Retrospective Evaluation of Empiric Vancomycin Therapy for Infectious Workups in Relation to Methicillin-Resistant Staphylococcus aureus (MRSA) Risk Factors in the Neonatal Intensive Care Unit

Introduction

Sepsis is one of the most common complications associated with neonatal intensive care, especially in premature and very low birth weight infants.^1,2 In the neonatal period, sepsis can be categorized as early-onset sepsis or late-onset sepsis (LOS). Generally, LOS is defined as sepsis occurring after the first 72 hours of life.^3,4 With a global incidence of approximately 2800 per 100,000 live births, and a mortality rate of nearly 17.6%, sepsis is one of the leading causes of morbidity and mortality in hospitalized neonates.⁵

The causative organisms in LOS are often coagulase-negative staphylococci species (CoNS), as well as Staphylococcus aureus.^6,7 There is a high rate of resistance among CoNS to beta-lactam antibiotics, making vancomycin the agent of choice in suspected LOS.^8,9 However, the risks associated with vancomycin use, namely toxicity and rising rates of resistance mechanisms, demonstrate the need for antimicrobial stewardship programs in intensive care settings.¹⁰ A review of studies in antimicrobial stewardship found that these programs can reduce the use of broad-spectrum antibiotics, as well as the incidence of health care–associated infections.¹¹ Several neonatal intensive care units (NICUs) have successfully reduced the use of vancomycin in the setting of neonatal sepsis with no significant differences in morbidity and mortality.^12–14

The American Academy of Pediatrics (AAP) has published guidance on how to approach antibiotic treatment of LOS; administration of antibiotics should occur as quickly as possible once concern for sepsis is established, but there are several accepted approaches in choosing an empiric agent.¹⁵ The 2024–2027 Report of the Committee on Infectious Diseases, otherwise known as the Red Book from AAP, currently recommends vancomycin as an empiric drug of choice for S aureus infections with unknown susceptibility.¹⁶ Many infants receive vancomycin as empiric therapy owing to a history of colonization and/or the presence of risk factors for methicillin-resistant Staphylococcus aureus (MRSA) as reported in the literature. These risk factors include, but are not limited to, indwelling foreign bodies such as central venous catheters or endotracheal tubes, parenteral nutrition, history of colonization, gestational age, birth weight, and place of birth (inborn vs outborn).^1,2,17–21 The purpose of this retrospective chart review is to compare the use of vancomycin with alternative Gram-positive agents in patients admitted to the NICU who are ≥3 days of age, based on the number of MRSA risk factors present. Secondarily, this study aims to compare rates of escalation and de-escalation in patients who receive at least 5 days or more of antimicrobial therapy. Finally, we present an algorithm based on the data in this study to aid clinicians in choosing empiric antimicrobial therapy for suspected LOS as well as de-escalating at an appropriate time to reduce exposure to vancomycin and antibiotics in general.

Materials and Methods

This was a retrospective study conducted at a 92-bed, Level IV NICU in Cincinnati, OH. Blood cultures obtained from patients admitted to the NICU between December 1, 2022, and May 31, 2023, were screened for inclusion. Patients were included if they were at least 3 days of age at time of culture and received vancomycin or an alternative Gram-positive–covering agent during the study period. Patients who received antibiotics preoperatively or for a reason other than a sepsis workup were excluded. Repeated cultures from the same patient were required to be at least 2 weeks apart—indicating a new infectious workup—to be included. Patients were analyzed in 2 groups: group 1 received vancomycin empirically and group 2 received an alternative Gram-positive agent (e.g., ampicillin, nafcillin, cefepime) empirically.

As shown in Figure 1, all patients in this institution’s NICU undergo weekly screening for Staphylococcus colonization, using nasal swabs. According to this institution’s most recent data, the rate of methicillin resistance in S aureus cultures was approximately 25%. Patients who test positive for methicillin-sensitive Staphylococcus aureus (MSSA) or MRSA receive 5 days of topical nasal mupirocin plus monthly treatment if they have a central line, external ventricular drain, or open wound. These patients also receive nasal mupirocin for 3 days prior to any surgery. In this unit, nafcillin is typically recommended over vancomycin for empiric Gram-positive coverage for LOS in infants without a history of MRSA colonization or MRSA infection.

