Advertisement

Diarrhoea in general practice: when should a Clostridium difficile infection be considered? Results of a nested case-control study

      Abstract

      Clostridium difficile infections (CDIs) are frequent in hospitals, but also seem to increase in the community. Here, we aim to determine the incidence of CDI in general practice and to evaluate current testing algorithms for CDI. Three Dutch laboratories tested all unformed faeces (12 714) for C. difficile when diagnostic testing (for any enteric pathogen) was requested by a general practitioner (GP). Additionally, a nested case-control study was initiated, including 152 CDI patients and 304 age and sex-matched controls. Patients were compared using weighted multivariable logistic regression. One hundred and ninety-four samples (1.5%) were positive for C. difficile (incidence 0.67/10 000 patient years). This incidence was comparable to that of Salmonella spp. Compared with diarrhoeal controls, CDI was associated with more severe complaints, underlying diseases, antibiotic use and prior hospitalization. In our study, GPs requested a test for C. difficile in 7% of the stool samples, thereby detecting 40% of all CDIs. Dutch national recommendations advise testing for C. difficile when prior antibiotic use or hospitalization is present (18% of samples). If these recommendations were followed, 61% of all CDIs would have been detected. In conclusion, C. difficile is relatively frequent in general practice. Currently, testing for C. difficile is rare and only 40% of CDI in general practice is detected. Following recommendations that are based on traditional risk factors for CDI, would improve detection of CDI.

      Keywords

      Introduction

      Clostridium difficile infection (CDI) is a common cause of hospital-acquired diarrhoea. Elderly patients with underlying diseases and recent antibiotic therapy are primarily affected, resulting in prolonged hospitalization and excess mortality [
      • Loo VG
      • Bourgault AM
      • Poirier L
      • et al.
      Host and pathogen factors for Clostridium difficile infection and colonization.
      ]. Recently, CDI was reported as an emerging disease outside healthcare facilities [
      • Dial S
      • Delaney JA
      • Barkun AN
      • Suissa S
      Use of gastric acid‐suppressive agents and the risk of community‐acquired Clostridium difficile‐associated disease.
      ]. Currently, more than a quarter of all CDI is estimated to be acquired in the community [
      • Khanna S
      • Pardi DS
      • Aronson SL
      • et al.
      The epidemiology of community‐acquired clostridium difficile infection: A Population‐Based Study.
      ]. In contrast to nosocomial CDI, patients in the community are younger, antibiotics are less frequently used and routes of exposure are often unknown. Consequently, over a third of these patients have no known risk factors for CDI [
      • Bauer MP
      • Veenendaal D
      • Verhoef L
      • Bloembergen P
      • van Dissel JT
      • Kuijper EJ
      Clinical and microbiological characteristics of community‐onset Clostridium difficile infection in the Netherlands.
      ,
      • Wilcox MH
      • Mooney L
      • Bendall R
      • Settle CD
      • Fawley WN
      A case‐control study of community‐associated Clostridium difficile infection.
      ]. This makes recognition of CDI problematic, especially because C. difficile is not widely tested for in general practitioners' practices [
      • McNulty CA
      • Lasseter G
      • Verlander NQ
      • et al.
      Management of suspected infectious diarrhoea by English GPs: are they right?.
      ].
      In 2009 a guideline was introduced in the UK, stating that all cases of diarrhoea among patients aged ≥2 years in the community should be tested for C. difficile unless good clinical or epidemiological reasons not to are present [
      • Department of Health and Health Protection Agency
      Clostridium difficile infection: How to deal with the problem. 2009.
      ]. Diarrhoea is common in general practice, reaching incidences of 200 per 10 000 person years [
      • de Wit MA
      • Kortbeek LM
      • Koopmans MP
      • et al.
      A comparison of gastroenteritis in a general practice‐based study and a community‐based study.
      ,
      • Wheeler JG
      • Sethi D
      • Cowden JM
      • et al.
      Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. The Infectious Intestinal Disease Study Executive.
      ], which makes comprehensive testing costly. Consequently, the UK guideline was modified in 2012 and advised testing of all diarrhoeal samples of elderly patients or patients with risk factors [
      • Department of Health
      Updated guidance of the daignosis and reporting of Clostridium difficile.
      ]. In most countries, including the Netherlands and the USA, guidelines for general practitioners still state that C. difficile should be suspected in patients with a recent hospitalization or antibiotic use [
      Tips from Other Journals: ‘Management of Infectious Diarrhea: IDSA Guideline’.
      , ], which may result in missed diagnoses.
      Although the need to characterize patients with CDI in the community is high, few studies focused on clinical presentation and additional characteristics of this patient group [
      • Wilcox MH
      • Mooney L
      • Bendall R
      • Settle CD
      • Fawley WN
      A case‐control study of community‐associated Clostridium difficile infection.
      ,
      • Chitnis AS
      • Holzbauer SM
      • Belflower RM
      • et al.
      Epidemiology of community‐associated Clostridium difficile infection, 2009 through 2011.
      ]. Additionally, studies often select diagnosed (and therefore recognized) patients only. Therefore, we decided to describe the occurrence of CDI in a laboratory-based cohort study, testing for C. difficile irrespective of whether the general practitioner requested C. difficile testing. Using this design, we aimed to determine the incidence of toxigenic C. difficile and to characterize patients with CDI. Additionally, we aimed to evaluate and guide current diagnostic algorithms.

