Epidemiology/Health Services Research ORIGINAL ARTICLE

Use of Dipeptidyl Peptidase-4 Inhibitors and the Reporting of Infections: A Disproportionality Analysis in the World Health Organization VigiBase

1,2 1,4 MARJOLEIN J. WILLEMEN, PHARMD RON H. MEYBOOM, PHD by reducing fasting and postprandial glu- 1,2 1,5 AUKJE K. MANTEL-TEEUWISSE, PHD TOINE C. EGBERTS, PHD 2,3 1,2 cose concentrations in patients with type 2 SABINE M. STRAUS, PHD HUBERT G. LEUFKENS, PHD diabetes (1). DPP-4 is assumed to have many other functions in the human physiology due — OBJECTIVE Dipeptidyl peptidase-4 (DPP-4) inhibitors are a new class of antidiabetic drugs. to its presence on the surface of many They inactivate incretin hormones but also have many other effects throughout the body, among different cell types, but these effects are which are effects on the immune system. This might result in an increased infection risk. This study assessed the association between use of DPP-4 inhibitors and the reporting of infections. still largely unknown. The role of DPP-4 in immune regulation is better defined RESEARCH DESIGN AND METHODS—A nested case-control was conducted using and includes induction of transforming VigiBase, the World Health Organization-Adverse Drug Reactions (WHO-ADR) database. The growth factor-b1 in activated T cells and base cohort consisted of ADRs for antidiabetic drugs (Anatomical Therapeutic Chemical code suppression of production of inflamma- A10). Cases were defined as ADRs of infection according to the Medical Dictionary for Regulatory fi tory cytokines by T cells (4), effects on cell Activities (MedDRA) classi cation system. All other ADRs were considered controls. Reporting growth, differentiation, and apoptosis odds ratios (RORs) were calculated to estimate the strength of the association between different (5,6). The immunomodulating effect has classes of antidiabetic drugs and the reporting of infections. given rise to concerns regarding a possible RESULTS—We identified 305,415 suspected ADRs involving antidiabetic drugs in 106,469 increase in the occurrence of infections case reports, of which 8,083 involved DPP-4 inhibitors monotherapy. Overall, the reporting of (1–3). infections was higher for patients using DPP-4 inhibitors compared with users of biguanides Nasopharyngitis, upper respiratory (ROR 2.3 [95% CI 1.9–2.7]). Reporting of upper respiratory tract infections (ROR 12.3 [95% CI tract (URTI), and related infections (acute – fi 8.6 17.5]) was signi cantly associated with use of DPP-4 inhibitors. bronchitis, pharyngitis, sinusitis, and rhi- CONCLUSIONS—This study indicates an increased reporting of infections, in particular nitis) were the most commonly reported upper respiratory tract infections, for users of DPP-4 inhibitors compared with users of other infections for the active substances com- antidiabetic drugs. However, the limitations of spontaneous reporting systems (e.g., underre- pared with the reference intervention in porting, the Weber-effect, reporting bias) should be taken into account. Therefore, further re- clinical trial programs (1–3). However, search is needed to evaluate this suspicion and the underlying mechanism. pooled analyses for vildagliptin and saxagliptin did not indicate an increased – Diabetes Care 34:369 374, 2011 risk of infections compared with the ref- erence group (7,8). In the three European ipeptidyl peptidase-4 (DPP-4) inhib- peptide-1 and glucose-dependent insuli- Union (EU) Risk Management Plans (a Ditors are a new class of antidiabetic notropic polypeptide) by DPP-4 inhibitors mandatory part of marketing applications drugs, with three products currently results in a rise in insulin from pancreatic since November 2005 [9]) for the ap- available on the market: sitagliptin, vilda- b-cells and a decrease in glucagon from proved DPP-4 inhibitors, “infections” gliptin, and saxagliptin (1–3). The inacti- pancreatic a-cells. As a consequence, were defined as important identified risks vation of incretin hormones (glucagon-like DPP-4 inhibitors improve glycemic control that require further evaluation. Postau- thorization safety studies specifically eval- uating the risk of hospitalization due to ccccccccccccccccccccccccccccccccccccccccccccccccc infections are currently being conducted From the 1Division of