Statewide Experience with Clostridium difficile Colitis in Academic and Non-Academic Medical Centers

Abstract

Background:

Clostridium difficile colitis is a nosocomial infection that can present as minor, readily treated symptoms or as fulminant colitis leading to death. Risk factors for C. difficile colitis have been defined, and certain populations of hospitalized patients appear to be particularly susceptible. However, most information on C. difficile colitis is from large tertiary-care medical centers. Whether the community hospital experience is similar to that of large referral centers is unknown.

Methods:

We abstracted all cases of C. difficile colitis (ICD-9-CM 008.45) for 2003–2005 from a state database and divided the hospitals into academic and nonacademic centers. Cases were stratified according to whether the colitis was listed as the primary presenting diagnosis or a secondary diagnosis. Demographic information, associated diagnoses, length of stay, and deaths were analyzed.

Results:

The incidence of C. difficile colitis increased from 2003 to 2005, and the majority of cases occurred at nonacademic centers. Patients in nonacademic centers more frequently had C. difficile colitis as the primary diagnosis, had a shorter length of stay, were older, and were more frequently women. The mortality rate was higher for secondary (8.5%) than for primary (2.6%; p < 0.05) C. difficile colitis, but there was no difference between academic and nonacademic centers.

Conclusions:

The incidence of C. difficile colitis is increasing in this statewide database. Compared with academic medical centers, nonacademic centers deal with a higher rate of primary C. difficile colitis that is associated with a lower mortality rate and shorter stay than secondary colitis.

C lostridium difficile-associated colitis is a common hospital-acquired infection in the United States [1]. In recent years, community-acquired cases of C. difficile colitis also have been identified [2]. Whether the latter represent situations where the organism was acquired in a community setting or simply cases where prior admission to the hospital was inadequately documented remains to be determined. Regardless, the incidence of C. difficile-associated colitis appears to be increasing [1,2]. Clostridial colitis can present with a spectrum of symptoms ranging from mild diarrhea and readily treated, uncomplicated disease to fulminant colitis that may necessitate emergency colectomy and is associated with a high mortality rate [2 –6].

An increasing incidence of C. difficile-associated colitis has been documented by data derived from single institutions [3], hospital consortiums [7], and samples selected from national databases [1]. Most of the reports evaluating its impact on outcome have been generated from large tertiary-care medical centers. These facilities often have programs, such as transplantation, that may be associated with a higher risk of C. difficile. Whether this experience with C. difficile colitis is reflective of the experience in community hospitals is unknown. We therefore evaluated the statewide experience with C. difficile colitis using a state discharge database and examined the impact of this disease at both academic medical centers and community (non-academic) hospitals.

Patients and Methods

The Kentucky Hospital Administration maintains a database of healthcare facility discharges from all institutions in the state. Data from 2003 to 2005 were abstracted. Patients who were discharged from rehabilitation hospitals, psychiatric hospitals, or facilities that were not acute care hospitals were excluded. We also excluded patients whose discharge destination was another inpatient facility, as these inter-hospital transfers would exist twice in the database for the same episode of illness. The study population was the subset of patients who had International Classification of Diseases (ICD)-9-CM 008.45 (intestinal infection due to Clostridium difficile or pseudomembranous colitis) listed as the principal diagnosis or as one of eight secondary diagnoses. Demographic data, including age, sex, diagnoses, procedures (up to six), and length of stay (LOS), were obtained. Discharge destination was collected also. Subjects were considered discharged to a skilled nursing facility if they were sent to a skilled nursing, intermediate care, long-term acute care, or hospice facility. All others were classified as discharged to rehabilitation facilities, home (with or without home care), or unknown.

We used ICD-9-CM codes to categorize patients into diagnostic groups when the principal diagnosis was not C. difficile colitis. We grouped these diagnoses into infectious (ICD-9 codes 001–139), neoplastic/hematologic (140–239, 280–289), endocrine/metabolic (240–279), neurologic/cerebrovascular (320–389, 430–439), cardiovascular (390–429, 440–459), respiratory/pulmonary (460–-519), gastrointestinal (527–579), genitourinary (580–629), pregnancy/perinatal/congenital (630–677, 740–779), skin/subcutaneous/musculoskeletal (680–739), trauma/poisoning (800–995), complications of medical care (996–999), and psychiatric/other (290–319, 780–799, V codes).

