Keywords
Hematological malignancies - infection - pulmonary infiltrate
Introduction
Pulmonary infections represent the most common and dreaded infectious complications
occurring during neutropenia phase of patients with cancer. The incidence of pneumonia
in high-risk patients (e.g., patients with acute leukemia or stem cell transplant
[SCT]) is 17%–24%.[1],[2] In spite of potent mold-active antifungal as well as broad-spectrum antibiotic therapy,
the clinical response is 60%–65%, whereas the infection-related fatality rate in these
patients may be as high as 38%.[2] Thus, pulmonary infections are associated with a significant prognostic deterioration
in patients with cancer. The Gram-positive organism is the predominant etiologies
isolated.[1] The attributed mortality is high, particularly in patients affected by an invasive
pulmonary mycosis. In patients with invasive pulmonary aspergillosis, the survival
rates had been decimal (10%) in patients undergoing SCT.[3] Early detection of pulmonary infiltrates has been demonstrated to improve the outcome
of (prompt) systemic antifungal treatment. Much effort has been made during the past
15 years to optimize diagnostic procedures to provide an early diagnosis of lung infections
in patients with malignancies.[4] We carried out the study to determine the etiology of pulmonary infiltrates in febrile
neutropenia episodes and correlate the radiological appearance of pulmonary infection
and microbiologically documented infection of febrile neutropenia.
Materials and Methods
This was a prospective study done at the Medical Oncology Ward, All India Institute
of Medical Sciences (AIIMS), New Delhi, during November 1, 2014, to June 30, 2016.
All febrile neutropenia episodes of hematological malignancy and hematopoietic SCT
patients admitted during the study period in the Medical Oncology Wards fulfilling
the inclusion and exclusion criteria were included. Patients with nonhematological
malignancy with febrile neutropenia with pulmonary infiltrates and with preexisting
pulmonary and cardiac dysfunction were excluded. The study approved by the Ethics
Committee of AIIMS, New Delhi. After obtaining written informed consent, eligible
patients were enrolled in the study and enrolment details related to personal and
demographic profile, underlying disease, and clinical presentation were noted. High-resolution
computerized tomography (HRCT) was done after enrolment. The imaging findings were
reviewed by an independent radiologist. Bronchoscopy was done and bronchoalveolar
lavage (BAL) fluid isolated for microbiological investigation on those patients with
pulmonary infiltrate on imaging. Revised Definitions of Invasive Fungal Disease from
the European Organization for Research and Treatment of Cancer/Invasive Fungal Infections
Cooperative Group and Mycoses Study Group (EORTC/MSG) Consensus Group criteria was
used.[5]
Results
A total of 55 patients were evaluated in the span of study duration. Half (51%) of
the reasons for ineligibility in our patients were hemodynamically instability and
low platelet count. Due to the risk of severe bleeding and unsustainability of bronchoscopy
procedure, most of our patients they rendered were ineligible. We had enrolled 27
patients after written consent. One patient failed at bronchoscopy procedure due to
respiratory distress. Of the 26 analyzable patients, majority of the patients belonged
to acute myeloid leukemia (32%) and non-Hodgkin's lymphoma (32%), acute lymphoid leukemia
(18%), Hodgkin's lymphoma (14%), and multiple myeloma (4%). We enrolled mostly adult
patients with febrile neutropenia (88%), the median age being 37 years. Sixty-eight
percent of our patients were male and the gender ratio was 2.14:1. The average duration
of illness was 9.8 days. The median duration of onset of febrile neutropenia to bronchoscopy
was 8.8 days. Fever was present in all patients. As most of our patients, the clinical
focus was the pulmonary infection, respiratory distress present in 69% of patients
at initial presentation. Another focus of infection was diarrhea (19%) and perianal
soft-tissue infection (7%). Other clinical signs and symptoms at presentation were
bleeding in 27% of patients; decreased urine output as defined by <1 ml/kg/h any time
during hospital stay was 38%. Baseline characteristics of all patients are given in
[Table 1].
