Abstract
Background
Allogeneic hematopoietic stem cell transplant (HSCT) recipients are at high risk for developing C. difficile infection (CDI). We studied the incidence, risk factors, NAP1/027 prevalence, and clinical outcomes including acute lower gastrointestinal graft versus host disease (GI GVHD) associated with early CDI in this population.
Methods
Retrospective review of patients who underwent allogeneic HSCT at MSKCC from January 1, 2005 – September 30, 2010. Early CDI was defined as infection occurring from day −10 to +40 from stem cell infusion.
Results
Among 793 patients who received allogeneic HSCTs, early CDI occurred in 11.9%; 56% cases were between day −5 to +5. Overall incidence was 25.2 cases/10,000 at risk days. There was a high prevalence of NAP1/027 strains during peak incidence (61% in 2008). NAP1/027 was the most common strain in both adults and pediatric cases (24 and 23 % respectively). CDI was clinically mild, including those due to NAP1/027. Metronidazole was the primary treatment for 91/94 patients, 7/8 cases refractory to metronidazole had no response to vancomycin; none were due to NAP1/027. Relapse of CDI was common (31%). The cumulative incidence of GI GVHD in patients with and without early CDI was 6.8% and 8.% respectively (p=0.5).
Conclusion
The majority of cases of CDI occurred during conditioning or immediately after transplant. Despite high prevalence of NAP 1/027, we found only mild disease. Most patients were treated successfully with metronidazole, irrespective of NAP-1/027 status. There was no significant association between early CDI and subsequent development of GI GVHD.
Keywords: Clostridium difficile infection, allogeneic hematopoietic stem cell transplant, NAP1/027 strain, immunocompromised host
INTRODUCTION
Patients undergoing allogeneic HSCT are particularly at risk for Clostridium difficile infection (CDI) due to the extended length of stay for transplant, frequent antibiotic use and immunocompromised state. C. difficile rates as high as 12.5–27% have been described in allogeneic HSCT recipients, a risk that is 15–20 times greater than other hospitalized patients (1–6), including solid organ transplant recipients.
Despite the rising incidence of CDI since the emergence of the NAP1/027 strain, the prevalence and impact of infection with this particular strain among allogeneic HSCT recipients has never been examined (7, 8). Although the frequency of CDI is much higher among allogeneic HSCT recipients when compared to autologous stem cell and solid organ transplant recipients, the disease is reported to be mild in allogeneic HSCT recipients (2, 9–12). Despite this, immunocompromising conditions have been associated with severe CDI. As a consequence some experts recommend vancomycin as first line therapy for CDI due to an inability to accurately apply clinical scoring systems that measure severity of CDI in this population (13, 14). CDI has also been implicated in the development of acute gastrointestinal graft vs host disease (GI GVHD), postulated to be triggered by disruption of mucosal barriers and release of proinflammatory cytokines. However, there has been no conclusive evidence supporting this hypothesis (1, 12, 15).
We examined the incidence, molecular epidemiology and clinical characteristics of early CDI especially in relation to occurrence of NAP1/027 strain. We also explore the potential relationship between early CDI and subsequent development of GVHD, with a focus on grade 2 or higher acute lower GI GVHD, in a large cohort of allogeneic HSCT recipients. The study population included patients that had received conventional, T-cell depleted (TCD) and cord blood allografts.
METHODS
Study Design and Patient Population
This is a retrospective review of 793 adult and pediatric patients who underwent allogeneic HSCT for hematologic malignancies at Memorial Sloan-Kettering Cancer Center (MSKCC) from January 1, 2005 until September 30, 2010. During this time, MSKCC was a 432-bed tertiary care facility in New York City with 19,000 annual admissions and 140,000 patient days. The adult bone marrow transplant unit at MSKCC is comprised of a 29-bed unit and the pediatric inpatient unit has 39 beds. All allogeneic HSCT patients are admitted to a private room and routinely placed under protective isolation (mask and gloves).
All demographic, clinical, and laboratory information were obtained from the institutional Clinical Research Database (CRDB). Additional information that is not available in CRDB was obtained by MD chart review. MSKCC Institutional Review Board reviewed the study and granted a HIPAA waiver of authorization.
Antibacterial Prophylaxis
Patients received antibacterial prophylaxis starting from day −2 of allogeneic HSCT until development of febrile neutropenia, overt infection, or engraftment, whichever occurred first. The choice of agent was levofloxacin for patients receiving a non-myeloablative conditioning regimen regardless of stem cell source. Ciprofloxacin was used for cord-blood recipients who underwent myeloablative or reduced intensity conditioning (RIC). Vancomycin prophylaxis was used for recipients of peripheral blood or bone marrow allografts who received a myeloablative or RIC regimen. Engraftment was defined as the first of 3 consecutive days with an absolute neutrophil count (ANC) >500 cells/µl. Fever was a temperature ≥ 38.0 °C.
