Neurophysiological Study of Peripheral Neuropathy after High-Dose Paclitaxel

Eligibility.

Patients with high-risk breast cancer (stage II with ≥10 axillary lymph nodes involved, stage IIIA, or IIIB carcinoma) were eligible to receive ACT with or without amifostine, whereas patients with responsive or stable stage IV disease were eligible to receive ACT with amifostine. The following were required: Karnofsky performance status of ≥80%; age of ≤60 years; bilirubin within normal range; AST and ALT ≤2 × the upper limit of normal; a measured creatinine clearance ≥70 ml/min; cardiac left ventricular ejection fraction ≥55%; a forced expiratory volume in 1 s of ≥2 liters; and diffusion capacity of carbon monoxide >60% of that predicted. Patients with bone marrow or central nervous system metastases were excluded. Although patients were not randomized between the two treatments (ACT alone or ACT with amifostine), the two protocols ran concurrently, had similar eligibility criteria, and sequential patients on each underwent neurophysiological studies. The same technologist performed all nerve conduction studies and quantitative sensory tests. All patients participating in these studies signed informed consent documents approved by the City of Hope Institutional Review Board.

Treatment Plan.

Granulocyte colony-stimulating factor (Filgrastim, Amgen Inc., Thousand Oaks, CA), 5 μg/kg bid, was administered s.c. On the 5th day, apheresis commenced and continued through collection of at least 2 × 10⁶/kg CD34+ peripheral blood progenitor cells. Cells were collected and immediately cryopreserved as reported earlier (20) . Ideal body weight was used when calculating doses. Doxorubicin (165 mg/m²) was administered over 96 h, between days −9 and −5, followed by 100 mg/kg cyclophosphamide over 2 h on day −5. A 24-h continuous i.v. infusion of 725 mg/m² paclitaxel was started on day −4. Amifostine (740 mg/m² over 10 min; ALZA Corporation, Mountain View, CA) was administered both before and 12 h into the paclitaxel infusion. Twenty-five percent of peripheral blood progenitor cells were reinfused on day −2 and 75% on day 0. Supportive care was provided as per institutional standards. Granulocyte colony-stimulating factor 5 μg/kg bid, i.v. was started with the first stem cell reinfusion and was continued until the absolute granulocyte count was ≥1 × 10⁹/liters for 3 consecutive days.

Assessment of Peripheral Neuropathy.

Neurological signs were recorded and neurophysiological tests (nerve conduction studies and quantitative sensory testing) obtained on consecutive patients before and 20–40 days after ACT chemotherapy. Peroneal nerve motor conduction studies were performed with the stimulating electrode at the ankle and the recording electrode 8 cm distally over the extensor digitorum brevis muscle. The sural nerve was stimulated antidromically in the posterior lower leg and recorded 14 cm distally at the lateral malleolus. Quantitative sensory testing of vibratory and cold detection thresholds was performed with the Case IV system one-time-period 4, 2, 1 Stepping Algorithm (21) . The point of stimulation was immediately distal to the base of the nail of the right large toe for vibration and the dorsal surface of the right foot for cold. After paclitaxel treatment, the relative changes in nerve conduction and quantitative sensory tests were recorded and expressed as the logarithm to the base 2 of ratios of post-transplant/pretransplant determinations (a logarithm ratio of 0 indicates no change, −1 indicates the value was reduced by half, +1 indicates an increase by a factor of 2, −2 indicates a reduction by a factor of 4, and so forth).

Peripheral neuropathy scores were obtained by a modification of the method by Chaudhry et al. (3) . As shown in Table 1⇓ , scores of 0–3 were given for each of the following seven categories: (category 1) sensory symptoms; (category 2) pin sensory loss; (category 3) vibration sensory loss; (category 4) strength; (category 5) deep tendon reflexes; (category 6) sural sensory nerve action potential (SNAP) amplitude; and (category 7) peroneal nerve compound muscle action potential (CMAP) amplitude. Percentage reduction of SNAP and CMAP were determined on the basis of the lower limit of the laboratory normal. Sum of scores could range from 0 (no impairment) to 21 (maximum impairment on this scale).

Pharmacokinetics.

For those patients in whom pharmacokinetic studies were performed, peripheral blood samples were obtained immediately before the start of the paclitaxel infusion, at 3, 6, 12, 20, and 24 h during the paclitaxel infusion, then at 15 and 30 min and 1, 2, 3, 6, 12, and 24 h after completion of the paclitaxel infusion. The concentration of paclitaxel in plasma was measured in accordance with a HPLC method published previously (22) . Pharmacokinetic data analyses of individual plasma drug concentration versus time curves were performed using ADAPT II software (Biomedical Simulations Resource, University of Southern California, Los Angeles, CA). Paclitaxel areas under the curves (AUCs) were estimated using linear trapezoids, with the terminal area extrapolated to infinity, using both the measured and fitted drug concentrations. Paclitaxel systemic clearances (CL_sys) were determined using the following relationship: CL_sys = dose/AUC. Peak plasma paclitaxel concentrations were defined as the actual paclitaxel levels measured immediately before the end of the infusion. The time that the plasma paclitaxel concentration remained above 0.05 μm was estimated by extrapolation of the model-derived fits of each individual patient’s data.

Statistical Analysis.

Not all patients provided blood samples for pharmacokinetic analysis, but all available data points were included in the data analysis. Wilcoxon Mann-Whitney rank-sum tests were used to compare determinations from patients treated with and without amifostine for pharmacokinetic measurements, duration of hematopoietic suppression or narcotic requirement, and changes from baseline of neurophysiological tests. Student’s t tests were used for comparison of neurological signs.

