High-dose methotrexate pharmacokinetics and outcome of children and young adults with osteosarcoma

Authors

  • Kristine R. Crews Pharm.D.,

    Corresponding author
    1. Department of Pharmaceutical Sciences, St. Judes Children's Research Hospital, Memphis, Tennessee
    2. University of Tennessee Health Science Center College of Pharmacy, Memphis, Tennessee
    • Department of Pharmaceutical Sciences, St. Jude Children's Research Hospital, Mail Stop 313, 332 N. Lauderdale, Memphis, TN 38105-2794
    Search for more papers by this author
    • Fax: (901) 525-6869

  • Tiebin Liu M.S.P.H.,

    1. Department of Biostatistics, St. Jude Children's Research Hospital, Memphis, Tennessee
    Search for more papers by this author
  • Carlos Rodriguez-Galindo M.D.,

    1. Department of Biostatistics, St. Jude Children's Research Hospital, Memphis, Tennessee
    2. University of Tennesse Health Science Center College of Medicine, Memphis, Tennesse
    Search for more papers by this author
  • Ming Tan Ph.D.,

    1. University of Tennesse Health Science Center College of Medicine, Memphis, Tennesse
    Search for more papers by this author
  • William H. Meyer M.D.,

    1. University of Tennesse Health Science Center College of Medicine, Memphis, Tennesse
    2. Department of Hematology-Oncology, St. Jude Children's Research Hospital, Memphis, Tennessee
    Search for more papers by this author
  • J. Carl Panetta Ph.D.,

    1. Department of Pharmaceutical Sciences, St. Judes Children's Research Hospital, Memphis, Tennessee
    Search for more papers by this author
  • Michael P. Link M.D.,

    1. Department of Pediatrics, Stanford University School of Medicine, Stanford, California
    Search for more papers by this author
  • Najat C. Daw M.D.

    1. University of Tennesse Health Science Center College of Medicine, Memphis, Tennesse
    2. Department of Hematology-Oncology, St. Jude Children's Research Hospital, Memphis, Tennessee
    Search for more papers by this author

Abstract

BACKGROUND

High-dose methotrexate (HDMTX) is used frequently in combination regimens that include nephrotoxic chemotherapy. The authors evaluated the impact of factors such as age and prior nephrotoxic agents on MTX pharmacokinetics in children and young adults with osteosarcoma and examined whether MTX pharmacokinetic parameters were associated with outcome.

METHODS

The authors evaluated MTX pharmacokinetics in 140 patients who were treated with 1083 courses of HDMTX on 3 consecutive studies of multiagent chemotherapy at a single institution. The influence of MTX pharmacokinetics on the outcome of 107 patients with localized disease was examined.

RESULTS

Mean peak MTX concentrations ≥ 1000 μM were achieved in 135 patients (96%). MTX clearance was decreased after cisplatin therapy (P = 0.01), after cisplatin in combination with ifosfamide therapy (P < 0.0001), and after MTX therapy (P = 0.003). In patients with localized osteosarcoma, a higher mean MTX area under the curve, a higher mean peak concentration of MTX, a longer mean time above a threshold concentration (500 μM), and a lower mean MTX clearance were associated with lower probability of event-free survival (EFS). Patients who had a mean peak MTX plasma concentration > 1500 μM were found to have a worse outcome (estimated 5-year EFS, 58.5% ± 6.7%) compared with patients who had a mean peak concentration ≤ 1500 μM (estimated 5-year EFS, 75.5% ± 6.6%; P = 0.02).

CONCLUSIONS

When HDMTX (12 g/m2) was used with multiagent therapy for patients with osteosarcoma, very high MTX exposures were associated with poorer outcome. The prospective evaluation of MTX pharmacokinetics and their relation to outcome in a large study is warranted to further substantiate the current findings and to elucidate the causative mechanism. Cancer 2004. © 2004 American Cancer Society.

Multiagent chemotherapy has improved cure rates dramatically for patients with osteosarcoma.1–4 Thus, high-dose methotrexate (HDMTX) with leucovorin rescue is used frequently in combination with nephrotoxic agents, such as cisplatin, carboplatin, and ifosfamide.1–6 Although HDMTX therapy has been associated with potentially life-threatening toxicity, the incidence of severe toxicity has decreased as a result of the monitoring of serum concentrations of methotrexate (MTX) and appropriate alterations in leucovorin dosing and hydration.7–9

Cisplatin causes dose-related renal insufficiency and has been shown to alter MTX pharmacokinetics.10–12 In a study of 11 patients with osteosarcoma, a cumulative cisplatin dose of 360 mg/m2 was associated with delayed MTX clearance;10 however, to our knowledge, the cumulative dose of cisplatin that affects MTX clearance has not been determined in a larger study. Ifosfamide is associated with clinically significant tubular and glomerular dysfunction.13, 14 Carboplatin, a second-generation cisplatin derivative, is potentially nephrotoxic, although to a lesser extent than cisplatin. The effects of cumulative doses of ifosfamide or carboplatin or the combination of these agents on MTX pharmacokinetics have not been reported to our knowledge. In addition, because MTX is a nephrotoxic agent itself, prior HDMTX therapy potentially may affect MTX pharmacokinetics.

