Abstract
We closely followed the pulmonary function of 150 consecutive high-risk breast cancer patients who underwent standard induction CAF (cyclophosphamide, doxorubicin, 5-fluorouracil) chemotherapy, followed by randomization to either standard-dose CPB (cyclophosphamide, cisplatin, bischloroethylnitrosourea [BCNU]) chemotherapy (SDC) or to high-dose CPB chemotherapy (HDC) with autologous bone marrow transplantation (ABMT) and peripheral blood progenitor cell support (PBPCS). Previously, we have described a delayed pulmonary toxicity syndrome (DPTS) which characterizes the pulmonary dysfunction after HDC and ABMT in this patient population. However, little is known concerning the role induction chemotherapy plays in its development. We found that after three cycles of induction CAF, the mean diffusing capacity of the lungs for carbon monoxide (Dl CO) significantly decreased by 12.6%. Additionally, in patients receiving HDC, the mean Dl CO further decreased to a nadir of 55.2 ± 14.1% which was significantly lower than those receiving SDC (nadir: 80.7 ± 12.3%). DPTS occurred in 72% of patients receiving HDC as compared with only 4% of patients receiving SDC. All individuals diagnosed with DPTS were treated with prednisone and the 2-yr follow-up of pulmonary function revealed a gradual improvement in mean Dl CO such that there were no differences between HDC and SDC groups at the end of the study. No mortality was attributable to pulmonary toxicity in either group. After induction chemotherapy, but before HDC, bronchoalveolar lavage (BAL) demonstrated significant elevations in interleukin-6 (IL-6), IL-8, neutrophils, and lymphocytes. We conclude that induction CAF produces asymptomatic pulmonary dysfunction and inflammation which may prime the lungs for further injury by HDC and predispose to the development of DPTS. Fortunately, in this specific ABMT patient population, the early and judicious use of prednisone appears to improve pulmonary function in patients who develop DPTS. Bhalla KS, Wilczynski SW, Abushamaa AM, Petros WP, McDonald CS, Loftis JS, Chao NJ, Vredenburgh JJ, Folz RJ. Pulmonary toxicity of induction chemotherapy prior to standard or high-dose chemotherapy with autologous hematopoietic support.
High-dose chemotherapy (HDC) with autologous bone marrow transplantation (ABMT) and/or peripheral blood progenitor cell support (PBPCS) is becoming a well-established method for treating high-risk primary breast cancer with extensive lymph node involvement (1-3). Most protocols call for induction chemotherapy, followed by alkylating agent-based consolidative chemotherapy and usually include external beam radiation therapy (RT) to the chest wall. Autologous bone marrow and/or peripheral blood progenitor cells (PBPC), along with granulocyte-colony stimulating factor (G-CSF) or granulocyte monocyte-colony stimulating factor (GM-CSF), are used for cellular support after consolidative chemotherapy. Several randomized trials are currently in progress to assess the effect on overall survival and disease-free survival in patients receiving such therapy (1, 4-6).
One of the most common nonhematological complications of these dose-intensive therapies, particularly those with bischloroethylnitrosourea (BCNU), is pulmonary toxicity, with a reported incidence of 39% to 64% (7-10). This usually presents as a dry cough, dyspnea with or without fever and hypoxemia, occurring weeks to months after consolidative chemotherapy, and, under some series, appears to respond well to corticosteroids. When recognized and treated early, this form of drug-induced interstitial pneumonitis has minimal mortality (7, 9, 10). Previously we have described the clinical features, diagnosis, and treatment of 45 patients with this form of interstitial pneumonitis seen after HDC with ABMT, and proposed the term “delayed pulmonary toxicity syndrome” (DPTS) to distinguish it from the idiopathic pneumonia syndrome (IPS), which has a higher mortality and earlier presentation (10).
In the current study, we closely followed the pulmonary function of a large group of high-risk breast cancer patients. All patients received 4 cycles of standard CAF (cyclophosphamide, doxorubicin, 5-fluorouracil) induction chemotherapy and were subsequently randomized to consolidative standard-dose CPB (cyclophosphamide, cisplatin, BCNU) chemotherapy (SDC) or to consolidative high-dose CPB chemotherapy with hematopoietic stem cell support. Our aim was, first, to characterize the effect of induction chemotherapy on pulmonary function and delineate bronchoalveolar lavage (BAL) cellular and biochemical changes. Second, we aimed to evaluate our current methods of diagnosing and treating DPTS.
Based on this analysis, we found standard CAF induction chemotherapy to have an adverse effect on lung function, and furthermore this regimen induced an inflammatory cellular response. Although all patients remained asymptomatic after induction CAF, significant changes were observed in their diffusing capacity of the lungs for carbon monoxide (Dl CO) and BAL parameters, indicating pulmonary toxicity even before they received HDC or SDC. In addition, we found a high incidence of DPTS in patients who received high-dose CPB chemotherapy but not in those randomized to standard-dose CPB chemotherapy. No mortality was attributable to DPTS in either group of patients and prednisone produced a dramatic improvement in both symptoms and lung function.
We retrospectively reviewed 150 consecutive patients enrolled at Duke University Medical Center, in a randomized trial of HDC with ABMT and PBPCS versus SDC for primary breast cancer with the involvement of 10 or more axillary lymph nodes. The study was approved by the Duke University institutional review board and written informed consent was obtained from each of the patients. Eligibility criteria for this protocol included: operable, histologically confirmed Stage IIA/IIB or IIIA primary, female breast adenocarcinoma with 10 or more positive axillary lymph nodes; less than 8 wk from last surgery; Karnofsky status 80 to 100%; no prior chemotherapy or radiation therapy; no other malignancy or comorbid disease; pulmonary function tests (PFT) with FVC, FEV1, and Dl CO greater than 60% predicted value; and negative bilateral bone marrow biopsies. In addition, we recruited five healthy, nonsmoking, adult females to serve as control subjects for bronchoscopy and BAL studies. After obtaining written informed consent, a detailed history, physical examination, and complete PFT were performed on each volunteer. Bronchoscopy and BAL were then performed by methods described subsequently.
