Japanese Journal of Clinical Oncology

INTRODUCTION

Bleomycin is one of the key drugs used for induction chemotherapy for testicular cancer. Pulmonary toxicity is the major and potentially fatal adverse side-effect of this drug. The incidence of pulmonary complication reaches 10% at cumulative bleomycin doses >450 mg (1). The cumulative dose of bleomycin used for the treatment of testicular cancer is usually <450 mg and serious pulmonary complications are not frequent at this dose. However, the results of several reports have suggested that patients treated with bleomycin were at risk of developing post-operative respiratory distress (2-4). Oxygen-exacerbated bleomycin toxicity is one of the proposed mechanisms accounting for serious post-operative respiratory complications (2,4). Age, prior lung disease, smoking, radiation therapy and impaired renal function have been considered as risk factors of bleomycin-induced pulmonary toxicity. Renal function is the most important marker, because bleomycin is usually used in combination with cisplatin in testicular cancer. In addition, several authors have reported cases of patients with severe pulmonary distress who received both recombinant human granulocyte colony-stimulating factor (G-CSF) and bleomycin, a result which suggested that G-CSF might augment bleomycin-induced pulmonary toxicity (5,6). In the present study, we retrospectively analyzed the effect of these factors on post-chemotherapy diffusion capacity (DLCO) and found that the elevation of the serum creatinine (s-Cr) level was the most significant predicting factor in the decrease in DLCO.

PATIENTS AND METHODS

We retrospectively analyzed the cases of 20 men treated for metastatic testicular cancer (11 patients with seminoma and nine with non-seminomatous testicular cancer) at Tsukuba University Hospital between 1990 and 1995. The median age of the patients was 36 years (range 18-66 years). Thirteen patients were classified as TNM clinical stage II and seven were stage III. Six of the stage III patients had lung metastases. All patients were treated with two to four cycles of a PVB regimen (7) or BEP regimen (8). The PVB regimen consisted of cisplatin 100 mg/m² intravenously (i.v.) on day 1, vinblastine 0.2 mg/m² i.v. on day 2, every 3 weeks. Bleomycin (30 IU) was given intramuscularly once a week during the chemotherapy. In the BEP regimen, etoposide 20 mg/m² i.v. on days 1-5 was used instead of vinblastine. Corticosteroid at doses of 250-500 mg was used in 16 patients to ameliorate chemotherapy-related nausea and vomiting. Recombinant human G-CSF was used in all but one patient at 250 µg/day by subcutaneous injection, when the white blood cell (WBC) count was below 2000/mm². The patient characteristics are summarized in Table 1.

Table 1. Patients' characteristics

	No. of patients
Tumor type
Seminoma	11
Non-seminoma	9
Stage
II	13
III	7
Lung metastases	6
Smoking history	10
Chemotherapy regimen
PVB	11
BEP	9
Corticosteroid for antiemetics	16
G-CSF for leukocytopenia	19

Statistical Analysis

The relation between the decrease in DLCO and risk factors was evaluated as follows. We defined the decrease in DLCO to below 75% of the predicted value as a binary outcome (0 = absent, 1 = present). The risk factors evaluated included age (0 = <30 yr, 1 = [ge]30 yr), bleomycin dose (0 = <300 mg, 1 = [ge]300 mg), cisplatin dose (0 = <500 mg, 1 = [ge]500 mg), renal injury (0 = absent, 1 = present), smoking history (0 = no, 1 = yes), leukocytosis with WBC counts of >20 000/mm³ (0 = absent, 1 = present), lung metastasis (0 = absent, 1 = present) and regimen (0 = BEP, 1 = PVB). We divided the population by the value of cumulative bleomycin dose of 300 mg because the most standard initial treatment of testicular cancer is three courses of PVB or BEP regimen which consist of cumulative bleomycin dose <300 mg. The value of the cumulative cisplatin dose was according to the median cumulative dose of cisplatin (537 mg).

