Introduction

The Infectious Diseases Society of America (IDSA) guidelines recommend that patients with febrile neutropenia (FN) who are at high risk of infection should be hospitalized for empiric administration of intravenous antibiotics [1]. Monotherapy with antipseudomonal β-lactams, such as cefepime, carbapenem, or piperacillin-tazobactam, is recommended as first-line empiric antibacterial therapy in such patients [1].

The efficacy of meropenem (MEPM) as empiric monotherapy for neutropenia in pediatric or adult patients with cancer has been assessed in randomized comparative trials against cefepime [2, 3], imipenem/cilastatin [4], ceftazidime with [5] or without amikacin [6], and piperacillin-tazobactam [7]. Because of those studies, MEPM is considered a standard antibiotic for the empiric treatment of FN in patients with hematologic malignancies [1, 8].

Doripenem (DRPM) is a relatively new carbapenem that was approved in Japan, the USA, and the European Union in 2005, 2007, and 2008, respectively. Both DRPM and MPEM possess broad-spectrum in vitro activity against many gram-positive, gram-negative, and anaerobic bacteria. DRPM is reportedly therapeutically noninferior to MEPM in adults with complicated intra-abdominal infections [9]. Some studies have shown that DRPM is a promising agent for the empiric treatment of sepsis in patients with neutropenia and hematologic malignancies [10], and a retrospective analysis [11] has demonstrated similar efficacies between MEPM and DRPM for the treatment of febrile patients with acute leukemia and myelodysplastic syndrome (MDS). However, data concerning the safety and efficacy of DRPM in such patients is limited. Thus, we conducted a trial to compare the safety and efficacy of DRPM and MEPM as first-line empiric antibacterial therapy of FN in patients with acute leukemia or MDS-refractory anemia with excess blasts (RAEB)-2.

Methods

Study design

This randomized, cooperative, open-label prospective trial included 133 FN patients with acute leukemia or MDS-RAEB-2 who received induction or consolidation chemotherapy between August 1, 2011, and December 31, 2013. The institutional review board of each hospital approved the study protocol, and written informed consent was obtained from each patient or legal guardian before enrollment. This trial was registered with the UMIN Clinical Trials Registry (UMIN-CTR #UMIN000006124).

Participants

Eligibility criteria included hospitalization for FN, ≥ 16 years of age, diagnosis of acute leukemia or MDS-RAEB-2, and no intravenous antibiotic therapy within 1 week of the study. Fever was defined as an oral temperature measurement of ≥ 38.0 °C [1] or axillary temperature of ≥ 37.5 °C [12] sustained over a 1-h period, and neutropenia was defined as a neutrophil count less than 0.5 × 109/L after chemotherapy. The exclusion criteria included hepatic or renal insufficiency, known allergy or other contraindication to the study drugs, and pregnancy or lactation.

Randomization on a 1:1 basis was performed on a centralized website for FN patients (Fig. 1).

Fig. 1
figure 1

Inclusion of patients and placement in groups. DRPM doripenem, MEPM meropenem

Full size image

Diagnostic evaluation

Before the start of DRPM or MERM therapy, bacterial examinations of the throat, blood, sputum (if available), and urine samples were performed; complete physical exams were conducted, after which chest radiographs were taken; and laboratory values including complete blood counts were recorded. Thereafter, bacterial cultures were collected as needed, and laboratory values (including complete blood counts) were assessed at least twice weekly. Diagnoses of catheter-related infections were confirmed by isolation of identical microorganisms from catheter materials and blood cultures. Coagulase-negative staphylococci or corynebacteria were accepted as significant pathogens only when isolated from at least two separate consecutive blood cultures. Chest CT scans were performed in cases of persistent FN in which antibiotics had been properly administered and fungal infections were suspected.

Study drug administration

DRPM or MEPM (1 g each) was administered intravenously over a 1-h period every 8 h. The study drugs were continued for at least 5 days, in the absence of toxicity, to evaluate for efficacy. Discontinuation for toxicity was at the individual investigator’s discretion. Although the criteria for discontinuation were not predefined, an independent data monitoring committee confirmed whether discontinuation of the study drug was appropriate.