All breast cancer patients received four, 28-day cycles of CAF (cyclophosphamide [600 mg/m2, intravenously, Day 1], doxorubicin [Adriamycin, 60 mg/m2, intravenously, Day 1], and 5-fluorouracil [600 mg/m2, intravenously, Days 1 and 8]), as induction chemotherapy. Any patient with absolute neutrophil counts (ANC) below 1,000 cells per mm3, before or during any cycle of induction CAF, received G-CSF (5 μg/kg/d, subcutaneously) as hematopoieitic support. At the end of the third cycle of induction CAF, patients were randomized to HDC or SDC arms of the protocol, if they had no disease progression and normal cardiac and pulmonary function. The patients randomized to the HDC arm received G-CSF (5 μg/kg/d, subcutaneously) for 8 d to prime their progenator cells for leukapheresis. Bone marrow was harvested after either the third or fourth cycles of induction CAF. They went on to receive HDC with cyclophosphamide (1,875 mg/m2, intravenously daily on Days −6, −5, −4), cisplatin (55 mg/m2, intravenously, continuously on Days −6, −5, −4), and BCNU (600 mg/m2, intravenously as a one-time infusion on Day −3). Reinfusion of ABMT was done on Day +1 and PBPC infusion on Days −1, 0, and +1. The patients randomized to the SDC arm received cyclophosphamide (300 mg/m2, intravenously, daily on Days −6, −5, and −4), cisplatin (30 mg/m2, intravenously, continuously on Days −6, −5, and −4), and BCNU (90 mg/m2, intravenously as a one-time infusion on Day −3). Patients on either arm received G-CSF if their ANC fell below 1,000 cells per mm3. Treatment protocol entailed standard local-regional radiotherapy in both arms, to be delivered to the chest wall and ipsilateral supraclavicular and internal mammary nodes at a minimum of 50.4 Gy, followed by a 10-Gy boost to the mastectomy scar. This was begun generally 6 wk after consolidative chemotherapy, but sometimes was delayed owing to cytopenias or pulmonary toxicity. Patients having disease progression after consolidative chemotherapy or pulmonary complications other than DPTS were included in the study (n = 22). However, patients who received additional chemotherapy or radiation therapy for disease progression/recurrence after consolidative chemotherapy, were excluded from the study.
Baseline PFT including FVC, FEV1, and Dl CO were performed prior to registration for the protocol. The PFT were repeated after completion of 3 cycles of induction CAF, which coincided with the time of randomization. Follow-up PFT were scheduled 6 wk after consolidative chemotherapy and then at regular intervals of 6 wk for the first 24 wk and then every 12 wk thereafter. Additional PFT were done if signs and/or symptoms of pulmonary toxicity developed. All PFT were performed according to the standards mandated by the American Thoracic Society (11, 12). Dl CO was corrected for hemoglobin using the formula: (1.7 × Hgb)/(9.38 + Hgb) (11). FVC, FEV1, and Dl CO were expressed as percentage predicted for the corresponding age, gender, and height of the patient (13, 14).
The clinicians in our Bone Marrow Transplant Clinic used the following guidelines for diagnosing pulmonary toxicity and initiating prednisone therapy: (a) no evidence/suspicion for infectious etiologies; (b) development of nonproductive cough, dyspnea with or without fever occurring several weeks to months after consolidative chemotherapy, and a decrease in Dl CO to less than 60% predicted; (c) decline in Dl CO to less than 50% predicted with or without symptoms; (d) for asymptomatic patients with Dl CO 50 to 60% predicted, treatment was left to the discretion of the treating physician.
Patients receiving prednisone were started at an initial dose of 60 mg/d for 14 d, then tapered to 50 mg/d for 5 d, 40 mg/d for 5 d, 30 mg/d for 5 d, 20 mg/d for 5 d, 10 mg/d for 5 d, 5 mg/d for 5 d, 2.5 mg/d for 5 d and then 2.5 mg every other day for 5 doses. Patients treated with prednisone also received concurrent pneumocystis prophylaxis with oral sulfamethoxazole–trimethoprim taken as 1 double-strength tablet every Monday, Wednesday, and Friday.
Fiberoptic bronchoscopy was performed on nine consecutive patients who were randomized to HDC, after they had received induction CAF, but prior to HDC, as well as in five healthy volunteers. All bronchoscopies were performed via the oral route using conscious intravenous sedation (meperidine, promethazine, midazolam) and topical anesthesia (2% lidocaine). After wedging the bronchoscope into both right middle lobe and lingula, serial lavage was performed with 60-ml aliquots of 0.9% sterile saline (Normosol; Abbott Laboratories, Abbott Park, IL) (total of 240 ml lavage per lobe), instilled through the bronchoscope and aspirated gently with a hand-held syringe. The bronchoalveolar lavage fluid (BALF) was filtered through sterile gauze, placed on ice, and processed immediately.
Cytocentrifuge (Cytospin3; Shandon Inc., Pittsburgh, PA) preparations were made from 200 μl of BALF, for differential cell counts using a Leukostat staining kit (Fisher Scientific, Pittsburgh, PA) and counting at least 300 cells. Aliquots of BALF were centrifuged at 500 g for 20 min at 4° C to obtain cell-free supernatant BALF. Cell-free supernatant BALF was used to determine cytokine, protein, and glutathione concentrations. Analysis of proinflammatory cytokines, including interleukin-6 (IL-6), IL-8, tumor necrosis factor-alpha (TNF-α), and soluble receptors to TNF (TNF-R1, TNF-R2), was performed at the Cytokine Core Laboratory at the University of Maryland, Baltimore, by sandwich-type ELISA. The sensitivities of these tests were IL-6 = 1.562 pg/ml, IL-8 = 1.953 pg/ml, TNF-α = 15.625 pg/ml, TNF-R1 = 3.9 pg/ml, and TNF-R2 = 15.625 pg/ml. Protein levels in BALF were determined using the Coomassie Plus assay kit (Coomassie Plus Protein Assay; Pierce Chemical Company, Rockford, IL). Total and oxidized glutathione concentrations were assayed by high-performance liquid chromatography (HPLC) (Hewlett-Packard 1100; Hewlett-Packard, Wilmington, DE) by methods previously described (15).