Patients.

Tables 2⇓ 3⇓ 4⇓ present comparisons of ACT subjects who received amifostine (n = 14, age range 30–62, median 45) and those who did not receive amifostine (n = 17, age range 29–56, median 42). Karnofsky score was 90 or 100 in all subjects, although the percentage of subjects at Karnofsky 90 was greater in the amifostine group (50% compared with 29%). All subjects without amifostine had stage II/III disease whereas six subjects receiving amifostine had stage IV disease. Of these six patients, two had complete response to chemotherapy, two had partial response, and two had stable disease before enrollment in the ACT plus amifostine protocol. Seventy to seventy-five percent of subjects in both groups had prior courses of either paclitaxel or docetaxel at standard doses. Of the patients with stage II-III disease, there was no significant difference in relapse-free survival between the two groups (P = 0.34; 95% confidence interval for relative hazard, 0.3–1.5) with median follow-up of surviving patients 22 months (range 1.4–33 months). Four of the 17 ACT only patients have died, although there have been no deaths to date among the eight stage II-III patients who received ACT plus amifostine. Of the six stage IV patients treated with ACT and amifostine, median progression-free survival was 18 months, although median survival is undefined at a median follow-up of 21 months (range 4–30 months). The therapeutic efficacy of ACT plus amifostine could not be determined in these six subjects because none had a change in measurable disease following high-dose therapy.

The Effect of Amifostine on Neurological Examination and Neurophysiological Tests after High-Dose Paclitaxel.

Fig. 1⇓ shows results of neurophysiological tests in 21 patients before and after high-dose paclitaxel. There was a decrease in amplitude of peroneal CMAP in all subjects treated with and without amifostine. As shown in Table 2⇓ , the mean decrease of peroneal amplitude expressed as logarithm ratio was less in the amifostine group (−1.1 compared with −2.5 without amifostine), but this difference was not statistically significant (P = 0.08). There was also a decrease in sural SNAP amplitude both in subjects with and without amifostine as well as an increase in sensory detection thresholds for vibration and cold (Fig. 1)⇓ . As shown in Table 2⇓ , the mean log ratios of sural SNAP amplitudes and the vibratory detection thresholds were essentially the same with and without amifostine. The mean increase in cold detection thresholds after paclitaxel was greater in the amifostine group, although the difference was not statistically significant. Of the 12 subjects treated without amifostine, nine had prior standard dose paclitaxel or docetaxel chemotherapy, and of the nine subjects receiving amifostine, six had prior paclitaxel or docetaxel. The degree of peroneal CMAP or sural SNAP amplitude drop-off or sensory threshold increase did not correlate with prior chemotherapy. Also, the extent of CMAP or SNAP amplitude drop-off did not predict the extent of threshold increases in quantitative sensory tests. Nerve conduction velocities and distal motor or sensory latencies were not consistently affected.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Nerve conduction tests and quantitative sensory tests pre- and post-high-dose paclitaxel with (solid lines) and without (dotted lines) amifostine. Peroneal amplitude of compound muscle action potential (CMAP), sural amplitude of sensory nerve action potential (SNAP), vibration detection threshold (VDT), and cold detection threshold (CDT).

Paired neurological examinations were performed in 13 subjects treated with ACT and amifostine and 11 subjects treated with ACT alone. All 24 subjects reported acroparesthesias after ACT. As shown in Table 2⇓ , all patients had absent ankle deep tendon reflexes after ACT. None had weakness in toes or ankles, and none had proprioceptive abnormalities in the toes. Decreased vibratory sensation (128C tuning fork) on the toes was reported in 91% of subjects without and 85% of subjects treated with amifostine, and decreased pin sensation on the feet was recorded in 45% of subjects treated without and 54% of subjects treated with amifostine.

Fig. 2⇓ shows changes in peripheral neuropathy composite scores in 17 subjects for which full clinical and neurophysiological data are available. Despite prior taxane exposure in 71%, only five subjects had pretransplant scores >0, and none had pretransplant scores >1. All subjects had an increase in peripheral neuropathy score after transplant with the maximum post-transplant score of 10 of a possible 21 on this scale. The increase in peripheral neuropathy scores was similar in patients receiving or not receiving amifostine.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Neuropathy scores pre- and post-high-dose paclitaxel. A, patients receiving doxorubicin, cyclophosphamide, paclitaxel (ACT) alone. B, patients receiving ACT and amifostine.

The Effect of Amifostine on High-Dose Paclitaxel Toxicity, Hematological Recovery, Narcotic Requirement, and Paclitaxel Pharmacokinetics.

Table 3⇓ lists percentage grade 3 or 4 toxicity in 17 patients treated with ACT alone and 14 patients treated with ACT plus amifostine. A comparable pattern of toxicity was present in both groups. Similarly, the mean times of hematopoietic cell recovery and requirement for narcotic medication for mucositis pain were not different in subjects with and without amifostine (Table 4)⇓ . The median number of days of narcotics (10 days) was the same in both groups, and the median number of days after transplant for the platelet count to exceed 50,000/μl (10 days after transplant) was also the same in both groups. The median number of days after transplant for granulocytes to exceed 500/μl was 9 or 10, and the median number of days for platelet independence was 5 or 6 for the groups treated without and with amifostine, respectively.

As shown in Table 4⇓ , there was greater variability of paclitaxel pharmacokinetics in subjects who received amifostine. Although the mean C_max value was approximately 20% higher in amifostine-treated patients, the difference was not statistically significant. Furthermore, there was no significant difference in paclitaxel AUC, clearance, or time above 0.05 μm (Table 4)⇓ .