Advertisement

The Risk Factors, Costs, and Survival Analysis of Invasive VRE Infections at a Medical Center in Eastern Taiwan

Open AccessPublished:November 08, 2016DOI:https://doi.org/10.1016/j.ijid.2016.11.005

      Highlights

      • Invasive VRE has a high case fatality rate.
      • Invasive VRE cases have a long hospital stay and higher costs of antibiotic use.
      • Active screening in oncology wards is suggested.

      Abstract

      Objective

      To analyze 48 cases the risk factors of vancomycin-resistant Enterococcus (VRE) infections, the antibiotic costs after infection, and the survival conditions.

      Design

      1:3 matched case-control study a medical center in the eastern Taiwan area. The case group, patients with VRE bacterial strains detected at the sterile sites, and the control group were randomly selected from invasive vancomycin-sensitive Enterococcus (VSE) infected patients at the nearest time point by taking the occurrence time of each VRE infection case as the reference time. Fisher exact tests were conducted in order to verify the existence of differences between the case and control groups; survival analysis was applied to explore the prognoses of the VRE infection cases.

      Results

      The mortality rate of the invasive VRE infection cases was 64.6%, which is obviously higher than that of the invasive VSE infection cases (39.4%); the fact of taking chemotherapy during a hospital stay as well as the use of third-generation cephalosporin, glycopeptides, and medicines of the metronidazole category before the infections, are the risk factors of future invasive VRE infections. Moreover, the antibiotic costs after the infections of invasive VRE infection cases are much higher than those of the VSE infection cases (the average daily cost is 3,433 new Taiwan dollars (NTD) vs. 1,742 NTD).

      Conclusions

      The history of receiving chemotherapy, the use of third-generation cephalosporin, glycopeptides, and medicines of the metronidazole category before the infections are the risk factors of VRE infections. The antibiotic costs after the infections of invasive VRE infection cases are much higher than those of the VSE infection cases.

      Keywords

      1. Introduction

      Vancomycin-resistant Enterococci (VRE) were first identified in 1988
      • Uttley A.H.C.
      • Collins C.H.
      • Naidoo J.
      • George R.C.
      Vancomycin-resistant enterococci.
      and have become a major nosocomial pathogen in Taiwan. Taiwan Nosocomial Infections Surveillance (TNIS) shows that the percentage of nosocomial Infections accounted for by VRE in ICUs in medical centers increased from 3% in 2003 to 24% in 2011. The study found cases of VRE infection crude mortality rate of 64% in National Taiwan University Hospital from 1993 to 2000. It demonstrates that hospital-acquired infections caused by VRE increases not only mortality rates but also costs of treatment and care. Nevertheless, there is no study on increased costs and little research on risk factors of VRE infections with large sample size in Taiwan.

      2. Methods

      From a medical center in Hualien, researchers collected bacterial culture data from 2007 to 2012. After all subjects under 18 years of age were excluded, cases and control groups were selected according to set requirements under a 1:3 matched case-control study.
      The variables data, as required by this study, were collected by reviewing the medical records and electronic medical records maintained by the hospital. First, Fisher exact tests were conducted in order to verify the existence of differences between the categorical variables of the risk factors of the VRE infection cases and those of VSE infection cases, as well as the existence of differences between the continuous variables of nonparametric analysis. Then, the prognoses of the VRE infection cases were explored through survival analysis. The statistical analysis all p-values < 0.05 was considered statistically significant.
      In the case group, patients with VRE bacterial strains detected at the normally sterile sites were taken as the individual cases, and the bacterial culture time was mainly taken as the infection time. In the control group, the individual cases were collected by taking the occurrence time of each VRE infection case as the reference time in order to search for invasive VSE infected patients at the nearest time point as the individual cases of the control group. The selection was made randomly in the event that there were more than two individual cases suitable for the control group.

      3. The collected data included

      3.1 Basic variables

      They include age, gender, place of residence, infection location (community, ward or intensive care unit (ICU)), medical care related infections or not, hospital discharge status (mortality or survival), and number of days of hospital stay.

