1 2 3 4 5 6 7 8 9 10 11 12 13 14 Q1 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55

Surgery for Obesity and Related Diseases ] (2015) 00–00

Original article

Preoperative glucocorticoid use and risk of postoperative bleeding and infection after gastric bypass surgery for the treatment of obesity Sigrid Bjerge Gribsholt, M.D.a,b,*, Elisabeth Svensson, Ph.D., M.Sc.b, Reimar Wernich Thomsen, Ph.D., M.D.b, Bjørn Richelsen, D.M.Sc., M.D.a, Henrik Toft Sørensen, D.M.Sc., Ph.D., M.D.b a

Department of Endocrinology and Internal Medicine, Aarhus University Hospital b Department of Clinical Epidemiology, Aarhus University Hospital Received November 28, 2014; accepted January 26, 2015

Abstract

Background:: Previous research suggests that patients using glucocorticoids may be at increased risk of postoperative bleeding and infection after major surgery. The objective was to investigate the association between preoperative glucocorticoid use and risk of bleeding and infection after Rouxen-Y gastric bypass surgery (RYGB). Setting: Nationwide cohort study of 13,195 patients, who underwent RYGB 2006–2012 using Danish population-based medical databases. Methods:: Information was obtained on current (redeemed prescription o60 d before surgery), recent (prescription 60–180 d before surgery), or no glucocorticoid use, and postoperative bleeding or infection within 30 days of surgery. We computed risk differences and odds ratios (ORs) as a measure of relative risk with 95% confidence intervals (95% CIs) for the association between glucocorticoid use and bleeding or infection, adjusting for gender, age, and co-morbidities by logistic regression. Results:: Among RYGB patients, 325 (2.5%) were current glucocorticoid users, and 365 (2.8%) were recent users. The risk of bleeding was increased in current users: 2.8% versus 1.6% among nonusers (risk difference: 1.2%, 95% CI: -.6, 3.0) corresponding to an adjusted OR of 1.5 (95% CI: .8, 3.0). For recent users, the adjusted OR for bleeding was 1.2 (95% CI: .5, 2.5). The risk of infection did not differ materially between current (1.8%), recent (1.0%) and nonusers (1.7%), corresponding to an adjusted OR of .9 (95% CI: .4, 2.1) among current versus nonusers. Conclusions:: Current use of glucocorticoids is associated with a slightly increased risk of postoperative bleeding, but not infection, after RYGB. No increased risks were found for recent users. (Surg Obes Relat Dis 2015;]:00–00.) r 2015 American Society for Metabolic and Bariatric Surgery. All rights reserved.

Keywords:

Bariatric surgery; Bleeding; Infection; Glucocorticoids; Cohort study

Synthetic glucocorticoids are potent antiinflammatory drugs widely used to attenuate the general immune response [1,2]. The antiinflammatory effect is utilized to treat several * Correspondence: Sigrid Bjerge Gribsholt, MD, PhD fellow, Department of Endocrinology and Internal Medicine, Aarhus University Hospital, Tage-Hansens Gade 2-4, 8000 Aarhus C, Denmark. E-mail: [email protected]

diseases, including chronic pulmonary diseases and rheumatic diseases. For many patients discontinuation is not feasible due to likely relapse of symptoms. In Denmark, 2% of the general population between 20 and 39 years and 4% of the population between 40 and 64 years were prescribed systemic glucocorticoids at least once during 2009 [3]. Glucocorticoid use has been associated with substantial postoperative complications including delayed wound

http://dx.doi.org/10.1016/j.soard.2015.01.017 1550-7289/r 2015 American Society for Metabolic and Bariatric Surgery. All rights reserved.