Pharmacoepidemiology and Clinical Pharmacology, Faculty of Science, Utrecht Institute for vildagliptin and saxagliptin (2,3). For for Pharmaceutical Sciences, Utrecht, the Netherlands; the 2Medicines Evaluation Board, The Hague, the 3 sitagliptin, the risk for infections will be Netherlands; the Department of Medical Informatics, Erasmus University Medical Center, Rotterdam, the Netherlands; the 4World Health Organization Collaborating Centre for International Drug Monitoring, further evaluated through an in-depth Uppsala Monitoring Centre, Uppsala, Sweden; and the 5Department of Clinical Pharmacy, University analysis of the safety results of the ongo- Medical Center Utrecht, Utrecht, the Netherlands. ing and planned clinical trials (1). Corresponding author: Aukje K. Mantel-Teeuwisse, [email protected]. Data on a possible direct relation Received 14 September 2010 and accepted 5 November 2010. DOI: 10.2337/dc10-1771 between diabetes mellitus and infections This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/ are inconclusive. Several studies investi- dc10-1771/-/DC1. gated a possible association between di- The opinions and conclusions are those of the authors and not necessarily those of the Uppsala Monitoring abetes mellitus and alterations of the Centre, the National Centers, or the World Health Organization. © 2011 by the American Diabetes Association. Readers may use this article as long as the work is properly immune system (10,11). Some epidemi- cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ ologic studies showed that these patients licenses/by-nc-nd/3.0/ for details. are at an increased risk for common care.diabetesjournals.org DIABETES CARE, VOLUME 34, FEBRUARY 2011 369 DPP-4 inhibitors and reporting of infections infections (12–15), but evidence from through a manual search were defined as the high-level group terms URTI, LRTI, and clinical trials is limited and inconsistent cases. All reports containing other ADRs UTI. Adjusted analyses were conducted (16). Disease progression may have an ef- were considered as controls. We grouped with all potential confounders included in fect on the occurrence of infections; thus, the infections on the first sublevel (high- the model. more severely ill patients might be at an level group terms) of MedDRA and looked More ADRs were usually reported for increased risk of infections (17). To our at URTI (e.g., sinusitis and nasopharyngi- onecasereport,andthereforeitwas knowledge, no studies have specifically tis), lower respiratory tract infections possible that one case report contained investigated the relation between the use (LRTI, e.g., bronchitis and pneumonia), more than one ADR of an infection. To of DPP-4 inhibitors and infections as ad- and urinary tract infections (UTI; e.g., cys- test the effect of multiple ADRs reported verse drug reactions (ADRs). Therefore, titis and pyelonephritis). Because of low in one case report, we analyzed the data the aim of the current study was to assess numbers, all other infections were com- on the level of case reports (one case the relation between different classes of bined. report generally represented one patient). antidiabetic drugs and the reporting of In the U.S., DPP-4 inhibitors are in- fi infections. Exposure de nition dicated for monotherapy for the treat- Exposure to antidiabetic drugs was the ment of diabetes mellitus, whereas in the RESEARCH DESIGN AND determinant that was investigated. Anti- EU, these medicines are only indicated for METHODS diabetic drugs were subclassified based combination therapies. To check whether on the Anatomical Therapeutic Chemical this affected the results of this study, we Setting and study design (ATC) classification system of the WHO performed a sensitivity analysis by ana- Data were obtained from the International (www.whocc.no): biguanides (ATC code lyzing the data for the U.S. and the rest of Drug Monitoring Program of the World A10BA), sulfonylurea derivatives (A10BB), the world separately. In addition, to study Health Organization (WHO). The WHO thiazolidinediones (A10BG), DPP-4 in- the effect of the type of reporter (health global individual case safety report (ICSR) hibitors (A10BH), insulins, and analogs care professional or consumer) on the database, VigiBase, is maintained by the (A10A). When multiple