Pentamidine for Prophylaxis against Pneumocystis jirovecii Pneumonia in Pediatric Oncology Patients Receiving Immunosuppressive Chemotherapy

Pneumocystis jirovecii pneumonia is a life-threatening opportunistic infection in children receiving immunosuppressive chemotherapy. Without prophylaxis, up to 25% of pediatric oncology patients receiving chemotherapy will develop Pneumocystis jirovecii pneumonia.

ABSTRACT

Pneumocystis jirovecii pneumonia is a life-threatening opportunistic infection in children receiving immunosuppressive chemotherapy. Without prophylaxis, up to 25% of pediatric oncology patients receiving chemotherapy will develop Pneumocystis jirovecii pneumonia. Trimethoprim-sulfamethoxazole is the preferred agent for prophylaxis against Pneumocystis jirovecii pneumonia. Pentamidine may be an acceptable alternative for pediatric patients unable to tolerate trimethoprim-sulfamethoxazole. A retrospective review was conducted of pediatric oncology patients who received ≥1 dose of pentamidine for Pneumocystis jirovecii pneumonia prophylaxis between January 2007 and August 2014. Electronic medical records were reviewed to determine the incidence of breakthrough Pneumocystis jirovecii pneumonia or discontinuation of pentamidine associated with adverse events. A total of 754 patients received pentamidine prophylaxis during the period. There were no cases of probable or proven Pneumocystis pneumonia, and 4 cases (0.5%) of possible Pneumocystis pneumonia. The incidence of possible breakthrough Pneumocystis pneumonia was not significantly different between subgroups based on age (<12 months [1.7%] versus ≥12 months [0.4%], P = 0.3), route of administration (aerosolized [0%] versus intravenous [1.0%], P = 0.2), or hematopoietic stem cell transplant status (transplant [0.4%] versus no transplant [0.8%], P = 0.6). Pentamidine was discontinued due to an adverse drug event in 23 children (3.1%), more frequently for aerosolized than for intravenous administration (7.6% versus 2.2%, respectively, P = 0.004). Intravenous or inhaled pentamidine may be a safe and effective second-line alternative for prophylaxis against Pneumocystis jirovecii pneumonia in children with cancer receiving immunosuppressive chemotherapy or hematopoietic stem cell transplantation.

INTRODUCTION

Pneumocystis jirovecii pneumonia (PCP) is a life-threatening opportunistic infection in children receiving immunosuppressive chemotherapy. PCP was the leading cause of death in pediatric patients with acute lymphoblastic leukemia (ALL) prior to prophylactic administration of antimicrobials (1). In the absence of prophylaxis, up to 25% of pediatric oncology patients receiving chemotherapy develop PCP (1, 2). The case fatality rate of untreated PCP outside the neonatal period is near 100%, and even with early recognition, improved diagnostic techniques, and appropriate antimicrobial therapy, a mortality rate of up to 32% has been reported in pediatric oncology patients (3, 4).

Randomized placebo-controlled trials have shown trimethoprim-sulfamethoxazole (TMP-SMX) administered twice daily, 3 days per week, is effective for the prevention of PCP, and so TMP-SMX is the preferred agent for PCP prophylaxis with a reported prevention rate of 93% to 100% (5). However, patients unable to tolerate TMP-SMX due to adverse drug reactions require prophylaxis with an alternative agent. Although other drugs with activity against Pneumocystis jirovecii have been identified, including dapsone, atovaquone, and pentamidine, limited data exist regarding the safety and efficacy of these in this population.

Small retrospective studies of aerosolized pentamidine for PCP prophylaxis in pediatric oncology patients have reported an efficacy similar to that of TMP-SMX (6, 7). The commonly reported adverse effects associated with aerosolized pentamidine include bronchospasm, wheezing, and cough, which may occur more commonly in patients with chronic respiratory diseases (6–8). The administration challenges, such as the requirement of a negative pressure room, specialized respiratory equipment, and personnel training, may have limited its use in some centers. However, the National Institute for Occupational Safety and Health recently modified its assessment of this drug and no longer considers pentamidine to be a hazardous medication (9). Aerosolized pentamidine is often avoided in young children due to concerns about their ability to receive the entire dose by this route (10).

