 (1.39)  0.10 (3.08) 42 0.70 (0.37)* 135 (83) 448 (340)& 447 (183) 5.8 (0.5) 182 (101) 183 (107) 46 (28) 800 (304) 127 (56) 14.6 (6.5) 1297 (465) 1793 (140) 2.12 (0.92)* 2.11 (2.43)  1.34 (1.67) 0.82 (2.46) Obese stone formers 43 (25) months (mean (s.d.)) 0.39 170 638 497 6.1 152 147 59 704 115 16.7 1329 1834 1.69 2.98  1.01  0.21 20 (0.20)*** (107) (340) (186) (0.8) (75) (71) (30)* (333) (51) (7.9) (186) (784) (0.38)** (2.77) (1.69) (3.45) Abbreviations: DG, delta Gibbs; RYGB, Roux-en-Y gastric bypass. *Po0.001 vs. no stone at 48 months. & Po0.05 vs. no stone. @ P ¼ 0.01 vs. no stone at o8 months. **Po0.05 vs. RYGB stone formers. ***Po0.005 vs. RYGB stone formers. Bold values indicate significant difference. 3.5 3 1.2 2.5 1 CaOxSS (DG) Urine oxalate (mmol/day) 1.4 0.8 0.6 0.4 2 1.5 1 0.5 0 0.2 –0.5 0 –1 0 6 12 18 24 Time since surgery/index date 30 0 6 12 18 24 Time since surgery/index date 30 Figure 3 | Changes in urine oxalate and CaOx SS after surgery. (a) Urinary oxalate increased subtly in all cases over time after bariatric surgery (~ solid diamonds, - - - - dashed line), and more dramatically in those who developed stones (D open triangles, ______ solid line). The mean urine oxalate was at the upper limit of the reference value (0.46 mmol/day) at all time points in obese controls who developed stones (J open circles, _. _. _. _ dash-dot line). (b) At all time points CaOx SS was highest in the post–bariatric-surgery patients who developed stones (D open diamonds, _____ solid line), but still at or above the reference mean (1.77 DG) in both obese controls with stones (J open circles, _. _. _ dash-dot line) and in post–bariatric-surgery patients without stones (~ solid diamonds, - - - - dashed line). oxalate excretion was characteristic of the post-bariatric stone-forming group, with oxalate excretions appreciably higher than that in non-stone-forming post–bariatric surgery patients and obese stone formers (Figure 3a). Overall, CaOx SS tended to decrease in the non-stone-forming postbariatric group, but remained high in those patients who experienced stone formation after surgery (Figure 3b). DISCUSSION The current study provides strong objective evidence that the risk for kidney stones is approximately doubled in patients 842 after RYGB compared with matched, non-operated, obese controls. This risk depends on the choice of specific type of bariatric surgery, being greatest in malabsorptive procedures, intermediate in standard RYGB, and least in restrictive procedures. Although it has been recognized that hyperoxaluria and kidney stones can occur after RYGB, these are some of the initial data to quantify the risk for stones in comparison with a group of obese, non-operated controls. Among those operated patients with available urinary data, hyperoxaluria was increasingly common, although not marked until B18 months after the procedure. Overall, Kidney International (2015) 87, 839–845 JC Lieske et al.: Kidney stones after bariatric surgery however, urinary saturations for calcium oxalate were increased at all time points after bariatric surgery. As expected, the urinary profile of obese controls with newonset stones differed, with hyperoxaluria being less of a feature. Previous studies suggested that hyperoxaluria develops in up to 50% of individuals after certain forms of bariatric surgery (e.g., RYGB)5–8 but not in others (e.g., those who have undergone restrictive procedures such as adjustable lap banding),9,10 and that patients with hyperoxaluria after bariatric surgery are at increased risk for nephrolithiasis.14 Perhaps the best previous quantitative data were estimated from a review of a database of private insurance claims that contained 4639 patients who were matched to obese controls.14 Over a median follow-up of B4 years, 7.65% of bariatric patients were seen to have developed a stone, compared with 4.63% of obese controls (odds ratio 1.71). The mean time from bariatric surgery to the stone event was 1.5 years, and RYGB patients were also more likely to have a surgical intervention for stones (odds ratio 3.65). In contrast, laparoscopic adjustable gastric band surgery or sleeve gastrectomy did not appear to increase stone risk.11,15 These findings are consistent with our current study. However, these studies were not population-based and also were not controlled for other risk factors (e.g., diabetes and hypertension). Enteric hyperoxaluria occurs in association with fat malabsorption and is believed to develop when oxalate in the diet is delivered to the colon without