Clostridium difficile is an important pathogen in healthcare facilities. Colonized or infected patients and spore-contaminated environments have been identified as sources for C. difficile infection (CDI). Patients generally develop CDI after exposure to broad-spectrum antibiotics.Reference Pires, Monteiro and Carneiro1, Reference McDonald, Gerding and Johnson2
The incidence of CDI in Latin America is likely to be underestimated due to low clinical suspicion as well as limited availability (and low sensitivity) of diagnostic tools.Reference Pires, Monteiro and Carneiro1, Reference Balassiano, Yates, Domingues and Ferreira3 Here we report the results of a large survey conducted to determine the frequency of diarrhea and CDI in hospitalized patients in Brazil.
Methods
This point-prevalence study involved adult patients (aged ≥18 years) with diarrhea admitted to 8 university hospitals in Brazil. Hospitals were located in 3 Brazilian state capitals: São Paulo, Curitiba, and Porto Alegre.
The study was conducted on 2 distinct dates: March 8, 2017 (summer), and July 12, 2017 (winter). Clinical and demographic data were collected for each patient, including date of onset of current episode of diarrhea, underlying diseases, and antimicrobial use (up to 30 days prior to hospitalization). Patients were excluded if they had been hospitalized in emergency rooms, pediatric wards, and dialysis units. The study was approved by the local ethics committees of the participating hospitals.
Stool samples were obtained from each enrolled patient. Samples were refrigerated at 4°C and sent to the reference laboratory within 24 hours (ie, the Molecular Biology Laboratory at Santa Casa de Misericordia de Porto Alegre). Only 1 fecal sample per patient was collected.
Culture for C. difficile was performed on fecal samples as follows. Samples were treated with absolute alcohol (1:1 proportion) at room temperature for 1 hour and subcultured on CM0601 C. difficile agar (Oxoid, Ontario, Canada), enriched with 7% blood horse, D-cycloserine and cefoxitin. The culture was incubated for 48 hours using an anaerobic generator (Genbox, bioMérieux SA, Marcy l’Etaile, France). Suspected colonies were identified at the species level by matrix-assisted laser desorption/ionization mass spectroscopy (MALDI-TOF/MS, Brucker Daltonics, Germany).
All fecal samples were investigated for the presence of toxin B (tcdB), binary toxin (cdtA), and deletion of 117 nucleotides on the tcdC gene using a commercial real-time polymerase chain reaction (PCR) kit (Xpert C. difficile test, Cepheid, Sunnyvale, CA) according to the manufacturer’s recommendations.Reference Kok, Wang, Thomas and Gilbert4
All patients with diarrhea and positive results for real-time PCR or culture plus MALDI-TOF were considered confirmed CDI cases. Statistical analyses were performed using JMP version 13.0.0 software (SAS Institute, Cary, NC).
Results
In the 2 days of study, we screened 6,374 patients and 153 presented with diarrhea. The point prevalence of diarrhea was 24.0 per 1,000 patient days (95% confidence interval [CI], 20.5–28.1).
Anaerobic culture was positive for 19 patients, 17 of whom had C. difficile confirmed by MALDI-TOF-MS. GeneXpert was positive for 12 of 153 patients. Therefore, CDI was confirmed in 19 subjects, with a point prevalence 3.0 per 1,000 patient days (95% CI, 1.9–4.8). The overall proportion of patients with CDI among screened patients with diarrhea was 11.8% (95% CI, 7.6–17.8).
The point prevalence of CDI among diarrhea patients did not differ between the 2 study periods, with 12 CDI cases (7.8%) in the summer and 7 cases (4.8%) in the winter. The characteristics of CDI patients and patients with diarrhea with other causes are detailed in Table 1. Median time to sample collection was shorter in CDI patients compared to non-CDI patients (8 vs 19 days; P = .041). Exposure to fluoroquinolones was the only variable associated with CDI risk (relative risk, 3.13; 95% CI, 1.07–9.19). Furthermore, 1 patient (5.2%) presented a hypervirulent C. difficile strain according to real-time PCR.
Table 1. Characteristics of Patients with Diarrhea in the Hospital: Comparison of Patients with Clostridium difficile infection (CDI) and Individuals with Diarrhea Due to Other Causes
a Unless otherwise specified.
Discussion
To the best of our knowledge, this is largest point-prevalence study conducted in Latin America to determine the prevalence and risk factors for CDI among hospitalized patients. In 2018, a study in a Brazilian university hospital compared the clinical characteristics among hospitalized adult patients who received antimicrobial treatments and developed CDI. In that study, comorbidity severity indices and the number of antibiotics used during hospitalization were strong independent predictors of nosocomial CDI.Reference Lopes Cançado, Silveira Silva and Rupnik5 In a study carried out in the Northeast Region of Brazil in high-risk cancer patients, a high rate of CDI was also found among inpatients with diarrhea: 23 of 48 patients (48%).Reference Costa, Carvalho and González6
In Latin America, Costa Rica published in 2008 the first study reporting the presence of C. difficile among adults. Samples were collected during a 13-month period, and the frequency of CDI was 31 of 104 (30%).Reference Balassiano, Yates, Domingues and Ferreira3
In our study, fecal culture plus MALDI-TOF were far more sensitive than PCR for the laboratorial diagnosis of CDI. The low sensitivity observed with GeneXpert could be attributed to levels of C. difficile that had fallen below the level of detection by PCR, to nontoxigenic strains, or to the inability of the kit to eliminate potential PCR inhibitors from stool samples.Reference Monot, Eckert and Lemire7
Clostridium difficile has several other virulence factors associated with adhesion and intestinal invasion that were not investigated in our study. Recent studies have shown variations in the genes of the toxins of PaLoc (pathogenicity locus), which can lead to false negative results in the PCR techniques currently used.Reference Monot, Eckert and Lemire7, Reference Merrigan, Venugopal and Roxas8 Another limitation of our study was the lack of a comorbidity score or a disease severity measurement, which may have influenced risk factor determination.
In conclusion, our findings suggest that Brazil has high rates of CDI, and criteria for surveillance and epidemiological monitoring are still not well structured. The use of fluoroquinolone as a risk factor for CDI enhances antimicrobial stewardship practices to prevent C. difficile–related disease.
Author ORCIDs
Alessandro C. Pasqualotto 0000-0002-6782-5395
Acknowledgments
We would like to thank all members of the research teams involved in this study, particularly and medical nurse students from Universidade Federal de Ciências da Saúde de Porto Alegre.
Financial support
This study was sponsored by an unrestricted research grant support from Merck Sharp & Dohme to Dr. Pasqualotto (grant no. MISP 53854).
Conflicts of interest
All authors report no conflicts of interest relevant to this article.
