Incidence, Risk Factors, and Clinical Effects of Recurrent Diverticular Hemorrhage: A Large Cohort Study

Gastroenterology 2018;-:1–12

Incidence, Risk Factors, and Clinical Effects of Recurrent Diverticular Hemorrhage: A Large Cohort Study Q6

Ravy K. Vajravelu,1,2 Ronac Mamtani,2,3 Frank I. Scott,2,4 Adam Waxman,2,3 and James D. Lewis1,2 1

Division of Gastroenterology, Department of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania; 2Center for Clinical Epidemiology and Biostatistics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania; 3 Abramson Cancer Center, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania; and 4Division of Gastroenterology, Department of Medicine, University of Colorado Denver School of Medicine, Aurora, Colorado

BACKGROUND & AIMS: Although recurrent diverticular hemorrhage is common, its incidence and risk factors have not been measured outside of small institutional cohorts. We analyzed the incidence of and risk factors for recurrent diverticular hemorrhage and whether discontinuing anticoagulation after diverticular hemorrhage is associated with ischemic stroke. METHODS: We performed a retrospective cohort study of patients enrolled in the OptumInsight Clinformatics database from 2000 through 2016. Incidence rates for initial and recurrent diverticular hemorrhage were calculated by identifying patients who had hospitalizations with a primary discharge diagnosis consistent with diverticular hemorrhage. The hazard ratios of second diverticular hemorrhage associated with anticoagulants or platelet aggregation inhibitors were calculated using Cox proportional hazards regression adjusted for demographics, comorbidities, and medication use. The hazard ratio for ischemic stroke among patients who discontinued anticoagulation after diverticular hemorrhage was calculated similarly. RESULTS: In the cohort analyzed, 14,925 patients had an initial diverticular hemorrhage; 1368 of these patients had a second episode. The unstandardized incidence rates of initial and second diverticular hemorrhage were 10.9 per 100,000 person-years (95% confidence interval [CI] 10.7–11.0) and 3625.6 per 100,000 person-years (95% CI 3436.0–3823.0). Platelet aggregation inhibitors were associated with second episodes

of diverticular hemorrhage (hazard ratio 1.47; 95% CI 1.15–1.88), whereas all classes of anticoagulation agents were not associated. Among patients with a potential indication for stroke prophylaxis, those who discontinued anticoagulation after the diverticular hemorrhage had an increased hazard of ischemic stroke (hazard ratio 1.93; 95% CI 1.17–3.19). CONCLUSIONS: In this retrospective cohort study, platelet aggregation inhibitors, but not anticoagulants, were associated with recurrent diverticular hemorrhage. Discontinuing anticoagulation was associated with increased hazard for ischemic stroke.

Keywords: Diverticular Disease; Antithrombotic Agents; Medical Decision Making; Pharmacoepidemiology.

Abbreviations used in this paper: CI, confidence interval; DOAC, directacting oral anticoagulant; HR, hazard ratio; ICD, International Classification of Diseases; IQR, interquartile range; IR, incidence rate; NSAID, nonselective nonsteroidal anti-inflammatory drug; Optum, OptumInsight Clinformatics; PAIs, platelet aggregation inhibitors; SSRIs, selective serotonin reuptake inhibitors. © 2018 by the AGA Institute 0016-5085/$36.00 https://doi.org/10.1053/j.gastro.2018.07.026

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce

CLINICAL AT

1 2 3 4 5 6 7 8 9 10 11 12 13 14 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 56 57 58 59 60

61 62 63 64 65 66 67 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

2

Gastroenterology Vol.

WHAT YOU NEED TO KNOW BACKGROUND AND CONTENT The incidence rate of recurrent diverticular hemorrhage has not been measured using population-level data. Additionally, the risk of re-bleeding is unknown among patients who resume anticoagulants or platelet aggregation inhibitors (PAIs). NEW FINDINGS CLINICAL AT

121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180

Vajravelu et al

The incidence rate of second diverticular hemorrhage was 300 times that of initial diverticular hemorrhage. PAIs, but not anticoagulants, were associated with recurrent hemorrhage, and discontinuation of anticoagulants was associated with increased risk of ischemic stroke. LIMITATIONS The risk factors for diverticular hemorrhage may have been confounded by over-the-counter aspirin and NSAID use, but a sensitivity analysis showed that this was unlikely to change the conclusions. IMPACT Despite high risk of recurrent hemorrhage, stopping anticoagulation after diverticular hemorrhage may not be justified as it appears to increase the risk of stroke without decreasing the risk of re-bleeding.

C

olonic diverticular hemorrhage is the most common cause of gastrointestinal bleeding distal to the stomach.1 Diverticular hemorrhage develops when the artery at the base of the diverticulum ruptures due to erosion of the vessel wall.2 Overall, diverticular hemorrhage is common, with a prevalence rate of 23.9 to 32.5 per 100,000 persons1,3 based on cross-sectional analyses of hospital-level discharge data in the United States. Patients with an initial episode of diverticular hemorrhage are presumed to be at increased risk for recurrent episodes of hemorrhage, but the incidence of a second episode has not been systematically evaluated. Current estimates of the proportion of patients who experience recurrent hemorrhage are derived from institutional cohorts and vary widely from 13.8% to 47.4%.4–12 Several risk factors, including aspirin use, nonsteroidal antiinflammatory drug (NSAID) use,13–17 obesity,18 hypertension,17 and atherosclerosis,19 have been associated with initial diverticular hemorrhage. There are limited data on the effect of platelet aggregation inhibitors (PAIs)8 and anticoagulants17 on diverticular hemorrhage, but they are thought to increase risk, like they do in other causes of gastrointestinal bleeding.20 A limitation of the existing literature is that it is difficult to isolate the effect of the identified risk factors for diverticular hemorrhage from their association with the development of diverticulosis. Disentangling this relationship is challenging because diverticular hemorrhage cannot occur without preexisting diverticulosis, diverticulosis is clinically silent until complications develop, and the relationship between risk factors, such as atherosclerosis and prescription drug use, with diverticulosis is confounded by aging.21

-,

No.

-

Some studies have attempted to overcome these challenges by matching patients with diverticular hemorrhage to controls with silent diverticulosis, but given that they are derived from institutional cohorts, they have not had enough power to make definitive conclusions about the range of potential risk factors.17 It is important to better understand the risk factors of diverticular hemorrhage because hospitalization for diverticular hemorrhage is associated with decompensation of comorbid conditions and increased 30-day mortality.22 For example, given the significant burden of cardiovascular comorbidities among patients with diverticular hemorrhage,17,22 precisely calculating the relative hazard of PAIs and anticoagulant medications for recurrent diverticular hemorrhage is necessary to inform decision making about whether the increased risk of hemorrhage outweighs the risk of thromboembolism.23,24 To address this question, we conducted a cohort study among patients with a history of a first diverticular hemorrhage using OptumInsight Clinformatics (Optum), a patient-level claims database from a national health insurer in the United States. Using this cohort of patients with known diverticulosis, we determined the incidence of recurrent diverticular hemorrhage, estimated the relative hazard of PAIs and anticoagulants for recurrent diverticular hemorrhage, and assessed the risk of ischemic stroke among patients who discontinue anticoagulation secondary to initial diverticular hemorrhage.