      Methods

       Study design

      The study was set in three medical microbiological laboratories: Stichting Huisartsen Laboratorium (Etten-Leur), the Laboratory for Medical Microbiology and Immunology of the St Elisabeth Hospital (Tilburg) and the Laboratory for Clinical Microbiology and Infectious Diseases of the Isala klinieken (Zwolle). These laboratories supply microbiological services to 832 general practices with together 2 810 830 patients. All general practitioners (GPs) were aware the study was being performed; two GPs declined participation and were not included in the study. Between 4 October 2010 and 31 January 2012, all unformed stool samples of patients aged ≥2 years, submitted by GPs, were prospectively tested for the presence of C. difficile toxin irrespective of whether the GP requested testing for C. difficile. Samples were excluded when a patient (i) had a prior positive test or (ii) was tested within the previous 30 days. An unformed stool was defined as ‘taking the shape of the container’ [
      • McDonald LC
      • Coignard B
      • Dubberke E
      • Song X
      • Horan T
      • Kutty PK
      Recommendations for surveillance of Clostridium difficile‐associated disease.
      ].
      Patients with a positive test for C. difficile toxin were defined as CDI. Using a nested case-control design, patients with CDI were matched for age (±5 years) and sex to two control patients. Control patients were selected from the cohort of toxin-negative patients and tested negative at most 1 week before the case patient. If a control patient was not available at that time, the first patient after the index date (date of CDI case) was selected. The study protocol was approved by the LUMC Medical Review Ethics Committee.

       Definitions and data collection

      We collected basic demographic data of all tested patients. One laboratory (Etten-Leur) additionally registered whether the C. difficile test was specifically requested by the GP. This was used to evaluate whether CDI testing was requested in current practice.
      After obtaining permission of the GP, questionnaires were sent to CDI patients and sampled controls. We contacted subjects by mail or telephone to request return of the questionnaire; this was done up to six times. Questions focused on medication and contact with infants or healthcare in the 3 months before diarrhoea, comorbidity in the year before diarrhoea, travelling history and proximity to other patients with diarrhoea. Frequency, viscosity and presence of bloody diarrhoea were ascertained at the height of the diarrhoeal episode. All variables, except for abdominal pain and fever, which were deemed too subjective, were included in univariate analysis. Follow-up of patients with CDI was done after 6 months by asking the GP about the initiated treatment for CDI, presence of relapses or death.

       Stool examinations

      The presence of toxin producing C. difficile was assessed by a cell cytotoxicity assay in Tilburg, which is still regarded as the reference standard [
      • Crobach MJ
      • Dekkers OM
      • Wilcox MH
      • Kuijper EJ
      European Society of Clinical Microbiology and Infectious Diseases (ESCMID): data review and recommendations for diagnosing Clostridium difficile‐infection (CDI).
      ,
      • Eastwood K
      • Else P
      • Charlett A
      • Wilcox M
      Comparison of nine commercially available Clostridium difficile toxin detection assays, a real‐time PCR assay for C. difficile tcdB, and a glutamate dehydrogenase detection assay to cytotoxin testing and cytotoxigenic culture methods.
      ]. The two other laboratories used an enzyme immunoassay (EIA) for toxins A and B (Premier toxins A&B, Meridian, Bioscience Inc., Cincinnati, OH, USA).
      Upon the request of the general practitioner, faeces were tested for diarrhoeal pathogens other than C. difficile. These pathogens were tested using local available tests (all PCR). Testing was possible for: bacterial pathogens (Salmonella spp., Shigella spp. and Campylobacter jejuni/coli), parasitic pathogens (Cryptosporidium, Giardia lamblia and Entamoeba histolytica) or viruses (norovirus) in all three laboratories. Additional tests were available upon request or if deemed clinically relevant based on patient data (data not shown). All microbiological results, including the result of the C. difficile toxin test, were reported to the GP.
      Stool samples that were positive for C. difficile in the initial test were cultured and isolates were typed with PCR ribotyping [
      • Bidet P
      • Lalande V
      • Salauze B
      • et al.
      Comparison of PCR‐ribotyping, arbitrarily primed PCR, and pulsed‐field gel electrophoresis for typing Clostridium difficile.
      ]. When an isolate could not be obtained, a PCR on the tcdB gene was performed on faeces to confirm the presence of toxigenic C. difficile [
      • van den Berg RJ
      • Vaessen N
      • Endtz HP
      • Schulin T
      • van der Vorm ER
      • Kuijper EJ
      Evaluation of real‐time PCR and conventional diagnostic methods for the detection of Clostridium difficile‐associated diarrhoea in a prospective multicentre study.
      ].

       Data analysis

      Incidence rates of diarrhoea and intestinal pathogens were calculated using the total number of person years at risk, which was calculated by multiplying the general practice population (the number of people serviced by the participating general practitioners, according to their patient list) by the period of study participation (between 12 and 15 months).
      Patients with Clostridium difficile infection and matched controls were compared using univariate conditional logistic regression. Results were displayed as matched odds ratios (mOR) with a 95% confidence interval (95% CI). Subsequently, all factors with a p-value of <0.10, except for symptoms, were included in multivariable analysis. Although these variables were complete in >92% of the CDI patients and controls, we used multiple imputation to account for missing values in multivariate analysis. This method is appropriate when predictors of the missing data are available (missing at random; MAR) [
      • Sterne JA
      • White IR
      • Carlin JB
      • et al.
      Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls.
      ]. All potential predictors of missing data, potential predictors of the outcome and the outcome itself were included in the imputation procedure. To include the matched variables (age and gender; both dichotomous) in the multivariate analysis, we performed case-control weighting [
      • Rose S
      • van der Laan MJ
      A Note on Risk Prediction for Case‐Control Studies. U.C. Berkeley Division of Biostatistics Working Paper Series. Working Paper 241.
      ]. This was possible due to the fact that the case-control study was nested in a cohort. Weights were determined by prevalence, age (continuous variable) and sex distribution of patients and controls compared with the original cohort. In patients, weights varied only marginally (between 1.2 and 1.4), because 78% of the diagnosed CDI patients participated in the case-control study. Weights of controls varied between 17 and 112 (mean 41), emphasizing the large sampling fraction and the relative over-representation of elderly patients due to matching.

       Evaluation of testing strategies

      We evaluated the current diagnostic practice of general practitioners by evaluating the samples for which the GP requested testing for C. difficile. This method was compared with the current advice in the Netherlands, the current advice in the UK and the former advice in the UK. The percentage of diarrhoeal patients that required testing according to the aforementioned recommendations, was calculated by using the prevalence of clinical characteristics in our weighted population of diarrhoeal patients and controls (e.g. prevalence of patients with antibiotic use or prior admission was calculated because these patients required C. difficile testing according to current Dutch recommendations). In the population that required testing, we determined the percentage of CDI (e.g. among patients with prior antibiotic use or an admission, 8% were CDI positive). Additionally, we determined the percentage of CDI patients that would have been tested by the algorithm (e.g. 60% of all CDI patients occurred in the group of patients with prior antibiotic use or an admission).
      We used SPSS version 20.0 (SPSS Inc., Chicago, IL, USA) and STATA software package 10.1 (StataCorp, College Station, TX, USA) for our analyses.

