Disproportionality Analysis of Safety with Nafcillin and Oxacillin with the FDA Adverse Event Reporting System (FAERS)

Antistaphylococcal penicillins such as nafcillin and oxacillin are among the first choices of treatment for severe invasive methicillin-susceptible Staphylococcus aureus (MSSA) infections, although there has been limited safety evaluations between individual agents.

ABSTRACT

Antistaphylococcal penicillins such as nafcillin and oxacillin are among the first choices of treatment for severe invasive methicillin-susceptible Staphylococcus aureus (MSSA) infections, although there has been limited safety evaluations between individual agents. Using the FDA Adverse Event Reports System (FAERS), oxacillin was observed to have a lower proportion of reports of acute renal failure (reporting odds ratio [ROR], 5.3 [95% confidence interval {CI}, 3.1 to 9.3] versus 21.3 [95% CI, 15.8 to 28.6], respectively) and hypokalemia (ROR, 0.7 [95% CI, 0.1 to 4.8] versus 11.4 [95% CI, 7.1 to 18.3], respectively) than nafcillin.

INTRODUCTION

Antistaphylococcal penicillins (ASPs) such as nafcillin and oxacillin are among the drugs of choice for severe and invasive methicillin-susceptible Staphylococcus aureus (MSSA) infections (1–4). Alternative agents such as cefazolin have been associated with improved safety compared to ASPs (5–9). However, comparative safety data between individual ASPs are limited and have shown possible improved safety with oxacillin among adults (10). A 2016 retrospective cohort study examining comparative safety outcomes between nafcillin and oxacillin found that patients prematurely discontinued therapy at a rate of 18% versus 2% due to adverse effects, respectively. Additionally, oxacillin exhibited a favorable safety profile with regard to rates of hypokalemia and acute renal injury and demonstrated similar rates of hepatotoxicity to nafcillin (10). To further evaluate the safety of nafcillin and oxacillin, we aimed to determine the relative proportion of adverse events reporting for these agents among the FDA Adverse Event Reporting System (FAERS) database.

(This project was presented in part at IDWeek 2018, 3 to 7 October, San Francisco, CA [Abstract 1407] [11].)

Adverse effect reports were reviewed from the FAERS database between October 2003 and March 2018. The FAERS database is a voluntary reporting and publicly available FDA database of adverse effects (AEs) that may be reported by health care professionals, consumers, and drug manufacturers (12). Rates of adverse effects obtained from the database are relative to all other drugs. Signals obtained from the database are thought to represent true adverse effects. As a sensitivity analysis, a direct comparison of the proportion of reports for adverse effects was made between agents. Reports that included medications indicated as the primary or secondary suspected cause were included, while those listed as interacting or concomitant were excluded. Duplicate reports were removed by retaining only the most recent update to each individual case. AEs were searched using the Medical Dictionary of Regulatory Activity (MedDRA) terminology. MedDRA preferred terms were used along with lower-level terms and related FAERS database variations to improve case retrieval (see “Methods - Search Terms” within the supplemental material). Safety terms included AE related to acute renal failure (ARF), rash, hypokalemia, hepatotoxicity, and neutropenia. Relative signals for AE were evaluated through disproportionality analysis of safety events for nafcillin and oxacillin. Measures of association for safety evaluation included reporting odds ratio (ROR) and proportional reporting ratio (PRR) (13). All analyses were performed in SAS 9.4 (SAS Institute, Cary, NC).

We included 10,058,187 unique AE reports between October 2003 and March 2018, of which nafcillin and oxacillin were noted for 349 and 309 reports, respectively. Reports of ARF were substantially more common with nafcillin (PRR, 18.2 [95% confidence interval {CI}, 14.2 to 23.4]) than with oxacillin (PRR, 5.1 [95% CI, 3.0 to 8.8]). Reports of rash were similar between nafcillin and oxacillin (PRR, 2.7 [95% CI, 1.6 to 4.7] versus 3.1 [95%, CI, 1.8 to 5.3], respectively). Hypokalemia reports were considerably more common with nafcillin than oxacillin (PRR, 10.8 [95% CI, 6.9 to 17.0] versus 0.7 [95% CI, 0.1 to 4.8], respectively). The reports of hepatotoxicity were similar among nafcillin and oxacillin (PRR, 3.7 [95% CI, 2.4 to 5.8] versus 4.6 [95% CI, 3.0 to 7.1], respectively). Finally, reports of neutropenia were similar between nafcillin and oxacillin (PRR, 3.9 [95% CI, 1.5 to 10.2] versus 5.5 [95% CI, 2.3 to 13.0], respectively). Similar observations were seen with RORs (Table 1). A sensitivity analysis reflected similar results (Table S1).