Methods Study Design and Data Source The Optum database was used for this retrospective cohort study. Optum is a patient-level medical claims database consisting of the inpatient, outpatient, pharmacy, procedure, and laboratory claims of approximately 77.2 million unique patients. Data from Optum have previously been used to study acute conditions25 and medication use.26,27 Optum contains inpatient diagnoses and procedures from 2004 to 2016 in addition to outpatient diagnoses, procedures, and prescriptions from 2000 to 2016. For this study, records after September 30, 2016, were excluded due to the possibility of incomplete claims at the time of data download (July 2017). Optum contains claims from patients coinsured with Medicaid from 2000 to 2010 and Medicare from 2000 to 2005. For encounters before October 1, 2015, diagnoses are classified using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes, and ICD-10 codes are used thereafter. Research using Optum has been classified as exempt by the Institutional Review Board of the University of Pennsylvania.

Study Population Patients who had experienced an initial episode of diverticular hemorrhage during follow-up in the database were included in the cohort if they all of met the following criteria: A. ICD diagnosis for diverticulosis of colon with hemorrhage as primary inpatient hospitalization discharge diagnosis at least 120 days after entry into the database. (ICD-9-CM code 562.12: diverticulosis of colon with hemorrhage, or ICD-10 code K57.31: diverticulosis of the

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce

181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240

241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300

2018

Recurrent Diverticular Hemorrhage

large intestine without perforation or abscess with bleeding). B. No ICD diagnosis for diverticulosis of colon with hemorrhage in any inpatient or outpatient claim in any diagnostic position before date of service in criterion A. C. No ICD diagnosis for inflammatory bowel disease or colorectal cancer before or within 2 years of the date of initial diverticular hemorrhage. These diagnoses were excluded to reduce the probability of including cases of lower gastrointestinal bleeding that may have been misclassified as diverticular hemorrhage (ICD codes listed in Supplementary Table 1). D. At least 1 year of follow-up in Optum without any gaps in enrollment. Diverticular hemorrhage is often a presumed clinical diagnosis because active diverticular bleeding is rarely witnessed during endoscopy. Thus, there is no gold standard for validation studies. To reduce the possibility of misclassification of the outcome of diverticular hemorrhage, we excluded ICD-9-CM code 562.13 (diverticulitis of colon with hemorrhage) and the corresponding ICD-10 code K57.33 (diverticulitis of the large intestine without perforation or abscess with bleeding) from the main analysis. ICD-9-CM code 562.12 (diverticulosis of the colon with hemorrhage) has previously been shown to have 97% positive predictive value for lower gastrointestinal hemorrhage when it is the primary hospital diagnosis.28 The minimum 120-day interval between enrollment and initial diverticular hemorrhage was incorporated into the inclusion criteria to provide an opportunity for a patient’s preexisting comorbidities at the time of enrollment in the database to be recorded.29

3

inhibitors (SSRIs) was recorded as a confounding variable.30–32 We used sensitivity analyses (described later in this article) to assess the impact of over-the-counter aspirin and NSAID use.

Outcome Definition The main study outcome was second episode of diverticular hemorrhage, identified by an ICD diagnosis code28 for diverticulosis of the colon with hemorrhage in the primary position during an inpatient hospitalization at least 30 days after the initial diverticular hemorrhage hospitalization. The 30-day interval was specified to identify distinct episodes of diverticular hemorrhage as opposed to identifying recrudescence of the initial episode. A sensitivity analysis was performed to assess the impact of this definition on the estimated incidence rates (IRs) (described later in this article).

Other Covariates Demographic covariates collected included patient age, sex, and race (documented in Optum as White, Black, Hispanic, Asian, and unknown). Diagnoses of interest were selected due to their association with diverticular hemorrhage or with medication exposures and were identified by the presence of at least one ICD diagnosis code in any position in any inpatient or outpatient claim. The full list of comorbidities is presented in the Supplementary Methods and Supplementary Table 1. All covariates were incorporated into the Cox proportional hazards model as time-updating variables. Age updated yearly, and diagnoses updated unidirectionally daily. New comorbid diagnoses were recorded so long as they occurred at least 1 day before the episode of recurrent diverticular hemorrhage. See Supplementary Figure 1 for a directed acyclic graph illustrating the relationships among the study exposures, outcome, and covariates.

Exposure Definition

Statistical Analysis

The primary exposures of interest were prescriptions for anticoagulants and PAIs. Anticoagulant use was recorded as a categorical variable with levels of (1) no anticoagulation, (2) direct-acting oral anticoagulants (apixaban, dabigatran, and rivaroxaban), (3) warfarin, and (4) subcutaneous heparin derivatives (heparin, dalteparin, and enoxaparin). The 3 adenosine diphosphate receptor inhibitors widely available in the United States (clopidogrel, prasugrel, and ticagrelor) were included in the PAI class. Exposures were identified using the pharmacy claims file. Prescription use at the time of initial diverticular hemorrhage was determined by presence of a prescription claim with days of medication supplied that overlapped with the admission date. Thereafter in follow-up, prescription status was incorporated into the analysis as a time-varying exposure that updated daily based on the days of medication supplied and refill status. For example, if a patient claimed a 30-day supply of a medication on March 1, he or she was considered exposed to the medication until March 31. If he or she next claimed a 30-day supply of the medication on April 3, he or she was considered unexposed on April 1 and April 2, but exposed from April 3 to May 3. If there were prescriptions for 2 anticoagulants dispensed on a single date, the medication recorded in the database second was excluded. This accounted for 576 instances of a total of 102,732 prescriptions (0.56%). Use of aspirin, NSAIDs, and selective serotonin reuptake

Stata version 15 (StataCorp, College Station, TX) was used for all statistical analyses. Continuous data are reported as medians with interquartile range (IQR). Categorical data are reported as counts and percentages. Results with fewer than 20 subjects are not reported to avoid re-identification. Rates of diverticular hemorrhage. For the IR of the initial episode of diverticular hemorrhage, the total follow-up time was calculated from the entire population of Optum that had at least 1-year follow-up without gaps in enrollment. IRs are primarily reported as unstandardized rates stratified by decade of life because the prevalence of diverticulosis increases with age and patients younger than 50 are at low risk for diverticular complications.21 IRs directly age- and sexstandardized to the 2010 US population33 are also presented for comparison with other studies. Unadjusted cumulative IRs for the second episode of diverticular hemorrhage were calculated using Kaplan-Meier survival analysis. To compare the rate of diverticular hemorrhage with the 2 prior studies that assessed the occurrence of diverticular hemorrhage using administrative hospital-level discharge data,1,3 prevalence rates were calculated using the total number of patients enrolled in Optum from January 1, 2000, to September 30, 2016. The prevalence rates are reported as both unstandardized rates and rates directly age- and sexstandardized to the 2010 US population.