sufficient calcium. This pathophysiology occurs when gastrointestinal disorders affect the mucosa of the ileum, when the ileum has been resected (or bypassed) leading to short bowel syndrome, or in association with pancreatic insufficiency. In recent times, a large number of patients have been recognized with hyperoxaluria and calcium oxalate stones after bariatric procedures performed for weight loss.5,16 Most commonly these procedures were RYGB, with or without other features to lead to relatively more malabsorption to enhance weight loss potential. Two examples in our current cohort are the biliopancreatic diversion/duodenal switch and the very, very long limb RYGB. It has been suspected that these patients might be at increased risk for oxalate-related complications because of a greater likelihood and degree of fat malabsorption.5,17 Indeed, the current study provides strong objective evidence that the more malabsorptive procedures have a greater risk of kidney stones compared with standard RYGB. The mechanism(s) by which RYGB patients develop hyperoxaluria is yet to be fully explained, especially because the entire ileum is in continuity with the enteric stream; however, it seems likely that the length of the common channel, at least in the small group of bariatric patients who undergo the distal malabsorptive so-called very, very long limb RYGB or the biliopancreatic diversion/duodenal switch, may predispose to clinically important fat malabsorption, leading to enteric hyperoxaluria. Previous studies have Kidney International (2015) 87, 839–845 clinical investigation suggested that the extent of hyperoxaluria corresponds to the degree of steatorrhea,18 as observed in the various disease conditions associated with fat malabsorption, including inflammatory bowel disease, ileal resection, and jejunoileal bypass.19–22 Interestingly, none of the 31 patients in our published Mayo Clinic series who were seen in the stone clinic with nephrolithiasis after RYGB reported symptoms of diarrhea.5 In the one stone clinic patient in whom fat malabsorption was assessed, 72-h fecal fat excretion was increased (57 g; normal o7 g), despite the absence of diarrhea. In a single case series, steatorrhea was reported in 18 of 45 patients after biliopancreatic bypass for obesity, although clinical symptoms were not reported.23 Not all patients with gastrointestinal disorders characterized by fat malabsorption develop kidney stones. A recent study identified 51 patients with fat malabsorption and compared 10 stone formers with 41 non-stone formers.24 The stone-forming group was characterized by greater urinary oxalate excretion (0.66 vs. 0.38 mmol per day), less urinary citrate excretion (309 vs. 607 mg/day), and greater calcium oxalate supersaturation (relative supersaturation for CaOx of 8.16 vs. 3.94). Interestingly, fecal elastase, serum betacarotene, and vitamin E were all less in the stone-former group, suggesting that the degree of fat malabsorption was a key differentiating feature. Our data are consistent with the paradigm that the degree of fat malabsorption is a key determinant of kidney stone risk. Because fecal fat collections are difficult to achieve, it will be interesting to see whether some panels of markers such as these can be used to identify those bariatric patients at greatest risk for stone formation. In addition, it would be interesting to extend these measurements to a group of idiopathic calcium oxalate stone formers in order to determine whether subclinical fat malabsorption is often or ever a contributing factor in that patient group as well. Kidney stones are not the only potential consequence of enteric hyperoxaluria after RYGB. Cases of oxalate nephropathy after RYGB have been clearly documented.25,26 The prevalence and risk factors for this severe outcome are less certain. In our clinics, several patients with enteric hyperoxaluria after bariatric surgery have progressed to kidney failure regardless of prescribed therapy.26 The presence of CKD before RYGB may be an important predisposing factor.25 In the current cohort the malabsorptive procedures, but not other forms of bariatric surgery, appeared to increase the risk for new-onset CKD. Clearly, only a relatively small percentage of the vast number of patients undergoing bariatric surgery have developed CKD and/or end-stage kidney disease to date. Reliable methods to detect which of these patients would benefit most from earlier identification and aggressive management so that renal failure can be prevented are clearly