Graft versus host disease
GVHD was monitored and managed as per MSKCC guidelines, and was graded according to the International Bone Marrow Transplant Registry classification (16). Acute lower GI GVHD was defined as GVHD that was diagnosed within the first 100 days after transplant.
Clostridium difficile Infection (CDI)
Case definition
A case of CDI was defined by clinical history compatible with CDI and a positive test. “Early” CDI was an infection that occurred from day −10 until day +40 of transplant. This timeline was derived from the median length of stay for first transplants during the study period. Complicated C. difficile infection was defined by severe disease characterized by any of the following - systemic sepsis, ICU admission, endoscopic evidence of pseudomembranous colitis or toxic megacolon.
Diagnostic methods
From January 2005 until September 2008, diagnosis of CDI at MSKCC was performed by the cytotoxin neutralization assay. After September 2008, a two step testing algorithm was implemented that included a first step glutamate dehydrogenase (GDH) assay followed by the cytotoxin neutralization assay for all GDH positive samples. C. difficile isolates were routinely stored for typing only after January 2008. Polymerase chain reaction (PCR) ribotyping and Multi Locus Sequence Typing (MLST) of isolates was performed as previously described (17, 18).
Treatment for CDI
First line treatment was metronidazole 500mg oral (or IV for patients unable to tolerate oral medications) every 8 hours for a total of 14 days or discontinuation of broad spectrum antibiotics, whichever was longer. Second line treatment was oral vancomycin 125 mg orally 4 times daily at the discretion of the treating physician.
Statistical Analysis
Incidence rates were calculated using the number of events divided by the total number of patient-days at risk. All second transplants were removed from the numerator and denominator when calculating incidence. Categorical variables were compared using the Chi-square test and continuous variables were compared using the Mann-Whitney rank-sum test for the baseline characteristics of early CDI. Simple logistic regression was used for the univariate analysis of risk factors of early CDI. Multiple logistic regression was used for multivariate analysis of risk factors of early CDI among adult patients.
Time-to-event analyses were conducted by Kaplan-Meier methods and log-rank test. Competing events for developing GVHD included patient death, relapse and second transplant. Patient-days-at-risk for acute lower GI GVHD were from day 0 of transplant until the development of acute GVHD of the lower GI tract (stage≥2), day +100 of transplant, second transplant, or death, whichever occurred first. A p-value < 0.05 was considered as statistical significant. All statistical analysis was performed using SAS version 9.2 (Cary, NC, USA).
RESULTS
A total of 793 allogeneic HSCTs were performed during the study period. This included 598 adults and 195 pediatric patients. Overall, 169 (21.3%) patients had at least one episode of CDI in the first year after transplant. Early CDI occurred in 94 (11.9%) transplant recipients, including 25 patients that were less than 21 years of age. Overall, the incidence of early CDI was 25.2 cases per 10,000 patient days at risk. The highest percentage of cases were observed in the years 2007 and 2008 (14.3 and 16.3%, respectively) and declined subsequently to 9.1 % in 2010 [P = 0.38]. No outbreaks were reported during the study period (Figure 1).
Figure 1.
Rate of early CDI among allogeneic HSCT recipients at MSKCC from January 2005 to September 2010.
CYT = Cytotoxin assay; GDH = Glutamate dehydrogenase assay
C. difficile characteristics and outcome
A total of 94 patients had early CDI, 56% of infections occurred from day −5 to day +5 (Figure 2). Table 1 shows the baseline characteristics of the study cohort and Supplementary Table 1 shows clinical characteristics of patients with early CDI. Diarrhea was mild (Grade 1) in the majority of cases (76.1%). Forty-four percent of patients were neutropenic and none developed acute renal failure at the time of CDI. Fourteen patients had imaging with abdominal X-ray series or computed tomographic scanning; 3 among these had evidence of colitis and none had ileus, toxic megacolon or evidence of perforation. There were no deaths attributed to CDI.
Figure 2.
Frequency distribution of timing of early CDI cases in relation to timing from stem cell infusion
Table 1.
