| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 Departments of Radiology (A.N.L., M.V.G., C.H.N., S.G.B.)
2 Health Research and Policy (R.M.W.)
3 Divisions of Bone Marrow Transplantation (W.W.H.)
4 Pulmonary and Critical Care (J.G.), S-072A, Stanford University Medical Center, Stanford, CA 94305-5105.
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
MATERIALS AND METHODS: Fifty-nine episodes of pulmonary infection were diagnosed in 52 (7.2%) of a consecutive series of 725 adult bone marrow transplant recipients. Causative organisms, time of diagnoses, radiographic patterns, and mortality rates were reviewed.
RESULTS: Cytomegalovirus and Aspergillus species were the two most common pathogens, accounting for 22 and 17 episodes, respectively. During the first 30 days after bone marrow transplantation, fungi caused the majority (nine [82%] of 11 episodes) of pulmonary infections; from days 31 to 100, viruses predominated (21 [62%] of 34 episodes). Recipients of allogeneic transplants had a higher probability of developing Cytomegalovirus pneumonitis than did the recipients of autologous and syngeneic transplants (P < .001). Radiographic findings of Cytomegalovirus pneumonia consisted of parenchymal opacification (90%) and innumerable nodules smaller than 5 mm (29%); in two patients, radiographs were normal. Nodules, masses, or nodules and masses, present in nine (69%) of the 13 patients with Aspergillus infection, were the most common radiographic findings in invasive aspergillosis. Bone marrow transplant recipients with a documented pulmonary infection were found to have a lower event-free survival than recipients without infection (P < .001).
CONCLUSION: Opportunistic pathogens account for the majority of pulmonary infections requiring invasive diagnosis and tend to manifest at predictable times in the course of events following recovery from bone marrow transplantation. Cytomegalovirus, the most common pathogen, causes a spectrum of radiographic findings that includes normal findings. Occurrence of a pulmonary infection is associated with an increased mortality rate.
Index terms: Aspergillosis, 60.2056 • Bone marrow, transplantation, 40.455 • Cytomegalovirus, 60.2066 • Lung, infection, 40.458, 60.201, 60.202, 60.2042, 60.205, 60.206, 60.2075, 60.21
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
Despite the institution of routine prophylaxis for common pathogenic agents and empiric therapy of febrile episodes in the early neutropenic period, pulmonary infections remain an important source of morbidity and mortality in the bone marrow transplantation population (4,5). To date, radiologic studies of pulmonary infections after bone marrow transplantation have been limited (6–9). The aim of our study was to assess the clinical and radiographic findings of pulmonary infections diagnosed by using invasive means (eg, bronchoalveolar lavage, transbronchial biopsy, percutaneous biopsy, open biopsy, or autopsy) in a consecutive series of 725 adult recipients of bone marrow transplants treated at our institution during an 8-year period.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
Clinical and Radiographic Follow-up
Patients received their preparatory regimen and transplant and recovered in the hospital, where they were examined by the medical team twice a day. Chest radiographs were obtained on a weekly basis in asymptomatic patients and more often if clinically indicated.
After discharge from the hospital, autologous and syngeneic transplant recipients were examined on a weekly basis until 30 days after transplantation, every 2 weeks until 60 days, monthly until 180 days, every 3 months until 365 days, and then on an annual basis. Allogeneic transplant recipients were examined daily until 100 days after transplantation, monthly until 180 days, every 3 months until 365 days, and then on an annual basis. Chest radiographs were obtained in all of the patients at all follow-up visits, except for in the allogeneic transplant recipients during the first 100 days after discharge from the hospital, during which time radiographs were obtained on a weekly basis.
Examination of a Febrile Patient
All febrile transplant recipients underwent clinical evaluation consisting of performance of a history and physical examination; in addition, blood culture analyses and chest radiographs were routinely ordered. Sputum and urine culture analyses were performed if clinically indicated.
Invasive procedures were indicated in patients with undiagnosed pulmonary disease. Percutaneous biopsy with computed tomographic (CT) guidance was performed for diagnosis of peripheral nodules or masses; bronchoalveolar lavage with or without accompanying transbronchial biopsy was performed in all other patients. Open or thoracoscopic biopsy was reserved for transplant recipients in whom there was a progressive, undiagnosed pulmonary process.