      3.2 Disease-related factors

      The variables included the severity of illness (using APACHE 2 score to calculate the disease severity status), related risk factors (including whether the patient had an operation two weeks before the inspection; whether the patient had a malignant tumor, poor renal function, chronic liver function disease, diabetes, organ transplantation, hemodialysis, or HIV infection; whether the patient used steroid or cancer chemotherapy, was incontinent, had pressure sores or leucopenia during the current stay at hospital); whether the patient had septicemia; and the types of antibiotics used before the infection and the Carlson comorbidity index (CCI), and invasive treatment items (48 hours before specimen collection, or the time of collecting the specimens and the information, such as whether a breathing apparatus, a central venous catheter indwelling, urethral catheter indwelling, or tracheal tube indwelling was used at the time or 48 hours before the time).
      • Tacconelli E.
      • Cataldo M.A.
      Vancomycin-resistant enterococci (VRE): transmission and control.

      3.3 Medical factors

      The variables included antibiotic cost before VRE infection, VRE treatment, and when the antibiotic treatment for VRE started (VRE treatment refers to the use of quinupristin-dalfopristin, linezolid, daptomycin, and tigecycline;
      • Huang Y.P.
      About the New Antibacterial for Treating Vancomycin-resistant Enterococci (VRE).
      and when the antibiotic treatment for VRE started is the timeframe between testing of VRE infection and the administration of antibiotic, which are within 24 hours after bacterial culture, between 24 to 48 hours after bacterial culture, and over 72 hours after bacterial culture, and no treatment).

      3.4 Prognostic factors

      The survival times of each individual case of the two groups were analyzed using the Kaplan-Meier test.

      3.5 Definitions of the variables of each category

      The invasive VRE infection cases were defined as the individual cases where the VRE bacterial strains were detected in blood, ascetic fluid, intra-abdominal abscess, pleural fluid, and joint fluid. Poor renal function was defined as creatinine clearance <60 mL/min; and poor liver function was defined when at least two of the following items were met: bilirubin concentration was >2.5 mg/dL; the aspartate aminotransferase (AST) or alanine aminotransferase (ALT) exceeded more than two times the normal value; and liver disease was diagnosed. Leucopenia was defined as white blood cell count being <500/mm3.
      • Lautenbach E.
      • Bilker W.B.
      • Brennan P.J.
      Enterococcal bacteremia: risk factors for Vancomycin resistance and predictors of mortality.
      To determine VRE, BD Phoenix Automated Microbiology System was used to test Enterococcus strains for Vancomycin minimum inhibitory concentration (MIC). VRE was confirmed when MIC result was >32 ug/Ml.
      The antibiotic treatment costs were calculated as follows: the types, dosages, frequencies and usage days of all antibiotics used; the types, dosages, and usage days of the antibiotics used before and after the infection, respectively, by taking the infection days as the tangent points; and then calculate the costs of antibiotics. The usage cost data of each antibiotic was collected by taking the unit price of each antibiotic, as stipulated by the National Health Insurance Administration in 2013,

      National Health Insurance Administration Ministry of Health and Welfare: National health insurance drug price adjustment paid job.http://www.nhi.gov.tw.Accessed2013/06/01.

      as the standard in order to add up the total usage costs of each antibiotic, in order to reduce the cost differences caused by currency inflation and medicine cost fluctuations.
      Cases of healthcare-associated infections (HAIs) were determined based on the 2009 revision of HAI Surveillance Definition of Taiwan Center for Diseases Control (CDC).