The admitting facilities were divided into academic and non-academic hospitals on the basis of the presence or absence of university-affiliated attending faculty and residents as well as of a substantial component of the institution's teaching programs and specialized services. Hospitals with free-standing residency programs not affiliated with a medical school and hospitals where university-affiliated residents rotated with community physicians were considered nonacademic. Veterans Affairs facilities do not submit to this database and were not included in this analysis.

Data are presented as the mean ± standard deviation (SD). Categorical variables were analyzed using chi-square and the Fisher exact tests. Analysis of variance was used to evaluate numerical variables. We defined incidence as the number of C. difficile cases per total number of discharges for each type of facility. Incidence rates were compared using Pearson chi-square analysis. This study was reviewed and approved by the University of Louisville Institutional Review Board.

Results

The total number of discharges from inpatient facilities and the number of patients in whom C. difficile colitis was diagnosed are listed in Table 1. The total number of discharges from inpatient facilities increased statewide from 2003 to 2005, as did the absolute number of C. difficile cases. The incidence of all types of C. difficile colitis diagnosed in Kentucky hospitals increased during this time, reaching a maximum of 0.55% of all inpatient admissions in 2005 (Table 1). Although C. difficile colitis was more common as a secondary than as the principal diagnosis, the incidence of C. difficile colitis as both the primary and secondary diagnosis increased from 2003 to 2005 (Table 1).

Table 1.

Prevalence of Clostridium difficile Colitis

Year	Total patients	Total colitis cases (%)	Primary colitis (%)	Secondary colitis (%)
2003	615,436	2,675 (0.43)	654 (0.11)	2,021 (0.33)
2004	615,488	3,043 (0.49)^a	761 (0.12)^b	2,282 (0.37)^a
2005	624,738	3,412 (0.55)^a,c	963 (0.15)^a,c	2,449 (0.39)^a,d

P < 0.0001 vs. 2003.

P < 0.005 vs. 2003.

P < 0.0001 vs. 2004.

p < 0.001 vs. 2004.

When we compared patients who had C. difficile colitis as the primary vs. secondary diagnosis, we found that the former were slightly younger and were more likely to be women (Table 2). Primary C. difficile colitis was more likely to be classified as an urgent or emergency admission, although the percentage of admissions categorized as elective was low for both groups (20.4% of elective admissions for primary disease; 23.1% of elective admissions for secondary disease). Patients who had C. difficile colitis listed as their primary diagnosis were more likely to be discharged home, had a lower mortality rate, and had a shorter total LOS than patients with C. difficile colitis listed as a secondary diagnosis (Table 2). Corresponding to the data on LOS, hospital charges were significantly lower for primary ($15,586 ± $21,164) than for secondary ($41,910 ± $63,141; p < 0.0001) C. difficile colitis. There was no difference from 2003 to 2005 in age, sex, mortality rate, or LOS for either primary or secondary C. difficile colitis (data not shown) with the exception that the average LOS for cases of secondary C. difficile colitis decreased slightly, from 12.7 ± 11.7 days in 2003 to 11.6 ± 10.2 days in 2005 (p = 0.003). There was a statistically significant increase in total hospital charges from 2003 ($13,652 ± $17,278) to 2005 ($16,281 ± $20,403; p = 0.023) for patients primary C. difficile colitis.

Table 2.

Characteristics of Patients with Clostridium difficile Colitis According to Type

	Primary colitis	Secondary colitis
Mean age ± standard deviation (years)	65.3 ± 20.4	67.5 ± 19.3^a
Percent male	33.6	40.4^a
Admission type (% emergency/urgent)	79.3	76.4^b
Mean length of stay ± standard deviation (days)	6.3 ± 6.1	12.1 ± 11.2^a
Mortality rate (%)	2.6	8.5^a
Discharge destination (%)
Home	74.0	51.2^a
Skilled nursing facility	22.1	36.5
Rehabilitation	1.3	3.8

P < 0.0001 vs. primary.