Specimens from invasive procedures were routinely sent to the laboratory for Gram staining, potassium hydroxide wet mount preparation, acid-fast bacillus staining, Grocott-Gomori methenamine–silver nitrate staining, aerobic bacterial culture, fungal culture, mycobacterial culture, and viral culture. Polymerase chain reaction and shell vial culture analyses were also performed for the detection of Cytomegalovirus.
Diagnostic Criteria
The inclusion criterion for patient enrollment in the study was diagnosis of a pulmonary infection by using invasive means, that is, by using bronchoscopic methods such as bronchoalveolar lavage or transbronchial biopsy, percutaneous biopsy, open biopsy, or autopsy during the retrospectively reviewed period of July 1987 to June 1995. Follow-up for transplant recipients ranged from a minimum of 1.0 to 7.6 years. Infections were categorized as to the time of diagnosis after bone marrow transplantation: less than or equal to 30 days, 31–100 days, 101–365 days, and more than 1 year. Bacterial pneumonia was diagnosed on the basis of positive culture results from bronchoalveolar lavage fluid obtained from an area of a radiographic abnormality and corroborated with positive blood culture results yielding the same organism (10) or histologic evidence of inflammation in transbronchial biopsy specimens from the same site. Legionella, Nocardia, respiratory syncytial virus, and influenza were considered pathogens when cultured from bronchoalveolar lavage fluid (11,12). Diagnosis of Cytomegalovirus pneumonia required positive culture results from bronchoalveolar lavage fluid samples in addition to identification of the characteristic cytopathic feature of intranuclear inclusions seen in either bronchoalveolar lavage fluid or biopsy tissue (13). Pneumocystis carinii pneumonia was diagnosed when the cystic form of the organism was identified with Grocott-Gomori methenamine–silver nitrate staining of a bronchoalveolar lavage fluid sample or lung biopsy specimen (13). One case of Pseudallescheria boydii pneumonia was diagnosed with concurrent growth of the organism on bronchoalveolar lavage fluid specimens and blood culture. Diagnoses of Candida and Aspergillus infections were established on the basis of either identification of the organism or positive culture results from lung tissue specimens obtained by means of percutaneous biopsy, open biopsy, or autopsy (13).
Radiologic Studies and Interpretation
For each infectious episode, the radiographs obtained before and closest temporally to the diagnostic invasive procedure were assessed specifically for the presence and distribution of parenchymal opacification, nodules, masses, reticular opacities, and pleural effusions. Parenchymal abnormalities were categorized as involving upper, middle, or lower lung zones; severity of parenchymal opacification was graded by assessing the silhouette of pulmonary vessels against the surrounding parenchyma as follows: Grade 1 was minimal haziness of the margins of the pulmonary vessels, grade 2 was obliteration of less than or equal to 50% of vessel margins, and grade 3 was obliteration of more than 50% of the margins of the pulmonary vessels. Nodules or masses were defined as discrete opacities smaller than or equal to 30 mm or larger than 30 mm, respectively; both were classified according to predominant size (smaller than or equal to 5 mm, 6–10 mm, 11–30 mm, 31–50 mm, or larger than 50 mm) and number (one, two to four, five to 10, more than 10, innumerable). Note was also made of the presence of cavitation.
Evaluation of radiographic studies was performed by three radiologists (A.N.L., M.V.G., C.H.N.) who, although aware that the study involved pulmonary infections after bone marrow transplantation, were otherwise blinded to clinical information and were unaware of the number and types of infections; studies were reviewed at the same time by the three observers, with conclusions reached in consensus.
Statistical Analysis
The probability of developing pulmonary infections in the allogeneic and autologous or syngeneic transplant recipients was calculated by using the Kaplan-Meier method (14), with censoring of deaths during the follow-up period. Event-free survival curves were also generated by using the Kaplan-Meier method, with events in the bone marrow transplant recipients defined as either death or relapse of original disease. Comparisons of the probabilities of infection and survival between transplant groups were performed by using the log-rank test.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
|
|
Radiographs were available for analysis in 21 patients. In two patients, the radiographs were normal. The most common radiographic finding of Cytomegalovirus pneumonia was parenchymal opacification, which was present in 19 (90%) of the 21 patients; the severity of opacification was classified as grade 1 in five patients (Fig 2a), grade 2 in seven patients (Fig 2b), and grade 3 in seven patients (Fig 2c). Innumerable associated nodules smaller than or equal to 5 mm were present in six (32%) of the 19 patients with abnormal radiographs. Opacification and nodules were found in a bilateral distribution in 16 (84%) of the 19 patients, involved a mean of 3.9 lung zones ± 1.7 [SD], and never affected the upper lung zone without concomitant involvement of the middle and lower zones on the ipsilateral side (Fig 3). Effusions were present in five patients; they were small in four patients and moderately sized in one patient. Effusions were bilateral in two patients.