If fevers persisted after the first 3–4 days of empiric therapy, co-administration with another antibiotic (usually a glycopeptide) was considered [1]. IDSA guidelines were followed for the addition of anti-MRSA agents in cases of clinically suspected serious catheter-related infections, as well as for the empirical addition of antifungal drugs in high-risk patients (i.e., those with persistent fever of unknown origin [FUO] despite treatment with 4–7 days of broad-spectrum antibiotics) [1].

Assessment criteria of efficacy and safety

The primary endpoint was defined as fever resolution (< 37 °C) within 3 to 5 days of initial treatment that was maintained for at least two consecutive days without change in therapy (i.e., “success rate”) [13]. Treatment failure was defined as no improvement or worsening of infection while receiving the initial study regimen, or the necessary addition of other antibacterial or antifungal drugs.

The secondary endpoints consisted of (1) resolution of fever by day 7 with or without treatment modification with other antibacterial or antifungal drugs (i.e., “overall response rate”); (2) resolution of fever by day 14; (3) survival rate on day 30; and (4) frequency of abnormal laboratory data and adverse events (AEs). An independent data monitoring committee assessed the AEs using Common Terminology Criteria for Adverse Events version 4.0 (CTCAE ver. 4) after each personal physician’s review and analysis of the data.

Sample size

Because no evidence has established the superiority of either study drug (DRPM or MEPM) for the treatment of FN, the sample size was determined using the efficacy of both drugs. In the Japanese population, the efficacy rates of DRPM (1.0 g every 8 h) and MEPM (1.0 g every 8 h) for FN in patients with sepsis are 69.2% and 42.0%, respectively [14, 15]. Therefore, we set the efficacy rates at 70% for DRPM and 40% for MEPM, and a power of 0.90 with 5.0% significance was used for the final analysis. The calculated total sample size (using JMP, version 8.0.2.2) was 126 patients (63 for each group).

Statistical analysis

The primary endpoint was evaluated using intention-to-treat analysis. Categorical variables were tested using the Wilcoxon rank-sum or chi-square tests. P values < 0.05 were considered statistically significant.

Results

Enrollment and patient characteristics

One hundred thirty-three patients recruited from six hospitals in Japan were equally assigned to the DRPM (n = 65) and MEPM (n = 68) groups (Fig. 1). There were no significant differences between the groups with regard to age, sex, underlying diseases, and chemotherapy treatment prior to onset of FN (Table 1). The DRPM group had one patient with MDS-RAEB-2 and the MEPM group had two patients with MDS-RAEB-2. Induction chemotherapy was performed in 43.1% (28/65) of the DRPM group and 47.1% (32/68) of the MEPM group; consolidation chemotherapy was performed in 55.4% (36/65) of the DRPM group and 51.5% (35/68) of the MEPM group. The antileukemic agents used for induction and consolidation chemotherapy were anthracycline and cytarabine. The median durations of treatment were similar (P = 0.071) between the DRPM and MPEM groups (9 days and 12 days, respectively; Table 1). No antibacterial agents were administered as prophylaxis in this study.

Table 1 Demographic and baseline characteristics
Full size table

Duration of neutropenia and neutrophil recovery

The median durations of neutropenia (neutrophil count < 500/μL) were 12 days (range, 1 to 80 days) and 15 days (range, 1 to 106 days) in the DRPM and MEPM groups, respectively (P = 0.1719). Neutrophil recovery (neutrophil count > 500/μL within 5 days after administration of DRPM or MEPM) was observed in 36.9% (24/65) of the DRPM group, and 22.1% (15/68) of the MEPM group (P = 0.0907). Neutrophil recovery with resolution of fever within 5 days of DRPM or MEPM administration did not significantly differ between the DRPM and the MEPM groups (P = 0.065).

Type of infection and causative microorganisms

Infections were detected in 33.8% (22/65) of DRPM patients and 39.7% (27/68) of MEPM patients (Table 2). Microbiologically documented infections (MDIs) occurred in 32.3% (21/65) of the DRPM group and 36.8% (25/68) of the MEPM group. Clinically documented infections (CDIs), including radiologically confirmed pulmonary infiltration and liver abscess diagnosed by computer tomography, were found in 3.1% (2/65) of the DRPM group (pneumonia: 2 cases) and 7.4% (5/68) of the MEPM group (pneumonia: 4 cases; liver abscess: 1 case; Table 2). FUO was determined for 66.2% (43/65) of the DRPM group and 60.3% (41/68) of the MEPM group (Table 2).