      4. Results

      There were a total of 50 invasive VRE infection cases during 2007-2012. Excluding the two cases where the age was under 18 years old (0 and 2 years old, respectively), the study team collected 48 cases for study. The control group collection was made according to the designed study method of 1:3, with 144 cases selected as the control group. However, two cases were not incorporated, as the related data of the two patients’ medical records could not be retrieved in time. In the end, 48 invasive VRE infection cases and 142 invasive VSE infection cases were collected for analysis.
      There were 31 males (accounting for 64.6% of the gender ratio) and 17 females (35.4%) in the invasive VRE infection cases; and 87 males (61.3%) and 55 females (38.7%) in the control group, which shows no statistically significant difference (p = 0.682) in gender between the two groups. Regarding the average age, that of the case group was 60.5 ± 18.2 (median: 59.5), and that of the control group was 65.7 ± 15 (median: 67), which shows no statistically significant difference (p = 0.078), although the age of the VSE group was older according to the age medians of the two groups.
      In terms of infection occurrence location, the VRE cases mostly occurred in ICUs with 24 people (accounting for 50%); thus, there were 41 (28.9%) VSE cases occurring in the ICUs. Chi-square testing found that the invasive VRE infection cases had a higher probability of infection in ICUs than the VSE infection cases, which reached the statistically significant level (p = 0.027). Regarding hospital discharge, there were 31 (64.6%) mortality cases by the time of hospital discharge in the VRE infection cases, and the number of VSE infection cases was 56 (39.4%). Chi-square testing found that the mortality rate of the VER cases was higher than that of the VSE cases, and reached a statistically significant level (p = 0.003).
      We found 52.1% (n = 25) of VRE infection cases used respirators when the infections occurred, which is significantly higher than the 30.3% (n = 43) of VSE infections cases (P = 0.0006). In addition, regarding the aspect of central catheter use, the VRE cases had a higher use rate of central venous catheters at 81.3% (n = 39), which is also higher than the 62.7% (n = 89) of the VSE infection cases (p = 0.018). With regard to having septic shock, the rate of the VRE infection cases was 43.8% (n = 21), which shows a significant difference (p = 0.028) from the 26.8% (n = 38) of VSE infection cases. Regarding the aspect of steroid use before the infection, the VRE infection cases had a higher use rate of 43.8% (n = 21), which is higher than the 24.6% (n = 35) of VSE infection cases, and reached a statistically significant level (p = 0.012). Regarding the CCI index and APACHE2 score, no significant difference was found between the data of the two groups after relevant testing (as shown in Table 1).
      Table 1Descriptive statistical analysis of data of the invasive VRE case group and the invasive VSE control group from a medical center in Hualien during 2007 - 2012
      VREVSEP-value
      (n = 48)(n = 142)
      Gender
      Male31(64.6)87(61.3)0.682
      Female17(35.4)55(38.7)
      Age60.5 ± 18.2

      (59.5)
      65.7 ± 15.0

      (67)
      0.057
      Infection location0.027*
      Community5(10.4)25(17.6)
      Ordinary ward19(39.6)76(53.5)
      ICU24(50)41(28.9)
      Enterococcus species < 0.0001*
      E.faecium46(95.8)40(28.2)
      E.faecalis2(4.2)102(71.8)
      Infection part0.026*
      Blood26(54.2)102(71.8)
      Ascitic fluid18(37.5)30(21.1)
      Pleural fluid3(6.5)4(2.8)
      Spinal fluid1(2.1)0(0)
      Other6(4.2)
      Medical care related infections23(47.9)62(43.7)0.608
      Operation11(22.9)43(30.3)0.328
      Malignant tumor16(33.3)52(36.6)0.681
      Poor renal function10(20.8)35(24.6)0.591
      Liver disease20(41.7)49(34.5)0.373
      Diabetes9(18.8)42(29.6)0.143
      Mechanical ventilation25(52.1)43(30.3)0.0006*
      Central line use39(81.3)89(62.7)0.018*
      Urinary catheter31(64.6)88(62.0)0.746
      Incontinence35(72.9)89(62.7)0.198
      Pressure sore9(18.8)25(17.6)0.858
      Leucopenia5(10.4)6(4.2)0.112
      Septic shock21(43.8)38(26.8)0.028*
      Using steroids21(43.8)35(24.6)0.012*
      Hemodialysis10(20.8)23(16.2)0.464
      CCI1.46 ± 1.53