TABLE 1.

Odds ratios of AEs with nafcillin and oxacillin

Results from the FAERS data indicate that oxacillin may have a better safety profile than nafcillin with regard to rates of ARF and hypokalemia. Reporting rates of rash, neutropenia, and hepatotoxicity were similar between the two agents. These results agree with the limited observational data available. Viehman et al. examined rates of AE and drug discontinuation in patients receiving either nafcillin or oxacillin for various MSSA infections in hospitalized patients. Nafcillin was associated with higher rates of kidney injury (18% versus 6%, respectively; P = 0.03) and hypokalemia (42% versus 17%, respectively; P < 0.0001) than was oxacillin. Additionally, rates of kidney injury were higher in the nafcillin group despite lower rates of concomitant nephrotoxic medications (88% versus 100%; P = 1.00). Rates of hepatotoxicity were found to be similar between the nafcillin and oxacillin groups, with increases in liver enzymes at 2% and 4% (P = 0.45), respectively (10). However, in contrast, Maraqa et al. found higher rates of hepatotoxicity with oxacillin compared to nafcillin (22% versus 0%, respectively; P < 0.001), higher rates of rash (31.7% versus 10.3%, respectively; P = 0.008) and similar rates of neutropenia (12.2% versus 17.2%, respectively; P = 0.49) for outpatient parenteral antimicrobial therapy (OPAT) in pediatric patients (14). These discordant results may reflect small sample sizes for adverse effect detection (n = 224 and n = 222 for Viehman and Maraqa, respectively) or different patient populations, and they may support the need for larger studies. A methodology using large, centralized data sources (e.g., Premier Healthcare or BD Insights Research Databases) may be better suited to meet the needs of detecting safety signals between these agents (15).

There is controversy regarding the use of cefazolin as a first-line agent for MSSA infections. Several recent studies have observed a favorable adverse effect profile and similar or improved efficacy with cefazolin compared to the ASPs, prompting consideration to regard it as first line (5–9, 16, 17). However, there is concern regarding the potential for cefazolin inactivation by beta-lactamases in the presence of high-inoculum infections, although its clinical impact is an ongoing debate (18–20). This scenario with nafcillin and oxacillin could be similar to that of the polymyxins described by Nation et al. in which two agents, previously considered interchangeable, were found to have clinically significant differences that subsequently created a preference (21). Therefore, the controversy of using cefazolin versus an ASP for MSSA infections may be lessened if oxacillin is found to have improved safety outcomes than previously thought. There are, however, data from subgroup analyses that have supported benefits in nephrotoxicity with cefazolin compared to oxacillin (6). Careful consideration of individual patient profiles and infections should be taken before choosing an agent.

There are some limitations in this study that are worth noting. First, due to the voluntary nature of FAERS reporting, the true incidence of AE is unknown. Biases inherent to the voluntary nature of reporting including overreporting, underreporting, and missing information are common to the database. Second, causal associations are not required before reporting AE to the FAERS database. Thus, confounding by other drugs or indications can occur. Third, laboratory data were not obtained to verify reports of AE. Definitions of AE are left to the reporting subject, and variations of definitions can occur. Lastly, few reports resulted in wide confidence intervals, and the results should subsequently be interpreted with caution. Despite these limitations, AE analysis using FAERS data can be used as hypothesis generating for future studies.

Our results support previous limited comparative safety observational data that oxacillin may be associated with overall improved safety compared to nafcillin, especially with regard to renal function and hypokalemia. With the similar costs and likely equal efficacy of these agents, clinicians may want to consider prescribing oxacillin over nafcillin if an ASP is indicated for an invasive MSSA infection (22, 23). However, in the context of the limitations of the FAERS database, causal associations cannot be made from these results, but rather support the need for further well-conducted comparative safety studies.