56 57 58 59 60 61 62 63 64 65 66 67

2

68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122

S.B. Gribsholt et al. / Surgery for Obesity and Related Diseases ] (2015) 00–00

healing, upper gastrointestinal bleeding, and increased risk of infections [1,4–6]. These complications are important due to their adverse effect on early morbidity, length of hospital stay, and mortality [1]. A registry-based study of 635,265 American patients undergoing a variety of operations reported higher risk of postoperative infections among glucocorticoid users compared with nonusers (odds ratios 1.7–2.5 ) [6]. A Danish study of 19,919 women undergoing breast cancer surgery found an increased risk of reoperations due to postoperative bleeding among current glucocorticoid users (reoperation risk: 4% among patients with redeemed prescription o90 d before surgery) compared with nonusers (reoperation risk: 2.5%) and former users 90þ days before (reoperation risk: 2.6%) [7], suggesting a rather acute effect of glucocorticoid use. For patients with morbid obesity, bariatric surgery, particularly Roux-en-Y gastric bypass (RYGB) surgery, is an effective weight loss treatment [8,9] but is also associated with postoperative complications [10–12]. Two recent registry-based observational studies and 1 randomized trial have reported absolute risks of 1.7%–2.9% for bleeding and .3%–2.9% for infection within 30 days after RYGB [10–12]. To our knowledge no previous studies have investigated the postoperative risk of bleeding or infection in patients undergoing RYGB who use glucocorticoids preoperatively. Since glucocorticoid use is frequent among young and middle-aged individuals and RYGB procedures have increased in recent years, any increased risk of postoperative bleeding or infections may have important clinical implications. We therefore investigated the association between glucocorticoid use and the risk of postoperative bleeding and infections after RYGB, examining both current and recent use.

Danish National Patient Registry (DNPR). The DNPR records information on nonpsychiatric inpatient hospitalizations and hospital outpatient clinic visits at all Danish hospitals [15]. All contacts are coded by treating physicians, using the International Classification of Diseases, tenth edition (ICD-10), since 1995. The registry compiles data on admission and discharge dates, diagnoses, surgical procedures, and municipality of residence [15]. It allows for 1 primary diagnosis code (condition that prompted patient admission and treatment course) and up to 20 secondary codes. We excluded patients with fewer than 30 days of follow-up (n ¼ 24) and foreign residents (n ¼ 37). Use of glucocorticoids Data on use of glucocorticoids were obtained from the Danish National Health Service Prescription Database (DNHSPD), a national database established by the Danish Regions together with the Department of Clinical Epidemiology, Aarhus University Hospital [16]. It contains information on prescription redemption dates and type of medication dispensed according to the Anatomical Therapeutic Chemical (ATC) Classification System (Appendix B) [16]. Drug exposures examined in our study were oral and intestinal-acting glucocorticoids. Redeemed prescriptions for these drugs were identified for all patients 180 days before RYGB. Patients were classified as nonusers if they had no redeemed prescription 180 days before the surgery date, as current glucocorticoid users if they had one or more redeemed prescriptions o60 days before the surgery date, and as recent glucocorticoid users if they had redeemed prescription for glucocorticoids 60–180 days before the surgery date. Postoperative bleeding and infection

Methods Setting The Danish National Health Care System provides the entire population of Denmark (5.6 million people) with taxsupported healthcare, including RYGB surgery, and with partial reimbursement for most prescribed drugs, including glucocorticoids. All residents are assigned a unique personal identification number either at birth or upon immigration, which allows unambiguous individual-level record linkage of the healthcare registries used in this study [13,14]. Participants This nationwide study included all patients undergoing first-time RYGB surgery (99% were laparoscopic operations) (Appendix A) in Denmark between January 1, 2006 and November 30, 2012 (N ¼ 13,195). Patients were identified by surgical procedure codes for RYGB in the

Patients were followed for 30 days after surgery. Data on primary or secondary diagnosis codes for postoperative gastrointestinal bleeding and infections were obtained from the DNPR during and after the index admission (Appendix C). The patients were categorized as having “postoperative bleeding”/“no postoperative bleeding” and/or as having “postoperative infection”/“no postoperative infection.” Mortality The Danish Civil Registration System (CRS) contains information on gender, age, civil personal registration number, date of birth, and date of death (if any) for each person. The register is complete from 1968 on [14]. Covariates We used the DNPR to summarize each patient’s medical history during the 2 years before RYGB, focusing on diagnoses included in the Charlson Comorbidity Index

123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177

Glucocorticoid Use and Complications After RYGB / Surgery for Obesity and Related Diseases ] (2015) 00–00