Most patients treated with G-CSF had episodes of leukocytosis with WBC counts >10 000/mm³. We therefore divided the population by leukocytosis with WBC counts of 20 000/mm³.

Chi-squared analyses were carried out for the decrease in DLCO and each risk factor. We then carried out logistic regression analyses (12) to test the relation between the decrease in DLCO and the risk factors. The risk factors in the logistic regression model were selected based on the results of chi-squared tests. We used the software program SAS version 6.11 for all statistical analyses.

RESULTS

The pretreatment vitalometry revealed no abnormality in 12 of the 13 patients tested. The other patient showed a mild decrease in the forced expiratory volume in 1 s. The pretreatment DLCO was measured in eight patients and none showed any abnormality.

The post-treatment DLCO was decreased to below 75% of the predicted values in nine patients, as shown in Table 2. The median cumulative bleomycin dose at the DLCO measurement was 270 mg. At the time that the DLCO was measured, only one patient (patient 1) had clinical symptoms and signs of lung toxicity. In the other 19 patients, the blood gas analyses, vitalometory and chest X-ray film revealed no abnormality at the time of evaluating the DLCO.

A significant elevation of the s-Cr level was observed in seven patients during chemotherapy. The peak levels of s-Cr were 1.4-4.0 mg/dl. The elevation of s-Cr was observed in the first course of chemotherapy in three patients, in the first and second courses in two patients and in the fourth course in one patient. In one patient (case 14), the s-Cr elevated to 1.8 mg/dl in the first and second courses; thereafter, a continuous mild elevation of s-Cr (up to1.5 mg/dl) was observed during chemotherapy. In the other six patients, the s-Cr level returned to the normal range within a few weeks. Hence post-treatment DLCO measurements were performed after the peak s-Cr elevation in all patients. No patient received hemodialysis. Two injections of bleomycin were omitted because of renal impairment in patients 1 and 5. Overall, five of these seven patients (patients 4, 5, 7, 14 and 20) received the bleomycin injection at least twice when the s-Cr was abnormally elevated.

Table 2. Bleomycin dose, renal injury, leukocytosis and post-treatment DLCO

Case	Bleomycin dose at DLCOmeasurement (mg)	Peak s-Cr level(mg/dl)	Peak WBC count(×104/mm3)	DLCO(% predicted value)
1	330	3.1	5.0	56.5
2	270	-*	-[dagger]	91.0
3	300	-	3.6	76.3
4	330	1.7	2.3	60.6
5	330	1.4	3.3	71.1
6	360	-	3.4	76.6
7	270	4.0	3.1	65.2
8	180	-	2.6	107.2
9	270	-	-	78.8
10	270	-	4.2	83.1
11	270	-	2.6	68.1
12	240	-	4.5	80.6
13	270	-	4.2	132.0
14	270	1.8	4.4	60.2
15	270	-	-	85.2
16	270	-	-	70.6
17	240	-	-	67.4
18	270	1.4	2.3	75.7
19	270	-	-	103.4
20	270	1.7	2.5	70.9

*No elevation of serum creatinine level during chemotherapy. ^[dagger]No leukocytosis with WBC count above 20 000/mm³ during chemotherapy.

Leukocytosis with WBC count >20 000/mm³ was observed in 14 patients. The leukocytosis occurred secondary to the G-CSF administration in all cases.

