Prevalence and predictors of pulmonary fungal infections in patients with acute leukemia and aggressive lymphomas: Implications for cancer care in developing countries

INTRODUCTION

Fungi rarely cause lower respiratory tract infections in immunocompetent patients because the host anatomical and innate immune defenses clear the fungal elements and prevent clinical disease.^[1,2] In patients with hematological malignancies, the pulmonary mucociliary defense may be impaired by chemotherapy or radiation, predisposing patients to fungal infections. Moreover, such patients may experience neutropenia from defective marrow function or cancer therapy, thereby further heightening their risks of invasive fungal infections. Pulmonary fungal infections (PFIs) can present primarily in the lungs or as manifestation of disseminated fungal infections.

PFIs account for most of the invasive fungal diseases reported in the literature. A report from the Center of Expertise in Mycology in the Netherlands shows that a majority of adults and children evaluated for severe fungal infections suffered from pulmonary infections or signs/symptoms.^[3] In a multicenter trial of posaconazole and voriconazole for the treatment of adult patients at risk for invasive aspergillosis, 80% of study patients had infections limited to the lower respiratory tract (primarily lung).^[4] Estimates from Nigeria among study population 12 years to 60 years showed 40% prevalence rate of pneumocystis pneumonia in children and 5-year period prevalence of 60,377 cases of chronic pulmonary aspergillosis in patients with pulmonary tuberculosis.^[5]

Due to the poor outcomes reported with invasive fungal infections,^[6-8] antifungal prophylaxis is increasingly being recommended for patients considered to be high risk for these infections. The Infectious Diseases Society of America has provided guidelines for antifungal prophylaxis in the following high-risk groups: patients undergoing intensive chemotherapy for acute myeloid leukemia or myelodysplastic syndromes, allogeneic stem cell transplant recipients, patients with prior history of invasive aspergillosis, and patients with anticipated prolonged neutropenia.^[9] The National Comprehensive Cancer Network recommends antifungal prophylaxis for patients with increased risk of fungal infections during cancer therapy.^[10] Unfortunately, universal adoption of these antifungal prophylaxis recommendations is almost impossible because practices are usually guided by local epidemiology, availability of drugs, and other economic conditions seen in low- and middle-income countries (LMICs).^[11,12]

A good understanding of the risk factors associated with PFI is therefore necessary if health professionals must improve outcomes and efficiently allocate resources. We thus aimed to evaluate the current prevalence and predictors of PFI in patients with cancer in a high-income country to further extrapolate what predictors may serve as guide for prioritizing antifungal prophylaxis where resources may be scarce.

MATERIALS AND METHODS

Data from the 2018 National Inpatient Sample (NIS) were used for the retrospective analyses. The NIS is the largest publicly available all-payer inpatient care database in the United States (US), sponsored by the Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. The NIS is designed to produce regional and national estimates of inpatient utilization, access, charges, quality, and outcomes in the USA. As a stratified sample, the 2018 NIS contains 20% (over 7 million) of all discharges from non-federal acute care hospitals in the US for the year 2018. When weighted, it estimates more than 35 million hospitalizations nationally. The 2018 NIS sampling frame includes data from 48 state-wide data organizations (47 states plus the district of Columbia).^[13] Each discharge record in the NIS contains patient-level (demographics, diagnoses, procedures, and mortality) as well as hospital-level data. Further information is available at https://www.hcup-us.ahrq.gov.

RESULTS

During the study period, a total of 205,525 hospitalizations met the inclusion criteria for the specified hematological malignancies. This cohort included 118,300 (57.5%) males, 142,223 (69.2%) Caucasians, and 18,641 (9.1%) African-Americans. The mean (SD) age of the study participants was 59.2 (±17.8) years, with a range of 18–90 years. The most common malignancies seen were non-Hodgkin’s lymphoma in 91,253 (44.4%) and AML in 69,262 (33.7%) patients. Up to 1248 (0.6%) of patients in the cohort had more than 1 hematological malignancy. Chemotherapy was administered in 56,930 (27.7%) hospitalizations, and 19,936 (9.7%) of the study participants had received bone marrow transplantation (BMT). A total of 177,985 (86.6%) hospitalizations were in urban teaching hospitals.

