Risk factors of the Clostridium difficile infection in patients with chronic kidney disease

Material and methods. We evaluated the medical records of all patients treated at the Department of Nephrology and Renal Transplantation of the Research and Development Center in the Provincial Specialist Hospital in Wrocław, Poland, between February 2009 and May 2012, who developed diarrhea, abdominal pain and/or fever within 72 h after admission. In patients with these symptoms, an enzyme cassette immunoassay was performed to detect antigens of C. difficile toxins A and B in stool.


Introduction
The Clostridium difficile (C.difficile, CD) infection is the most frequent cause of diarrhea in patients treated with antibiotics who were admitted to hospital. 1 Whereas the course of the disease is mild in most cases, in some instances it may lead to severe dehydration, septic shock or even death. 25][6] In recent years, a number of reports have appeared, indicating an increased incidence and morbidity of CD-associated infections in patients with chronic kidney disease (CKD). 7,8These observations corroborated current clinical experience in our Department, which is why we sought to assess the incidence, morbidity and risk factors of the CD infection in our CKD patients.

Material and methods
We evaluated the medical records of the patients hospitalized in the Department of Nephrology and Renal Transplantation of the Research and Development Center in the Provincial Specialist Hospital in Wrocław, Poland, between February 2009 and May 2012, who during their hospital stay developed symptoms indicating CD-associated enterocolitis.Qualifying symptoms were as follows: diarrhea, abdominal pain and/ or fever within at least 72 h after admission to hospital.In all patients meeting these criteria, a rapid enzyme cassette immunoassay was performed, detecting antigens of toxins A and B of C. difficile in stool (TOX A/B QUIK CHEK ® ; Techlab, Blackburg, USA).The sample material was taken with a spatula to the test tube, transferred to the test cassette and read after 15 min of incubation at room temperature.
Detailed data was accrued, pertaining to the patients' age, gender, concomitant diseases, and pharmacotherapy with the emphasis on antibiotics, PPIs and antidepressants.Laboratory tests were done in the hospital lab, using standard methods, including the measurement of serum creatinine and albumin concentrations.The shortened Modification of Diet in Renal Disease (MDRD) equation was used to calculate the glomerular filtration rate (eGFR).
Numerical data was expressed as means and standard deviations (SD).Statistical analysis was performed utilizing the STATISTICA v. 12 software (StatSoft Inc., Tulsa, USA).Normal distribution verified with the Kolmogorov-Smirnov test enabled the assessment of the differences between the 2 groups with Student's t test, the homogeneity of variations being checked with Fisher's test.In the case of nonparametric distribution, statistical importance of the differences was evaluated with the use of the Mann-Whitney U test.For quantitative data, the χ 2 analysis was done.Statistical significance cut-off level was set at p = 0.05.

Results
A total of 207 patients were enrolled in the study (104 males and 103 females), aged 20-91 years (64.5 years ±15.79),hospitalized at the Department of Nephrology In our study, longer hospitalization time and lower initial serum albumin concentration significantly increased the risk of infection (Table 1).
Moreover, it was demonstrated that the patients who died during the hospital stay not only more frequently tested positive for CD toxins (Table 1), but also had lower serum albumin concentration at admission, were older, were given more antibiotics during hospitalization, and their hospital stay lasted longer (Table 2).The mean morbidity coefficient for the CD infection was 12.5 per 1,000 hospitalizations.
It was not observed in the study group that lower eGFR, body weight, treatment with PPIs, H2-receptor blockers, immunosuppression, or statins, and the presence of diabetes increased significantly the risk of the CD infection.Furthermore, the length of antibiotic therapy and the number of used antibiotics did not increase considerably the risk of infection.
The stratification of CKD patients with symptoms of acute enterocolitis referring to the classes of CKD is presented in Table 3.

Discussion
The prevalence of the CD infection in the results presented herein: 12.5/1000 hospitalizations is strikingly high when compared to the American (9/1000) and European (4.1/1000)populations, respectively. 9,10Similarly, in one of the few Polish publications related to the CD infection in nephrology wards, the prevalence of 9.9/1000 hospitalizations was reported. 8High prevalence of the CD infection in CKD patients of our Department confirms our observations, indicating a higher prevalence of the infection in nephrological patients when compared both to the general population and to patients hospitalized in other wards. 6,11t also indicates that CKD, especially class 5, is an independent risk factor of the CD infection.The duration of hospitalization and lower serum albumin concentration at admission are significantly related to a higher frequency of the CD infection, as demonstrated in our study.Moreover, we documented an increased risk of death in CD-infected patients.In our study, age, prolonged hospitalization time and lower serum albumin concentration were proven as risk factors of death.Some authors underline the importance of malnutrition, especially among older patients as one of the main agents of mortality.Thus, it is important to focus on proper nutritional therapy in patients with the CD infection. 12On the other hand, the study under discussion failed to confirm the importance of many established risk factors of the CD infection, such as prolonged use of antibiotics, PPIs and immunosuppressant drugs. 4,5,10,11This may be attributable to the fact that the control group in our study consisted of CD-negative patients with diarrhea, instead of all persons hospitalized in the analyzed period (the available data on the treatment of these patients was incomplete).
Recent reports concerning the impact of CKD on the prevalence of the CD infection are inconsistent.A higher risk of infection has been observed in chronically dialyzed patients, 7 whereas this relation has not been confirmed in CKD subjects not requiring RRT. 13 At least 1 report showed a higher risk of the CD infection in patients with both acute and chronic renal injury. 14In the present report, it was not documented that reduced eGFR augmented the risk of the CD infection; however, the patients tested positive for CD toxins had on average eGFR lower by 9.7 mL/min/1.73m 2 .Moreover, the investigated group of patients was dominated by persons with class 5 CKD:

Compliance with ethical standards
All procedures applied in the studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee, and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.For this type of study, formal consent was not required.

Table 1 .
Clinical features of patients tested positive and negative for CD (mean values) M -male; F -female; PPIs -proton pump inhibitors; eGFR -glomerular filtration rate; * statistically significant.

Table 2 .
Clinical characteristics of patients deceased during hospitalization (mean values)

Table 3 .
Classes of CKD in patients demonstrating the symptoms of acute enterocolitis