|
|
|
|
In the first case of respiratory syncytial viral pneumonia, the chest radiograph was normal in appearance. In the second case, 20–30 nodules of 6–10 mm were distributed in bilateral upper and middle lung zones, were associated with grade 3 parenchymal opacification in the left upper and middle lung zones, and were accompanied by small, bilateral pleural effusions (Fig 4). The single episode of influenza B pneumonia demonstrated 20–30 nodules of 6–10 mm distributed in bilateral middle and lower lung zones without associated findings.
|
Radiographs were available for analysis in 14 patients; one patient had documented simultaneous infection with Aspergillus and Pseudomonas. Of the 13 patients with isolated Aspergillus infection, nine (69%) had nodules, masses, or nodules and masses as the most common radiographic findings. The number of discrete opacities were one (n = 3), two to four (n = 2), five to 10 (n = 2), and more than 10 (n = 2); the predominant sizes of opacities were smaller than or equal to 5 mm (n = 1), six to 10 mm (n = 4), 11–30 mm (n = 3), and larger than 3 cm (n = 1). Grade 2 or 3 parenchymal opacification occurred in six (46%) of the 13 patients and was associated with nodules, masses, or nodules and masses in two patients. Cavitation was not identified in any patient. The parenchymal abnormalities were found in a bilateral distribution in eight (62%) of the 13 patients and involved a mean of 2.7 lung zones ± 1.9 (Fig 5). Isolated upper lung zonal involvement without concomitant involvement of ipsilateral middle and lower lung zones occurred in six (46%) patients. Small bilateral effusions and a unilateral pleural effusion were present in three patients and one patient, respectively.
|
Other Opportunistic Fungal Infections
Candida and P boydii each caused one episode of pneumonia in allogeneic transplant recipients at 66 and 16 days after transplantation, respectively. Grade 3 parenchymal opacification involved all lung zones in the case of candidal pneumonia and the right middle and bilateral lower lung zones in the case of P boydii pneumonia (Fig 6).
|
Radiographs were available for analysis in all of the 11 patients with solitary bacterial pneumonia. The most common radiographic finding was parenchymal opacification, which was present in 10 (91%) of the 11 patients; grades of opacification were 1 (n = 2), 2 (n = 3), and 3 (n = 5). In two patients with Pseudomonas pneumonia, 6–10-mm nodules were found in the same zones involved by parenchymal opacification. Parenchymal masses measuring 3.0 and 7.5 cm occurred in patients with Nocardia pneumonia (Fig 7) and Pseudomonas pneumonia, respectively. Parenchymal cavitation in association with bacterial pneumonia was observed only in the patient with Nocardia pneumonia. The parenchymal abnormalities demonstrated a unilateral distribution in eight (73%) of the 11 patients and involved a mean of 2.2 lung zones ± 1.8. In the three patients with bacterial pneumonia resulting in bilateral parenchymal disease, the pneumonia was caused by Pseudomonas (n = 2) (Fig 8) or Legionella (n = 1). Isolated upper lung zonal involvement without concomitant involvement of ipsilateral middle and lower lung zones occurred in three (27%) patients. Small, bilateral pleural effusions were present in both patients with Legionella pneumonia; small, unilateral effusions were associated with the cases of Haemophilus influenzae and Enterobacter pneumonia.
|
|
In one patient, the radiographic findings consisted of grade 3 parenchymal opacification diffusely involving all lung zones with an associated small right pleural effusion. In the other patient, the radiographic findings consisted of grade 1 parenchymal opacification in the right lower lung zone and six associated ill-defined 6–10-mm nodules scattered in the left upper and bilateral middle and lower lung zones.
Survival Statistics
Twenty-three (39%) of 59 infections were associated with patient death within 30 days of diagnosis; opportunistic fungal infections were the group most highly associated with 30-day mortality (12 [63%] of 19 cases). Patients with a documented pulmonary infection had a significantly lower event-free survival (median, 0.33 years) than patients without pulmonary infections (median, 2 years; P < .001) (Fig 9).