      (1.0)
      1.71 ± 1.70

      (1.0)
      0.401

      APACHE217.8 ± 6.8

      (16.0)
      17.0 ± 7.2

      (16.0)
      0.453

      Total length of stay, median55.4 ± 47.7

      (46)
      38 ± 34

      (25)
      0.009*

      Hospital stay days after infection32.5 ± 34.3

      (21.5)
      24.2 ± 25.9

      (15.5)
      0.082

      Hospital discharge status
      Mortality31(64.6)56(39.4)0.003*
      Survival17(35.4)86(60.6)
      Note: * p < 0.05; the () in the category variable is the percentage; data presented as continuous variables are the average ± standard deviation (median). VRE: vancomycin-resistant Enterococcus; VSE: vancomycin sensitive Enterococcus; ICU: intensive care unit; CCI: Charlson comorbidity index; APACHE 2: acute physiological and chronic health evaluation 2.
      With regard to antibiotic usage types before infection, it was found that the average types of antibiotics used before the infection by the VRE infection cases were 4.2 ± 2.7 types (the median was four kinds), while those by the VSE infection cases were 2.0 ± 2.1 types (the median was one kind), which shows the more the types of antibiotics are used, the bigger the chance of VRE infection (p > 0.0001).
      In terms of antibiotic usage costs during a hospital stay, the average daily antibiotic usage cost of VRE infected patients during their stay in the hospital was 2,666 ± 2,077.9 new Taiwan dollars (NTD) (median: 2,183.3 NTD), and that of VSE infected patients was 1,504 ± 1,398 NTD (median: 1,216.3 NTD), showing that antibiotic costs during a hospital stay, in the event that the patient is infected with drug-resistant Enterococcus, are significantly higher than those of the non-drug-resistant Enterococcus cases (p < 0.0001). The same result was found from a comparison of antibiotic usage costs after an infection. The average daily antibiotics usage cost of the VRE cases after an infection was 3,433 ± 2,966 NTD (median: 2,984.6 NTD), while that of the VSE infection cases was 1,742 ± 1,611 NTD (median: 1425.2 NTD). p < 0.0001 was obtained upon testing the data of the two groups, and reacheed a statistically significant level (as shown in Table 2).
      Table 2Descriptive statistical analysis of related antibiotic usage status and costs of the invasive VRE case group and the invasive VSE control group from a medical center in Hualien during 2007 – 2012
      VREVSEP-value
      (n = 48)(n = 142)
      Types of antibiotics used before infection4.2 ± 2.7

      (4)
      2.0 ± 2.1

      (1)
       < 0.0001*

      Antibiotic types used after infection4.3 ± 2.7

      (4.5)
      3.2 ± 2.3

      (3)
      0.006*

      Average daily antibiotic cost during the hospital stay2666.1 ± 2077.9

      (2183.3)
      1503.7 ± 1397.9

      (1216.3)
       < 0.0001*

      Average daily antibiotic cost before infection1589.9 ± 1458.5

      (1136.6)
      902.9 ± 1275.8

      (472.4)
       < 0.0001*

      Average daily antibiotic cost after infection3433.3 ± 2965.5

      (2984.6)
      1742.1 ± 1611.6

      (1425.2)
       < 0.0001*

      Types of antibiotics used before infection
      Aminoglycosides22(45.8)9(6.3)0.611
      Antifungals21(43.8)13(9.2)0.058
      1st-generation Cephalosporins43(89.6)11(7.7)0.315
      2nd-generation Cephalosporins22(45.8)5(3.5)0.375
      3rd-generation Cephalosporins35(72.9)20(14.1)0.026*
      4th-generation Cephalosporins4(8.3)4(2.8)0.115
      Glycopeptides12(25)17(12.0) < 0.0001*
      Glycylcyclines5(10.4)7(4.9)0.013*
      Macrolides4(8.3)2(1.4)0.645
      Oxazolidione0(0)2(1.4)0.063
      Penicillins32(72.9)25(17.6) < 0.0001*
      Quinolones25(52.1)21(14.8) < 0.0001*
      Sulfonamides3(6.3)4(2.8)0.049*
      Tetracyclines0(0)2(1.4)0.063
      Colistin5(10.4)4(2.8)0.235
      Metronidazole12(25)18(12.7) < 0.0001*
      Carbapenem36(75)28(19.7) < 0.0001*
      beta-lactamase Inhibitors5(10.4)10(7.0)0.001*
      Note:* p < 0.05; the () in the category variable is the percentage; data presented as continuous variables are the average ± standard deviation (median). VRE: vancomycin-resistant Enterococcus; VSE: vancomycin sensitive Enterococcus.
      Survival was analyzed using a Kaplan-Meier curve on the invasive VRE infection cases and VSE infection cases. The average survival days of the VRE cases was 52 days, while that of VSE infection cases was 61 days, which was not significant (as shown in Fig. 1), and had an overall test result of p = 0.246.
      Figure thumbnail gr1
      Fig. 1Survival analysis of the invasive VRE infection case group and of the control group of invasive VSE infection cases from a medical center in Hualien during 2007 – 2012. Note: VRE: vancomycin-resistant Enterococcus.
      Among the 190 patients in the collected invasive Enterococcus infection cases, 103 (54.2%) survived and 87 (45.8%) did not. Cox regression analysis was performed on the data of the two groups, and the p-value of the overall model was tested to be lower than 0.001, indicating that the overall Cox regression model was significant. The APACHE2 score was higher when the patient was found to have diabetes, leucopenia, or using the 4th-generation cephalosporin; glycopeptides and carbapenem before an infection constituted a risk factor affecting the invasive Enterococcus prognosis. However, the drug–resistant or non-drug–resistant nature of the bacterial strain infected does not constitute a risk factor affecting the prognosis. (p = 0.45) (Table 3).
      Table 3Survival and mortality predictor data analysis of the invasive Enterococcus infection cases from a medical center in Hualien during 2007 - 2012
      SurvivalMortalityHRP-value
      (n = 103)(n = 87)(95%CI)
      Gender1.0(0.56-1.82)0.984
      Male42(40.8)30(34.5)
      Female61(59.2)57(65.5)
      Age63.5 ± 16.3