Baseline characteristics of patients with and without early CDI
| Total | CDI | non-CDI | P value | ||
|---|---|---|---|---|---|
| Number of patients (%) | 793 (100) | 94 (12) | 699 (88) | ||
| Sex | 0.88 | ||||
| Male | 453 (57) | 53 (12) | 400 (88) | ||
| Female | 340 (43) | 41 (12) | 299 (88) | ||
| Age | |||||
| Mean (range) | 39.2 (0 – 73) | 37.9 (0.7 – 69.3) | 39.4 (0 – 73) | 0.59 | |
| Adult | 585 (74) | 67 (71) | 518 (74) | 0.56 | |
| Pediatric | 208 (26) | 27 (29) | 181 (26) | ||
| Source | 0.32 | ||||
| Peripheral blood | 581 (73) | 64 (68) | 517 (74) | ||
| Cord blood | 114 (15) | 14 (15) | 100 (14) | ||
| Bone marrow | 98 (12) | 16 (17) | 82 (12) | ||
| T-cell depletion | 0.45 | ||||
| No | 383 (48) | 42 (45) | 341 (49) | ||
| Yes | 410 (52) | 52 (55) | 358 (51) | ||
| Underlying disease | 0.16 | ||||
| Leukemia | 427 (54) | 58 (14) | 369 (86) | ||
| AML | 243 (31) | 37 (15) | 206 (85) | ||
| ALL/Biphenotypic | 138 (17) | 20 (14) | 118 (86) | ||
| CLL | 23 (3) | 0 (0) | 23 (100) | ||
| CML/MPD | 23 (3) | 1 (4) | 22 (96) | ||
| MDS | 91 (11) | 12 (13) | 79 (87) | ||
| NHL | 132 (17) | 11 (8) | 121 (92) | ||
| HD | 39 (5) | 3 (8) | 36 (92) | ||
| MM | 38 (5) | 2 (5) | 36 (95) | ||
| NMHD | 66 (8) | 8 (12) | 58 (88) | ||
| Donor | 0.65 | ||||
| MRD | 280 (35) | 30 (11) | 250 (89) | ||
| MUD | 232 (29) | 31 (13) | 201 (87) | ||
| Mismatched | 281 (35) | 33 (12) | 248 (88) | ||
| Conditioning regimen | |||||
| Myeloablative | 495 (62) | 65 (13) | 430 (87) | ||
| Reduced intensity | 152 (19) | 18 (12) | 134 (88) | ||
| Non-myeloablative | 146 (18) | 11 (8) | 135 (92) | 0.18 | |
| Total body irradiation | 0.07 | ||||
| No | 362 (46) | 42 (45) | 320 (46) | ||
| Low dose | 124 (16) | 8 (8) | 116 (16) | ||
| Full dose | 307 (39) | 44 (47) | 263 (38) |
PBSC: peripheral blood stem cell; BM: bone marrow; AML: acute myeloid leukemia; ALL: acute lymphoblastic leukemia; CLL: chronic lymphocytic leukemia; CML: chronic myelogenic leukemia; MPD: myeloproliferative disorders; HD: Hodgkin’s disease; MDS: myelodysplastic syndrome; MM: multiple myeloma; NHL: non-Hodgkin’s lymphoma; NMHD: non-malignant hematological disease; MRD: matched-related donor; MMRD: mismatch-related donor; MUD: match-unrelated donor; RIC: reduced-intensity conditioning; TBI: total body irradiation
The majority of the patients received metronidazole (91/94) at a dose of 500 mg three times daily for 10–14 days or until alternate agent was used. For 8/91 patients, treatment was changed to vancomycin due to persistent diarrhea (median time to treatment switch = 5.5 days; range 3–8 days). The time of onset of CDI for all but one case with refractory diarrhea was between day −7 and day 0. Seven of these patients had persistent diarrhea despite changing to vancomycin and antimotility agents were eventually used for symptom relief. In 4/7 cases, retesting for CDI was done before initiating antimotility agents and was negative in all cases. For 1 patient, diarrhea resolved after changing to vancomycin without concomitant use of antimotility agents. Molecular typing was done on isolates retrieved from 7/ 8 patients; none were infected with the NAP1/027 strain.
Concomitant antibacterial agents for fever and neutropenia prophylaxis or management of concurrent infection were used in 89% of the patients during CDI treatment (Table 2). Twenty-nine (31%) patients had relapse of CDI. The median time to relapse was 79 days. Nine (9.6%) patients had early relapse within 8 weeks after the index episode.
Table 2.
Clinical characteristics and outcomes associated with NAPI vs non-NAP1 strains
| Characteristic N (%) |
Non NAP 1 (N=32) |
NAP 1 (N=10) |
P value* |
|---|---|---|---|
| Male | 20 (62.5%) | 7 (70.0%) | 1.0 |
| Age (years, median) | 39.0 | 49.5 | 0.77 |
| Pediatric Patient | 10 (31.3%) | 3 (30.0%) | 1.0 |
| Fever | 12 (37.5%) | 2 (20.0%) | 0.45 |
| Abdominal Pain | 8 (25.0%) | 2 (20.0%) | 1.0 |
| Diarrhea (>grade 2)** | 6 (18.8%) | 0 (0%) | 0.31 |
| Treatment change | 8 (25%) | 0 (0%) | 0.16 |
| Concomitant antibiotics | 29 (90.6%) | 10 (100%) | 1.0 |
| Relapse | 8 (25.0%) | 5 (50.0%) | 0.24 |
| Time to infection (days, median) | −0.5 | 1.0 | 0.38 |
| Creatinine > 1.2 | 2 (6.3%) | 0 (0%) | 1.0 |
| ANC < 500 | 15 (46.9%) | 4 (40.0%) | 1.0 |
Based on Fisher’s Exact for dichotomous variables and Wilcoxon-Mann-Whitney for continuous variables
> grade 2 diarrhea = More than 6 stools/day
Strain characteristics by PCR ribotyping and Multi locus sequence typing (MLST)
Isolates from 42 out of 58 infections that occurred after January 2008 were available for analysis by MLST and PCR ribotyping (Figure 3 & Supplementary Figure 1). Storage of samples was random and not determined by clinical severity of disease. The overall prevalence of NAP1/027 among the typed isolates from the study years 2008–2010 was 24%. The highest frequency of infections by NAP1/027 co-related with the peak incidence of CDI in the year 2008, when 61% (8/13) of all typed isolates were NAP1/027. The percentage of NAP1 isolates in the study years 2009 & 2010 were 6.2 and 7.6%, respectively.
