Comparison of Adverse Drug Reactions Between Patients Treated With Ceftaroline or Ceftriaxone: A Single-Center, Matched Cohort Study

Abstract

Background

Little information is available on the relative tolerability of ceftaroline versus other cephalosporins in clinical practice. We sought to compare adverse drug reactions (ADRs) associated with ceftaroline with those associated with ceftriaxone in hospitalized patients.

Materials and Methods

This was a retrospective, single-center matched cohort (according to age, indication, and duration) study of patients treated with ceftaroline or ceftriaxone at the VA St Louis Health Care System between 29 October 2010 and 28 March 2017, to compare rates of ADRs between the agents. Patients included received ≥2 doses of either medication to treat osteomyelitis, acute bacterial skin and skin structure infection, blood stream infection, pneumonia, infective endocarditis, septic arthritis, prosthetic joint infection, or empyema. The primary and secondary outcomes were the composite of any ADR during therapy and any ADR leading to premature discontinuation of therapy. The ADRs evaluated included rash, neutropenia, acute kidney injury, eosinophilia, thrombocytopenia, transaminitis, and hyperbilirubinemia.

Results

After matching, 50 patients per group were included and analyzed. An ADR occurred in 20% (10 of 50) of patients treated with ceftriaxone and 16% (8 of 50) of those treated with ceftaroline (P = .60). Two percent (1 of 50) of those treated with ceftriaxone and 16% (8 of 50) treated with ceftaroline had therapy discontinued owing to an ADR (P = .03). The most common ADR was eosinophilia (3 of 50) in the ceftriaxone group and rash (5 of 50) in the ceftaroline group. Ceftaroline therapy was identified as an independent risk factor for an ADR requiring premature discontinuation (odds ratio, 10.2; 95% confidence interval, 1.19–87.8; P = .03).

Conclusions

Although there was no difference in the rates of ADRs between patients in the ceftriaxone and ceftaroline groups, significantly more ceftaroline-treated patients required premature discontinuation.

Ceftaroline is a broad-spectrum, expanded-generation cephalosporin with activity against methicillin-resistant Staphylococcus aureus (MRSA) that was approved by the US Food and Drug Administration in 2010 for the treatment of community-acquired bacterial pneumonia and acute bacterial skin and skin structure infections (ABSSSI) [1]. Initially ceftaroline appeared generally well tolerated, with few agent-specific adverse drug reactions (ADRs) reported, but postmarketing data suggested that neutropenia may be more common with prolonged exposure to ceftaroline [2–10].

This postmarketing safety data, along with anecdotal experience at the VA St Louis Health Care System, prompted a local facility review of all patients treated with ceftaroline to determine the frequency and type of ADRs. Of 75 patients who received ceftaroline an ADR developed in 13 (17.3%), requiring discontinuation of the drug in 10. The most common ADR was rash, occurring in 9.3% (7 of 75), with neutropenia occurring in 4% (3 of 75) [11].

This 17.3% overall ADR rate with ceftaroline was higher than other published reports, and we sought to compare this cohort with patients during the same time period treated with ceftriaxone, the most commonly used cephalosporin at the VA St Louis Health Care System [11]. The hypothesis was that ADRs would develop in a significantly higher proportion of patients treated with ceftaroline, compared with those treated for similar indications and durations with ceftriaxone.

MATERIALS AND METHODS

This was a retrospective, single-center cohort study conducted at the VA St Louis Health Care System, including all patients who received ≥2 doses of either ceftaroline or ceftriaxone between 29 October 2010 and 28 March 2017 for the treatment of osteomyelitis, ABSSSI, blood stream infection, pneumonia, infective endocarditis, septic arthritis, prosthetic joint infection, or empyema (each diagnosis had to be explicitly described in a progress note written by an infectious diseases physician). High-dose therapy was identified for each group and was defined as treatment with ceftriaxone at ≥2 g/d of or ceftaroline at 600 mg every 8 hours (or the equivalent based on renal function). Patients could be included only once in a single arm of the cohort, determined by which medication was received first during the study period and met inclusion criteria. Patients could not be receiving concomitant antibiotic therapy.

