Predictors of prolonged vancomycin-resistant enterococci colonization in acute stroke patients admitted to an intensive care unit: A retrospective cohort study.

ABSTRACT: To investigate the factors affecting the duration of vancomycin-resistant enterococci (VRE) colonization in stroke patients.A total of 52 stroke patients with VRE colonization were enrolled. We divided the groups into several factors and confirmed whether each factor affected VRE colonization. Independent t test, bivariate correlation analysis, and Cox proportional hazards model were used to confirm statistical significance.Among 52 patients, 28 were ischemic stroke and 24 were hemorrhagic stroke. The mean duration of the VRE colonization was 39.08 ± 44.22 days. The mean duration of VRE colonization of the ischemic stroke patients was 25.57 ± 30.23 days and the hemorrhagic stroke patients was 54.83 ± 52.75 days. The mean intensive care unit (ICU) care period was 15.23 ± 21.98 days. Independent sample t test showed the hemorrhagic stroke (P < .05), use of antibiotics (P < .01), oral feeding (P < .01) were associated with duration of VRE colonization. Bivariate correlation analysis showed duration of ICU care (P < .001) was associated with duration of VRE colonization. Cox proportional hazard model showed oral feeding (P = .001), use of antibiotics (P = .003), and duration of ICU care (P = .001) as independent factors of duration of VRE colonization.Careful attention should be given to oral feeding, duration of ICU care, and use of antibiotics in stroke patients, especially hemorrhagic stroke patients, for intensive rehabilitation at the appropriate time.

Introduction

Enterococci is a normal intestinal flora and is one of the most important causes of nosocomial infection.[ Vancomycin-resistant enterococci (VRE) was first identified in Europe in 1986. Recently, the increasing use of antibiotics has increased the number of patients infected with VRE and has become a major pathogen worldwide.[ Although the incidence of VRE colonization is uncertain, the results of a meta-analysis study on the incidence of VRE colonization in patients admitted to the intensive care unit (ICU) showed that the incidence rate was 5.3% in the ASIA population.[ Nationwide surveillance in Korea has also shown increasing resistance of Enterococcus to vancomycin, ranging from 2.9% in 1997 to 17%–21% in 2007.[ VRE colonization has been identified in most gastrointestinal tracts, but has also been identified in skin and genitourinary tracts.[ Health care workers’ hands are now identified as the major source of transmission.[ Stroke is classically characterized as a neurological deficit attributed to an acute focal injury of the central nervous system by a vascular cause, including cerebral infarction, intracerebral hemorrhage, and subarachnoid hemorrhage.[ Brain damage due to lack of blood supply causes various disability.[ Stroke patients should undergo rehabilitation at the same time as acute care and should be focused on minimizing the sequelae of brain damage through more active rehabilitation after acute care.[ Stroke patients with VRE colonization are undergoing isolation and contact caution care to prevent infection.[ For this reason, VRE colonization is one of the important factors that makes active rehabilitation treatment difficult. Therefore, minimizing the duration of VRE colonization is an important factor to reduce patient disability. However, there is a lack of research on what factors currently affect VRE clearance in stroke patients. Previous studies have shown that antibiotic use, ICU care, affects VRE colonization.[ Dysphagia and gait disturbance are common in patients with stroke. Therefore, in this study, we checked whether there was a change in the duration of VRE colonization according to feeding and ambulation, including the previously known antibiotic use and ICU care duration. In addition, duration differences of VRE colonization by underlying disease, comorbidity burden were also studied.

Methods

Subject

The study included patients admitted to the Kosin University Gospel Hospital from 2013 to 2017. This study was approved by our institutional review board, and the requirement for written consent was waived (IRB No. KUGH 2019-06-010). Patients were diagnosed with stroke and were admitted to the ICU and received a rectal swab at admission. The inclusion criteria were: patients with ischemic or hemorrhagic stroke, patients in whom VRE was cultured through a rectal swab, and patients who have finally confirmed VRE negative. All patients were assumed to be VRE-negative prior to admission. Patients with a stroke onset more than 6 months or patients with insufficient medical records were excluded. A total of 97 patients were included in the study and 13 patients were excluded after 6 months of stroke. Thirty-two patients were excluded because of the difficulty in confirming accurate records of ambulation and feeding type. A total of 52 patients were studied (Fig. 1).