Although intravenous (i.v.) pentamidine is an FDA-approved treatment modality for PCP, it is considered a second-line therapy (11). The reported adverse effects include dysglycemia, hypotension, phlebitis, fatigue, dysgeusia, nephrotoxicity, electrolyte imbalances, allergic reactions, hepatotoxicity, and pancreatitis (12, 13). Recent studies have reported acceptable breakthrough PCP rates (0% to 1.3%) with the use of i.v. pentamidine for PCP prophylaxis in immunocompromised children receiving chemotherapy or status post hematopoietic stem cell transplant (HSCT) or solid organ transplant (SOT) (10, 14, 15).

To date, no studies have directly compared aerosolized and i.v. pentamidine for PCP prophylaxis in pediatric oncology patients. This retrospective analysis was performed to describe our experience using aerosolized and i.v. pentamidine for PCP prophylaxis in children receiving immunosuppressive chemotherapy at St. Jude Children's Research Hospital.

RESULTS

Patient characteristics.

A total of 754 patients received 3,991 doses of aerosolized or i.v. pentamidine for PCP prophylaxis during the study period. The median number of doses received was 3 (range, 1 to 33). The patient demographic characteristics are summarized in Table 1. Sixty children received their initial dose of pentamidine while less than 1 year of age (8%); 117 received their initial dose while less than 2 years of age (15.6%). Of the total number of patients receiving pentamidine, the majority received the i.v. formulation (Table 1). Most children receiving aerosolized pentamidine were given the maximum dose (300 mg/dose every 4 weeks), while two children (<5 years of age) received approximately 7 mg/kg/dose every 4 weeks. Aerosolized pentamidine was administered by s Respirgard II nebulized system. Intravenous pentamidine was administered at 3 to 4 mg/kg/dose infused over 60 min every 4 weeks. HSCT recipients represented 30% (n = 229) of the study population; the majority of non-HSCT recipients were being treated for nonhematologic malignancies (n = 348, 46%).

TABLE 1.

Baseline demographics

Demographic	Administration route			Total
	Aerosolized	i.v.	Aerosolized and i.v.
No. of patients	158	508	88	754
Median pentamidine doses per patient (n [range])	2 (1–31)	3 (1–31)	3 (1–33)	3 (1–33)
Median age at first dose (yr [range])	14.3 (4.5–22.7)	4.8 (0.08–24)	14.4 (2.9–24.8)	8 (0.08–24)
Sex (n [%])
Male	96	274	50	420 (55.7)
Female	62	234	38	334 (44.3)
Patient age (yr) at first dose (n [%])
<1		60		60 (8.0)
1 to <2		57		57 (7.6)
≥2	158	391	88	637 (84.4)
Hematopoietic stem cell transplant (n [%])	34	158	37	229 (30.4)
Solid tumor/neuro-oncology (n [%])	55	262	31	348 (46)
Leukemia/lymphoma (n [%])	69	88	20	177 (23.6)

Probable or proven PCP.

No episodes of probable or proven PCP were identified during the study period (Table 2). Four patients (0.5% of patients; 0.03 cases per 1,000 patient days) met criteria for possible PCP (Table 2). Although these patients presented with a compatible clinical syndrome, none had documented PCP testing and all died shortly after presentation without having PCP treatment initiated. These patients had received prophylaxis with i.v. pentamidine; case descriptions are shown in Table S2 in the supplemental material.

TABLE 2.

Breakthrough Pneumocystis pneumonia

Comparison	No. of patients	No. (%) with probable or proven PCP	No. (%) with possible PCP	P valuea	Possible PCP
					Rate per 1,000 days	95% CI
All patients	754	0 (0)	4 (0.5)		0.03	0.009 to 0.07
By age (yr)
<1 (infants)	60	0 (0)	1 (1.7)	0.21b	0.08	0.004 to 0.4
<2	117	0 (0)	1 (0.9)	0.59	0.05	0.002 to 0.2
≥2	637	0 (0)	3 (0.5)		0.03	0.006 to 0.07
By dosage route
Aerosolized only	158	0 (0)	0 (0.0)	0.26	0	0 to <0.2
i.v. only	508	0 (0)	4 (0.8)		0.04	0.0006 to 0.05
Aerosolized and i.v.	88	0 (0)	0 (0.0)		0	0 to <0.2
HSCT
Recipients	228	0 (0)	1 (0.4)	0.81	0.02	0.001 to 0.1
Nonrecipients	526	0 (0)	3 (0.6)		0.03	0.008 to 0.08

^aChi-square or Fisher's exact tests were performed for subgroup statistical analyses.