P = 0.037 vs. primary.

For patients with a primary diagnosis of C. difficile colitis, the mortality rate was the same for men (3.0%) and women (2.3%; p = 0.33), and LOS was statistically similar (men 6.2 ± 6.4 days, women 6.4 ± 6.0 days; p = 0.56). A total of 34 patients (1.4%) with primary C. difficile colitis had colon surgery performed during their hospitalization, and the frequency of colon surgery in this group did not change with time (data not shown). However, in patients with primary disease, the mortality rate was significantly greater in those patients who underwent colon surgery (29.4%) than in those who did not (2.2%). There was no statistical difference in the mortality rate associated with colon surgery over time (data not shown). Patients who underwent colon surgery tended to be older (71.8 ± 12.4 years) than those who did not (65.2 ± 20.4 years), but this difference did not reach statistical significance (p = 0.062).

For patients with a secondary diagnosis of C. difficile colitis, the mortality rate was similar for men and women (8.7% vs. 8.4%, respectively), but men had a statistically significantly longer LOS (men 12.9 ± 12.3 days, women 11.6 ± 10.4 days; p < 0.0001). This longer LOS was associated with statistically significantly greater hospital charges for men ($48,100 ±$73,300) than for women ($37,700 ± $54,800; p < 0.0001). The principal diagnoses for patients who had secondary C. difficile colitis are listed in Table 3. Patients who had secondary C. difficile colitis most commonly had a principal diagnosis of respiratory/pulmonary disease (18.8%), but C. difficile colitis was associated with all diagnostic groups. There were changes in the distribution of principal diagnoses in patients with secondary C. difficile colitis from 2003 to 2005, but these changes were numerically small, appeared to be clinically insignificant, and were statistically significant primarily because of the large sample (data not shown).

Table 3.

Principal Diagnosis in Patients with Secondary Clostridium difficile Colitis

Disease	n (%)
Respiratory/pulmonary	1,272 (18.8)
Gastrointestinal	879 (13.0)
Infectious	830 (12.3)
Cardiovascular	780 (11.6)
Endocrine/metabolic	689 (10.2)
Genitourinary	561 ( 8.3)
Other/psychiatric	453 ( 6.7)
Neoplastic/hematologic	380 ( 5.6)
Complications of medical care	292 ( 4.3)
Trauma/poisoning	221 ( 3.3)
Neurologic/cerebrovascular	173 ( 2.6)

Because much of the literature on C. difficile colitis has been generated from tertiary-care medical centers that have service lines such as transplantation or trauma that may pose a greater risk for C. difficile colitis, we divided the institutions into academic and non-academic facilities. Whereas the total incidence of C. difficile colitis cases remained stable in the academic centers over time, the incidence of C. difficile colitis in non-academic facilities increased from 0.43% of all patients in 2003 to 0.56% in 2005 (Table 4). Because of the greater number of patients treated in non-academic centers, the change in yearly incidence was associated with a decrease in the proportion of all C. difficile colitis cases treated in academic centers (15.5% in 2003 to 12.5% in 2005). The incidence of C. difficile colitis as a primary diagnosis was higher in nonacademic facilities. The frequency of primary C. difficile colitis increased over time in both academic and non-academic facilities (Table 4). Academic facilities had a higher percentage of secondary C. difficile colitis cases than non-academic facilities except in 2005, when the incidence rates were similar (Table 4).

Table 4.

Prevalence of Clostridium difficile Colitis in Academic and Nonacademic Medical Centers

	Year	Total cases	Total C. difficile cases (%)	Primary C. difficile (%)	Secondary C. difficile (%)
Academic	2003	88,063	414 (0.47)	42 (0.05)	372 (0.42)
	2004	86,700	431 (0.50)	58 (0.07)	373 (0.43)
	2005	86,586	425 (0.49)	74 (0.07)^a	351 (0.41)
Nonacademic	2003	527,373	2,261 (0.43)	612 (0.12)^b	1,649 (0.31)^b
	2004	528,788	2,612 (0.49)^a	703 (0.13)^b,c	1,908 (0.36)^a,b
	2005	538,152	2,987 (0.56)^a,b,d	889 (0.17)^a,b,d	2,098 (0.39)^a,e

P < 0.003 vs. 2003.