antidiabetic drugs outcome, we performed a sensitivity anal- Uppsala Monitoring Centre and contains were reported for a certain ADR, this was ysis in which the reports were assessed summaries of suspected spontaneous case classified as combination therapy, irrespec- according to the type of reporter. Statisti- reports originally summated by health tive of whether a drug reported was classi- cal analysis was done using SPSS 16.0 care professionals and patients to national fied as “suspected” or as a comedication. statistical software (SPSS Inc., Chicago, IL). pharmacovigilance centers in 98 coun- tries worldwide. As of May 2010, this Potential confounding factors RESULTS—From the WHO VigiBase database contained .5 million case re- Potential confounding factors retrieved we identified 305,415 suspected ADRs ports of suspected ADRs regarding spe- from the case reports included age and related to the use of antidiabetic drugs in cific, but anonymous, patients. The gender of the patient, reporting year, 106,469 case reports in the study period reports contain administrative data, pa- reporting region (Europe, North America, 1999 through 2009. Patients were a mean tient data, ADR data, medication data, rest of the world), and reporter type, age of 59.7 years (SD, 14.3) and 59.6% and additional information. The informa- including physician, pharmacist, other care- were women. A total of 288,434 reports tion in these reports is not homogenous, at giver, pharmaceutical company (indirectly (94.4%) reports involved one antidiabetic least with regard to origin, completeness obtained from a health care professional), drug, 14,057 (4.6%) reported a combi- of documentation, or the likelihood that and patient/consumer. Concomitant use nation of two antidiabetic drugs, and the suspected drug caused the adverse of medication affecting the immune sys- 2,924 (1.0%) reported three or more events (18). ADRs are coded according tem, defined as reporting one of these drugs antidiabetic drugs (Table 1). Overall, the to the Adverse Reaction Terminology as a concomitant drug for an ADR, was most commonly reported infections on (WHO-ART) and Medical Dictionary for taken into account when recorded, in- the level of MedDRA lower-level terms Regulatory Activities (MedDRA; www. cluding antibiotics (ATC code J01), corti- were pneumonia (11.8%), nasopharyngi- who-umc.org). costeroids for systemic use (H02), and tis (10.1%), UTIs (6.2%), infection not This study was designed as a nested immunosuppressants (L04). otherwise specified (5.5%), sinusitis case-control study. The base cohort con- (5.1%), and bronchitis (4.8%). All other sisted of all ADRs associated with the use Data analysis types of infection were reported in of any antidiabetic drug (Anatomical Descriptive statistics were used to sum- ,4.5% of the reports related to infections. Therapeutic Chemical [ATC] code A10), marize the baseline characteristics of the A total of 242 infections were reported including oral antidiabetic drugs and in- case reports. Unconditional logistic re- as a MedDRA term in 212 case reports sulins, in the period 1999 through 2009. gression analysis was used to estimate the for DPP-4 inhibitors, of which 188 strength of the association between use of (88.7%) reported one infection. Of the Definition of cases and controls antidiabetic drugs and reporting of infec- 24 case reports (11.3%) with multiple Cases were defined as ADRs classified as tions and expressed as reporting odds infections, 12 (50%) reported a nonspe- an infection. Infections were defined by ratios (RORs) with corresponding 95% cific infection term, such as “infection” or means of MedDRA adverse reaction CIs. Biguanides were the reference group. “URTI,” combined with a more specific terms, including all relevant high-level Owing to low numbers, combination infection term, such as “nasopharyngi- group terms and lower-level terms. All therapy was analyzed on an aggregated tis” or “cystitis” (see the Supplementary infections from the System Organ Class level, not on an individual drug-drug Data for a summary of the infections ac- (SOC) “Infections and Infestations” and in- combination level. We focused on infec- cording to the MedDRA lower-level fections reported in other SOCs identified tions in general and more specifically on term).

