lable at ScienceDirect

American Journal of Infection Control 42 (2014) 459-60

Contents lists avai

American Journal of Infection Control

journal homepage: www.aj ic journal .org

American Journal of Infection Control

Brief report

Clostridium difficile infection in Brazil: A neglected problem?

Renata N. Pires RN a,b, Alexandre A. Monteiro BSc a, Lilian C. Carneiro PhD a,Ludmila F. Baethgen PhD a, Rejane Tavares PhD c, Carla S. Lincho MDd,Steven Park e, David Perlin PhD e, Edison M. Rodrigues Filho MDb,d,Alessandro C. Pasqualotto MD, PhD a,b,*a Federal University of Health Sciences of Porto Alegre, Porto Alegre, RS, Brazilb Santa Casa Hospital Complex, Porto Alegre, RS, Brazilc Federal University of Pelotas, Pelotas, RS, BrazildConceição Hospital Group, Porto Alegre, RS, Brazile Public Health Research Institute, New Jersey Medical School, Newark, NJ

Key Word:Epidemiology

* Address correspondence to Alessandro C. PasqBiology Laboratory, Santa Casa de Misericordia dependência 155, 8� andar, 90035-075, Porto Alegre, Br

E-mail address: pasqualotto@ufcspa.edu.br (A.C. PSupported by an independent research grant by M

Investigator Study Program IIS No. 38784, USA).Conflicts of interest: None to report.

0196-6553/$36.00 - Copyright � 2014 by the Associahttp://dx.doi.org/10.1016/j.ajic.2013.10.012

The epidemiology of Clostridium difficile infection is virtually unknown in Brazil. In this prospective study,8.3% of patients with nosocomial diarrhea were found to have toxigenic strains of C difficile in their feces.The relevant risk factors for Clostridium difficile infection were receipt of solid organ transplantation andseptic shock.

Copyright � 2014 by the Association for Professionals in Infection Control and Epidemiology, Inc.Published by Elsevier Inc. All rights reserved.

Clostridium difficile infection (CDI) is the leading infectious cause A standardized clinical report formwas used for data collection.

of nosocomial diarrhea in the developed world.1 CDI is associatedwith an increased morbidity, prolonged hospitalization, andincreased costs. However, very little is known about the epidemi-ology of CDI in Latin American countries.2 The incidence of CDI mayactually be underestimated in these areas. Here, we investigatedthe frequency of CDI in a cohort of patients in 2 large hospitals inBrazil. Risk factors for CDI were characterized.

METHODS

This was a multicenter, prospective, observational study inwhich patients admitted to 2 large teaching hospitals locatedwith adistance of 1.5 miles in Porto Alegre, southern Brazil. Hospital 1 is atransplant hospital with 65 beds; hospital 2 is a tertiary care hos-pital with 800 beds, including 40 beds in the intensive care unit.Adult (>18 years old) patients showing � 3 liquid stools over a 24-hour period were considered for inclusion. For hospital 2, onlypatients admitted to the intensive care unit were studied. Partici-pation in the study was conditional on the signature of an informedconsent.

ualotto, MD, PhD, MolecularPorto Alegre, Avenue Inde-azil.asqualotto).erck Sharp & Dhome (Merck

tion for Professionals in Infection C

Clinical and demographic data were obtained, including the pres-ence of comorbidities, previous (last 100 days) hospitalizations andantibiotic use, immunosuppression, receipt of chemotherapy, anduse of acid suppressors. Laboratory results performed on the daypatients entered the study were also recorded.

For the purposes of this study, patients were considered ashaving CDI in the presence of a positive enzyme-linked immuno-sorbent assay (ELISA) test for toxins A and B (Premier Toxins A & B;Meridian Bioscience, Cincinnati, OH), in association with a either apositive culture for C difficile (C difficile agar-CLO; bioMérieux,Basingstoke, UK) or positive real-time polymerase chain reaction(qPCR) test. All tests were performed in fresh fecal samples. DNAwas extracted from stools using a commercial kit (QIAamp DNAStool Mini Kit; Qiagen, Valencia, CA). qPCR tests were performedwith specific primers and probes (molecular beacons) aiming forthe tcdB gene, both provided by the Public Health Research Institute(Newark, NJ). The qPCR assays were performed on an ABI 7500thermocycler (Applied Biosystems, Foster City, CA). All reactionscontained negative (water) and positive controls (a specific 1,185-base pair PCR product of the tcdB gene).