      (66)
      65.5 ± 15.6

      (66)
      1.0(0.98-1.03)0.597
      Infection location0.236
      ICU18(17.5)47(54)Reference
      Ordinary ward67(65)28(32.2)0.4(0.17-1.18)0.103
      Community18(17.5)12(13.8)0.64(0.27-1.48)0.295
      Bacterial strain
      E.faecium34(33)52(59.8)Reference
      E.faecalis69(67)35(40.2)0.8(0.36-1.66)0.511
      Medical care related infections41(39.8)44(50.6)0.7(0.34-1.28)0.218
      Operation30(29.1)24(27.6)1.5(0.74-3.04)0.262
      Malignant tumor37(35.9)31(35.6)1.0(0.46-2.16)1.000
      Poor renal function22(21.4)23(26.4)0.9(0.44-1.95)0.835
      Liver disease35(34.0)34(39.1)1.6(0.76-3.30)0.220
      Diabetes31(30.1)20(23.0)0.3(0.15-0.61)0.001*
      Mechanical ventilation22(21.4)46(52.9)0.8(0.37-1.88)0.664
      Central line use54(52.4)74(85.1)1.8(0.77-4.11)0.179
      Urinary catheter56(54.4)63(72.4)0.6(0.22-1.58)0.290
      Chemotherapy6(5.8)6(6.9)0.5(0.10-2.21)0.335
      Incontinence53(51.5)71(81.6)1.9(0.72-5.06)0.190
      Pressure sore15(14.6)19(21.8)1.1(0.54-2.33)0.768
      Leucopenia5(4.9)6(6.9)3.6(1.92-6.70)0.000*
      Septic shock11(10.7)48(55.2)3.0(0.76-11.9)0.116
      Using steroids20(19.4)36(41.4)1.9(0.91-3.80)0.091
      Hemodialysis10(9.7)23(26.4)1.5(0.69-3.32)0.307
      CCI1.6 ± 1.7

      (1)
      1.7 ± 1.6

      (1)
      1.1(0.84-1.34)0.603
      APACHE214.9 ± 6.2

      (15)
      19.9 ± 7.1

      (18)
      1.1(1.01-1.10)0.022*
      VRE17(16.5)31(35.6)0.8(0.38-1.53)0.450
      Types of antibiotics used before infection
      Aminoglycosides19(18.4)12(13.8)0.8(0.31-1.96)0.603
      Antifungals11(10.7)23(26.4)0.9(0.37-2.04)0.754
      1st-generation Cephalosporins36(35)18(20.7)0.7(0.33-1.34)0.257
      2nd-generation Cephalosporins15(14.6)12(13.8)0.7(0.27-2.00)0.547
      3rd-generation Cephalosporins24(23.3)31(35.6)0.9(0.42-1.83)0.730
      4th-generation Cephalosporins2(1.9)6(6.9)0.2(0.04-0.83)0.027*
      Glycopeptides11(10.7)18(20.7)0.4(0.14-0.92)0.033*
      Glycylcyclines3(2.9)9(10.3)3.3(0.86-12.92)0.082
      Macrolides2(1.9)4(4.6)3.6(0.89-14.64)0.073
      Oxazolidinone0(0)2(2.3)0.8(0.05-12.00)0.857
      Penicillins23(22.3)34(39.1)1.0(0.50-1.99)0.985
      Quinolones17(16.5)29(33.3)0.9(0.45-1.71)0.700
      Sulfonamides4(3.9)3(3.4)1.4(0.25-7.72)0.705
      Tetracyclines1(1)1(1.1)1.5(0.14-16.92)0.727
      Colistin3(2.9)6(6.9)0.6(0.14-2.14)0.391
      Metronidazole10(9.7)20(23)1.8(0.78-4.02)0.174
      Carbapenem28(27.2)36(41.4)0.3(0.13-0.61)0.001*
      betalactamine6(5.8)9(10.3)1.0(0.31-2.96)0.947
      Note: * p < 0.05; the () in the category variable is the percentage; data presented as continuous variables are the average ± standard deviation (median). VRE: vancomycin-resistant Enterococcus; VSE: vancomycin sensitive Enterococcus; ICU: intensive care unit; CCI: Charlson comorbidity index; APACHE 2: acute physiological and chronic health evaluation 2. HR: hazard ratio.
      The risk factors of VRE infections were explored by logistic regression analysis, where the variables that may affect the drug-resistant factors were placed into the model (gender, age, infection occurrence location, nosocomial infection or not, use of a central catheter or not during the infection, respirator usage, taking chemotherapy or not during the hospital stay, being diagnosed as having septic shock or not, previous use of steroids, usage of 3rd-generation cephalosporins, glycopeptides, glycylcyclines, penicillins, quinolones, metronidazole, and carbapenem before the infection). The R2 of the overall model was 0.283. The test result of overall fitness was p < 0.001, which reached the statistically significant level (as shown in Table 4).
      Table 4Logistic regression model analysis of risk factors affecting drug resistance
      VREVSERegression Test Analysis
      (n = 48)(n = 142)OR(95%CI)P-value
      Gender
       Male31(64.6)87(61.3)Reference
       Female17(35.4)55(38.7)1.2(0.48-2.82)0.734
      Age60.5 ± 18.2

