Incidence of Pancreatitis andPancreatic Cancer in aRandomized ControlledMulticenter Trial (SAVOR-TIMI53) of the Dipeptidyl Peptidase-4Inhibitor SaxagliptinDiabetes Care 2014;37:2435–2441 | DOI: 10.2337/dc13-2546

OBJECTIVE

To determine the incidence of pancreatitis and pancreatic cancer in the SAVOR-TIMI 53 trial.

RESEARCH DESIGN AND METHODS

A total of 16,492 type 2 diabetic patients ‡40 years old with established cardio-vascular (CV) disease or CV risk factors were randomized to saxagliptin or placeboand followed for 2.1 years. Outcome measures were investigator reported withblinded expert adjudication of total pancreatitis (acute and chronic) and reportedcases of pancreatic cancer.

RESULTS

Trial investigators reported 35 events of pancreatitis in each treatment arm in 63patients (33 [0.40%] in the saxagliptin arm and 30 [0.37%] in control arm), with ahazard ratio (HR) of 1.09 (95% CI 0.66–1.79, P = 0.80). Adjudication confirmedpancreatitis in 24 patients (26 events) in the saxagliptin arm (0.29%) and 21patients (25 events) in placebo arm (0.26%), with an HR of 1.13 (0.63–2.06, P =0.77). Cases of definite acute pancreatitis were confirmed in 17 (0.2%) vs. 9(0.1%) (HR 1.88 [0.86–4.41], P = 0.17), definite plus possible pancreatitis in 22vs. 16 (HR 1.36 [0.72–2.64], P = 0.42), and chronic pancreatitis in 2 vs. 6 (HR 0.33[0.05–1.44], P = 0.18) in the saxagliptin and placebo arms, respectively. No differ-ences in time to event onset, concomitant risk factors for pancreatitis, investigator-reported causality from study medication or disease severity, and outcomewere found between treatment arms. The investigators reported 5 and 12cases of pancreatic cancer in the saxagliptin and placebo arms, respectively(HR 0.42 [0.13–1.12], P = 0.09).

CONCLUSIONS

In the SAVOR-TIMI 53 trial, within 2.1 years of follow-up, risk for pancreatitis intype 2 diabetic patients treated with saxagliptin was low and apparently similar toplacebo, with no sign of increased risk for pancreatic cancer. Further studies areneeded to completely resolve the pancreatic safety issues with incretin-basedtherapy.

1DiabetesUnit, Department ofMedicine, HadassahHebrew University Hospital, Jerusalem, Israel2TIMI Study Group, Cardiovascular Division,Brigham and Women’s Hospital and HarvardMedical School, Boston, MA3AstraZeneca Research and Development,Wilmington, DE4AstraZeneca Research and Development,Molndal, Sweden5Bristol-Myers Squibb, Princeton, NJ6University Medicine Greifswald, Greifswald,Germany

Corresponding author: Itamar Raz, ntv502@netvision.net.il.

Received 31 October 2013 and accepted 8 May2014.

Clinical trial reg. no. NCT01107886, clinicaltrials.gov.

This article contains Supplementary Data onlineat http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc13-2546/-/DC1.

A slide set summarizing this article is availableonline.

© 2014 by the American Diabetes Association.Readers may use this article as long as the workis properly cited, the use is educational and notfor profit, and the work is not altered.

Itamar Raz,1 Deepak L. Bhatt,2

Boaz Hirshberg,3 Ofri Mosenzon,1

Benjamin M. Scirica,2

Amarachi Umez-Eronini,2 KyungAh Im,2

Christina Stahre,4 Alona Buskila,1

Nayyar Iqbal,5 Norton Greenberger,2 and

Markus M. Lerch6

Diabetes Care Volume 37, September 2014 2435

CLIN

CARE/ED

UCATIO

N/N

UTR

ITION/PSYC

HOSO

CIAL

Several clinical features associated withtype 2 diabetes and obesity are recog-nized risk factors for acute pancreatitis(1). Recently, concerns have been raisedregarding long-term consequences ofincretin therapy in type 2 diabetic pa-tients. These concerns include possibletriggers for acute pancreatitis and initi-ation or acceleration of histologicalchanges, suggesting silent or clinicalchronic pancreatitis with preneoplasticlesions and potentially pancreatic can-cer (2,3). Although some reports havesupported the linkage of type 2 diabetestreatment with dipeptidyl peptidase-4(DPP-4) inhibitors or GLP-1 receptor ag-onists and increased risk of pancreatitis(4,5), others have not supported thisassociation (6–9).The U.S. Food and Drug Administra-

tion requires the investigation of cardio-vascular (CV) safety of new drugs fordiabetes (10,11). Therefore, several ran-domized controlled CV safety studieswith incretin-based therapies are ongo-ing or have recently been completed(12–15). These studies provide an op-portunity to further evaluate additionalsafety events of interest, including ef-fects on the pancreas.