For the purpose of this study, 2 groups of patients werecompared: patients with diarrhea because of C difficile infectionand patients with diarrhea because of other causes. Categoricalvariables were compared with c2 or Fisher exact test. Quantitativevariables were treated with Student t or Mann-Whitney U test. Pvalues of � .05 were considered statistically significant. Variablesassociated with P values of < .20 at univariate analysis were

ontrol and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved.

Table 1Univariate analysis of risk factors for C difficile infection among patients withnosocomial diarrhea

Characteristics Non-CDI, n (%) CDI, n (%) P value

Age, median (range), y 63 (19-94) 62 (19-76) .96Sex, male 53 (60.2) 7 (87.5) .25Admission to the transplant hospital 24 (27.3) 5 (62.5) .05Chronic renal failure on hemodialysis 15 (17.0) 2 (25.0) .63Solid organ transplantation 20 (22.7) 4 (50.0) .10Immunosuppressive therapy 21 (23.9) 4 (50.0) .20Previous antibiotic use 84 (87.5) 5 (62.5) .06Use of opiates 61 (70.1) 4 (50.0) .26Previous surgery 47 (53.4) 7 (87.5) .08Use of H2 receptor antagonists 48 (55.2) 2 (25.0) .14Use of proton pump inhibitors 49 (59.0) 5 (62.5) 1.00Septic shock at presentation 6 (6.8) 3 (37.5) .03

Table 2Multivariate analysis of risk factors for C difficile infection

Variables P value Odds ratio 95% Confidence interval

Shock septic at presentation .01 14.6 2.09-102.13Solid organ transplantation .04 6.1 1.04-35.63

R.N. Pires et al. / American Journal of Infection Control 42 (2014) 459-60460

selected for inclusion in the multivariate model, using Enter se-lection. Data analysis was performed with SPSS, version 19.0 (SPSS,Inc, Chicago, IL).

RESULTS

A total of 96 patients was studied over a 12-month period(2011-2012). Of these, 60 were men (62.5%), with a median age of63 years old (range,19-94). Most patients were in hospital 2 (n¼ 67,69.8%). The median length of hospital stay was 68.5 days (range,7-339). Twenty-four patients were solid organ transplant re-cipients. Twenty-five patients (26.0%) received immunosuppres-sive drugs. The use of proton pump inhibitors was identified in54 (56.3%), and 65 (67.7%) received opiates.

The frequency of CDI in this cohort was 8.3% (8/96). Median ageof patients with CDI was 62 years old (range, 19-87), and the ma-jority was male (n ¼ 7, 87.5%). Table 1 shows a comparison of riskfactors for CDI in patients with nosocomial diarrhea. In CDI group,4 patients (50.0%) were solid organ transplant recipients, and 75.0%had undergone a liver transplantation. Only 1 patient (12.5%) hadcommunity acquired CDI. Immunosuppressive therapy were usedin 50.0% (n ¼ 4), and 50.0% (n ¼ 4) also received opiates. Medianlength of hospital stay before the diagnosis of CDI was 35 days.Among transplant patients, the incidence of C difficile was 17.2%(n ¼ 5), whereas, in critically ill patients, the incidence of CDI was4.5% (n ¼ 3). The in-hospital mortality of CDI was 50.0%, in com-parison with 39.8% for CDI-negative patients (P ¼ .418).

Table 2 shows the results for the multivariate analysis for riskfactors for CDI. Independent risk factors for CDI were septic shock atpresentation (P ¼ .007) and solid organ transplantation (P ¼ .045).

DISCUSSION

In this study, the frequency of CDI was 8.3% in patients withhealth care-associated diarrhea. This is the first prospective studyconducted in Brazil aimed to determine the frequency of CDI inadult critically ill patients and transplant recipients. Most of thestudies conducted so far in Brazil included children only3,4 withlimited data being available for adults.5

Even though CDI is a major problem in the hospital setting,several risk factors for community-acquired CDI have been identi-fied.6 Using a standard definition for community-acquired CDI7 (ie,disease occurringwithin 48 hours of hospitalization), only 1 patientin our study was found to satisfy this criterion.