Table 1
Baseline characteristics of 26 febrile neutropenic patients
|
Parameter
|
Observation (range)
|
|
Age (year)
|
37.5±19.2 (2-65)
|
|
Sex (male:female)
|
2.1:1
|
|
Presentation duration (days)
|
9.8±5.2 (4-24)
|
|
Fever duration (days)
|
8.8±4.4 (4-20)
|
|
Fever (%)
|
26/26 (100)
|
|
Tachypnea/respiratory distress (%)
|
18/26 (69)
|
|
Perianal/soft-tissue infection (%)
|
2/26 (7)
|
|
Loose stool/vomiting (%)
|
5/26 (19)
|
|
Decreased urine output (%)
|
10/26 (38)
|
|
Bleeding (%)
|
7/26 (27)
|
Laboratory characteristics
Patients were evaluated regarding the complete blood count, renal function test, and
hepatic function tests. All our patients were febrile neutropenia; 90% of our patients
were anemic (<8 g/dL) requiring transfusion. The median leukocyte count was 1 × 109/L (range 0.2–1.45 × 109/L. The median platelet count was 30 × 109/L (20–50 × 109/L) and 27% of all patients presented with bleeding at the presentation from any site.
As acute kidney injury (AKI) is the most common complication of sepsis, 10% of our
patients identified as AKI and three of our patients required hemodialysis. As the
acute leukemic patients generally nutritionally debilitated, the median albumin level
of our patients was 2.9 g/dL. Hepatic dysfunction due to sepsis or leukemia or drug-induced
presented in 10% patients.
Imaging evaluation
HRCT was carried out on the patients with febrile neutropenia with pulmonary infection.
The objective findings were lobar/lobular consolidation (50%), nodule (27%), ground
glass opacity (23%), and effusion (11.5%) All the radiologic images were reviewed
by independent radiologists. After radiologic review, 39% were classified as bacteria
and 31% of the population classified as a fungal infection. Thirteen per cent of the
patients were classified as mixed infection as per radiology.
Bronchoalveolar lavage quality assessment
BAL quality assessment was done using physical as well as microbiological techniques.
Chamberlin's criteria were used for quality assessment. Inadequate BAL sample was
found in 5% of all samples collected. The complications of BAL are as follows: The
procedure of bronchoscopy was well tolerated by the patients and the problems encountered
(tachycardia – 11, hypoxia – 4, and throat pain – 8) were mild and reversible. One
patient developed episodes of epistaxis postbronchoscopy, which was managed with conservative
treatments.
Serum and bronchoalveolar lavage galactomannan
Galactomannan was analyzed using enzyme-linked immunosorbent assay and 46% of our
patients diagnosed as EORTC-MSG probable/possible invasive aspergillosis group. Serum
galactomannan test came positive (optical density (OD) > 0.5) in 8/26 (30.7%) and
BAL galactomannan positive (OD >1) in 4/26 (15.3%). A total of 17 microbiological
isolates identified by BAL bacterial culture and 8 organisms isolated by blood culture.
BAL fungal stain and culture identified two Aspergillus flavus species and one Aspergillus terreus species. We also stained and cultured the entire specimen for TB GeneXpert and Ziehl–Neelsen
stain and Lowenstein–Jensen media culture but none of them. We did stain for Norcardia and Actinomycosis and anaerobic culture as a case-to-case basis. Cytomegalovirus polymerase chain reaction
(PCR) was done for two patients. PCR for Pneumocystis carinii came positive for one patient [Table 2]. We correlate between the radiologic findings with the microbiological determination.
Patients with fungal pneumonia the diagnostic etiology by radiology review a positive
correlation exists with serum or BAL galactomannan elevation (r = +0.46) and with documented fungal isolation (r = +0.37). Patients were followed up up to 30-day postadmission. More than half of
our patients recovered with antibiotic and inotrope support and 41% of our patients
died. For majority of cases, the cause of death was due to multiorgan failure. The
most common diagnosis at death was the refractory shock (80%), respiratory failure
(15%), and pulmonary hemorrhage (5%).
