- Email: [email protected]
Carboplatin versus cisplatin in solid tumors: An analysis of the literature
Annals of Oncology 9: 13-21, 1998. © 1998 Kluwer Academic Publishers. Printed in the Netherlands.
Review Carboplatin versus cisplatin in solid tumors: An analysis of the literature J. Lokich & N. Anderson The Cancer Center of Boston in Boston, Plymouth and Framingham, MA, USA
or equivalent to carboplatin in therapeutic efficacy in all five tumors but was associated with an increased toxicity profile for Background: Introduced into clinical usage in 1992 as a plati- gastrointestinal, renal and neurologic effects. Conclusions: For some tumors, cisplatin appears to be num analogue with a distinctively different toxicity profile from cisplatin, carboplatin has become a commonly preferred superior to carboplatin in terms of therapeutic effectiveness agent over cisplatin. The comparative therapeutic efficacy of (germ cell tumors, bladder cancer, head and neck cancer), the two agents remains controversial however, prompting an while for others, effectiveness is comparable (lung cancer, analysis of the phase III trials in ovarian cancer and other ovarian cancer). Toxicity profiles are distinctly different for the two analogues however, generally favoring carboplatin. The tumors in which the two were compared. Methods: Clinical trials comparing carboplatin with cis- issue of potential carboplatin underdosing related to the lack platin, both as single agents and in combination with other of physiologic dose calculations (utilizing the AUC [area under agents, were analyzed within the tumors for which platinum the curve] method) in the comparative trials of cis- versus has become a standard or commonly employed agent. A Med- carboplatin is probably not clinically important since a dose line search identifying the randomized trials and references response effect has not been established for carboplatin or for cisplatin. The selection of the optimal platinum analogue to be from these reports were collated for analysis. Results: The original clinical comparative trials as well as employed is dependent on the type of tumor, the treatment literature reviews and commentaries were reviewed. Five solid intention (palliative vs. curative) and the other component tumors were identified within which comparative trials had drugs being used in combination. been conducted: ovary, 10 trials; lung, 2 trials; head and neck, 2 trials; germ cell tumors, 3 trials and 1 trial in bladder cancer. Depending upon the end point selected, cisplatin was superior Key words: carboplatin, cisplatin, randomized trials Summary
Introduction
Platinum analogues have become the mainstay of treatment for many tumors including ovarian cancer, lung cancer (both non-small-cell and small-cell), germ cell tumors, head and neck cancer, bladder cancer and to a lesser degree breast cancer and gastric cancer. Cisplatin was the initial platinum analogue that was introduced into clinical practice with a toxicity profile characterized by nausea and vomiting, renal dysfunction and neuroand ototoxicity. Carboplatin was the second clinically important platinum analogue developed with a substantially different toxicity profile which included bone marrow suppression (particularly thrombocytopenia) and a lesser incidence of gastrointestinal, renal and neurotoxicity. Carboplatin is commonly considered to be the therapeutic equivalent of cisplatin while offering the additional benefit of reduced gastrointestinal effects permitting an ease of outpatient administration and obviating the requirement for hydration and anti-emetic therapy. Therapeutic activity for carboplatin has been identified in all the tumors for which cisplatin has had a role with 'comparable' therapeutic effects reported in germ cell tumors [1], ovary [2], bladder cancer [3], small-
cell and non-small-cell lung cancer [5], head and neck cancer [6] and mesothelioma [7]. In addition, antitumor activity for carboplatin has been established in resistant acute leukemia [8]. In 1994, Ruckdeschel stated that "carboplatin (rather than cisplatin) is recommended for chemotherapy for palliative or non curative intent" in a non protocol clinical arena due to its "more favorable toxicity profile and lower overall cost" [9]. On the other hand, Comis has stated that "cisplatin currently (1994) remains the recommended agent of choice for treatment of several malignancies" [10]. Cisplatin is being replaced in common clinical practice by carboplatin and the toxicologic and pharmacologic differences between the two drugs may justify such a substitution provided therapeutic efficacy is not compromised. The initial studies comparing cisplatin and carboplatin in ovarian cancer suggested that the agents were comparable therapeutically although the observation was not consistent across all trials. In contrast, randomized trials comparing carboplatin and cisplatin in combination regimens in nonsmall-cell lung cancer, bladder cancer, head and neck cancer and other tumors have indicated that cisplatin is superior to carboplatin in some instances.
