Clostridium difficile is the most common cause of nosocomial diarrhea in the United States. C. difficile infection (CDI) among acute care hospitalized patients ranges between 1 and 10/1000 discharges (0.1%-1%). At the VA Caribbean Healthcare System, the rate of CDI was traditionally less than 2.5%, by the end of 2004, this rate increased to 3.5%. The increasing trend was also associated with an increase in the severity and mortality of the CDI cases. In November 2004, the crude mortality of patients with CDI was as high as 30%. Multiple comorbidity indices are available, but none has been tested to predict mortality in CDI. A prospective study by Pepin et al1 showed an association between complicated CDI with megacolon and shock requiring vasopressor therapy. Leucocytosis was a strong independent factor associated with an increased risk of complicated disease. In this study, fever was common in complicated CDI, although no association was found in multivariate analysis. As reported by Zuckerman et al, severe pseudomembranous colitis is also associated with low albumin gradient ascites.2 More recently, Loo et al3 reported that C. difficile strains with binary toxin genes and the partial deletion in the tcdC gene were associated with increased virulence and high morbidity and mortality. To try to identify the cases with an increased morbidity and mortality, we retrospectively evaluated several of the previously described risk factors in addition to presenting signs, symptoms, laboratory, and radiographic data that could be associated to the increased morbidity and mortality in our study population. The purpose of this study was to develop a Severity Score Index (SSI) that could predict the outcome of patients with CDI.
Records of patients identified with C. difficile toxin positive at the clinical microbiology laboratory for the period of October 1, 2004, to January 31, 2005, were reviewed. Patients were considered as having CDI if presented with acute onset diarrhea and a stool sample positive for C. difficile toxin A/B. For study purpose, we define diarrhea as watery stool for at least 72 hours based on nurse's and primary physician's documentation. Information regarding age, sex, antibiotic use (Fig. 1) and type of antibiotic (Fig. 2), comorbidities (chronic obstructive pulmonary disease, congestive heart failure, malignancy, end-stage renal disease), previous hospitalization within 6 months, feeding tube during hospital stay, proton pump inhibitors (PPI), and histamine-2 blockers within the last 6 months, among others, were reviewed. This information was validated, reviewing the electronic medical record, including progress notes and prescription orders and reffils of medications. A control group was randomly selected from patients with similar comorbidities, documented diarrhea, and negative toxin assay.
Data were analyzed with a developed SSI, which consisted of 9 criteria in the presence of CDI. The criteria considered were as follows: altered mental status, abdominal pain or distention, tachycardia, hypotension, fever, leukocytosis or leukopenia, or increased band forms; ascites or colitis documented by imaging studies; hypoalbuminemia; and admission or transfer to an intensive care unit (Table 1). Mild disease was define as CDI with 1 to 3 criteria, moderate with 4 to 6, and severe with ≥7 criteria. The association between mortality and the developed SSI was evaluated. The data were analyzed using descriptive statitics such as frecuency, percent, means, standard errors, and standard deviations to analyze the characteristics that were measured. χ2 analysis was used to observe possible associations between groups with α < 0.05. Differences in means were estimated by t test, with decision criteria at α < 0.05. SPSS for windows release 13.0.0 (SPSS Inc, Chicago, Ill) was the software used to process the data.
A total of 101 records were reviewed: 51 cases with CDI and 50 controls. All were male with similar comorbidities (Table 2), with a mean age of 77.2 years for CDI cases and 73.5 years for controls. Risk factors with significant association to the development of CDI were as follows: previous antibiotics and PPI use (P < 0.005). The fourth-generation cephalosporin, cefepime, and fluroquinolones demonstrated the strongest association to CDI. No significant association was found with the use of feeding tubes, preexisting comorbidities, or use of H2-blockers. Regarding the SSI, leukocytosis or leukopenia or increased band forms percent were reported in 71% of the patients who developed CDI. Fever and tachycardia were also strongly associated with CDI (P < 0.05). Low albumin and hypotension were more common in the CDI group than in the control group, although no statistical significance was found. Among those with an SSI with less than 3 criteria (mild infection), the CDI mortality was 4% (Fig. 3).For those with moderate infection (4-6 criteria), the mortality rate increased to 45.5%. Patients with severe infection (≥7 criteria) had a mortality rate of 75% (P < 0.05).
Our study confirms the strong association between the use of antibiotics and the subsequent development of CDI (Fig. 1). Cefepime and the fluoroquinolones were the antibiotics most commonly associated CDI (Fig. 2). Contrary to previously published data, we did not find an association between the use of a feeding tube and the development of CDI. There was a statistically significant correlation between the use of PPI and the development of CDI (P < 0.05). Our SSI was able to predict CDI infection severity and mortality based on the following scores: mild infection (1-3 criteria), moderate infection (4-6 criteria), and ≥7 as severe infection. Those patients with a higher SSI were more likely to die while hospitalized. The SSI demonstrated to be a powerful and useful instrument to predict mortality in our patient population. The development of fever, tachycardia, leukocytosis or increased in >10% bands, and the presence of colitis seem to be the most important warning signs and predictors of early mortality in our studied population. Patients with severe infection had a mortality of 75.5%. Limitation of our study was that it is retrospective, more virulent form was not identified, and all our controls and cases were male patients. These findings could not be extrapolated to female patients. A prospective validation study is underway to confirm the predictive power of the SSI for CDI.
The authors thank Mr David Rosario for the statistics expertise.
1. Pepin J, Valiquette L, Alavary M. Clostridium difficile-
associated diarrhea in a region of Quebec from 1991 to 2003: a changing pattern of disease severity. CMAJ
2. Zuckerman E, Kanel G, Ha C, et al. Low albumin gradient ascites complicating severe pseudomembranous colitis. Gastroenterology
3. Loo V, Poirier L, Miller M. A predominantly clonal multi-institutional outbreak of Clostridium difficile
-associated diarrhea with high morbidity and mortality. N Engl J Med
4. Bartlett JG, Chang TW, Gurwith M, et al. Antibiotic associated pseudomembranous colitis due to toxin-producing clostridia. N Engl J Med
5. Aas J, Gessert CE, Bakken JS. Recurrent Clostridium difficile
colitis: case series involving 18 patients treated with donor stool administered via a nasogastric tube. Clin Infect Dis
6. Barlett JG. Antibiotic associated diarrhea. N Engl J Med
7. Gorbach S. Antibiotics and Clostridium difficile
. N Engl J Med
8. Kelly CP, Pothoulakis C, LaMont JT, et al. Clostridium difficile
Colitis. N Engl J Med
9. Mandell GL, Bennett J, Dolin R, et al. Principles and Practice of Infectious Diseases
. 6th ed. 2005:1249-1262.
10. Gaynes R, Rimland D, Killum E, et al. Outbreak of Clostridium difficile
infection in a long term care facility: association with Gatifloxacin use. Clin Infect Dis
11. Johnson S, Samore MH, Farrow KA, et al. Epidemics of diarrhea caused by clindamycin-resistant strain of Clostridium difficile
in four hospital. N Engl J Med
© 2008 Lippincott Williams & Wilkins, Inc.
12. Mayfield J, Leet T, Miller J, et al. Effectiveness of environmental and infection control programs to reduce transmission of Clostridium difficile
. Clin Infect Dis