P < 0.02 vs. academic.

P < 0.05 vs. 2003.

P < 0.0001 vs. 2004.

P < 0.02 vs. 2004.

The mortality rate associated with C. difficile colitis was similar in academic and non-academic medical centers (Table 5) and was similar in the two classes of facilities for both primary and secondary C. difficile (data not shown). Patients treated in non-academic facilities were older, more frequently women, and more likely to have primary C. difficile colitis and had a shorter overall LOS (Table 5). The LOS for primary C. difficile colitis was similar for both academic (6.3 ± 7.1 days) and nonacademic (6.3 ± 7.1 days) centers (p = 0.93), but the LOS was different for secondary colitis (academic 16.8 ± 15.3 days; non-academic 11.2 ± 10.0 days; p < 0.0001). Patients treated at non-academic facilities were more likely to be discharged to skilled nursing facilities and less likely to be discharged home or to rehabilitation facilities (Table 5).

Table 5.

Characteristics and Outcomes of C. difficile Colitis at Different Types of Facilities

	Academic	Nonacademic
Mean age ± standard deviation (years)	53.7 ± 23.7	69.1 ± 18.0^a
Percent male	47.4	37.2^a
Type of disease (%)
Primary	13.7	28.0^a
Secondary	86.3	72.0
Mean length of stay ±standard deviation (days)	15.3 ± 14.9	9.8 ± 9.3^a
Mortality rate (%)	8.0	6.8
Discharge destination (%)
Home	63.7	56.1^a
Skilled nursing facility	22.7	34.3
Rehabilitation facility	5.6	2.8

P < 0.0001 vs. academic.

Discussion

Clostridium difficile colitis is an increasingly common nosocomial infection that has reached epidemic proportions in some parts of North America [8,9]. Several risk factors have been defined and include age, medical co-morbidities, use of antibiotics, use of drugs that alter gastric acidity, and an environment where C. difficile may be spread through contact such as hospitals [2,10,11]. Several reports document the fact that the incidence of C. difficile colitis is increasing. Many of these reports have come from large tertiary-care medical centers [3,6,8,12,13] or from national databases that do not stratify facilities according to size, acuity of their patient populations, or the type of clinical programs that might influence the risk of contracting C. difficile [1,14,15]. Whether the experience with C. difficile colitis reported from tertiary-care centers is reflective of that in community or rural hospitals is unknown. It also is unknown if there are differences in either the patients or the disease process for C. difficile colitis cases treated in different facilities. We examined a state discharge database to begin to address these issues.

We found that the percentage of hospitalizations in which C. difficile colitis was diagnosed increased in a statistically significant manner, from 0.43% to 0.55% of all hospital discharges, from 2003–2005. This increase represents almost 400 new cases annually from Kentucky acute-care hospitals. Increases were seen for patients who had C. difficile colitis listed as the primary diagnosis necessitating hospitalization, as well as those who had C. difficile colitis listed as a secondary diagnosis. Patients with secondary C. difficile colitis were older, were more frequently men, had a longer LOS, and were more likely to be discharged to rehabilitation centers or nursing facilities. The mortality rate associated with secondary C. difficile colitis was higher (8.5%) than that of primary colitis (2.6%) and was stable over time. Colectomy was associated with a higher mortality rate with primary C. difficile colitis, which is consistent with other reports of surgical resection for severe disease [3 –5]. Only patients undergoing colectomy who had a diagnosis of primary C. difficile colitis were examined in an attempt to minimize the number of cases in which the colectomy was performed for another indication and the C. difficile colitis was a postoperative complication treated medically. More specific information would be required, however, to ascertain the indication for colectomy in these cases.

