Use of culture- and ELISA-based toxin assay for detecting Clostridium Difficile, a neglected pathogen: A single-center study from a tertiary care setting

Introduction

Clostridium difficile is a Gram-positive spore-bearing anaerobic bacillus increasingly associated with both community- and hospital-acquired colitis and diarrhea. It is the most common identifiable bacterial cause of nosocomial diarrhea associated with antibiotic use and one of the most common anaerobic infections.[1] C. difficile-associated diarrhea (CDAD) is a life-threatening disease with an attributable mortality of 6%–15% and up to 25% in frail elderly people.[2] The clinical presentations in increasing order of severity include asymptomatic carriage, colitis without pseudomembrane formation, pseudomembranous colitis (PMC), and fulminant colitis.[3] Damage to the normal colonic flora serves as a prerequisite for infection by the bacterium.[4] The incidence of CDAD is on the rise and a newer ribotype, NAP1/BI/027, has been found to be responsible for multiple recent outbreaks in various parts of the world.[56]

The diagnosis of C. difficile infection includes the detection of toxin A/B in stool specimens by direct enzyme immunoassay (EIA), culture of pathogen from the stool specimens using a selective agar cycloserine-cefoxitin fructose agar (CCFA), tissue culture assay to demonstrate the cytopathic effect of toxin on various cell lines which is seen as cell rounding, detection of glutamate dehydrogenase (GDH) an enzyme produced by C. difficile, and nucleic acid amplification tests for the toxin. Among these, tissue culture assay is considered as the gold standard.[7] However, detection of toxin using ELISA is preferred by diagnostic laboratories because of its ease. ELISA-based cytotoxin assays are reported to have sensitivity and specificity ranging from 50%–90% to 70%–95%, respectively.[8]

Considering the changing epidemiology of CDAD, it is important to diagnose it as early as possible both to prevent transmission as well as for better patient management. With few reports from India on this disease,[910111213] the present study was planned to throw more light on the prevalence and utility of laboratory diagnostic methods for CDAD.

Materials and Methods

After obtaining permission from the Ethics Committee (No-EC/154/2011), a prospective case–control study was carried out, from January 2012 to December 2013 in a tertiary care hospital. Assuming that C. difficile infection rate is 10% among the patients with antibiotic-associated diarrhea (AAD) and alpha error 5%, we needed to enroll 144 cases. Further assuming that 5% of the samples were contaminated, so it was decided to enroll 150 cases along with 50 controls. We could not add more controls due to financial constraints for the study. Medicine and allied departments and pediatric departments were requested to send stool samples from patients who satisfy inclusion criteria after taking written informed consent of study patients or their parents/guardians and controls. Furthermore, the investigator visited different wards to identify patients fitting inclusion criteria. All patients from medical and pediatric wards of any age/gender were included who satisfied the following inclusion criteria: (1) history of antibiotic use either in the previous month or recently since 5 days and who developed diarrhea and (2) patients with PMC detected on lower gastrointestinal endoscopy referred for C. difficile detection and no other recognized etiology of diarrhea.[14] Controls were those patients admitted during the study period who had taken antimicrobials for at least 5 days but did not develop diarrhea. Diarrhea was defined as six watery stools over 36 h or three unformed stools in 24 h for 2 days or eight unformed stools over 48 h.[15] Patients who had diarrhea during the first 72 h of admission in a hospital, neonates, and psychiatric patients were excluded from the study. A detailed study pro forma was filled up for each one of them, which included age, gender, severity of diarrhea[16] with duration, association with other symptoms such as abdominal pain, fever, any other illness, ward and unit of admission, antibiotic history, other significant laboratory investigations, duration of hospital stay, and provisional diagnosis.

Results

One hundred and fifty patients who gave a history suggestive of antibiotic-associated diarrhea were included in the study. Of these, 31 were children (<12 years and minimum age 4 years) and 119 were adults. Fifty age- and gender (for adult group)-matched controls were also taken from the same hospital setting. Out of 150 patients in the study group, C. difficile was isolated from the stool of four patients (Three adults and one child). Thirteen fecal samples tested positive for toxin A + B by EIA (11 adults and two children). All the four samples tested positive by culture were found to carry C. difficile toxin by ELISA. No fecal sample from the control group was positive for C. difficile either by culture or ELISA. No additional yield was obtained by enrichment with RCM. No additional yield was obtained by reincubation after 48 h.