needed. Limitations of this study include its retrospective nature and lack of randomization. Further, the results apply primarily to Caucasians. In addition, kidney stones, CKD, and comorbidities were determined by diagnosis codes, and 843 clinical investigation laboratory data were available for only a subset of all patients with or without kidney stones. The incidence and prevalence of CKD post-bariatric surgery might also have been underestimated because of the effects of weight loss on creatinine generation and serum creatinine levels.28 It is also possible that kidney stones or CKD may have been more readily diagnosed in the bariatric surgery group if they were undergoing more medical care and diagnostic procedures. However, we were able to study a population-based, matched cohort of obese patients who had or had not undergone bariatric surgery. Furthermore, detailed lab data were available in a sizable number of post-bariatric-surgery patients, including many without kidney stones. In conclusion, the current study suggests that obese patients who undergo RYGB have an increased risk for kidney stones that is approximately double that of obese, nonoperated controls. About one in five individuals will develop a stone within 10 years after these bariatric procedures. Urinary risk factors include hyperoxaluria, low urine volume, and hypocitraturia. Patients with malabsorptive bariatric procedures appear at greatest risk for stones but are also at increased risk for new-onset CKD. Future studies are needed to develop better treatment strategies and identify those at greatest risk for this complication. MATERIALS AND METHODS Data on all patients with a history of bariatric surgery at Mayo Clinic between the years 2000 and 2011 were taken from our institutional bariatric surgical registry, which included the type and date of procedure. Patients without research authorization, Olmsted County residency, or preoperative BMI greater than 35 kg/m2 were excluded. Using the Rochester Epidemiology Project12 controls were sampled from among all Olmsted County residents with a BMI greater than 35 kg/m2 who were seen at Mayo Clinic during the study period. Using validated matching algorithms27 surgery cases were individually matched exactly to controls in terms of sex and index year (BMI date in controls closest to preoperative BMI in surgery patients) and for BMI within ±3. Further, penalty weights were assigned to large age and BMI differences. With these criteria, 759 of the 762 cases were matched, with 96% having an age within 5 years. Over 95% of the Olmsted County population is seen by a local health-care provider over any 3-year period,12 and thus the Rochester Epidemiology Project was used to capture follow-up kidney stone and CKD events in both bariatric surgery patients and controls. Kidney or bladder stones were identified using the International Classification of Diseases-9 codes 592, 594, and 274.11 as reported previously.29 The initial dates of comorbidities (hypertension, diabetes, sleep apnea, osteoarthritis, and CKD) were identified from International Classification of Diseases-9 codes as described previously.29 Clinical laboratory test results, including results of 24-h urine studies, were taken from the electronic medical record. In the bariatric surgery group, urine chemistries were obtained as part of routine follow-up visits beginning B6 months post surgery or potentially at the time of a nephrology stone clinic visit if they developed stones. Urine studies from obese controls were only available at the time of a nephrology stone clinic visit. 844 JC Lieske et al.: Kidney stones after bariatric surgery Statistical analyses The associations of bariatric surgery with a subsequent kidney stone event and CKD were assessed using Kaplan–Meier plots and Cox proportional hazard models with adjustments for age, sex, and other baseline comorbidities. Subjects with prevalent kidney stones were excluded from the analyses of new-onset stones. Risk factor effects are presented as hazard ratios of the event rates and 95% confidence intervals. All tests were two-sided with alpha level 0.05. All analyses used SAS v9.3 software (SAS Institute, Cary, NC). DISCLOSURE All the authors declared no competing interests. ACKNOWLEDGMENTS This study was supported by the Mayo Clinic O’Brien Urology Research Center (U54 DK100227), the Rare Kidney Stone Consortium (U54KD083908), a member of the NIH Rare Diseases Clinical Research Network (RDCRN), funded by the NIDDK and the National Center for Advancing Translational Sciences (NCATS), the Mayo Hyperoxaluria Center, and the Mayo Foundation. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. Carlsson LM, Peltonen M, Ahlin S et al. Bariatric surgery and prevention of type 2 diabetes in Swedish obese subjects. N Engl J Med 2012; 367: 695–704. Sjostrom L, Peltonen M, Jacobson P et al. Bariatric surgery and long-term cardiovascular events. JAMA 2012; 307: 56–65. Nguyen NT, Nguyen B, Gebhart A et al. Changes in the makeup of bariatric surgery: a national increase in use of laparoscopic sleeve gastrectomy. J Am Coll Surg 2013; 216: 252–257. Nguyen NT, Masoomi H, Magno CP et al. Trends in use of bariatric surgery, 2003-2008. J Am Coll Surg 2011; 213: 261–266. Sinha MK, Collazo-Clavell ML, Rule A et al. Hyperoxaluric nephrolithiasis is a complication of Roux-en-Y gastric bypass surgery. Kidney Int 2007; 72: 100–107. Asplin JR, Coe FL. Hyperoxaluria in kidney stone formers treated with modern bariatric surgery. J Urol 2007; 177: 565–569. Duffey BG, Alanee S, Pedro RN et al. Hyperoxaluria is a long-term consequence of Roux-en-Y Gastric bypass: a 2-year prospective longitudinal study. J Am Coll Surg 2010; 211: 8–15. Duffey BG, Pedro RN, Makhlouf A et al. Roux-en-Y gastric bypass is associated with early increased risk factors for development of calcium oxalate nephrolithiasis. J Am Coll Surg 2008; 206: 1145–1153. Penniston KL, Kaplon DM, Gould JC et al. Gastric band placement for obesity is not associated with increased urinary risk of urolithiasis compared to bypass. J Urol 2009; 182: 2340–2346. Semins MJ, Asplin JR, Steele K et al. The effect of restrictive bariatric surgery on urinary stone risk factors. Urology 2010; 76: 826–829. Chen T, Godebu E, Horgan S et al. The effect of restrictive bariatric surgery on urolithiasis. J Endourol 2013; 27: 242–244. Melton LJ 3rd. History of the Rochester Epidemiology Project. Mayo Clin Proc 1996; 71: 266–274. Rule AD, Lieske JC, Li X et al. The ROKS nomogram for predicting a second symptomatic stone episode. J Am Soc Nephrol 2014 (e-pub ahead of print). Matlaga BR, Shore AD, Magnuson T et al. Effect of gastric bypass surgery on kidney stone disease. J Urol 2009; 181: 2573–2577. Semins MJ, Matlaga BR, Shore AD et al. The effect of gastric banding on kidney stone disease. Urology 2009; 74: 746–749. Lieske JC, Kumar R, Collazo-Clavell ML. Nephrolithiasis after bariatric surgery for obesity. Semin Nephrol 2008; 28: 163–173. Puzziferri N, Blankenship J, Wolfe BM. Surgical treatment of obesity. Endocrine 2006; 29: 11–19. McLeod RS, Churchill DN. Urolithiasis complicating inflammatory bowel disease. J Urol 1992; 148: 974–978. Andersson H, Bosaeus I. Hyperoxaluria in malabsorptive states. Urol Int 1981; 36: 1–9. Modigliani R, Labayle D, Aymes C et al. Evidence for excessive absorption of oxalate by the colon in enteric hyperoxaluria. Scand J Gastroenterol 1978; 13: 187–192. Stauffer JQ. Hyperoxaluria and intestinal disease. The role of steatorrhea and dietary calcium in regulating intestinal oxalate absorption. Am J Dig Dis 1977; 22: 921–928. Kidney International (2015) 87, 839–845 clinical investigation JC Lieske et al.: Kidney stones after bariatric surgery 22. 23. 24. 25. Earnest DL, Johnson G, Williams HE et al. Hyperoxaluria in patients with ileal resection: an abnormality in dietary oxalate absorption. Gastroenterology 1974; 66: 1114–1122. Ocon Breton J, Perez Naranjo S, Gimeno Laborda S et al. Effectiveness and complications of bariatric surgery in the treatment of morbid obesity. Nutr Hosp 2005; 20: 409–414. Siener R, Petzold J, Bitterlich N et al. Determinants of urolithiasis in patients with intestinal fat malabsorption. Urology 2013; 81: 17–24. Nasr SH, D’Agati VD, Said SM et al. Oxalate nephropathy complicating Roux-en-Y Gastric Bypass: an underrecognized cause of irreversible renal failure. Clin J Am Soc Nephrol 2008; 3: 1676–1683. Kidney International (2015) 87, 839–845 26. 27. 28. 29. Nelson WK, Houghton SG, Milliner DS et al. Enteric hyperoxaluria, nephrolithiasis, and oxalate nephropathy: potentially serious and unappreciated complications of Roux-en-Y gastric bypass. Surg Obes Relat Dis 2005; 1: 481–485. Bergstralh EJ, Kosanke JL, Jacobsen SJ. Software for optimal matching in observational studies. Epidemiology 1996; 7: 331–332. Lieske JC, Callazo-Clavell ML, Sarr MG et al. Gastric bypass surgery and measured and estimated GFR in women. Am J Kid Dis 2014; 64: 663–665. Rule AD, Bergstralh EJ, Melton LJ 3rd et al. Kidney stones and the risk for chronic kidney disease. Clin J Am Soc Nephrol 2009; 4: 804–811. 845