370 DIABETES CARE, VOLUME 34, FEBRUARY 2011 care.diabetesjournals.org Willemen and Associates

Table 1—Baseline characteristics of all Table 2—Crude and adjusted RORs for any infection spontaneous reports for antidiabetic drugs – in the WHO VigiBase (1999 2009) ROR (95% CI) Drug Reports (N) Reports of infections (N) Crude Adjusted* Spontaneous reports Biguanides 21,763 289 Reference Reference Variable (N = 305,415) SU derivatives 16,675 258 1.2 (1.0–1.4) 1.1 (0.9–1.3) TZDs 57,814 919 1.2 (1.1–1.4) 1.3 (1.1–1.5) Patient age, mean (SD), DPP-4 I 8,083 242 2.3 (1.9–2.7) 2.3 (1.9–2.9) years 59.7 (14.3) Insulins 80,347 1,703 1.6 (1.4–1.8) 1.5 (1.3–1.7) Age (missing) 63,212 (20.7) OAD + OAD 11,991 155 1.0 (0.8–1.2) 0.9 (0.7–1.1) Female sex 182,130 (59.6) OAD + insulin 2,066 48 1.8 (1.3–2.4) 1.5 (1.1–2.2) Sex (missing) 9,100 (3.0) $3 ADs 2,924 39 1.0 (0.7–1.4) 1.1 (0.9–1.4) Reporter ADs, antidiabetic drugs; DPP-4 I, dipeptidyl peptidase inhibitors; OAD, oral antidiabetic drug; SU, sulfo- Health care professional 92,896 (30.4) nylurea; TZD, thiazolidinediones. *Adjusted for age, sex, reporting year, reporting region, reporter type, and Nonhealth care comedication affecting the immune system: antibiotics ( J01), corticosteroids for systemic use (H02), and professional 129,287 (42.3) immunosuppressants (L04). Study/literature 218 (0.1) Unknown 45,308 (14.8) 1.6 [95% CI 1.4–1.8]), and the combina- The crude ROR was lower for DPP-4 Other 37,706 (12.3) tion of any oral antidiabetic drug and in- inhibitors but was still significantly in- Reporting year sulin therapy (ROR 1.8 [95% CI 1.3–2.4]) creased (ROR 1.6 [95% CI 1.3–1.9]) in 1999–2004 93,255 (30.5) were all statistically significant associated the analysis of the data on the level of 2005–2009 212,160 (69.5) with ADR reports of infections compared case reports. The RORs for other antidia- Region with biguanides. A slightly increased re- betic drugs did not change, except for in- Europe 19,252 (6.3) porting of infections for patients using sulin. The ROR for insulin monotherapy United States 273,079 (89.4) thiazolidinediones was found (ROR 1.2 increased to 2.1 (95% CI 0.8–2.4). The Other 13,084 (4.3) [95% CI 1.1–1.4]), but no increased re- sensitivity analyses showed that the coun- Antidiabetic drugs porting of infections was found for sulfo- try from which the case reports originates involved nylurea derivatives (ROR 1.2 [95% CI and the type of reporter did not have a Monotherapy 288,434 (94.4) 1.0–1.4]), combination therapy of two major effect on the results (data not Biguanides 21,763 (7.1) oral antidiabetic drugs (ROR 1.0 [95% shown). The point estimates changed Sulfonylurea derivative 16,675 (5.5) CI 0.8–1.2]), or for concomitant use of only slightly, but because of the decreased Thiazolidinediones 57,814 (18.9) three or more antidiabetic drugs (ROR numbers the confidence intervals became Dipeptidyl peptidase-4 1.0 [95% CI 0.7–1.4]). Adjustment for wider. inhibitors 8,083 (2.6) the potential confounding factors did Insulin 80,347 (26.3) not affect the results. CONCLUSIONS—This study showed Dual therapy 14,057 (4.6) Assessment of different types of in- that infections were approximately two 2 oral antidiabetics 11,991 (3.9) fections (Table 3) showed an increased times more frequently reported for DPP-4 Biguanide, reporting of URTI (ROR 12.3 [95% CI inhibitors compared with biguanides in sulfonylurea derivative 