Main presenting features of both Groups A and B are listed in Table 1.

Table 1 Main presenting clinical features of patients in the two antibiotic-associated diarrhea groups

Age and gender distribution in the two study groups and controls was analyzed. Maximum cases were from 31 to 45 years age group, males (nine) were more than females (four).

Evaluation of stool samples for routine parameters was done as shown in Table 2.

Table 2 Evaluation of stool specimens for routine parameters

Of 13 positive cases of C. difficile, two patients died. Mortality was attributed to chronic renal failure in one patient and septicemia in another. Seven patients responded on stopping the initial antibiotic which included a third-generation cephalosporin and/or an aminoglycosides or clindamycin. Two patients responded successfully to treatment with oral metronidazole and two more patients required an additional oral vancomycin therapy.

Discussion

The present study on 150 patients with AAD and 50 controls was carried out to determine the prevalence of CDAD using culture and toxin assay. In the present study, 8.67% of suspected AAD cases were either culture positive or toxin assay positive for C. difficile. The culture positivity rate was 3.34%. Low culture positivity rates have been documented in other studies as given in Table 3. The higher rates in some of the studies may be attributed to a low sample size[21] or due to the bias of results with a study done on children of age group of 5–12 years as participants.[18] Children are reported to have higher colonization rates of C. difficile.[1]

Table 3 Culture and ELISA positivity rates of Clostridium difficile in various studies

Lower culture positivity rates can be due to delay in sample transportation to the laboratory, inefficient management of anaerobiosis due to repeated subculture of the isolate which leads to loss of viability, C. difficile being overgrown by many other microorganisms on CCFA, and dilutional effects of diarrhea as culture is dependent on the presence of spores or viable vegetative cells.[24] Culturing of nondiarrheal stools also leads to false-negative results.[25]

In the present study, the specimens were transported and processed within an hour of collection. Anaerobiosis was maintained during incubation of the specimens according to the established standards, repeated subculture of the isolates was avoided, and all the stool specimens were diarrheic. Hence, the lower isolation rate in the present study may be due to sampling error inherent to uneven distribution of C. difficile in the fecal samples or other unexplained reasons. Only one stool specimen per patient was processed. Inclusion of multiple stool specimens may have increased the yield but was not attempted due to cost constraints.

Currently, stool culture is the method least employed by hospitals owing to limitations of cost and turnaround time of approximately 48 h. No additional yield was obtained after incubating for further 48 h or by enriching it on RCM media. The turnaround time in the present study was also 3 days. In addition, the accuracy of this method varies considerably in different laboratories because the methods and culture media are not standardized. The culture media used in the present study were CCFA. Various other media and their recovery rates, which have been evaluated for the isolation of C. difficile include (Chrome–ID C. difficile agar 98%), CLO (C. difficile selective agar 97%), BBL (B. D. chrome agar for C. difficile, 97%), (Cefsulodin Cycloserine Egg Yolk [CCEY] Agar, 98%), Oxoid C. difficile selective agar (98%), and CCEY/L (CCEY Agar with lysozyme 96%.[26] CCFA has been used in the present study because it is effective, economical, and readily available selective media.

Infectious Disease Society of America (IDSA) currently recommends that a positive stool culture followed by identification of a toxigenic isolate, as performed by an experienced laboratory, provides the standard against which other clinical test results should be compared (SHEA IDSA).[14] In the present study, 8.67% of stool samples were positive for C. difficile toxin A + B. All the cultured isolates were toxigenic. The ELISA positivity rates range from 10% to 17% in various Indian studies[2728] and up to 30%[29] in Western literatures. ELISA has a good specificity; however, 100–1000 pg of both toxin A and toxin B must be present for the test to be positive.[27] The kit used in the present study has the capacity to detect approximately 2 ng/ml of toxin A and 3 ng/ml of toxin B. Hence, a false-negative rate of 10%–20% may occur.[25] A false-negative result can also be caused by proteolytic decomposition of the toxins due to inappropriate storage of the sample.[14] In the present study, as the ELISAs were run in batches, the stool specimens were stored at −20°C immediately after culturing, and hence the possibility of toxin decomposition is negligible. Toxin assay has the advantage of lower turnaround time and ease of performance; however, the high cost per single test may necessitate batching of samples due to usage of other kit components such as sample diluent, wash buffer, and controls. The cost per test for toxin assay in the present study was approximately Rs 200/. This does not include the cost of equipment and workforce but only that of the consumables. As per the kit literature, the assay has a limitation of not giving accurate results on a concentrated sample. In the present study, samples were not concentrated. Furthermore, multiple samples over time are indicated to improve yield, but this could not be achieved due to difficulty in follow-up and cost constraints.