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce

CLINICAL AT

-

301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360

4

Gastroenterology Vol.

Risk factors for recurrent diverticular hemorrhage. Adjusted hazard ratios (HRs) for risk factors for

CLINICAL AT

361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420

Vajravelu et al

recurrent diverticular hemorrhage were calculated using Cox proportional hazards regression. The model was adjusted for demographics and comorbid diagnoses using a disease risk score for diverticular hemorrhage to balance covariates among the exposure groups. Details of the disease risk score creation and its incorporation into the Cox model are presented in the Supplementary Methods. The Cox model was adjusted for use of aspirin, NSAIDs, SSRIs, PAIs, direct-acting oral anticoagulant (DOACs), warfarin, heparin, the type of anticoagulant at initial diverticular hemorrhage, and the disease risk score. The initial Cox model included an interaction term for anticoagulation type and chronic kidney disease, because chronic kidney disease may influence the pharmacologic effect of anticoagulants due to altered pharmacokinetics. The interaction term was not statistically significant by the Wald test, so it was removed from subsequent models. The Cox proportional hazards model was stratified with homogeneous covariate effects before and after the date of first DOAC approval by the Food and Drug Administration (October 19, 2010). Censoring occurred if the patient was no longer enrolled in the database or underwent colectomy (codes in Supplementary Table 2, Panel 1). As the goal of the analysis was to create an explanatory model adjusting for all potential confounders, covariates that were not statistically significant in the main model were not eliminated. Violations of the proportional hazards assumption were assessed using Schoenfeld residuals. The hazard of recurrent diverticular hemorrhage when using multiple medications simultaneously was assessed through linear combinations from the Stata lincom command.

Sensitivity Analyses To assess the robustness of the study results, several additional analyses were performed. First, because the 2 previously published studies that calculated the rates of diverticular hemorrhage using administrative data used different ICD-9-CM criteria for the definition of diverticular hemorrhage (562.12 only3 vs 562.12 and 562.131), IRs and risk factors for recurrent hemorrhage were also calculated in a sensitivity analysis where diverticular hemorrhage was identified by an expanded list of ICD codes (Supplementary Table 3). Second, to determine the impact of excluding episodes of recurrent diverticular hemorrhage that occurred within 30 days of the prior diverticular hemorrhage, we calculated IRs in which these episodes were included in the analysis. Third, to determine if anticoagulation or PAIs were associated with severe episodes of diverticular hemorrhage, an analysis in which the outcome was diverticular hemorrhage requiring a red blood cell transfusion during the recurrent hemorrhage was conducted (ICD codes for transfusion in Supplementary Table 4). In this analysis, requirement of blood transfusion was considered a proxy for severity of hemorrhage. Fourth, to reduce the impact of other causes of lower gastrointestinal hemorrhage misclassified as diverticular hemorrhage, an analysis in which diverticular hemorrhage with documented lower intestinal endoscopy was considered the outcome (ICD procedure codes for lower intestinal endoscopy in Supplementary Table 5). Finally, to assess whether the HR estimates for

-,

No.

-

anticoagulation were sensitive to misclassification of aspirin and NSAID use, simulation studies varying aspirin and NSAID exposure were performed (detailed in Supplementary Methods).

Thrombotic Sequelae of Anticoagulation Discontinuation To assess the clinical impact of discontinuing anticoagulation after the initial episode of diverticular hemorrhage, an analysis assessing the hazard of ischemic stroke among patients who were taking anticoagulation and had a potential indication for stroke prophylaxis at the time of initial diverticular hemorrhage was performed. The cohort was formed from the group of patients who were identified using the expanded diverticular hemorrhage ICD list. Indications for stroke prophylaxis were defined as ICD codes for atrial fibrillation, atrial flutter, hypercoagulable state, and heart valve replacement (ICD codes in Supplementary Table 1). The outcome was a diagnosis of ischemic stroke in the primary diagnostic position (ICD-9 codes 434.x and ICD-10 codes I63.x) during an inpatient encounter. The validity of these codes has been previously described.34 This cohort was selected to study the impact of discontinuing anticoagulation because its members often have strong and persistent indications for anticoagulation.35–37 Indications for anticoagulation, such as deep venous thrombosis, pulmonary embolism, and postsurgical venous thromboembolism prophylaxis were not considered because the duration of therapy is variable and difficult to ascertain in administrative data. Unlike the main analysis of recurrent diverticular bleeding, anticoagulation status was recorded as exposure to any anticoagulant. Patients were categorized into the following 2 groups to describe the sequence of their anticoagulant use before and after the initial episode of diverticular hemorrhage: anticoagulated to none (ie, discontinued) and anticoagulated to anticoagulated (ie, continued). Anticoagulation group was a time-varying exposure that updated daily based on the date of anticoagulation prescription and the days of medication dispensed. Cox proportional hazards regression was used to compare the HRs of patients who discontinued vs continued anticoagulation, adjusted for aspirin use, NSAID use, age, and CHADS2-VASc (Congestive heart failure, Hypertension, Age [
65 ¼ 1 point,
75 ¼ 2 points], Diabetes, and Stroke/transient ischemic attack [2 points]–Vascular disease) score.38 Exposure to PAIs was not used for adjustment because no ischemic strokes occurred during exposure to PAIs. Patients were censored on second diverticular hemorrhage or leaving the database. We also calculated the HR of recurrent diverticular hemorrhage for use of any anticoagulant among this cohort to compare this risk to the risk of ischemic stroke from discontinuing anticoagulation. To help ensure that the results of the analysis were not due to residual confounding, a falsification test was performed to assess the hazard of discontinuing anticoagulation with new diagnosis of osteoarthritis after the initial diverticular bleeding episode. There is no known association between anticoagulation and osteoarthritis.

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce

421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480

481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540

2018

Recurrent Diverticular Hemorrhage

5

CLINICAL AT

-

Figure 1. Study creation.

cohort

Results Cohort Characteristics A total of 14,925 patients who met the inclusion criteria and experienced an initial episode of diverticular hemorrhage during follow-up in the database were identified (Figure 1). The characteristics of the cohort at the time of initial diverticular hemorrhage are presented in Table 1. The median age was 78 (IQR 70–82), and 50.4% were women. The median follow-up time was 2.0 years (IQR 0.9–4.0). The proportion of ICD codes used to identify the cohort are presented in Supplementary Tables 6 and 7. Medication refilling patterns among patients in the cohort are presented in Supplementary Table 8. Relative to the proportion of patients using each type of medication before the initial diverticular bleeding episode, the proportion using each medication after the episode was nearly unchanged (Supplementary Table 9). Of the 14,925 patients who experienced initial diverticular hemorrhage, 26 (0.2%) had arterial embolization during hospitalization and 310 (2.1%) underwent colectomy.