RESEARCH DESIGN AND METHODS

The Saxagliptin Assessment of VascularOutcomes Recorded in Patients with Di-abetes Mellitus–Thrombolysis in Myo-cardial Infarction 53 (SAVOR-TIMI 53)trial evaluated the long-term CV safetyand efficacy of saxagliptin in patientswith type 2 diabetes at high risk for CVevents (16). The TIMI Study Group andHadassah Medical Organization de-signed the trial in collaboration withthe sponsors. SAVOR-TIMI 53 wasa randomized, double-blind, placebo-controlled, multicenter internationalintervention trial conducted betweenMay 2010 and June 2013 in 788 medicalcenters in 26 countries. Patients withtype 2 diabetes .40 years of age withestablished CV disease or$55 (men) and$60 (women) years of age with one ormore CV risk factors were randomized1:1 to receive 5 mg/day saxagliptin orplacebo. In patients with moderate orsevere renal failure, the saxagliptindose was decreased to 2.5 mg/day. Theprotocol was approved by the ethicscommittees of all participating centers,and all patients gave written informedconsent before enrollment. The full

eligibility criteria have been previouslypublished (16). In short, eligible patientshad a documented history of type 2 di-abetes, had an HbA1c between 6.5% and12.0% (48–108 mmol/mol) during the 6months before enrollment in the study,and were taking any antidiabetic drugexcept incretin-based therapy. To qualifyfor the established CV disease criteria,study subjects had to be at least 40 yearsold and have a history of a clinical eventsecondary to atherosclerosis involvingthe coronary, cerebrovascular, or pe-ripheral vascular system. To qualifybased on multiple risk factor criteria,study subjects had to be at least 55(men) or 60 (women) years old andhave at least one of the following addi-tional risk factors: dyslipidemia, hyper-tension, or active smoking. Patientswere ineligible if they were currently orpreviously (within 6 months) treatedwith an incretin-based therapy or hadend-stage renal disease (long-term dial-ysis, renal transplant, or serum creati-nine level .6.0 mg/dL). The baselinecharacteristics of the SAVOR-TIMI 53 par-ticipants were previously reported (17).

Patients were followed at the studysites by aminimumof oneoffice visit every6 months with either an additional phonecall or office visit every 3 months. The pa-tients also continued their usual medicalcare with their respective treating physi-cians, and the sites were instructed tomaintain contact with these physicians.

Pancreatitis events were reported bythe investigator as an adverse event(AE), a serious AE, and an event of spe-cial interest (EOSI). We systematicallycollected all events, investigators re-ceived specific training, and specific pro-spective electronic case report formswere used for event collection.

Pancreatitis events were blindly adju-dicated by an external expert commit-tee, which included two pancreaticdisease experts (N.G. and M.M.L.). Theclassification of the pancreatitis eventswas decided on by an external expertcommittee and was in accordance withinternationally accepted guidelines (18–20). Reported cases were classified intofour categories: definite acute pancrea-titis, possible acute pancreatitis, chronicpancreatitis, or unlikely to be pancreati-tis. The predefined end point was thetotal number of adjudicated cases ofpancreatitis. The diagnosis of acute pan-creatitis was regarded as definite when

at least two of the following three crite-ria from internationally accepted guide-lines (18) were met: 1) acute onset ofpersistent, severe, epigastric pain oftenradiating to the back; 2) elevation in se-rum lipase or amylase to three times ormore the upper limit of normal at therespective institution; and 3) character-istic findings of acute pancreatitis on di-agnostic imaging (transabdominalultrasound, contrast-enhanced com-puted tomography, or magnetic reso-nance imaging). To allow for greatersensitivity for a potential associationwith pancreatic disorders, we used asecond category of possible acute pan-creatitis (18), which included patients inwhom reported atypical abdominalsymptoms were not characteristic ofpancreatitis (but were without an alter-native explanatory diagnosis) plus atleast one of the following: 1) elevatedserum lipase or amylase levels to threetimes or more the upper limit of normal,2) characteristic findings of acute pan-creatitis on diagnostic imaging, or 3) ahistory of previous pancreatitis. Patientswere classified as having chronic pan-creatitis when either medical recordsdocumented or cross-sectional imagingconfirmed the diagnosis. Symptomswere regarded as unlikely to be pancre-atitis when none of the these definitionsapplied. Patients with definite or possi-ble pancreatitis were assessed as havingsevere disease when single or multipleorgan failure persisted for .48 h(19,20). Known risk factors for pancrea-titis, including alcohol use, gallstonedisease (including prior cholecystec-tomy), hypertriglyceridemia, hypercal-cemia, recent endoscopic retrogradecholangiopancreatography, trauma, andconcomitant drugs known to be associ-ated with pancreatitis, were also evalu-ated. In addition, investigators wereinstructed to treat pancreatitis accordingto local guidelines and to continue or dis-continue the study drug according totheir clinical judgment. Of note, historyof pancreatitis was not a contraindicationfor participation in the trial.

Pancreatic cancer cases were col-lected as AE and serious AE and, like allcancer cases, as EOSI. We collected asmuch information as possible on eachcase, including medical records and pa-thology and imaging results. Each reportwas reviewed for accuracy and com-pleteness by blinded study physicians

2436 SAVOR-TIMI 53dPancreatitis/Pancreatic Cancer Diabetes Care Volume 37, September 2014

of both the TIMI Study Group and thesponsor.