      (59.5)
      65.7 ± 15.0

      (67)
      1.0(0.95-1.0)0.195
      Infection location
       ICU24(50)41(28.9)Reference
       Ordinary ward19(39.6)76(53.5)0.8(0.23-2.86)0.744
       Community5(10.4)25(17.6)3.0(0.61-14.2)0.177
      Medical care related infections23(47.9)62(43.7)0.6(0.22-1.50)0.258
      Mechanical ventilation25(52.1)43(30.3)2.2(0.62-7.60)0.229
      Central line use39(81.3)89(62.7)1.2(0.39-3.62)0.769
      Chemotherapy6(12.5)6(4.2)6.3(1.10-37.0)0.042*
      Septic shock21(43.8)38(26.8)1.3(0.49-3.50)0.592
      Using steroids21(43.8)35(24.6)0.6(0.19-1.81)0.359
      CCI1.46 ± 1.53

      (1.0)
      1.71 ± 1.70

      (1.0)
      0.8(0.60-1.10)0.174
      APACHE217.8 ± 6.8

      (16.0)
      17.0 ± 7.2

      (16.0)
      1.0(0.91-1.05)0.503
      Types of antibiotics used before infection
       3rd-generation Cephalosporins35(72.9)20(14.1)2.7(1.04-7.14)0.040*
       Glycopeptides12(25)17(12)4.2(1.36-12.91)0.013*
       Glycylcyclines5(10.4)7(4.9)0.6(0.12-2.80)0.492
       Penicillins32(72.9)25(17.6)2.5(0.93-6.69)0.069
       Quinolones25(52.1)21(14.8)2.0(0.77-5.44)0.150
       Metronidazole12(25)18(12.7)4.0(1.32-11.96)0.014*
       Carbapenem36(75)28(19.7)2.4(0.92-6.30)0.075
      Note: data presented as continuous variables are the average ± standard deviation (median); the () in the category variable is the percentage; *p < 0.05; CI = confidence interval; OR = odds ratio. VRE: vancomycin-resistant Enterococcus; VSE: vancomycin sensitive Enterococcus; ICU: intensive care unit; CCI: Charlson comorbidity index; APACHE 2: acute physiological and chronic health evaluation 2.
      Our results show that the use of 3rd-generation cephalosporins, glycopeptides, and metronidazole-category medicines before the infection of a patient who had cancer chemotherapy during a hospital stay could effectively predict and explain the presence of VRE. The odds ratio of patients infected with the vancomycin-resistant enterococci who received chemotherapy was 6.3, p = 0.042, while that of patients who received third-generation cephalosporin-category medicines before infection was 2.7, p = 0.040; 4.2, p = 0.013 for those receiving glycopeptide-category medicines; and 4.0, p = 0.014 for those receiving metronidazole (as shown in Table 4).
      The factors influencing survival of VRE infection cases were analyzed using Cox regression. After gender, age, CCI, and APACHE2 score were adjusted, the overall model was p = 0.001, showing a significant overall Cox regression model. It was found from the data that the risk of mortality in cases not receiving VRE antibiotic treatment was 5.2 times that in cases receiving treatment, which reached a statistically significant level (p < 0.0001). Moreover, each one-point increase in the APACHE2 score caused the risk of mortality to increase by 1.1 times, which reached a statistically significant level (p = 0.047) (as shown in Table 5).
      Table 5Analysis of the factors of invasive VRE infection cases affecting survival using Cox regression
      SurvivalMortalityTest Analysis
      (n = 17)(n = 31)HR(95%CI)P-value
      Gender1.7(0.76-3.94)0.189
      Male11(64.7)20(64.5)
      Female6(35.3)11(35.5)
      Age56 ± 18.2