Figure 3.
Distribution of MLST types from 42 CDI diagnosed from January 2008 to September 2010. Red triangles represent infections due to NAP1 /027 strain and blue boxes non-NAP1 infections, each horizontal bar connects unique ST types.
There was concordance between the MLST and ribotyping results; all 10 isolates identified as 027 in our study were identified as ST-1 by the MLST scheme. The thirty- two non- NAP1 strains were characterized by MLST, the most common among these was -ST -2 (n=5) followed by ST-3 and 42 (n=3 each); ST8,10,11,43,44,58 (2 each) and isolated strains belonging to ST types 4,16,34,41,46,54,103,123 and 191. Although majority of the NAP1 infections were detected in the second half of 2008, none of these or other patients diagnosed with common strains overlapped in time or space, suggesting that hospital based transmission did not occur during the time of transplant (Figure 3).
The 42 typed isolated included 13/16 (81%) infections from pediatric transplant recipients and 29/42 (69%) infections diagnosed from adult transplant recipients during this time period. NAP1/027 strains accounted for approximately a quarter infections in both populations (23% in pediatric and 24% in adults). Univariate analysis did not reveal any significant differences in clinical severity or outcomes associated with CDI infection due to the NAP1 /027 strain when compared to non-NAP1/027 strains (Table 2).The only notable finding was the higher relapse rate among NAP1/027 infections, however this did not reach statistical significance.
Relationship between CDI and acute lower GI GVHD
To explore the relationship between early CDI and subsequent development of lower GI GVHD, we examined the probability of developing acute lower GI GVHD [Grade ≥ 2 in the first 100 days] in patients with and without early CDI. We only evaluated patients in whom CDI preceded the development of GI GVHD. Overall, 46 study patients were diagnosed with acute lower Gl GVHD in the first 100 days after transplant, 4 of these patients had early CDI. The cumulative incidence of acute lower GI GVHD in patients with early CDI was 6.8% compared to 8% in patients without early CDI (p=0.5) (Figure 4a).The incidence of acute lower GI GVHD was highest in the group of patients undergoing conventional transplant who did not have early CDI, although the difference did not reach statistical significance (Figure 4b).
Figure 4.
a. Cumulative incidence of acute lower GI GVHD (Grade ≥2) in patients with and without early CDI
b. Cumulative incidence of acute lower GI GVHD (Grade ≥2) in patients with and without early CDI sorted by transplant type
CONV = conventional; TCD = T-cell depleted; CDI = Clostridium difficile infection
Transplant characteristics and related risk factors
We compared the characteristics of 94 cases of early CDI with the 699 allogeneic HSCT recipients without early CDI to identify risk factors for early CDI (Table 3). No statistically significant associations were found between CDI and demographic characteristics including, age and sex, or transplant variables such as underlying disease, HLA matching status, conditioning-intensity, use of total body irradiation (TBI)-containing conditioning regimens, or nonmyeloablative conditioning regimen. The lack of significant association persisted in the adjusted logistic regression model for CDI. Antibiotic prophylaxis with fluoroquinolones and vancomycin was examined and no associated risk of CDI was identified.
Table 3.