2,413 (0.8) 8.6–17.5]) for the DPP-4 inhibitors, the WHO Vigibase. In particular, URTIs, Biguanide, whereas the reporting for LRTI, UTI, including nasopharyngitis and sinusitis, thiazolidinediones 1,087 (0.4) and other infections was not increased. were reported more frequently for DPP-4 Sulfonylurea derivatives, In addition, an increased reporting of inhibitors, although the reporting of thiazolidinediones 866 (0.3) URTI was found for users of thiazolidine- URTI was also increased for users of Oral antidiabetic + diones (ROR 2.3 [95% CI 1.7–3.3]), and thiazolidinediones, insulin monotherapy, insulin 2,066 (0.7) for concomitant use of three or more an- and concomitant use of three or more Triple therapy 2,924 (1.0) tidiabetic drugs (ROR 2.5 [95% CI 1.3– antidiabetic drugs, but to a much lesser Concomitant medication 4.7]). Slightly increased RORs were found extent than for the DPP-4 inhibitors. Antibiotics (J01) 5,072 (1.7) for the use of insulin monotherapy and A hypothesis resulting from the cur- Immunosuppressants the reporting of URTI (ROR 1.5 [95% rent study is that the effect of DPP-4 (L04) 2,013 (0.7) CI 1.1–2.2]) and UTI (ROR 1.7 [95% CI inhibitors results in a slight imbalance of Corticosteroids for 1.1–2.5]). The reporting of LRTI (ROR the immune system that causes an in- systemic use (H02) 4,527 (1.5) 1.8 [95% CI 1.4–2.3]) and other infec- creased risk of common, less severe in- All data except patient age are shown as number (%). tions (ROR 1.7 [95% CI 1.4–2.0]) was fections such as (viral) upper respiratory also increased for insulin monotherapy. infections. This is supported by the re- For the combination of an oral antidia- sults of the pivotal randomized clinical betic drug and insulin, increased report- trials that also reported increased num- Table 2 reports the RORs of infections ing of UTI (ROR 3.5 [95% CI 1.7–7.2]) bers of common infections rather than (overall) per antidiabetic drug compared and other infections (ROR 1.7 [95% CI serious infections (1–3). As far as we are with biguanides. The use of DPP-4 inhib- 1.1–2.6]) was noted. The other associa- aware, no studies reporting serious infec- itors as monotherapy (ROR 2.3 [95% CI tions between antidiabetic drugs and in- tions in association with the use of DPP-4 1.9–2.7]), insulins as monotherapy (ROR fections were nonsignificant (Table 3). inhibitors have been reported. At this care.diabetesjournals.org DIABETES CARE, VOLUME 34, FEBRUARY 2011 371 DPP-4 inhibitors and reporting of infections

Table 3—Crude RORs for specific infections

URTI LRTI UTI Other infections Drug N ROR (95% CI) N ROR (95% CI) N ROR (95% CI) N ROR (95% CI) Biguanides 38 Reference 65 Reference 30 Reference 166 Reference SU derivatives 35 1.2 (0.8–1.9) 58 1.2 (0.8–1.7) 35 1.5 (0.9–2.5) 141 1.1 (0.9–1.4) TZDs 233 2.3 (1.7–3.3) 232 1.4 (1.0–1.8) 113 1.4 (1.0–2.1) 367 0.8 (0.7–1.0) DPP-4 I 171 12.3 (8.6–17.5) 20 0.8 (0.5–1.4) 13 1.2 (0.6–2.3) 51 0.8 (0.6–1.2) Insulins 215 1.5 (1.1–2.2) 425 1.8 (1.4–2.3) 186 1.7 (1.1–2.5) 1,015 1.7 (1.4–2.0) OAD + OAD 33 1.6 (1.0–2.5) 33 0.9 (0.6–1.4) 29 1.7 (1.0–2.9) 65 0.7 (0.5–0.9) OAD + insulin 4 1.1 (0.4–3.1) 8 1.3 (0.6–2.7) 10 3.5 (1.7–7.2) 27 1.7 (1.1–2.6) $3 ADs 13 2.5 (1.3–4.7) 8 0.9 (0.4–1.9) 6 1.5 (0.6–3.6) 16 0.7 (0.4–1.2) ADs, antidiabetic drugs; DPP-4 I, dipeptidyl peptidase inhibitors; OAD, oral antidiabetic drug; SU, sulfonylurea; TZDs, thiazolidinediones.