14 We undertook a review of randomized trials comparing carboplatin and cisplatin published through 1997 in a spectrum of solid tumors using the methodological guidelines suggested by Weed [11] for a literature review paper. Our specific purpose was to determine if the literature supported a superiority of one analogue over the other or an equivalency of clinical efficacy of the two analogues. We used a Medline search to identify all English language published papers describing a clinical trial comparison of cisplatin and carboplatin as well as reference lists from those papers. Only papers which represented randomized trials were selected for analysis with the criteria for evaluating the quality of studies based upon sufficient patient numbers to validate a statistical analysis of the data base. The numbers of published clinical trials within each tumor category selected were limited in some tumors eliminating the possibility of a meta analysis and necessitating a narrative analysis and summary. Pharmacologic and toxicologic comparison of cisplatin and carboplatin Cisplatin and carboplatin demonstrate the identical mechanism of cell kill, functioning as non classical alkylating agents by binding to cellular DNA to form cross links. The quantitative level of platinum DNA adduct formation correlates with the degree of cell killing. Knox et al. demonstrated that equivalent tumor cytotoxicity is dependent upon achieving an equivalent level of DNA binding [12]. Carboplatin has a greater degree of chemical stability than cisplatin resulting in a lesser reactivity with DNA; as a consequence, a higher dose of carbobplatin is necessary to obtain comparable anti-tumor effects in experimental systems. In in vivo experimental anti tumor systems, carboplatin is 8 to 45 times less potent than cisplatin although the spectrum of experimental tumor system activity is similar for the two agents [13]. The therapeutic dose of carboplatin relative to the therapeutic dose of cisplatin has generally been described as a ratio of 4:1 (400-500 mg/m2 versus 100 mg/m 2 ) based on clinical studies in ovarian cancer in which these doses achieved a relative equivalency. The toxic effects of carboplatin are more precisely predicted utilizing the formula developed by Calvert et al. [14] and subsequently modified by Jodrell et al. [15]. In this formula, a specific area under the curve (AUC) is established as a goal for the carboplatin plasma concentration related to the creatinine clearance since 90% of a carboplatin dose is excreted by the kidneys. An optimal single agent carboplatin AUC of 5-7 mg/ml x minutes has been suggested as a basis for optimal dose calculation. Most prospective randomized trials comparing carboplatin and cisplatin have employed a fixed dose for carboplatin delivery and have not used the Calvert formula. This represents an important consideration in comparing the two agents within such
Table 1 Comparison of pharmacology and toxicity patterns of carboplatin and cisplatin.
Pharmacology Metabolism Excretion Kinetics Tl/2Beta Tl/2 Gamma Toxicity3
a b
Cisplatin
Carboplatin
Inactivated by sulfhydryl groups 25% excreted; 9:1 urine: bile
No significant metabolism 90% renal recovery
60 minutes 24 hours Gastrointestinal Renal Insufficiency Neuropathy
80-100 minutes 22-40 hours Myelosupressionb
In order of severity and dose limiting effect (see text). Characteristically delayed with nadir at day 21 to 28.
