Active Surveillance Cultures and Decolonization to Reduce NICU Staphylococcus aureus Infections

Abstract

Background and Objectives

Staphylococcus aureus (S. aureus) is a common cause of healthcare associated infections (HAI) in neonates. Our objectives were to examine the impact of S. aureus decolonization on the incidence of S. aureus infection and to measure the prevalence of mupirocin resistance.

Methods

We retrospectively identified neonates admitted to a tertiary care NICU between April 1 2011 and September 30 2014. We compared rates of MSSA-positive cultures and infections before and after implementation of active surveillance culture and decolonization intervention for MSSA-colonized neonates. We used two measurements to identify the primary outcome, NICU-attributable MSSA: 1) any culture sent during routine clinical care that grew MSSA and 2) any culture that grew MSSA and met criteria of the NHSN's HAI surveillance definitions. S. aureus isolates were tested for mupirocin susceptibility. To determine the impact of the intervention on MSSA infection, we estimated incidence rate ratios using interrupted time series models.

Results

Pre- and post-intervention, 1523 neonates (29,220 patient-days) and 1195 neonates (22,045 patient-days) were admitted to the NICU, respectively. There was an immediate reduction in mean quarterly incidence rate of NICU-attributable MSSA-positive clinical cultures, of more than 60% (IRR=0.36, 95% CI 0.19, 0.70) after implementation of the intervention and MSSA positive culture rates continued to decrease by 21% per quarter (IRR 0.79 95% CI 0.74, 0.84). MSSA infections also decreased immediately following the intervention implementation (IRR=0.27; 95% CI=0.10, 0.79). No mupirocin resistance was detected.

Conclusions

Active surveillance cultures and decolonization may be effective in decreasing S. aureus infections in NICUs.

Introduction

Healthcare-associated infections (HAIs) are responsible for significant morbidity and mortality in hospitalized neonates. Preterm and low birth weight neonates have an increased susceptibility to infections due to an immature immune system, increased duration of hospitalization and increased need for invasive procedures.^1–4 The CDC estimates that there are more than 33,000 HAIs in US NICUs every year.⁵ Neonates with HAIs have increased length of hospital stay and increased healthcare costs.^2,6,7 Despite appropriate therapy, neonatal infections can have long-term sequelae including adverse neurodevelopmental and growth outcomes.^8,9

Staphylococcus aureus (S. aureus) is the second most common cause of HAIs and late-onset sepsis in neonates, second only to coagulase-negative Staphylococcus (CONS).^10,11 In addition to the high burden of S. aureus, antibiotic-resistant S. aureus strains, especially methicillin-resistant S. aureus (MRSA) have become endemic in many NICUs.^4,12,13 Despite enhanced infection control measures and strategies, S. aureus remains a threat to neonates. Current recommendations to prevent MRSA transmission and infections in the NICU include identifying colonized neonates and placing them on contact precautions, cohorting, hand hygiene, and in some cases, decolonizing colonized neonates and/or healthcare workers.^14,15 These strategies have focused on MRSA prevention and have overlooked potentially preventable methicillin-susceptible S. aureus (MSSA) disease. Recent data suggest that invasive MSSA infections occur 2 ½ times more frequently than invasive MRSA infections in neonates, and MSSA infections have comparable morbidity and mortality in this high-risk population.^3,16–19 Our objectives were to examine the impact of MSSA decolonization on the incidence of MSSA infections and to measure the prevalence of mupirocin resistance in a level IV NICU.

Patients and Methods

Setting and Design

The Johns Hopkins Hospital (JHH) is a tertiary-care academic medical center with an embedded 200-bed Children's Center that houses a 45-bed, level IV NICU. We retrospectively identified a cohort of neonates admitted to the NICU between April 1 2011 and September 30 2014. We performed a quasi-experimental, pre-post study to compare rates of MSSA-positive cultures and infections before and after implementation of active surveillance cultures (ASC) and decolonization of MSSA-colonized neonates. The Johns Hopkins Institutional Review Board approved this retrospective cohort study with a waiver of informed consent.