178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232

(CCI) [17]. Patients were grouped according to overall level of co-morbidity, defined as a modified CCI score of 0 (low), 1–2 (moderate) and 3þ (severe) (Appendix C). Chronic pulmonary diseases (i.e., asthma, chronic obstructive pulmonary disease [COPD], and others) and connective tissue diseases (i.e., rheumatoid arthritis, spondylitis, and others), which constitute the 2 most important medical indications for glucocorticoid treatment [18], were considered as the joint variable “glucocorticoid-associated diseases.” Statistical analysis We tabulated co-morbidity and demographic variables according to glucocorticoid use (current, recent, no use). In crude analyses, we assessed all covariates (gender [male, female], age [o34 yr, 35–44 yr, 45þ yr], CCI score [0, 1– 2, 3þ], and glucocorticoid-associated diseases [yes, no]) as potential confounders using logistic regression analysis. We calculated risk differences (RDs) and odds ratios (ORs) as a measure of relative risk, and associated 95% confidence intervals (95% CI) comparing risk of bleeding or infection associated with current use and recent use versus no glucocorticoid use. To assess the effect of the individual confounders, we performed 3 stepwise cumulative adjustments: the first model adjusted for gender, the second adjusted for gender and glucocorticoid-associated diseases, and the third adjusted for gender, glucocorticoid-associated diseases, and age. Statistical analyses were performed using STATA 13.0 (StataCorp LP, College Station, TX, USA). The study was approved by the Danish Data Protection Agency (record number 2012-41-0793).

3

Results Characteristics of the study cohort We identified 13,202 patients who fulfilled our inclusion and exclusion criteria. Within 30 days of follow-up 7 patients died. They were all nonusers and were excluded from the further analyses. Of the remaining 13,195 patients, 325 (2.5%) were current glucocorticoid users and 365 (2.8%) were recent users. There were more women among current users (83%) and recent users (84%) than among nonusers (77%). Current glucocorticoid users were older: 48% were in the oldest age group (45þ yr), compared with 41% of recent users and 33% of nonusers. Current glucocorticoid users also had substantially more co-morbidity: 44% had a CCI score higher than 0, mainly due to a high frequency of glucocorticoid-associated diseases (33%). CCI scores higher than 0 were less frequent among recent users (35%) and nonusers (21%). Glucocorticoid-associated diseases were found in 34% of the recent users and in 7% of nonusers (Table 1).

Postoperative bleeding after RYGB The risk of postoperative bleeding was 2.8% (95% CI: 1.3, 5.2) among current users, 1.9% (95% CI: .8, 3.9) among recent users, and 1.6% (95% CI: 1.4, 1.8) among nonusers (Table 2). Thus the RD between current and nonusers was 1.2% (95% CI: -.6, 3.0). Compared with nonusers, current users had increased odds of postoperative bleeding after RYGB (crude OR ¼ 1.8; 95% CI: .9, 3.4). The OR decreased to 1.5 (95% CI: .8, 3.0) after adjusting

Table 1 Baseline characteristics of the 13,195 gastric bypass–operated patients, according to glucocorticoid use No glucocorticoid use (n ¼ 12505) Gender Female 9587 (76.7%) Male 2918 (23.3%) Age o34 3736 (29.9%) 35–44 4680 (37.4%) 45þ 4089 (32.7%) Year of operation 2006–07 1202 (9.6%) 2008–09 4136 (33.1%) 2010 4035 (32.3%) 2011–12 3132 (25.1%) Charlson Comorbidity Index 0 9912 (79.3%) 1–2 2357 (18.9%) 3þ 236 (1.9%) Diagnoses from the CCI associated with glucocorticoid treatment Chronic pulmonary disease 640 (5.1%) Connective tissue disease 218 (1.7%) Any major glucocorticoid-associated disease 839 (6.7%) RYGB ¼ Roux-en-Y gastric bypass; CCI ¼ Charlson Comorbidity Index.

Recent users (n ¼ 325)

Current users (n ¼ 365)

All RYGB-operated (n ¼ 13195)

307 (84.1%) 58 (15.9%)

268 (82.5%) 57 (17.5%)

10162 (77.0%) 3033 (23.0%)

80 (21.9%) 137 (37.5%) 148 (40.6%)

65 (20.0%) 103 (31.7%) 157 (48.3%)

3881 (29.4%) 4920 (37.3%) 4394 (33.3%)

37 (10.1%) 139 (38.1%) 108 (29.6%) 81 (22.2%)

43 (13.2%) 107 (32.9%) 97 (29.9%) 78 (24.0%)