Statistical AnalysisAll analyses were performed with an in-tention to treat all patients who under-went randomization. Continuous variablesare expressed as mean 6 SD, medianand interquartile range, or minimumand maximum values; categorical varia-bles are expressed as frequencies andpercentages. x2 or Fisher exact tests forcategorical variables and Wilcoxon ranksum tests for continuous variables arereported. Incidence of pancreatitis wascalculated as the number of patientswith pancreatitis divided by the totalpopulation at risk in the beginning ofthe study. In addition, the cumulative in-cidence of pancreatitis at the end ofstudy was reported by Kaplan-Meierplots (Supplementary Fig. 1A–D). Todetermine the relative risk of pancreati-tis and pancreatic cancer between thesaxagliptin and placebo arms, Cox pro-portional hazards modeling using profilelikelihood estimations with stratificationby study design variables, baseline re-nal impairment, and baseline CV risk

group was performed. SAS version 8.2and above (SAS Institute Inc., Cary, NC)was used for all analyses. P , 0.05 wasconsidered statistically significant.

RESULTS

Between May 2010 and December2011, 16,492 patients were randomizedand participated in the trial (12,959 pa-tients with established CV disease [78%vs. 79% in the saxagliptin vs. placeboarms, respectively] and 3,533 with mul-tiple risk factors [22% vs. 21% in thesaxagliptin vs. placebo arms, respec-tively]). Baseline characteristics of ran-domized patients were well balancedbetween treatment arms, as previouslypublished (17) (mean [SD] age 65.0 [8.5]years, male sex 66.9%, BMI 31.2 [5.6]kg/m2, median diabetes duration10.3 [8.9] years, mean HbA1c 8.0%[1.4] [64 (9.8) mmol/mol]). The num-bers of patients taking diabetic medi-cations at baseline were as follows:insulin, 6,832 (41.4%); sulfonylurea,6,633 (40.2%); metformin, 11,473(69.6%); and none (antidiabetic drug–naıve patients), 735 (4.5%). CV medica-tions at baseline included statins,

12,917 (78.3%); aspirin, 12,404 (75.2%);ACE inhibitors/angiotensin receptorblockers, 12,995 (78.8%); and b-blockers,10,162 (61.6%). Overall, the SAVOR-TIMI53 patients had a high prevalence of CVrisk factors, including dyslipidemia(71.2%), hypertension (81.8%), BMI $30kg/m2 (53.6%), BMI $40 kg/m2 (7.5%),estimated glomerular filtration rate#50 mL/min (15.6%), and currentsmoking (13.5%).

The total observation time was16,884 person-years in the saxagliptingroup and 16,761 person-years in pla-cebo group. The study drug was prema-turely discontinued less frequently inpatients assigned to saxagliptin thanto placebo (1,527 [18.4%] vs. 1,705[20.8%], respectively, P, 0.001). A finalvital status was assessed in 99.1% of pa-tients. Twenty-eight patients were lostto follow-up. The median follow-up was2.1 years (interquartile range 1.8–2.3years), and maximum follow-up was 2.9years.

Investigators reported a total of 35events of pancreatitis in 33 patients inthe saxagliptin arm and 35 events in 30patients in the placebo arm (hazard ratio

Table 1—Investigator-reported and adjudication-confirmed pancreatitis in SAVOR-TIMI 53

Saxagliptin overall(n = 8,280)

Placebo overall(n = 8,212)

P value(Fisher exact test) HR (95% CI)

Investigator-reported pancreatitisPatients with pancreatitis events 33 (0.40) 30 (0.37) 0.80 1.09 (0.66–1.79)Number of events of pancreatitis 35 35 0.99 (0.90–1.10)

Adjudication-confirmed pancreatitisPatients with any pancreatitis event 24 (0.29) 21 (0.26) 0.77 1.13 (0.63–2.06)Number of events of pancreatitis 26 25 1.03 (0.93–1.15)Patients with definite acute pancreatitis events 17 (0.2) 9 (0.1) 0.17 1.88 (0.86–4.41)Patients with acute pancreatitis, definite or possible 22 (0.3) 16 (0.2) 0.42 1.36 (0.72–2.64)Patients with chronic pancreatitis events 2 (0.02) 6 (0.07) 0.18 0.33 (0.05–1.44)

Patients with any pancreatitis event*Duration of events (days)n 21 22 d d

Mean (SD) 13.7 (28.3) 38.5 (103.4) d d

Median (minimum, maximum) 7.0 (3.0, 135.0) 12.0 (2.0, 494.0) d d

Action taken to study drugNot applicable 4 (15.4) 5 (20.0) d d

Dose not changed 16 (61.5) 13 (52.0) d d

Drug interrupted 2 (7.7) 1 (4.0) d dDrug withdrawn 4 (15.4) 6 (24.0) d d

Outcome of the AEResolved 21 (80.8) 21 (84.0) d d

Recovering 3 (11.5) 1 (4.0) d dNot resolved 2 (7.7) 1 (4.0) d d

Resolved with sequelae 0 1 (4.0) d d

Fatal 0 1 (4.0) d d

Data are n (%) unless otherwise indicated. HR and 95% CIs from a Poisson regression model. *Summary is at the event level; denominator is the totalnumber of events for each treatment group. Eight pancreatic cases missing data on duration of events.