Table 2
Species distribution of microbiologically documented patients
|
Microbiological etiology isolated
|
n (percentage of isolated etiology)
|
|
Acinetobacter
|
2/17 (11.7)
|
|
Escherichia coli
|
2/17 (11.7)
|
|
Klebsiella pneumonia
|
2/17 (11.7)
|
|
Pseudomonas
|
2/17 (11.7)
|
|
Burkholderia
|
1/17 (6)
|
|
Mixed Gram-negative bacteria
|
4/17 (23
|
|
Gram-positive bacteria
|
0/17 (0)
|
|
Aspergillus flavus
|
2/17 (11.7)
|
|
Aspergillus terreus
|
1/17 (6)
|
|
Pneumocystis jirovecii
|
1/17 (6)
|
Discussion
Chemotherapy for hematological malignancies cause profound suppression of immune system
and adversely affects the gastrointestinal mucosal integrity, and hence, they are
at a heightened risk for invasive infection due to colonizing bacteria or fungi that
translocate across intestinal mucosal surfaces. Profound prolonged neutropenia is
most likely to occur in the preengraftment phase of allogeneic hematopoietic cell
transplantation, and patients undergoing induction chemotherapy for acute leukemia
are at risk of serious infection as compared to another chemotherapy schedule.
Pulmonary infection is the most common site of infection in patients with hematologic
malignancies with febrile neutropenia.[2],[6] We carried out the present study to evaluate the pulmonary infection in neutropenic
patients using BAL and compare with the conventional method of diagnosis, e.g., clinical,
radiology, blood, and BAL culture. The uniqueness of study population as it included
the selected group of febrile neutropenic patients with hematologic malignancy. Fever
in neutropenic patients is an emergency as it is almost always indicative of an underlying
infective etiology. However, localizing signs may be subtle or absent. Respiratory
signs were present in three-quarter of the patients in the present study, indicating
that signs and symptoms of respiratory infection are blunted in many of these patients,
as they are immunosuppressed. In contrast to the global trends of Gram-positive bacteria
as the predominant infecting agent in febrile neutropenic patients, most of our isolated
were Gram-negative organism 13/17 (76%). Gram-positive bacteria were not isolated
separately in our study. This may be partly due to the fact that effective antibiotics
active against Gram-positive bacteria were incorporated early as part of empirical
antibiotic therapy and the pattern of infection has a significant geographic variation.
Among the Gram-negative pathogens isolates, in 4/13 patients, extended-spectrum beta-lactamases
positive bacteria were isolated. This points out to the emergence of these types of
organisms as an important cause of infection in these patients. This also highlights
the importance of using antibiotics that are capable of neutralizing the β-lactamase,
produced by these organisms and incorporating them into upfront empirical antibiotic
combinations. Therefore, each institution must have an antibiotic policy based on
local prevalence and sensitivity patterns. In this aspect, BAL forms an important
diagnostic tool for these patients to optimize antimicrobial treatment. In 2 out of
17 patients in whom organism was isolated, multiple bacteria or fungus were grown
simultaneously. This may represent actual infection with more than one bacterium or
may be due to contaminants. One of the disadvantages of the fiber optic bronchoscopy
(FOB) is the possible contamination during its introduction through the nasal passage.
Therefore, it is important to be careful during the procedure of collection and transport
of the BAL specimen. Surveillance cultures from the upper airways and correlation
with the BAL isolates may be helpful in distinguishing between contamination from
the upper airway and true infection of the lower respiratory tract. Quantitative estimation
of the cultures also would be helpful in this regard.
Invasive mycoses are one of the major causes of mortality in patients with febrile
neutropenia. Patients undergoing treatment for hematologic malignancies have an estimated
cause-specific mortality due to invasive fungal infections (IFI) of 35%.[3],[7],[8] In an EORTC/MSG study, factors significantly associated with the development of
IFI/invasive mold infection are prolonged neutropenia, prolonged use of systemic corticosteroid,
T-cell suppressive therapy, and recent SCT (<6 months before episode).[9] Similarly, relative or absolute T-cell deficiency or suppression by drugs is the
major risk factor for the development of symptomatic pneumocystis infection.[10] Fungi could be grown in culture from three patients. However, none of the patients
stained positive for fungi from the direct KOH examination indicates the relative
insensitivity of KOH preparation. Twenty-two patients have received amphotericin-B
as part of empirical treatment before bronchoscopy, which might have been a reason
for KOH preparation being negative. The sensitivity of the organism to antifungals
drugs was not done in our study. Strong clinical or radiological evidence of fungal
infection should be treated with the full course of antifungal therapy, as BAL cultures
can be sterile for fungi, especially if the patient has already been started on antifungal
therapy empirically. In our study, one patient came positive for P. carinii. The radiology was corroborative of these findings. She recovered from the Pneumocystis pneumonia after starting of trimethoprim-based therapy.