      Results

      During the study period, 12 714 unformed stool samples met the study's inclusion criteria (Fig. 1). The incidence of diarrhoea in which investigation of faeces was requested was 44 per 10 000 person years. Patients were on average 41.3 years old and the majority was female (57.4%) (Table 1).
      Figure thumbnail gr1
      FIG. 1Patient inclusion chart.
      TABLE 1Age, gender and incidence of intestinal pathogens in unformed stool samples with a test request from the general practitioner
      Samples (N = 12 714)
      No. of cases% of all samplesRate per 10 000 person years (95% CI)No. of samples tested
      Female gender730257.4
      Age, mean (±SD)41.3 (23.2)
      Diagnosed pathogens
      Campylobacter coli/jejuni10568.33.67(3.45–3.90)10 598
      Giardia lamblia4543.61.58(1.44–1.73)8954
      Salmonella spp.1981.60.69(0.60–0.79)10 598
      Clostridium difficile1941.50.67(0.58–0.78)12 714
      Shigella spp.1140.90.40(0.33–0.47)10 598
      Cryptosporidium1070.80.37(0.31–0.45)8954
       Norovirus750.60.26(0.21–0.32)1374
      Entamoeba histolytica20.00.01(0.00–0.02)6720
      Pathogen other than C. difficile278621.99.68(9.33–10.05)12 566
      All samples were tested for C. difficile, whereas other pathogens were tested upon request of the general practitioner. All laboratories used a PCR to detect the pathogens: Campylobacter [
      • Lund M
      • Nordentoft S
      • Pedersen K
      • Madsen M
      Detection of Campylobacter spp. in chicken fecal samples by real‐time PCR.
      ,
      • Schuurman T
      • de Boer RF
      • van Zanten E
      • et al.
      Feasibility of a molecular screening method for detection of Salmonella enterica and Campylobacter jejuni in a routine community‐based clinical microbiology laboratory.
      ], Salmonella [
      • Schuurman T
      • de Boer RF
      • van Zanten E
      • et al.
      Feasibility of a molecular screening method for detection of Salmonella enterica and Campylobacter jejuni in a routine community‐based clinical microbiology laboratory.
      ,
      • Malorny B
      • Paccassoni E
      • Fach P
      • Bunge C
      • Martin A
      • Helmuth R
      Diagnostic real‐time PCR for detection of Salmonella in food.
      ], Shigella [
      • Vu DT
      • Sethabutr O
      • Von SL
      • et al.
      Detection of Shigella by a PCR assay targeting the ipaH gene suggests increased prevalence of shigellosis in Nha Trang, Vietnam.
      ], Giardia lamblia [
      • Bruijnesteijn van Coppenraet LE
      • Wallinga JA
      • Ruijs GJ
      • Bruins MJ
      • Verweij JJ
      Parasitological diagnosis combining an internally controlled real‐time PCR assay for the detection of four protozoa in stool samples with a testing algorithm for microscopy.
      ,
      • Verweij JJ
      • Blange RA
      • Templeton K
      • et al.
      Simultaneous detection of Entamoeba histolytica, Giardia lamblia, and Cryptosporidium parvum in fecal samples by using multiplex real‐time PCR.
      ,
      • Jothikumar N
      • da Silva AJ
      • Moura I
      • Qvarnstrom Y
      • Hill VR
      Detection and differentiation of Cryptosporidium hominis and Cryptosporidium parvum by dual TaqMan assays.
      ], Cryptosporidium [
      • Bruijnesteijn van Coppenraet LE
      • Wallinga JA
      • Ruijs GJ
      • Bruins MJ
      • Verweij JJ
      Parasitological diagnosis combining an internally controlled real‐time PCR assay for the detection of four protozoa in stool samples with a testing algorithm for microscopy.
      ,
      • Verweij JJ
      • Blange RA
      • Templeton K
      • et al.
      Simultaneous detection of Entamoeba histolytica, Giardia lamblia, and Cryptosporidium parvum in fecal samples by using multiplex real‐time PCR.
      ], Entamoeba histolytica [
      • Bruijnesteijn van Coppenraet LE
      • Wallinga JA
      • Ruijs GJ
      • Bruins MJ
      • Verweij JJ
      Parasitological diagnosis combining an internally controlled real‐time PCR assay for the detection of four protozoa in stool samples with a testing algorithm for microscopy.
      ,
      • Verweij JJ
      • Blange RA
      • Templeton K
      • et al.
      Simultaneous detection of Entamoeba histolytica, Giardia lamblia, and Cryptosporidium parvum in fecal samples by using multiplex real‐time PCR.
      ] and Norovirus [
      • Kageyama T
      • Kojima S
      • Shinohara M
      • et al.
      Broadly reactive and highly sensitive assay for Norwalk‐like viruses based on real‐time quantitative reverse transcription‐PCR.
      ].

       Incidence of C. difficile infection

      Of 12 714 stool samples, 194 (1.5%) were positive for C. difficile (incidence of 0.67 per 10 000 patient years). In Tilburg, a cell cytotoxicity assay (considered as reference standard) was used to diagnose CDI. Here, 54 tests were found to be positive among 3009 diarrhoeal samples (1.8%; 103 per 10 000 patient years).
      Ninety-nine per cent of the stool samples were also tested for the presence of pathogens other than C. difficile (12 566/12 714), which were identified in 21.9% (2786/12 714) of all samples: in 22.1% of the CDI-negative samples (2763/12 520) and in 11.8% (23/194) of the CDI-positive samples. The most frequently found co-pathogen in CDI-positive samples was Campylobacter coli/jejuni (n = 10; 5%). In total, Campylobacter coli/jejuni and Giardia lamblia were found in 8.3% (1056/12 714; 3.67 per 10 000 person years) and 3.6% (454/12 714; 1.58 per 10 000 person years) of all samples, respectively. Salmonella spp. was found in percentages similar to C. difficile: 1.6% (198/12 714; 0.69 per 10 000 person years).

