Colonization with vancomycin-resistant enterococci in chronic hemodialysis patients

Colonization With Vancomycin-Resistant Enterococci in Chronic Hemodialysis Patients Mary-Claire Roghmann, MD, MS, Jeffrey C. Fink, MD, MS, Louis Polish, MD, Towanda Maker, RN, Jeanine Brewrink, RN, J. Glenn Morris, Jr, MD, MPH & TM, and Paul D. Light, MD ● Vancomycin-resistant enterococcus (VRE) has been identified with increased frequency in dialysis populations, but the risk factors for VRE colonization are not well defined in hemodialysis patients. Patients from a universityaffiliated outpatient dialysis center had surveillance stool or rectal cultures for VRE during April 1994 and January 1996. The combined cohort of 168 patients was followed-up for all-cause mortality, subsequent hospitalization, and VRE infection. Demographic and risk factor information, including age, gender, race, diabetes, coronary artery disease (CAD), and human immunodeficiency virus (HIV) infection, were collected on all patients. Sixteen patients had surveillance cultures grow vancomycin-resistant Enterococcus faecium or E faecalis (VREF), and nine additional patients had clinical cultures positive for VREF. The median follow-up time for patients with positive surveillance or clinical cultures for VREF was 421 days versus 423 days for those without VREF. Patients with positive surveillance cultures for VREF had less time on hemodialysis before screening (median ⫽ 207 days v 822 days; P ⬍ 0.01), and more hospitalization in the year before screening (median ⫽ 19 days v 3 days, P ⬍ 0.01) compared with those without VREF. Patients with VREF colonization were more likely to develop infection with VREF (25% v 1%, P ⬍ 0.01) than those without VREF colonization. However, adjusting for age, diabetes, coronary artery disease, and acquired immune deficiency syndrome (AIDS) using Cox-proportional hazards models, the presence of VREF on screening culture was not associated with increased risk of death (RR ⫽ 1.1, P ⫽ 0.86). Thus after adjusting for other comorbidities, VREF colonization was not associated with increased mortality. Patients with end-stage renal disease (ESRD) on hemodialysis who are hospitalized are more likely to have VREF, but longer duration on hemodialysis was not associated with presence of this organism. This suggests that VRE transmission occurs predominantly in the inpatient setting. r 1998 by the National Kidney Foundation, Inc. INDEX WORDS: Enterococcus; drug effects; drug resistance, microbial; hemodialysis patients; vancomycin; epidemiological methods.

I

NFECTIONS ARE COMMON in hemodialysis patients and contribute significantly to the morbidity and mortality of this population. Vancomycin-resistant enterococci (VRE) are gram-positive bacteria that have been shown to colonize chronic hemodialysis populations.1-4 Previous studies have also identified hemodialysis as a risk factor for both VRE colonization and infection.5,6 The risk factors for VRE colonization in outpatients receiving chronic hemodialysis are unknown. In 1993, we had four hemodialysis patients with bloodstream infections caused by VRE. Although highly resistant to antibiotics, enterococci are relatively avirulent compared with other gram-positive organism such as Staphylococcus aureus. All patients had central venous catheters, From the University of Maryland School of Medicine, Baltimore, MD. Received November 13, 1997; accepted in revised form March 3, 1998. Address reprint requests to Mary-Claire Roghmann, MD, MS, Baltimore VA Medical Center (BT111), 10 N Greene St, Baltimore, MD 21201. E-mail: [email protected]