Paired t tests were performed to compare the change in Dl CO after induction CAF. Comparisons between the two arms of the protocol were performed using two-way analysis of variance. Linear regression analysis was performed to evaluate correlation between parameters. Nonparametric statistical analysis was used to compare the results of BAL. These results are expressed as median and ranges, and comparisons were made using the Mann-Whitney rank sum test. Statistical analysis was performed on an IBM-compatible personal computer using SigmaStat statistical software (Jandel Corporation, San Rafael, CA). A p value of 0.05 was considered to be statistically significant.
Demographic and lung function data are shown in Table 1. Patients randomized to the HDC and SDC arms of the protocol were similar with regard to age, race, and smoking status. The mean age at the time of consolidative chemotherapy was 44.7 ± 8.25 and 44.5 ± 7.55 yr for the HDC and SDC arms, respectively. The majority of patients in both groups were Caucasian. Only 15 patients in the HDC arm and six in the SDC arm were smoking at the time of protocol entry. Patients with unknown smoking history were not evaluated at protocol entry for tobacco use. Baseline lung function values of both groups recorded before protocol entry were not statistically different.
| HDC (n = 75) | SDC (n = 75) | |||
|---|---|---|---|---|
| Age, yr | 44.7 (8.25) | 44.5 (7.55) | ||
| Race | ||||
| Caucasian | 64 (85.3%) | 65 (86.7%) | ||
| African-American | 10 (13.3%) | 7 (9.3%) | ||
| Hispanic | 1 (1.3%) | 1 (1.3%) | ||
| Oriental | 0 | 2 (2.7%) | ||
| Smoking status | ||||
| Current, pack-years | 15 | 6 | ||
| 22 (5–50)† | 38 (5–60)† | |||
| Ever, pack-years | 16 | 19 | ||
| 9 (2–50)† | 10 (2–40)† | |||
| Never | 37 | 45 | ||
| Unknown | 7 | 5 | ||
| Follow-up, wk | 112 (27–331)† | 84 (28–332)† | ||
| Baseline PFT | ||||
| FEV1, L | 2.78 (0.51) | 2.78 (0.45) | ||
| % predicted | 94.5% | 94.8% | ||
| FVC, L | 3.43 (0.60) | 3.56 (0.50) | ||
| % predicted | 96.8% | 99.7% | ||
| Dl CO, ml/min/mm Hg | 21.41 (3.79) | 21.48 (3.11) | ||
| % predicted | 101.9% | 101.7% |
Five healthy, nonsmoking, adult females were recruited to serve as control subjects for bronchoscopy and BAL. Their mean age was 30.1 ± 9.5 yr at the time of bronchoscopy. Four volunteers were Caucasian and one was African-American. PFT performed on them demonstrated normal function with mean FEV1 = 93.4 ± 5.76% predicted, mean FVC = 97.5 ± 4.7% predicted, and mean Dl CO = 108.8 ± 8.56% predicted. Detailed history and physical examination on all volunteers revealed no significant abnormalities.
Effects of CAF induction chemotherapy. We compared the Dl CO of 150 patients before and after they received induction CAF. Mean Dl CO before induction CAF was 101.8% predicted and this fell to 91.3% predicted after 3 cycles of induction CAF, a significant change of 12.6% (p < 0.001). All patients remained asymptomatic despite the decrease in diffusion capacity. Of the 150 patients studied, 30 (20%) had an apparent increase in Dl CO after induction CAF (Figure 1). However, the mean increase was only 8.0 ± 5.9% as compared with the 120 patients in whom the mean Dl CO decreased by 15.1 ± 10.1%.
Fig. 1. Effects of induction CAF on pulmonary function. The change in Dl CO in 150 patients undergoing 3 cycles of CAF induction chemotherapy is expressed as percentage change. The two time points used in this calculation are the Dl CO before protocol entry and on completion of 3 cycles of induction CAF. The mean decrease ± SEM in Dl CO was 12.6 ± 1.3% over these two time points and median decrease was 12.0%.
[More] [Minimize]To delineate the contribution an individual's baseline lung function has on this change in Dl CO, we performed linear regression analysis of the decline in Dl CO after induction CAF versus the individual's baseline Dl CO before induction CAF and found no significant correlation between the two parameters (data not shown). Multiple linear regression analysis was unable to correlate any relation between tobacco use and age with this decline in Dl CO after induction CAF.
Effects of consolidative chemotherapy. Results of serial PFT on patients randomized to both groups are shown in Figure 2. The patients in the HDC arm had a significant decrease in lung function. The mean nadir Dl CO for the HDC arm was 55.2 ± 14.1% at 15 to 18 wk postchemotherapy compared with 80.7 ± 12.3% at 6 to 15 wk for the SDC arm (p < 0.001). Fifty-four patients (72%) who received HDC had drug-induced interstitial pneumonitis and were treated with oral corticosteroids, as opposed to only three patients (4%) in the SDC arm. At every time point measured, the mean Dl CO of patients randomized to HDC was significantly lower than SDC, except at protocol entry, at randomization, and at Weeks 108 and 132.
Fig. 2. Long-term follow-up of pulmonary function in patients receiving induction CAF followed by either SDC or HDC consolidative CPB. Baseline Dl CO was recorded before protocol entry. Median length of follow-up was 112 and 84 wk for HDC and SDC, respectively. The number of individual subjects evaluated at each time point is shown adjacent to each corresponding symbol. All values are mean Dl CO; error bars indicate SEM. Asterisks indicate time points at which the mean Dl CO of the two randomized groups differs significantly; p < 0.05.