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
In the bone marrow transplantation population, infection occurs as a direct result of transplantation-induced immune suppression. Because of the predictable course of immunosuppression and recovery, time lines have been established that focus the differential diagnosis on pathogens most likely to occur at sequential points in the bone marrow transplantation process (2,4,18). In the preengraftment period (days 0–30), profound neutropenia and damaged mucosal membranes are the predominant defects in host defense, which predispose the patient to bacterial and fungal infections (2,18). In the postengraftment period (days 31–100), infections occur secondary to impairment of both cellular and humoral immunity, with Cytomegalovirus as the single most important pathogen (3,18). Patients with acute graft-versus-host disease remain susceptible to bacterial and fungal infections as a result of the immunosuppressive agents for treatment of the disease and the intrinsic damaging effects of the disease on immune recovery and mucosal barriers. In the late posttransplantation period (on and after day 101), infection in the absence of chronic graft-versus-host disease is uncommon (18). Return of normal immunologic function usually occurs by 12 months after transplantation; however, in patients with graft-versus-host disease, the defects may persist indefinitely (2,3).
In our series, the established time lines closely reflected the observed pathogens causing pulmonary infection during each period. In the preengraftment period, opportunistic fungi accounted for 82% of infections; in the postengraftment period, viruses accounted for 62% of infections; and in the late posttransplantation period, there was no predominant pathogen, but the largest group of infections were those caused by bacteria (43% [six of 14]). Only one case of bacterial pneumonia was observed from days 0 to 30; in this case, Pseudomonas aeruginosa was cultured as an organism coexisting with Aspergillus at autopsy. In our experience, radiographically evident bacterial pneumonia is rare in bone marrow transplant recipients in the preengraftment period; this is likely because of our use of empiric antibiotics in neutropenic patients.
Allogeneic transplant recipients were shown to have a significantly higher probability of pulmonary infection than autologous and syngeneic transplant recipients, with the greatest difference in the probability of infection occurring in the postengraftment period. Between days 30 and 100, one of the major risk factors for infection is the presence of acute graft-versus-host disease (2,19); autologous and syngeneic transplant recipients who are not at risk of graft-versus-host disease experience fewer infections after engraftment than do allogeneic transplant recipients.
Viruses were the most common group of pathogens in this series, accounting for 25 (42%) of the 59 episodes of pulmonary infection after bone marrow transplantation. Similar to findings in other series (15,17,20), we found Cytomegalovirus to be the single most frequent etiologic agent; it caused 37% of all documented cases of pneumonia in our study. Cytomegalovirus infection, defined by viral shedding, may result from primary infection or, more commonly, reactivation in the bone marrow transplant population (18). We observed a significantly higher probability of Cytomegalovirus pneumonitis in the allogeneic transplant recipients than in the autologous and syngeneic transplant recipients. While there is no documented difference in the incidence of Cytomegalovirus infection among transplant types, clinically apparent disease is more common in allogeneic transplant recipients, with the greater susceptibility likely attributable to the immunosuppressive effects of graft-versus-host disease and its prophylaxis and treatment (3,18). Other major risk factors predisposing patients to the development of Cytomegalovirus pneumonia are positive serologic status of the recipient or donor, advanced age, and use of total-body irradiation or multiagent conditioning regimens (5,18).
The radiologic findings of Cytomegalovirus pneumonia have been variably described as consisting of lobar consolidation (21), diffuse and focal parenchymal haziness (22,23), and diffuse micronodules (24). Normal findings, as observed in two patients with biopsy-proved Cytomegalovirus pneumonia in our series, have also been documented at both radiography (23) and CT (24). In agreement with findings of two prior studies (22,23) performed in cardiac and lung transplant recipients, we found parenchymal opacification to be the most common radiographic manifestation of Cytomegalovirus pneumonia; it was present in 90% of affected patients. Associated micronodules and a bilateral distribution of parenchymal findings occurred in 32% and 84% of patients with abnormal radiographs, respectively.
We observed a distinct predilection for the parenchymal findings of Cytomegalovirus pneumonia to be distributed in the middle and lower lung zones; isolated upper lung zonal involvement without concomitant involvement of the ipsilateral middle and lower lung zones never occurred in our series. Shreeniwas et al (23) and Kang et al (24) reported similar distributions of parenchymal abnormalities in 14 (93%) of 15 and nine (90%) of 10 cases of Cytomegalovirus pneumonia, respectively. As hematogenous dissemination of Cytomegalovirus occurs commonly in the bone marrow transplant population (25), we hypothesize that the observed basilar predominance of pulmonary infection is in part related to normal inequalities of pulmonary blood flow that result in greater perfusion to the lower lung zones.