       CDI vs controls with diarrhoea

      Within the cohort of 12 714 samples, we performed a nested case-control study. One hundred and fifty-two of 194 CDI patients (78%) completed the questionnaire and were matched for age and gender to 304 controls. Participating CDI patients were on average 52.3 years old (standard deviation 22.5); 61% of them were female. Symptoms of diarrhoea started in the community in 94% (n = 143). Three patients (2%) developed symptoms in a long-term care facility and six (4%) developed diarrhoea during hospitalization but were diagnosed after discharge. Compared with controls, CDI patients more often had severe symptoms (bloody stools, watery or frequent diarrhoea), underlying diseases, prior hospitalization and prior use of antibiotics (univariate analysis; Table 2). A third of the CDI patients (n = 58; 39%) did not use antibiotics nor were previously hospitalized; 14% of the CDI patients (n = 22) had no underlying diseases, hospitalization or medication use prior to diarrhoea. CDI patients reported abdominal pain and fever in 77% and 31%, respectively; controls reported these symptoms in 75% and 20%, respectively.
      TABLE 2Clinical characteristics of CDI patients and matched control patients, analysed with conditional logistic regression analysis
      CharacteristicsCDI cases (N = 152)Controls (N = 304)Crude analysis
      N%N%mOR95% CIp‐Value
      Symptoms
       Bloody stools3625.24415.71.821.07–3.090.03
       Watery diarrhoea11978.320768.11.711.08–2.710.02
       Frequency of diarrhoea >8 times6844.77524.92.391.59–3.61<0.01
       Time to visit GP <1 month9664.516556.31.400.94–2.100.10
      Medication
       Antibiotics8255.04916.68.154.57–15.5<0.01
       Other medication9260.516656.11.260.81–1.980.31
       PPI/antacid4329.16021.11.590.99–2.550.06
       Statin2516.94014.11.380.74–2.580.31
       NSAID117.4248.40.800.37–1.730.57
       DM106.8196.71.030.46–2.280.95
       Immunosuppression117.4124.21.720.74–4.020.21
       Diuretics, antihypertensives4730.97625.21.480.87–2.530.15
      Underlying diseases
       Any disease9059.212039.72.641.66–4.20<0.01
       Circulatory system diseases1811.83411.31.090.54–2.190.81
       Respiratory system diseases2415.8268.61.901.08–3.360.03
       Cancer106.672.33.601.21–10.70.02
      Environment
       Previous admission2818.4217.03.161.67–5.99<0.01
       Family member with diarrhoea74.8238.00.580.25–1.350.20
       Infant <2 years old4027.69732.20.750.47–1.200.23
       Visited foreign country
       In western world1615.44318.40.790.40–1.560.50
       Outside western world1514.44117.50.770.38–1.580.48
      The crude analysis was done by univariate conditional logistic regression, which takes into account the matched factors ‘age’ and ‘gender’. Variables with a p‐value <0.10 (n = 9) supplemented with age and sex were included in the multivariate analysis (Table 3).
      Six variables had a p-value of <0.10 in univariate analysis and were possible predictive factors of CDI. These were included in multivariate analysis together with age and gender. Age ≥50 years, an underlying disease in the year before start of diarrhoea and hospitalization in the preceding 3 months and cancer in the preceding year had a strong association with CDI. Antibiotic use in the preceding 3 months had the strongest association with CDI (Table 3).
      TABLE 3Multivariable analysis using weighted logistic regression analysis
      FactorsMVA
      OR95% CIp‐value
      Age ≥501.410.79–2.520.25
      Gender1.180.70–1.990.53
      Antibiotics6.883.97–11.9<0.01
      PPI/antacid1.100.56–2.080.77
      Any disease1.801.00–3.230.05
      Respiratory system diseases1.250.51–3.060.63
      Cancer4.041.47–11.1<0.01
      Previous admission1.660.75–3.680.21
      In the multivariate analysis (MVA) we included all possible predictive factors for CDI with a p‐value of <0.10 according to the crude analysis and ‘age’ and ‘gender’. Symptoms were not included in the MVA. We adjusted the variables in this table for age, gender, antibiotics, PPI/antacid, respiratory system diseases, cancer and previous admission. The variable ‘any disease’ was not adjusted for ‘respiratory system diseases’ and ‘cancer’ as these variables were included in ‘any disease’.

       Performance of testing algorithms

      According to data from one laboratory (Etten-Leur), general practitioners request a test for CDI in 7% of submitted samples (543/8338). These samples included 40% of all diagnosed CDI patients in this study. Currently, the advice regarding testing for C. difficile in general practice in the Netherlands is to test all patients with diarrhoea and recent antibiotic use or hospitalization. As 18% of the patients in the study recently used antibiotics or were hospitalized, this advice would lead to testing of 18% of all diarrhoeal patients, detecting 61% of all CDI patients. In the United Kingdom, all diarrhoeal patients aged ≥65 years or patients with recent antibiotic use or a recent hospitalization are advised to be tested. Implementing this strategy in our study population would result in detection of two-thirds of all CDI patients, whereas it would require testing 31% of all diarrhoeal samples.

       Confirmation of C. difficile

      Of the 152 patients with CDI, the presence of C. difficile could be confirmed by PCR ribotyping or a positive tcdB PCR in 68% (n = 103). Types 002 and 078 (both n = 11; 11%) were most frequently found; type 001 (8%), 005 (6%), 014 (8%), 015 (9%) and 126 (4%) were other frequently found PCR ribotypes. The virulent type 027 that caused many outbreaks in hospitals [
      • Loo VG
      • Poirier L
      • Miller MA
      • et al.
      A predominantly clonal multi‐institutional outbreak of Clostridium difficile‐associated diarrhea with high morbidity and mortality.
      ] was isolated in one patient with frequent relapses and prior long-term hospitalization. Thirty-five stool samples were not available for confirmation testing. The majority of the CDI patients in the case-control study had C. difficile as the only detected pathogen (130 of 152; 86%).