1282 4382 4240 3291

236 (64.7%) 120 (32.9%) 9 (2.5%)

181 (55.7%) 123 (37.9%) 21 (6.5%)

10329 (78.3%) 2600 (19.7%) 266 (2.0%)

74 (20.3%) 20 (5.5%) 86 (23.6%)

102 (31.4%) 12 (3.7%) 108 (33.2%)

821 (6.2%) 250 (1.9%) 1033 (7.8%)

(9.7%) (33.2%) (32.1%) (24.9%)

233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 T1252 253 254 255 256 257 258 259 T2 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287

S.B. Gribsholt et al. / Surgery for Obesity and Related Diseases ] (2015) 00–00

4

288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342

Table 2 Odds ratios and 95% confidence intervals (CI) for the association between glucocorticoid use and 30-d postoperative bleeding in 9904 patients undergoing gastric bypass surgery n (%) No glucocorticoid use 12,505 (94.8) Recent glucocorticoid 365 (2.8) users * Current 325 (2.5) glucocorticoid user †

No bleeding Postoperative Absolute risk Risk difference (%) bleeding (%) % (95% CI) (95% CI)

Unadjusted Adjusted OR Adjusted OR Adjusted OR OR (95% CI) (95% CI) ‡ (95% CI) § (95% CI) ¶

12,306 (98.4) 358 (98.1)

199 (1.6)

1.6 (1.4, 1.8) Referent

Referent

Referent

Referent

Referent

7 (1.9)

1.9 (.8, 3.9) .3 (-.1, 1.7)

1.2 (.6, 2.5)

1.3 (.6, 2.7)

1.2 (.6, 2.6)

1.2 (.5, 2.5)

316 (97.3)

9 (2.8)

2.8 (1.3, 5.2) 1.2 (-.6, 3.0)

1.8 (.9, 3.4)

1.8 (.9, 3.6)

1.6 (.8, 3.1)

1.5 (.8, 3.0)

CI ¼ confidence interval; OR ¼ odds ratio. Redeemed prescription 60–180 d before surgery. † Redeemed prescription o60 d before surgery. ‡ Adjusted for gender. § Adjusted for gender and glucocorticoid-associated diseases. ¶ Adjusted for gender, glucocorticoid-associated diseases, and age. *

for gender, glucocorticoid-associated diseases and age (Model 3). Diseases appeared to be the strongest confounder of the association between glucocorticoids and bleeding. The crude OR for recent glucocorticoid users compared to nonusers was 1.2 (95% CI: .6, 2.5) and the adjusted OR (Model 3) was 1.2 (95% CI: .5, 2.5). All the estimates had poor precision. Infection after RYGB

T3

Risk of infection did not differ between current (1.8%, 95% CI: .7, 4.0), recent (1.0%, 95% CI: .3, 2.7) and nonusers (1.7%, 95% CI: 1.5, 2.0) (see Table 3). The crude OR of infection among current users compared with nonusers was 1.1 (95% CI: .5, 2.4). After adjusting for age, gender, and glucocorticoid-associated diseases (model 3), the OR decreased to .9 (95% CI .4, 2.1) (Table 3). The

crude OR for recent users compared with nonusers was .6 (95% CI: .2, 1.7). The OR was .6 (95% CI: .2, 1.6) after full adjustment (Model 3). All the estimates had poor precision.

Discussion In this population-based nationwide study, 2.5% of patients undergoing RYGB had redeemed a prescription for glucocorticoids within 60 days before their operation. These patients had an increased risk of bleeding, although the absolute risk increase was low, and they had no increased risk of infection compared to nonusers. In contrast, recent users of glucocorticoids had no materially increased risk of either bleeding or infection compared with nonusers, suggesting no protracted glucocorticoid effect on adverse outcomes.