C difficile toxin detection by ELISA remains largely employed as adiagnostic test for CDI, but recent data have shown that qPCR is farmore sensitive in this context.8 Conversely, we observed that thedetection of C difficile toxins by ELISA had a greater clinical

relevance, and it may have a reduced cost, in comparison withqPCR. Among the main findings of this study are the identificationof independent risk factors for CDI, including septic shock at diseasepresentation and receipt of a solid organ transplant. A recent studyanalyzed the inclusion of renal dysfunction or chronic renal failureas a marker of severity, and renal dysfunction was identified as anindependent predictor of mortality in CDI patients.9 Another studyfound that the incidence of CDI in liver recipients ranged between3% and 19%, considerably higher than the expected incidence forother hospitalized patients (ie, below 1%).1

Our study had some limitations. In particular, some cases weremissing during the study mostly because of unawareness of healthcare workers on the actual incidence of CDI in our population,resulting in a delay or no communication of suspicious cases.

Our data reinforce the results of other studies inwhich C difficilewas responsible for a large number of cases of nosocomial diarrhea.The major relevance of the study relies on providing epidemiologicdata for CDI, an infection that has been largely neglected in LatinAmerica. Among patients with diarrhea, the identification of pa-tients at high risk for CDI such as those with septic shock and solidorgan transplant recipients may facilitate the implementation ofinfection control measures until a proper diagnosis of CDI is per-formed. qPCR has demonstrated high sensitivity in screening pa-tients with diarrhea for CDI.

Acknowledgment

The authors thank all of those involved in sample collection andpatient recruitment, particularly Marion W. Trombetta and DanielaBranco.

References

1. Dubberke ER, Burdette SD, AST Infectious Diseases Community of Practice.Clostridium difficile infections in solid organ transplantation. Am J Transplant2013;13:42-9.

2. Graf K, Cohrs A, Gastmeier P, Kola A, Vonberg RP, Mattner F, et al. An outbreak ofClostridium difficile associated disease (CDAD) in a German university hospital.Eur J Clin Microbiol Infect Dis 2009;28:543-5.

3. Balassiano IT, Yates EA, Domingues RM, Ferreira EO. Clostridium difficile: aproblem of concern in developed countries and still a mystery in Latin America.J Med Microbiol 2012;61:169-79.

4. Alcides AP, Brazier JS, Pinto LJ, Balassiano IT, Boente RF, de Paula GR, et al. NewPCR ribotypes of Clostridium difficile detected in children in Brazil. Antonie VanLeeuwenhoek 2007;92:53-9.

5. Balassiano IT, dos Santos-Filho J, Vital-Brasil JM, Nouér SA, Souza CR, Brazier JS,et al. Detection of cross-infection associated to a Brazilian PCR-ribotype ofClostridium difficile in a university hospital in Rio de Janeiro, Brazil. Antonie VanLeeuwenhoek 2011;99:249-55.

6. Trifilio SM, Pi J, Mehta J. Changing epidemiology of Clostridium difficile-associ-ated disease during stem cell transplantation. Biol Blood Marrow Transplant2013;19:405-9.

7. Association for Professionals in Infection Control and Epidemiology. APIC guideto preventing Clostridium difficile infections. New York [NY]: Association forProfessionals in Infection Control and Epidemiology; 2013.

8. Deshpande A, Pasupuleti V, Rolston DD, Jain A, Deshpande N, Pant C, et al.Diagnostic accuracy of real-time polymerase chain reaction in detection ofClostridium difficile in the stool samples of patients with suspected Clostridiumdifficile infection: a meta-analysis. Clin Infect Dis 2011;53:e81-90.

9. Shah DN, Bhatt NS, Welch JK, Koo HL, Garey KW. Defining acute renaldysfunction as a criterion for the severity of Clostridium difficile infection inpatients with community-onset vs hospital-onset infection. J Hosp Infect 2013;83:294-9.