Infection Control and Prevention Program

The JHH NICU has a program of ASC and decolonization of MRSA-colonized neonates to prevent MRSA transmission and infections as previously described.^4,20 Nares swabs are performed weekly by nurses to identify MRSA-colonized neonates. In addition, nares swabs are performed at the time of NICU admission for neonates transferred from other hospitals or admitted from home. In April 2013, the program was expanded to include ASC to identify and decolonize MSSA-colonized neonates (`the intervention') in addition to MRSA-colonized neonates due to occurrences of serious MSSA infections. Decolonization consisted of mupirocin applied to the nares twice a day for 5 days and 2 baths with 2% chlorhexidine gluconate-impregnated cloths administered 48 hours apart for infants greater than 36 weeks gestational age or greater than 4 weeks chronological age or daily for 5 days for infants greater than 2 months chronologic age^4,20.

Data Collection and Outcome Ascertainment

We searched a computerized surveillance system (TheraDoc, Premier, Inc) to identify patients with surveillance cultures and cultures sent during clinical care that grew S. aureus during the study period. The primary study outcome was NICU-attributable MSSA. We compared two measurements to identify NICU-attributable MSSA: 1) MSSA clinical culture defined as any clinical culture sent as part of clinical care that grew MSSA; 2) MSSA infection defined as any clinical culture that grew MSSA and met the NHSN's surveillance definition for a specific HAI.²¹ We reviewed medical records of patients whose cultures sent during clinical care grew MSSA. NHSN definitions for HAIs were applied by a trained observer (V.O.P.) who applied definitions consistently over the study period, to distinguish infection from colonization. MRSA cultures were similarly adjudicated to assess for secular changes. Cultures were further classified into present-on-admission if they were collected less than three days after admission to the NICU or NICU-attributable if they were obtained 3 days or more after admission to the NICU.

Laboratory Methods

Prior to April 2013, surveillance swabs were plated on MRSA Select (BD Diagnostics, Sparks, MD). In April 2013, surveillance swabs were also plated on BBL™ CHROMagar™ Staph aureus (BD Diagnostics, Sparks, MD), and beginning April 2014, swabs were plated on SaSelect™ (Bio-Rad Laboratories, Hercules, CA) in addition to MRSA Select. After 24 hours of incubation, mauve-colored colonies or pink- to orange-colored colonies respectively were confirmed as S. aureus by Gram's stain and coagulase testing. To monitor for emergence of mupirocin resistance among MSSA isolates, we tested consecutive isolates obtained from neonates after introduction of MSSA decolonization for mupirocin susceptibility using Etest (bioMerieux SA, Marcy-L'Etoile, France). We determined minimum inhibitory concentrations (MICs) using standard CLSI methodology.²² Mupirocin susceptibility was defined as an MIC < 4 μg/mL, low-level mupirocin resistance (LLMR) if the MIC was between 8 and 64 μg/mL, and high-level mupirocin resistance (HLMR) if the MIC was > 512 μg/mL.²³

Statistical Analysis

We compared characteristics of neonates admitted during the pre-intervention and post-intervention periods using chi-square tests for categorical variables and the two-sample t-tests for continuous variables. The outcomes were calculated as the quarterly incidence rate of NICU-attributable MSSA clinical cultures and the quarterly incidence rate of NICU-attributable MSSA infections expressed as the number of outcomes in a quarter per 1000 patient-days. We measured the impact of the intervention on S. aureus in the NICU by first comparing the mean quarterly incidence rates during the pre- and post-intervention periods using Poisson regression and then using a quasi-experimental interrupted-time series (ITS) model for the log-transformed quarterly incidence rates.²⁴ From the ITS, the effect of the intervention is reported as a) the “immediate” effect of the intervention as the relative change in quarterly incidence rate comparing the first quarter of the post-intervention period to the last quarter of the pre-intervention period, and b) the “sustained” effect of the intervention as the relative change in the quarterly incidence rate per quarter during the post-intervention period. Due to the small numbers of MRSA positive cultures, we quantified the impact of the intervention by only comparing the mean quarterly incidence rates of MRSA pre- and post-intervention, using Poisson regression. To determine the robustness of our findings, we carried out a sensitivity analysis by varying the start time of the intervention period in the ITS model. Data were maintained in Microsoft Access 2007 (Microsoft) and analyzed using StataSE (ver 13.1; StataCorp) and Microsoft Excel 2007 (Microsoft). The interrupted time-series models were fit using the itsa module in Stata.