The primary outcome evaluated was any documented ADR experienced during therapy. The Naranjo algorithm was used to assess all potential ADRs, and only events scoring ≥4 were included in the analysis [12]. This algorithm defines a medication’s association with an adverse event, based on a score ranging from 0 to 13. A score <2 indicates that an ADR is doubtful; scores of 2–4, 5–8, and >9 indicate possible, probable, and definite ADRs, respectively [12]. A score of 4 was included only if the answer to question 5 on the Naranjo algorithm (“Are there alternative causes that could have caused the reaction?”) was “yes.”

The adverse reactions evaluated were rash (determined by a review of each patient’s electronic medical record and searching progress notes during treatment with the medication in question for documentation of a rash), development of neutropenia (defined as any absolute neutrophil count after doses of <1500/μL for either medication), development of acute kidney injury (defined as a 2-fold increase in serum creatinine or a decrease in creatinine clearance [calculated using the Cockcroft-Gault method] of 50% from the baseline values after doses of either medication), eosinophilia (defined as any absolute eosinophil count >500/μL after doses of either medication ), thrombocytopenia (defined as any platelet count <150 000/μL after doses of either medication), anemia (defined as a decrease in baseline hemoglobin value ≥2 g/dL), transaminitis (defined as aspartate aminotransferase or alanine aminotransferase levels ≥3 times the upper limit of normal), or hyperbilirubinemia (defined as any total bilirubin value ≥3.0 mg/dL after doses of either medication).

The secondary outcome of the study was the rate of any of the above ADRs that lead to antibiotic discontinuation. To deem that premature discontinuation was due to an ADR, there must have been a progress note stating that the suspect antibiotic was likely the cause, and the ADR must have abated with discontinuation of the antibiotic.

All potential patients were identified through a review of pharmacy records. Once the ceftaroline cohort was identified, a ceftriaxone-treated patient was identified for 1:1 matching, using the nearest-neighbor method and accounting for age, treatment indication, and duration of therapy to better balance the arms of the cohort. To best match the 75 unique patients treated with ceftaroline outlined in the previous evaluation [11], it was determined that ≥312 ceftriaxone-treated patients should be screened. Once 312 ceftriaxone-treated patients were identified, the propensity score matching model, restricted by a caliper of 0.5 of the standard deviation, would be completed. If the matching model retained significance, no additional patient data would be collected.

Categorical variables were tested using χ ² or Fisher exact tests, and continuous variables were tested using independent t or Wilcoxon rank sum tests, as appropriate. Because of the small sample size, and overall low expected rate of ADRs, the factor of ceftaroline treatment, high-dose therapy (for ceftaroline and ceftriaxone), and number of reported allergies at therapy initiation were directly included in a multivariate logistic regression to determine whether any of these factors were independently associated with premature discontinuation. Significance was defined as a 2-sided α value <.05. All statistical testing was performed using IBM-SPSS software, version 22.0 (IBM).

RESULTS

Three hundred twelve patients treated with ceftriaxone and 75 treated with ceftaroline between 29 October 2010 and 28 March 2017 met inclusion criteria. The a priori parameter of significance of the matched cohort was met, so no additional ceftriaxone-treated patients were screened for inclusion. With use of the nearest-neighbor method and with ceftriaxone- and ceftaroline-treated patients matched based on age, indication for therapy, and duration of therapy, 50 patients remained in each arm of the cohort.

Baseline characteristics of the matched cohort are shown in Table 1; there were no significant differences between groups for any characteristics evaluated. Most patients were being treated for osteomyelitis, an ABSSSI, or pneumonia, and the mean duration of therapy was 17 days in the ceftaroline arm and 14.5 days in the ceftriaxone arm. There were no significant differences in the number of reported allergies between groups. Nine patients treated with ceftaroline and 14 treated with ceftriaxone had β-lactam allergies listed in their medical records at the time of therapy initiation (P = .23), 7 penicillin-class and 2 cephalosporin allergies in the ceftaroline and 12 penicillin-class and 2 cephalosporin allergies in the ceftriaxone group.