Rates of Diverticular Hemorrhage In Optum, the unstandardized and age- and sexstandardized IRs of initial diverticular hemorrhage were 10.9 (95% confidence interval [CI] 10.7–11.0) and 18.9 (95% CI 18.9–18.9) per 100,000 person-years respectively. The unstandardized IR of initial hemorrhage increased as the population aged, ranging from 0.3 (95% CI 0.2–0.3) per

100,000 person-years in those aged 30 to 39 to 49.6 (95% CI 48.4–50.9) per 100,000 person-years in those 80 and older (Table 2). To compare the rates of diverticular hemorrhage derived from Optum with those from the other studies that used administrative data,1,3 the prevalence rate of initial diverticular hemorrhage was also calculated. The unstandardized prevalence of initial diverticular hemorrhage was 19.6 (95% CI 19.2–19.9) per 100,000 persons and the age- and sex-standardized prevalence was 36.3 (95% CI 35.7–37.0) per 100,000 persons. Among those with an initial hemorrhage, the unstandardized IR for a second episode was 3625.6 (95% CI 3436.0–3823.0), and the unstandardized IR increased from 2614.4 (95% CI 2135.9–3200.1) per 100,000 person-years to 4792.1 (95% CI 4210.0–5454.8) per 100,000 personyears as age-group at initial hemorrhage increased from 50 to 59 years to 80 years and older (Table 2). The unstandardized IR of a third episode of diverticular hemorrhage was 10,848.3 (95% CI 9625.2–12,183.7) per 100,000 person-years. The IRs of each subsequent episode were progressively higher (Table 3).

Characteristics of Second Episode of Diverticular Hemorrhage Among patients experiencing an initial episode of diverticular hemorrhage, 1368 (9.2%) experienced a second episode. The median time between the first and second hemorrhage was 1.2 years (IQR 0.5–2.5). In the unadjusted Kaplan-Meier analysis, the cumulative incidence

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce

541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600

6

Gastroenterology Vol.

Table 1.Cohort Characteristics at Date of Initial Diverticular Hemorrhage (n ¼ 14,925) Median

IQR

78 n

70–82 %

7748 7177 14,925

51.9 48.1 100.0

9739 2385 1258 277 1266 14,925

65.2 16.0 8.4 1.9 8.5 100.0

13,286 237 1185 217 14,925

84.9 1.4 12.2 1.5 100.0

167 3184 1762 847

1.1 21.3 11.8 5.7

4095 733 2340 3520 198 4389 6131 447 158 12,914 126 175 6966 4072 535 706 2702 1110

27.4 4.9 15.7 23.6 1.3 29.4 41.1 3.0 1.1 86.5 0.8 1.2 46.7 27.3 3.6 4.7 18.1 7.4

-,

No.

-

8 (0.6%) underwent arterial embolization and 64 (4.7%) underwent colectomy.

Risk Factors for Second Diverticular Hemorrhage Age

CLINICAL AT

601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660

Vajravelu et al

Sex Female Male Total Race White Black Hispanic Asian Unknown Total Anticoagulation None DOACs Warfarin Heparin Total Prescriptions Aspirin NSAIDs SSRIs Platelet aggregation inhibitors Comorbidities Atrial fibrillation Atrial flutter Carotid artery stenosis Chronic kidney disease Cirrhosis Congestive heart failure Coronary artery disease Heart valve replacement Hypercoagulable state Hypertension Leukemia Lymphoma Osteoarthritis Peripheral artery disease Pulmonary embolism Rheumatoid arthritis Stroke Venous thromboembolism

of a second episode of diverticular hemorrhage was 4.7% (95% CI 4.3–5.0%) at 1 year. At 2 and 5 years, the cumulative incidences were 8.3% (95% CI 7.8%–8.9%) and 15.7% (95% CI 14.8%–16.6%) respectively (Figure 2). During the hospitalization for the second episode of diverticular hemorrhage, 707 (51.7%) received a transfusion. Of the 1368 patients who had a second episode of diverticular hemorrhage, 978 (71.5%) underwent colonoscopy or flexible sigmoidoscopy during the second hospitalization. Among the patients who had a second hemorrhage, 719 (52.5%) underwent lower endoscopy during the first and second hospitalizations and 232 (17.0%) underwent lower endoscopy during the second admission only. Of the 1368 patients who experienced a second diverticular hemorrhage,

The overall exposure time to anticoagulants was 522.9 person-years for DOACs, 3349.9 person-years for warfarin, 680.7 person-years for subcutaneous heparin, and 33,152.2 person-years for no anticoagulation. The overall exposure time to SSRIs and PAIs was 4733.2 and 1174.0 person-years respectively. PAI exposure was significantly associated with second diverticular hemorrhage (HR 1.47, 95% CI 1.15–1.88), whereas each category of anticoagulation was not (Table 4). In the analysis of simultaneous exposure to more than 1 class of medication, the HR of using PAIs with aspirin was 1.84 (95% CI 1.13–3.00). Other statistically significant combinations included PAIs with heparin (HR 1.88, 95% CI 1.08–3.25), PAIs with aspirin and warfarin (HR 1.73, 95% CI 1.02–3.02), and PAIs with aspirin and heparin (HR 2.34, 95% CI 1.17–4.70) (Supplementary Table 10).

Sensitivity Analyses In the sensitivity analysis using the expanded ICD code list to identify episodes of diverticular hemorrhage, 17,212 patients with initial diverticular hemorrhage were identified. The unstandardized IR was 12.6 (95% CI 12.4–12.8) per 100,000 person-years, compared with 10.9 (95% CI 10.7–11.0) per 100,000 person-years in the primary analysis. The age- and sex-adjusted IR was 22.0 (95% CI 22.0– 22.0) per 100,000 person-years, compared with 18.9 (95% CI 18.9–18.9) in the primary analysis (Supplementary Table 11, Panel 1). Statistically significant risk factors for recurrent diverticular hemorrhage were unchanged from the primary analysis (Supplementary Table 12). In the sensitivity analysis including recurrent episodes of diverticular hemorrhage that occurred within 30 days of prior episode, the unstandardized IR of a second diverticular hemorrhage was 4592.6 (95% CI 4377.2–4815.8) compared with 3625.6 (95% CI 3436.0–3823.0) in the main analysis. The IRs of subsequent diverticular hemorrhage episodes in this analysis are presented in Supplementary Table 11, Panel 2. In the unadjusted Kaplan-Meier analysis, the cumulative incidence of a second episode of diverticular hemorrhage at 1, 2, and 5 years was 7.0% (95% CI 6.6%–7.5%), 10.6% (95% CI 10.0%–11.1%), and 17.8% (95% CI 16.8%–18.7%), respectively. In the analysis considering second diverticular hemorrhage with blood transfusion as the outcome, only PAI therapy (HR 1.45, 95% CI 1.02–2.04) was associated (Supplementary Table 13). In the sensitivity analysis considering diverticular hemorrhage with endoscopy as the outcome, again, only PAI therapy (HR 1.49, 95% CI 1.06–2.11) was associated (Supplementary Table 14). These results were consistent with the primary analysis. In the sensitivity analyses assessing the impact of overthe-counter aspirin and NSAID use, if the exposure time to aspirin among only the patients who experienced diverticular bleeding increased by 20% (44.5 to 53.5 person-years),