View Figure

• Download as Powerpoint
• Download Figure

Results

Baseline Characteristics and Primary Outcome.

Table 1 compares the baseline characteristics across the 2 groups. Patients in group 1 had a lower gestational age (29.58 vs 33.17 weeks; p < 0.001) and birth weight (1.32 vs 2.13 kg; p < 0.001) than those in group 2 at baseline. Group 1 patients also had a higher postnatal age than group 2 (58 vs 23 days; p < 0.001). Dosing weight was not significantly different between the 2 groups. A total of 143 blood cultures from 95 patients were included in the final analysis; 30 patients had at least 2 cultures included, and 11 patients had 3 or more cultures included.

Table 1.Baseline Characteristics

View Fullscreen

Table 2 details the rates of each MRSA risk factor between the groups. Patients who started on vancomycin had higher rates of every MRSA risk factor included in this study, except outborn status. The most common risk factors in the vancomycin group were low birth weight (80.4% vs 58.7%; p = 0.009), prematurity (84.3% vs 63.0%; p = 0.008), and length of stay over 10 days (80.4% vs 41.3%; p < 0.001). Overall, the total number of MRSA risk factors per patient was higher in group 1 than group 2 (4.88 vs 2.53; p < 0.001).

Table 2.MRSA Risk Factors

View Fullscreen

Secondary Outcome.

A total of 19 patients in group 1 and 31 patients in group 2 continued antibiotic therapy for at least 5 days. Of these, 8 patients in each group either escalated or de-escalated from their empiric agent of choice. Figure 2 describes the distribution of MRSA risk factors present in these patients. Patients who de-escalated from vancomycin had an average of 5.7 (SD, ±1.7) MRSA risk factors, and patients who escalated to vancomycin had an average of 2.9 (SD, ±1.9) risk factors. Additionally, only 9 of the 143 cultures (6%) grew organisms; most were CoNS, with 2 cultures positive for MRSA.

View Figure

• Download as Powerpoint
• Download Figure

For patients in group 1 who received at least 5 days of therapy, 11 patients did not de-escalate from vancomycin. Only 4 patients had cultures requiring vancomycin; this included 2 resistant Staphylococcus epidermidis bacteremias, 1 MRSA respiratory culture, and 1 MRSA wound culture. The latter 2 patients had blood cultures during their respective workups, which qualified them for inclusion, but ended up being negative. Therefore, 7 patients had the opportunity to de-escalate to a narrower agent. For patients in group 2 who received at least 5 days of therapy, only 3 of the 8 patients who escalated to vancomycin had cultures requiring escalation. All of the positive cultures in group 2 were blood cultures that grew CoNS, and the blood cultures requiring escalation were all resistant S epidermidis species.

Discussion

The results of this study helped to identify patterns of vancomycin prescribing for the indication of LOS in our NICU. Patients who were started on vancomycin were appropriately found to have more MRSA risk factors than those who were started on alternative Gram-positive–covering antibiotics. Empiric vancomycin was prescribed more in patients who a had low birth weight, prematurity, a central line, endotracheal tube, other indwelling lines, longer durations of hospitalization, parenteral nutrition, history of positive MRSA blood culture, and history of positive MRSA nares colonization. Outborn status was not found to be a statistically significant risk factor, likely owing to the few number of patients who were born at this institution during our study period. A meta-analysis on the incidence of MRSA colonization found that colonization was more prevalent in outborn than inborn patients.¹ Outborn status, specifically in relation to MRSA risk factors, has not been studied in patients with LOS, as more data exist in patients who develop sepsis within 3 days of birth.^1,27 This study also found a relatively low rate of de-escalation from vancomycin in patients who had negative cultures; this may be because this study was not able to capture the full clinical picture at the time of blood culture, which may have had more influence on the selection of antimicrobial agent and duration than patient-specific risk factors.