In neutropenic patients, only about 10% of conventional chest radiographs show abnormalities,
whereas CT in these patients reveals pathological findings in 50% of cases.[11] For evaluation of pneumonia, HRCT with subsequent guided BAL has been recommended
as the most sensitive technique.[12] HRCT is the preferred sequence of choice due to its thin collimation, more sensitive
in detecting small nodule <5 mm, and no contrast media exposure. The relative sensitivity
and specificity of HRCT to detect bacterial infection were 84.78% and 93.84%, while
the for the fungal infection, it is 95.2% and 96.7%, respectively.[13] We correlate between the radiologic findings with the microbiological determination.
In patients with fungal pneumonia, the diagnostic etiology by radiologic review, a
positive correlation exists with serum or BAL galactomannan elevation (r = +0.46) and with documented fungal isolation (r = +0.37). The possible reasons for the lower strength of correlation were due to
the low yield of fungal culture and low sample size.
The median duration of days between onset of symptoms and bronchoscopic evaluation
was 9.8 days. This delay was due to the fact that most of our patients who developed
pneumonitis had a hemodynamically unstable and poor performance status for withstanding
the bronchoscopy procedure. Furthermore, we had to ensure a platelet count of at least
50,000/mm3 before bronchoscopy. Most of our patients (90%) received single donor platelet
before bronchoscopy; this was difficult to achieve as most of the patients had received
intensive cytotoxic therapy and subsequently were severely myelosuppressive. Therefore,
in achieving an adequate platelet count, the procedure was often delayed. A well-coordinated
approach and aggressive platelet support should help in decreasing the time interval
between onset of symptoms and BAL. Overall, the bronchoscopy procedure was well tolerated
with minor reversible complications. We had a diagnostic yield of 65% similar as compared
to the previously published study [Table 3].[14] However, the therapeutic utility or the change in the management was low as we usually
start our therapy empirically. The identification of organisms ensured that the complete
duration of antimicrobial therapy could be given to these patients as part of definitive
therapy. Therefore, the actual benefit to the patients may have been higher than indicated.
Furthermore, if these organisms are not eradicated fully, they could form part of
resistant colonization. These could be the potential source of further morbidity in
future chemotherapy.
Table 3
Comparison of previous study evaluating pulmonary infiltrate in febrile neutropenic
patients
|
Study
|
Gilbert et al.[14]
|
Present study
|
|
TB - Tuberculosis
|
|
Sample size
|
144
|
26
|
|
Population
|
Hematopoietic stem cell transplant population
|
Hematologic malignancy and stem cell transplant
|
|
Diagnostic yield
|
52.5%
|
65%
|
|
Etiology
|
Bacterial (31%), fungal (15%), alveolar hemorrhage (11%)
|
76% Gram-negative
|
|
17.6% fungal
|
|
0% Gram-positive/TB
|
|
Changes in therapy
|
59%
|
25%
|
|
Comments
|
11% noninfectious causes
|
Anaerobe: Negative TB: Negative
|
Conclusion
Evaluation of pulmonary infiltrates in febrile neutropenia patients helps in optimizing
antimicrobial therapy. Knowledge of prevalence of organisms and their sensitivity
patterns would also help in formulating effective empiric antibiotic protocols. This
would help in decreasing the morbidity and also in cutting the cost of antibiotic
treatment. Our finding also emphasizes the importance of appropriate empirical antimicrobial
therapy, which forms the mainstay of treatment for these patients. Our finding suggests
that BAL has got several potential advantages in evaluating patients with hematological
malignancies. However, it must be emphasized that the effectiveness of BAL depends
on several local factors. The presence of good microbiological and cytopathological
backup is crucial in analyzing the specimen. HRCT accurately correlates with the microbiological
isolation. In summary, our experience with FOB and BAL in patients with hematological
malignancies pulmonary infiltrates is encouraging. It has been a learning experience,
and the results obtained would be of benefit in planning further studies.
