Evaluation of the Safety of Cefazolin Monotherapy in Outpatient Parenteral Antimicrobial Therapy at a Large Academic Medical Center

Abstract

In outpatient parenteral antimicrobial therapy, weekly lab monitoring is often standard practice, even for antimicrobials with low complication rates. We reviewed 420 patients treated with cefazolin: lab abnormalities associated with readmission or treatment changes occurred in 2% of patients. Reduced-frequency monitoring may be reasonable for patients treated with cefazolin outpatient.

Outpatient parenteral antimicrobial therapy (OPAT) is a multidisciplinary program designed to monitor patients on IV antimicrobials outside the hospital. OPAT programs are associated with reductions in hospitalizations [1, 2] and healthcare costs [3] while maintaining clinical outcomes [4]. Given their outpatient nature, and the potential adverse effects of antimicrobials, care coordination, and safety monitoring comprise a key component of OPAT programs.

Current guidelines from the Infectious Diseases Society of America (IDSA) recommend routine safety labs for all patients enrolled in OPAT [5]. These guidelines are based on known antimicrobial toxicities [6, 7], and findings that monitoring programs reduce hospital readmission rates [8] and errors in therapy administration [9]. However, despite improving safety, the uniform application of these monitoring recommendations may not sufficiently account for the varying risks of adverse events among individual antimicrobials. The 2018 IDSA OPAT guidelines recommend that safety labs be drawn on a weekly basis for all patients regardless of antimicrobial choice, but acknowledge that “data are insufficient to make evidence-based recommendations about specific tests and specific frequencies of monitoring for individual antimicrobials” [5].

This high frequency of monitoring is potentially disproportionate for patients receiving low-risk therapies. Effective monitoring must balance the need for the early detection of treatment complications against minimizing patient burdens and avoiding unnecessary healthcare utilization.

Cephalosporins are widely used in OPAT and carry a lower risk of adverse effects compared to other classes of antibiotics [6]. Cefazolin, in particular, is commonly prescribed for the treatment of soft tissue, bloodstream, and bone infections, including staphylococcal or streptococcal infections, and is thought to carry few complication risks. The most common adverse events are eosinophilia, leukopenia, and drug hypersensitivity reactions (HSRs) [10]. To that end, OPAT safety lab monitoring for patients treated with cefazolin serves primarily for the early detection of eosinophilia and other findings that may distinguish asymptomatic eosinophilia from HSRs. Previous studies in patients treated with cefazolin monotherapy in OPAT programs have found adverse event rates between 1% and 5%, though many of these studies are limited by small cefazolin cohorts [6, 11, 12]. Moreover, some complications, such as rash, can be clinically evident without lab monitoring.

To better describe the frequency of lab abnormalities detected by routine monitoring and the role they serve in informing treatment decisions, we performed a single-center retrospective review of patients enrolled in OPAT treated with cefazolin monotherapy at a large academic medical center.

METHODS

Study Design

A retrospective review was performed to identify patients enrolled in the OPAT program and prescribed cefazolin monotherapy at Brigham and Women's Hospital (BWH), Boston, Massachusetts, USA, between February 2018 and February 2023. Patients discharged on multiple antimicrobial agents were excluded. This study was approved by the Mass General Brigham Institutional Review Board.

Patient Consent Statement

This study involved retrospective analysis of deidentified data and did not necessitate individual patient consent per the Mass General Brigham Institutional Review Board.

OPAT Workflow

The OPAT program at BWH is comprised of an infectious diseases (ID) clinician, ID pharmacist, nurse practitioner, 3 registered nurses, and an administrative assistant. Initial evaluation and referral to the OPAT program was performed by the BWH ID consult team. Laboratory collections were performed by a visiting nurse association or at a regional clinical laboratory. Results were faxed to the OPAT team or transmitted electronically to the patient's electronic medical record. Laboratory monitoring and management of intervisit issues, such as symptomatic changes, were performed by the OPAT team, and any abnormal results, clinical issues, or proposed changes were communicated to the responsible ID clinicians for approval.

Outcomes and Data Collection

Deidentified patient data were retrospectively obtained from the BWH OPAT database maintained in the Research Electronic Data Capture (REDCap) platform. Patients are manually entered into the REDCap database at the time of OPAT referral.

The primary outcome was the incidence of OPAT-related adverse events, defined as lab abnormalities detected after discharge and while on therapy, or Clostridioides difficile infection. Weekly safety labs were examined over the course of OPAT and included: change in creatinine (increase of 0.5 mg/dL or a ≥50% relative increase from the lowest value recorded during admission), increased alanine transaminase (>100 units/L), neutropenia (<1000 cells/mm³), eosinophilia (>500 cells/mm³), and thrombocytopenia (<100 K/μL or a ≥ 50% decrease relative to the higher of the last platelet count during admission or the first outpatient platelet count). Additional treatment complications adjudicated from chart review included change in antimicrobials, unplanned hospital readmission, or death during OPAT. Secondary outcomes included Clostridioides difficile infection or a new rash while on OPAT. Time to detection of lab abnormalities was calculated by measuring the number of days since hospital discharge to the date of the first lab draw abnormality. Lab results originating from rehospitalizations were excluded from analysis. Descriptive statistics were performed using R 4.5.

RESULTS

Our study included 420 patients enrolled in OPAT and treated with cefazolin monotherapy (2 g of IV cefazolin every 8 hours with adjustment for renal function). The median treatment duration was 5.6 weeks. The median age was 61.7 years and the most common indications for treatment were bacteremia and osteomyelitis (Table 1).

Table 1.