care.diabetesjournals.org Raz and Associates 2437

[HR] 1.09 [95% CI 0.66–1.79]) (Table 1and Supplementary Table 1). Pancreati-tis was adjudicated in 24 patients (26events) receiving saxagliptin and 21 (25events) receiving placebo (HR 1.13 [95%CI 0.63–2.06], Fisher exact test P = 0.77)(Table 1 and Supplementary Fig. 1A).One case was adjudicated first as possi-ble acute pancreatitis and subsequentlyas chronic pancreatitis. Thirty-eight pa-tients were adjudicated to have eitherdefinite or possible acute pancreatitis(22 [0.3%] vs. 16 [0.2%], HR 1.36 [95%CI 0.72–2.64], Fisher exact test P = 0.42)in the saxagliptin and placebo arms, re-spectively (Table 1 and SupplementaryFig. 1B). Of these, 26 patients were ad-judicated to have definite acute pancre-atitis (17 [0.2%] vs. 9 [0.1%], HR 1.88[95% CI 0.86–4.41], Fisher exact testP = 0.17) in the saxagliptin and placeboarms, respectively (Table 1 and Supple-mentary Fig. 1C). Repeated cases of pan-creatitis were reported in two patientsin the saxagliptin arm and three patientsin the placebo arm. Of these five pa-tients, one in the saxagliptin arm waspermanently discontinued from thestudy drug after the first event, yet hehad another occurrence of pancreatitis,and the second patient was continuedon the study drug. In the placebo arm,two patients were permanently discon-tinued from the study drug after the firstevent, yet one had two recurrentevents, and one was continued on

the study drug. Preexisting risk factorsfor pancreatitis were identified in 18(81.8%) patients with acute pancreatitis(definite and possible) receiving saxa-gliptin and in 15 patients (93.8%) withacute pancreatitis receiving placebo(Fig. 1 and Supplementary Table 2).Risk factors in general and the numberof risk factors per patient were distrib-uted similarly between the two treat-ment arms (Supplementary Table 2and Fig. 1). The blinded pancreatitis ad-judication committee considered 19cases as unlikely to be pancreatitis (9in the saxagliptin arm and 10 in the pla-cebo arm). All these cases did nothave typical abdominal pain, increasedpacreatic enzymes (more than threetimes the upper limit of normal), orimaging consistent with pancreatitis.The event was recovered in 16 (7 inthe saxagliptin arm and 9 in the placeboarm) and not recovered in 2 in thesaxagliptin arm (1hadhepatic cellular car-cinoma) and 1 in the placebo arm (whohad metastatic pancreatic cancer).

After inclusion in the study, the timeto onset of symptoms of all adjudicateddefinite or possible acute pancreatitis,definite acutepancreatitis, or chronic pan-creatitis did not differ between treatmentarms. Events occurred at various timepoints during the treatment period, withno cluster of events in either treatmentarm (Fig. 2). Investigators’ assessmentsof study medication causality in all

pancreatitis cases in the saxagliptin andplacebo arms were similar (nine vs. nine,respectively), as was the frequency atwhich the study medication was discon-tinued due to pancreatitis (six vs. seven,respectively). Adjudicated reports of se-vere pancreatitis and deaths were simi-lar between the saxagliptin and placeboarms (one case of severe pancreatitis ineach arm, and no vs. one death in thesaxagliptin vs. placebo arms, respec-tively) (Table 1). The mean (SD) durationof events was 13.7 (28.3) vs. 38.5 (103.4)days in saxagliptin- vs. placebo-treatedpatients, respectively. The recoveryrate from pancreatitis was similar be-tween the two treatment arms (21[80.8%] vs. 21 [84.0%] were resolved, 2[7.7%] vs. 1 [4.0%] were not resolved,and 0 vs. 1 [4%] were resolved with se-quelae in the saxagliptin vs. placeboarms, respectively). The number of pa-tients who continued on the study med-ication despite pancreatitis was 16(61.5%) in the saxagliptin arm comparedwith 13 (52.0%) in the placebo arm (Ta-ble 1). The proportion of patients withmore than two on-treatment AEs ofpancreatitis was generally similar be-tween the two treatment groups (one[,0.1%] patient in the saxagliptin groupand two [,0.1%] in the placebo group).

Chronic pancreatitis was reported bythe investigators in five patients receiv-ing saxagliptin and four receiving pla-cebo; adjudication confirmed chronic

Figure 1—Most common risk factors for pancreatitis identified in patients with pancreatitis, intention-to-treat population. Con., control; HyperTG,hypertriglyceridemia.

2438 SAVOR-TIMI 53dPancreatitis/Pancreatic Cancer Diabetes Care Volume 37, September 2014

pancreatitis in two patients receivingsaxagliptin (0.02%) and six receiving pla-cebo (0.07%) (HR 0.33 [95%CI 0.05–1.44],Fisher exact test P = 0.18)] (Sup-plementary Fig. 1D). Pancreatic cancerwas reported in 5 patients in thesaxagliptin arm and 12 in the placeboarm (HR 0.42 [95% CI 0.13–1.12], Fisherexact test P = 0.09), including 1 patientwith a neuroendocrine tumor in the pla-cebo arm.