       Six months follow-up

      Of 122 CDI patients with known follow-up (80.3%), the majority (n = 96; 78.7%) was treated for the infection: monotherapy with metronidazole was most frequently used (n = 85; 88.5%); six patients were treated with vancomycin (6.3%), three with a combination of both (3.1%). Thirty patients (24.6%) had recurrent diarrhoea within 6 months, which was confirmed by a positive toxin test in 36.7%. Within 6 months, six CDI patients (3.9%) were hospitalized because of diarrhoea and four died (2.6%). In one patient (0.6%) CDI contributed to the cause of death.

      Discussion

       Incidence of CDI in general practice

      This study determined the incidence of C. difficile in a large sample of microbiological test requests from general practitioners. One out of 66 diarrhoeal episodes was positive for C. difficile (1.5%), which was comparable to the incidence of Salmonella spp. Earlier studies reported similar incidences of CDI (1.5–2.1% [
      • Bauer MP
      • Veenendaal D
      • Verhoef L
      • Bloembergen P
      • van Dissel JT
      • Kuijper EJ
      Clinical and microbiological characteristics of community‐onset Clostridium difficile infection in the Netherlands.
      ,
      • Wilcox MH
      • Mooney L
      • Bendall R
      • Settle CD
      • Fawley WN
      A case‐control study of community‐associated Clostridium difficile infection.
      ]; 0.7–2.5 per 10 000 person years [
      • Dial S
      • Delaney JA
      • Barkun AN
      • Suissa S
      Use of gastric acid‐suppressive agents and the risk of community‐acquired Clostridium difficile‐associated disease.
      • Khanna S
      • Pardi DS
      • Aronson SL
      • et al.
      The epidemiology of community‐acquired clostridium difficile infection: A Population‐Based Study.
      • Bauer MP
      • Veenendaal D
      • Verhoef L
      • Bloembergen P
      • van Dissel JT
      • Kuijper EJ
      Clinical and microbiological characteristics of community‐onset Clostridium difficile infection in the Netherlands.
      • Wilcox MH
      • Mooney L
      • Bendall R
      • Settle CD
      • Fawley WN
      A case‐control study of community‐associated Clostridium difficile infection.
      • McNulty CA
      • Lasseter G
      • Verlander NQ
      • et al.
      Management of suspected infectious diarrhoea by English GPs: are they right?.
      • Department of Health and Health Protection Agency
      Clostridium difficile infection: How to deal with the problem. 2009.
      • de Wit MA
      • Kortbeek LM
      • Koopmans MP
      • et al.
      A comparison of gastroenteritis in a general practice‐based study and a community‐based study.
      • Wheeler JG
      • Sethi D
      • Cowden JM
      • et al.
      Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. The Infectious Intestinal Disease Study Executive.
      • Department of Health
      Updated guidance of the daignosis and reporting of Clostridium difficile.
      Tips from Other Journals: ‘Management of Infectious Diarrhea: IDSA Guideline’.
      • Chitnis AS
      • Holzbauer SM
      • Belflower RM
      • et al.
      Epidemiology of community‐associated Clostridium difficile infection, 2009 through 2011.
      • McDonald LC
      • Coignard B
      • Dubberke E
      • Song X
      • Horan T
      • Kutty PK
      Recommendations for surveillance of Clostridium difficile‐associated disease.
      • Crobach MJ
      • Dekkers OM
      • Wilcox MH
      • Kuijper EJ
      European Society of Clinical Microbiology and Infectious Diseases (ESCMID): data review and recommendations for diagnosing Clostridium difficile‐infection (CDI).
      • Eastwood K
      • Else P
      • Charlett A
      • Wilcox M
      Comparison of nine commercially available Clostridium difficile toxin detection assays, a real‐time PCR assay for C. difficile tcdB, and a glutamate dehydrogenase detection assay to cytotoxin testing and cytotoxigenic culture methods.
      • Bidet P
      • Lalande V
      • Salauze B
      • et al.
      Comparison of PCR‐ribotyping, arbitrarily primed PCR, and pulsed‐field gel electrophoresis for typing Clostridium difficile.
      • van den Berg RJ
      • Vaessen N
      • Endtz HP
      • Schulin T
      • van der Vorm ER
      • Kuijper EJ
      Evaluation of real‐time PCR and conventional diagnostic methods for the detection of Clostridium difficile‐associated diarrhoea in a prospective multicentre study.
      • Sterne JA
      • White IR
      • Carlin JB
      • et al.
      Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls.
      • Rose S
      • van der Laan MJ
      A Note on Risk Prediction for Case‐Control Studies. U.C. Berkeley Division of Biostatistics Working Paper Series. Working Paper 241.
      • Loo VG
      • Poirier L
      • Miller MA
      • et al.
      A predominantly clonal multi‐institutional outbreak of Clostridium difficile‐associated diarrhea with high morbidity and mortality.
      • Hensgens MP
      • Keessen EC
      • Squire MM
      • et al.
      Clostridium difficile infection in the community: a zoonotic disease?.
      • Noren T
      • Akerlund T
      • Back E
      • et al.
      Molecular epidemiology of hospital‐associated and community‐acquired Clostridium difficile infection in a Swedish county.
      • Karlstrom O
      • Fryklund B
      • Tullus K
      • Burman LG
      A prospective nationwide study of Clostridium difficile‐associated diarrhea in Sweden. The Swedish C. difficile Study Group.
      • Lambert PJ
      • Dyck M
      • Thompson LH
      • Hammond GW
      Population‐based surveillance of Clostridium difficile infection in Manitoba, Canada, by using interim surveillance definitions.
      • Hirschhorn LR
      • Trnka Y
      • Onderdonk A
      • Lee ML
      • Platt R
      Epidemiology of community‐acquired Clostridium difficile‐associated diarrhea.
      • Kuntz JL
      • Chrischilles EA
      • Pendergast JF
      • Herwaldt LA
      • Polgreen PM
      Incidence of and risk factors for community‐associated Clostridium difficile infection: a nested case‐control study.
      ]), with the exception of a study from the UK that reported virtually no CDI in general practice [
      • Tam CC
      • Rodrigues LC
      • Viviani L
      • et al.
      Longitudinal study of infectious intestinal disease in the UK (IID2 study): incidence in the community and presenting to general practice.
      ]. The latter UK study confirmed our relatively low rate of salmonellosis (1.8 per 10 000 patient years using faecal culture), but should be interpreted with caution because exclusion criteria such as recent travel and diarrhoeal illness lasting over 2 weeks resulted in the analysis of 45% (991/2203) of all diarrhoeal episodes. Although we included all diarrhoeal samples that were sent to a laboratory, the incidence of CDI in our study could be underestimated if diarrhoeal samples of patients with CDI were not sent to a laboratory and the disease had a self-limiting course.
      Our study included 12 714 diarrhoeal episodes and showed that CDI is relatively common among diarrhoeal stool samples and should be included in the differential diagnosis of infectious diarrhoea in general practice.