Risk factor analysis of early CDI: Univariate and multivariate analysis for early CDI
| Univariate | Multivariate | |||||
|---|---|---|---|---|---|---|
| Crude OR | P value | AOR ^a | P value | |||
| Sex | ||||||
| Male | REF | REF | ||||
| Female | 0.85 (0.51–1.44) | 0.55 | 0.69 (0.40–1.20) | 0.18 | ||
| Source | ||||||
| Cord blood | REF | REF | ||||
| BM | 0.84 (0.22–3.25) | 0.96 | 0.52 (0.10–2.71) | 0.93 | ||
| PBSC | 0.76 (0.39–1.49) | 0.62 | 0.31 (0.09–1.06) | 0.09 | ||
| T-Cell depletion | ||||||
| No | REF | REF | ||||
| Yes | 1.34 (0.81–2.24) | 1.09 (0.52–2.30) | 0.81 | |||
| Underlying disease | ||||||
| ALL/Biphenotypic | REF | REF | ||||
| AML | 1.25 (0.56–2.77) | 0.91 | 1.31 (0.55–3.09) | 0.91 | ||
| CLL | NA | NA | ||||
| CML/MPD | NA | NA | ||||
| HD | 0.36 (0.07–1.79) | 0.94 | 0.34 (0.06–2.07) | 0.93 | ||
| MDS | 0.87 (0.33–2.29) | 0.92 | 0.99 (0.33–2.95) | 0.91 | ||
| MM | 0.33 (0.07–1.63) | 0.94 | 0.34 (0.06–2.10) | 0.93 | ||
| NHL | 0.56 (0.22–1.43) | 0.93 | 0.52 (0.16–1.70) | 0.93 | ||
| NMHD | NA | NA | ||||
| Donor | ||||||
| Matched related | REF | REF | ||||
| Unrelated Identical | 1.19 (0.65–2.18) | 0.43 | 1.20 (0.64–2.27) | 0.09 | ||
| Mismatch | 0.92 (0.49–1.74) | 0.55 | 0.47 (0.19–1.18) | 0.06 | ||
| Conditioning regimen | ||||||
| Non-myeloablative | REF | REF | ||||
| Myeloablative | 2.39 (1.13–5.06) | 0.04 | 0.51 (0.10–2.76) | 0.49 | ||
| RIC | 1.86 (0.80–4.32) | 0.56 | 0.54 (0.10–2.90) | 0.60 | ||
| TBI | No | REF | REF | |||
| Low | 0.46 (0.20–1.09) | 0.03 | 0.48 (0.10–2.38) | 0.32 | ||
| Full | 1.34 (0.78–2.30) | 0.02 | 1.18 (0.53–2.63) | 0.33 | ||
| Antibiotic Prophylaxis | ||||||
| None | REF | REF | ||||
| Vancomycin only | 0.75 (0.31–1.83) | 0.93 | 0.54 (0.20–1.45) | 0.61 | ||
| FQ Only | 0.42 (0.12–1.54) | 0.18 | 0.40 (0.10–1.62) | 0.34 | ||
| Vancomycin +FQ | 0.92 (0.38–2.26) | 0.33 | 0.64 (0.24–1.74) | 0.84 | ||
OR: odds ratio; AOR: adjusted odds ratio; REF: reference group; CI: confidence interval; PBSC: peripheral blood stem cell; BM: bone marrow; ALL: acute lymphoblastic leukemia; AML: acute myeloid leukemia; CLL: chronic lymphocytic leukemia; CML: chronic myelogenic leukemia; MPD: myeloproliferative disorders; HD: Hodgkin’s disease; MDS: myelodysplastic syndrome; MM: multiple myeloma; NHL: non-Hodgkin’s lymphoma; NMHD: non-malignant hematological disease; RIC: reduced-intensity conditioning; TBI: total body irradiation, Low (<200); FQ = Fluoroquinolone
DISCUSSION
Allogeneic HSCT recipients are among the most susceptible hosts for CDI (1, 3, 12, 15). Our study demonstrates that most CDIs occur around the time of conditioning and in the first week following stem cell infusion. We report early CDI rates over a wide peri transplant period (day −10 to day +40) and a slightly higher rate than previous studies from major transplant centers that have examined rates of CDI in the first month after allogeneic HSCT [5.6–8.35 % in study by Alonso et al.] and Europe [6.4% in study by Willems et al.](1, 6). Differences in study population, antimicrobial prophylaxis, and testing practices (e.g. varying policies regarding rejection of formed stools for testing) may account at least in part for the difference in rates. MSKCC is a referral center and about 50% of patients who undergo transplant have received pretransplant care elsewhere. Often prior history of CDI in the pretransplant phase of illness cannot be ascertained. The majority of patients (>90%) undergoing allogeneic HSCT at MSKCC are tested for C. difficile irrespective of stool consistency, leading to ascertainment bias.
Although NAP1 was the most commonly isolated strain in our cohort, no epidemiologic evidence of patient-to-patient transmission was found and no outbreaks occurred during the study period. Overall 19 ST types were found among the typed isolates with NAP 1/027 being the most common strain. However, the prevalence of this strain varied over time and overall during the study period, wide diversity of infecting strains was encountered. This finding, along with the timing of infection, suggest that most CDI seen early after transplant occur in previously colonized patients and establishment of colonization likely occurs prior to transplant. Prospective studies using PCR for screening at MSKCC have identified colonization rates as high as 39 % in this population, with detection mostly early after transplant. Approximately 70 % of patients’ who developed early CDI in this cohort were previously colonized with tcdB+ strains of C. difficile (19). Further evidence supporting this notion comes from a recent study done at Karmanos Cancer Center where allogeneic HSCT recipients were screened for C. difficile upon admission and at weekly intervals thereafter. Fourteen percent of patients were found to be colonized with C. difficile on initial screen and accounted for the majority of the patients who developed infection subsequently. Only 8/107 patients who tested negative on initial screen were later suspected to have nosocomial acquisition of C. difficile (20).
