SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgement
  9. References

Aliment Pharmacol Ther31, 754–759

Summary

Background  Proton pump inhibitors (PPIs) and H2 receptor antagonists (H2RAs) may play an important role on the onset of Clostridium difficile-associated disease (CDAD) in adults. The impact of Clostridium difficile on children treated with gastric acid-suppressing agents remains unknown.

Aim  To investigate the relationship between CDAD and exposure to acid suppressive therapy in hospitalized paediatric patients.

Methods  We reviewed the medical records of children, with a diagnosis of protracted diarrhoea and abdominal pain, whose stool was analysed for C. difficile toxins. We identified 68 patients with CDAD. For each patient, we randomly selected one control subjects with stool analysis negative for C. difficile. Comorbid illnesses, previous hospitalizations, antibiotics, corticosteroids, immunosuppressants and gastric acid suppressing exposures were recorded.

Results  The use of PPI was significantly higher in C. difficile positive group compared with C. difficile negative group [odds ratio (OR): = 4.5; 95% confidence interval (CI) = 1.4–14.4]. We also found a trend for the use of H2RAs in patients infected by C. difficile compared with C. difficile negative comparison group (OR: = 3.8; 95% CI = 0.7–18.9).

Conclusions  Children exposed to PPIs therapy seem to be at higher risk for the development of Clostridium difficile-associated disease.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgement
  9. References

The prevalence of Clostridium difficile infection is in the process of changing.1Clostridium difficile is an anaerobic spore-forming bacillus that produces pathogenic exotoxins, including Toxin A and Toxin B. Disease transmission occurs in humans by the ingestion of bacteria by a susceptible host. Following colonization with toxigenic C. difficile, individuals may become asymptomatic carriers2 or develop colonic disease. This process is supposed to be mediated by the spore form of the pathogen. Clinical features of C. difficile-associated disease (CDAD) can range from mild, self-limited diarrhoea, to severe colitis with cramping, haematochezia, pseudomembrane formation and intestinal perforation.3, 4 In adults, a recent study demonstrates a mortality rate of over 60% in patients with a complication suggesting that patients destined to do poorly are being recognized too late or are treated with inadequate treatment.5 Over the last decade, the incidence of CDAD has progressively increased in hospital and community settings in North America and Europe.1, 6 Possible contributing factors could be the emergence of more virulent strain, increased use of fluoroquinolone antibiotics and the development of resistance to antibiotics.6, 7 Factors associated with altered enteric flora increase the risk of C. difficile colonization.8 In addition to established risk factors, such as exposure to antibiotics, age, severe underlying illness, hospitalization, and immunosuppressants,9 several recent studies have reported an association between the use of gastric acid-suppressing agents and C. difficile infection in both hospitalized adult patients and the general population.10–14 Gastric acidity constitutes a major defence mechanism against ingested pathogens, and loss of normal stomach acidity has been associated with colonization of the normally sterile upper gastrointestinal tract. Proton pump inhibitors (PPIs) and H2 receptor antagonists (H2RAs) are potent drugs suppressing gastric acid secretion; thus, they could be implicated as a cause of altered intestinal flora of the upper gastrointestinal tract and lead to complications such as malabsorption, enteric infections and infections outside the gastrointestinal tract.15 At present, the impact of C. difficile on paediatric patients treated with gastric acid-suppressing agents is still unknown. We therefore assessed whether the use of PPIs and H2RAs is associated with an increased risk of C. difficile infection in paediatric population.

Patients and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgement
  9. References

We performed a retrospective, single-centre, observational, case-control study. We identified and reviewed the medical records of all children aged 1–18 years who were admitted to the Department of Pediatrics, University of Naples ‘Federico II’ from June 2005 to July 2009 with a diagnosis of protracted diarrhoea and with abdominal pain whose stools were analysed for C. difficile Toxin A and/or Toxin B. The diagnosis of protracted diarrhoea was based on a decrease in consistency and an increase in frequency of bowel movements to ≥3 per 24 h, persisting for more than 15 days. Data related to C. difficile infections were obtained from the Microbiology Laboratory of our Hospital. Stool samples were also screened for other bacterial, viral and parasitic species. The total number of patients who underwent stool analysis for C. difficile during the study period was 910. We identified 68 patients with CDAD as diagnosed by the evidence of C. difficile toxins in the stools and presence of protracted diarrhoea and abdominal pain. For those patients who were admitted multiple times for CDAD infection, we only analysed data from their first hospital admission for the purposes of this study. The date of hospital admission for CDAD served as the index date for analysis. All patients infected by C. difficile were affected by comorbid illnesses: 43 patients by Inflammatory Bowel Disease (IBD) and 25 presented other disorders (food allergy, gastro-oesophageal reflux disease, functional gastrointestinal disorders, genetic diseases). For each patient, we randomly selected one control subject from the remaining 842 patients with the same symptoms, but with stool analysis negative for C. difficile toxins and for other pathogenic species, thereby obtaining a comparison group of 68 children. In the comparison group, children presented the following comorbid illnesses: 31 were affected by IBD, 37 presented other disorders (food allergy, gastro-oesophageal reflux disease, functional gastrointestinal disorders, genetic diseases). Control subjects were assigned the same index date as that of their corresponding case-patient.