       When should we consider CDI and request a test?

      Dutch GPs are advised to test all patients with prior antibiotic use or hospitalization for CDI. Currently, GPs do not follow these recommendations and test only 7% of all diarrhoeal patients, detecting 40% of all CDI patients. This large proportion of undiagnosed patients with CDI is in our opinion undesirable, as all CDI patients had diarrhoeal complaints and nine patients (5.8%) experienced a complicated course (hospitalization or death within 6 months). A similar course was observed in community-based studies [
      • Wilcox MH
      • Mooney L
      • Bendall R
      • Settle CD
      • Fawley WN
      A case‐control study of community‐associated Clostridium difficile infection.
      ,
      • Hirschhorn LR
      • Trnka Y
      • Onderdonk A
      • Lee ML
      • Platt R
      Epidemiology of community‐acquired Clostridium difficile‐associated diarrhea.
      ]; however, as most CDI patients in these studies were treated for CDI, we expect the number of complicated courses to be higher when CDI is undiagnosed and therefore untreated. In our study, complicated courses were also experienced by patients without traditional risk factors (3/9; 33.3%), which underlines the necessity for diagnosis.
      Because testing of all unformed stool samples, as was the former UK advice, requires a large budget, this is currently probably not achievable in most laboratories and general practices. Our study confirms that well-known risk factors for nosocomial CDI (antibiotic use and hospitalization) are present in only 61% of the patients with CDI in the community. As shown in multivariate analysis, the clinical presentation of patients with CDI differs from other causes of diarrhoea, as they frequently have bloody stools, watery diarrhoea and many stools daily. Therefore, we suggest including clinical symptoms in a future prediction model for CDI. For now, we recommend following current Dutch guidelines or the current UK advice in the Netherlands. This would result in detection of 61% or 72% of all CDI, respectively, which would clearly outperform current practice.

       Strengths and weaknesses

      We are the first to provide a complete overview of incidence, clinical characteristics and testing strategies of CDI in general practitioners' practice. The size of the cohort and high participation rate (78%), and the early and thorough follow-up of the questionnaire, provide a stable base for our conclusions. Furthermore, we were able to confirm C. difficile with PCR ribotyping in two-thirds of the cases with a positive toxin test, which enabled us to compare types circulating in general practice with those causing disease in hospitals. Similar types were seen in general practice and hospitals in the Netherlands during the study period [
      Fifth Annual Report of the National Reference Laboratory for Clostridium difficile (May 2010 to May 2011) and results of the sentinel surveillance.
      ,
      Sixth Annual Report of the National Reference Laboratory for Clostridium difficile (May 2011 to May 2012) and results of the sentinel surveillance.
      ]. As recent evidence suggests that direct transmission of C. difficile between hospitalized patients is not the prime route of transmission [
      • Didelot X
      • Eyre D
      • Cule M
      • et al.
      Microevolutionary analysis of Clostridium difficile genomes to investigate transmission.
      ], the large overlap of PCR ribotypes in both settings strengthen the hypothesis of movement of C. difficile between both settings.
      Our study also has limitations. Firstly, we restricted our study to samples that were sent to a laboratory. Our conclusions are therefore not necessarily generalizable to settings with different testing criteria. Although Dutch GPs request laboratory diagnostics in 10–20% of gastroenteritis consultations [
      • van den Brandhof WE
      • Bartelds AI
      • Koopmans MP
      • Van Duynhoven YT
      General practitioner practices in requesting laboratory tests for patients with gastroenteritis in the Netherlands, 2001–2002.
      ] and 20–30% of the GPs in the UK request testing [
      • Wheeler JG
      • Sethi D
      • Cowden JM
      • et al.
      Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. The Infectious Intestinal Disease Study Executive.
      ,
      • Noone A
      • Cossar J
      • Spence G
      • Allardice G
      • Girdwood T
      Gastrointestinal infections presenting in general practice in Scotland.
      ], testing criteria in other countries could differ. Secondly, testing strategies in our study include the ‘reference standard’ and an enzyme immunoassay (EIA), which has a limited negative and positive predictive value in the community [
      • Wilcox MH
      • Planche T
      Clostridium difficile infection.
      ]. Missing cases due to a false-negative toxin test could have resulted in an underestimation of the incidence of CDI. However, the incidence according to the reference standard (used in Tilburg) was even higher. The large sampling fraction in the case-control study makes it unlikely that false-negative patients were included as controls. However, false-positive cases might have occurred. In the majority of the CDI cases (n = 130, 86%) no pathogens other than C. difficile were found. Additionally, in 13 of the 22 CDI cases with a co-pathogen, the presence of toxigenic C. difficile was confirmed by PCR ribotyping. Therefore, we assume that bias due to to false-positive cases is limited. Thirdly, we would like to stress that the results of Table 4 are dependent on the test that was used. In a setting where different tests for C. difficile are used, sensitivity and specificity and therefore the measured incidence of CDI (and the weighted case-control analyses of Table 4) can differ. Nonetheless, our conclusion regarding present insufficient testing and suggestions for future testing are strong and will hold in a setting with a different test.
      TABLE 4Performance of seven different algorithms for testing diarrhoeal samples for Clostridium difficile in general practice
      Test algorithm for CDI in diarrhoeal samples from the communitySettingPatients testedPositive resultsDetection of CDI
      % of all unformed stool samples% of all tested samples% of all positives
      ≥2 yearsFormer advice UK (2009)1001.5100
      ≥65 years, after AB use or hospitalizationCurrent advice UK (2012)313.572
      After AB use or hospitalizationCurrent advice NL185.061
      Doctor's current practiceCurrent practice NL78.140
      These percentages are based on the weighted analysis of all CDI patients and controls (n = 12 714). AB, antibiotic.