[More] [Minimize]To understand the effect of induction CAF on lung function after HDC, we performed linear regression analysis of percent change in Dl CO after induction CAF versus percent change in Dl CO after HDC (Figure 3). This analysis demonstrated a relationship between an individual's decrease in Dl CO after induction CAF with its further decrease after HDC. This highly significant relationship (p < 0.001) indicated that individuals with a larger decrease in Dl CO after induction CAF, experienced a more precipitous decrease in Dl CO following HDC.
Fig. 3. Linear regression analysis of the relationship between pulmonary function after induction CAF and after HDC. The abscissa represents the percent change in Dl CO after induction CAF (time points used: Dl CO before protocol entry and Dl CO on completion of 3 cycles of induction CAF). The ordinate represents the percent change in Dl CO after HDC (time points used: Dl CO on completion of 3 cycles of induction CAF and nadir Dl CO post-HDC). The slope of this line was significantly different from zero (p < 0.001). The dotted lines indicate 95% confidence limits.
[More] [Minimize]We analyzed in detail the Dl CO of 75 patients who were subsequently randomized to HDC. Two groups of patients, one whose Dl CO decreased more than 30% after induction CAF, and another group which had an apparent increase in Dl CO more than 10% after induction CAF, were chosen. Fifteen patients met the decrease in Dl CO by 30% criteria and six patients met the increase by 10% criteria. Among these 15 patients whose Dl CO decreased by more than 30%, no differences were found with regard to age, tobacco use, and baseline Dl CO values, when compared with the larger randomized group. However, 13 (87%) of these 15 patients subsequently developed DPTS, requiring prednisone. Six patients met the apparent increase by 10% criteria and two of these were current smokers at the time of baseline Dl CO but stopped smoking after registration for the protocol. Their increase may be due to cessation of smoking, resulting in decrease in CO blood levels and subsequent increase in Dl CO (11). From this group, only three patients (50%) subsequently developed DPTS and received prednisone.
We also analyzed the long-term follow-up of FVC and FEV1 in patients receiving induction CAF followed by either HDC or SDC (Figure 4). At all time points analyzed, the ratio of FEV1/FVC remained within normal limits (HDC: 0.78 to 0.80; SDC: 0.77 to 0.79). However, a significant decrease in FVC and FEV1 was observed up to 48 wk after HDC, as compared with patients receiving SDC. Notably, no difference was observed between the two groups after induction CAF. Thus induction CAF produced a significant reduction in Dl CO alone, whereas HDC substantially reduced both lung diffusion capacity and vital capacity.
Fig. 4. Long-term follow-up of FVC and FEV1 in patients receiving induction CAF followed by either HDC or SDC consolidative CPB. The ratio of FEV1 to FVC at each time point is indicated adjacent to each corresponding symbol. All values are mean FVC or FEV1; error bars indicate SEM.
[More] [Minimize]Of the 54 patients (72%) who were diagnosed and treated for DPTS in the HDC arm, 18 patients were treated on the basis of symptoms only. Their Dl CO was greater than 60% predicted, but these patients presented with symptoms typical of DPTS and empiric treatment with prednisone was initiated. We compared the Dl CO prior to prednisone, to the highest Dl CO during or immediately after treatment. Prednisone produced a significant increase in Dl CO (p < 0.001, n = 37) in all these patients with a mean rise in Dl CO of 12.7%. Patients in whom a Dl CO was not obtained within a week of initiating prednisone were excluded from the analysis. Eleven patients received a prolonged course of prednisone for symptoms during the tapering of the medication, and six patients received a second course of treatment for recurrent symptoms. Only three patients randomized to the SDC arm received prednisone. These three patients had symptoms of DPTS, but their Dl CO was 53%, 73%, and 84% predicted when prednisone was initiated. All three responded quickly to prednisone and the dose was tapered early.
To further evaluate the role prednisone plays in treatment of DPTS after HDC, we compared the mean Dl CO of patients who received prednisone to the mean Dl CO of patients who did not receive prednisone (Figure 5). Twenty-one patients did not receive prednisone and their mean nadir Dl CO was 77.5% predicted at 15 to 18 wk post-HDC. The peak mean Dl CO in these patients was 87.9% predicted at 48 wk post-HDC, but the difference between nadir and peak mean Dl CO was not statistically significant (p = 0.111). In contrast, the 54 patients who received prednisone had a mean nadir Dl CO of 54.4% predicted at 15 to 18 wk post-HDC. However, their mean Dl CO increased significantly to 64.4% predicted by 48 wk post-HDC (p = 0.001), and continued to increase to 72.1% predicted at 108 wk post-HDC.
Fig. 5. Long-term follow-up of pulmonary function, comparing 21 patients who did not receive prednisone after HDC with 54 patients who did receive prednisone after HDC for DPTS. All 21 patients who were not treated with prednisone received RT in accordance with the schedule. Only three patients who received prednisone were able to initiate and complete RT in accordance with the schedule. The number of individual subjects at each time point is shown adjacent to each corresponding symbol. All values are mean Dl CO with error bars representing SEM. Asterisks indicate time points at which the mean Dl CO of the two groups differs significantly; p < 0.05.
[More] [Minimize]We explored the role RT played in the development of DPTS. The protocol entails that RT be initiated 6 wk after HDC or SDC. However, among the 75 patients receiving HDC, only 24 were able to start and complete the RT course in accordance with protocol requirements. Of the remaining patients, initiation of RT was delayed owing to pulmonary toxicity in 36 patients, and in 12 patients RT was begun at 6 wk post-HDC but was interrupted because of development of pulmonary symptoms. In these 12 patients, RT was resumed after prednisone controlled the pulmonary symptoms. They received prednisone through the entire course of RT and had no further symptoms. In three patients receiving HDC, radiation therapy was interrupted owing to dermatological and other complications of RT.
In contrast, 70 patients receiving SDC did not have any pulmonary complications before, during, or subsequent to receiving RT. All three patients who developed DPTS after SDC, were symptomatic before RT, and were receiving prednisone when RT was initiated. However, in spite of pulmonary toxicity, RT was initiated at 6 wk post-SDC, without exacerbation of pulmonary symptoms. In two patients receiving SDC, radiation therapy was interrupted owing to dermatological complications.