Table 3 Odds ratios and 95% confidence intervals (CI) for the association between glucocorticoid use and 30-d postoperative infection in 9904 patients undergoing gastric bypass surgery n (%) No glucocorticoid use Recent glucocorticoid users * Current glucocorticoid user †

No infection Postoperative (%) infection (%)

Absolute risk Risk difference Unadjusted Adjusted OR Adjusted OR Adjusted OR (95% CI) § (95% CI) ¶ % (95% CI) (95% CI) OR (95% CI) (95% CI) ‡

12,505 12,287 (98.3) 218 (1.7) (94.8) 365 (2.8) 361 (98.9) 4 (1.1)

1.7 (1.5, 2.0) Referent

Referent

Referent

Referent

Referent

1.0 (.3, 2.7) -.06 (-1.7, .4)

.6 (.2, 1.7)

.6 (.2, 1.7)

.6 (.2, 1.6)

.6 (.2, 1.6)

325 (2.5) 319 (98.2)

1.8 (.7, 4.0) .1 (-1.3, 1.6)

1.1 (.5, 2.4)

1.1 (.5, 2.4)

1.0 (.5, 2.4)

.9 (.4, 2.1)

6 (1.9)

CI ¼ confidence interval; OR ¼ odds ratio. * Redeemed prescription 60–180 d before surgery. † Redeemed prescription o60 d before surgery. ‡ Adjusted for gender. § Adjusted for gender and glucocorticoid-associated diseases. ¶ Adjusted for gender, glucocorticoid-associated diseases, and age.

343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397

Glucocorticoid Use and Complications After RYGB / Surgery for Obesity and Related Diseases ] (2015) 00–00

398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452

Our study corroborates and extends knowledge about the increased risk of postoperative complications in patients currently using glucocorticoids. Our finding of a modest association between current glucocorticoid use and an increased risk of bleeding in RYGB patients accords with the observations of Winther Lietzen et al. They reported that current but not recent use of glucocorticoids increased the risk of reoperation due to bleeding among elderly women after mastectomy [7], possibly due to poorer wound healing associated with glucocorticoid use. This finding is supported by our finding of a RD of 1.2% of bleeding among current users compared with nonusers within 30 days after RYGB. Previous studies have found a bleeding risk after RYGB between 1.7% and 2.3% [10,11]. These studies contained no information on the correlation between medical treatment and postoperative bleeding, hindering comparisons with our results. In a Danish study of 34,641 colorectal cancer patients, Ostenfeldt et al. found an increased mortality risk after surgery among current glucocorticoid users (mortality risk ratio = 1.28 versus nonusers), but did not find an increased risk among recent users (mortality risk ratio = .92 versus nonusers) [1]. This corroborates our findings that the adverse effects of glucocorticoid use may be rapidly reversible and diminish 60–90 days after discontinuation. Our study did not show any substantially increased risk of infections among glucocorticoid users. This is in contrast with the results of Ismael et al. [6], who found a 2-fold increased risk of postoperative infections among glucocorticoid users without adjusting for co-morbidity [6]. The patients in that study were older than patients undergoing RYGB and they underwent surgery for different reasons. Moreover, no adjustments were performed considering comorbidity, whereas in our analysis the differences between the groups diminished after adjusting for glucocorticoidassociated diseases. The main strengths of the present study include its population-based design, encompassing all patients undergoing RYGB in Denmark during the study period, and a large sample size although the absolute number of adverse outcomes was still relatively low. Our cohort was welldefined due to the use of comprehensive population-based healthcare databases with complete follow-up, reducing selection problems seen in other studies [14]. In addition, accurate and independently recorded data on prescriptions and diagnoses eliminated recall bias and allowed adjustment for potential confounders [15,16]. This study also has limitations, and several issues should be considered when interpreting our results. The potential for misclassification of glucocorticoid use is a concern, as the DNDRP contains no information on actual drug intake and adherence. Some patients may have received glucocorticoids to use “as needed” should they develop disease symptoms. Since we did not have information on medical indications for the prescriptions we do not know whether the patients were