       Clinical relevance

      Although it has several limitations, our study illustrates that CDI should be included in the differential diagnosis of infectious diarrhoea in general practice, even when the patient was not recently using antibiotics, is young and has no comorbidity. Additionally, it highlights that current Dutch testing strategies are insufficient. We recommend following current Dutch guidelines or the current UK advice in the Netherlands, which outperform current practice without testing a large number of samples.

      Acknowledgements

      We thank Celine Harmanus and Ingrid Sanders and all local technicians for performing the diagnostic tests and confirmation tests.

      Funding

      This work was supported by a grant from ZonMw (grant number 4726).

      Reporting

      This study was reported according to the STROBE guidelines.

      Details of Contributors

      M. Hensgens: contributed to the collection of the data and design of study, performed all analyses and produced the first draft of the article. O. Dekkers: designed the study, contributed to the epidemiological analyses and revised drafts of the article. A. Demeulemeester: contributed to the collection of the data and coordinated the study in Etten-Leur. P. Bloembergen and A. Buiting: coordinated the study in Zwolle and Tilburg, respectively, and revised drafts of the manuscript. B. van Benthem: contributed to the design of the study and revised drafts of the manuscript. S. Le Cessie: assisted with and performed parts of the statistical analysis of the study and revised drafts of the manuscript. E. Kuijper: designed the study and revised the drafts of the article.

      Transparency Declaration

      None to declare.