Figure 1

Number of patients in the study.

Methods

The specimens were collected through a rectal swab. VRE colonization was confirmed by culturing the specimen in vancomycin-containing selective medium. When rectal culture was positive, the patient was treated with isolation and periodic rectal swabs were performed every week. After first identification of VRE-negative in the colonization test, 2 more tests were performed and the patient was released from the isolation if it was confirmed as 3 consecutive negative.

The test was performed periodically from the day VRE was confirmed. VRE clearance duration was set up to the first inspection date when the negative was confirmed in 3 consecutive times.

Age, sex, body mass index, stroke type, ICU care duration, feeding type, ambulation, underlying disease, comorbidity burden, and antibiotic usage were determined as factors affecting VRE colonization duration. The stroke type was divided into ischemic stroke and hemorrhagic type. Feeding type was divided into oral feeding and tubal feeding group. It was classified as oral feeding group if dietary intake was possible through oral cavity regardless of viscosity of solid, soft, etc. It was classified as tube feeding group when nasogastric tube, percutaneous endoscopic gastrostomy, percutaneous radiologic gastrostomy. Ambulation was divided into walking group and immobile group. Patients were classified as walking groups if they were able to walk, regardless of whether they were using canes or walkers, and were classified as immobile group when riding a wheelchair or bed ridden state. Walking groups were classified as patients who could walk regardless of whether they were using cane or walker, and were classified as immobile group when wheelchair ambulation or bed ridden state. Patients who can walk but sometimes need a wheelchair were classified into the walking group. Diabetes mellitus and immune diseases were considered among the underlying diseases, and the groups were divided according to the history of diagnosis. Comorbidity burden was calculated using the Charlson comorbidity index, and groups were divided by scores from 0 to 5, and patients with scores of 6 or higher were assigned to the same group. The use of antibiotics was included only when used during the VRE colonization period. When antibiotics were used regardless of oral or intravascular administration, patients were classified into antibiotics administered group. ICU duration also included only those periods for which ICU care was given during the VRE colonization period.

Statistics

Continuous data were presented using mean and standard deviation, and independent t test was performed to statistically confirm the difference between the 2 groups. The log rank P value was calculated by the Kaplan-Meier method. In order to confirm the correlation between variables and VRE colonization, bivariate correlation analysis was performed and Pearson correlation coefficient was calculated to confirm the association. Factors affecting the duration of VRE colonization were calculated using competing risk analysis, with death considered as the competing risk.

Result

A total of 52 patients were included in this study. Among 52 patients, 23 were male and 29 were female. The mean age was 65.63 ± 13.45 years. There were 28 patients with ischemic stroke and 24 patients with hemorrhagic stroke. The mean body mass index was 23.17 ± 4.21. The mean duration of the VRE colonization was 39.08 ± 44.22 days. The mean duration of VRE colonization of the ischemic stroke patients was 25.57 ± 30.23 days and the hemorrhagic stroke patients was 54.83 ± 52.75 days. The mean ICU care period was 15.23 ± 21.98 days (Table 1). On the day that VRE colonization was confirmed as negative, 27 patients were able to oral feed, and 12 patients were able to walk. During the VRE colonization, 36 patients were treated with antibiotics. Independent sample t test showed the hemorrhagic stroke (P < .05), use of antibiotics (P < .01), oral feeding (P < .01) were associated with duration of VRE colonization. Kaplan-Meier method showed that stroke type, feeding, and ambulation had a significant effect on VRE colonization period (Table 2). Sex and ambulation were not statistically significant. Bivariate correlation analysis showed duration of ICU care (P < .01) was associated with duration of VRE colonization (Table 3). Age and body mass index were not statistically significant. Competing risk analysis model showed oral feeding (P < .01), use of antibiotics (P < .01) and duration of ICU care (P < .01) as independent factors of duration of VRE colonization (Table 4).

Table 1

General characteristics of all subjects.

Characteristics	Value
Age	65.63 ± 13.45
Sex
Male	23
Female	29
Stroke type
Ischemic stroke	28
Hemorrhagic stroke	24
Feeding type
Oral feeding	27
Tube feeding	25
Ambulation
Walking group	12
Immobile group	40
Underlying disease
Diabetes mellitus	31
Immune disease	11
Charlson comorbidity index
0	21
1	4
2	18
3	4
4	2
5	0
6	3
Use of antibiotics
Administrated	36
Non-administrated	16

Table 2

Comparison of duration of VRE colonization.