5

acutely or chronically treated. However, several studies have shown good accordance between self-reported drug use or general practitioner-reported use and filled prescriptions in Denmark [19–21]. Patients filling prescriptions more than 180 days before surgery who used glucocorticoids intermittently might have been misclassified as nonusers. We also lacked data on glucocorticoids dispensed in hospitals. However, it seems unlikely that patients would have an elective and invasive operation performed shortly after receiving intravenous or intramuscular glucocorticoids due to serious illness. Moreover, any misclassification of drug exposure would bias any outcome differences between the exposure groups toward the null, and not explain the associations we observed. Inaccurate coding of postoperative complications may have influenced the observed prevalence of bleeding and infection. However, such misclassification is likely unrelated to glucocorticoid exposure and would thus also bias the estimates toward the null [22]. Only bleeding episodes severe enough to lead to hospital referral by the patients’ general practitioners were included in our study. We lacked direct clinical data on bleeding severity. It is possible that part of the association between glucocorticoid use and bleeding-related hospitalizations was due to increased severity of bleeding caused by glucocorticoids, rather than increased bleeding risk per se [23]. Our nonrandomized study may be vulnerable to uncontrolled and unmeasured confounding [22], particularly confounding by indication for glucocorticoid therapy. We reduced this risk by adjusting for the most frequent glucocorticoid-associated diseases. Nonetheless, almost 70% of the glucocorticoid users had no glucocorticoidassociated diseases recorded in the hospital registries. These patients were likely treated solely by general practitioners, who do not record diagnoses in the DNPR. As observed by Tøttenborg, et al., patients with less severe stages of COPD are likely to be followed in general practice in Denmark, while patients with severe COPD are followed in hospital clinics [24], and residual confounding thus may have led to an overestimation of the observed associations in our study. Finally, we lacked data on lifestyle factors such as smoking. Importantly, however, the risk of bleeding was not materially elevated among noncurrent yet recent users of systemic glucocorticoids, who are likely to have underlying diseases and lifestyles relatively similar to those of current users. In conclusion, in this large population-based study, current use of glucocorticoid treatment was associated with an increased risk of postoperative bleeding and possibly also an increased risk of postoperative infections, while no material associations were observed for recent use. Thus, discontinuation of glucocorticoid treatment 60–90 days before RYGB may reduce these postoperative complications. The elevated risk must be weighed against the potential beneficial effect of the glucocorticoid treatment [22]. The absolute risk of complications is low, but for patients for whom discontinuation is impossible, increased

453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507

6

508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 Q3 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557

S.B. Gribsholt et al. / Surgery for Obesity and Related Diseases ] (2015) 00–00

surveillance in the early postoperative period may be beneficial. Disclosures None of the authors declare any conflict of interest. Funding sources: Grant from the Central Denmark Region and the Novo Nordisk Foundation The study was supported by the Program for Clinical Research Infrastructure (PROCRIN) established by the Lundbeck Foundation and the Novo Nordisk Foundation. Acknowledgments The authors thank Stine Skovbo, M.Sc. and Dóra Körmendiné Farkas, M.Sc., Department of Clinical Epidemiology, Aarhus University Hospital, for help with data extraction and management. Appendix A Supplementary data Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j. soard.2015.01.017. References [1] Ostenfeld EB, Erichsen R, Thorlacius-Ussing O, Riis AH, Sorensen HT. Pre-admission use of glucocorticoids and 30-day mortality following colorectal cancer surgery: a population-based Danish cohort study. Aliment Pharmacol Ther 2014;39(8):843–53. [2] Johannesdottir SA, Horvath-Puho E, Dekkers OM, et al. Use of glucocorticoids and risk of venous thromboembolism: a nationwide population-based case-control study. JAMA Intern Med 2013;173 (9):743–52. [3] Medstat.dk [homepage on the internet] SSI, the Danish Ministry of Health. [updated 2014 June 7, cited July 2014] Available from: http:// www.medstat.dk/. [4] Nielsen GL, Sorensen HT, Mellemkjoer L, et al. Risk of hospitalization resulting from upper gastrointestinal bleeding among patients taking corticosteroids: a register-based cohort study. Am J Med 2001;111(7):541–5. [5] Durmus M, Karaaslan E, Ozturk E, et al. The effects of single-dose dexamethasone on wound healing in rats. Anesth Analg 2003;97 (5):1377–80. [6] Ismael H, Horst M, Farooq M, Jordon J, Patton JH, Rubinfeld IS. Adverse effects of preoperative steroid use on surgical outcomes. Am J Surg 2011;201(3):305–8. (discussion 308–9). [7] Winther Lietzen L, Cronin-Fenton D, Garne JP, Kroman N, Silliman R, Lash TL. Predictors of re-operation due to post-surgical bleeding in breast cancer patients: a Danish population-based cohort study. Eur J Surg Oncol 2012;38(5):407–12. [8] Sjostrom L. Review of the key results from the Swedish Obese Subjects (SOS) trial—a prospective controlled intervention study of bariatric surgery. J Intern Med 2013;273(3):219–34.