Table 1.

Baseline Patient Characteristics by Treatment Group

Characteristic	Ceftaroline (n = 50)	Ceftriaxone (n = 50)	P Value
Age, mean (SD), y	63.4 (11.3)	65.4 (9.5)	.35
Male sex, no. (%)	50 (100)	48 (96)	.50
White race, no. (%)	41 (82)	37 (74)	.33
Weight, median (IQR), kg	98.6 (30.4)	93.7 (52.7)	.78
Platelet count, mean (SD), cells/μL	275.6 (117.4)	262.5 (131.5)	.61
Leukocyte count, mean (SD), cells/μL	10.4 (4.7)	10.9 (5.3)	.64
Duration of therapy, mean (SD), d	17.0 (15.8)	14.5 (16.9)	.44
Serum creatinine level at initiation, mean (SD), mg/dL	1.5 (1.2)	1.5 (1.3)	.91
No. of reported allergies at therapy initiation, no. (%)
0	19 (38)	12 (24)	.17
1	11 (22)	10 (20)
2	8 (16)	15 (30)
3	4 (8)	5 (10)
≥4	8 (16)	8 (16)
High-dose therapy,a no. (%)	1 (2)	8 (16)	.03
Treatment indication, no. (%)
ABSSSI	11 (22)	12 (24)
Bloodstream infection	5 (10)	2 (4)
Empyema	1 (2)	1 (2)
Infective endocarditis	1 (2)	2 (4)
Meningitis	0 (0)	1 (2)
Prosthetic joint infection	2 (4)	0 (0)
Pneumonia	6 (12)	17 (34)
Osteomyelitis	17 (34)	12 (24)
Septic arthritis	2 (4)	1 (2)
Combination of the above	5 (10)	2 (4)

Abbreviations: ABSSSI, acute bacterial skin and skin structure infection; IQR, interquartile range; SD, standard deviation.

^aHigh-dose therapy was defined as 600 mg every 8 hours (or equivalent for renal function) for ceftaroline and ≥2 g/d for ceftriaxone

Overall, 18 patients of the 100 in the cohort experienced an ADR, 9 of which necessitated premature discontinuation of the antibiotic. Eight patients (16%) treated with ceftaroline experienced an ADR, compared with 10 (20%) treated with ceftriaxone (P = .60). The most commonly documented ADR was rash (n = 5) in the ceftaroline and eosinophilia (n = 3) in the ceftriaxone group. A complete list of observed ADRs is presented in Table 2.

Table 2.

Adverse Drug Reactions by Treatment Group

	Patients, No. (%)
ADRs	Ceftaroline (n = 50)	Ceftriaxone (n = 50)	P Value
Any ADR	8 (16)	10 (20)	.60
ADR leading to treatment discontinuation	8 (16)	1 (2)	.03
Type of ADR
Rash	5 (10)	1 (2)
Acute kidney injury	1 (2)	2 (4)
Neutropenia	1 (2)	1 (2)
Eosinophilia	0 (0)	3 (6)
Thrombocytopenia	0 (0)	2 (4)
Bronchospasm	1 (2)	0 (0)
Transaminitis	0 (0)	1 (2)

Abbreviation: ADR, adverse drug reaction.

ADRs necessitating premature antibiotic discontinuation occurred more frequently in patients treated with ceftaroline than in those treated with ceftriaxone, occurring in 16% (8 of 50) versus 2% (1 of 50), respectively (P = .03). Of the 8 ceftaroline-treated patients requiring therapy discontinuation, rash occurred in 5, neutropenia in 1, acute kidney injury in 1, and 1 patient developed bronchospasm. The only ceftriaxone-treated patient whose ADR required premature discontinuation had neutropenia.

Only ceftaroline therapy was found to be independently associated with therapy discontinuation in the regression analysis (odds ratio, 10.2; 95% confidence interval, 1.19–87.8; P = .03). Complete results of the multivariate regression analysis are provided in Table 3.