Several studies have investigated the relation between peak serum concentrations of MTX and histologic tumor response or the disease-free survival of patients with osteosarcoma. Delepine et al. first reported that individualized MTX doses that were designed to achieve a serum MTX concentration ≥ 1000 μM at the end of a 6-hour infusion led to improved histologic response in patients with osteosarcoma.15 In a more recent study, a threshold peak concentration of 700 μM after a 6-hour infusion of HDMTX (8–12 g/m2) was associated with an improved histologic response.5 In a retrospective analysis of 198 patients in osteosarcoma Austrian-Swiss Osteosarcoma Study Group (COSS) studies COSS-80, COSS-82, and COSS-86, Graf et al. showed that a threshold concentration of 1000 μM after a 4-hour HDMTX infusion (12 g/m2) was associated with a greater probability of disease-free survival and histologic tumor response for the COSS-80 study only.12

In the current study, we report the effects of factors such as age and prior treatment with cisplatin, cisplatin in combination with ifosfamide, carboplatin in combination with ifosfamide, and HDMTX on MTX pharmacokinetics in a large group of pediatric patients who were treated for osteosarcoma. In addition, we characterize the relation between pharmacokinetic parameters of MTX and outcome.

MATERIALS AND METHODS

Patients

From records in the solid tumor data base of St. Jude Children's Research Hospital, we identified 165 patients who were enrolled on 3 consecutive studies (MIOS, OS86, and OS91) between June 1982 and July 1997 for treatment of newly diagnosed osteosarcoma. MTX pharmacokinetic and chemotherapy dosing data were available for 140 of 165 patients (Table 1). The median patient age was 14.5 years (range, 3.2–24.1 years). The 25 patients who were not included in the analysis did not differ from the group studied in terms of age, gender, primary disease site (extremity or other site), histologic tumor features, or the percentage of patients with localized disease (P > 0.12).

Table 1. Characteristics of 140 Patients with Osteosarcoma Treated on 3 Consecutive Trials at St. Jude Children's Research Hospital
CharacteristicNo. of patients (%)
Race 
 Black33 (23.6)
 White99 (70.7)
 Hispanic6 (4.3)
 Other2 (1.4)
Gender 
 Female62 (44.3)
 Male78 (55.7)
Protocol 
 MIOS36 (25.7)
 OS8648 (34.3)
 OS9156 (40.0)
Extent of disease 
 Localized and resectable107 (76.4)
 Unresectable primary5 (3.6)
 Metastatic28 (20.0)
Primary tumor site 
 Extremity126 (90.0)
 Nonextremity14 (10.0)
Tumor histologic features 
 Osteosarcoma, not further specified56 (40.0)
 Osteoblastic osteosarcoma53 (37.9)
 Chondroblastic osteosarcoma12 (8.6)
 Telangiectatic osteosarcoma7 (5.0)
 Fibroblastic osteosarcoma6 (4.3)
 Other6 (4.3)

The chemotherapy schemas for the protocols are presented in Figure 1. MIOS was a multiinstitutional study in which patients were assigned randomly to groups that either received or did not receive adjuvant chemotherapy comprised of cisplatin, doxorubicin, HDMTX, bleomycin, cyclophosphamide, and dactinomycin after surgery.1 Data for patients who refused randomization and were treated with chemotherapy on this study were included in our analyses. OS86 was an institutional study in which ifosfamide was used as up-front window therapy that was followed by HDMTX, cisplatin, doxorubicin, and additional ifosfamide.6 OS91 was an institutional study in which the same agents were used as in the OS86 study, except that carboplatin replaced cisplatin.6 In OS91, the activity of the combination of carboplatin and ifosfamide given as up-front window therapy was evaluated. The design and pharmacokinetic evaluations for each study were approved by the Institutional Review Board.

Figure 1.