Demographic characteristics including age, gender, diagnosis on admission and comorbid conditions were collected. Exposure to risk factors for C. difficile infection in adults such as use of antibiotics, PPIs, H2RAs, corticosteroids, mesalazine (mesalamine), immunosuppressants (azathioprine and methotrexate) and hospitalization within 3 months from the time C. difficile detection were investigated.

The primary outcome of this study was to estimate the relative risk of CDAD associated with acid suppressive therapy in children. Secondary outcomes were to determine the relative risk of other factors already known to be associated with CDAD in adults, including antibiotics, immunosuppression and comorbidities.

The study was approved by the Institutional Review Board of the University of Naples, ‘Federico II’.

Microbiological methods

In this study, human stool specimens were analysed for the detection of C. difficile toxins by ImmunoCard Toxins A & B (Meridian Bioscience, Cincinnati, OH, USA), a rapid, qualitative, horizontal-flow enzyme immunoassay (EIA). The assays were performed according to the manufacturer’s instructions. This enzyme immunoassay has a sensitivity of 83% ± 6.7% and a specificity of 95 ± 1.6% for C difficile infection screening.

Statistical analysis

Variables were screened for their distribution and appropriate parametric or non parametric tests were adopted as required. Cross-tabulations were evaluated by using the Fisher test and χ2test. Statistical significance was predetermined as < 0.05. To measure the association between CDAD in hospitalized patients and the primary and secondary factors, we conducted chi-squared test for matched pairs and Multivariate conditional logistic regression analysis was used to explore odds associated with C. difficile positivity. The dependent variable was C. difficile positivity, while the effect of all the above-mentioned variables were analysed by a stepwise procedure. SPSS version 15 (SPSS Inc, Chicago, Illinois, USA) was used for all statistical analysis.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgement
  9. References

During the study period, we identified 68 patients affected by CDAD (46/22, M/F; mean age 9.4 y; range 1.1–17.8) admitted to the Department of Pediatrics, University of Naples ‘Federico II’. The comparison group consisted of 68 children (40/28, M/F; 7.6 y; range 1.2–17.5) with stool analysis negative for C. difficile. Baseline characteristics of the study population are reported in Table 1. We observed that the use of PPIs was significantly higher in C. difficile positive group compared with C. difficile negative group [22.1% vs. 5.9%; χ= 7.4; = 0.006; odds ratio (OR) = 4.52; 95% confidence interval (CI) = 1.4–14.4]. At the same time, we found a trend for the higher use of H2RAs in patients infected by C. difficile compared with C. difficile negative group, but this finding did not rise to statistical significance [10.3% vs. 2.9%; χ= 2.9; = 0.08; odds ratio (OR) = 3.78; 95% confidence interval (CI) = 0.7–18.9]. Global exposure to acid-suppressing drugs (PPIs and H2RAs) was significantly higher in patients with CDAD (20/68, 29.4%) in comparison with children with C. difficile negative stool analysis (6/68; 8.82%) [χ= 9.2; = 0.02; (OR) = 3.33; 95% confidence interval (CI) = 1.4–7.7] (Table 2). With regard to antibiotic exposure, we did not identify a significant difference between the C. difficile positive group and the C. difficile negative group (32.4% vs. 30.9%; χ= 0.34; = 0.5). No significant association was found between the C. difficile positive group and the C. difficile negative group in relation to the use of azathioprine (17.6% vs. 13.2%; χ= 0.5; = 0.3), methotrexate (1.5% vs. 2.9%; χ= 0.3; = 0.5) and mesalazine (30.9% vs. 30.9%; χ= 0; = 0.5). The C. difficile positive group was marginally more exposed to steroids than the C. difficile negative group, but the difference was not statistically significant (20.6% vs. 10.3%; χ= 2.7; = 0.07).