trials and is discussed in a subsequent section. The pharmacology and toxicity patterns of carboplatin and cisplatin are described in Table 1. A major difference in the two analogues is in their metabolism and excretion. In vivo, carboplatin demonstrates no significant catabolism while cisplatin is inactivated by sulfhydryl groups. Generally, 90% of an injected dose of carboplatin is recovered in the urine while only 25% of cisplatin is excreted unaltered. The general pharmacokinetic pattern as reported by Reed et al. [13] indicates that cisplatin and carboplatin are relatively similar. However, there is less protein binding of carboplatin and more sophisticated pharmacologic studies separating ultra filterable platinum (UF) and total platinum reveal that carboplatin has a longer half-life of UF platinum with a T 1/2 beta of 120 minutes versus 36 minutes for cisplatin [16]. Nonetheless, the terminal half-life is comparable for the two drugs at 5.8 and 5.4 days. With regard to toxicity, carboplatin is less nephrotoxic and less emetogenic than cisplatin and neurotoxicity and ototoxicity are virtually absent. Myelosuppression is the major toxic effect of carboplatin and combining carboplatin with other cytotoxic agents may be complicated since the pattern of myelosuppression, particularly thrombocytopenia may be delayed to a nadir at day 21 to 28. In contrast, the major toxicity for cisplatin has been nausea and vomiting and generalized gastrointestinal effects including post platinum diarrhea. For the most part, this pattern of toxicity is minimized with the use of serotonin antagonists and with dose fractionation of the cisplatin. Hydration and dose fractionation mitigate most of the nephrotoxic effect of the cisplatin. The neuropathic effects for cisplatin are relatively common and are related to the cumulative dose administered. The combination of simultaneous administration of carboplatin and cisplatin has been employed in a number of clinical trials [17-19], based upon the fact that the toxicity profiles are dissimilar for the two agents and since tumor cell killing is dependent upon maximizing DNA adduct formation, higher total platinum doses
15 have been postulated to yield higher levels of cell killing. Three studies combining cis- and carboplatin delivered a maximum carboplatin dose of 280, 300 and 350 mg/ m2 with the paired cisplatin dose of 75,100 and 150 mg/ m2 respectively. No significant increase in tumor response was observed in spite of the substantial increase in UF platinum in studies in ovarian cancer [19] and head and neck cancer [18]. The logistical complexities of these double platinum regimens combined with the lack of an important therapeutic advantage observed has discouraged further study of the analogue combination in spite of the interesting if not compelling rationale for its use. Clinical trials The tumors for which platinum analogues have become standard chemotherapy broaches a wide spectrum. Platinum is part of the first line chemotherapy regimen for five solid tumors including: ovarian cancer, germ cell tumors, head and neck cancer, both small-cell and nonsmall-cell lung cancer and bladder cancer. Additionally, selected tumors of the gastrointestinal tract (specifically esophageal, gastric and anal cancers) are commonly treated with platinum regimens although the role of 5-fluorouracil is arguably more important. Cisplatin also is active in breast cancer although it is infrequently utilized as first or second line therapy. Randomized trials comparing cis- and carboplatin have been reported for the five common tumors for which a platinum analogue is routinely administered as part of first line therapy.
Table 2. A comparison of response rates to cis- or carboplatin single agent or combination chemotherapy regimens as first line therapy in ovarian cancer. Study [reference]
No. patients
Response rate %
Dose mg/m Carbo
Cis
Carbo
Cis
Single agent Taylor [20] Adams et al. [21] Mangioni et al. [22]
131 88 173
400 400 400
100 100 100
38.4 67.5 57.3
53.8 47.5 71.6
Combination studies NCI Canada 1992 [23] EORTC 1988 [24] SWOG 1992 [25] Italy 1991 [26] ARTAC 1992 [27] Mayo Clinic 1989 [28] Netherlands 1997 [29]
417 342 291 164 144 103 190
300 350 300 200 300 150 AUC5
75 100 100 50 75 60 75
59 36.2 61 62 47 No datsi Equal
57 45.9 52 66 73 a
' P = 0.008.