Results

During the 24 month pre-intervention period, 1,523 neonates were admitted to the NICU accounting for 29,220 patient days. In the 18 months after the MSSA screening and decolonization program began, 1,195 neonates were admitted to the NICU accounting for 22,045 patient days. Of the neonates admitted in the post-intervention period, 899 (75.2%) were screened for MSSA colonization and 89 had a surveillance culture grow MSSA. Of those MSSA-colonized neonates, 72 (78.7%) were treated with mupirocin and chlorhexidine per protocol. Neonates admitted in the pre- and post-intervention periods demonstrate minor differences in demographic and clinical characteristics that are unlikely to impact findings (Table 1). There were 683 female neonates admitted during the pre-intervention period and 521 during the post intervention period (44.8% vs. 43.7%, p=0.55); median birth weight during the pre-intervention period was 2820g (IQR = 1360g) and 2860g (IQR=1390) during the post-intervention period (p=0.51); median length of NICU stay (IQR) was 7.2 days (17.7) pre-intervention compared with 6.5 days (16.5) post-intervention (p=0.20).

Table 1.

Characteristics of neonates admitted in the NICU during pre-intervention (April 1, 2011 through March 31, 2013) and post-intervention (April 1, 2013 through September 30, 2014)

Patient characteristics	Pre-intervention period N=1524(%)	Post-intervention period N=1193(%)	P-value
Female (%)	683 (44.8)	521 (43.7)	0.55
Race
Asian	58 (3.8)	54 (4.5)	0.35
Black or African-American	710 (46.6)	522 (43.8)	0.14
White	577 (37.9)	445 (37.3)	0.76
Other	179 (11.8)	172 (14.5)	0.04
Birth Weight, median (IQR)	2820 (1360)	2860 (1390)	0.51
Length of NICU Stay, median (IQR)	7.2 (17.7)	6.5 (16.5)	0.20
Inborn	1287 (84.5)	968 (81.1)	0.02
Mortality	49 (3.2)	37 (3.1)	0.87
Quarterly Device Utilization Ratio	0.47	0.48	0.64
Mean Quarterly Patient Days	3653	3674	0.76

During the study period, 83 patients had 153 NICU-attributable S. aureus clinical isolates. Clinical isolates included 142 (92.8%) MSSA and 11 (7.2%) MRSA cultures. Forty-three of 142 (30.3%) MSSA cultures met the NHSN's definition for a HAI. Sites of MSSA infection included blood stream infections (n=14, 32.6%), lower respiratory tract (n=12, 27.9%), skin and soft-tissue (n=8, 18.6%), pneumonia (n=3, 7.0%), conjunctivitis (n=3, 7.0%), meningitis (n=1, 2.3%), phlebitis (n=1, 2.3%), and intra-abdominal infection (n=1, 2.3%).

During the pre-intervention period there were 106 MSSA-positive clinical cultures compared to 36 MSSA positive clinical cultures during the post-intervention period. Overall, incidence rate of MSSA clinical cultures was 3.62 per 1000 patient days during the pre-intervention period compared to 1.62 per 1000 patient days during the post-intervention period (IRR = 0.45; 95% CI = 0.22, 0.92). In the quarter following introduction of an ASC and decolonization protocol, MSSA clinical culture incidence rates decreased by an estimated 64% (IRR= 0.36; 95% CI = 0.19, 0.70). This reduction in MSSA clinical culture incidence rates was sustained during the post-intervention period with an estimated quarterly decrease of 21% (IRR = 0.79; 95% CI = 0.74, 0.84). On carrying out a sensitivity analysis with varying start dates for the post-intervention period (the start date was adjusted backwards into the pre-intervention period by one quarter at a time for four quarters), a decrease in the post-intervention slope of MSSA clinical culture rates was observed at each of the tested start dates; however, a statistically significant immediate drop in level of MSSA clinical culture rates only occurred at the actual start date.

Thirty-one MSSA infections (per NHSN criteria) occurred during the pre-intervention period compared to 12 MSSA infections during the post-intervention period. Overall, incidence rate of MSSA infections was 1.07 per 1000 patient-days during the pre-intervention period compared to 0.55 per 1000 patient-days during the post intervention period (IRR = 0.51; 95% CI = 0.14, 1.82). Immediately following the intervention, the incidence rate of MSSA infections decreased in level by an estimated 73% (IRR=0.27; 95% CI=0.10, 0.79). There was not a sustained reduction in incidence rates of MSSA infections during the post-intervention period (IRR = 0.83 95% CI = 0.62, 1.12).