The diagnosis of PFI was made in 1635 hospitalizations, resulting in a prevalence of 0.8%. Of the patients with PFI, aspergillosis was noted in 1315 (80.4%) of cases while candidiasis was noted in 195 (11.9%) cases. The distribution of PFI based on causative pathogens is shown in [Figure 1].

Close

Figure 1:

Distribution of pulmonary fungal infections in study cohort.

Export to PPT

AML was the primary hematological malignancy in 1050 (64.2 %) of those who had PFI. Clinical variables compared between patients with and without PFI are presented in [Table 1].

The PFI group, compared to the non-PFI group, was younger (mean age 56.8 vs. 59.3 years, P = 0.013), more likely to be non-Caucasian (38.1% vs. 30.8%, P = 0.023), more likely to have Charlson comorbidity index ≥3 (63.6% vs. 54.8%, P = 0.002), more likely to have a diagnosis of AML (64.2% vs. 33.5%, P = 0.000), had longer hospital stay (mean length of stay of 23 vs. 9 days, P = 0.000), and had higher mean total hospital charges ($422,972 vs. $124,976, P = 000). As shown in [Table 2], in-hospital mortality rate was higher for the PFI group versus the non-PFI group (16.5% and 5.5%, respectively, P = 0.000). Similarly, patients with PFI had higher rates of bacteremia, sepsis, and organ failure (respiratory and renal) compared to patients without PFI.

Univariable analysis revealed that PFI was associated with higher odds of acute respiratory failure requiring advanced respiratory support, bacteremia, sepsis, septic shock, and inhospital mortality [Table 3].

On multivariable analysis [Table 4], predictors associated with PFI were Hispanic or native American origin (OR = 1.71; 95% CI: 1.21–2.42), Charlson comorbidity index above ≥3 (OR = 1.52; 95% CI: 1.16–2.00), neutropenia (OR = 1.97; 95% CI: 1.58–2.46), malnutrition (OR = 2.30; 95% CI: 1.75–3.01), bone marrow transplant status (OR = 2.28; 95% CI:1.53–3.39), and a diagnosis of AML (OR = 3.42; 95% CI: 2.60–4.50). In a multivariable model adjusting for confounders, PFI was an independent predictor for inhospital mortality (OR = 2.33; 95%CI: 1.74–3.13, P = 0.000).

DISCUSSION

Among patients receiving cancer therapy, PFIs are an important cause of morbidity and mortality as shown in this current study. Even though PFIs constitute a significant proportion of invasive fungal infections, few studies have comprehensively described the burden of PFI in patients with hematologic malignancies. In this present study, we leveraged data from a large US inpatient database to evaluate prevalence, resource utilization, and outcomes of PFI. Our data show low prevalence, but significant morbidity and mortality from PFI in patients with acute leukemia and aggressive lymphomas. Our study also identified several risk factors for PFI and confirmed PFI as a driver for higher health-care resource use.

PFI prevalence rates reported in the literature vary with the population being studied, the climate, and the clinical practice patterns. In our study cohort, a PFI prevalence of 0.8% is plausible given that antifungal prophylaxis for high-risk patients is a common practice in the US.^[9,10] In other clinical contexts, where prophylaxis is not routine or conditions are more favorable for fungal growth, prevalence rates up to 20% are typical.^[16-20] When breakthrough infections are further factored into the prevalence, the burden is higher among high-risk patients. A study in Nigeria evaluated 216 participants with presumptive tuberculosis, where 62% were male while 10% were children <20 years and they reported a PFI prevalence of 73.6%, with aspergillosis being the predominant pulmonary fungal pathogen.^[21] Their reported risk factors for PFI were: presence of PTB, prolonged (>3 weeks) use of antibiotics, pet possession, and cigarette smoking. Although not assessed in the current study, studies done in LMIC should assess for these important additional risk factors in patients at risk of PFI.