comparable for the two analogues. No difference in median or progression free survival was observed between the arms in any of the three trials but the data are complicated by the fact that crossover to the alternative platinum analogue was permitted in all trials. In the seven combination chemotherapy studies involving cis- versus carboplatin, the alkylating agent cyclophosphamide was employed in six in which two trials added an anthracycline. Paclitaxel was the companion agent in one trial. The dose of carboplatin was 150 mg/m2 in the Mayo Clinic study (although dose escalation was employed as part of the design) but was 300 mg/m2 or more in four studies and one utilized the Ovarian cancer trials AUC method. The seven studies involved almost 1700 Ten randomized trials comparing cis-and carboplatin in patients, predominantly with suboptimal residual tuthe treatment of ovarian cancer have been reported mor. In two of the studies [26, 27], the pathologic between 1989 and 1997 [20-29] (Table 2) with the most complete response rate for cisplatin was superior to that recent review and analysis of these trials appearing in of carboplatin but was not statistically significant. Over1993 [29]. A meta analysis of chemotherapy in ovarian all survival and response rates were statistically equivacancer [30] reported in 1991 and a general review of the lent for cis- and carboplatin in all trials with two excepsame topic in 1993 [31] identified 11 trials comparing cis- tions. The Mayo Clinic trial which was closed to patient and carboplatin but many of the trials had limited follow entry after an interim analysis demonstrated a statistiup at that time and two of the eleven were either cally significant disease free survival advantage for the unpublished or had limited patient numbers. cisplatin arm [28]. This trial also employed a strategy Of the 10 now mature trials, three involved a com- of dose equivalency based on toxicity for the cis- and parison of single agent carbo and cisplatin in previously carboplatin arms. In this trial, 103 patients were ranuntreated patients one of which was updated in 1994. domized and the interim analysis revealed a median Almost 400 patients were entered in the three single progression free survival of 17 months for the cisplatin agent studies which were carried out predominantly in containing arm versus 12 months for the carboplatin patients with suboptimal residual disease. The carbopla- containing arm (P = 0.04). Median overall survival was tin dose was fixed in all three studies with a ratio of 4:1 also superior for the cisplatin arm (27 vs. 20 months) to the cisplatin dose. Although not statistically signifi- but was not statistically significant (P - 0.14). The cantly different, the overall response rate was higher in carboplatin patients on therapy at time of study closure the cisplatin arms in two of the three studies and in the were offered crossover to cisplatin therapy and within third study the overall response rate was higher in the this group of 21 patients, the survival pattern reverted to carboplatin arm. Pathologic complete response rates that of patients receiving cisplatin as their initial therwere similar and the median response durations were apy.
16 The other trial demonstrating a significant therapeutic difference between the analogues and favoring cisplatin was reported by Belpomme et al. in abstract form for the French ARTAC ovarian cancer study group [27]. The pathologic CR and overall response rate was higher for the cisplatin containing arm 33% CR and 73% PR versus 15% and 47% (P - 0.008) respectively for the carboplatin arm. In addition, median survival was different and statistically significant 27.9 months vs. 20.6 months (P - 0.028) favoring cisplatin. This trial has not as yet been reported in a peer reviewed journal. Three of the remaining five trials of combination chemotherapy comparing cis- and carboplatin are remarkably similar across the trials in terms of response rates (59%, 61%, 62% RR for carboplatin and 57%, 52%, 66% RR for cisplatin) with apparent equivalency for the analogues. One early trial [24] resulted in lower overall response rates (36% and 45%) and cisplatin yielded the higher respone rate but the difference was not statistically significant. A preliminary report from the Netherlands has reported on a comparison of paclitaxel 175 mg/m2 with either cisplatin 75 mg/m2 or carboplatin to an AUC of 5 [29]. This trial does not have the experimental design flaws or problems of the ongoing GOG trial (vide infra) and is reported with an interim analysis only but efficacy was not significantly different between the two arms. Another method of comparing cis- and carboplatin is through an analysis of data from trials in which patients are treated in sequence with the alternative platinum analogue after developing refractory or resistant disease to one analogue. Thus, if a patient is treated first with cisplatin and develops resistance (or becomes refractory), the second line or crossover treatment is carboplatin and vice versa. In fact, crossover was allowed in most of the randomized trials reviewed. Two reports