time, the magnitude of the effects of DPP- or decreased reporting (physicians do not Another explanation for our results 4 inhibitors on the immune system may report ADRs that are already mentioned can be that diabetes itself, and its pro- not be compared with the magnitude of in Summaries of Product Characteristics) gression, are often associated with an the effects as seen, for example, with bio- might have occurred. In absolute terms, increased risk of infections. Although logic agents, resulting in rather serious the number of reported infections is low: the literature is not yet conclusive on infections such as tuberculosis or histoplas- only 3% of the cases for DPP-4 inhibitors this, we could not exclude this possible mosis due to tumor necrosis factor-a an- reports involved an infection (242 reports association; therefore, the current study tagonists (19,20). With the current data, of infection of 8,083 case reports). was limited to case reports of antidiabetic however, it was not possible to further dif- Furthermore, differences in classifica- medicines only, thereby eliminating the ferentiate between infections of different tion strategies or misclassification may effect of the disease itself. Some studies, nature and viral, bacterial, or fungal causes. have occurred by the translation from however, suggest that more severe diabe- The strength of this study is that the clinical terminology to the classification tes itself is also associated with higher risk WHO VigiBase allows studying the asso- systems used by the WHO Vigibase of infections (14,22), although this is not ciation between use of antidiabetic drugs (WHO-ART or MedDRA terms). We do supported by strong evidence. Because and infections outside the highly controlled not expect, however, that this misclassifi- disease severity is therefore possibly asso- environment of clinical trials. Neverthe- cation is different for different antidiabetic ciated with both exposure and outcome, less, some limitations of this study need medicines, and the effect of this nondiffer- this can be a confounding factor. to be addressed: ential misclassification on the results of DPP-4 inhibitors are indicated as a First, besides the known issue of this study will therefore be limited. In second- or third-line therapy in combi- underreporting in spontaneous reporting several case reports, more than one in- nation with other oral antidiabetic drugs systems (21), the reporting pattern of fection term was recorded, mainly con- according to treatment guidelines in dif- ADRsmaydifferbetweennewandold sisting of a specific infection term (e.g., ferent parts of the world (23,24). There- drugs, with the most vigorous monitoring nasopharyngitis) and an a-specific term fore, patients who are treated with DPP-4 at the time of marketing and shortly there- (e.g., URTI), which might influence the inhibitors may in general be more se- after, as described by Weber (21). The results. However, our analysis at the case- verely ill compared with patients being DPP-4 inhibitors and thiazolidinediones report level showed that the overall results treated with, for example, biguanides or were introduced in the study period did not change. sulfonylurea derivatives. In the current (1999 through 2009), which might ex- Unfortunately, we were not able to study, however, most of the case reports plain the relatively large number of re- analyze the risk of infections for combi- indicated a DPP-4 inhibitor was the only ports for those drugs. However, it is nation therapy on an individual drug- antidiabetic drug, which indicates mono- unknown whether the type of ADRs that drug combination level owing to the low therapy with DPP-4 inhibitors. A ques- are reported changes over time and how number of cases reporting each possible tion prompted by this study is whether this affects the results of this study. None- combination therapy. This is probably the patients treated with DPP-4 inhibitors theless, adjustment for year of reporting becausecomedicationisonlypoorlyre- monotherapy were indeed the more se- did not affect the results. ported in most case reports. verely ill patients, because one antidia- Second, the results of this study may Finally, we reasoned that the different betic drug was sufficient for these be subject to reporting bias, because indications for the DPP-4 inhibitors in the patients. In addition, for users of a com- infections are listed in both the EU and U.S. versus the rest of the world, and a bination of an oral antidiabetic and in- U.S. Summaries of Product Character- possible effect of the type of reporter sulin or combination therapy of two istics for all three DPP-4 inhibitors. This (health care professional, consumer, or antidiabetics (i.e., usually the more se- may have led to differential monitoring industry) on the case reports, might have verely ill patients), we did not found an and reporting of infections for the DPP-4 influenced the results of this study. How- increased risk of infections. This is in line inhibitors compared with other antidia- ever, excluding reports from the U.S. and with the studies that did not find an as- betic drugs. Increased reporting (physi- excluding reporting from consumers and sociation between diabetes mellitus and cians report ADRs that are likely to occur) industry did not affect the results. infection.