This study evaluated multiple risk factors for MRSA, based on previous findings in the literature. Our results are generally consistent with the findings of other studies that evaluated MRSA risk factors in the neonatal population. Balamohan and colleagues²¹ found MRSA colonization to be a risk factor for infection when adjusted for length of stay and gestational age, and a case-control study by El Manouni El Hassani and colleagues²³ found that parenteral nutrition for more than 10 days was associated with higher odds of developing LOS irrespective of an identified pathogen. Additionally, similar to this study’s findings, Stoll and colleagues²⁵ found that rates of LOS were inversely related to birth weight and gestational age and positively correlated with length of hospital stay and duration of intravascular access. Though our study did not evaluate duration of intravascular access, significantly more patients with central lines were initiated on vancomycin empirically. Finally, prematurity was the most prevalent risk factor in patients who were started on vancomycin. Many studies have found a strong correlation between lower gestational age at birth and risk of infection, while others have found links between lower gestational age and the development of comorbidities such as bronchopulmonary dysplasia, cognitive deficits, and necrotizing enterocolitis.^7,24,28,29

Antimicrobial stewardship is an important tool to use in the NICU setting, and several practices have been established to help promote stewardship principles. For example, nares swabs can be helpful in determining colonization status in suspected and confirmed MRSA infections. One hospital found that educating providers on the use of nasal screening to guide vancomycin de-escalation resulted in both an increase in the use of nasal swabs and a decrease in vancomycin use per 1000 patient-days.³⁰ In a pharmacist-led nasal MRSA screening for adult intensive care patients, Diep and colleagues³¹ saw a decrease in the average duration of vancomycin and number of vancomycin concentrations drawn. Other data have shown that nasal S aureus surveillance can help detect and prevent infection in high-risk neonates.^32,33

Decolonization using topical agents such as mupirocin may also prove to be an important tactic in reducing exposure to MRSA and developing infection. Kotloff and colleagues³⁴ conducted a study on the safety and efficacy of topical mupirocin at multiple body sites (including intranasally) in eligible NICU patients, and found that most patients achieved decolonization within 3 days of completion of therapy. Another retrospective cohort study completed by Pierce and colleagues³⁵ found that among MRSA-colonized neonates, the hazard ratio for developing a Gram-positive infection was 64% lower in patients who received topical mupirocin than in those who did not. This institution’s Staphylococcus screening policy, as illustrated in Figure 1, has helped to identify patients who are at higher risk of developing a MRSA infection. This study demonstrated that history of positive MRSA blood culture, as well as colonization, is used at our institution to determine empiric therapy.

The risk of nephrotoxicity related to vancomycin use is also an important factor to consider when choosing empiric therapy. The AWAKEN study by Jetton and colleagues,³⁶ which included patients at this institution, found an overall acute kidney injury (AKI) prevalence of about 29.9% in NICU patients, which increased to 47.9% in patients between 22 and 29 weeks’ gestational age. This study also found that patients with AKI had a mortality odds ratio of 4.6 and an increase in the average length of stay by 8.8 days, compared with patients who did not have AKI. Several other studies have found that acute kidney injury in neonates is associated with higher rates of morbidity and mortality, and the risk is greater in patients with sepsis, prematurity, and very low birth weight.^37,38 Because nephrotoxic drug exposure may be considered a modifiable risk factor for AKI, institutions who use programs such as Baby NINJA may be more inclined to choose agents based on risk of nephrotoxicity.^39,40 Indicators of neonatal renal function in this study were not evaluated at time of review, which may have affected providers’ decision-making. However, our study emphasizes the need for using stewardship strategies in the management of antimicrobial therapy in LOS. Considering risk factors such as gestational age and birth weight can potentially help mitigate risk of acute kidney injury in these patients. Additionally, increasing monitoring of renal function and vancomycin levels may be sufficient when considering initiation of vancomycin where there is a clear benefit.