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce

661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720

-

Recurrent Diverticular Hemorrhage

7

Table 2.IRs of Initial and Second Episodes of Diverticular Hemorrhage per 100,000 Person-Years Stratified by Age Group at Time of Initial Episode of Diverticular Hemorrhage Initial diverticular hemorrhage (n ¼ 38.5 million, 142.1 million person-years)

Second diverticular hemorrhage (n ¼ 14,925, 37,705.8 person-years)

Age group

Events

IR

95% CI

Events

IR

95% CI

30–39 40–49 50–59 60–69 70–79 80 and older

37 267 978 2475 5382 6164

0.3 1.4 4.5 12.3 36.4 49.6

0.2–0.3 1.2–1.6 4.2–4.8 11.8–12.8 35.4–37.4 48.4–50.9

— — 94 316 711 229

— — 2614.4 3233.5 3932.9 4792.1

— — 2135.9–3200.1 2895.9–3610.4 3654.1–4232.8 4210–5454.8

NOTE. Cells with counts less than 20 are suppressed to avoid re-identification.

the HR of aspirin would change from nonsignificant to significant (HR 1.57, 95% CI 1.08–2.28). If the exposure time to aspirin among patients who experienced diverticular bleeding increased by 2546% (to 1177.5 personyears), the HR of PAIs would change from significant to nonsignificant (HR 1.28, 95% CI 0.99–1.64). For NSAIDs, if the exposure time among patients who experienced diverticular bleeding increased by 8% (from 616.0 to 668.1 person-years), the HR of NSAIDs would change from nonsignificant to significant (HR 1.17, 95% CI 1.03–1.32). If the NSAID exposure time among patients who experienced diverticular bleeding increased by 176% (to 1701.6 personyears), the HR of PAIs would change from significant to nonsignificant (HR 1.31, 95% CI 0.99–1.72). There was no level of differential misclassification of over-the-counter aspirin or NSAID use that changed the statistical significance of the HRs for DOACs, warfarin, or heparin (Supplementary Tables 15 and 16).

Thrombotic Sequelae of Anticoagulation Discontinuation Of the 17,212 patients with a first diverticular hemorrhage identified by the expanded ICD list, 1542 (9.2%) were prescribed anticoagulation and had a potential indication for ischemic stroke prophylaxis at the time of initial diverticular hemorrhage. Of these patients, 983 (63.7%)

continued or eventually resumed anticoagulation. The median time to an anticoagulant prescription refill was 40 days (IQR 19–85). Subsequent to the initial hemorrhage, ischemic stroke occurred in 64 (4.1%) patients. Among those patients who experienced an ischemic stroke, the median number of days from initial diverticular hemorrhage to stroke was 392 (IQR 105–679). These results are similar to previously reported annualized risks of stroke in patients with atrial fibrillation.39 Of the 64 patients who had an ischemic stroke, 35 (54.7%) were not on anticoagulation at the time of their stroke. Discontinuing anticoagulation was associated with increased relative hazard of ischemic stroke (HR 1.93, 95% CI 1.17–3.19; Table 5). Thirteen of these patients (37.1%) eventually restarted anticoagulation. The median number of days from the stroke to starting anticoagulation was 115 days (IQR 44–295). In the cohort of 1542 patients on anticoagulation at the time of initial diverticular hemorrhage with a potential indication for ischemic stroke prophylaxis, the relative hazard of recurrent diverticular hemorrhage was not elevated among those who continued any anticoagulant after the first hemorrhage (HR 0.98, 95% CI 0.79–1.22; Supplemental Table 12). In the falsification test, discontinuing anticoagulation was not associated with new diagnosis of osteoarthritis, indicating that the results of the ischemic stroke analysis were unlikely to be due to residual confounding.

Table 3.IRs of Diverticular Hemorrhage per 100,000 Person-Years Unstandardized

Standardized

Episode

Events

IR

95% CI

IR

95% CI

1 2 3 4 5

14,925 1368 285 76 28

10.9 3625.6 10,848.3 16,441.3 24,970.1

10.7–11.0 3436.0–3823.0 9625.2–12,183.7 12,953.8–20,578.7 16,592.4–36,088.7

18.9 109.1 314.0 486.2 718.4

18.9–18.9 103.4–115.1 278.6–352.6 383.1–608.6 477.4–1038.3

NOTE. The IR for each episode is conditional on having a prior episode of diverticular bleeding. For example, among patients who have previously experienced 2 episodes of diverticular bleeding, the unstandardized IR of a third episode is 10,848.3 per 100,000 person-years.

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce

CLINICAL AT

721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780

2018

781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840

8

Gastroenterology Vol.

-,

No.

-

Table 5.Thrombotic Sequelae of Discontinuing Anticoagulation After the Initial Episode of Diverticular Hemorrhage Among Patients With a Potential Indication for Ischemic Stroke Prophylaxis With Anticoagulation (n ¼ 1542) Ischemic stroke (64 events)

Discontinued anticoagulation Aspirin NSAIDs Age CHADS2-VASc

CLINICAL AT

841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900

Vajravelu et al

Figure 2. Unadjusted Kaplan-Meier survival curve for second episode of diverticular hemorrhage among patients with an initial episode of diverticular hemorrhage.

Discussion Recurrent diverticular hemorrhage is a common complication of colonic diverticulosis. In this study, the IR of a second episode of diverticular hemorrhage was estimated to be 3625.6 per 100,000 person-years, and the risk of recurrent hemorrhage increased with older age at initial diverticular hemorrhage. Additionally, PAIs were demonstrated to be a pharmacologic risk factor for recurrence of diverticular hemorrhage, but anticoagulant medications were not. Patients with a potential indication for anticoagulation to prevent ischemic stroke had increased relative hazard for ischemic stroke if they discontinued anticoagulation following diverticular hemorrhage. These nationally representative estimates of the incidence of and risk factors for recurrent diverticular bleeding may serve as an important step for developing management plans for patients with indications for PAI or anticoagulation use and a history of diverticular hemorrhage.

Table 4.Risk Factors for Second Diverticular Hemorrhage

Exposures of interest PAIs Anticoagulants DOACs Warfarin Heparins Confounding covariates Aspirin NSAIDs SSRIs Anticoagulants at first bleed DOACs Warfarin Heparins Disease risk score a

P < .05.