CONCLUSIONS

The frequency of pancreatitis cases pre-viously reported in epidemiologicalstudies of diabetic patients aged .40years versus the general populationwas 54.0 vs. 30.1 per 100,000 person-years, respectively (adjusted rate ratio1.77 [95% CI 1.46–2.15]) (7). These num-bers, however, are imprecise because ofinvestigators’ reliance on epidemiologi-cal data rather than on prospective clin-ical studies. The SAVOR-TIMI 53 trial isthe first prospective study in which pan-creatitis was predefined as an EOSI.We systematically collected informa-tion regarding all pancreatitis events,using a similar process to that for theCV end points, and used an indepen-dent, blinded adjudication committeeto ensure the quality of these adju-dicated end points. This rigorous pro-cess provides important data regarding

pancreatitis in type 2 diabetic patients,particularly regarding incretin-basedtherapy. To increase sensitivity formilder cases of pancreatitis, we usedboth the definite and the possible defi-nition. However, because amylase andlipase levels were not routinely exam-ined throughout the trial, and clinicalvisits were performed every 6 months(with a telephone call visit every 3months in between), subclinical pancre-atitis might have been missed. Post hocanalysis of studies that examined thelevels of lipase and amylase in over-weight/obese patients versus type 2 di-abetic patients found a higher incidenceof increased enzyme levels and highervariability within the same patient,among type 2 diabetic patients (21).The use of DPP-4 inhibitors might leadto a further increase and fluctuation inlevels of amylase and lipase in asymp-tomatic patients (22). The clinical im-pact, however, of a diabetes-related aswell as a drug-induced increase in amy-lase and lipase levels and their relationto subclinical pancreatitis are unclear.

The predefined end point of all casesof pancreatitis was similarly distributedbetween treatment arms (26 vs. 25 inthe saxagliptin vs. placebo arms, respec-tively; HR 1.03 [95% CI 0.93–1.15]).However, there were numerical differ-ences between the treatment arms

regarding the subtypes of pancreatitis.In the saxagliptin arm, more cases ofdefinite acute pancreatitis were found(17 vs. 9, HR 1.88 [95% CI 0.86–4.41])and fewer cases of possible (5 vs. 7, HR0.85 [95% CI 0.27–2.56]) and chronic (2vs. 6, HR 0.33 [95% CI 0.05–1.44]) pan-creatitis. None of these differencesreached statistical significance, how-ever. The meaning of statistical non-significance with such a small numberof cases is limited. In addition, the nu-merical imbalance in the cases of defi-nite acute pancreatitis that were alsoobserved in the EXAMINE (Examinationof Cardiovascular Outcomes: Alogliptinvs. Standard of Care in Patients withType 2 Diabetes Mellitus and Acute Cor-onary Syndrome) trial (12 [0.4%] vs. 8[0.3%] unadjudicated cases of pancrea-titis in the alogliptin vs. placebo arm,respectively) (9) should not be dis-missed without further investigation.

A possible slight increase in the riskfor pancreatitis should be balancedwith the favorable course and outcomeof the rare pancreatitis cases. Bothstudy arms had similar clinical outcomesas expressed by the length of hospitali-zation and similar high rates of recoveryregardless of continuation or discon-tinuation of study drug, and no fatali-ties occurred in the saxagliptin arm.The investigators’ perception of most

Figure 2—Scatter plots of adjudication-confirmed pancreatitis, ITT population by category. Subject E7836013 (saxagliptin) was confirmed as havingdefinite acute pancreatitis after possible acute pancreatitis and is presented twice in the figure. ITT, intention to treat.

care.diabetesjournals.org Raz and Associates 2439

pancreatitis events as mild might ex-plain the high continuation rate of studydrug in both the saxagliptin (61.5%)and the placebo (52%) arms. Finally,the low rate of recurrent events (twoin the saxagliptin arm and three in theplacebo arm), regardless of continua-tion or discontinuation of study drug,is also reassuring.Patients with type 2 diabetes have an

increased risk of developing acute pan-creatitis, chronic pancreatitis, and pan-creatic cancer (1,23) compared with thegeneral population and even more sowhen known confounders, such as obe-sity (24) and CV disease (25), exist.Moreover, medications often used totreat diabetes comorbidities have beenshown in population-based case-controlstudies to increase the risk of develop-ing pancreatitis, including statins (oddsratio [OR] 1.44), ACE inhibitors (OR 1.5),antidepressants (OR 2.8), and nonste-roidal anti-inflammatory drugs (OR 2.7)(26). Previously published retrospectivestudies suggested that the relationshipbetween incretin-based therapy andpancreatitis was nonrandomized andtherefore uncontrolled for baselineknown and unknown risk factors forpancreatitis. Pancreatitis cases werenot blindly adjudicated, which mayhave led to a bias of overreportingamong the incretin-treated patients(4,5,27). The current study is the firstin our knowledge to prospectively adju-dicate pancreatitis, and therefore, itprovides a more precise estimate ofthe actual rate.Whereas acute pancreatitis must be