Table 2 Types of infections in each study group
Full size table

No differences in types of infections were observed between the two groups. The majority of MDIs were bacterial or fungal blood stream infections (BSIs), some of which were documented venous catheter-related infections. Gram-positive bacteria were detected in 77.3% (17/21) and 69.2% (18/24) of BSIs in the DRPM and MEPM groups, respectively; gram-negative bacteria were detected in 18.2% (4/21) and 26.9% (7/24) of BSIs in the DRPM and MEPM groups, respectively; and fungi were detected in 4.5% (1/21) and 3.9% (1/24) of BSIs in the DRPM and MEPM groups, respectively (Table 3). Superinfection with gram-positive and gram-negative bacteria was observed in one of the BSIs in the DRPM group and two of the BSIs in the MEPM group (Table 3). The most common causative microorganisms were Streptococcus spp. and Staphylococcus spp., including methicillin-resistant staphylococci (Table 3).

Table 3 Microorganisms in blood infections
Full size table

Efficacy

The success rate (resolution of fever within 3–5 days without any modification of therapy) was higher in the DRPM group (60.0%; 39/65) than in the MEPM group (45.6%; 31/68), although the difference was not significant (P = 0.136; Table 4). Subtype analysis revealed that patients with FUO had a success rate of 69.8% (30/43) and 58.5% (24/41) in the DRPM and MEPM groups, respectively (P = 0.397); and those with MDIs had a success rate of 42.9% (9/21) and 29.2% (7/24) in the DRPM and MEPM groups, respectively (P = 0.518; Table 4). We noted that the success rates did not depend on neutrophil recovery: success rates for the DPRM group were 58.3% (14/24) and 60.9% (25/41), with and without neutrophil recovery, respectively; success rates for the MEPM group were 33.3% (5/15) and 49.0% (26/53), with and without neutrophil recovery, respectively (P = 0.233, P = 0.346; Table 5).

Table 4 Success rates according to infection type
Full size table
Table 5 Success rates according to neutrophil recovery
Full size table

The most frequently added antimicrobial drugs were glycopeptides, which were added as empiric or targeted antibacterial therapy of persistent or recurrent fever in patients with MDIs caused by gram-positive bacteria, including methicillin-resistant species (Table 6). In all, 40.0% (26/65) of the DRPM group and 32.3% (22/68) of the MEPM group received glycopeptides. Antifungal medications were added as empiric or preemptive therapy of persistent or recurrent fevers believed to be caused by fungal infections; those medications were administered to 21.5% (14/65) of the DRPM group and 13.2% (9/68) of the MEPM group (Table 6).

Table 6 The frequency of cases needed for anti-MRSA agent and/or antifungal agent
Full size table

By day 7, the overall response rate was significantly better in the DRPM group (78.4%, 51/65) than in the MEPM group (60.2%, 41/68; P = 0.037). The cumulative number of afebrile cases by day 14 was 87.7% (57/65) and 83.8% (57/68) in the DRPM and MEPM groups, respectively (P = 0.523; Table 7). Survival rates 30 days after the start of administration were 98.5% (64/65) and 100% (68/68) in the DRPM and MEPM groups, respectively (P = 0.304; Table 7).

Table 7 Assessment of secondary endpoints
Full size table

Safety

The rates of AEs did not significantly differ between the DRPM (38.5%, 25/65) and MEPM groups (41.2%, 26/68; P = 0.749; Table 8). The AEs were grades 1–2 for all but one patient. Although all patients were able to continue taking their respective study drugs until at least day 5, one patient discontinued on day 3 because of resolution of fever.

Table 8 Adverse events of each study group
Full size table

Discussion

To the best of our knowledge, this is the first prospective, randomized study to compare the safety and efficacy of DRPM to that of MEPM for the treatment of FN in patients with acute leukemia or MDS-RAEB-2.