Table 3.

Multivariate Logistic Regression Analysis for Factors Associated with Premature Discontinuation

Variable	OR (95% CI)	P Value
Ceftaroline therapy	10.2 (1.19–87.8)	.03
High-dose therapy	1.15 (.16–14.8)	.72
No. of reported allergies at therapy initiation	1.07 (.76–1.49)	.70

Abbreviations: CI, confidence interval; OR, odds ratio.

DISCUSSION

To date, this is the largest matched, retrospective cohort focusing on comparing ADRs between ceftaroline and a commonly used cephalosporin, ceftriaxone. In this study it was determined there was no difference in the rate of documented occurrence of any ADR between ceftaroline and ceftriaxone (16% vs 20%; P = .60), but ADRs leading to premature discontinuation of therapy occurred more frequently in patients treated with ceftaroline (16% vs. 2%; P = .03). Ceftaroline therapy was the only factor independently associated with the development of an ADR leading to premature discontinuation (odds ratio,10.2; 95% confidence interval, 1.19–87.8).

In evaluating early ADR data with ceftaroline, the CANVAS 1 and 2 trials identified that rash occurred in 3.2% of ceftaroline-treated patients; overall, 3% of patients required therapy discontinuation owing to any ADR [13]. In the FOCUS 1 and 2 trials it is not specified how many patients had rash, but 4.4% of ceftaroline-treated patients did require premature discontinuation owing to an ADR [14]. In the present evaluation, the rate of rash (compared only with the CANVAS trials) was 2–3 times higher and the rate of significant ADRs (compared with both CANVAS and FOCUS trials) was about 4–5 times higher. This could be related to the duration of therapy; the mean duration of ceftaroline treatment in the present evaluation was 17 days, compared with 8.3 and 6.6 days, respectively in the CANVAS and FOCUS trials [13, 14].

Postmarketing data have provided more information regarding rates of ADRs with use of ceftaroline. In a phase III trial evaluating high-dose ceftaroline (600 mg every 8 hours) versus vancomycin and aztreonam for complicated skin and skin structure infections, rash events occurred in 7.9% (of ceftaroline-treated patients 40 of 506) and 9.4% of those receiving vancomycin and aztreonam (24 of 255); 12% of patients in the study had ceftaroline prematurely discontinued [15]. In the present evaluation, the rash event rate and rate of premature discontinuation (10% and 16%) are similar to what was observed by Dryden et al [15]. All patients in that study [15] received high-dose therapy, but only 1 patient in our present cohort was treated with high-dose ceftaroline.

A study by Athans et al [16] compared outpatient ceftaroline use with vancomycin for the treatment of osteoarticular infections. The median duration of ceftaroline therapy was 39 days (longer than in the present evaluation), and 12% of patients treated with ceftaroline required premature discontinuation owing to an ADR; rash occurred in 10% of ceftaroline-treated patients (5 of 50) [16]. The longer duration of therapy, total number of patients, and premature discontinuation and rash rates demonstrated by Athans et al [16] are similar to what was found in the current evaluation. Finally, a retrospective study published by Blumenthal et al [7] evaluated all ADRs associated with ceftaroline use at 2 centers in Massachusetts over a 2-year span. Overall, an ADR occurred in 21% of patients (20 of 96); rash developed in 9.3% (9 of 96). The median duration of treatment was 13 days (similar to that in the present study), and patients with an allergy history, including a history of β-lactam allergies, were not more likely to experience an ADR [7]. Similarly, the present findings did not identify the number of allergies as an independent predictor of ceftaroline discontinuation.

In other postmarketing studies, ceftaroline has been associated with neutropenia; however, in the current matched cohort, only 1 patient in each group met the criteria for this ADR. In the postmarketing data, the rate of neutropenia has ranged from 1.4% to 18% in ceftaroline-treated patients and is possibly related to duration of therapy [2–10]. It is worth noting that in the complete ceftaroline cohort previously published, 3 cases of neutropenia were identified [11]. No cases of neutropenia were described by Dryden et al [15] or Athans et al [16], and Blumenthal et al [7] identified 5 cases (5.2%) [7].