In addition to herpesviruses such as Cytomegalovirus, community respiratory viruses can also play an important role in the cause of respiratory diseases in the bone marrow transplant population. On the basis of a prospective surveillance study by Whimbey et al (26), in which bone marrow transplant recipients who developed upper respiratory tract infections were followed up over two consecutive 6-month winter seasons, the three most common organisms cultured were respiratory syncytial virus (49%), influenza virus (18%), and picornaviruses (18%). Fifty-eight percent of the infections detected in the study by Whimbey et al (26) were complicated by pneumonia, with an associated mortality rate of 51%. Respiratory syncytial viral pneumonia is almost always preceded by symptoms of upper respiratory tract infection such as rhinorrhea, nasal congestion, sore throat, and cough (26,27). Respiratory syncytial viral infections occur with peak incidence in the winter and spring (27); our two cases of respiratory syncytial viral pneumonia occurred in February and April. The radiographic findings of respiratory syncytial viral pneumonia are nonspecific; in a study of eight infected bone marrow transplant recipients (27), two showed normal or unchanged radiographic findings, two had localized infiltrates, and four had diffuse interstitial infiltrates. The radiographic findings in our two cases consisted of normal findings and a bilateral, nodular pattern associated with focal areas of consolidation.
Opportunistic fungi constituted the second most common group of pathogens in this series, accounting for 19 (32%) of the 59 episodes. Aspergillus, the most common fungal organism (it was responsible for 17 [89%] of the 19 cases), had a significantly higher probability of causing infection in allogeneic than in autologous and syngeneic transplant recipients. Prolonged neutropenia and corticosteroid use are major risk factors for the development of invasive aspergillosis (3,28). In the late posttransplantation period, we observed three cases of pulmonary aspergillosis that involved two allogeneic transplant recipients who were receiving steroids and cyclosporine for the prevention and treatment of graft-versus-host disease and one autologous transplant recipient who had neutropenia from chemotherapy for relapsed non-Hodgkin lymphoma.
In our study, the radiographic findings of invasive aspergillosis consisted of nodules, masses, or nodules and masses (69%); parenchymal opacification (46%); or a combination of these findings (15%). In agreement with findings of previous studies by Kuhlman et al (29) and Haramati et al (30), we found that parenchymal abnormalities tended to be multiple rather than solitary. An upper lobe predominance of invasive aspergillosis has previously been documented in cardiac transplant recipients (22,30) and, in 46% of infected patients in our study, isolated upper lung zonal involvement occurred. Although we found a slight predominance of the pulmonary lesions due to Aspergillus in the right lung (20 [57%] of 35 affected lung zones), this right-sided predilection was not as striking as the 74% of lesions previously reported (30).
Two additional cases of opportunistic fungal pneumonia caused by Candida and P boydii were observed in our series; both were associated with patient death. Candida species are a frequent cause of fungemia in the bone marrow transplant recipient, with the portal of entry believed to be the gastrointestinal tract or indwelling catheters (5). Factors that predispose bone marrow transplant recipients to candidal infections include increased age (31,32), increased duration of granulocytopenia (31), acute graft-versus-host disease (32), and use of irradiation as part of the preparative regimen (31). The radiographic findings (33) of pulmonary candidiasis consist predominantly of air-space disease with a segmental, lobar, or bilateral diffuse distribution, as was observed in our case. At autopsy, the lungs of our affected patient demonstrated an organizing pneumonia associated with innumerable 5-mm hemorrhagic nodules, which are pathologic findings previously reported to be typical of hematogenously disseminated candidal infection (34).
P boydii is a ubiquitous soil fungus that has been reported to rarely cause pulmonary infection in the bone marrow transplant recipient (35,36). Similar to Aspergillus, this organism has a propensity for angioinvasion and appears as dichotomous branching hyphae at histologic examination (35,36). Differentiation between these two organisms, which requires culture, is important because the treatment differs; amphotericin B, the treatment of choice for aspergillosis, is often not effective against P boydii (35,36). The radiographic findings of P boydii pneumonia typically consist of focal or multifocal sites of consolidation that may be associated with air crescent formation (35,37).
Bacteremia is a common complication in the preengraftment period, occurring in 10%–59% of bone marrow transplant recipients (18,38). Gram-negative bacilli arising from the gastrointestinal tract or oral mucosa have been the predominant etiologic group (5,18); however, with the use of selective oral antimicrobial prophylaxis, as is practiced in our institution, an increasing number of gram-positive infections caused by Staphylococcus and Streptococcus have been recognized (5,18). Because of the early empiric use of broad-spectrum antibiotics during periods of fever, bacterial pneumonia is infrequently diagnosed during the preengraftment period (3).