r 1998 by the National Kidney Foundation, Inc. 0272-6386/98/3202-0009$3.00/0 254

and all recovered with discontinuation of their vascular catheter. We initiated this study because of the new appearance of this organism as a cause of bloodstream infection. In this study, we followed a cohort of chronic hemodialysis outpatients for over a year to determine the prevalence of VRE colonization, risk factors for VRE colonization, the incidence of VRE infection, and overall mortality. METHODS The patient population consisted of outpatients who received hemodialysis three times a week at an offsite dialysis center affiliated with the University of Maryland Medical System. Patients receiving dialysis (n ⫽ 168) had surveillance stool cultures for VRE to determine whether they had gastrointestinal colonization with VRE done initially (April 1994) and repeated 20 months later (January 1996). Stool cultures were plated on selective media for recovery of VRE.7 Vancomycin-resistant enterococci such as E gallinarum and E cassiflavus were excluded because they are not considered human pathogens. Clinical cultures, which grew VRE during the study period, were also recorded. Demographic information including age, gender, and race were collected on all patients. In addition, the presence of diabetes (DM), coronary artery disease (CAD), and human immunodeficiency virus (HIV) infection were determined. Diabetes was defined as a diagnosis of diabetes or any glucose greater than 200 or therapy with insulin or oral

American Journal of Kidney Diseases, Vol 32, No 2 (August), 1998: pp 254-257

RISK FACTORS FOR VRE IN HEMODIALYSIS PATIENTS

hypoglycemic agent according to medical record or patient’s physician. CAD was defined as a diagnosis of coronary artery disease or any form of angina or prior myocardial infarction according to medical record or patient’s physician. HIV infection was defined as a positive HIV test according to medical record or patient’s physician. Acquired immune deficiency syndrome (AIDS) was defined according to Centers for Disease Control (CDC) criteria.8 Hospitalization during the year before the first screening culture was also recorded. Patients were followed-up from initial surveillance culture until death or the end of the study for frequency of hospitalization, VRE infection, and survival. VRE infection were distinguished from VRE colonization by standard definitions for nosocomial infections.9 In general, infection is distinguished from colonization by the adverse reaction of the host to the bacteria. For example, a urinary tract infection is distinguished from asymptomatic bacteriuria by the presence of fever, urgency, frequency, dysuria, or suprapubic tenderness. All statistical analysis was performed using SAS statistical software (Carey, NC). For preliminary analysis, summary statistics included medians and ranges for continuous variables, and proportions or relative risks for categorical variables. Group medians were compared using a Wilcoxon rank-sum test, and proportions by Pearson’s chi-squared tests. Logistic regression was used to calculate the odds ratio for VRE colonization while adjusting for multiple variables simultaneously. Time to death from initial determination of VRE colonization were plotted by Kaplan-Meier plots, and groups were compared using log-rank tests. Cox regression was used to calculate risk ratios and to adjust for other covariables (age, CAD, DM, AIDS) that might affect survival.

RESULTS

The cohort consisted of 168 people receiving chronic outpatient hemodialysis and is described in Table 1. Characteristics of people with and without VRE colonization on outpatient surveillance cultures are compared in Table 2. Duration of follow-up was similar in patients with and without VRE colonization (median, 424 v 422 days, NS). People with VRE colonization were significantly more likely to have been in the hospital before the culture. People with VRE colonization had received dialysis for a shorter period before the culture compared with people without VRE colonization. In a logistic regression model adjusting for other comorbid conditions, including hospitalization, AIDS (OR, 1.4; P ⫽ 0.70), and hemodialysis for more than 2 years (OR, 0.4; P ⫽ 0.15), only prior hospitalization of more than 4 days was an independent risk factor (OR, 4.1; P ⫽ 0.04). People with VRE colonization were significantly more likely to develop a VRE infection

255 Table 1. Characteristics of Chronic Hemodialysis Patients Study Population (n ⫽ 168) No. Median

Age in years, median (range) Men Black Reason for renal failure Diabetes Hypertension Glomerulonephritis Other Unknown Diabetes Coronary artery disease HIV seropositive AIDS Prior hemodialysis in days, median (range) Hospitalization in year before first screening culture in days (%) Hospital days in year before first screen, median (range) Duration of follow-up in days, median (range) VRE colonization as detected by surveillance cultures VREF colonization as detected by clinical cultures while hospitalized VRE infection Died during follow-up

55 97 156

% (range)

(23-86) 58 93

40 68 20 20 20 67 45 16 9

24 40 12 12 12 40 27 10 5

667

(28-6,941)