No deaths were attributable to DPTS in either arm of the study. During the duration of the study, eight patients in the HDC arm and seven patients in the SDC arm expired. These deaths were caused mostly as a result of disease progression and infrequently owing to complications of chemotherapy other than pulmonary toxicity
Table 2 summarizes BAL parameters measured in nine patients receiving 4 cycles of induction CAF as compared with five healthy volunteers. Significantly higher percentages of neutrophils (p = 0.03) and lymphocytes (p = 0.006) were seen in the BALF of patients when compared with healthy volunteers. The cellularity of the BALF in these patients correlated with significant elevations in two proinflammatory cytokines: IL-6 and IL-8 (Figure 6). Calculated correlation coefficients were: IL-8/neutrophils: R = 0.803, p = 0.009; IL-6/lymphocytes: r = 0.928, p < 0.001. Both IL-6 and IL-8 were below detection limits in the healthy volunteers but were significantly higher (IL-6: p = 0.045, IL-8: p = 0.045) in the patients. Also, a significant decrease in TNF-R1 was seen in the BALF of patients (p = 0.033). TNF-α levels were below ELISA detection limits in both patients and healthy volunteers. No differences were observed in BALF total proteins, total glutathione, and oxidized glutathione among the two groups.
| Patients (n = 9) | p Value | Healthy Volunteers (n = 5) | ||||
|---|---|---|---|---|---|---|
| % Return of BALF | 56.0 (42.0–64.0) | 0.072 | 64.0 (58.0–77.0) | |||
| Macrophage, % | 58.7 (15.3–87.3) | 0.008† | 88.7 (76.7–90.0) | |||
| Neutrophil, % | 2.7 (0.8–12.4) | 0.030† | 0.0 (0.0–1.8) | |||
| Lymphocyte, % | 23.0 (7.7–78.3) | 0.006† | 4.0 (0.3–8.0) | |||
| Eosinophil, % | 0.7 (0.0–1.7) | 0.159 | 0.0 (0.0–0.7) | |||
| Protein, μg/ml | 91.1 (32.5–166.7) | 0.594 | 69.5 (46.9–76.7) | |||
| GSH, μmol/L | 0.3 (0.1–1.0) | 0.286 | 0.4 (0.3–1.1) | |||
| GSSG, μmol/L | 0.0 (0.0–0.5) | 0.790 | 0.1 (0.0–0.5) | |||
| IL-6, pg/ml | 2.5 (0.0–14.2) | 0.045† | 0.0 (0.0–0.2) | |||
| IL-8, pg/ml | 2.2 (0.0–22.6) | 0.045† | 0.4 (0.0–1.8) | |||
| TNF-R1, pg/ml | 35.9 (0.0–85.3) | 0.033† | 69.1 (53.6–101.5) | |||
| TNF-R2, pg/ml | 32.3 (25.4–105.3) | 0.594 | 49.2 (31.9–53.7) |
Fig. 6. Linear regression analysis demonstrating the relationship between IL-6, IL-8, and BALF cellular patterns. A strong and significant correlation was found between IL-6 and lymphocytes and between IL-8 and neutrophils. The dotted lines represent 95% confidence levels.
[More] [Minimize]Pulmonary complications represent a major cause of morbidity and mortality in patients who undergo bone marrow transplantation (15-17). In particular, allogeneic bone marrow transplantation appears to be associated with a higher risk of pulmonary complications as compared with autologous or syngeneic bone marrow transplantation (18). This may be attributed to the presence of graft versus host disease (GVHD), the use of methotrexate to control GVHD, and to the use of total body or high-dose-rate radiation therapy (18). Both infectious and noninfectious complications contribute significantly to morbidity and mortality after allogeneic bone marrow transplantation (15-17). Wingard and colleagues established the causes in 113 cases of nonbacterial, nonfungal interstitial pneumonitis (18). Fifty percent were idiopathic, 40% were caused by cytomegalovirus (CMV) alone or by CMV plus another viral pathogen, and the remaining 10% were caused by Pneumocystis carinii or viruses other than CMV. Idiopathic pneumonia syndrome (IPS) refers to this subgroup of noninfectious pneumonitis seen after bone marrow transplantation (19).
In contrast, the pulmonary toxicity seen in recipients of autologous bone marrow or peripheral stem cell support more likely represents the result of pretransplantation conditioning regimens, including induction chemotherapy, HDC, growth factor support, and possibly post-transplantation radiation therapy (19). In a study of lung pathology in patients with primary breast cancer treated with HDC and ABMT, open-lung and transbronchial biopsies showed alveolar septal thickening with fibrosis, atypical Type II pneumocytes, and pulmonary endothelial cell injury, characteristic of drug toxicity (7). We have previously proposed the term DPTS to distinguish this form of pulmonary toxicity seen after HDC with ABMT and/or PBPCS from IPS (10). In this specific population, DPTS has a higher incidence, lower mortality and appears steroid-responsive, all in apparent contrast to IPS.
In the current study, involving a large population of high-risk breast cancer patients undergoing HDC with ABMT and PBPCS, we have highlighted some early aspects of pulmonary toxicity not previously addressed. All patients enrolled in the study received 4 cycles of CAF induction chemotherapy before receiving their consolidative chemotherapy with either high-dose or standard-dose CPB. Given this setting, one role of induction CAF is to ascertain the responsiveness of the cancer to chemotherapeutic agents, and any patient having disease progression/recurrence after induction CAF, was excluded from further participation in the protocol. We found a significant reduction in pulmonary function occurring in these patients after induction CAF. The mean decrease in Dl CO of the 150 patients studied was 12.6% after 3 cycles of induction CAF.