Prevalence of CDAD is around 2%–4% in patients without diarrhea and 7%–30% in patients with diarrhea in different hospital-based studies.[17182127] In the present study, the prevalence of C. difficile was 8.67% in hospitalized diarrhea patients and 0% in nondiarrhea controls. All the cases and the controls were on antibiotics. Gupta and Yadav[30] have reported C. difficile isolation rate of 25.3% in hospitalized patients with diarrhea and 4.3% in controls admitted for other ailments. Niyogi et al.[10] have reported 4% in hospitalized patients with diarrhea and 2.7% in nondiarrhea controls. Bhattacharya et al.[27] isolated C. difficile as a sole pathogen from 7.3% of 233 patients with acute diarrhea. Vaishnavi et al.[28] reported 30% positivity for C. difficile toxin in hospitalized patients of all age groups receiving single to multiple antibiotics for various diseases, but only in 7% of patients not receiving antibiotics. Some recent studies estimated a prevalence rate of 10%,[20] 14%,[18] and 17%.[13] The isolation of C. difficile in nondiarrhea controls in other studies may be related to colonization. Colonization by C. difficile in asymptomatic adults depends on the presence of long-standing disease, contact with suspected patient of CDAD, and length of hospital stay which increases the chances of contact with spores.[1] Low carriage rates in asymptomatic adults in the present study may be due to very low number of CDAD patients, thereby minimizing exposure risk, inclusion of nondiarrheal controls, and incorporation of all age groups rather than only pediatric population which shows high carriage rate.

Among patients positive for C. difficile, fever, abdominal pain, and diarrhea have been reported to be more common.[31] In the present study, abdominal pain and abdominal distension were significant predictors of the disease (P = 0.03), whereas severe diarrhea and fever were not. Thompson et al.[29] had reported fever, abdominal pain, and distension as the predominant symptoms. Similar results have been documented in other studies.[1718] Liquid stool (P = 0.005) and Gram-positive bacilli with oval subterminal spores (P < 0.001) were also considered as significant predictors of disease.

CDAD has been reported to be more common in women and older patients.[12] Studies from India have reported varying male-female ratios. In the present study, among 13 positive cases, nine were males (60.9%) and four were females. A maximum number of positive cases were found in the age group of 30–45 years followed by those more than 45 years of age. Similar male preponderance and higher age association have been reported in other studies from India.[122120] The increased risk of acquiring C. difficile infection in the elderly may be due to age-related changes in fecal flora, immune senescence, impaired ability of neutrophils to phagocytose and kill C. difficile and decrease in the capacity of serum to neutralize toxins with increasing age, or the presence of other underlying diseases.

The present study had certain limitations. Due to cost constraints, the sample size of the controls was restricted to 50. Due to the absence of facilities for tissue culture, cytotoxin assay for C. difficile which is considered the gold standard could not be carried out.

Since other reports highlight the increasing prevalence of C. difficile as the causative agent of AAD, it is important to lay some protocol for correct isolation of the organism in cases of AAD. To increase the yield of laboratory detection, one potential strategy recommended by IDSA is a two-step method that uses EIA detection of GDH as initial screening test followed by the cell cytotoxicity assay or toxigenic culture as the confirmatory test for GDH-positive stool specimens only. As reported in literature, antibiotic-associated C. difficile diarrhea is posing a new challenge to health-care settings with its changing epidemiology and emerging hypervirulent strain NAP1. Unfortunately, it is still a neglected diagnosis due to tedious anaerobic methods, costly serological techniques, and nonavailability of tissue culture laboratory.

Conclusion

Culture- and toxin based assays may not detect all cases of CDAD. Based on the results of the present study, culture does not provide any additional yield over toxin assay. Better diagnostic modalities would be required to prove CDAD.