Chemotherapy on studies MIOS, OS86, and OS91. S: surgery; BCD: bleomycin (15 U/m2/day for 2 days), actinomycin D (0.6 mg/m2/day for 2 days), and cyclophosphamide (600 mg/m2/day for 2 days); M: methotrexate (12 g/m2); a: doxorubicin (30 mg/m2/day for 3 days); ap: doxorubicin (50 mg/m2 on Day 1) and cisplatin (100 mg/m2 on Day 1); I: ifosfamide (1.6 g/m2/day for 5 days with mesna); AP: cisplatin (100 mg/m2 on Day 1) and doxorubicin (37.5 mg/m2/day on Days 2 and 3); IC: ifosfamide (2.65 g/m2/day for 3 days) and carboplatin (560 mg/m2 on Day 1); A: doxorubicin (25 mg/m2/day for 3 days).

Chemotherapy

All patients received HDMTX (at a dose of 12 g/m2) intravenously over 4 hours with leucovorin rescue. Hydration regimens in each protocol were as follows. Patients on the MIOS and OS86 studies received intravenous hydration fluids (4.5 L/m2) starting 6 hours before the administration of HDMTX and continuing for 24 hours after HDMTX administration was completed. Patients on the OS91 study received oral hydration fluids (2.5–3.0 L/m2) the day before HDMTX therapy and intravenous and oral hydration fluids (4.2 L/m2) starting 2 hours before the initiation of HDMTX therapy and continuing for 24 hours after HDMTX therapy was completed. In each study, patients received sodium bicarbonate for urine alkalinization. Supplemental sodium bicarbonate was administered in the OS86 and OS91 studies if the urine pH fell below 6.5; urine pH was not monitored in the MIOS study. All three protocols included specific instructions for leucovorin rescue. In the MIOS study, leucovorin was given at a dose of 15 mg every 6 hours beginning 24 hours after the initiation of the HDMTX infusion for a total of 10 doses. In the OS86 study, leucovorin (given as 15 mg/m2 per dose) was given every 6 hours starting 24 hours after the initiation of the HDMTX infusion for a total of 10 doses. In the OS91 study, leucovorin was given every 6 hours, and its administration was begun 30 hours after the initiation of HDMTX at a dose of 30 mg/m2 per dose for 2 doses, 15 mg/m2 per dose for 2 doses, 10 mg/m2 per dose for 1 dose, and 5 mg/m2 per dose for 1 dose. One additional dose of 5 mg/m2 was given 72 hours after the initiation of HDMTX infusion. Irrespective of the protocol, patients received increased leucovorin dosing for clinical toxicity or elevated MTX concentrations at specified times after infusion; leucovorin doses were increased for 24-hour MTX concentrations in excess of 10 μM in all 3 studies. Chemotherapy courses after HDMTX were scheduled to be given at 7-day intervals. The exception was Course 10 of HDMTX in the MIOS study; the next chemotherapy course was scheduled to be given 21 days after Course 10.

Pharmacokinetic Analyses

Serial blood samples were obtained at the end of the 4-hour infusion of HDMTX and at 24 hours and 48 hours after the initiation of the infusion; in addition, some patients had samples obtained at 10 hours, 16 hours, and 28 hours after the initiation of the infusion. The serum MTX concentrations were measured by a fluorescence polarization immunoassay (TDx System; Abbott Laboratories, Abbott Park, IL). The serum concentration-versus-time data for each HDMTX course were fit to a 2-compartment model using a Bayesian estimation algorithm with pediatric population priors10, 16 as implemented in ADAPTII (Biomedical Simulations Resource, University of Southern California, Los Angeles, CA). Systemic clearance was calculated by multiplying the volume of distribution of the central compartment by the elimination rate constant.17 The area under the concentration-time curve (AUC) from the beginning of infusion until 48 hours later, the serum MTX concentrations at the end of the infusion (Cmax) and at 24 hours after the beginning of the infusion (i.e., the 24-hour MTX concentration), and the time during which the MTX concentration exceeded thresholds of 700 μM and 500 μM were estimated using the concentration-versus-time data extrapolated from the estimated pharmacokinetic parameters. The mean MTX Cmax value for an individual patient was the mean of the serum concentrations at the end of each infusion. The mean MTX AUC value for each patient was the mean of the AUCs associated with each infusion. Creatinine clearance was estimated from serum creatinine concentrations as outlined by Schwartz et al.18 (for children age ≤ 17 years) or by Cockcroft and Gault19 (for men and women age > 18 years).