After marrow engraftment, bacterial pneumonia is usually caused by gram-positive organisms associated with indwelling catheters or gram-negative bacilli in patients with graft-versus-host disease (5,18). Persistent deficits in cellular immunity predispose patients to intracellular pathogens such as Listeria monocytogenes and to encapsulated organisms such as Streptococcus pneumoniae; persistent deficits in humoral immunity predispose patients to intracellular pathogens such as Nocardia species and to encapsulated organisms such as H influenzae (3,38). Each of these four bacteria were responsible for at least one infection in our series. Of the seven bacterial pneumonias diagnosed after day 30 in our allogeneic transplant group, 86% were caused by gram-negative organisms, with Pseudomonas responsible for 57% of cases.
The radiographic findings of bacterial pneumonia are believed to be identical in immunocompetent and immunocompromised patients and manifest most frequently as localized areas of consolidation (39,40). In a study by Lossos et al (41) in which the radiographic findings of 52 episodes of bacterial pneumonia in bone marrow transplant recipients were reviewed, infiltrates involved more than one lobe or segment in only 12 cases (23%). In agreement, we found that parenchymal opacification, the most common radiographic manifestation—it was present in 91% of our 11 cases of bacterial pneumonia—was relatively localized; it involved a mean of 2.2 lung zones and was unilateral in distribution in 73% of cases. The three cases of bilateral parenchymal involvement were caused by Pseudomonas and Legionella, organisms that have previously been documented to cause more diffuse patterns of disease (42,43).
The incidence of P carinii pneumonia in bone marrow transplant recipients has decreased sharply with the introduction of routine prophylaxis with trimethoprim-sulfamethoxazole (5,18). Both of the two patients with P carinii pneumonia in this series were no longer receiving prophylaxis at the time of the infection. The radiographic findings of P carinii pneumonia in immunocompromised patients without the acquired immunodeficiency syndrome, or AIDS, most commonly consist of a diffuse, bilateral reticular or granular pattern that can progress to alveolar consolidation (44). The first of our observed cases demonstrated this typical pattern of parenchymal involvement. The second case demonstrated multiple small nodules associated with a focal area of parenchymal opacification; these are atypical features that have previously been reported to occur in up to one-half of patients (45).
In our series, we found that bone marrow transplant recipients with a documented pulmonary infection had a significantly lower event-free survival than recipients without pulmonary infection. Fungi were the pathogenic group most highly associated with mortality within 30 days of diagnosis. Mortality due to pulmonary aspergillosis in bone marrow transplant recipients is high and has been reported to occur in up to 85% of cases (3,28); the high mortality rate associated with this infection is thought to be related to difficulties in achieving an early diagnosis (13,28) and to the limited efficacy of antifungal treatments (3,5). Seven (28%) of 25 cases of viral pneumonia, all caused by Cytomegalovirus, were associated with mortality within 30 days of diagnosis. Early diagnosis of Cytomegalovirus pneumonia is believed to improve survival (5,19); combination therapy with ganciclovir and the intravenous administration of immune globulin has resulted in survival rates of 30%–75% in bone marrow transplant recipients (19).
A limitation of our study was underestimation of the true prevalence of pulmonary infections in our patient population. As we specifically selected bone marrow transplant recipients who required an invasive procedure for diagnosis, cases of pneumonia treated empirically or diagnosed by using sputum culture alone were not included in the study. To ensure the validity of our radiographic descriptions, strict diagnostic criteria that allowed differentiation between colonization and infection by detected organisms were applied, which again resulted in the exclusion of some infectious episodes. Analysis of radiographs acquired closest to the time of diagnosis probably biased our results so that they reflect later stages of infection; however, it is particularly at this point in the diagnostic work-up in bone marrow transplant recipients that the radiologist's opinion will be actively sought. It should be recognized that, although our study has focused solely on infectious pulmonary complications after bone marrow transplantation, differential diagnosis for any bone marrow transplant recipient with respiratory symptoms or radiographic abnormalities must also include noninfectious diseases such as pulmonary edema, drug toxic effects, pulmonary hemorrhage, and relapse of original disease.