      (52)
      63 ± 18

      (61)
      1.0(0.99-1.03)0.493
      CCI1.2 ± 1.5

      (0)
      1.6 ± 1.6

      (1)
      1.0 (0.78-1.31)0.954
      APACHE215.8 ± 6.6

      (15)
      18.9 ± 6.7

      (18)
      1.1 (1.00-1.13)0.047*
      VRE treatmentTreated14(82.4)22(71)Reference
      Not treated3(17.6)9(29)5.2(2.08-12.90)<0.0001*
      Note:* p < 0.05; the () in the category variable is the percentage; data presented as continuous variables are the average ± standard deviation (median). VRE: vancomycin-resistant Enterococcus; VSE: vancomycin sensitive Enterococcus; CCI: charlson comorbidity index; APACHE 2: acute physiological and chronic health evaluation 2.
      In addition, exploration of the influence of the early use of related antibiotic treatment on the survival of VRE infection cases found that, after Kaplan-Meier testing, there was no significant difference (p = 0.752) in the adoption of related early antibiotic treatment on survival days (Table 6).
      Table 6Influence of the treatment time of VRE infection cases on survival
      SurvivalMortality
      (n = 14)(n = 22)P-value
      Treatment time0.752
      Treatment within 24 hours2(14.3)8(36.4)
      Treatment within 24-48 hours3(21.4)1(4.5)
      Treatment after 72 hours9(64.3)13(59.1)
      Note: VRE: vancomycin-resistant Enterococcus.