The NAP1/027 strain first appeared in the US and North America almost a decade ago (7, 21). Infection due to this strain has been associated with a higher incidence and severity of infection. The strain is a hyper toxin producer which has been associated with the increased severity of illness observed with NAP 1/027 infection (22). The impact of infection due to this particular strain has not been described in previous studies of allogeneic HSCT recipients, except in the study by Willems et al. in which no cases were attributed to it (6). At MSKCC, we observed a doubling of the rate of CDI between the years 2006 and 2007 (Figure 1), without any substantial change in the number of transplants performed or the testing method for CDI. Although strains from these study years (2006 and 2007) were not examined by molecular typing, the NAP1/027 prevalence among tested strains during peak C. difficile incidence in our cohort was 61%. This suggests that the high rates observed from 2007–2009 were likely driven by high prevalence of this strain, which was endemic at many medical centers across the US during these years.
The disease observed in out cohort was uniformly mild, irrespective of the infecting strain type and no poor outcomes were associated with NAP1/027 infection except was higher relapse rate, however the number of cases were small to reach statistical significance. While most cases of CDI had resolution of diarrhea with metronidazole, non response was noted in 8 patients in whom treatment was changed to vancomycin. Diarrhea persisted, however, and antimotility agents were used for symptom relief in 7 of these patients. Diarrhea in the peri-transplant period can occur due to a variety of factors, and physician knowledge about the epidemic NAP1/027 strain may have influenced decisions regarding treatment change. Among 7 patients in whom treatment failure with metronidazole was suspected and strain typing could be performed, none had infection with the NAP1 /027 strain. These cases were not deemed to be metronidazole failures. They occurred during conditioning or stem cell infusion and the persistent symptoms were attributed to the regimens rather than refractory CDI. Vancomycin was used as first line therapy in only 3 cases. Comparison of vancomycin with metronidazole for symptom resolution and relapse, therefore, could not be made due to the small number of patients in the former group. Similar to other studies of allogeneic HSCT recipients, relapse of CDI was common and occurred in 31% of patients.
Our study has several other important findings, 1. Risk factor analysis for early CDI suggested that patients with AML and those who received cord blood grafts had a higher risk of CDI, but the association did not persist in multivariate analysis. While studies have postulated mucosal injury resulting from conditioning regimen, especially TBI, as a risk factor for CDI, we did not find a significant association even with full dose TBI(6). 2. The development of acute lower GI GVHD after CDI has led some to speculate a causal relationship between the two conditions. However, we found that the cumulative frequency of acute lower GI GVHD in our cohort was not significantly different in the group of patients with and without early CDI. The association, if any, may not have become overt due to the use of T-cell depletion and the lower incidence of GVHD at our center. No difference was observed when examined by transplant type including in persons receiving cord blood transplants and TBI-containing conditioning regimens, although only a small number of patients were diagnosed with acute lower GI GVHD in these groups [Figure 3b]. The reverse association of acute lower GI GVHD as a risk factor for CDI evaluated in other reports (2, 12, 15) has to be interpreted cautiously. Testing bias in this population and pitfalls of current diagnostic methodologies especially molecular based testing which detects the toxin gene rather than the toxin can contribute to overestimates of clinically significant infection.
The limitations of this study are its retrospective design, the lack of strain typing of all 94 cases of CDI and the fact that pre-transplant history of CDI could not be ascertained for approximately half of the patients in our cohort who received pre transplant care elsewhere. Antibiotics used for treatment during the transplant were not evaluated as an independent variable, since >90% of patients received antibiotic prophylaxis or treatment of infections in the first 40 days after transplant. Diarrhea occurs relatively commonly in the early transplant period due to the effects of conditioning regimens and thus the patients are tested for enteric pathogens frequently. It is possible that some patients who were carriers of C. difficile had the diarrheal illness ascribed to CDI when in fact the conditioning regimen caused the symptoms, leading to an overestimation of rates.
Our findings are especially relevant due to the widespread implementation of molecular based testing methods by most transplant centers. Recently, PCR based screening using serial stool samples from patients undergoing allogeneic HSCT at our center recognized 37 patients with toxigenic C. difficile, among which only 16 were clinically suspected to have CDI (19). When used for diagnosis of CDI, molecular methods offer many advantages over widely used enzyme immunoassay (EIA) and cytotoxin neutralization assays, including higher sensitivity and rapid turnaround time, detection of colonization is much higher with these assays that detect the toxin gene. With doubling of C. difficile detection rates reported by many centers after implementation of PCR(3, 23), interpretation of a positive test in this population will have to be carefully evaluated in prospective studies that employ screening strategies and rigorously implemented clinical scoring systems to overcome the diagnostic challenge of distinguishing between active infection and colonization(3).