^bAnalysis of patients <1 year versus patients ≥1 year of age.

Pentamidine discontinuation and documented adverse events.

Pentamidine was discontinued due to adverse drug events in 23 children (3.1%). The most common adverse drug event was anaphylaxis, which occurred in seven patients (0.9%). Two of these patients received only aerosolized pentamidine, 4 received only i.v. pentamidine, and 1 received both aerosolized and i.v. pentamidine prior to the event. No further dosing of pentamidine was attempted in these patients. Additional adverse events leading to the discontinuation of pentamidine included respiratory complications such as wheezing, bronchospasm, shortness of breath and cough, neuropathy such as tingling or numbness, hypotension, and an altered mental status. Respiratory complications were more common in patients receiving aerosolized pentamidine than in those receiving i.v. pentamidine (2.5% versus 0.2%, respectively, P = 0.01). Three patients receiving i.v. pentamidine experienced hypotension, which led to the discontinuation of the drug. There were no cases of hypotension among those receiving aerosolized pentamidine. Of the 23 patients who required discontinuation, only 3 received 6 months or longer of therapy. The adverse events leading to pentamidine discontinuation stratified by the route of administration are summarized in Table 3.

TABLE 3.

Adverse events leading to discontinuation of pentamidine

Adverse event	No. of patients
	Aerosolized pentamidine (n = 12)	i.v. pentamidine (n = 11)
Anaphylaxis	3	4
Respiratory complicationsa	4	1
Hypotension	0	3
Pancreatitis	2	1
Tingling/numbness	1	2
Altered mental status	1	0
Undescribed event	1	0

^aRespiratory complications include bronchospasm, cough, wheeze, and/or shortness of breath.

Twenty-four children (3.2%) developed laboratory findings with/without signs/symptoms consistent with pancreatitis while receiving pentamidine, and pentamidine was discontinued in 3 cases. Pancreatitis was observed in 5.1% of patients receiving aerosolized pentamidine and in 3.2% of patients receiving i.v. pentamidine (P = 0.3). Grade 2 and grade 3 pancreatitis were seen in 11 patients (1.5%) and 13 patients (1.7%), respectively. Of the 24 patients who developed pancreatitis, seven were concomitantly receiving medications which have been reported to induce pancreatitis; five were receiving an asparaginase or thiopurine product, one was receiving ponatinib and then imatinib, and one was receiving infliximab. The remaining 17 patients who developed pancreatitis were not concomitantly receiving a medication known to induce pancreatitis.

DISCUSSION

This study represents the largest cohort study of pediatric oncology patients receiving pentamidine as a routine second-line PCP prophylaxis and the only study to describe clinical experience with both aerosolized and i.v. pentamidine administrations.

No patients in this cohort developed probable or proven PCP while receiving pentamidine prophylaxis during the study time frame, and all patients with possible PCP had presumptive alternative diagnoses. Although atypical forms of pulmonary PCP have been reported in adult patients with HIV receiving inhaled pentamidine, the broad definition of possible PCP used in this study includes these cases, so this should not adversely affect the generalizability of this finding (16). These findings compare favorably with previously published retrospective studies analyzing either aerosolized or i.v. pentamidine for PCP prophylaxis. Two other studies reported no incidents of breakthrough PCP among pediatric oncology patients receiving aerosolized pentamidine for PCP prophylaxis (6, 7). The reported incidence of breakthrough PCP among pediatric oncology and solid organ transplant patients ranges from 0% to 1.3% overall (10, 14, 15, 17–20). Our findings are comparable to breakthrough PCP incidence rates reported for TMP-SMX (5). PCP incidence rates stratified by route of administration and HSCT status were also comparable to previously published reports examining these populations (6, 7, 15, 18). The consistency of these findings strengthens the evidence for the efficacies of both aerosolized and i.v. pentamidine as PCP prophylaxis.