      5. Discussion

      Invasive VRE infection does cause increased mortality, and hospitalization costs should not be underestimated. The risk factor of VRE infection is a history of receiving chemotherapy (p = 0.042) (Table 4) before an infection after controlling for age, gender, and disease severity. It is therefore suggested that active screening in the future should be started in oncology wards, as the risk of patient mortality could be effectively reduced if the VRE colonization could be confirmed early and related isolation and protection measures were immediately taken to reduce the invasive infection. Regarding the bacteremia data, the crude mortality rate was 73.1%, indicating that the mortality rate of bacteremia would be higher than other invasive infections, which is in accord with studies by the National Taiwan University Hospital in 2001
      • Jean S.S.
      • Fang C.T.
      • Wang H.K.
      • Hsue P.R.
      • Chang S.C.
      • Luh K.T.
      Invasive infections due to vancomycin-resistant enterococci in adult patients.
      and Chiayi Chang Gung Memorial Hospital in 2010.
      • Dai W.L.
      • Wang C.C.
      • Ku Y.P.
      • et al.
      Vancomycin-resistant Enterococcus: A five-year Experience at a Teaching Hospital in Southern Taiwan.
      After age, gender, and disease severity were controlled, we found that the use of 3rd-generation cephalosporins, glycopeptides, and metronidazole-category medicines before the infection had a higher correlation with VRE infection (p < 0.05)(Table 4), which is similar to other studies (Dai et al., 2010).
      • Dai W.L.
      • Wang C.C.
      • Ku Y.P.
      • et al.
      Vancomycin-resistant Enterococcus: A five-year Experience at a Teaching Hospital in Southern Taiwan.
      Many people think the early use of the correct antibiotics for treatment can increase survival rate. The lack of an effect of an early antibiotic treatment could be related to the relatively low number of patients evaluated in the study. According to this study, most antibiotic treatments start after 72 hours, accounting for 64.3%, while those starting within 24 hours account for 14.3%. However, Kaplan-Meier analysis demonstrated that the adoption of early antibiotic treatment did not significantly affect patients prognosis (p = 0.752) (Table 6), which is consistent with the study results of Han et al. in 2009.
      • Han S.H.
      • Chin B.S.
      • Lee H.S.
      • Jeong S.J.
      • Choi H.K.
      • Kim C.O.
      • et al.
      Vancomycin-resistant enterococci bacteremia: Risk factors for mortality and influence of antimicrobial therapy on clinical outcome.
      In addition to this, we also found that for survival analyzed on the invasive VRE infection cases and VSE infection cases, which was not significant (as shown in Fig. 1), the main reason may be the small number of study samples. Therefore, although the number of survival days of the VRE cases was significantly lower than that for the VSE infection cases, it did not reach the statistically significant level.
      This study analyzed the risk factors of infection, prognosis status, and antibiotic usage costs by taking invasive VRE infection cases as the main object of study, and is the first in Taiwan to conduct related studies on the antibiotic costs of invasive VRE infections. In this study, the invasive VRE infections were mainly of E.faecium, accounting for 95.8% (Table 1), which is much higher than the invasive E.faecium infection rate (42.9%) found in the five year study by Chiayi Chang Gung Memorial Hospital,
      • Dai W.L.
      • Wang C.C.
      • Ku Y.P.
      • et al.
      Vancomycin-resistant Enterococcus: A five-year Experience at a Teaching Hospital in Southern Taiwan.
      but closer to the 100% rate found by the National Taiwan University Hospital (Dai et al., 2010).
      • Dai W.L.
      • Wang C.C.
      • Ku Y.P.
      • et al.
      Vancomycin-resistant Enterococcus: A five-year Experience at a Teaching Hospital in Southern Taiwan.
      Our study was limited by the fact that it was a retrospective study of medical records, thus some deviations in data might occur in the event that some situations were not noted on the medical records. Moreover, since the data under study belong to the same hospital, and the number of samples is small, the representativeness of the sample would be affected.
      Currently, there is no standard treatment for VRE infection. A prospective randomized placebo-controlled trial should be designed to determine the optimal antibiotic choice to improve patient survival and outcome.

      Acknowledgements

      None reported.
      Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
      Conflicts of interest: None.
      Ethical Approval: The protocol has been approved by the Tzu Chi General Hospital Research Ethics Committee. The committee is organized under, and operates in accordance with, the Good Clinical Practice guidelines and governmental laws and regulations.

      References

        • Tacconelli E.
        • Cataldo M.A.
        Vancomycin-resistant enterococci (VRE): transmission and control.
        International Journal of Antimicrobial Agents (ELSEVIER). 2008; 31: 99-106
        • Huang Y.P.
        About the New Antibacterial for Treating Vancomycin-resistant Enterococci (VRE).
        The Chinese Pharmaceutical Journal. 2009; 25: 32-39
        • Lautenbach E.
        • Bilker W.B.
        • Brennan P.J.
        Enterococcal bacteremia: risk factors for Vancomycin resistance and predictors of mortality.
        Infection control and hospital epidemiology(ICHE). 1999; 20: 318-323
      1. National Health Insurance Administration Ministry of Health and Welfare: National health insurance drug price adjustment paid job.http://www.nhi.gov.tw.Accessed2013/06/01.

        • Jean S.S.
        • Fang C.T.
        • Wang H.K.
        • Hsue P.R.
        • Chang S.C.
        • Luh K.T.
        Invasive infections due to vancomycin-resistant enterococci in adult patients.
        J Microbiol Immunol Infect. 2001; 34: 281-286
        • Dai W.L.
        • Wang C.C.
        • Ku Y.P.
        • et al.
        Vancomycin-resistant Enterococcus: A five-year Experience at a Teaching Hospital in Southern Taiwan.
        Infection Control Journal. 2010; 20: 23-32
        • Han S.H.
        • Chin B.S.
        • Lee H.S.
        • Jeong S.J.
        • Choi H.K.
        • Kim C.O.
        • et al.
        Vancomycin-resistant enterococci bacteremia: Risk factors for mortality and influence of antimicrobial therapy on clinical outcome.
        Journal of Infection. 2009; 58: 182-190
        • Uttley A.H.C.
        • Collins C.H.
        • Naidoo J.
        • George R.C.
        Vancomycin-resistant enterococci.
        Lancet. 1988; 1: 57-58