regarded as an incident condition thatoccurred during the treatment periodof the study, the same argument doesnot hold for chronic pancreatitisbecause of the difficulty in excluding un-diagnosed, preexisting chronic pancrea-titis; therefore, a possible associationwith the respective treatment arm isless stringent (28). Following cases ofchronic pancreatitis is important be-cause data suggest that patients withchronic pancreatitis are at increasedrisk of pancreatic cancer (29–31). Inthe SAVOR-TIMI 53 trial, eight cases ofchronic pancreatitis were confirmed byadjudication (two in the saxagliptinarm and six in the placebo arm), andall but one were reported after .300days of exposure. Pancreatic cancerwas reported in 5 patients in the

saxagliptin arm and 12 (including 1with neuroendocrine tumor) in the pla-cebo arm (Fisher exact test P = 0.09).Although a reduction in pancreatic can-cer risk has been suggested for antidia-betic drugs like metformin (32,33),such a protective effect is not assumedfor saxagliptin. However, concerns aboutan increased pancreatic cancer or neuro-endocrine tumor risk, as proposed byothers in patients under relatively shortexposure to DPP-4 inhibitors (2), couldnot be confirmed in the current study. Itshould be noted that the median follow-up period in the SAVOR-TIMI 53 trialwas 2.1 years. Larger studies with longerfollow-up periods are needed to confirmthe findings.

The SAVOR-TIMI 53 trial did not com-pletely resolve the issue of whetherDPP-4 inhibitors are associated with aslight increase in the risk of pancreatitis(mainly milder cases). The possible riskshould be properly judged in the contextof the favorable outcome of these casesand the overall proven safety record ofthis group of drugs. This question will befurther determined by the results ofother ongoing DPP-4 inhibitors prospec-tive outcome trials (34,35).

The SAVOR-TIMI 53 data do notsupport a pancreatic cancer signal. It isunlikely, however, that the effect ofDPP-4 inhibitors on cancer risk and pro-gression, particularly at specific cancersites, will be fully addressed with ran-domized controlled clinical trials be-cause of the follow-up time limitation.The combination of the ongoing largeoutcome studies with a well-balancedmeta-analysis and continued largedata-based surveys might further clarifythis issue.

Acknowledgments. SAVOR-TIMI 53 ExecutiveCommittee: Eugene Braunwald (study chair),Deepak L. Bhatt (co-principal investigator),Itamar Raz (co-principal investigator), Jaime A.Davidson, Robert Frederich (nonvoting), BoazHirshberg (nonvoting), Ph. Gabriel Steg.Funding and Duality of Interest. AstraZenecaand Bristol-Myers Squibb (BMS) funded thestudy.

The sponsors were involved in the study designand data collection and analysis. They weregiven the opportunity to review the manuscriptbut not to change any aspect.

I.R. discloses the following relationships: advi-sory board for Novo Nordisk, AstraZeneca/BMS,Sanofi, Merck Sharp & Dohme (MSD), and EliLilly; consultant for AstraZeneca/BMS, Insuline,

and Andromeda Biotech Ltd.; and speaker’sbureau for Novo Nordisk, AstraZeneca/BMS,Sanofi, MSD, Eli Lilly, and Novartis. D.L.B.discloses the following relationships: advisoryboard for Elsevier PracticeUpdate Cardiology,Medscape Cardiology, and Regado Biosciences;board of directors for Boston VA Research In-stitute and Society of Cardiovascular PatientCare; chair of the American Heart AssociationGet With The Guidelines Steering Committee;data monitoring committees for Duke ClinicalResearch Institute, Harvard Clinical Research In-stitute, Mayo Clinic, and Population Health Re-search Institute; honoraria from the AmericanCollege of Cardiology (editor, CardioSource clin-ical trial summaries), Belvoir Publications (edi-tor in chief, Harvard Heart Letter), Duke ClinicalResearch Institute (clinical trial steering com-mittees), Harvard Clinical Research Institute(clinical trial steering committee), HMP Com-munications (editor in chief, Journal of InvasiveCardiology), Population Health Research Insti-tute (clinical trial steering committee), SlackPublications (chief medical editor, Cardiology To-day’s Intervention), WebMD (continuing medicaleducation steering committees); other editorial ac-tivities for Clinical Cardiology (associate editor) andthe Journal of the American College of Cardiology(section editor, pharmacology); research grantfunding from Amarin, AstraZeneca/BMS, Eisai,Ethicon, Medtronic, Roche, Sanofi, and TheMedicines Company; and unfunded researchfor FlowCo, PLx Pharma, and Takeda. B.H. andC.S. are employees of AstraZeneca. O.M. disclosesthe following relationships: research grant sup-port through Hadassah University Hospitalfunded by Novo Nordisk; speaker’s bureau forAstraZeneca/BMS, Novo Nordisk, Eli Lilly,Sanofi, Novartis, and MSD; advisory board forAstraZeneca/BMS, Novo Nordisk, Eli Lilly, Sanofi,and Novartis; and grants paid to her institutionas study physician by AstraZeneca/BMS. B.M.S.discloses the following relationships: researchgrant support through Brigham and Women’sHospital from AstraZeneca/BMS, Daichi-Sankyo,GlaxoSmithKline, Johnson & Johnson, Gilead,Eisai, Arena, and Merck; consulting fees fromGilead, Arena, Eisai, Lexicon, BMS, and St. Jude’sMedical; and editorial board for Elsevier(PracticeUpdate Cardiology). A.U.-E. and K.I. dis-close research grant support to the TIMI StudyGroup from AstraZeneca. A.B. is the HadassahUniversity Hospital Diabetes Unit Project Man-ager for the SAVOR-TIMI 53 trial. N.I. is an em-ployee of BMS. M.M.L. discloses the followingrelationships: advisory board for Eli Lilly, Roche,Solvay, Abbott, MSD, and AstraZeneca; consul-tant for British Biotech, Bioserve, Centogene,Metanomics, Takeda, AstraZeneca/BMS,Sanochemia, Alpinia, and Novartis; and speaker’sbureau forMenarini, Recordati, Falk, Nordmark,Solvay, Abbott, AbbVie, and Roche. No otherpotential conflicts of interest relevant to thisarticle were reported.Author Contributions. I.R. contributed to thestudy concept, design, and supervision; dataanalysis and interpretation; and drafting of themanuscript. D.L.B., B.H., and B.M.S. contributedto the study concept, design, and supervision;data analysis and interpretation; and criticalrevision of the manuscript for important in-tellectual content. O.M. contributed to the data