Table 1.   Baseline characteristics
 Case groupControl groupchi-square test*P value
  1. IBD, inflammatory bowel disease; GD, genetic diseases; FGIDs, functional gastrointestinal disorders; GERD, gastro-oesophageal reflux disease; FA, food allergy.

Mean age (years)9.4 (range 1.1–17.8)7.6 (range 1.2–17.5)1.32N.S.
Gender  1.13N.S.
 Male46 (67.6)40 (58.8)  
 Female22 (32.4)28 (41.2)  
Comorbidities (%)
 IBD43 (63.2)31 (45.6)8.80.017
 GD4 (5.9)5 (7.4)1.19N.S.
 FGIDs16 (2.3)24 (3.52)2.26N.S.
 GERD2 (2.9)2 (2.9)0N.S.
 FA3 (4.4)6 (8.8)1.07N.S.
Table 2.   Distribution of matching factors for C. difficile infection case-control pairs
VariablesCase group N = 68Control group N = 68chi-squared test*P valueOR95% CI
  1. * Fisher exact test as appropriate.

PPI (%)15 (22.1)4 (5.9)7.40.0064.521.4–14.4
H2RAs (%)7 (10.3)2 (2.9)2.90.083.780.7–18.9
Gastric acid Suppressive Th. (%)20 (29.4)6 (8.82)9.20.023.331.47–7.8
Antibiotic (%)22 (32.4)21 (30.9)0.03N.S.1.070.5–2.2
Corticosteroids (%)14 (20.6)7 (10.3)2.70.072.250.8–6
Azathioprine (%)12 (17.6)9 (13.2)0.5N.S.1.40.5–3.5
Mesalazine (%)21 (30.9)21 (30.9)0N.S.10.4–2
Methotrexate (%)1 (1.5)2 (2.9)0.3N.S.0.40.04–5.5
Hospitalization (%)34 (50)20 (29.4)6.00.012.41.4–4.8
Comorbidities (%)      
IBD43 (63.2)31 (45.6)8.80.0172.051.03–4.07

As regards comorbid illnesses, we found a higher frequency of IBD in patients affected by CDAD in the C. difficile positive group compared with the C. difficile negative group (63.2% vs. 45.6%; χ= 8.8; = 0.017). In addition, we observed a higher frequency of previous hospitalizations in the C. difficile positive group compared with the C. difficile negative group (50% vs. 29.4%; χ= 6.0; = 0.01) (Table 2).

In a multivariate logistic regression analysis, the presence/absence of C. difficile in the stools was used as dependent variable. The variable that contributed most significantly to the risk of C. difficile infection was the use of PPIs [= 0.008; (OR) = 1.2; 95% confidence interval (CI) = 1.04–1.39]. The only other variable which was associated with C. difficile infection was diagnosis of IBD [= 0.03; (OR) = 1.1; 95% confidence interval (CI) = 1.0–1.2]. Antibiotics, H2RAs, other comorbidities and previous hospitalization did not contribute significantly to the risk of infection.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgement
  9. References

In this paediatric hospital-based study, we found that patients who received PPIs therapy are at increased risk of contracting C. difficile infection. Patients who were affected by CDAD were exposed more often to PPIs therapy than those in the control group. Our findings support the hypothesis that PPIs therapy is an important risk factor for contracting C. difficile infection in children. Dial et al., found in two different studies that the use of PPIs was independently associated with an increased risk of C. difficile infection.9, 10 Recently, Aseeri et al. reported that patients receiving PPIs were 3.6 times more likely to develop CDAD respect to controls,16 but it is still unclear whether or not the use of PPIs is a risk factor for CDAD.17 In a recent review, Leonard et al., concluded after analysing data coming from 25 different papers that there is an association between risk of enteric infections, including C. difficile infection and the use of acid suppression.18 The mechanism by which gastric acid-suppressing drugs contribute to an increased risk of CDAD is yet to be defined, but the implicit mechanism of suppression of the protective function of gastric acidity is strongly suspected.