372 DIABETES CARE, VOLUME 34, FEBRUARY 2011 care.diabetesjournals.org Willemen and Associates

Besides infections in general, the as- In conclusion, the results of this study 4. Reinhold D, Biton A, Goihl A, et al. Dual sociation between diabetes and UTI has show that there is an increased reporting inhibition of dipeptidyl peptidase IV and also been described (12,14,15,22), al- of infections for users of DPP-4 inhibitors aminopeptidase N suppresses inflamma- fi fi tory immune responses. Ann N Y Acad Sci though we did not nd signi cantly in- compared with users of other antidiabetic – creased reporting of ADRs for UTIs and drugs. Although the limitations of spon- 2007;1110:402 409 5. Ansorge S, Bank U, Heimburg A, et al. use of any of the antidiabetic drugs. Only taneous reporting systems (e.g., under- Recent insights into the role of dipeptidyl for combination therapy of an oral antidi- reporting, the Weber -effect, reporting fi aminopeptidase IV (DPIV) and amino- abetic and insulin did we nd a slightly bias) should be taken into account, physi- peptidase N (APN) families in immune increased reporting for UTI. Channeling cians and patients nevertheless should functions. Clin Chem Lab Med 2009;47: of this combination therapy toward the remain vigilant on the occurrence of 253–261 more severely ill diabetes patients, for infections and continue to report infec- 6. Thompson MA, Ohnuma K, Abe M, which the use of insulin combined with tions as possible ADRs. Infections may be Morimoto C, Dang NH. CD26/dipeptidyl an oral antidiabetic drug is a marker (25), related to diabetes, but a direct effect of peptidase IV as a novel therapeutic target fi for cancer and immune disorders. Mini may possibly explain this nding. This is the medication on the occurrence of – in line with studies suggesting that the infections should be considered. Rev Med Chem 2007;7:253 273 7. Williams-Herman D, Engel SS, Round E, severity of the disease may play a role in et al. Safety and tolerability of sitagliptin in the occurrence of UTI (14,22). Unfortu- — clinical studies: a pooled analysis of data nately, the subset of the VigiBase used Acknowledgments All authors declare no conflict of interest relevant to the subject from 10,246 patients with type 2 diabetes. in this study did not contain information matter or materials discussed in the article. BMC Endocr Disord 2010;10:7 on the severity of the underlying disease, The division of Pharmacoepidemiology and 8. Ligueros-Saylan M, Foley JE, Schweizer A, so the effect of this phenomenon in the Clinical Pharmacology employing authors Couturier A, Kothny W. An assessment current study remains unclear. M.J.W., A.K.M.-T., H.G.L., and T.C.E. has re- of adverse effects of vildagliptin versus Postmarketing evaluation of safety ceived unrestricted funding for pharmacoepi- comparators on the liver, the pancreas, the concerns raised during the preregistration demiological research from GlaxoSmithKline, immune system, the skin and in patients phase of medicines is import for the the private-public funded Top Institute Pharma with impaired renal function from a large assessment of the benefit-risk balance of (www.tipharma.nl, includes co-funding from pooled database of Phase II and III clinical universities, government, and industry), the trials. Diabetes Obes Metab 2010;12:495– drugs. This study adds to the knowledge 509 of this specific safety issue for the DPP-4 Dutch Medicines Evaluation Board, and the Dutch Ministry of Health. No other potential 9. European Union Volume 9A of the rules inhibitors. The results of the study, using conflicts of interest relevant to this article were governing medicinal products in the Eu- data reported to National Pharmacovigi- reported. ropean Union: guideline on pharmaco- lance Centres, are in line with the findings M.J.W. analyzed data, interpreted results, vigilance for medicinal products for human from the clinical trial program that and wrote the manuscript. A.K.M.