have appeared in which carboplatin or cisplatin was either the rechallenging agent or the crossover agent [32, 33]. In the study from the Royal Marsden Hospital, 15 of 43 crossover patients responded (34%) as opposed to a 9% response rate in rechallenged patients i.e., the same platinum analogue was used upon relapse [32]. Ten patients achieved a complete response to carboplatin after cisplatin but there was no statistically significant difference in response rate according to which analogue was administered first. In the other crossover study, 57 patients were analyzed; 24 received cisplatin and 33 received carboplatin as the crossover platinum therapy [33]. Response rates were higher with cisplatin (25% vs. 9% for carboplatin) and crossover responses were observed only with cisplatin as second line in patients not responding to first line platinum therapy (3 of 12 for cisplatin vs. 0 of 11 for carboplatin). Most of these randomized trials of single agent and combination chemotherapy were completed prior to the advent of new agents and effective second-line therapy for ovarian cancer including paclitaxel and topotecan. Thus, the state-of-the-art for the treatment of ovarian cancer in 1997 has changed substantially with more
recent trials indicating that intraperitoneal cisplatin (with i.v. cyclophosphamide) is superior to intravenous cisplatin in terms of survival [34] and that paclitaxel plus cisplatin is superior to cyclophosphamide plus cisplatin [35]. In fact, the future direction of chemotherapy for ovarian cancer is being guided by phase I and phase II trials investigating the three drug combination of paclitaxel, cyclophosphamide and cisplatin [36] with cisplatin perferred over carboplatin in such trials because of its reduced marrow suppression but up front combinations including topotecan may preferentially use carboplatin because of the renal effect of cisplatin. A definitive conclusion with regard to the question of cis- vs. carboplatin in ovarian cancer may still be elusive. Markman has suggested that carboplatin and cisplatin well may be therapeuticaly equivalent in .^optimal residual disease but that does not necessarily extend to optimal residual disease [37]. In the randomized trials listed in Table 2, the majority of patients had suboptimal residual tumor. The ongoing Gynecologic Oncolcogy Group (GOG) trial in optimal residual stage III ovarian cancer compares cisplatin with carboplatin, the latter utilizing an AUC of 7.5. Paclitaxel is added to both arms with different doses and infusion times 135 mg/m2 over 24 hours versus 175 mg/m2 over three hours. This trial may permit a definitive comparison of cis- and carboplatin but the use of different paclitaxel doses and schedules may create problems for the final analysis.
Non-ovarian cancer trials
In the four solid tumors for which platinum has become a standard component of combination chemotherapy, eight studies have been devoted to comparing cisplatin and carboplatin in a randomized trial design (Table 3). Carboplatin dosage varied from 250 mg/m2 to 500 mg/ m2 as a fixed dose depending upon the type of tumor or the combination of drugs employed but two trials did use an AUC guided carboplatin dose. Head and neck cancer
Two prospective randomized trials address the issue of cisplatin versus carboplatin in head and neck cancer [38, 39]. The study from the Southwest Oncology Group (SWOG) involved 277 patients with the companion drug being 5-fluorouracil (5-FU) administered as a 96-hour infusion. Single agent methotrexate was employed as a control arm. Ototoxicity and renal toxicity were substantially greater in the cisplatin arm compared to the carboplatin arm and uncharacteristically, hematologic toxicity was also greater in the cisplatin group. In spite of the greater response rate in the cisplatin group (32% vs. 21%), the response duration and median survival was similar for both cisplatin and carboplatin arms as well as the single agent methotrexate arm. The second study in head and neck cancer comparing cisplatin and carbo-
17 Table 3. Carboplatin versus cisplatin in lung, head and neck, bladder and germ cell tumors and lung cancer. Study [reference]
Head and neck cancer Andres et al. [38] Forestiere et al. [39]
No. patients
Dose mg/m 2
Response rate %
Carbo
Cis
Carbo
Cis
95 277
400 300
100 100
96 32
76 21
m2; however, it would seem unlikely that in the context of this particular combination of drugs that carboplatin dose escalation would be possible in view of the 58% incidence of leukopenia. Germ cell tumors