When we stratified hospitals into academic and non-academic facilities, we found the overall incidence of C. difficile colitis increased in non-academic hospitals but was stable in academic medical centers. Academic medical centers had a lower incidence of primary C. difficile colitis and a higher incidence of secondary C. difficile colitis. The incidence of secondary C. difficile colitis did not change in academic medical centers, but the incidence of primary C. difficile colitis increased slightly from 2003 to 2005. Non-academic medical centers experienced increases in both primary and secondary C. difficile colitis over this same time period. Patients with C. difficile colitis treated in non-academic centers were older, more frequently women, and had a shorter hospital LOS.

The reasons for the differences between academic and non-academic facilities need to be interpreted carefully. Because facilities designated “academic” in this analysis include the two Level 1 trauma centers, the two transplant centers, and the major children's hospitals in the state, one would expect that some of the differences would be attributable to these specialized patient populations. In these facilities, a substantial number of patients would be referred for specialized services that would frequently define the principal diagnosis. This fact may contribute to the relatively lower frequency of patients in academic medical centers having primary C. difficile colitis (Table 4). One could also speculate that tertiary-care centers that accept complicated cases, higher-acuity patients, or patients with more medical co-morbidities may have longer LOS or a higher mortality rate because of the complexity of these cases. It is possible that these factors increase the susceptibility of patients to C. difficile overgrowth or simply increase the duration of exposure of these patients to a C. difficile-rich hospital environment. Whether these or other factors are involved will be difficult to discern using an administrative database such as this.

We cannot tell from our data whether the higher frequency of C. difficile colitis represents greater awareness of the disease by physicians, nurses, and other healthcare providers; more coding of the disease by medical records staff over time; or a true increase in the frequency with which new patients are being found to have symptomatic C. difficile colitis. Healthcare facility awareness programs and infection control programs have become more common as concern about this disease has increased and the importance of spread of C. difficile by contact has become evident [10]. These programs can help prevent the spread of C. difficile colitis by sensitizing healthcare workers to the importance of barrier precautions for infected or colonized patients. These programs also can condition healthcare workers to search for the disease, sometimes in minimally symptomatic patients who may have multiple reasons for having diarrhea while hospitalized. The use of barrier precautions, antibiotic management protocols, surveillance, or isolation practices may all be different in academic and non-academic facilities, may impact the prevalence of C. difficile colitis, and cannot be defined by our analysis. Our data demonstrate that, regardless of the etiology, C. difficile colitis is being diagnosed more frequently in Kentucky hospitals, and this phenomenon will need to be addressed.

Our data demonstrate increases in both primary and secondary C. difficile colitis in nonacademic medical centers. This finding should not be interpreted as meaning that the infection control practices or healthcare practices at these institutions are suboptimal. Many patients who present with C. difficile colitis could have been hospitalized previously and therefore exposed to a C. difficile-rich environment [10], and this database does not allow us to identify prior hospitalization. Given the rural nature of the state, as well as referral patterns to academic medical centers, it is conceivable that patients were exposed to C. difficile colitis while being treated at academic medical centers, returned to their local communities, and presented to a local hospital when symptoms of active colitis became apparent. This sequence of events could increase the prevalence of primary C. difficile colitis in non-academic medical centers. It also is possible that infection control policies or protocols may be more advanced in nonacademic facilities or that healthcare workers are more proactive in diagnosing colitis, therefore increasing detection of C. difficile-positive patients. A more thorough investigation of all risk factors, including prior hospitalization, or examining strain similarities of infected patients would be required to define the etiology of these differences.

It must be acknowledged that our findings may underestimate the true prevalence of C. difficile colitis in this population. We utilized a database of hospital discharges, so these figures do not account for patients who were managed in an outpatient setting. The Kentucky Hospital Association database lists eight secondary diagnoses in addition to the primary diagnosis, and the order of diagnoses is dependent on medical records coding practices. Many hospital software programs allow the inclusion of more secondary diagnosis codes than are present in this database. Patients who were treated for C. difficile colitis but had the disease listed after the ninth diagnosis would not have been captured by this analysis. This same limitation, as well as the absence in this database of a measure of medical diseases, such as the Charlson index, hinders our ability to analyze potentially important co-morbidities. Our analysis may also underestimate the mortality rate associated with C. difficile colitis, as only the hospital mortality rate, as opposed to the overall or 30-day mortality rate, was recorded. Whether the death is attributable directly to C. difficile colitis, as opposed to associated diseases or co-morbidities, cannot be distinguished from this database and would require additional data to define.