Statistical Analyses

A mixed-effects model for repeated measures20, 21 was used to assess whether variables such as prior treatment with the different nephrotoxic agents, the interval after HDMTX courses, or increased leucovorin dosing were associated significantly with the following pharmacokinetic parameters: MTX AUC, MTX clearance, and 24-hour MTX concentration. A mixed-effects model was used to assess whether the actual leucovorin dose administered differed by study. To assess whether the effects of prior nephrotoxic agents on MTX pharmacokinetics were independent of prior HDMTX therapy, we performed multiple regression analyses based on the mixed-effects model. Least-squares estimation was used to analyze the effect of age on each pharmacokinetic parameter. The data were log-transformed when the original data were not distributed normally. For each HDMTX course, we calculated the cumulative prior dose of cisplatin alone, cisplatin and ifosfamide, or carboplatin and ifosfamide. The number of days between one HDMTX course and the next administered course of chemotherapy was calculated. An interval > 7 days (or 21 days for Course 10 of HDMTX in the MIOS study) was considered a delay in chemotherapy.

Survival was defined as the time from the date of diagnosis to last follow-up or death due to any cause. The period of event-free survival (EFS) was defined as the time from the first HDMTX administration to the time of disease progression or recurrence, second malignancy, death due to any cause, or the last follow-up. Survival and EFS distributions were estimated by using the method of Kaplan and Meier22; associated standard errors were calculated according to the method of Peto et al.23 Factors were examined as predictors of survival and EFS through the use of the Cox proportional hazards regression model24 and the exact log-rank test; for factors with more than two subgroups, the Mantel–Haenszel test was used.25 Risk ratios and 95% confidence intervals were estimated from Cox proportional hazards regression models.

RESULTS

Serum Concentrations Achieved with HDMTX Administration

We analyzed the MTX pharmacokinetics using data collected from 140 patients who together received 1083 HDMTX courses (Table 2). A mean MTX Cmax ≥ 1000 μM was attained in 135 patients (96%). The overall mean Cmax (± standard error) for the 140 patients was 1605 μM ± 56 μM. Of the 1083 individual peak concentrations, 1013 (94%) were above the putative critical threshold of 1000 μM.12, 15 MTX Cmax was found to be correlated highly with AUC, time above a threshold concentration of 700 μM, time above a threshold concentration of 500 μM, and 24-hour MTX concentrations and was found to be correlated inversely with MTX clearance (P < 0.001). Estimated creatinine clearance was correlated positively with MTX clearance (P = 0.02) and inversely with MTX Cmax (P = 0.05).

Table 2. High-Dose Methotrexate Courses and Pharmacokinetic Parameters for Patients Enrolled on Protocol MIOS, OS86, or OS91
ParameterProtocolAll patients
MIOSOS86OS91
  1. Cmax: serum concentration at the end of infusion; HDMTX: high-dose methotrexate; MTX: methotrexate; SEM: standard error of the mean; AUC: area under the curve.

Total no. of patients364856140
No. of patients (%) with    
 Mean Cmax < 1000 μM1 (3)2 (4)2 (4)5 (4)
 Mean Cmax ≥ 1000 μM35 (97)46 (96)54 (96)135 (96)
HDMTX courses3593443801083
No. of HDMX courses (%) with    
  Cmax < 1000 μM (%)15 (4)17 (5)38 (10)70 (6)
  Cmax ≥ 1000 μM (%)344 (96)327 (95)342 (90)1013 (94)
 MTX Cmax (μM)    
  Mean1840.01623.31365.91604.8
  SEM149.763.745.456.2
 MTX AUC (μM*hr)    
  Mean11,999.49589.98374.99962.6
  SEM1438.0333.0409.5511.0
 MTX clearance (mL/min/m2)    
  Mean47.153.858.453.2
  SEM0.92.40.80.9
 24-hr MTX concentration (μM)    
  Mean13.710.211.511.8
  SEM1.20.92.61.0

Effect of Age on MTX Pharmacokinetics

The dose of HDMTX was adjusted according to age in multiple previous osteosarcoma studies.26–28 In those studies, a dose of 12 g/m2 was used for children age < 12 years, and a dose of 8 g/m2 was used for older patients. We examined the effect of age at the time of diagnosis on MTX pharmacokinetics. Analysis of age at the time of diagnosis as a continuous variable showed that age did not appear to significantly influence MTX Cmax (P = 0.3), AUC (P = 0.8), or clearance (P = 0.7) but that age was associated significantly with the 24-hour MTX concentration (P = 0.003). The 24-hour MTX concentrations in patients age ≥ 12 years were significantly greater compared with the concentrations in patients age < 12 years (P < 0.001).