2440 SAVOR-TIMI 53dPancreatitis/Pancreatic Cancer Diabetes Care Volume 37, September 2014

analysis and interpretation, drafting of themanuscript, and critical revision of the manu-script for important intellectual content. A.U.-E.and K.I. contributed to the statistical analysis.C.S. and N.I. contributed to the study supervi-sion, data analysis and interpretation, draftingof the manuscript, and critical revision of themanuscript for important intellectual content.A.B. contributed to the administrative, techni-cal, and material support; data acquisition,analysis, and interpretation; and critical revisionof the manuscript for important intellectualcontent. N.G. contributed to the data acquisi-tion, analysis, and interpretation. M.M.L. con-tributed to the data acquisition, analysis, andinterpretation and drafting of the manuscript.I.R. is the guarantor of this work and, as such,had full access to all the data in the study andtakes responsibility for the integrity of the dataand the accuracy of the data analysis.

References1. Noel RA, Braun DK, Patterson RE, BloomgrenGL. Increased risk of acute pancreatitis and bil-iary disease observed in patients with type 2diabetes: a retrospective cohort study. DiabetesCare 2009;32:834–8382. Butler AE, Campbell-Thompson M, Gurlo T,Dawson DW, Atkinson M, Butler PC. Markedexpansion of exocrine and endocrine pancreaswith incretin therapy in humans with increasedexocrine pancreas dysplasia and the potentialfor glucagon-producing neuroendocrine tu-mors. Diabetes 2013;62:2595–26043. Butler PC, Elashoff M, Elashoff R, Gale EA. Acritical analysis of the clinical use of incretin-based therapies: Are the GLP-1 therapies safe?Diabetes Care 2013;36:2118–21254. Dore DD, Seeger JD, Arnold Chan K. Use of aclaims-based active drug safety surveillance sys-tem to assess the risk of acute pancreatitis withexenatide or sitagliptin compared to metforminor glyburide. Curr Med Res Opin 2009;25:1019–10275. Singh S, Chang HY, Richards TM, Weiner JP,Clark JM, Segal JB. Glucagonlike peptide 1-based therapies and risk of hospitalization foracute pancreatitis in type 2 diabetes mellitus apopulation-based matched case-control study.JAMA Intern Med 2013;173:534–5396. Nauck MA. A critical analysis of the clinicaluse of incretin-based therapies: the benefits byfar outweigh the potential risks. Diabetes Care2013;36:2126–21327. Gonzalez-Perez A, Schlienger RG, RodrıguezLA. Acute pancreatitis in association with type 2diabetes and antidiabetic drugs: a population-based cohort study. Diabetes Care 2010;33:2580–25858. Engel SS, Williams-Herman DE, Golm GT,et al. Sitagliptin: review of preclinical and clini-cal data regarding incidence of pancreatitis. Int JClin Pract 2010;64:984–9909. Bonner-Weir S, In’t Veld PA, Weir GC. Re-analysis of study of pancreatic effects of incretin