Determining the need for broad-spectrum antibiotics can vary across institutions, based on policies and local antibiograms, but there is strong evidence to support the use of protocols to promote stewardship. At one hospital, Chiu and colleagues⁴¹ developed a NICU vancomycin use guideline for their institution that does not allow vancomycin initiation until the condition of the infant receiving an alternative agent, such as nafcillin or oxacillin, clinically deteriorates after 48 hours of sterile cultures, or the cultures grow MRSA or CoNS. This guideline reduced both the number of infants exposed to vancomycin as well as the use of vancomycin per 1000 patient-days.^41,42 At the time of data collection, this institution’s NICU did not have a guideline for the use of vancomycin. Therefore, the data in this study have been used to create a NICU vancomycin algorithm that may be helpful to institutions that do not currently have the means to guide antibiotic choices through established policies or protocols.

As shown in Figure 3, the proposed algorithm is based on our finding that patients who were started on vancomycin had 4.88 risk factors for MRSA on average. MRSA colonization and culture history are well described in the literature as being highly correlated to MRSA infection, and patients with either of these risk factors should receive vancomycin empirically. According to the proposed algorithm, vancomycin can be considered in patients with 5 or more of the risk factors listed in our study. Although hemodynamic instability was not recorded in our patients, it is included in the algorithm to account for patients who may not have enough risk factors but have a higher acuity. This addition is also supported by the AAP Red Book guidance on invasive staphylococcal infections, which recommends vancomycin empirically for life-threatening S aureus infections with unknown susceptibility.¹⁶ Finally, our algorithm recommends narrowing based on cultures and susceptibilities, as well as discontinuing antibiotics after 36 hours of negative culture findings. While we did not capture time to discontinuation, other studies have used between 24 and 48 hours to rule out the need for antibiotics.^43–46

View Figure

• Download as Powerpoint
• Download Figure

When applied to this study’s patients, the algorithm highlights the need for reduction in both the initial exposure and duration of therapy for vancomycin, especially considering the low incidence of MRSA (1%) found during the study timeframe. In group 1 patients, 33 of 51 workups (65%) would have resulted in initiation of vancomycin when using the algorithm without accounting for hemodynamic instability. Of the patients in group 1 who received vancomycin for at least 5 days, 6 of the 11 patients (55%) would have been initiated on vancomycin according to our algorithm, and only the 4 patients with positive culture findings would have received vancomycin for the full treatment course.

Our study was limited by being conducted at a single center, which may influence factors such as local resistance patterns or provider prescribing preferences. Owing to the retrospective nature of this study, it was difficult to capture patient acuity at the time of infectious workup, which would have been valuable information when determining appropriateness of antibiotic choice. Infectious markers such as white blood cell count, C-reactive protein, and procalcitonin were also not evaluated in this study, which may be useful in determining empiric antibiotic choice and length of therapy according to previous findings.^26,39,47

Limitations also exist for the proposed vancomycin use algorithm. While we recommend that all institutions use an algorithm or protocol to conserve the use of vancomycin for patients with LOS, local antibiograms should be used to evaluate rates of MSSA/MRSA incidence and guide therapy. This algorithm should be used only as a guideline, as medical treatment of LOS should occur through shared decision-making with the care team and the patient’s family.

Conclusions

Many risk factors for MRSA infection are present in infants started on vancomycin empirically at our institution, and these risk factors are used to determine empiric antibiotic regimens in LOS at this institution. Using a history of MRSA infection or colonization, as well as MRSA risk factors, is important in choosing an empiric antimicrobial agent. This institution had opportunities to avoid vancomycin as well as reduce the length of vancomycin exposure according to the proposed algorithm, largely owing to the lack of patients with positive culture results and incidence of MRSA, and future efforts will be focused on narrowing therapy 36 hours after cultures are obtained. Further studies are needed to determine the true nature of the relationship between the studied risk factors and the incidence of MRSA infection to optimize empiric antibiotic choices.