Kim et al. reported a concerning trend toward an increased risk of breakthrough PCP in younger patients receiving i.v. pentamidine for PCP prophylaxis, with an incidence of 6.5% (1.1% per patient month) for PCP in 31 children <2 years of age and 9.1% (1.9% per patient month) in children <2 years who underwent HSCT (10). Kruizinga et al. reported a 0.9% breakthrough rate for PCP in 106 pediatric patients receiving pentamidine prophylaxis, which is comparable to the breakthrough rate for TMP-SMX (15). The one patient in that study who had a breakthrough infection was a 1-year-old at a status post HSCT of 18 weeks (15). Overall, subsequent studies have not corroborated this concern, reporting no incidence of PCP in HSCT recipients <2 years of age receiving i.v. pentamidine every 2 or 4 weeks (15, 18). Furthermore, Solodokin et al. reported no incidence of PCP in 24 hematology/oncology patients <2 years of age who received i.v. pentamidine for PCP prophylaxis, but HSCT status was not reported (20). We found no cases of probable or proven PCP in 117 children <2 years of age, of whom 23 were HSCT recipients. We did identify one case of possible PCP in an 11-month-old who received a second HSCT at our institution. This patient presented with respiratory failure and subsequently died from diffuse alveolar hemorrhage without the exclusion of an infectious etiology. It has been suggested that young children (<2 years of age) who receive HSCT may be at a higher risk for breakthrough PCP due to the continuous immunosuppression which impairs the clearance of P. jirovecii early in life (10, 17). A very large sample size would be required to compare the risk of breakthrough PCP between different prophylaxis regimens, even in this high-risk population. In the present study, i.v. pentamidine appeared to be an acceptable agent for the prevention of PCP in HSCT recipients, including those <2 years of age.

The reported discontinuation rates of pentamidine due to adverse drug events in previously published studies range from 0.7% to 6%; these events included anaphylaxis, tachycardia, nausea, and pancreatitis (7, 14, 15). Similarly, the present study found an overall rate of pentamidine discontinuation due to an adverse drug event of 3%. Anaphylaxis was the most common adverse effect leading to discontinuation and occurred following both i.v. and aerosolized therapies. Discontinuation due to respiratory complaints occurred at approximately the same frequency as anaphylaxis in those given aerosolized therapy. Respiratory complications leading to discontinuation appeared to be less frequent among those receiving i.v. pentamidine. None of the patients who developed respiratory complications from aerosolized pentamidine had a history of reactive airway disease, a finding which is different than that reported by Macesic et al. (8). However, two of their reported cases had undergone HSCT, possibly contributing to the respiratory compromise. No patients in this study had pentamidine discontinued due to tachycardia, but pancreatitis was the reason for discontinuation in three patients (0.4%), which is similar to the rate reported by Clark et al. (0.6%) (14). The rate of pancreatitis is notable; however, a sensitive definition was utilized to ensure we captured all potential cases of pancreatitis.

There were some limitations of this study due to its retrospective nature. While portions of the medical record were electronic throughout the duration of time included in this study, our institution converted to fully electronic medical records in 2010, approximately midway through the study period. This conversion created a potential for inconsistency in our ability to capture all pertinent data from records prior to this transition. A comparison of the rates of discontinuation due to an adverse event during the two periods did not demonstrate a significant difference (3.3% versus 2.6%), suggesting that the documentation of these events occurred in similar fashions during both periods. The data regarding the discontinuation rate comparison between the two periods are summarized in Table S3 in the supplemental material. Additionally, inconsistencies in the documentation of symptoms associated with PCP, diagnostic tests, and diagnostic rationale may have created a negative detection bias. To address this potential bias, we opted to utilize the more sensitive surveillance definition rather than a specific definition for PCP (see Table S1) to capture all possible cases. There was no active comparator group who received TMP-SMX or other agents in this study, and so the published rates of breakthrough PCP in pediatric oncology patients are described. Despite the absence of probable or proven PCP in this cohort study, we have seen proven breakthrough PCP infections during prophylaxis with pentamidine at our institution at other times, and so clinicians should be mindful that this can occur. Finally, this study was conducted at a single institution; the rationale for the use of pentamidine, the choice of route of administration, and the discontinuation of pentamidine due to adverse events may be influenced by characteristics specific to this institution or the individual clinicians involved in the care of this cohort of patients.

In summary, while TMP-SMX remains the gold standard for PCP prophylaxis in pediatric oncology patients, we found that both aerosolized and i.v. pentamidine had rates of breakthrough PCP comparable to those published for TMP-SMX, making pentamidine appear to be an effective alternative for PCP prophylaxis. Further, i.v. or aerosolized pentamidine had an acceptable safety profile, with low rates of events necessitating the discontinuation of this agent. In the cohort study, neither infant patients nor HSCT recipients, who are at a high risk for developing infections, developed probable or proven PCP, suggesting no difference in effectiveness in these high-risk populations. However, clinicians should remain aware that breakthrough PCP can occur with pentamidine, just as it may occur with other therapies.