In studies where patients admitted to the intensive care unit were the target study population, the rates of invasive aspergillosis ranged from 0.33% to as high as 19.0%.^[22-29] Higher rates of aspergillosis reported in lung malignancies can be explained by likely tumor infiltration, mucosal injury inflicted by radiotherapy, and immunosuppression from chemotherapy.^[30-33]

The causative fungal organisms and underlying malignancies found in our study are consistent with what has been reported.^[34-37] Aspergillosis was the most common type of PFI in our cohort as has been reported by others.^[16,34-37] In a few reports where Candida species were the prevalent cause of PFI, the population studied were either not predominantly cancer patients^[38] or had a low proportion of PFI.^[6] Other fungal infections reported in our study were Candidiasis, Mucormycosis, and Cryptococcus, and these are typical pathogens among immunosuppressed patients.^[34,39] In comparison to other hematological malignancies, AML poses a higher risk for invasive fungal infections.^[16,40-42]

Clinical variables found to increase the odds of PFI in our study were BMT status, neutropenia, and acute myeloid leukemia. These are likely due to the impairment of both cell line of defense against fungal infections which are the macrophages and neutrophils,^[30] with the degree of immune suppression associated with odds of PFI from our study. Furthermore, chemotherapy, as well as immunosuppressive agents given for BMT, can both induce neutropenia and cause pulmonary mucociliary impairment, predisposing these patients to fungal infections. The patients may also experience neutropenia from defective marrow function, thereby further heightening their risks of invasive fungal infections. Hispanic or native American origin, malnutrition, and multiple comorbidities were also associated with higher odds. Characteristics such as malnutrition, poorly managed multiple comorbidities, and origin from a minority population group have been found by many studies to negatively affect health and mortality.^[43-47] Neutropenia and malnutrition are known risk factors for fungal infections,^[12,43].Unfortunately, these characteristics reflect the health conditions of many cancer patients living in LMIC.

Interestingly, a prior history of asthma or chronic obstructive pulmonary disease was not a significant risk factor for PFI, an observation that has been reported by at least one other study.^[30]

As expected, patients with PFI were severely ill and had increased risks for bacteremia, sepsis, septic shock, renal, and respiratory failure. Consequently, these patients had longer hospital stay, hospital costs, and mortality. Although the burden of cancer is increasing globally, a significant portion of this burden is borne by LMICs.^[48-51]

While access to cancer care remains one of the greatest needs in LMIC, other drivers of worse cancer outcomes include poor health-seeking behavior of patients, high out-of-pocket medical costs, insufficient training and retention of oncology professionals, unavailability of novel cancer therapeutics, diagnostics, and infrastructure.^[52] Because cancer research is not a top priority for many governments of LMIC, very few patients receive cancer treatments within the context of well-designed clinical trials. For the oncology health professionals in LMIC, lack of access to up-to-date medical literature and cancer care guidelines may preclude the use of appropriate medications even when they are available. As such, cancer care in many parts of LMIC is largely driven by availability and affordability rather than standard guidelines.

Regardless of the peculiar circumstances in any LMIC, patients with hematological malignancies can be successfully treated, with careful planning. Cancer therapies should be administered within or in proximity to facilities that have the trained staff or are equipped to care for critically ill patients since such patients are prone to developing severe complications which may require intensive care as was noted in the current study. If available, antifungal prophylaxis should especially be offered to patients at high risk for fungal infections. Empiric antifungal therapy should include a mold-active agent pending microbiological confirmation of causative agent. Therapeutic drug monitoring and molecular typing of strains may become necessary if compliance, breakthrough infection or drug toxicity is of concern. All nutritional deficiencies and comorbidities should be properly addressed in patients undergoing cancer treatments. If cancer patients develop PFI, they will likely require multidisciplinary care from oncologists, infectious disease physicians, intensivists, and pulmonologists, among others. Provision of medications and cancer research ought to be high priority for policy-makers in LMIC.

There are some limitations to our study. Due to the retrospective design, we could not determine if PFIs were limited to the lungs or a part of disseminated fungal disease. Second, administrative databases like the NIS lack data on medication use, and so the use of antifungal prophylaxis by all study participants could not be firmly ascertained, and thus, it is uncertain that these cases were primary fungal infection or breakthrough infections. Finally, our study design could not distinguish between probable, possible, and proven fungal infections.

CONCLUSION

In the current era of increasing use of prophylactic antifungals in at-risk patients, our study shows that PFIs remain a major cause of morbidity and mortality in patients with acute leukemia and aggressive lymphomas. Where resources are scarce, patients with the identified high-risk characteristics should be given priority for antifungal prophylaxis, and surveillance with clinical, molecular, and drug monitoring measures when possible. In LMIC, patients with these hematological malignancies should ideally be managed at referral centers equipped for critically ill patients, due to associated medical complications.