Although asymptomatic, BALF from these patients demonstrated a modest inflammatory response with significant elevations in IL-6, IL-8, neutrophils, and lymphocytes. Both IL-6 and IL-8 are grouped together, along with IL-1 and TNF, as proinflammatory cytokines having significant relevance in lung pathologies. IL-6 is a pleiotropic cytokine having protean biological effects. In the lung, IL-6 is produced predominantly by fibroblasts (20), alveolar macrophages (21), and large-vessel endothelium (22) in response to a wide variety of stimuli including infection, inflammation, and trauma. IL-6 may, individually or in combination with IL-1 and TNF, serve to initiate local and systemic inflammatory processes, including T-cell proliferation, immunoglobulin production, fever, and acute-phase protein production (23). Considering these facts, IL-6 may be regarded as a nonspecific marker of lung injury and we found a 5-fold increase in IL-6 in the BALF of the patients, signifying lung injury.
IL-8 has been shown to be chemoattractant to neutrophils and lymphocytes at nanomolar (24) and picomolar (25) concentrations, respectively. The elicitation of neutrophils into the lung constitutes an essential element of the inflammatory response that is common in host defense. Although their presence is usually a hallmark of acute inflammation, they are nevertheless often involved in immunopathologic conditions associated with a number of pulmonary diseases (26, 27).
A major proinflammatory cytokine, TNF-α, was not detectable by ELISA in the BALF supernatant of the patients. This is not surprising, owing to the fact that most TNF-α is membrane bound and whatever small amount is secreted, is neutralized by TNF antagonists (28-30). However, we speculate that enhanced TNF activity was present based on a significant decrease in TNF-R1. Soluble receptors to TNF can potentially downregulate the proinflammatory effects of TNF by binding to TNF within the extracellular environment, thereby preventing TNF from reaching its target cells (28-30). However, whether these immunological events are a protective reaction to stress or represent dysregulated autocrine and paracrine pathways in inflammation remains to be elucidated.
At our institution, cyclophosphamide was the only agent included in the induction chemotherapy protocol which has been previously known to cause pulmonary toxicity. At the tissue level, cyclophosphamide increases the deposition of collagen in the lung (31) and probably causes oxidative damage by increasing the generation of reactive oxygen species by alveolar macrophages (32) and depleting glutathione stores (32). The alveolar epithelial lining fluid is known to have the highest concentration of glutathione of any extracellular compartments in the body (33). Oxidation or depletion of this antioxidant may be a marker of oxidative stress. However, we found neither an increase in oxidized glutathione nor a decrease in total glutathione in the BALF of patients receiving induction CAF. This indicates that mechanisms other than oxidative stress were responsible for induction CAF–induced pulmonary toxicity.
We explored other factors that may contribute to induction CAF–associated pulmonary toxicity but found no correlation between tobacco use, age, and baseline lung function. Induction CAF may act to prime the lungs for the more acute toxicity which follows HDC. A highly significant, albeit moderately strong correlation, between decreased pulmonary function after induction CAF and development of DPTS favors this hypothesis. Further, individual review of patients' PFT reveals a much higher incidence of DPTS (87%) in patients whose Dl CO decreased by more than 30% after induction CAF. On the other hand, patients with apparent increase in Dl CO more than 10% after induction CAF had only a 50% incidence of DPTS.
After consolidative HDC, the decrease in pulmonary function was directly related to the arm to which the patients randomized. In the HDC arm, 72% of patients were treated with steroids for DPTS, compared with only 4% of patients in the SDC arm. Also, the mean nadir Dl CO of patients randomized to receive HDC was much lower (55.2% predicted), compared with the mean nadir Dl CO of the SDC arm (80.7% predicted). Development of DPTS in patients receiving HDC did not correlate with tobacco use, age, or baseline lung function values in our study. Interestingly, induction CAF specifically decreased Dl CO without affecting vital capacity, whereas CPB from HDC decreased Dl CO as well as vital capacity, but to a much greater extent. This may indicate one of two possibilities. First, the mechanism of induction CAF–induced pulmonary toxicity may be distinct from that of HDC–induced pulmonary toxicity, or second, the entire process of lung damage, starting with induction CAF, may be a continuum of a single mechanism, exacerbated by HDC, resulting in reduced lung volume and diffusion capacity.
The major difference between the two arms of the protocol, which may be responsible for the higher incidence of DPTS in patients receiving HDC, was the dose of chemotherapeutic agents. Although cyclophosphamide toxicity is claimed to be independent of dosage, BCNU has been shown to have a dose-dependant toxicity profile (34, 35). However, these studies have suggested that BCNU has minimal pulmonary toxicity at doses below 1,000 mg/m2. The patients in our protocol received 600 mg/m2 during HDC, making it less likely for BCNU to be the sole source of pulmonary toxicity. The synergism between BCNU and cyclophosphamide to cause pulmonary toxicity at the given doses has been proposed by some investigators (7, 9). This is based on the observation that both drugs deplete reduced glutathione and impair antioxidant defenses.
RT may contribute to the development or exacerbation of DPTS but does not appear to be a major factor. Twelve patients receiving HDC developed DPTS during RT, requiring cessation of RT and treatment with prednisone. On the contrary, none of the patients receiving SDC developed pulmonary symptoms with respect to RT. The commencement of RT was delayed in patients who developed DPTS until prednisone therapy improved and stabilized the symptomatology. In general, when RT was started, these patients were maintained on a low level of prednisone and their pulmonary symptoms/function appeared to remain stable. Of 29 breast cancer patients receiving similar HDC and stem cell support, Wilczynski and coworkers (10) demonstrated an average reduction of 13.8% in Dl CO, but could not demonstrate any relationship between the decrease in Dl CO and the time at which RT was received (relative to HDC). We interpret these data to support the conclusion that RT, given in this setting, plays a minor role in the pathogenesis of DPTS.