Because other infection control measures in the unit would have impacted MRSA and MSSA rates over time, MRSA clinical cultures and infections were assessed to help confirm an independent effect on the addition of screening and decolonizing MSSA carriers. There was not a difference in the rate of positive MRSA cultures and MRSA infections comparing the post- to pre- intervention periods. There were 8 positive MRSA clinical cultures during the pre-intervention period and 3 positive MRSA clinical cultures during the post-intervention period. Of the 11 NICU-attributable clinical cultures that grew MRSA during the study period only two (18.18%), one each in the pre- and post-intervention periods met the NHSN's definition for a specific HAI. The mean quarterly incidence rate of NICU-attributable MRSA positive clinical cultures was 0.27 and 0.16 per 1000 patient-days during the pre- and post-intervention periods, respectively (IRR = 0.60; 95% CI = 0.05, 7.77).

Of the first 85 neonates that had a surveillance or clinical culture grow MSSA post-intervention, 65 had an isolate available for mupirocin susceptibility testing. None of the 65 tested MSSA isolates were resistant to mupirocin. The median mupirocin MIC was 0.19 μg/mL.

Discussion

S. aureus remains a major cause of HAIs and late-onset sepsis in neonates. Active surveillance cultures and decolonization for MSSA successfully decreased S. aureus disease in our NICU. Quarterly incidence rates of MSSA-positive clinical cultures and MSSA infections decreased more than 50% immediately following implementation of this strategy. Quarterly incidence rates of MSSA-positive cultures continued to decrease by nearly 30% per quarter for the remainder of the study period. To our knowledge, this is the first study to evaluate the impact of active surveillance and targeted decolonization for decreasing MSSA burden in the NICU in a non-outbreak setting.

ASC coupled with decolonization has been shown to be an effective strategy in decreasing S. aureus transmission and infection.^12,20,25 Current guidelines suggest that decolonization may be considered in high risk neonates during an MRSA outbreak or to combat endemic MRSA when other strategies have failed.²⁶ Recommendations are less clear for MSSA and few data exist on the safety and efficacy of decolonization in this population.²⁷ Delaney et al found that after instituting a mupirocin prophylaxis regimen for 7 years, incidence rate of S. aureus (MSSA and MRSA) infections in their NICU decreased from 1.88 per 1000 patient-days to 0.33 per 1000 patient-days.²⁸ They found no mupirocin resistant S. aureus isolates.

While MRSA has been the target of most NICU S. aureus prevention and control programs, MSSA may cause comparable morbidity and mortality and is likely more prevalent in most centers. ^16–19 Ericson and colleagues recently reported that MSSA was responsible for 2.5 times more infections than MRSA.¹⁹ In Shane and colleagues' study of 8,444 VLBW neonates with S. aureus bacteremia or meningitis, MSSA was nearly thrice as prevalent as MRSA and both strains were associated with high mortality.³ In our study, MSSA accounted for greater than 90% of all S. aureus clinical isolates and a third of all MSSA infections were BSIs. A higher absolute burden of disease and mortality from MSSA compared to MRSA strains justifies refocusing prevention strategies to include MSSA in addition to MRSA.

Our findings are consistent with previous reports of a low prevalence of mupirocin resistance among S. aureus isolates from mupirocin treated neonates. Hitomi et al described the use of universal decolonization using mupirocin as a strategy for eradicating an outbreak of MRSA in their NICU.²⁹ As discussed above, Delaney and colleagues reported treating all neonates in their NICU with intranasal mupirocin for 7 years.²⁸ Mupirocin resistance was not observed in either study. During our recent study of mupirocin resistance among MRSA isolates from hospitalized neonates, we found a low prevalence (3 of 84 isolates, 3.6%) of low level mupirocin resistance and no isolates with high level resistance to mupirocin.²³ Although high level mupirocin resistance has been associated with treatment failure, the clinical significance of low level mupirocin resistance is unclear.³⁰ Similarly, in this study, we performed susceptibility testing on 65 available isolates from the first 85 neonates that had a culture grow MSSA post intervention and found no mupirocin-resistant MSSA isolates. Acquisition of mupirocin resistance is often a concern when considering a more aggressive decolonization strategy such as one that includes decolonizing MSSA-colonized neonates. However, while mupirocin resistance has been reported following widespread use in hospitalized adults^31–33, this has not been reported in neonates.