HR

95% CI

1.47

1.15–1.88a

0.68 0.95 1.28

0.33–1.41 0.74–1.22 0.78–2.08

1.25 1.02 0.93

0.82–1.91 0.90–1.16 0.78–1.10

0.23 0.95 1.08 1.07

0.05–0.96 0.76–1.20 0.60–1.95 1.05–1.08a

Osteoarthritis (166 events)

HR

95% CI

HR

95% CI

1.93

1.17–3.19a

1.32

0.97–1.81

1.37 0.83 1.03 1.30

0.19–9.96 0.44–1.60 0.99–1.07 1.09–1.57a

0.58 0.83 1.00 0.95

0.08–4.19 0.55–1.26 0.98–1.02 0.86–1.07

NOTE. The HRs for the risk of ischemic stroke/osteoarthritis are adjusted for aspirin use, NSAID use, age, and CHADS2VASc score. Estimates were not adjusted for platelet aggregation inhibitor use because no ischemic strokes occurred in 86.4 person-years of exposure. Osteoarthritis was a falsification endpoint to test for indications of residual confounding in the ischemic stroke analysis. CHADS2-VASc, Congestive heart failure, Hypertension, Age [
65 ¼ 1 point,
75 ¼ 2 points], Diabetes, and Stroke/ transient ischemic attack [2 points]–Vascular disease. a P < .05.

Given that several studies27,40,41 have demonstrated that anticoagulant medications are associated with initial allcause gastrointestinal hemorrhage, one might expect to observe a strong positive association among the classes of anticoagulants and recurrent diverticular hemorrhage. To our knowledge, prior studies have not systematically studied the association of diverticular hemorrhage and anticoagulant use. It is possible that the pathophysiology of diverticular hemorrhage may explain the lack of association between anticoagulants and recurrent diverticular hemorrhage in our study. Because diverticular hemorrhage involves arterial rupture, there is a mechanistic difference relative to other causes of gastrointestinal hemorrhage. For example, the mechanism of anticoagulant-related bleeding in upper gastrointestinal ulcers and intestinal malignancy is postulated to be related to exacerbation of bleeding from systemic and local anticoagulation at sites of existing mucosal injury.42,43 Because this mucosal injury does not exist in colonic diverticula,44 diverticular bleeding may not be as strongly related to anticoagulation as other causes of gastrointestinal hemorrhage. It is unknown whether the observed association with PAIs in this study is a result of medication-induced mucosal injury45 at the diverticular dome or neck that exposes the artery to the colonic lumen to cause eventual rupture, but this should be explored in future studies. The study results presented here are an initial step to guide the development of clinical pathways to manage diverticular bleeding. This study showed that anticoagulants were not associated with recurrent hemorrhage and that discontinuing anticoagulation for stroke prophylaxis

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce

901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960

961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020

2018

Recurrent Diverticular Hemorrhage

increased the relative hazard of ischemic stroke. These results are consistent with recent studies that demonstrated high rates of comorbid complications related to lower gastrointestinal bleeding,22 including recurrent thromboembolism23,24 and increased 90-day mortality when anticoagulation is discontinued at hospital discharge.46 Furthermore, this study demonstrated that after 2 prior episodes of diverticular hemorrhage, the annual risk of recurrent hemorrhage per patient was approximately 11%. Future studies should assess the impact of different surgical thresholds for colectomy for prevention of future diverticular hemorrhage on overall survival and quality of life.47

Strengths and Limitations Prior studies of diverticular hemorrhage using administrative data used hospital-level discharge data,1,3 thus preventing assessment of recurrent hemorrhage. Prior studies of recurrent diverticular hemorrhage through institutional cohorts have widely varying IR estimates, likely due to small sample sizes and loss to follow-up.4–10 This study used a patient-level administrative database to overcome the limitations of the prior literature through a large sample size of approximately 15,000 patients with longitudinal medical claims data. This allowed for granular assessment of several time-varying exposures. By focusing on patients who already had established diverticulosis, this study controls for confounding related to the development of diverticulosis. Other strengths of this study include the use of directed acyclic graphs and disease risk scores to adjust for confounding from several comorbidities associated with bleeding propensity and prescription for anticoagulants, PAIs, aspirin, and NSAIDs. However, there are potential caveats to consider when interpreting these study results. First, although a disease risk score was used to adjust for potential confounding from comorbidities and channeling, it is possible that there was unmeasured confounding. The most likely source of unmeasured confounding is misclassification of exposure to aspirin and NSAIDs from over-the-counter use. The potential magnitude of this confounding was assessed using simulation studies adjusting the amount of differential misclassification of aspirin and NSAID use. Although the main analysis did not demonstrate a statistically significant association of aspirin and NSAID use with recurrent diverticular bleeding, the sensitivity analyses showed that if we were able to account for small amounts of over-the-counter use, the HRs for these medications would reach traditional thresholds for statistical significance. This result is consistent with prior research demonstrating that aspirin and NSAID use is associated with diverticular hemorrhage.13–17 Similarly, when accounting for over-the-counter use as an unmeasured confounder, the hazard estimates for the classes of anticoagulant drugs were not meaningfully changed for all levels of over-the-counter aspirin or NSAID misclassification and the hazard estimates for PAIs were altered only at implausibly high levels of differential misclassification (the 1368 patients who ultimately had recurrent

9

hemorrhage would have to take 1133 person-years of overthe-counter aspirin whereas the 13,557 patients who did not have hemorrhage would have to take 0 person-years of over-the-counter aspirin). These results indicate that the HRs for anticoagulants and PAIs in the main analysis were not confounded by over-the-counter aspirin and NSAID use. Additionally, it is possible that there was misclassification of the study outcome, diverticular hemorrhage, in the claims data. To assess the impact of potential misclassification, we conducted several sensitivity analyses varying the definition of the outcome to consider hemorrhage requiring transfusion, hemorrhage requiring endoscopy, and diverticular hemorrhage identified by an expanded list of ICD codes. None of these analyses meaningfully changed the results. In clinical practice, diverticular hemorrhage is usually a presumed diagnosis when another source of bleeding is not identified and diverticulosis is present. Diverticulosis can be diagnosed with colonoscopy or radiographic tests. In this cohort, 71.5% of patients who experienced a recurrent hemorrhage underwent lower endoscopy, which is within the range of 50.3% to 80.8% reported from prior studies that used administrative data to study lower gastrointestinal bleeding.9,48 Why the other patients did not undergo lower endoscopy could not be ascertained, but could include the absence of severe bleeding and recent imaging confirming diverticulosis. Notably, the cumulative incidences of recurrent diverticular hemorrhage of 4.7% at 1 year, 8.3% at 2 years, and 15.7% at 5 years were at the lower end of the range of cumulative IRs that had been previously reported from institutional cohorts.4–10 The strict definition for identifying recurrent diverticular hemorrhage that excluded rehospitalizations for diverticular hemorrhage within 30 days of the initial hemorrhage may explain the lower cumulative IR in this study. This definition was used to identify new hemorrhage events rather than recrudescence of hemorrhage of an incompletely healed bleeding diverticulum. Additionally, this study demonstrated that the prevalence rate of initial diverticular hemorrhage was similar to rates identified in the prior studies of diverticular hemorrhage calculated using administrative data.1,3 This lends support to the reliability of the data from the Optum database. Finally, because the inpatient data from the Optum database uses ICD procedure codes instead of the more granular Current Procedural Terminology, we were unable to assess how endoscopic interventions such as hemoclip placement or electrocautery at the bleeding diverticulum alter the risk of recurrent hemorrhage. However, because the main analysis excluded recurrent hemorrhage episodes within 30 days of the initial hemorrhage and endoscopic intervention for an actively bleeding diverticulum is rare,49 we expect that this limitation did not profoundly alter the study results. In conclusion, in this study of patients who had previously experienced an initial episode of diverticular hemorrhage, the incidence per 100,000 person-years of a second