-T. super- use. European Commission, 2010. Avail- included a much more selected patient vised data analysis, interpreted results, and able from http://ec.europa.eu/health/files/ population. However, more research is reviewed and edited the manuscript. S.M.S., eudralex/vol-9/pdf/vol9a_09-2008_en.pdf needed to further evaluate the clinical and R.H.M., T.C.E., and H.G.L. interpreted results, 10. Duncan BB, Schmidt MI. The epidemiol- regulatory consequences of this finding, contributed to discussion, and reviewed and ogy of low-grade chronic systemic in- edited the manuscript. flammation and type 2 diabetes. Diabetes such as severity of the infections. Technol Ther 2006;8:7–17 As a result of the observed increased This study was presented in abstract form at the 26th International Conference on Pharma- 11. Alexandraki K, Piperi C, Kalofoutis C, risk in RCTs, the Risk Management Plans coepidemiology and Therapeutic Risk Manage- Singh J, Alaveras A, Kalofoutis A. In- for DPP-4 inhibitors also address a possi- ment, Brighton, U.K., 19–22 August 2010. flammatory process in type 2 diabetes: the ble increased risk of infections. The def- The authors are indebted to the national role of cytokines. Ann N Y Acad Sci 2006; inition of the outcome (“infection”)inthe centers contributing data to the WHO In- 1084:89–117 postauthorization safety studies that are ternational Drug Monitoring Programme. 12. Muller LM, Gorter KJ, Hak E, et al. In- being conducted as part of the Risk Man- Patrick C. Souverein of the Department of creased risk of common infections in pa- Pharmacoepidemiology and Clinical Pharma- tients with type 1 and type 2 diabetes agement Plans is, therefore, of particular – importance. Because our study points out cology, Utrecht Institute for Pharmaceutical mellitus. Clin Infect Dis 2005;41:281 288 that several types of infections (URTIs Sciences, Utrecht University, is highly ac- 13. Bertoni AG, Saydah S, Brancati FL. Di- such as sinusitis and nasopharyngitis) knowledged for his help with the data ex- traction. abetes and the risk of infection-related are more frequently reported than others, mortality in the U.S. Diabetes Care 2001; the outcome should not be limited to in- 24:1044–1049 fections in general but should also take References 14. Boyko EJ, Fihn SD, Scholes D, Chen CL, specific types of infections into account. 1. Committee for Medicinal Products for Normand EH, Yarbro P. Diabetes and the In addition, the nonserious infections Human Use. European Public Assessment risk of acute urinary tract infection among seem to be neglected in the postauthori- Report (EPAR) for Januvia. London, UK, postmenopausal women. Diabetes Care – zation safety studies because all studies CHMP, 2009 2002;25:1778 1783 aim to investigate the risk of serious in- 2. Committee for Medicinal Products for 15. Shah BR, Hux JE. Quantifying the risk fections (1–3). Although from a reg- Human Use. European Public Assessment of infectious diseases for people with Report (EPAR) for Galvus. London, UK, diabetes. Diabetes Care 2003;26:510– ulatory viewport the focus on serious CHMP, 2007 513 infections is understandable, the effect 3. Committee for Medicinal Products for 16. Joshi N, Caputo GM, Weitekamp MR, of (recurrent) nonserious infections on Human Use. European Public Assessment Karchmer AW. Infections in patients with the quality of life can also be consider- Report (EPAR) for Onglyza. London, UK, diabetes mellitus. N Engl J Med 1999; able. CHMP, 2010 341:1906–1912 care.diabetesjournals.org DIABETES CARE, VOLUME 34, FEBRUARY 2011 373 DPP-4 inhibitors and reporting of infections

17. Rayfield EJ, Ault MJ, Keusch GT, Brothers 21. Weber JCP. Epidemiology of adverse re- 24. Rodbard HW, Jellinger PS, Davidson JA, MJ, Nechemias C, Smith H. Infection and actions to nonsteroidal antiinflammatory et al. Statement by an American Associa- diabetes: the case for glucose control. Am J drugs. New York, Raven Press, 1984 tion of Clinical Endocrinologists/Ameri- Med 1982;72:439–450 22. Boyko EJ, Fihn SD, Scholes D, Abraham L, can College of Endocrinology consensus 18. Edwards IR, Olsson S. The WHO In- Monsey B. Risk of urinary tract infection panel on type 2 diabetes mellitus: an ternational Drug Monitoring Program. and asymptomatic bacteriuria among di- algorithm for glycemic control. Endocr Amsterdam, Elsevier Science, 2003 abetic and nondiabetic postmenopausal Pract 2009;15:540–559 19. Keane J, Gershon S, Wise RP, et al. Tu- women. Am J Epidemiol 2005;161:557– 25. Turner RC, Cull CA, Frighi V, Holman RR; berculosis associated with infliximab, a 564 UK Prospective Diabetes Study (UKPDS) tumor necrosis factor alpha-neutralizing 23. National Institute for Health and Clinical Group. Glycemic control with diet, sul- agent. N Engl J Med 2001;345:1098– Excellence. Type 2 diabetes - newer agents fonylurea, metformin, or insulin in patients 1104 (update). London, National Institute for with type 2 diabetes mellitus: progres- 20. Hamilton CD. Infectious complications of Health and Clinical Excellence, 2008. sive requirement for multiple therapies treatment with biologic agents. Curr Opin Available from http://www.nice.org.uk/ (UKPDS 49). JAMA 1999;281:2005– Rheumatol 2004;16:393–398 nicemedia/live/12165/44320/44320.pdf 2012

374 DIABETES CARE, VOLUME 34, FEBRUARY 2011 care.diabetesjournals.org