In our patient population developing DPTS, prednisone produced a dramatic improvement in symptoms and reversal of pulmonary toxicity as witnessed by a 12.6% increase in the mean Dl CO after treatment. Interestingly, 18 patients (33%) who developed DPTS after they received HDC, had a Dl CO greater than 60% predicted, but yet they were symptomatic. Long-term follow-up of pulmonary function shows a gradual increase in the mean Dl CO of patients receiving HDC, and eventually the mean Dl CO of these patients was not significantly different from the mean Dl CO of patients receiving SDC. Our patient population showed no mortality attributable to DPTS, which we attribute to our early diagnosis and prompt treatment with prednisone. The 54 patients who received prednisone for DPTS had significantly greater increase in mean Dl CO compared with patients who did not receive prednisone. Based on these data and our experiences, we propose guidelines for diagnosing and treating DPTS after HDC with ABMT and/or PBPCS (Figure 7). It should be noted that these guidelines were developed for the specific patient population described previously. To fully evaluate the role prednisone plays in treatment of DPTS would require a prospective randomized trial. However, this is unlikely to occur based on the positive clinical response of patients with DPTS to prednisone and ethical considerations of giving these severely affected individuals placebo medications.
Fig. 7. Guidelines for the diagnosis and management of delayed pulmonary toxicity after HDC and autologous bone marrow/peripheral blood stem cell support. In all cases, the Dl CO should be corrected for hemoglobin. Ideally, a patient's Dl CO should be compared with his or her baseline Dl CO (Dl CO performed before receiving chemotherapy). If a baseline Dl CO is not available, Dl CO percent predicted should be used. See text for prednisone taper schedule.
[More] [Minimize]In conclusion, CAF induction chemotherapy alone, decreases pulmonary function (as measured by Dl CO), increases pulmonary alveolar cellular inflammation, and may predispose to further HDC-induced delayed pulmonary toxicity. Although DPTS presents in a high percentage of individuals, judicious and expeditious use of prednisone appears to improve a significant amount of pulmonary function with a trend to normalization. Whether prophylactic treatment strategies will prevent DPTS in this patient population altogether, remains to be determined.
The authors thank Ken Kuzenski for excellent computer database support, faculty and staff of the Duke University Bone Marrow Transplant Program, participating breast cancer patients, the Duke University Pulmonary Function Laboratory staff, and Drs. Joe Govert and Neil MacIntyre for discussion and critical reading of this manuscript. They also acknowledge the support of the Duke University Comprehensive Cancer Center.
Supported in part by National Institutes of Health Grants HL55166 and ES/HL- 08698, and American Heart Association Grant in Aid to R.J.F. and 2P01CA47741 to N.J.C.
| 1. | Peters W. P., Ross M., Vredenburgh J. J., Meisenberg B., Marks L. B., Winer E., Kurtzberg J., Bast J. R. C., Jones R., Shpall E., Wu K., Rosner G., Gilbert C., Mathias B., Coniglio D., Petros W., Henderson I. C., Norton L., Weiss R. B., Budman D., Hurd D.High-dose chemotherapy and autologous bone marrow support as consolidation after standard-dose adjuvant therapy for high-risk primary breast cancer. J. Clin. Oncol.11199311321143 |
| 2. | Antman K. H., Rowlings P. A., Vaughan W. P., Pelz C. J., Fay J. W., Fields K. K., Freytes C. O., Gale R. P., Hillner B. E., Holland H. K., Kennedy M. J., Klein J. P., Lazarus H. M., McCarthy P. L., Saez R., Spitzer G., Stadtmauer E. A., Williams S. F., Wolff S., Sobocinski K. A., Armitage J. O., Horowitz M. M.High-dose chemotherapy with autologous hematopoietic stem-cell support for breast cancer in North America. J. Clin. Oncol.15199718701879 |
| 3. | Rahman Z. U., Hortobagyi G. N., Buzdar A. U., Champlin R.High-dose chemotherapy with autologous stem cell support in patients with breast cancer. Cancer Treat. Rev.241998249263 |
| 4. | Hurd D. D., Peters W. P.Randomized, comparative study of high-dose (with autologous bone marrow support) versus low-dose cyclophosphamide, cisplatin, and carmustine as consolidation to adjuvant cyclophosphamide, doxorubicin, and fluorouracil for patients with operable stage II or III breast cancer involving 10 or more axillary lymph nodes (CALGB Protocol 9082). Cancer and Leukemia Group B. J. Natl. Cancer. Inst. Monogr.1919954144 |
| 5. | Aisner J., Cirrincione C., Perloff M., Perry M., Budman D., Abrams J., Panasci L., Muss H., Citron M., Holland J.et alCombination chemotherapy for metastatic or recurrent carcinoma of the breast—a randomized phase III trial comparing CAF versus VATH versus VATH alternating with CMFVP: Cancer and Leukemia Group B Study 8281. J. Clin. Oncol.13199514431452 |
| 6. | Schwartzberg L. S., Birch R., West W. H., Tauer K. W., Wittlin F., Leff R., Campos L., Rymer W., Carter P., Mangum M., Greco F. A., Hainsworth J., Raefsky E., Blanco R., Buckner C. D., Weaver C. H.Sequential treatment including high-dose chemotherapy with peripheral blood stem cell support in patients with high-risk stage II-III breast cancer: outpatient administration in community cancer centers. Am. J. Clin. Oncol.211998523531 |
| 7. | Todd N. W., Peters W. P., Ost A. H., Roggli V. L., Piantadosi C. A.Pulmonary drug toxicity in patients with primary breast cancer treated with high-dose combination chemotherapy and autologous bone marrow transplantation. Am. Rev. Respir. Dis.147199312641270 |