In previous studies, the success rate of MEPM monotherapy on day 3 ranged from 35 to 87.5% for patients with FN. For clinical studies in which more than 40% of patients had acute leukemia, MEPM monotherapy has shown relatively low success rates: 56% [5], 37.0% [16], 35% [17], 44% [6]. For clinical studies that primarily included patients with malignant lymphoma and solid tumors—but not leukemia—MEPM monotherapy has shown relatively high success rates. In our study, nearly all patients (96.9%) had acute leukemia and were undergoing remission induction or consolidation therapies. We found that the success rate of MEPM treatment was 45.6% and that of DRPM treatment was 60.0%. Remarkably, the success rate did not depend on neutrophil recovery; both drugs were effective whether or not neutrophil recovery occurred within 5 days after administration of DRPM or MEPM (MEPM, 49.0%; DRPM, 60.9%). This result suggests that the clinical efficacy of DRPM was noninferior to that of MERM for the treatment of FN associated with hematological malignancies such as acute leukemia.

The distribution of infection types in our study differed from that of other studies. We observed a higher proportion of patients with FUO (63.2% vs. reported incidence of 29.3–60.9%), a lower proportion of patients with CDIs (5.3% vs. reported incidence of 10–24.7%), and a similar proportion of MDIs (34.6% vs. reported incidence of 18.1–53.6%). Sub-analysis of our patients revealed that DRPM-treated patients with FUOs and MDIs had success rates of 69.8% and 42.9%, respectively. In comparison, MEPM-treated patients with FUOs and MDIs had success rates of 58.5% and 29.2%, respectively. Our results were similar to those of previous studies in which patients with FUOs had higher response rates than those with MDIs.

The relatively low success rate in patients with MDIs was likely due to the high frequency of catheter-related infections. Those infections were primarily caused by gram-positive bacteria, some of which were predictably insensitive to the study drugs. We detected a total of 48 BSIs that were caused by 11 species of microorganisms. Thirty-five of the BSIs were caused by gram-positive bacteria (72.9%) and 11 were caused by gram-negative bacteria (22.9%). Considering the predominance of staphylococci and streptococci in patients with MDIs in our study, we recommend attention be paid to gram-positive bacteria in catheter-related infections.

Recent trends in the epidemiology of microorganisms isolated from cancer patients with bacteremia indicate a predominance of gram-negative bacteria and the global presence of antibiotic resistance in gram-negative and gram-positive bacteria [18]. An epidemiological study of bacteremia among FN patients with hematologic malignancies who had not received antibiotic prophylaxis was performed in Japan from 2006 to 2009 [19]; it demonstrated that 48.1% of the bacterial isolates are gram-negative, and 45.5% are gram-positive [19]. Nevertheless, vancomycin is not considered standard first-line empiric antibiotic therapy for FN because many randomized studies have shown no significant reduction in duration of fever and overall mortality [20].

The overall response rate on day 7 revealed that DRPM (78.4%) was significantly more effective than MERM (60.2%; P = 0.037). The reason for that is not known; however, it might be explained by the fact that DRPM is slightly more potent than MEPM against gram-positive aerobic bacteria [21].

Per our study protocol, anti-MRSA agents were added to the study drug regimen in 36.1% (48/133) of all patients as empiric or targeted antibiotic therapy for persistent or recurrent fever from MDIs, some of which were caused by gram-positive bacteria, including methicillin-resistant species. Antifungal agents were added to the study drug regimen in 17.3% (23/133) of all patients as empiric or preemptive antifungal therapy for persistent or recurrent fever believed to be caused by fungal infections. Because the frequency of addition of anti-MRSA and antifungal agents did not differ between the DRPM and MEPM groups, we do not believe that therapy modifications influenced the results. In most cases of modified treatments, chest X-rays were routinely performed on the first or second day of FN; chest CT scans were not routinely performed. All treatment modifications were considered treatment failure. Nevertheless, fevers resolved in 85.7% of all patients by day 14, and the survival rates on day 30 were similar between both groups. These data suggest that DRPM was as effective as MEPM.

The frequencies of AEs did not significantly differ between the DRPM and MEPM groups. Most of the AEs were grade 1–2 hepatic and renal dysfunctions, as was expected. None of the patients required drug discontinuation, and infection-related mortality was not observed. These results suggest that both drugs were well-tolerated by patients with high-risk FN.

In conclusion, the efficacy of monotherapy was similar between the DRPM and MEPM groups, but co-administration of anti-MRSA and/or antifungal drugs with DRPM was significantly more effective than DRPM alone. Thus, empiric therapy with DRPM or MEPM provides a safe and well-tolerated option for the treatment of FN in patients with acute leukemia and RAEB.