therapy: methodological deficiencies. DiabetesObes Metab. In press10. Food and Drug Administration, Center forDrug Evaluation and Research. Guidance for in-dustry: diabetes mellitusdevaluating cardio-vascular risk in new antidiabetic therapies totreat type 2 diabetes [article online], 2008. Avail-able from http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm071627.pdf. Accessed 19 January201111. Hirshberg B, Raz I. Impact of the U.S. Foodand Drug Administration cardiovascular assess-ment requirements on the development ofnovel antidiabetes drugs. Diabetes Care 2011;34(Suppl. 2):S101–S10612. Scirica BM, Bhatt DL, Braunwald E, et al.;SAVOR-TIMI 53 Steering Committee and In-vestigators. Saxagliptin and cardiovascular out-comes in patients with type 2 diabetes mellitus.N Engl J Med 2013;369:1317–132613. White WB, Cannon CP, Heller SR, et al.; EX-AMINE Investigators. Alogliptin after acute cor-onary syndrome in patients with type 2diabetes. N Engl J Med 2013;369:1327–133514. Rosenstock J, Marx N, Kahn SE, et al. Ratio-nale and design of the CAROLINA trial: an activecomparator CARdiOvascular Outcome Study ofthe DPP-4 Inhibitor LINAgliptin in Patients withType 2 Diabetes At High Cardiovascular risk (Ab-stract). Diabetes 2011;60(Suppl. 1):A30315. Bethel MA, Green J, Califf RM, Holman RR.Rationale and design of the Trial Evaluating Car-diovascular Outcomes with Sitagliptin (TECOS)(Abstract). Available from http://professional.diabetes.org/Abstracts_Display.aspx?TYP=1&CID=74327. Accessed 1 October 201316. Scirica BM, Bhatt DL, Braunwald E, et al. Thedesign and rationale of the Saxagliptin Assess-ment of Vascular Outcomes Recorded in Pa-tients with Diabetes Mellitus-Thrombolysis inMyocardial Infarction (SAVOR-TIMI) 53 study.Am Heart J 2011;162:818–82517. Mosenzon O, Raz I, Scirica BM, et al. Base-line characteristics of the patient population inthe Saxagliptin Assessment of Vascular Out-comes Recorded in patients with diabetes mel-litus (SAVOR)-TIMI 53 trial. Diabetes Metab ResRev 2013;29:417–42618. Banks PA, Bollen TL, Dervenis C, et al.; AcutePancreatitis Classification Working Group. Clas-sification of acute pancreatitisd2012: revisionof the Atlanta classification and definitions byinternational consensus. Gut 2013;62:102–11119. LerchMM. Classifying an unpredictable dis-ease: the revised Atlanta classification of acutepancreatitis. Gut 2013;62:2–320. Bradley EL 3rd. A clinically basedclassification system for acute pancreatitis.Summary of the International Symposium onAcute Pancreatitis, Atlanta, Ga, September 11through 13, 1992. Arch Surg 1993;128:586–59021. Malloy J, Gurney K, Shan K, Yan P, Chen S.Increased variability and abnormalities in pan-creatic enzyme concentrations in otherwise

asymptomatic subjects with type 2 diabetes. Di-abetes Metab Syndr Obes 2012;5:419–42422. Tokuyama H, Kawamura H, Fujimoto M,et al. A low-grade increase of serum pancreaticexocrine enzyme levels by dipeptidyl peptidase-4inhibitor in patients with type 2 diabetes. Di-abetes Res Clin Pract 2013;100:e66–e6923. Giovannucci E, Harlan DM, ArcherMC, et al.Diabetes and cancer: a consensus report. Dia-betes Care 2010;33:1674–168524. Martınez J, Johnson CD, Sanchez-Paya J, deMadaria E, Robles-D ıaz G, Perez-Mateo M.Obesity is a definitive risk factor of severityand mortality in acute pancreatitis: an up-dated meta-analysis. Pancreatology 2006;6:206–20925. Bexelius TS, Ljung R, Mattsson F, LagergrenJ. Cardiovascular disease and risk of acute pan-creatitis in a population-based study. Pancreas2013;42:1011–101526. Nitsche C, Maertin S, Scheiber J, Ritter CA,Lerch MM, Mayerle J. Drug-induced pancreati-tis. Curr Gastroenterol Rep 2012;14:131–13827. Dore DD, Bloomgren GL, WentenM, et al. Acohort study of acute pancreatitis in relation toexenatide use. Diabetes Obes Metab 2011;13:559–56628. Mayerle J, Hoffmeister A, Werner J, Witt H,Lerch MM, Mossner J. Chronic pancreatitisddefinition, etiology, investigation and treatment.Dtsch Arztebl Int 2013;110:387–39329. Lowenfels AB, Maisonneuve P, Cavallini G,et al.; International Pancreatitis Study Group.Pancreatitis and the risk of pancreatic cancer.N Engl J Med 1993;328:1433–143730. Ekbom A, McLaughlin JK, Karlsson BM, et al.Pancreatitis and pancreatic cancer: a population-based study. J Natl Cancer Inst 1994;86:625–62731. Duell EJ, Lucenteforte E, Olson SH, et al.Pancreatitis and pancreatic cancer risk: a pooledanalysis in the International Pancreatic CancerCase-Control Consortium (PanC4). Ann Oncol2012;23:2964–297032. Li D, Yeung S-CJ, Hassan MM, Konopleva M,Abbruzzese JL. Antidiabetic therapies affect riskof pancreatic cancer. Gastroenterology 2009;137:482–48833. Currie CJ, Poole CD, Gale EAM. The influ-ence of glucose-lowering therapies on cancerrisk in type 2 diabetes. Diabetologia 2009;52:1766–177734. Green JB, Bethel MA, Paul SK, et al. Ratio-nale, design, and organization of a randomized,controlled Trial Evaluating Cardiovascular Out-comes with Sitagliptin (TECOS) in patients withtype 2 diabetes and established cardiovasculardisease. Am Heart J 2013;166:983–98935. Rosenstock J, Marx N, Kahn SE, et al. Car-diovascular outcome trials in type 2 diabetesand the sulphonylurea controversy: rationalefor the active-comparator CAROLINA trial.Diab Vasc Dis Res 2013;10:289–301

care.diabetesjournals.org Raz and Associates 2441