In conclusion, in bone marrow transplant recipients, opportunistic pathogens account for the majority of pulmonary infections requiring invasive diagnosis and tend to occur at predictable times in the bone marrow transplantation course. Allogeneic transplant recipients have a higher probability of developing pneumonia than autologous and syngeneic transplant recipients. Cytomegalovirus and Aspergillus are the two most frequent organisms, and both are associated with a high mortality rate. Although some trends in the radiographic patterns of pneumonia caused by different etiologic groups were observed, radiographic findings are variable and often nonspecific. Recognition of these clinical and radiographic features may aid in the diagnosis and treatment of pulmonary infections in bone marrow transplant recipients.
|
APPENDIX |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
Fractionated total-body irradiation was administered in doses of 1,200–1,320 cGy. Patients who underwent autografting received chemotherapy-only preparatory regimens if they had had prior exposure to radiation therapy (which precluded a radiation-containing preparatory regimen), if posttransplantation consolidative radiation therapy to sites of bulky disease was planned, or if fractionated total-body irradiation was considered to be too morbid, given the patient age or performance status. Patients who underwent allografting for severe aplastic anemia or myelodysplastic syndrome received chemotherapy-only regimens. Patients with hematopoietic malignancies received radiation therapy and chemotherapy prior to allografting unless they had had prior radiation exposure or were age 50 years or older or unless fractionated total-body irradiation was considered to be too morbid, given the patient performance status. All patients who received matched unrelated bone marrow or partially matched related bone marrow transplants underwent a fractionated total-body irradiation–based preparatory regimen.
Bone marrow harvest of allogeneic donors and autograft recipients was performed under anesthesia, as previously described (46). Autologous bone marrow was used as the sole source of hematopoietic reconstitution prior to May 1990. Starting in May 1990, autologous peripheral blood progenitor cells were collected from patients with breast cancer and multiple myeloma prior to high-dose therapy. Patients with Hodgkin disease underwent peripheral blood progenitor cell collection starting in December 1991. For each disease, peripheral blood progenitor cells were used as the sole source of hematopoietic reconstitution after several months of using both autologous marrow and peripheral blood progenitor cells for autografting. Autograft recipients with a diagnosis of non-Hodgkin lymphoma or chronic lymphocytic leukemia underwent bone marrow collection, which was purged with a panel of anti–B cell antibodies or anti–T cell antibodies and complement. Peripheral blood progenitor cells were obtained by means of apheresis following treatment with cyclophosphamide (4 g/m2) and granulocyte colony-stimulating factor. The product of the 1 day's apheresis collection, enriched by the collection of low-density cells after the application to a Percoll gradient and purged with antibodies and complement, was used as the autologous graft. Patients with acute nonlymphocytic leukemia or biphenotypic leukemia underwent apheresis of autologous peripheral blood progenitor cells starting in August 1995, whereas bone marrow purged with chemotherapy or antibodies was used as the autograft prior to that date.
Immunosuppressive Regimens and Infection Prophylaxis
All allogeneic transplant recipients received immunosuppressive therapy to prevent graft-versus-host disease. Graft-versus-host disease prophylaxis was determined by the preparatory regimen. Patients who received fractionated total-body irradiation and etoposide were randomly assigned to receive either a two- or three-drug graft-versus-host disease regimen. The two-drug regimen utilized cyclosporin A (Sandimmune [Sandoz Pharmaceuticals, Hanover, NJ]; 5 mg/kg/d load for 5 days then 3 mg/kg/d) starting 2 days before marrow transplantation and methotrexate (15 mg/m2) on day 1, with the dosage decreased (10 mg/m2) on days 3, 6, and 11. The three-drug regimen utilized the same doses of cyclosporin A beginning on 2 days before bone marrow transplantation, the same doses of methotrexate on days 1, 3, and 6, and methylprednisolone (0.25 mg/kg administered intravenously twice a day) starting on day 7. When patients were able to tolerate oral medication, they were converted to oral dosing of cyclosporin A (5 mg/kg administered twice a day) and prednisone (0.5 mg/kg orally twice a day); this regimen continued on a fixed tapering schedule until 6–12 months after transplantation, although some allograft recipients continued to require immunosuppression for longer periods. Patients with severe aplastic anemia and patients who received matched unrelated or partially matched related bone marrow transplants received cyclosporine A and methotrexate, as described. Patients who received busulfan-containing regimens or fractionated total-body irradiation, etoposide, and cyclophosphamide received cyclosporin A and prednisone for graft-versus-host disease prophylaxis.