Patient Characteristics and Adverse Events After Discharge

Variable	No. (%)
Patient Information	420 (100%)
Sex
Male	252 (60.0%)
Female	168 (40.0%)
Age (years)
<50	111 (26.4%)
50–59	81 (19.3%)
60–69	109 (30.0%)
70–79	84 (20.0%)
≥80	35 (19.3%)
Referring service
Oncology/Transplant	62 (14.8%)
Other	358 (85.2%)
Source of Infectiona
Bacteremia	208 (49.5%)
Osteomyelitis	70 (16.7%)
Prosthetic joint infection	48 (11.4%)
Endocarditis	45 (10.7%)
Skin/soft tissue infection	33 (7.9%)
Septic arthritis	27 (6.4%)
Postoperative spine infection	22 (5.2%)
Otherb	165 (39.3%)
Discharge Location	…
Home	305 (72.6%)
Facility	115 (27.4%)
Adverse Events After Discharge
Lab abnormalityc	18 (4.3%)
Eosinophilia	13 (3.1%)
Creatinine elevation	4 (1.0%)
Transaminase elevation	2 (0.5%)
Neutropenia	1 (0.2%)
Thrombocytopenia	0 (0%)
Other adverse eventsb	13 (3.1%)
Rash	12 (2.9%)
Clostridioides difficile infection	1 (0.2%)
Treatment complicationb	55 (13.1%)
Antimicrobial agent change	25 (6.0%)
Readmission	39 (9.3%)
Death	2 (0.5%)
Cause of death	2 (0.5%)
Metastatic cancer	1 (0.2%)
Pulmonary hypertension	1 (0.2%)
Treatment Complications in Patients with Lab Abnormalities
Treatment complication	9 (2.1%)
Antimicrobial agent change	5 (1.2%)
Readmission	3 (0.7%)
Antimicrobial agent change and Readmission	1 (0.2%)
Death	0 (0%)

^aPercentages do not add to 100% due to possibility of multiple diagnoses.

^bAdditional infectious sources with <5% prevalence included epidural abscess, pneumonia/empyema, intracranial infection, vascular graft infection, fracture fixation, intra-abdominal infection, urinary tract infection, diabetic foot infection, and meningitis.

^cRates of individual events do not sum to rates of total events due to some patients having multiple adverse events.

Weekly safety labs detected abnormalities for 18 patients (4.3%) in our cohort, including 13 cases of eosinophilia (3.1%) and 4 cases of serum creatinine elevation (1%). Other adverse events detected included 12 patients with a rash and one case of Clostridioides difficile infection. Interestingly, rash cooccurred with eosinophilia in only 1 patient in our cohort.

To detect cases where safety labs might have contributed to a change in clinical decisions, we recorded cases where patients were switched from cefazolin to an alternative antimicrobial agent or readmitted to the hospital during the OPAT period. There were 39 readmissions (9.3%), and 25 antimicrobial changes (6%). However, only 9 patients who experienced readmission or medication change had a preceding lab abnormality (2.1%). There were 2 deaths in our cohort over the OPAT period. Chart review revealed that neither was related to complications from OPAT or associated with a detected safety lab abnormality.

Finally, we sought to understand how soon after discharge lab abnormalities were detected through weekly monitoring. We found that two-thirds of the 18 lab abnormalities in our cohort were detected within the first 2 weeks after discharge, with 4 in the first week (22.2%) and 8 in the second week (44.4%). Of the remaining 6 lab abnormalities, 3 (16.7%) were detected in the last week of treatment.

DISCUSSION

In this single-center retrospective study evaluating the utility of OPAT safety lab monitoring for patients treated with cefazolin monotherapy, we found that few patients experienced a lab abnormality detected by safety labs (4.3%). Additionally, safety labs coincided with changes in treatment plans in <2% of patients and readmission in <1%, indicating that safety labs contributed to changes in clinical decision-making for few patients. These findings coincide with previous studies looking at the safety profile of cefazolin that found adverse event rates between 1% and 5%, primarily consisting of eosinophilia, nephrotoxicity, and HSRs [6, 11–13]. Further, as our data show, decisions to alter antimicrobial treatment often result from clinical changes, such as the development of a rash, that are evident with or without concomitant lab abnormalities. The low rate of lab abnormalities and the fact that abnormalities typically develop early in the treatment course presents the possibility that weekly monitoring may be unnecessary, especially in the later weeks of therapy.

Our study has several strengths. First, our study population was larger than most comparable studies investigating the effects of cefazolin monotherapy [11, 12]. Second, our study focused on adverse events that occurred during OPAT, excluding confounding events that occurred during acute care. Third, our study included rates of readmission and antimicrobial agent change, allowing us to determine a ceiling on the maximum number of patients for whom safety lab monitoring potentially altered treatment course. Fourth, our study measured the time to detection of lab abnormalities, which can inform the optimal frequency of laboratory monitoring.

Our study also has several limitations. The study was performed at a single center, which poses limitations on the ability to extrapolate to other centers. However, as our center is a quaternary care facility treating a wide variety of patients including immunocompromised patients, our finding of a low prevalence of treatment complications should be reassuring for centers with a similar or less complex patient population. We were also limited to the clinical data incorporated in the OPAT database. This prevented us from performing a risk stratification of patients to detect features that may be associated with higher rates of adverse events.

In conclusion, cefazolin in OPAT is well-tolerated and is associated with few adverse events. We observed that routine lab monitoring detected lab abnormalities in less than 5% of patients; less than half of these experienced a change in antimicrobial agent, indicating that lab monitoring rarely altered clinical course. We believe these findings support the idea that less frequent (ie, every other week) safety labs may be reasonable in patients treated with cefazolin monotherapy. We encourage continued investigation into more personalized approaches to safety monitoring in OPAT programs.