Disease transmission occurs via the ingestion of C. difficile by a susceptible host; this process is believed to be mediated by the spore form of the pathogen, which is acid-resistant. A potential explanation could be researched in the fact that the increase in gastric pH, caused by acid suppressive therapy, aids the survival of the vegetative form, which is killed by acid. A recent study conducted by Jump et al. found that the vegetative forms of C. difficile are shed in faeces in significant quantities prior to the initiation of treatment of CDAD and remain viable in room air on moist surfaces and in gastric contents with elevated pH.19 Thus, the vegetative form, which has not been considered important in the past, could play a role in pathogenesis.

Most studies primarily emphasize the role of PPIs exposure on the onset of CDAD, while data on H2RAs exposure are conflicting.9–12, 16 Dial et al. estimated an adjusted risk ratio for CDAD with the use of H2RAs to be 2.0.9 However, Aseeri et al. did not find an association between the exposure to H2RAs and the onset of CDAD.16 Similarly, in our analysis, PPIs therapy was confirmed to be a relevant risk factor for CDAD, with a crude risk ratio of 4.8, whereas our study did not have sufficient power to confirm an association between H2RAs therapy and the onset of C. difficile infection, but a clear trend was shown. Patients exposed to H2RAs showed a borderline increase in risk. It is inconclusive as to whether or not this result could be explained by the fact that H2RAs are less frequently prescribed or simply by a different mechanism of action. Therefore, further studies with H2RAs seem to be necessary.

Our results were strengthened, however, by conducting exact-conditional logistic regression. By using a multivariate model and comparing C. difficile positive group with the comparison group, we were able to control for most common risk factors (gastric acid suppressors, antibiotics, previous hospitalizations, comorbidities, corticosteroids) reported to be associated with CDAD in adult hospitalized patients and to verify which were the most predictive factors for the onset of C. difficile infection. We found that PPIs was the most significant predictor of CDAD, with an adjusted risk ratio of 1.2. The only other factor which seemed to contribute significantly to the onset of CDAD was a diagnosis of IBD. Issa et al. recently reported that C. difficile infection has significantly increased in adult IBD patients.20–22 We have already demonstrated that IBD is a relevant risk factor for C. difficile infection in paediatric patients.23 In the current study, the prevalence of IBD was higher in patients with C. difficile than in the C. difficile negative control group. The multivariate model confirmed that patients affected by IBD are at increased risk of developing CDAD. None of the others factors contributed to its development.

Systemic immunosuppression is considered a known risk factor for CDAD.24–26 In our study, both the case and the control groups are composed of a consistent number of IBD patients; thus, we tried to investigate a possible correlation between the use of methotrexate and azathioprine in IBD patients and C. difficile infection. However, when we compared the case group with the C. difficile negative group, no significant association was found. Previous exposure to antibiotics was almost considered a prerequisite for the diagnosis of CDAD in adults.9, 24, 27, 28 Unlike in adults with C.difficile disease, even if there are few studies in paediatric age, prior antibiotic exposure does not seem to be a prerequisite for disease development in children.29 Our study did not find antibiotics to be a notable risk factor for paediatric C. difficile infection, as previously demonstrated.23

Hospitalized patients are thought to be the primary target of C. difficile. Hospitals and long-term care facilities represent crucial reservoirs of C. difficile spores.2, 30, 31 We found a significantly higher frequency of previous hospitalization in patients affected by CDAD. However, the multivariate analysis did not confirm hospitalization as a true risk factor and as such, the higher frequency of previous hospitalization, found in patients infected with C. difficile, should be considered a confounding variable.

We also investigated whether or not corticosteroid exposure had a role in the onset of CDAD. Exposure was not significantly higher when comparing C. difficile positive group with the C. difficile negative group. These data were confirmed by the multivariate model which excluded steroids as predictors of the occurrence of CDAD.

Clostridium difficile-associated disease has become an important public health issue. The association of CDAD with gastric acid suppression has become increasingly evident over the last few years reflecting the high usage of these therapies in hospitalized patients.32–34

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgement
  9. References

We have shown in a retrospective case-control study that the use of PPIs therapy represents an important risk factor for CDAD not only in adults but in children as well. Despite the obvious benefits of anti-secretory drugs in the prevention and treatment of upper gastrointestinal symptoms, health-care providers should consider the risks involved in anti-secretory therapy, including that of C. difficile infection. A case-control design is ideal to estimate the risk associated with a specific exposure, but it has the limitation of being retrospective with his possible, albeit improbable, pitfalls. Additional prospective studies will be necessary to support these findings further.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgement
  9. References