To reduce the incidence of gram-negative septicemia during marrow aplasia, oral neomycin sulfate (1 g three times a day) and vancomycin (250 mg three times a day) were administered. P carinii prophylaxis consisted of intravenous administration of trimethoprim-sulfamethoxazole (160–800 mg) twice a day or one inhalation of pentamidine (300 mg) via a nebulizer during the preparatory regimen until 2 days before transplantation. Subsequently, one dose of inhaled pentamidine or oral trimethoprim-sulfamethoxazole (160–800 mg) was started on days 30–42 at twice a day for 2 days per week for 1 month for the autotransplant recipients. Allogeneic transplant recipients received inhaled pentamidine (300 mg) via a nebulizer every 4 weeks until immunosuppression was discontinued. For fungal prophylaxis, all patients received low-dose amphotericin B at 0.15 mg/kg/d starting on day 1; allogeneic transplant recipients also received nebulized amphotericin B (20 mg) daily.
All allogeneic transplant recipients with Cytomegalovirus seropositivity or donor seropositivity underwent bronchoalveolar lavage for Cytomegalovirus shell vial culture on posttransplantation day 35. Beginning in August 1993, bronchoalveolar lavage fluid was also analyzed by means of polymerase chain reaction for the detection of Cytomegalovirus transcripts. Biweekly blood samples from day 21 to 56, then weekly blood samples from day 57 to 100 were obtained for Cytomegalovirus detection by means of polymerase chain reaction, shell vial culture, and viral culture. If patients had positive results for Cytomegalovirus but did not demonstrate active Cytomegalovirus disease, they received preemptive therapy with ganciclovir induction at 5 mg/kg every 12 hours for 14 days, then once a day for 5 d/wk until day 100, in addition to intravenous immune globulin at 500 mg/kg every 14 days. If patients demonstrated clinical Cytomegalovirus disease in the form of pneumonitis or enteritis, they were treated with ganciclovir induction for 14 days then for 5 days per week until day 120. In addition, intravenous immune globulin was administered daily for 4 days, then every other day for 12 doses, then weekly until day 120.
Antibiotic Coverage and Work-up during Neutropenia
At an absolute neutrophil count of less than or equal to 500/µL, all transplant recipients received 1 g of vancomycin every 12 hours. At the first fever spike (at a temperature of more than 38.5°C) or with a persistent low-grade fever (temperature of 38.0°–38.5°C for 18–24 hours), two sets of blood culture results were obtained and the patient started to receive 2 g of ceftazidime every 8 hours. For continued neutropenic fever, usually an aminoglycoside was also administered or, instead ceftazidime was changed to piperacillin-tazobactam (3.375–4.5 g every 6 hours), imipenem (500 g every 6 hours), or aztreonam (2 g every 8 hours) administered intravenously. Repeat blood culture results were not obtained until 48–72 hours from the time of prior culture. For continued fever, the dosage of amphotericin B was increased from prophylactic levels.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, A.N.L.; study concepts, A.N.L., M.V.G., W.W.H., R.M.W.; study design, A.N.L., R.M.W.; definition of intellectual content, A.N.L., W.W.H., R.M.W.; literature research, A.N.L.; clinical studies, A.N.L., M.V.G., C.H.N.; data acquisition, A.N.L., M.V.G., C.H.N., S.G.B., J.G.; data analysis, A.N.L., M.V.G., C.H.N., S.G.B.; statistical analysis, A.N.L., R.M.W.; manuscript preparation, A.N.L.; manuscript editing and review, A.N.L., M.V.G., C.H.N., S.G.B., W.W.H., R.M.W., J.G.
Received May 29, 1998;
revision requested July 16, 1998; revision received July 29, 1998;
accepted October 19, 1998.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. Franquet, S. Rodriguez, R. Martino, A. Gimenez, T. Salinas, and A. Hidalgo Thin-Section CT Findings in Hematopoietic Stem Cell Transplantation Recipients with Respiratory Virus Pneumonia Am. J. Roentgenol., October 1, 2006; 187(4): 1085 - 1090. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. L. Escuissato, E. L. Gasparetto, E. Marchiori, G. de Melo Rocha, C. Inoue, R. Pasquini, and N. L. Muller Pulmonary Infections After Bone Marrow Transplantation: High-Resolution CT Findings in 111 Patients Am. J. Roentgenol., September 1, 2005; 185(3): 608 - 615. [Abstract] [Full Text] [PDF]
|
|
|