	N	Duration of VRE colonization	P value	Log rank P value
Sex
Male	23	46.52 ± 48.02	.284	.331
Female	29	33.17 ± 40.84
Stroke type
Ischemic	28	25.57 ± 30.23	<.05∗	<.05∗
Hemorrhagic	24	54.83 ± 52.75
Feeding type
Oral feeding	27	18.15 ± 24.39	<.01∗∗	<.01∗∗
Tube feeding	25	61.68 ± 49.86
Ambulation
Walking group	12	24.75 ± 34.52	.204	.080
Immobile group	40	43.38 ± 46.25
Underlying disease
With DM	31	38.52 ± 43.26	.311	.127
Without DM	21	41.97 ± 40.11
With ID	11	40.06 ± 37.88	.216	.117
Without ID	41	27.32 ± 31.23
Charlson comorbidity index
0	21	31.37 ± 41.65	.341	.231
1	4	28.66 ± 36.51
2	18	36.28 ± 43.97
3	4	39.52 ± 37.04
4	2	38.41 ± 40.23
5	0
6	3	39.41 ± 40.55
Use of antibiotics
Administrated	36	51.39 ± 48.22	<.01∗∗	<.01∗∗
Non-administrated	16	11.38 ± 6.40

Table 3

Correlation between variables and duration of VRE colonization (bivariate correlation analysis).

Variables	B	HR (95% CI)	P value
Age	0.004	1.00 (1.00–1.00)	.199
Sex	0.003	1.00 (1.00–1.01)	.267
Duration of ICU care	–0.02	0.98 (0.97–0.99)	<.01∗∗

Table 4

Factors associated with duration of VRE colonization (competing risk analysis model).

Variables	HR (95% CI)	P value
Age	1.04 (0.96–1.07)	.279
Sex	1.07 (0.47–1.87)	.675
BMI	0.78 (0.64–1.05)	.346
Hemorrhagic stroke	0.56 (0.20–1.25)	.211
Oral feeding	4.09 (1.73–7.81)	<.01∗∗
Ambulation	0.60 (0.28–1.83)	.310
Duration of ICU care	0.93 (0.89–0.97)	<.01∗∗
Underlying disease - DM	0.41 (0.16–1.76)	.446
Underlying disease - ID	0.54 (0.11–1.44)	.511
Charlson comorbidity index	0.98 (0.43–1.66)	.133
Use of antibiotics	0.31 (0.10–0.62)	<.01∗∗

Discussion

The incidence of VRE colonization is increasing.[ Despite the limitation of active rehabilitation treatment due to isolation, there is no study of which factors affect the duration of VRE colonization in stroke patients. Active rehabilitation for an appropriate period of time is crucial for stroke patients to return to society. Therefore, it is necessary to study the factors that reduce the duration of VRE colonization. In this study, duration of ICU care, oral feeding, and antibiotic use were found to affect duration of VRE.

Hemorrhagic stroke patients are generally more severe and have a higher mortality rate.[ Intracerebral hemorrhage is an independent predictor of poor neurologic outcome.[ Hemorrhagic stroke patients were found to have a longer duration of VRE colonization compared to patients with ischemic stroke in independent t test. However, it was not statistically significant in the Cox proportional hazards model. The patients with hemorrhagic stroke were severe and the general condition was bad, which induced prolonged ICU care duration, tubal feeding, and use of antibiotics. Perhaps these factors affected the prolonged VRE-colonization of hemorrhagic stroke patients.

Previous studies have shown that the use of antibiotics is a risk factor for VRE colonization.[ In this study, 36 patients were using antibiotics and antibiotic use was identified as an independent factor during VRE colonization (P = .003). This is thought to be the result of the use of antibiotics to alter the state of intestinal bacteria and cause overgrowth of VRE colonization.[ In patients with stroke, the incidence of all nosocomial infection was 30% and the most common infection was pneumonia.[ If any infection occurs, the duration of VRE colonization may be prolonged due to the administration of antibiotics. Therefore, it is necessary to pay attention to the infection in the stroke patient and to reduce unnecessary antibiotic use after completed infection treatment.