[9] Courcoulas AP, Christian NJ, Belle SH, et al. Weight change and 558 health outcomes at 3 years after bariatric surgery among individuals 559 with severe obesity. JAMA 2013;310(22):2416–25. 560 [10] Carlin AM, Zeni TM, English WJ, et al. The comparative effectiveness of sleeve gastrectomy, gastric bypass, and adjustable gastric 561 banding procedures for the treatment of morbid obesity. Ann Surg 562 2013;257(5):791–7. 563 [11] Hernandez-Boussard T, Downey JR, McDonald K, Morton JM. 564 Relationship between patient safety and hospital surgical volume. 565 Health Serv Res 2012;47(2):756–69. [12] Schauer PR, Bhatt DL, Kirwan JP, et al. Bariatric surgery versus 566 intensive medical therapy for diabetes–3-year outcomes. N Engl J 567 568 Med 2014;370(21):2002–13. [13] Frank L. Epidemiology. When an entire country is a cohort. Science 569 2000;287(5462):2398–9. 570 [14] Schmidt M, Pedersen L, Sorensen HT. The Danish Civil Registration 571 System as a tool in epidemiology. Eur J Epidemiol 2014;29(8):541–9. [15] Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient 572 573 Register. Scand J Public Health 2011;39(7 Suppl):30–3. [16] Johannesdottir SA, Horvath-Puho E, Ehrenstein V, Schmidt M, 574 Pedersen L, Sorensen HT. Existing data sources for clinical epidemi- 575 ology: The Danish National Database of Reimbursed Prescriptions. 576 Clin Epidemiol 2012;4:303–13. [17] Thygesen SK, Christiansen CF, Christensen S, Lash TL, Sorensen 577 HT. The predictive value of ICD-10 diagnostic coding used to assess 578 Charlson comorbidity index conditions in the population-based 579 Danish National Registry of Patients. BMC Med Res Methodol 580 2011;11:83. 581 [18] Truven Health Analytics Inc. Prednisolone. Indications. [database on the internet] Micromedex 2.0. 2014 [updated YYYY Month DD, Q4582 Accessed 2014 October 20] Available from: http://www.microme- 583 dexsolutions.com.ez.statsbiblioteket.dk:2048/micromedex2/librarian/ 584 ND_T/evidencexpert/ND_PR/evidencexpert/CS/2DCD3E/ND_App585 Product/evidencexpert/DUPLICATIONSHIELDSYNC/04446C/ 586 ND_PG/evidencexpert/ND_B/evidencexpert/ND_P/evidencexpert/ 587 PFActionId/evidencexpert.DisplayDrugpointDocument?doc588 Id¼478490&contentSetId¼100&title¼Prednisolone&servicesTitle¼Prednisolo589 ne&topicId¼dosingAndIndicationsSection&subtopicId¼fdaSection. 590 [19] Wogelius P, Poulsen S, Sorensen HT. Validity of parental-reported 591 questionnaire data on Danish children's use of asthma-drugs: a comparison with a population-based prescription database. Eur J 592 593 Epidemiol 2005;20(1):17–22. [20] Lokkegaard EL, Johnsen SP, Heitmann BL, et al. The validity of self- 594 reported use of hormone replacement therapy among Danish nurses. 595 Acta Obstet Gynecol Scand 2004;83(5):476–81. 596 [21] Johannesdottir SA, Maegbaek ML, Hansen JG, Lash TL, Pedersen L, Ehrenstein V. Correspondence between general practitioner-reported 597 medication use and timing of prescription dispensation. Clin Epide- 598 miol 2012;4:13–8. 599 [22] Sorensen HT, Lash TL, Rothman KJ. Beyond randomized controlled 600 trials: a critical comparison of trials with nonrandomized studies. 601 Hepatology 2006;44(5):1075–82. [23] Olsen M, Christensen S, Riis A, Thomsen RW. Preadmission use of 602 systemic glucocorticoids and 30-day mortality following bleeding 603 peptic ulcer: a population-based cohort study. Am J Ther 2010;17;17 604 (1). (23–9. 24). 605 [24] Tottenborg SS, Thomsen RW, Nielsen H, Johnsen SP, Frausing 606 Hansen E, Lange P. Improving quality of care among COPD outpatients in Denmark 2008–2011. Clin Respir J 2013;7(4):319–27. 607