98

58

4

(0-136)

422

(16-1,049)

16

10

9 5 54

4 3 31

during the study period. Five VRE infections were detected during the study period, for an incidence of 2.3 per 100,000 days followed (5 of 215,779 days). Four VRE infections were bloodstream infections, and one was a wound infection. The four bloodstream infections were temporally associated with the patients’ deaths, occurring within 6 weeks of death. All VRE infections occurred during hospitalization for an unrelated event. Survival analysis and Cox regression were used to determine the effect of VRE colonization on survival. Because of potential selection bias from detection in the hospital, patients with VRE detected only in clinical cultures were not categorized as VRE colonized in the survival analysis. In univariate analysis, age older than 68 years, coronary artery disease, and HIV status were significantly associated with a decreased sur-

256

ROGHMANN ET AL Table 2. Characteristics of Chronic Hemodialysis Patients With and Without VRE Colonization on Outpatient Surveillance Cultures VRE Positive (n ⫽ 16) No.

Age in years, median (range) Men Black Diabetes Coronary artery disease HIV seropositive AIDS Hospitalization in year before first screening culture in days, % Hospitalization in year before first screening culture in days, median (range) Duration of hemodialysis before first screening culture in days, median (range) Hospitalization during follow-up in days, median (range) VRE infection Died during follow-up

vival. VRE colonization as detected by screening culture did not significantly increase the risk of death (see Table 3). The relative risk of death given VRE colonization remained low (see Table 4) when Cox regression was used to adjust for the effect of known predictors of mortality in hemodialysis patients. DISCUSSION

In this study, VRE colonization as detected by surveillance cultures was present in 10% of a cohort of outpatient hemodialysis patients. This is similar to the prevalence of VRE colonization reported from other dialysis centers. Tokars et al2 reported that 12% of units had at least one or more patients colonized with VRE in a national survey of dialysis units and that VRE colonized patients were more likely to dialyze in nonprofit, Table 3. Risk Ratios for Mortality in Univariate Analysis in Chronic Hemodialysis Outpatients

60 10 16 6 4 2 2 13

%

(28-75) 63% 100% 38% 25% 13% 13% 81%

19

(0-136)

322 12 4 8

(55-6,421) (0-52) 25% 50%

Risk Ratio

VREF by screening culture Age ⬎68 years DM CAD HIV AIDS

1.9 2.4 1.5 2.1 2.4 5.8

95% CI

0.9 1.4 0.8 1.2 1.1 2.4

4.1 4.1 2.5 3.8 5.4 13

No.

P Value

%

55 87 140 61 41 14 7 85

(23-86) 57% 92% 40% 27% 9% 5% 56%

NS NS NS NS NS NS 0.18 0.05

3

(0-94)

⬍0.01

(28-6,921) (0-99) 1% 30%

0.07 NS ⬍0.01 0.10

729 5 1 46

hospital-based centers in the Northeast. Gregory and Holly3 reported a VRE colonization prevalence of 10% in 144 hemodialysis patients requiring hospitalization.3 Humur et al4 also found that 10% of patients were VRE colonized in active surveillance of more than 400 dialysis patients in Canada.4 Most of the VRE-colonized patients in their survey were nonambulatory ill patients. Though acute hemodialysis may be a risk factor for VRE infection or colonization in hospitalized patients,5,6 we did not show that VRE colonization in this outpatient hemodialysis population was associated with increased mortality. This is consistent with a prior study that showed that VRE infections were not associated with an increased mortality compared with VSE infections when severity of illness was adjusted.10 However, because of limitations in sample size, Table 4. Risk Ratios for Mortality Adjusting for Multiple Risk Factors in Chronic Hemodialysis Outpatients

Cox-Proportional Hazards Variable

VRE Negative (n ⫽ 152)