Effect of Prior Cisplatin Treatment on MTX Pharmacokinetics

For patients on the MIOS study, the pharmacokinetic parameters of HDMTX given before cisplatin administration were compared with those of HDMTX given after the administration of at least 2 doses of cisplatin (cumulative cisplatin dose, 200–400 mg/m2). The MTX AUC and the 24-hour MTX concentration for courses administered after cisplatin therapy were significantly higher than those for courses administered before cisplatin therapy; MTX clearance was significantly lower after cisplatin therapy (P < 0.001). This effect on MTX clearance was observed in patients who had received a minimum cumulative cisplatin dose of 200 mg/m2 (P < 0.001) (Fig. 2). When prior HDMTX therapy was included along with prior cisplatin therapy in a multiple regression analysis, the effect of prior cisplatin therapy on MTX AUC (P = 0.003), 24-hour concentration (P = 0.03), and clearance (P = 0.01) remained significant.

Figure 2.

Methotrexate (MTX) clearance estimates for high-dose MTX Courses 1–8 (before cisplatin) on the MIOS study and for Courses 9 and 10 after cisplatin (cumulative dose, 200 mg/m2) administration. Boxes delimit the first and third quartiles (median inside); bars represent the minimum and maximum values.

Effect of Cumulative Doses of Ifosfamide and Cisplatin on MTX Pharmacokinetics

In the OS86 study, 3 courses of ifosfamide (cumulative ifosfamide dose, 24 g/m2) were administered before the first 3 HDMTX courses (Fig. 1). After the third course of HDMTX, cisplatin (100 mg/m2) and a fourth course of ifosfamide (8 g/m2; cumulative ifosfamide dose, 32 g/m2) were given. For HDMTX courses administered after ifosfamide (32 g/m2) and cisplatin (100 mg/m2), the AUCs were significantly higher (P < 0.001), and the MTX clearance was lower (P = 0.05) compared with those values for HDMTX courses administered after ifosfamide (24 g/m2) without cisplatin had been given (Fig. 3). When prior HDMTX therapy was included along with cumulative ifosfamide and cisplatin doses in a multiple regression analysis, the effect of prior ifosfamide (32 g/m2) and cisplatin (100 mg/m2) therapy on MTX clearance remained significant (P < 0.0001).

Figure 3.

Methotrexate (MTX) clearance estimates for high-dose MTX Courses 1–3 on the OS86 study after ifosfamide (cumulative dose, 24 g/m2) and for Courses 4–6 after ifosfamide (cumulative dose, 32 g/m2) and cisplatin (cumulative dose, 100 mg/m2). Boxes delimit the first and third quartiles (median inside); bars represent the minimum and maximum values.

Effect of Cumulative Doses of Ifosfamide and Carboplatin on MTX Pharmacokinetics

In the OS91 study, 3 courses of ifosfamide and carboplatin (cumulative ifosfamide dose, 24 g/m2; cumulative carboplatin dose, 1680 mg/m2) were administered before the first 3 HDMTX courses (Fig. 1). After the third course of HDMTX, a fourth course of ifosfamide and carboplatin was given (cumulative ifosfamide dose, 32 g/m2; cumulative carboplatin dose, 2240 mg/m2). MTX clearance was not altered after this fourth course (P > 0.48); however, MTX clearance was found to be decreased significantly after a fifth course of ifosfamide and carboplatin (cumulative ifosfamide dose, 40 g/m2; cumulative carboplatin dose, 2800 mg/m2; P = 0.02). When prior HDMTX therapy was included along with cumulative ifosfamide and carboplatin doses in a multiple regression analysis, the effect of prior ifosfamide and carboplatin therapy on clearance was no longer significant (P = 0.94).

Effect of Cumulative Doses of HDMTX on MTX Pharmacokinetics

Because HDMTX Courses 1–8 were given before any other nephrotoxic agents in the MIOS study, we were able to analyze the effect of repeated HDMTX administration on the pharmacokinetics of MTX for these courses. As the number of prior HDMTX courses received increased, MTX AUC increased (P = 0.008) and MTX clearance decreased (P = 0.003).