| 8. | Jones R. B., Matthes S., Shpall E. J., Fisher J. H., Stemmer S. M., Dufton C., Stephens J. K., Bearman S. I.Acute lung injury following treatment with high-dose cyclophosphamide, cisplatin, and carmustine: pharmacodynamic evaluation of carmustine. J. Natl. Cancer Inst.851993640647 |
| 9. | Chap L., Shpiner R., Levine M., Norton L., Lill M., Glaspy J.Pulmonary toxicity of high-dose chemotherapy for breast cancer: a non-invasive approach to diagnosis and treatment. Bone Marrow Transplant.20199710631067 |
| 10. | Wilczynski S. W., Erasmus J. J., Petros W. P., Vredenburgh J. J., Folz R. J.Delayed pulmonary toxicity syndrome following high-dose chemotherapy and bone marrow transplantation for breast cancer. Am. J. Respir. Crit. Care Med.1571998565573 |
| 11. | American Thoracic SocietySingle-breath carbon monoxide diffusing capacity (transfer factor). Am. J. Respir. Crit. Care Med.152199521852198 |
| 12. | American Thoracic SocietyStandardization of spirometry. Am. J. Respir. Crit. Care Med.152199511071136 |
| 13. | Crapo R. O., Morris A. H., Gardner R. M.Reference spirometric values using techniques and equipment that meet ATS recommendations. Am. Rev. Respir. Dis.1231981659664 |
| 14. | Crapo R. O., Morris A. H.Standardized single breath normal values for carbon monoxide diffusing capacity. Am. Rev. Respir. Dis.1231981185189 |
| 15. | Robbins R. A., Floreani A. A., Buchalter S. E., Spurzem J. R., Sisson J. H., Rennard S. I.Pulmonary complications of transplantation. Annu. Rev. Med.431992425435 |
| 16. | Breuer R., Lossos I. S., Berkman N., Or R.Pulmonary complications of bone marrow transplantation. Respir. Med.871993571579 |
| 17. | Ettinger N. A., Trulock E. P.Pulmonary considerations of organ transplantation: part 2. Am. Rev. Respir. Dis.1441991213223 |
| 18. | Wingard J. R., Mellits E. D., Sostrin M. B., Chen D. Y.-H., Burns W. H., Santos G. W., Vriesendorp H. M., Beschorner W. E., Saral R.Interstitial pneumonitis after allogeneic bone marrow transplantation: nine-year experience at a single institution. Medicine671988175186 |
| 19. | Clark J. G., Hansen J. A., Hertz M. I., Parkman R., Jensen L., Peavy H. H.Idiopathic pneumonia syndrome after bone marrow transplantation. Am. Rev. Respir. Dis.147199316011606 |
| 20. | Elias J. A., Lentz V.IL-1 and tumor necrosis factor synergistically stimulate fibroblast IL-6 production and stabilize IL-6 messenger RNA. J. Immunol.1451990161166 |
| 21. | Kotloff R. M., Little J., Elias J. A.Human alveolar macrophage and blood monocyte interleukin-6 production. Am. J. Respir. Cell Mol. Biol.31990497505 |
| 22. | Loppnow H., Libby P.Adult human vascular endothelial cells express the IL6 gene differentially in response to LPS or IL1. Cell Immunol1221989493503 |
| 23. | Zitnik, R. J., and J. A. Elias. 1993. Interleukin-6 and the lung. In J. Kelley, editor. Cytokines of the Lung, Lung Biology in Health and Disease. Marcel Dekker, Inc., New York. 229–280. |
| 24. | DeMarco D., Kunkel S. L., Strieter R. M., Basha M., Zurier R. B.Interleukin-1 induced gene expression of neutrophil activating protein (interleukin-8) and monocyte chemotactic peptide in human synovial cells. Biochem. Biophys. Res. Commun.1741991411416 |
| 25. | Larsen C. G., Anderson A. O., Appella E., Oppenheim J. J., Matsushima K.The neutrophil-activating protein (NAP-1) is also chemotactic for T lymphocytes. Science243198914641466 |
| 26. | Hunninghake G. W., Gadek J. E., Lawley T. J., Crystal R. G.Mechanisms of neutrophil accumulation in the lungs of patients with idiopathic pulmonary fibrosis. J. Clin. Invest.681981259269 |
| 27. | Hunninghake G. W., Garrett K. C., Richerson H. B., Fantone J. C., Ward P. A., Rennard S. I., Bitterman P. B., Crystal R. G.Pathogenesis of the granulomatous lung diseases. Am. Rev. Respir. Dis.1301984476496 |
| 28. | Hofsli E., Lamvik J., Nissen-Meyer J.Evidence that tumour necrosis factor (TNF) is not constitutively present in vivo: the association of TNF with freshly isolated monocytes reflects a rapid in vitro production. Scand. J. Immunol.281988435441 |
| 29. | Ulich T. R., Guo K. Z., Irwin B., Remick D. G., Davatelis G. N.Endotoxin-induced cytokine gene expression in vivo: II. Regulation of tumor necrosis factor and interleukin-1 alpha/beta expression and suppression. Am. J. Pathol.137199011731185 |
| 30. | Beutler B., Greenwald D., Hulmes J. D., Chang M., Pan Y. C., Mathison J., Ulevitch R., Cerami A.Identity of tumour necrosis factor and the macrophage-secreted factor cachectin. Nature3161985552554 |
| 31. | Morse C. C., Sigler C., Lock S., Hakkinen P. J., Haschek W. M., Witschi H. P.Pulmonary toxicity of cyclophosphamide: a 1-year study. Exp. Mol. Pathol.421985251260 |
| 32. | Cooper J. A. D., Marrill W. W.Modulation of endoperoxide product levels and cyclophosphamide induced injury by glutathione repletion. J. Appl. Physiol.67198923162322 |
| 33. | Cantin A. M., North S. L., Hubbard R. C., Crystal R. G.Normal alveolar epithelial lining fluid contains high levels of glutathione. J. Appl. Physiol.631987152157 |
| 34. | Durant J. R., Norgard M. J., Murad T. M., Bartolucci A. A., Langford K. H.Pulmonary toxicity associated with bischloroethylnitrosourea (BCNU). Ann. Intern. Med.901979191194 |
| 35. | Selker R. G., Jacobs S. A., Moore P. B., Wald M., Fisher E. R., Cohen M., Bellot P.1,3-Bis(2-chloroethyl)-1-nitrosurea (BCNU)- induced pulmonary fibrosis. Neurosurgery71980560565 |