Data on the cost effectiveness of ASC and decolonization for prevention of S. aureus are limited. You et al, using decision analysis modeling, recently examined the potential clinical outcomes and cost of ASC for MRSA with and without decolonization in Hong Kong NICUs.³⁴ Even at very low levels of decolonization efficacy for prevention of MRSA infections in colonized neonates, decolonization, when combined with ASC was both cost-saving and effective in decreasing incidence of MRSA infections and MRSA-associated mortality. For programs already collecting surveillance cultures for MRSA colonization, there is little additional expense to identify MSSA colonized infants. Additional studies are needed to determine the cost-effectiveness of S. aureus surveillance and decolonization programs in the NICU.

Despite its benefits, decolonization may negatively impact the developing neonatal nasal microbiome.³⁵ Although chlorhexidine was well-tolerated in our population with no skin toxicity observed, the broad antimicrobial activity of mupirocin and chlorhexidine could predispose the neonate to colonization by more harmful pathogens. When considering decolonization as an infection control strategy in the NICU, therefore, the risks must be carefully weighed against benefits. Neonates should be closely monitored for acquisition of mupirocin resistance and replacement of the nasal flora by fungi and other pathogens. By targeting neonates for decolonization through ASC rather than universal treatment, the risks are minimized to those infants who may have the most benefit.

The primary outcome (positive clinical culture) has been used before as a surrogate outcome for MRSA infection and burden.³⁶ This outcome likely overestimates the burden of disease, but our conclusions are supported by the similar observed reduction in MSSA infections that met the NHSN's surveillance definition for a specific HAI. All quasi-experimental studies are at risk of influence by unobserved changes in practices over time. To account for this possibility, the burden of MRSA was evaluated during the study period and no change was observed reflecting the targeted nature of this intervention. The decrease in incidence rates of MSSA-positive clinical cultures and NHSN-defined MSSA infections may represent a return to baseline rates following an outbreak or a regression to the mean. However, during sensitivity analyses, while shifting the start of the intervention period backwards did not change the post-intervention slope significantly, the immediate drop in rates of MSSA was only statistically significant at the actual intervention start time. Compliance with the intervention was only 78%, but this is partially due to the fact that some neonates were discharged from the NICU before culture results were reported. Finally, although median birth weight is comparable between the pre-and post-intervention period, it is possible that there is a difference in the number of very low birth weight neonates in each period but we think this is unlikely.

Preterm and low birth weight neonates are particularly vulnerable to S. aureus infections. ASC and decolonization may be effective in decreasing the burden of S. aureus in NICUs and preventing infections and should not be limited to MRSA colonized neonates. Additional studies are needed to confirm the impact of decolonization on reducing MSSA infections among hospitalized neonates and to monitor for unanticipated consequences.

Figure 1.

Mean quarterly incidence of MSSA before and after implementation of an ASC and decolonization protocol. The dashed, horizontal and solid, horizontal lines represent the incidence rate of MSSA averaged over the pre- and post-intervention periods respectively and the dotted, vertical line (beginning of the 9^th study quarter) marks the start of the intervention.

Table2a.

Distribution of clinical culture categories

Specimen Category (n)	Period
	Pre*	Post**	Total
Abscess drainage	2	0	2
Blood	16	3	19
Body fluid, other	0	1	1
Broncho alveolar lavage	3	0	3
Catheter tip	1	0	1
Cerebrospinal fluid	1	0	1
Eye	7	3	10
Other	2	1	3
Peritoneal fluid	0	1	1
Sputum, non-cystic fibrosis	59	23	82
Urine	5	1	6
Wound	10	3	13
Total	106	36	142

^*Pre-intervention period (24 months)

^**Post-intervention period (18 months)

Table2b.

Distribution of infections

Category of Infection (n)	Period
	Pre*	Post**	Total
BSI	11	3	14
CNSI	1	0	1
CVSI-VASC	1	0	1
EENTI	2	1	3
IAB	0	1	1
LRTI	10	2	12
PNEU	1	2	3
SSTI	5	3	8
Total	31	12	43

^*Pre-intervention period (24 months)

^**Post-intervention period (18 months)