MTX Pharmacokinetics and Patient Outcome

Patients with localized and resectable primary tumors (n = 107 patients) were included in the survival analyses. Patients with metastatic disease were excluded from these analyses because metastatic disease is a poor indicator of prognosis, and separate survival analyses for patients with metastatic disease were not performed because of the small number of these patients (n = 28 patients). The number of HDMTX courses included in the survival analyses was 911. Fifty patients (46.7%) were female and 79 patients (73.8%) were white. The median age at the time of initial diagnosis of osteosarcoma was 14.5 years (range, 3.2–23.6 years). The most common tumor histologies were osteosarcoma, not further specified (n = 45 patients; 42.1%) and osteoblastic osteosarcoma (n = 40 patients; 37.4%); the tumor histologies for the remaining patients included chondroblastic osteosarcoma, fibroblastic osteosarcoma, and telangiectatic osteosarcoma. The median length of follow-up since the time of diagnosis for all survivors was 11.1 years (range, 3.6–18.5 years), and 95% of survivors had been contacted within 1 year from the start of the survival analyses. Forty patients (37.4%) experienced an event (disease recurrence or progressive disease, second malignancy, or death) after the first course of HDMTX; only 1 of those 40 patients developed a second malignancy. The median time from the first course of HDMTX to the first event was 1.6 years (range, 0.1–14.2 years). The 5-year survival estimate for the 107 patients was 73.7% ± 4.5%, and the 5-year EFS estimate was 66.8% ± 4.8% (Fig. 4). The EFS and survival probabilities for the 107 patients did not differ by treatment protocol (P = 0.53 and P = 0.69, respectively) (Fig. 5). In addition, differences with regard to age, gender, primary tumor site (extremity or other site), and tumor histologic features did not appear to affect EFS and survival probabilities for these patients (P > 0.1).

Figure 4.

The probability of survival and event free survival for 107 patients who had localized osteosarcoma and were enrolled on the MIOS, OS86, or OS91 study. The estimates were calculated using the Kaplan–Meier method.

Figure 5.

The probability of (A) survival and (B) event free survival by treatment protocol for 107 patients with localized osteosarcoma.

All 107 patients were included in the analyses of MTX pharmacokinetics and survival. Unexpectedly, the mean Cmax value was found to be related inversely to EFS. Patients who had mean Cmax values > 1500 μM (the median of the mean Cmax values for individual patients) fared worse compared with patients who had mean Cmax values ≤ 1500 μM (P = 0.02) (Fig. 6). Similar relations were observed for mean AUC, time during which the MTX concentration exceeded the threshold of 700 μM, and time during which the MTX concentration exceeded the threshold of 500 μM, when each group was dichotomized on the basis of the median value of each parameter (Table 3). Lower mean MTX clearance was associated with a lower probability of EFS (P = 0.02). Patients who received > 3 HDMTX courses with Cmax values ≥ 1500 μM had a worse outcome (estimated 5-year EFS, 57.7 ± 6.7%) compared with patients who received ≤ 3 such courses (estimated 5-year EFS, 74.2% ± 6.6%; P = 0.04). The mean estimated creatinine clearance at the start of each HDMTX course was not found to be associated with outcome (P = 0.60).

Figure 6.

The probability of event-free survival in patients with osteosarcoma who had mean serum methotrexate (MTX) concentrations at the end of the infusion (Cmax) > 1500 μM and for patients who had mean MTX Cmax values ≤ 1500 μM.

Table 3. Association between Methotrexate Pharmacokinetic Parameters and Event-Free Survival Probability
ParameterNo. of patientsP valueaFive-yr EFS probability (%)RR (95% CI)
  • EFS: event-free survival; RR: relative risk; 95% CI: 95% confidence interval; Cmax: serum concentration at the end of infusion; MTX: methotrexate; AUC: area under the curve.

  • a

    P values were obtained by Cox proportional hazards regression analysis.

  • b

    Reference group.

Mean MTX Cmax    
 ≤ 1500 μMb5475.5 ± 6.62.2
 > 1500 μM530.02058.5 ± 6.7(1.1–4.2)
Mean log AUC    
 ≤ 9.07b5480.9 ± 6.02.5
 > 9.07530.00653.5 ± 6.8(1.3–5.0)
Mean time above 700 μM    
 ≤ 5.25 hrb5977.4 ± 6.02.2
 > 5.25 hr480.01354.2 ± 7.1(1.2–4.2)
Mean time above 500 μM    
 ≤ 6.78 hrb5578.0 ± 6.22.0
 > 6.78 hr520.03055.6 ± 6.9(1.1–3.9)
Mean clearance    
 ≤ 51.5 mL/min/m2b5455.6 ± 6.70.5
 > 51.5 mL/min/m2530.02479.1 ± 6.3(0.2–0.9)
Mean 24-hr MTX concentration    
 ≤ 7.02 μMb5369.6 ± 6.91.1
 > 7.02 μM540.7164.4 ± 6.7(0.6–2.1)

To determine whether MTX pharmacokinetic factors were associated with a delay in the delivery of scheduled chemotherapy, we evaluated the relation between pharmacokinetic parameters and the number of days between the end of one HDMTX infusion and the start of the next course of chemotherapy. The number of days between each HDMTX course and the next chemotherapy administration was correlated positively with 24-hour MTX concentration for individual HDMTX courses (P < 0.001) and with the time during which the MTX concentration exceeded a threshold of 500 μM (P = 0.01).