MATERIALS AND METHODS

A single-center, retrospective chart review of pediatric oncology patients who received at least one dose of pentamidine for PCP prophylaxis at St. Jude Children's Research Hospital between 1 January 2007 and 31 August 2014 was conducted. Patients were identified using a fully integrated electronic medical record (EMR). Patients were considered eligible if they were treated for a hematologic or solid tumor malignancy or received an HSCT for the treatment of a malignancy. Patients with nonmalignant hematologic disorders were excluded. This study was approved by the St. Jude Children's Research Hospital Institutional Review Board prior to the initiation of data collection.

Primary objective.

The primary objective was to determine the incidence of probable or proven breakthrough PCP while receiving pentamidine for PCP prophylaxis. At St. Jude, patients with suspected PCP or with unexplained respiratory illness refractory to standard therapy typically undergo bronchoalveolar lavage to identify the presence of pneumocystis. During the study period, PCR gradually replaced immunofluorescence as the predominant test of PCP (21). Potential PCP cases were identified by the review of patient EMRs, ICD-9 codes (136.3), pharmacy and pathology records, and death summaries. Identified cases were classified as possible, probable, or proven PCP using a surveillance definition (Table S1 in the supplemental material). A compatible clinical syndrome was defined as a lower respiratory tract infection in an immunocompromised patient with at least one of the following: cough, shortness of breath, increased work of breathing, hypoxia, or lung crepitations. If imaging was performed, ground glass or consolidative opacities, cystic changes, linear-reticular opacities, solitary or multiple nodules, or parenchymal cavities were considered consistent with PCP. Immunocompromised patients were defined as those within 6 months of completion of chemotherapy or corticosteroid therapy of greater than 20 mg prednisone equivalent per day or those within 6 months of HSCT.

Secondary objectives.

Subgroup analyses were performed to compare the incidence of breakthrough PCP on the basis of age at first administration (<12 months versus ≥12 months of age), route of administration (aerosolized versus i.v.), and HSCT status. The safety and tolerability of pentamidine were assessed using the incidence of adverse events directly leading to the discontinuation of pentamidine and the incidence of pancreatitis within the at-risk time period, defined as 28 days after the last administration. Adverse events were identified by a review of the completed medical record. The review focused on clinical notes and medication allergy/adverse reaction information, with attention to those notes recorded following each administration of the drug. Pancreatitis was also identified by capture of ICD-9 coding consistent with this diagnosis, by review of laboratory results demonstrating elevated amylase and lipase, and by diagnostic imaging results with a notation of the finding. Pancreatitis was defined using the common terminology criteria for adverse events (CTCAE) version 4.03 definitions for elevated pancreatic enzymes and pancreatitis (22). Episodes of pancreatitis were evaluated for prior administration of concomitant medications also known to cause pancreatitis. Subgroup analyses were also performed to compare the incidence of adverse drug events and the development of pancreatitis on the basis of the route of administration.

The EMR was queried to identify patients who received pentamidine dosing for PCP prophylaxis (9 mg/kg/dose or a maximum of 300 mg/dose via nebulizer or 3 to 4 mg/kg/dose i.v. every 4 weeks). Queries were performed to capture each patient's primary admitting service, patients who were assigned ICD-9 codes for PCP and acute or chronic pancreatitis, the date such ICD-9 codes were assigned, amylase and lipase levels measured within the reporting window, and the date of each patient's death (if applicable). Individual medical records were reviewed for reported allergies to TMP-SMX and/or pentamidine, adverse events associated with or attributed to pentamidine, and the cause of death (if applicable). The following search terms were used to review clinical documents within the EMR: Septra, sulfamethoxazole-trimethoprim; Pentam, pentamidine; PCP, Pneumocystis pneumonia; Pneumocystis carinii DNA assay; Pneumocystis jirovecii DNA assay; BAL, bronchoalveolar lavage; and pancreatitis.

Statistical analysis.

Descriptive statistics were used to summarize patient demographic data, the incidence of probable or proven breakthrough PCP while receiving pentamidine, and the incidence rates for secondary objectives. We constructed exact 95% confidence intervals for the incidence of PCP infection and the incidence of adverse events. The chi-square or Fisher's exact test (for small sample sizes) was used to compare the incidence of possible PCP between subgroups, with a P value of <0.05 signifying statistical significance. All statistical tests were computed with Statistica 64 (TIBCO, Palo Alto, USA).