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce

CLINICAL AT

-

1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080

10

CLINICAL AT

1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140

Vajravelu et al

Gastroenterology Vol.

episode of diverticular hemorrhage was 3625.6 (95% CI 3436.0–3823.0). The IR increased with older age at initial hemorrhage and with more episodes of prior diverticular hemorrhage. PAI use was a risk factor for a second episode of diverticular hemorrhage, whereas anticoagulant use was not. Moreover, among patients with a potential indication for stroke prophylaxis, discontinuing anticoagulation after the initial episode of diverticular hemorrhage was associated with approximately twofold increased risk of ischemic stroke. These results may serve as a launching point for future clinical management algorithms for recurrent diverticular hemorrhage. The creation of such pathways is increasingly important, as the prevalence of diverticular hemorrhage is expected to increase as the global population ages.50

Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at https://doi.org/10.1053/ j.gastro.2018.07.026.

References 1. Laine L, Yang H, Chang SC, et al. Trends for incidence of hospitalization and death due to GI complications in the United States from 2001 to 2009. Am J Gastroenterol 2012;107:1190–1195; quiz 1196. 2. Meyers MA, Alonso DR, Gray GF, et al. Pathogenesis of bleeding colonic diverticulosis. Gastroenterology 1976; 71:577–583. 3. Wheat CL, Strate LL. Trends in hospitalization for diverticulitis and diverticular bleeding in the United States From 2000 to 2010. Clin Gastroenterol Hepatol 2016; 14:96–103.e1. 4. McGuire HH Jr, Haynes BW Jr. Massive hemorrhage for diverticulosis of the colon: guidelines for therapy based on bleeding patterns observed in fifty cases. Ann Surg 1972;175:847–855. 5. McGuire HH Jr. Bleeding colonic diverticula. A reappraisal of natural history and management. Ann Surg 1994;220:653–656. 6. Poncet G, Heluwaert F, Voirin D, et al. Natural history of acute colonic diverticular bleeding: a prospective study in 133 consecutive patients. Aliment Pharmacol Ther 2010;32:466–471. 7. Niikura R, Nagata N, Yamada A, et al. Recurrence of colonic diverticular bleeding and associated risk factors. Colorectal Dis 2012;14:302–305. 8. Aoki T, Nagata N, Niikura R, et al. Recurrence and mortality among patients hospitalized for acute lower gastrointestinal bleeding. Clin Gastroenterol Hepatol 2015;13:488–494.e1. 9. Longstreth GF. Epidemiology and outcome of patients hospitalized with acute lower gastrointestinal hemorrhage: a population-based study. Am J Gastroenterol 1997;92:419–424.

-,

No.

-

10. Aytac E, Stocchi L, Gorgun E, et al. Risk of recurrence and long-term outcomes after colonic diverticular bleeding. Int J Colorectal Dis 2014;29:373–378. 11. Taki M, Oshima T, Tozawa K, et al. Analysis of risk factors for colonic diverticular bleeding and recurrence. Medicine (Baltimore) 2017;96:e8090. 12. Joaquim N, Caldeira P, Antunes AG, et al. Risk factors for severity and recurrence of colonic diverticular bleeding. Rev Esp Enferm Dig 2017;109:3–9. 13. Strate LL, Liu YL, Huang ES, et al. Use of aspirin or nonsteroidal anti-inflammatory drugs increases risk for diverticulitis and diverticular bleeding. Gastroenterology 2011;140:1427–1433. 14. Yuhara H, Corley DA, Nakahara F, et al. Aspirin and nonaspirin NSAIDs increase risk of colonic diverticular bleeding: a systematic review and meta-analysis. J Gastroenterol 2014;49:992–1000. 15. Tsuruoka N, Iwakiri R, Hara M, et al. NSAIDs are a significant risk factor for colonic diverticular hemorrhage in elder patients: evaluation by a case-control study. J Gastroenterol Hepatol 2011;26:1047–1052. 16. Nagata N, Niikura R, Aoki T, et al. Impact of discontinuing non-steroidal antiinflammatory drugs on long-term recurrence in colonic diverticular bleeding. World J Gastroenterol 2015;21:1292–1298. 17. Yamada A, Sugimoto T, Kondo S, et al. Assessment of the risk factors for colonic diverticular hemorrhage. Dis Colon Rectum 2008;51:116–120. 18. Strate LL, Liu YL, Aldoori WH, et al. Obesity increases the risks of diverticulitis and diverticular bleeding. Gastroenterology 2009;136:115–122.e1. 19. Okamoto T, Watabe H, Yamada A, et al. The association between arteriosclerosis related diseases and diverticular bleeding. Int J Colorectal Dis 2012;27:1161–1166. 20. Holster IL, Valkhoff VE, Kuipers EJ, et al. New oral anticoagulants increase risk for gastrointestinal bleeding: a systematic review and meta-analysis. Gastroenterology 2013;145:105–112.e15. 21. Peery AF, Keku TO, Martin CF, et al. Distribution and characteristics of colonic diverticula in a United States screening population. Clin Gastroenterol Hepatol 2016; 14:980–985.e1. 22. Rubin JN, Shoag D, Gaetano JN, et al. Risk factors for 30-day hospital readmission for diverticular hemorrhage [published online ahead of print July 21, 2017]. J Clin Gastroenterol https://doi.org/10.1097/MCG. 0000000000000883. 23. Witt DM, Delate T, Garcia DA, et al. Risk of thromboembolism, recurrent hemorrhage, and death after warfarin therapy interruption for gastrointestinal tract bleeding. Arch Intern Med 2012;172:1484–1491. 24. Sengupta N, Feuerstein JD, Patwardhan VR, et al. The risks of thromboembolism vs. recurrent gastrointestinal bleeding after interruption of systemic anticoagulation in hospitalized inpatients with gastrointestinal bleeding: a prospective study. Am J Gastroenterol 2015; 110:328–335. 25. Ma GK, Brensinger CM, Wu Q, et al. Increasing incidence of multiply recurrent Clostridium difficile infection in the

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce

1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200

1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231Q4 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260