The actual leucovorin dose administered per course did not differ by treatment protocol (P = 0.63). We evaluated the relation between MTX pharmacokinetics and leucovorin dose. HDMTX courses after which the leucovorin dose was increased over the standard, protocol-specified dose had significantly higher Cmax (P = 0.003), AUC (P < 0.0001), time above a threshold concentration of 700 μM (P < 0.0001), time above a threshold concentration of 500 μM (P < 0.0001), and 24-hour MTX concentration (P < 0.0001), and lower clearance (P < 0.0001) compared with courses that were followed by the standard leucovorin dose.

A Cmax value ≥ 1000 μM was attained in 856 of the 911 courses (94%) of HDMTX administered to patients with localized disease. Only 2 patients who were included in the survival analyses had mean MTX Cmax values < 1000 μM. One of those 2 patients was alive and event free 14.1 years after diagnosis; the other patient had progressive disease 6.6 months after the first course of HDMTX and died 9.9 months after diagnosis.

DISCUSSION

The current study findings show that a MTX dose of 12 g/m2 infused over 4 hours consistently resulted in a Cmax value ≥ 1000 μM in children and young adults with osteosarcoma (mean peak concentrations ≥ 1000 μM were achieved in 96% of patients). Cmax was found to be correlated highly with other measures of MTX systemic exposure, including AUC, time during which the MTX concentration exceeded 700 μM, time during which the MTX concentration exceeded 500 μM, and 24-hour MTX concentration; Cmax was found to be correlated inversely with clearance. Factors that affected MTX pharmacokinetics were prior cisplatin therapy, prior cisplatin in combination with ifosfamide therapy, and prior HDMTX therapy. Minimum cumulative doses identified as affecting MTX clearance were 200 mg/m2 for cisplatin alone, 32 g/m2 for ifosfamide alone, and 100 mg/m2 for cisplatin in combination. The effects of prior cisplatin therapy and prior cisplatin in combination with ifosfamide therapy on MTX clearance were independent of the effect of prior HDMTX therapy.

The current study finding that cisplatin affects MTX pharmacokinetics is consistent with the report of Graf et al. that prior cisplatin therapy is associated with higher MTX exposures; however, those authors did not detect a significant influence of the addition of ifosfamide to cisplatin on MTX pharmacokinetics.12 Because of the chemotherapy schedules of the protocols, we were unable to assess the effect of doses of cisplatin l< 200 mg/m2 (in particular, the effect of 100 mg/m2 of cisplatin alone) or the single-agent effect of ifosfamide or carboplatin on MTX pharmacokinetics within individual patients. Therefore, it remains to be determined whether the addition of ifosfamide to cisplatin (100 mg/m2) contributed to the observed effect on MTX pharmacokinetics or whether this effect would be noted with this dose of cisplatin alone.

Conflicting reports exist regarding the relation between age and MTX pharmacokinetics. Wang et al. reported that Cmax values and 24-hour MTX concentrations after a 6-hour infusion of HDMTX in young patients (ages 5–17 years) with osteosarcoma were lower compared with those in adult patients (ages 18–70 years); the authors attributed their findings to a greater volume of distribution and a shorter early half-life in children.26 In contrast, Graf et al. found no age effect on MTX pharmacokinetics.12 In the current study, age did not appear to affect MTX Cmax or clearance, but older age (≥ 12 years) was associated with higher 24-hour MTX concentrations. This finding suggests that, within the adolescent and young adult population studied, patients age ≥ 12 years may be at risk of elevated 24-hour MTX concentrations and should be monitored appropriately for toxicity. Adjusting the dose of HDMTX according to age was done in previous osteosarcoma studies26–28 but not in more contemporary studies. Our results suggest that such routine adjustment of HDMTX dose is not necessary.