To determine the risk factors for a decrease in DLCO, we evaluated the eight covariates listed in Table 3 by a logistic procedure. As shown in Table 3, renal injury detected by the elevation of s-Cr was the only significant risk factor (P = 0.018) by the chi-squared test. Based on the results of the chi-squared tests, we performed a logistic regression analysis using bleomycin dose, cisplatin dose, renal injury detected by the elevation of s-Cr and leukocytosis as contributing factors. The results indicated that renal injury was an independent risk factor for a decrease in the DLCO (P = 0.049, odds ratio = 22.3) (Table 4). The odds ratio of cisplatin dose was 5.04 (95% confidence limits 0.22-115.3), which was second to renal injury. However, cisplatin dose was not found to be a significant independent risk factor by logistic regression analysis (P = 0.311). In addition, the bleomycin dose ([ge]300 mg vs <300 mg) was not an independent risk factor in this analysis (P = 0.994). The effect of bleomycin dose was also evaluated using the cumulative bleomycin dose per unit body area. The average bleomycin dose per unit body area at the DLCO measurement for all patients was 162 mg/m². The bleomycin dose per unit body area was <162 mg/m² in six of the nine patients with decreased DLCO. There was no significant difference in cumulative bleomycin dose per unit body area at the DLCO measurement between the decreased DLCO patients and the normal DLCO patients (157 ± 36 and 165 ± 28 mg/m², respectively).

After the measurement of DLCO, 16 patients including eight patients with decreased DLCO underwent retroperitoneal lymph node dissection for residual tumor. No patients developed respiratory failure post-operatively. Five patients received additional chemotherapy (BEP in two cases and high dose chemotherapy consisting of etoposide, carboplatin and cyclophosphamide in three cases and methotrexate, actinomycin-D and CDDP in two cases). No clinical lung toxicity was observed during or after the additional chemotherapy.

Table 3. Determination of risk factors for the decrease in DLCO by chi-squared test

Factor	P value
Age	0.442
Bleomycin dose	0.442
Cisplatin dose	0.064
Renal injury	0.018
Smoking history	0.654
Leukocytosis	0.496
Lung metastases	0.769
Regimen	0.964

Table 4. Determination of independent risk factors for the decrease in DLCO by logistic analysis

Factor	P value	Odds ratio(95% confidence limits)
Bleomycin dose	0.994	0.99 (0.04-22.1)
Cisplatin dose	0.311	5.04 (0.22-115.3)
Renal injury	0.049	22.3 (1.02-487.3)
Leukocytosis	0.333	0.20 (0.01-5.28)

DISCUSSION

The incidence of bleomycin pulmonary toxicity is related to the total dose of bleomycin, age, prior lung disease, concomitant radiotherapy and the administration of other cytotoxic drugs and oxygen therapy. In induction chemotherapy for testicular cancer, in which the PVB and BEP regimens are generally used, the total bleomycin dose is usually <450 mg. Hence clinical bleomycin pulmonary toxicity is rare in induction chemotherapy. However, several authors have reported post-operative respiratory distress after induction chemotherapy with a total bleomycin dose <450 mg (2-4). Goldiner et al. (2) pointed out the possible causal role of intra-operative and post-operative oxygen therapy in exacerbating bleomycin pulmonary toxicity. The details of pre-operative pulmonary function were not reported in most of these cases; however, Ingrassia et al. (4) described one case suggesting the presence of subclinical bleomycin pulmonary toxicity with decreased pre-operative DLCO (67% of the predicted value). This finding indicates that the evaluation of respiratory function is necessary before adjuvant surgery, even when the total bleomycin dose is <450 mg. The DLCO is an easy and useful indicator of subclinical lung toxicity during bleomycin treatment and was reported to predict the lung toxicity of bleomycin in a prospective study (13).

In the present study, only one of the 20 patients (5%) developed typical signs and symptoms of lung toxicity. In contrast, a significant but subclinical decrease in DLCO was observed in nine of the 20 patients (45%) after 2-4 cycles of chemotherapy. When a significant decrease in DLCO was detected, we routinely consulted the anesthesiologist about the intra-operative respiratory management. No patient developed post-operative respiratory distress.