Notably, this study found very mild CDI disease, in contrast to the severe infections reported in solid organ transplant recipients. Even the NAP1/027 strain, which accounted for most infections in our cohort, was not associated with severe illness in this population, perhaps demonstrating the importance of host factors in the development of severe disease. Although conventional C. difficile severity scoring systems are not applicable in our cohort, metronidazole appears to be safe and effective in this population, despite widespread use of concomitant antibacterial agents. The impact of treatment choice on reducing relapse of C. difficile disease needs to be formally investigated in prospective studies. Unlike previous reports we found no association between early CDI and subsequent development of acute lower GI GVHD in our cohort (1, 12, 15).
Supplementary Material
Acknowledgments
Grant Support: This work was supported by National Institute of Health/National Institute of Allergy and Infectious Diseases career development award to MK [K23 AI083880]
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
This study demonstrates the high incidence of CDI early after allogeneic HSCT with wide diversity among infecting strains. Despite high prevalence of NAP1/027, disease is mild but relapses are common. No association is found between CDI and subsequent development of GI GVHD.
Supplemental material
Supplemental Figure 1. PCR ribotpying pattern of 42 C. difficile strains from January 2008 to September 2010.
Supplemental Table 1. Clinical characteristics of patients with early CDI
Conflict of interest: None for all authors
REFERENCES
- 1.Alonso CD, Treadway SB, Hanna DB, Huff CA, Neofytos D, Carroll KC, et al. Epidemiology and outcomes of Clostridium difficile infections in hematopoietic stem cell transplant recipients. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2012 Apr;54(8):1053–1063. doi: 10.1093/cid/cir1035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Chopra T, Chandrasekar P, Salimnia H, Heilbrun LK, Smith D, Alangaden GJ. Recent epidemiology of Clostridium difficile infection during hematopoietic stem cell transplantation. Clinical transplantation. 2011 Jan-Feb;25(1):E82–E87. doi: 10.1111/j.1399-0012.2010.01331.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kamboj M, Son C, Cantu S, Chemaly RF, Dickman J, Dubberke E, et al. Hospital-onset Clostridium difficile infection rates in persons with cancer or hematopoietic stem cell transplant: a C3IC network report. Infection control and hospital epidemiology : the official journal of the Society of Hospital Epidemiologists of America. 2012 Nov;33(11):1162–1165. doi: 10.1086/668023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Loo VG, Poirier L, Miller MA, Oughton M, Libman MD, Michaud S, et al. A predominantly clonal multi-institutional outbreak of Clostridium difficile-associated diarrhea with high morbidity and mortality. The New England journal of medicine. 2005 Dec 8;353(23):2442–2449. doi: 10.1056/NEJMoa051639. [DOI] [PubMed] [Google Scholar]
- 5.Samore MH, DeGirolami PC, Tlucko A, Lichtenberg DA, Melvin ZA, Karchmer AW. Clostridium difficile colonization and diarrhea at a tertiary care hospital. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 1994 Feb;18(2):181–187. doi: 10.1093/clinids/18.2.181. [DOI] [PubMed] [Google Scholar]
- 6.Willems L, Porcher R, Lafaurie M, Casin I, Robin M, Xhaard A, et al. Clostridium difficile infection after allogeneic hematopoietic stem cell transplantation: incidence, risk factors, and outcome. Biology of blood and marrow transplantation : journal of the American Society for Blood and Marrow Transplantation. 2012 Aug;18(8):1295–1301. doi: 10.1016/j.bbmt.2012.02.010. [DOI] [PubMed] [Google Scholar]
- 7.Pepin J, Valiquette L, Alary ME, Villemure P, Pelletier A, Forget K, et al. Clostridium difficile-associated diarrhea in a region of Quebec from 1991 to 2003: a changing pattern of disease severity. CMAJ : Canadian Medical Association journal = journal de l'Association medicale canadienne. 2004 Aug 31;171(5):466–472. doi: 10.1503/cmaj.1041104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Valiquette L, Low DE, Pepin J, McGeer A. Clostridium difficile infection in hospitals: a brewing storm. CMAJ : Canadian Medical Association journal = journal de l'Association medicale canadienne. 2004 Jul 6;171(1):27–29. doi: 10.1503/cmaj.1040957. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Boutros M, Al-Shaibi M, Chan G, Cantarovich M, Rahme E, Paraskevas S, et al. Clostridium difficile colitis: increasing incidence, risk factors, and outcomes in solid organ transplant recipients. Transplantation. 2012 May 27;93(10):1051–1057. doi: 10.1097/TP.0b013e31824d34de. [DOI] [PubMed] [Google Scholar]