Cox Proportional Hazards

P Value

0.08 ⬍0.01 0.19 0.01 0.03 ⬍0.01

Risk Ratio

VRE by screening culture Age ⬎68 years DM CAD AIDS

1.1 2.0 1.4 1.8 7.8

95% CI

0.4 1.1 0.7 1.0 2.8

2.8 3.9 2.5 3.3 22

P Value

0.86 0.03 0.33 0.06 0.01

RISK FACTORS FOR VRE IN HEMODIALYSIS PATIENTS

we cannot state that VRE colonization is not associated with an increased mortality. Our results (RR 1.1; 95% CI, 0.4-2.8) are consistent with a risk of mortality of 0.4 to 2.8. Our results are similar to prior estimates of VRE prevalence in hemodialysis units. Of the factors we studied, only prior hospitalization was a significant risk factor for VRE colonization, suggesting that nosocomial transmission is a major factor in VRE acquisition. However, there were three VRE-colonized patients who had not been hospitalized in the year before the surveillance culture, indicating that acquisition in the hemodialysis unit or community is possible although patients have been known to be colonized with VRE for periods greater than 1 year.11,12 Despite this, our study implicates hospitalization as a major risk factor for VRE colonization in this outpatient population. REFERENCES 1. Woodford N, Morrison D, Johnson AP, Briant V, George RC, Cookson B: Application of DNA probes for rRNA and vanA genes to investigation of a nosocomial cluster of vancomycin-resistant enterococci. J Clin Microbiol 31:653658, 1993 2. Tokars J, Jarvis W, Stivelman J, Farrino M: Vancomycin-resistant enterococci and vancomycin use in hemodialysis centers. J Am Soc Nephrol 7:1446, 1996 3. Gregory M, Holly J: Incidence of vancomycinresistant enterococcal colonization in hospitalized hemodialysis patients. J Am Soc Nephrol 7:1448, 1996 4. Humur A, Dedier H, Campbell I, Bruntry J, Willey B, Lour D, Lior L, Litt M, Garcia M, Gillis G, Jolley B, Kennedy C, Carly J: Clinical and microbiologic analysis of a vancomycin-resistant enterococcus outbreak in a dialysis

257

population. 36th Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC). New Orleans, LA, 1996, abstr J003 5. Edmond M, Wallace S, Pfaller M, Jones R, Wenzel R: Surveillance at 49 U.S. Medical Centers for Vancomycinresistant Enterococcal Bacteremia. Infectious Diseases Society of America (IDSA) 34th Annual Meeting, New Orleans, LA, 1996, session 28, poster 49 6. Tornieporth N, Roberts R, John J, Hafner A, Riley L: Risk factors associated with vancomycin-resistant Enterococcus faecium infection or colonization in 145 matched case patients and control patients. Clin Infect Dis 23:767-772, 1996 7. Morris JG Jr, Shay DK, Hebden JN, McCarter RJ Jr, Perdue BE, Jarvis W, Johnson JA, Dowling TC, Polish LB, Schwalbe RS: Enterococci resistant to multiple antimicrobial agents, including vancomycin: Establishment of endemicity in a university medical center. Ann Intern Med 123:250-259, 1995 8. Anonymous. 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR 41:1-19, 1992 9. Garner J, Jarvis W, Emori T, Horan T, Hughes J: CDC definitions for nosocomial infections, 1988. Am J Infect Control 16:128-140, 1988 10. Shay DK, Maloney SA, Montecalvo M, Banerjee S, Wormser GP, Arduino MJ, Bland LA, Jarvis WR: Epidemiology and mortality risk of vancomycin-resistant enterococcal bloodstream infections. J Infect Dis 172:993-1000, 1995 11. Roghmann M, Morris J Jr: Longitudinal Follow-up of Cancer Patients Colonized by Vancomycin Resistant Enterococcus. Infectious Diseases Society of America (IDSA) 34th Annual Meeting, New Orleans, LA, 1996, session 28, poster 51 12. Roghmann M, Quaiyumi S, Johnson J, Schwalbe R, Morris J Jr: Recurrent vancomycin-resistant Enterococcus faecium bacteremia in a leukemia patient with persistent VRE colonization for two years. Clin Infect Dis 24:514-515, 1997