We examined the effect of MTX pharmacokinetics on EFS and survival probability. In our comparison of the EFS and survival probabilities of patients from the three studies, we calculated EFS and survival estimates only for patients who were enrolled at a single institution of the multiinstitutional MIOS study. Although the EFS estimate for this subgroup of patients enrolled at St. Jude (5-year EFS estimate, 60.0% ± 8.1%) does not necessarily reflect the EFS estimate for all patients on the MIOS study (6-year EFS estimate, 61%),29 the results appear comparable. It is interesting to note that the EFS and survival estimates for patients with localized osteosarcoma who were treated at St. Jude on the 3 studies during a period of approximately 15 years did not differ by treatment protocol. This lack of difference suggests that the addition of ifosfamide to a regimen containing cisplatin, doxorubicin, and HDMTX does not result in survival benefit. Conversely, the substitution of carboplatin for cisplatin in a regimen containing ifosfamide, doxorubicin, and HDMTX does not appear to result in lower survival probability.6

Many reports have evaluated age as a prognostic factor in the treatment of osteosarcoma. Early studies revealed that young age indicated a poor prognosis.15, 30–33 However, those studies included MTX doses that were lower than those used in contemporary regimens. Since the use of HDMTX (12 g/m2) has become widespread, several studies have reported no difference in the outcomes of patients of various ages.15, 27, 32, 34–39 Our study confirmed that age is not associated with outcome.

Because a mean peak MTX concentration of 1000 μM was achieved in nearly all patients in our survival analysis, we could not examine the relation between this previously identified threshold peak concentration of MTX and outcome. Overall, 96% of patients had a mean MTX Cmax value ≥ 1000 μM. This percentage corresponds to that reported by Graf et al. for the COSS-86 study, in which this threshold was achieved in 94% of patients who received cisplatin-based chemotherapy and restricted fluid hydration (3 L/m2) over 24 hours after HDMTX administration.12

For patients with localized osteosarcoma (in almost all of whom a mean serum MTX concentration of 1000 μM was achieved at the end of the HDMTX infusions), EFS probability was associated with mean MTX clearance and was associated inversely with mean MTX AUC, mean Cmax value, and mean time during which the MTX concentration exceeded 500 μM. It is interesting to note that patients who had mean Cmax values that exceeded 1500 μM fared worse than patients who had mean Cmax values < 1500 μM. These findings were counterintuitive, and the causative mechanism for the association between very high MTX exposures and lower survival estimates is unclear. One possible explanation is that prolonged exposure to high MTX concentrations or poor renal function reflected by high MTX concentrations may have led to decreased dose intensity of other active agents in the combination regimens. A detrimental effect of decreased dose intensity is supported by the results of the first European Osteosarcoma Intergroup study, which showed that the disease-free survival probability for patients who were treated with cisplatin and doxorubicin was greater compared with that for patients who received cisplatin and doxorubicin at a reduced dose intensity in combination with HDMTX (8 g/m2).40 In our study, higher 24-hour MTX concentrations and longer times during which the MTX concentration exceeded 500 μM were associated with a delay in the delivery of scheduled chemotherapy. Another possible explanation is that increased leucovorin dosing in patients with very high MTX exposures may have compromised the antitumor effect of MTX.41, 42 In the current study, higher MTX Cmax values and MTX exposures were associated with an increase in leucovorin dosing.

In pediatric patients with acute leukemia, it has been shown that the efficacy and toxicity of MTX are related to serum concentrations and to exposure time. Because little is known regarding the optimum systemic exposure to MTX for the treatment of osteosarcoma, we evaluated the following as possible predictors of outcome: AUC and time during which the MTX concentration remained above certain threshold concentrations. The time during which the MTX concentration exceeded 500 μM was found to be correlated inversely with EFS estimates and was correlated positively with a delay in administration of the next scheduled course of chemotherapy. Because the time during which MTX concentration exceeded 500 μM was correlated strongly with MTX Cmax, it may be more practical to measure and report Cmax as a surrogate marker for this time above a threshold concentration of 500 μM in individual patients. Our results suggest that, when HDMTX (12 g/m2) is given in the context of multiagent chemotherapy that includes cisplatin, pharmacokinetic monitoring is not necessary to ensure the achievement of a threshold MTX peak of 1000 μM. However, pharmacokinetic monitoring for the individualization of HDMTX doses to avoid peak MTX concentrations ≥ 1500 μM may be warranted.

The results of the current study demonstrating a correlation between MTX pharmacokinetics and patient outcome were consistent for all the pharmacokinetic parameters studied, and have important implications for the treatment of patients with osteosarcoma, the majority of whom are given HDMTX-containing regimens. After the delivery of HDMTX, higher mean Cmax concentrations, higher exposures, and lower mean clearance of MTX were associated with poorer outcome. The current study findings underscore the importance of incorporating careful pharmacokinetic monitoring into future osteosarcoma treatment protocols. Prospective analysis of MTX pharmacokinetics in a large cohort of patients is warranted to substantiate our findings further and to elucidate the causative mechanism by which very high MTX exposures are associated with poorer outcome.

Acknowledgements

The authors thank Mickey A. Cain, Nancy Kornegay, and Debbie L. Poe for data management and Julia Cay Jones for scientific editing of the article.

Ancillary