The mean ICU care period was 15.23 days and was confirmed to be independent of duration of VRE colonization (P = .001). Previous studies have shown that a longer ICU care period will increase the duration of VRE colonization because of the longer exposure to the pathogen.[ The duration of ICU care was increased by the severity or complication of the stroke, and the longer the ICU care duration, the longer the duration of VRE colonization. Therefore, attention should be paid to the occurrence of general condition deterioration and complication, which can lead to a longer ICU care duration. If the patient's condition is restored, transferring quickly from the ICU to the general ward will help reduce the duration of VRE colonization. The correlation between ICU care duration and VRE colonization duration was statistically significant. The single factor analysis Pearson correlation coefficient was 0.647, but the B value was low at 0.035 in the Cox hazard ratio. When ICU care period was prolonged, oral feeding was not performed in many cases, and complication such as infection occurred frequently and antibiotics were used more frequently. These factors are thought to have affected the outcome.

Stroke patients who were unable to receive oral feeding had a consciousness degradation or swallowing difficulty due to their large lesions or general condition deterioration. However, feeding is a factor that affects the recovery of the condition and it is important to start oral feeding early. Tube feeding can increase the probability of infection with antibiotic-resistant bacteria and can be a source of infection.[ In this study, 27 patients were able to receive oral feeding, and oral feeding was found to be a factor in reducing the duration of VRE colonization (P < .001). Therefore, an active swallowing rehabilitation is important in order to start feeding early, and it is important to quickly identify the patient's dietary allowance through 3 spoon test, VFSS, etc, and to start a proper diet for swallowing disorder.

Ambulation was not statistically significant (P = .204). This result suggests that patients who receive ICU care may actually have ambulation, but most of them are immobilized for stability.

Strengthening

In this study, it was confirmed that it is important to pay attention to feeding, antibiotic use, and ICU care period to reduce VRE colonization duration of stroke patients. This is the first study of factors affecting the duration of VRE colonization in stroke patients. Previous studies have investigated the risk factors for VRE colonization. However, this study investigated the factors that influence the duration of colonization in patients who already have VRE. Therefore, it may help to reduce the duration of VRE colonization in patients.

Limitation

This study was limited in obtaining the necessary information as a retrospective study. Although the duration of VRE colonization may vary due to the complications of the underlying disease of the patients, the medical record lacks information on underlying diseases. In the case of oral feeding, dietary intake such as solid diet and soft diet is used. However, due to lack of records, we could not classify feeding type into subdivisions. In the case of using antibiotics, there were many cases in which various kinds of antibiotics were used at the same time or changed while using 1 antibiotic instead of using only 1 antibiotic. Therefore, it was difficult to consider the duration of antibiotic use. The duration of the VRE colonization may vary depending on the type of antibiotic, since the bacteria that can be eradicated depends on the type of antibiotic. Therefore, subdivision of the antibiotic class will be necessary in the next study. Also, time-varying covariance due to changes during repeated follow-up was not considered.[ And this study was not externally validated.

Conclusion

This study showed that stroke type, feeding type, duration of ICU care, and use of antibiotics affect the duration of VRE colonization in stroke patients. It is important that stroke patients be released from isolation and receive intensive rehabilitation treatment. Therefore, careful attention should be given to oral feeding, duration of ICU care, and use of antibiotics in stroke patients, especially hemorrhagic stroke patients, for intensive rehabilitation at the appropriate time.

Author contributions

Conceptualization: Han Eum Choi, Jae Hyun Lee, Young-Joo Sim, Ho Joong Jeong, Ghi Chan Kim.

Data curation: Han Eum Choi, Ghi Chan Kim.

Formal analysis: Han Eum Choi.

Investigation: Ghi Chan Kim.

Methodology: Ghi Chan Kim.

Project administration: Ghi Chan Kim.

Software: Ghi Chan Kim.

Supervision: Jae Hyun Lee, Young-Joo Sim, Ho Joong Jeong.

Validation: Ghi Chan Kim.

Writing – original draft: Han Eum Choi, Jae Hyun Lee.

Writing – review & editing: Han Eum Choi, Young-Joo Sim, Ho Joong Jeong, Ghi Chan Kim.