2018

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

Recurrent Diverticular Hemorrhage

United States: a cohort study. Ann Intern Med 2017; 167:152–158. Wunsch H, Wijeysundera DN, Passarella MA, et al. Opioids prescribed after low-risk surgical procedures in the United States, 2004–2012. JAMA 2016;315:1654–1657. Abraham NS, Noseworthy PA, Yao X, et al. Gastrointestinal safety of direct oral anticoagulants: a large population-based study. Gastroenterology 2017; 152:1014–1022.e1. Siddique J, Ruhnke GW, Flores A, et al. Applying classification trees to hospital administrative data to identify patients with lower gastrointestinal bleeding. PLoS One 2015;10. e0138987. Lewis JD, Bilker WB, Weinstein RB, et al. The relationship between time since registration and measured incidence rates in the General Practice Research Database. Pharmacoepidemiol Drug Saf 2005;14:443–451. de Abajo FJ, Garcia Rodriguez LA, Montero D. Association between selective serotonin reuptake inhibitors and upper gastrointestinal bleeding. BMJ 1999;319:1106–1109. Masclee GM, Valkhoff VE, Coloma PM, et al. Risk of upper gastrointestinal bleeding from different drug combinations. Gastroenterology 2014;147:784–792.e9; quiz e13–14. Cheng YL, Hu HY, Lin XH, et al. Use of SSRI, but not SNRI, increased upper and lower gastrointestinal bleeding: a nationwide population-based cohort study in Taiwan. Medicine (Baltimore) 2015;94:e2022. United States Census Bureau. Age and sex composition: 2010. 2010 Census Briefs. US Department of Commerce - Econoics and Statistics Administration, 2011. Jones SA, Gottesman RF, Shahar E, et al. Validity of hospital discharge diagnosis codes for stroke: the Atherosclerosis Risk in Communities Study. Stroke 2014; 45:3219–3225. Chiang CE, Naditch-Brule L, Murin J, et al. Distribution and risk profile of paroxysmal, persistent, and permanent atrial fibrillation in routine clinical practice: insight from the real-life global survey evaluating patients with atrial fibrillation international registry. Circ Arrhythm Electrophysiol 2012;5:632–639. Hart RG, Kanter MC. Hematologic disorders and ́ ́ ischemic stroke. A selective review. Stroke 1990; 21:1111–1121. Nishimura RA, Otto CM, Bonow RO, et al. 2017 AHA/ ACC focused update of the 2014 AHA/ACC guideline for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. J Am Coll Cardiol 2017;70:252–289. Coppens M, Eikelboom JW, Hart RG, et al. The CHA2DS2-VASc score identifies those patients with atrial fibrillation and a CHADS2 score of 1 who are unlikely to benefit from oral anticoagulant therapy. Eur Heart J 2013;34:170–176. Gage BF, Waterman AD, Shannon W, et al. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA 2001;285:2864–2870.

11

40. Miller CS, Dorreen A, Martel M, et al. Risk of gastrointestinal bleeding in patients taking non-vitamin k antagonist oral anticoagulants: a systematic review and meta-analysis. Clin Gastroenterol Hepatol 2017; 15:1674–1683.e3. 41. Coleman CI, Sobieraj DM, Winkler S, et al. Effect of pharmacological therapies for stroke prevention on major gastrointestinal bleeding in patients with atrial fibrillation. Int J Clin Pract 2012;66:53–63. 42. Vaduganathan M, Bhatt DL. Gastrointestinal bleeding with oral anticoagulation: understanding the scope of the problem. Clin Gastroenterol Hepatol 2017;15: 691–693. 43. Flack KF, Desai J, Kolb JM, et al. Major gastrointestinal bleeding often is caused by occult malignancy in patients receiving warfarin or dabigatran to prevent stroke and systemic embolism from atrial fibrillation. Clin Gastroenterol Hepatol 2017;15:682–690. 44. Wedel T, Barrenschee M, Lange C, et al. Morphologic basis for developing diverticular disease, diverticulitis, and diverticular bleeding. Viszeralmedizin 2015;31: 76–82. 45. Uotani T, Sugimoto M, Nishino M, et al. Ability of rabeprazole to prevent gastric mucosal damage from clopidogrel and low doses of aspirin depends on CYP2C19 genotype. Clin Gastroenterol Hepatol 2012;10:879–885. e2. 46. Patel P, Nigam N, Sengupta N. Lower gastrointestinal bleeding in patients with coronary artery disease on antithrombotics and subsequent mortality risk. J Gastroenterol Hepatol 2018;33:1185–1191. 47. Chen CY, Wu CC, Jao SW, et al. Colonic diverticular bleeding with comorbid diseases may need elective colectomy. J Gastrointest Surg 2009;13:516–520. 48. Strate LL, Ayanian JZ, Kotler G, et al. Risk factors for mortality in lower intestinal bleeding. Clin Gastroenterol Hepatol 2008;6:1004–1010; quiz 955-. 49. Jensen DM, Machicado GA, Jutabha R, et al. Urgent colonoscopy for the diagnosis and treatment of severe diverticular hemorrhage. N Engl J Med 2000; 342:78–82. 50. Lutz W, Sanderson W, Scherbov S. The coming acceleration of global population ageing. Nature 2008; 451:716–719.

Received May 29, 2018. Accepted July 22, 2018. Reprint requests Address requests for reprints to: Ravy K. Vajravelu, MD, MSCE, University of Pennsylvania, Division of Gastroenterology, 3400 Civic Center Boulevard, Perelman Center for Advanced Medicine–South Pavilion 7th floor, Philadelphia, Pennsylvania 19104–6021. e-mail: Q1 [email protected]; fax: 215–349–5915. Acknowledgments Author contributions: Ravy K. Vajravelu and James D. Lewis had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Ravy K. Vajravelu and James D. Lewis. Acquisition, analysis, and interpretation of data: All authors. Drafting of the manuscript: Ravy K. Vajravelu and James D. Lewis. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: Ravy K. Vajravelu and James D. Lewis. Obtained funding: Frank I. Scott, Ronac Mamtani, James D. Lewis. Study supervision: James D. Lewis.

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce

CLINICAL AT

-

1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320

12

CLINICAL AT

1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380

Vajravelu et al

Gastroenterology Vol.

Conflicts of interest These authors disclose the following: Ronac Mamtani has served as a consultant for Genentech-Roche. The scope of work was unrelated to this research. Frank I. Scott has served as a consultant to Janssen (manufacturer of rivaroxaban) and Merck. The scope of work was unrelated to this research. He has received research funding from Takeda, also unrelated to this research. James D. Lewis has served as a consultant to AbbVie, Amgen, Bridge Biotherapeutics, Celgene, Gilead, Janssen (manufacturer of rivaroxaban), Johnson & Johnson Consumer Inc, Lilly (manufacturer of prasugrel), Merck, Nestle Health Science, Pfizer (manufacturer of dalteparin), Samsung Bioepis, Takeda, and UCB. The scope of work for each of these

-,

No.

-

was unrelated to this research. He has received research funding from Nestle Health Science and Takeda, also unrelated to this research. The Q2 remaining authors disclose no conflicts.

1381 1382 1383 Funding Grant support: Ravy K. Vajravelu: National Institutes of Health (NIH)/National 1384 Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) T321385 DK007066; Ronac Mamtani: NIH/National Cancer Institute K23-CA187185; 1386 Frank I. Scott: NIH/NIDDK K08-DK095951; Adam Waxman: NIH/National Institute of General Medical Sciences T32-GM075766; and James D. Lewis: 1387 Q3 Q5 NIH/NIDDK K24-DK078228. 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440

FLA 5.5.0 DTD
YGAST62010_proof
8 October 2018
9:38 am
ce