The logistic regression analyses revealed that the renal injury detected by the elevated s-Cr was the only independent risk factor for the decrease in DLCO (P = 0.049, odds ratio = 22.3). The total bleomycin dose and bleomycin dose per unit body area were not found to be significant factors at the relatively low dose used in the induction chemotherapy. The leukocytosis induced by G-CSF also did not seem to influence the DLCO. There was no difference in the peak WBC count or in the duration of leukocytosis during chemotherapy between the normal DLCO and the decreased DLCO patients (data not shown). Several authors have reported severe pneumonitis in patients on bleomycin and G-CSF and suggested a synergic effect of G-CSF on bleomycin-induced pulmonary toxicity (5,6). Some controversy surrounds the effect of G-CSF (14); the possible effects of G-CSF are thought to be mediated by the recruitment of neutrophils or released reactive oxygen species (6,15). We cannot draw any conclusion as to whether the synergism is present or not based on the results of the present study, but it seems that the WBC count in the peripheral blood cannot predict the risk of developing pulmonary injury enhanced by the G-CSF effect. Further investigation is needed on this point, because the interval between WBC counting and G-CSF administration was different among the cases in the present study.

The present analyses showed that renal impairment was a significant risk factor for the decrease in post-treatment DLCO. A correlation between renal function and DLCO was also revealed in the eight patients in whom both pre- and post-treatment DLCO were measured. In four of these eight patients, the normal pretreatment DLCO was significantly decreased after treatment and three of the four patients revealed a transient elevation of s-Cr during chemotherapy. Results of the present study are consistent with those of a previous study which demonstrated the association of renal function and DLCO during PVB treatment for testicular cancer (16). One of the proposed underlying mechanisms is the alteration of bleomycin pharmacokinetics, because bleomycin is known to be cleared principally by renal excretion. Hall et al. (17) found that the plasma clearance of bleomycin was correlated with s-Cr concentration. Crooke et al. (18) reported that the terminal elimination half-life of bleomycin increased exponentially as the creatinine clearance (Ccr) decreased, when the Ccr decreased to <25-35 ml/min. In the present retrospective study, the influence of bleomycin clearance was not clear because the Ccr level at the time of bleomycin injection was not evaluated. Regarding bleomycin exposure, however, the total bleomycin dose or bleomycin dose per unit body area did not influence the DLCO at the low dose used in the present study. Therefore, we propose that the correlation between renal impairment and the decrease in DLCO is not simply the result of an increase in the cumulative bleomycin exposure.

Sleijfer et al. (19) recently proposed another mechanism for the decrease in DLCO after bleomycin-containing combination chemotherapy for testicular cancer. They observed a decrease in DLCO in both patients who received treatment with BEP and those who received EP (etoposide and cisplatin). A significant decrease in the DLCO in the BEP patients compared with the EP patients was observed only after four cycles of BEP treatment. Sleijfer et al. proposed that capillary alteration caused by cisplatin and/or etoposide may have been responsible for the decrease in DLCO in the patients who received EP. They also pointed out that the pulmonary capillary blood volume and the vital capacity might be specific indicators of bleomycin-induced pulmonary damage. The effect of etoposide on the DLCO is not considered to have been significant in the present study, because no difference in the DLCO was observed between the BEP- and the PVB-treated patients. Although statistically not significant, the odds ratio of cumulative cisplatin dose for decrease in DLCO was 5.04 (95% confidence limits: 0.22-115.3), which was second to renal injury. Hence it is possible that cisplatin-induced vascular damage might be involved in the mechanism of the association between renal injury and the decrease in DLCO.

In conclusion, we have demonstrated that a decrease in DLCO is frequently seen in patients treated with induction chemotherapy with PVB and BEP for testicular cancer and that elevated s-Cr during chemotherapy is a significant indicator of a decrease in DLCO. The mechanism underlying the decrease in DLCO remains unclear and there is a possibility that the change in DLCO might not directly indicate bleomycin-induced lung toxicity. Although further investigation is needed, we recommend that a detailed pulmonary function assessment be performed during and after induction chemotherapy, especially when an elevation of the serum creatinine level is seen during the chemotherapy.