In the present evaluation the ADRs most commonly encountered with ceftriaxone were eosinophilia (n = 3) and acute kidney injury (n = 2). According to the package insert for ceftriaxone, the rate of eosinophilia is 6% and <1% of patients experience an elevation in serum creatinine; rash is reported to occur in 1.7% [17]. In a systematic review by Moskovitz [18] published in 1984, including 153 individual studies with ceftriaxone, 8.1% of patients (215 of 2640) experienced an ADR, with rash or pruritis occurring in 2.1% (55 of 2640) [18]. A study by Mandell et al [19] from 1989 found that 3.1% of patients treated with ceftriaxone (5 of 160) required premature discontinuation owing to ADRs, and rash occurred in 1.8% (3 of 160) [19]. More recently, a 2015 analysis of various β-lactams for outpatient parenteral antibiotic therapy by Lee et al [20] reported the lowest rate of ADR with ceftriaxone, 1.5% (4 of 104), with rash occurring in only 1 patient. Finally, a study by Le Turnier et al [21], evaluating the tolerability of high-dose ceftriaxone for central nervous systems infections in 189 patients, found ADRs in only 17; no patients were reported to have a rash, and the ADR led to discontinuation in only 1 patient. The tolerability of ceftriaxone is reflected in the present evaluation, and consistent with the previous literature, by the low number of ADRs leading to premature discontinuation (n = 1).

The ADRs most associated with β-lactam antibiotics are immune-mediated hypersensitivity reactions [22, 23]. In the current cohort it was more common for ceftaroline-treated patients to experience a hypersensitivity reaction than for those treated with ceftriaxone. It is known that, even in patients with documented β-lactam allergies, side chains attached to the β-lactam ring play a key role in immunogenicity [23]. There are significant differences between ceftaroline and ceftriaxone side chains, with the former being adjusted to incorporate MRSA activity. This is also borne out when one considers the agents’ molecular weights— ceftaroline at 762.75 g/mol and ceftriaxone at 661.60 g/mol [17, 24]. It is possible that the differing side chain and larger molecule size (in terms of molecular weight) make ceftaroline more immunogenic than ceftriaxone and therefore more likely to induce hypersensitivity reactions.

The current evaluation is not without limitations. Because it was retrospective, the assessment of allergies and ADRs was dependent on the accuracy of documentation. However, to help mitigate this limitation, all potential allergies and ADRs were evaluated using the Naranjo algorithm, and only those with a score of ≥4 (indicating possible, probable, or definite ADR) were included in the analysis [12]. In addition, all charts were reviewed for allergies by pharmacists trained in infectious diseases (J. W. J., X. T., and R. P. M.). Moreover, only 1 patient in the ceftaroline group received high-dose therapy, making it difficult in the present evaluation to determine whether the ceftaroline dose affects ADR rates. Finally, it is likely that the inclusion of eosinophilia led to the identification of clinically irrelevant ADRs in the ceftriaxone population.

In conclusion, the current study is, to date, the only matched cohort study comparing ADR rates between ceftaroline and another commonly used cephalosporin antibiotic. This evaluation is representative of the real-world use of these antimicrobials, with many patients receiving treatment for off-label indications and for durations longer than what is recommended in the package inserts. Our present findings suggest that ceftaroline-treated patients, compared with patients of a similar age receiving ceftriaxone for similar indications and duration, experienced significantly more ADRs necessitating premature discontinuation; premature discontinuation was necessary in 100% (8 of 8) of ceftaroline-treated patients who experienced an ADR, compared with 10% (1 of 10) of those treated with ceftriaxone. The most commonly occurring ADR leading to discontinuation was rash. Further studies with a larger, more diverse patient cohort seem reasonable based on the current findings, but it is worth noting that patients receiving courses of ceftaroline longer than what is recommended in the package insert may be more likely to experience ADRs that require premature discontinuation, and those ADRs are not limited to neutropenia.