      References

        • Loo VG
        • Bourgault AM
        • Poirier L
        • et al.
        Host and pathogen factors for Clostridium difficile infection and colonization.
        N Engl J Med. 2011; 365: 1693-1703
        • Dial S
        • Delaney JA
        • Barkun AN
        • Suissa S
        Use of gastric acid‐suppressive agents and the risk of community‐acquired Clostridium difficile‐associated disease.
        JAMA. 2005; 294: 2989-2995
        • Khanna S
        • Pardi DS
        • Aronson SL
        • et al.
        The epidemiology of community‐acquired clostridium difficile infection: A Population‐Based Study.
        Am J Gastroenterol. 2011; 56: 1401-1406
        • Bauer MP
        • Veenendaal D
        • Verhoef L
        • Bloembergen P
        • van Dissel JT
        • Kuijper EJ
        Clinical and microbiological characteristics of community‐onset Clostridium difficile infection in the Netherlands.
        Clin Microbiol Infect. 2009; 15: 1087-1092
        • Wilcox MH
        • Mooney L
        • Bendall R
        • Settle CD
        • Fawley WN
        A case‐control study of community‐associated Clostridium difficile infection.
        J Antimicrob Chemother. 2008; 62: 388-396
        • McNulty CA
        • Lasseter G
        • Verlander NQ
        • et al.
        Management of suspected infectious diarrhoea by English GPs: are they right?.
        Br J Gen Pract. 2014; 64: e24-e30
        • Department of Health and Health Protection Agency
        Clostridium difficile infection: How to deal with the problem. 2009.
        (last accessed 23 July 2014).
        • de Wit MA
        • Kortbeek LM
        • Koopmans MP
        • et al.
        A comparison of gastroenteritis in a general practice‐based study and a community‐based study.
        Epidemiol Infect. 2001; 127: 389-397
        • Wheeler JG
        • Sethi D
        • Cowden JM
        • et al.
        Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. The Infectious Intestinal Disease Study Executive.
        BMJ. 1999; 318: 1046-1050
        • Department of Health
        Updated guidance of the daignosis and reporting of Clostridium difficile.
        (last accessed 21 May 2014).
      1. Tips from Other Journals: ‘Management of Infectious Diarrhea: IDSA Guideline’.
        Am Fam Physician. 2001; 64 (2012.): 1065-1066
        • Nederlands Huisartsen Genootschap
        Standaarden.
        (last accessed 4 March 2014).
        • Chitnis AS
        • Holzbauer SM
        • Belflower RM
        • et al.
        Epidemiology of community‐associated Clostridium difficile infection, 2009 through 2011.
        JAMA Intern Med. 2013; 173: 1359-1367
        • McDonald LC
        • Coignard B
        • Dubberke E
        • Song X
        • Horan T
        • Kutty PK
        Recommendations for surveillance of Clostridium difficile‐associated disease.
        Infect Control Hosp Epidemiol. 2007; 28: 140-145
        • Crobach MJ
        • Dekkers OM
        • Wilcox MH
        • Kuijper EJ
        European Society of Clinical Microbiology and Infectious Diseases (ESCMID): data review and recommendations for diagnosing Clostridium difficile‐infection (CDI).
        Clin Microbiol Infect. 2009; 15: 1053-1066
        • Eastwood K
        • Else P
        • Charlett A
        • Wilcox M
        Comparison of nine commercially available Clostridium difficile toxin detection assays, a real‐time PCR assay for C. difficile tcdB, and a glutamate dehydrogenase detection assay to cytotoxin testing and cytotoxigenic culture methods.
        J Clin Microbiol. 2009; 47: 3211-3217
        • Bidet P
        • Lalande V
        • Salauze B
        • et al.
        Comparison of PCR‐ribotyping, arbitrarily primed PCR, and pulsed‐field gel electrophoresis for typing Clostridium difficile.
        J Clin Microbiol. 2000; 38: 2484-2487
        • van den Berg RJ
        • Vaessen N
        • Endtz HP
        • Schulin T
        • van der Vorm ER
        • Kuijper EJ
        Evaluation of real‐time PCR and conventional diagnostic methods for the detection of Clostridium difficile‐associated diarrhoea in a prospective multicentre study.
        J Med Microbiol. 2007; 56: 36-42
        • Sterne JA
        • White IR
        • Carlin JB
        • et al.
        Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls.
        BMJ. 2009; 338: b2393
        • Rose S
        • van der Laan MJ
        A Note on Risk Prediction for Case‐Control Studies. U.C. Berkeley Division of Biostatistics Working Paper Series. Working Paper 241.
        (last accessed 22 October 2012).
        • Loo VG
        • Poirier L
        • Miller MA
        • et al.
        A predominantly clonal multi‐institutional outbreak of Clostridium difficile‐associated diarrhea with high morbidity and mortality.
        N Engl J Med. 2005; 353: 2442-2449
        • Hensgens MP
        • Keessen EC
        • Squire MM
        • et al.
        Clostridium difficile infection in the community: a zoonotic disease?.
        Clin Microbiol Infect. 2012; 18: 635-645
        • Noren T
        • Akerlund T
        • Back E
        • et al.
        Molecular epidemiology of hospital‐associated and community‐acquired Clostridium difficile infection in a Swedish county.
        J Clin Microbiol. 2004; 42: 3635-3643
        • Karlstrom O
        • Fryklund B
        • Tullus K
        • Burman LG
        A prospective nationwide study of Clostridium difficile‐associated diarrhea in Sweden. The Swedish C. difficile Study Group.
        Clin Infect Dis. 1998; 26: 141-145
        • Lambert PJ
        • Dyck M
        • Thompson LH
        • Hammond GW
        Population‐based surveillance of Clostridium difficile infection in Manitoba, Canada, by using interim surveillance definitions.
        Infect Control Hosp Epidemiol. 2009; 30: 945-951
        • Hirschhorn LR
        • Trnka Y
        • Onderdonk A
        • Lee ML
        • Platt R
        Epidemiology of community‐acquired Clostridium difficile‐associated diarrhea.
        J Infect Dis. 1994; 169: 127-133
        • Kuntz JL
        • Chrischilles EA
        • Pendergast JF
        • Herwaldt LA
        • Polgreen PM
        Incidence of and risk factors for community‐associated Clostridium difficile infection: a nested case‐control study.
        BMC Infect Dis. 2011; 11: 194
        • Tam CC
        • Rodrigues LC
        • Viviani L
        • et al.
        Longitudinal study of infectious intestinal disease in the UK (IID2 study): incidence in the community and presenting to general practice.
        Gut. 2012; 61: 69-77
      2. Fifth Annual Report of the National Reference Laboratory for Clostridium difficile (May 2010 to May 2011) and results of the sentinel surveillance.
        (last accessed 1 March 2012).
      3. Sixth Annual Report of the National Reference Laboratory for Clostridium difficile (May 2011 to May 2012) and results of the sentinel surveillance.
        (last accessed 4 March 2014).
        • Didelot X
        • Eyre D
        • Cule M
        • et al.
        Microevolutionary analysis of Clostridium difficile genomes to investigate transmission.
        Genome Biol. 2012; 13: R118
        • van den Brandhof WE
        • Bartelds AI
        • Koopmans MP
        • Van Duynhoven YT
        General practitioner practices in requesting laboratory tests for patients with gastroenteritis in the Netherlands, 2001–2002.
        BMC Fam Pract. 2006; 7: 56
        • Noone A
        • Cossar J
        • Spence G
        • Allardice G
        • Girdwood T
        Gastrointestinal infections presenting in general practice in Scotland.
        Health Bull (Edinb). 2000; 58: 286-300
        • Wilcox MH
        • Planche T
        Clostridium difficile infection.
        BMJ. 2009; 338: b2528
        • Lund M
        • Nordentoft S
        • Pedersen K
        • Madsen M
        Detection of Campylobacter spp. in chicken fecal samples by real‐time PCR.
        J Clin Microbiol. 2004; 42: 5125-5132
        • Schuurman T
        • de Boer RF
        • van Zanten E
        • et al.
        Feasibility of a molecular screening method for detection of Salmonella enterica and Campylobacter jejuni in a routine community‐based clinical microbiology laboratory.
        J Clin Microbiol. 2007; 45: 3692-3700
        • Malorny B
        • Paccassoni E
        • Fach P
        • Bunge C
        • Martin A
        • Helmuth R
        Diagnostic real‐time PCR for detection of Salmonella in food.
        Appl Environ Microbiol. 2004; 70: 7046-7052
        • Vu DT
        • Sethabutr O
        • Von SL
        • et al.
        Detection of Shigella by a PCR assay targeting the ipaH gene suggests increased prevalence of shigellosis in Nha Trang, Vietnam.
        J Clin Microbiol. 2004; 42: 2031-2035
        • Bruijnesteijn van Coppenraet LE
        • Wallinga JA
        • Ruijs GJ
        • Bruins MJ
        • Verweij JJ
        Parasitological diagnosis combining an internally controlled real‐time PCR assay for the detection of four protozoa in stool samples with a testing algorithm for microscopy.
        Clin Microbiol Infect. 2009; 15: 869-874
        • Verweij JJ
        • Blange RA
        • Templeton K
        • et al.
        Simultaneous detection of Entamoeba histolytica, Giardia lamblia, and Cryptosporidium parvum in fecal samples by using multiplex real‐time PCR.
        J Clin Microbiol. 2004; 42: 1220-1223
        • Jothikumar N
        • da Silva AJ
        • Moura I
        • Qvarnstrom Y
        • Hill VR
        Detection and differentiation of Cryptosporidium hominis and Cryptosporidium parvum by dual TaqMan assays.
        J Med Microbiol. 2008; 57: 1099-1105
        • Kageyama T
        • Kojima S
        • Shinohara M
        • et al.
        Broadly reactive and highly sensitive assay for Norwalk‐like viruses based on real‐time quantitative reverse transcription‐PCR.
        J Clin Microbiol. 2003; 41: 1548-1557