In conclusion, our data demonstrate that the frequency of C. difficile colitis being diagnosed in Kentucky hospitals is increasing. The prevalence of C. difficile colitis as the primary presenting diagnosis as well as a secondary diagnosis is increasing. Whereas most publications on C. difficile colitis have been generated at academic medical centers, the majority of cases in this dataset came from non-academic facilities. Clostridium difficile colitis is not just a disease of tertiary-care referral or academic medical centers but affects patients in all hospitals. Healthcare workers in all settings should be familiar with the disease and knowledgeable regarding its diagnosis and management.

Footnotes

Acknowledgment

The authors acknowledge the Kentucky Hospital Association as the source of the data used in this analysis.

Author Disclosure Statement

The authors have no financial conflicts of interest to disclose.

References

Zerey

, Paton

, Lincourt

et al. The burden of Clostridium difficile in surgical patients in the United States. Surg Infect, 2007; 8:553–556.

Kelly

, LaMont

. Clostridium difficile: More difficult than ever. N Engl J Med, 2008; 359:1932–1940.

Dallal

, Harbrecht

, Boujoukas

et al. Fulminant Clostridium difficile: An underappreciated and increasing cause of death and complications. Ann Surg, 2002; 235:363–372.

Koss

, Clark

, Sanders

et al. J. The outcome of surgery in fulminant Clostridium difficile colitis. Colorectal Dis, 2005; 8:149–154.

Lamontagne

, Labbe

, Haeck

et al. Impact of emergency colectomy on survival of patients with fulminant Clostridium difficile colitis during an epidemic caused by a hypervirulent strain. Ann Surg, 2007; 245:267–272.

Greenstein

, Byron

, Zhang

et al. Risk factors for the development of fulminant Clostridium difficile colitis. Surgery, 2008; 143:623–629.

Burger

, Fry

, Fusco

et al. Multihospital surveillance of nosocomial methicillin-resistant Staphylococcus aureus, vancomycin-resistant Enterococcus, and Clostridium difficile: Analysis of a 4-year data-sharing project, 1999–2002. Am J Infect Control, 2006; 34:458–464.

McDonald

, Killgore

, Thompson

et al. An epidemic, toxin gene-variant strain of Clostridium difficile. N Engl J Med, 2005; 353:2433–2441.

Loo

, Poirier

, Miller

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.

10.

DuPont

, Garey

, Caeiro

, Jiang

. New advances in Clostridium difficile infection: Changing epidemiology, diagnosis, treatment, and control. Curr Opin Infect Dis, 2008; 21:500–507.

11.

Wren

, Ahmed

, Jamal

, Safadi

. Preoperative oral antibiotics in colorectal surgery increase the rate of Clostridium difficile colitis. Arch Surg, 2005; 140:752–756.

12.

Pepin

, Valiquette

, Alary

et al. Clostridium difficile-associated diarrhea in a region of Québec from 1991–2003: A changing pattern of disease severity. Can Med Assoc J, 2004; 171:466–472.

13.

Song

, Bartlett

, Speck

et al. Rising economic impact of Clostridium difficile-associated disease in adult hospitalized patient population. Infect Control Hosp Epidemiol, 2008; 9:823–828.

14.

Archibald

, Banerjee

, Jarvis

. Secular trends in hospital-acquired Clostridium difficile disease in the United States, 1987–2001. J Infect Dis, 2004; 189:1585–1589.

15.

Zilberberg

, Shorr

, Kollef

. Increase in adult Clostridium difficile-related hospitalizations and case-fatality rate, United States, 2000–2005. Emerging Infect Dis, 2008; 14:929–931.