- 10.Mitu-Pretorian OM, Forgacs B, Qumruddin A, Tavakoli A, Augustine T, Pararajasingam R. Outcomes of patients who develop symptomatic Clostridium difficile infection after solid organ transplantation. Transplantation proceedings. 2010 Sep;42(7):2631–2633. doi: 10.1016/j.transproceed.2010.04.057. [DOI] [PubMed] [Google Scholar]
- 11.Bobak D, Arfons LM, Creger RJ, Lazarus HM. Clostridium difficile-associated disease in human stem cell transplant recipients: coming epidemic or false alarm? Bone marrow transplantation. 2008 Dec;42(11):705–713. doi: 10.1038/bmt.2008.317. [DOI] [PubMed] [Google Scholar]
- 12.Dubberke ER, Reske KA, Srivastava A, Sadhu J, Gatti R, Young RM, et al. Clostridium difficile-associated disease in allogeneic hematopoietic stem-cell transplant recipients: risk associations, protective associations, and outcomes. Clinical transplantation. 2010 Mar-Apr;24(2):192–198. doi: 10.1111/j.1399-0012.2009.01035.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Treatment of Clostridium Difficile Infection. The Medical Letter. 2011 Feb 21;53(1358) [PubMed] [Google Scholar]
- 14.Zar FA, Bakkanagari SR, Moorthi KM, Davis MB. A comparison of vancomycin and metronidazole for the treatment of Clostridium difficile-associated diarrhea, stratified by disease severity. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2007 Aug 1;45(3):302–307. doi: 10.1086/519265. [DOI] [PubMed] [Google Scholar]
- 15.Chakrabarti S, Lees A, Jones SG, Milligan DW. Clostridium difficile infection in allogeneic stem cell transplant recipients is associated with severe graft-versus-host disease and non-relapse mortality. Bone marrow transplantation. 2000 Oct;26(8):871–876. doi: 10.1038/sj.bmt.1702627. [DOI] [PubMed] [Google Scholar]
- 16.Rowlings PA, Przepiorka D, Klein JP, Gale RP, Passweg JR, Henslee-Downey PJ, et al. IBMTR Severity Index for grading acute graft-versus-host disease: retrospective comparison with Glucksberg grade. British journal of haematology. 1997 Jun;97(4):855–864. doi: 10.1046/j.1365-2141.1997.1112925.x. [DOI] [PubMed] [Google Scholar]
- 17.Bidet P, Barbut F, Lalande V, Burghoffer B, Petit JC. Development of a new PCR-ribotyping method for Clostridium difficile based on ribosomal RNA gene sequencing. FEMS microbiology letters. 1999 Jun 15;175(2):261–266. doi: 10.1111/j.1574-6968.1999.tb13629.x. [DOI] [PubMed] [Google Scholar]
- 18.Griffiths D, Fawley W, Kachrimanidou M, Bowden R, Crook DW, Fung R, et al. Multilocus sequence typing of Clostridium difficile. Journal of clinical microbiology. 2010 Mar;48(3):770–778. doi: 10.1128/JCM.01796-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kinnebrew YJL Melissa A, Jenq Robert R, Lipuma Lauren, Littmann Eric R, Gobourne Asia, No Daniel, van den Brink Marcel, Pamer Eric G, Taur Ying. Clostridium difficile Colonization and Infection During Allogeneic Hematopoietic Stem Cell Transplantation. PLOS One. 2014 doi: 10.1371/journal.pone.0090158. (with/pending revision). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Jain MCC Tania, BA, Banavasi Harsha Vardhan Reddy, MBBS, Urday-Cornejo Varinia, MD, Liubakka Alyssa, Cutright Jessica, BS, Salimnia Hossein, PhD, Lephart Paul, PhD, Revankar Sanjay, MD, Chandrasekar Pranatharthi, MD, Chopra Teena, MD MPH, Alangaden George., MD . ID week. San Francisco, CA: 2013. Oct 2–6, Prospective Evaluation of Toxigenic and Non-toxigenic Strains of Clostridium difficile Colonization and Infection among Allogenic Hematopoietic Stem Cell Transplant Recipients, 2010–2012. 2013;Abstract number 627 (Oral) [Google Scholar]
- 21.Dallal RM, Harbrecht BG, Boujoukas AJ, Sirio CA, Farkas LM, Lee KK, et al. Fulminant Clostridium difficile: an underappreciated and increasing cause of death and complications. Annals of surgery. 2002 Mar;235(3):363–372. doi: 10.1097/00000658-200203000-00008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Warny M, Pepin J, Fang A, Killgore G, Thompson A, Brazier J, et al. Toxin production by an emerging strain of Clostridium difficile associated with outbreaks of severe disease in North America and Europe. Lancet. 2005 Sep 24–30;366(9491):1079–1084. doi: 10.1016/S0140-6736(05)67420-X. [DOI] [PubMed] [Google Scholar]
- 23.Longtin Y, Trottier S, Brochu G, Paquet-Bolduc B, Garenc C, Loungnarath V, et al. Impact of the type of diagnostic assay on Clostridium difficile infection and complication rates in a mandatory reporting program. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2013 Jan;56(1):67–73. doi: 10.1093/cid/cis840. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.