Platinum chemotherapy represented a major advance in the treatment and curability of germ cell tumors from 57 71 70 41 250 the time of its initial introduction into clinical use. Germ cell tumor Platinum based combination chemotherapy regimens Bajorinset al. [41] 500 100 90 265 for this tumor have included VPB (adding vinblastine Bokemeyer et al. [42] 76 81 54 AUC 5 100 and bleomycin), VIP (adding etoposide and ifosfamide) Horwich et al. [43] 87 44 598 AUC5 100 and BEP (adding bleomycin and etoposide). In the 1990s Lung cancer etoposide and cisplatin forms the basis of virtually all Skalos et al. [45] 143 300 100 57 58 regimens for the treatment of germ cell cancer. Klatersky et al. [46] 120 27 202 16 325 A large multi-center study in germ cell tumors compared etoposide with either cisplatin or carboplatin in a randomized design [41]. The dose for carboplatin in this platin also employed 5-FU infusion but for a 120-hour study was the highest of any of the randomized trials period and the dose of carboplatin was increased by 1/3 comparing carboplatin at 500 mg/m2 with cisplatin at to 400 mg/m2. Cycles were repeated at 21 days in 100 mg/m2. The response rates, as would be expected in contrast to the SWOG study in which the carboplatin this responsive tumor, were extremely high with 88% of arm was repeated at 28 days. The study was interrupted patients achieving a CR on the cisplatin arm and 90% with only 95 patients accrued because of the significant achieving a complete response on the carboplatin arm. superiority of the cisplatin arm compared to the carbo- Importantly, however, the group receiving carboplatin platin arm in response rate, disease free survival and had a 12% incidence of relapse compared to only 3% in overall survival. The contrasting results of the two trials the group receiving cisplatin. Although no difference in may be related to the fact that the smaller study did not overall survival was established, the event free and include patients with metastatic or recurrent cancer but relapse free survival was inferior for the patients treated with carboplatin with a P-value equal to 0.02 and 0.005 both studies demonstate an advantage for cisplatin. respectively. Furthermore, hematologic toxicity was substantially greater in the patients receiving carboplatin and although neuropathy and ototoxicity were increased Bladder cancer in the cisplatin arm no grade 3 or grade 4 toxicity was seen. The authors conclude that cisplatin remains as the Platinum has become a standard part of the MCV standard platinum analogue for germ cell tumors. regimen (methotrexate, cisplatin and vinblastine) in bladder cancer which is a outgrowth of the original A more recent randomized trial compared cisplatin M-VAC program that included the anthracycline doxo- versus carboplatin in a three-drug combination with rubicin. In a small study carried out by Petrioli et al., etoposide and bleomycin in a group of 54 patients [42]. the European version of M-VAC designated as MVEC This study utilized an AUC of 5 mg/ml x min for (epirubicin replacing doxorubicin), was employed in a carboplatin dose calculation. Response rates were comrandomized comparison of cisplatin versus carboplatin parable for both arms (81% vs. 76%); however, relapse [40]. Of potentially critical importance is the fact that was higher in the carboplatin arm at 32% (vs. 13% for the cisplatin arm employed a recognized therapeutic cisplatin) and tumor related mortality was higher for the single agent dose (70 mg/m2) while the carboplatin carboplatin arm, four patients vs. one patient. In the dose was only slightly over one-half of the single agent companion editorial by Horwich and Bliss [43], the dose at 250 mg/m2 maintaining a carbo to cisplatin authors describe their original phase II trial of carboratio of 3.6:1. Toxicities were substantially different be- platin with etoposide and bleomycin as well as their tween the two arms with nephrotoxicity and neurotox- report of a phase III trial in which carboplatin is comicity at 54% and 58% respectively in the cisplatin arm by pared to cisplatin in the same three drug combination cycle six compared to only 16% and 10% in the carbo- [44]. In this large study involving 598 patients, the platin arm. In contrast, in the cisplatin arm, only 37% response rate with cisplatin was 94% vs. 87% for carbohad grade 2 to 4 leukopenia by cycle six compared to platin (P = 0.009) and failure free rates at one year of 58% in the carboplatin arm. The difference in response 91% vs. 77% respectively. The authors concluded that rates (71% cisplatin vs. 41% carboplatin) was statisti- carboplatin in this combination chemotherapy regimen cally significant. One may challenge the difference as was inferior to cisplatin and remains the standard first being related to the low dose of carboplatin at 250 mg/ line platinum analogue for metastatic germ cell tumors. Bladder cancer Petrioli et al. [40]
18 Lung cancer The combination of etoposide and cisplatin has become a widely employed standard chemotherapy regimen in both small cell and non-small-cell lung cancer. Carboplatin is active in both small cell carcinoma as a single agent as well as in non small cell carcinoma as part of some combination regimens including ICE (with ifosfamide and etoposide) and paclitaxel plus carboplatin. In a recent large study, Skalos et al. reported on results in small-cell lung cancer in patients with limited or extensive stage disease who received etoposide at a dose of 300 mg/m 2 and cisplatin or carboplatin at 100 and 300 mg/m2 respectively [45]. Response rates were similar in the two arms at 57% and 58%. Median survival was essentially the same for both groups at 12.5 months for cisplatin versus 11.8 months for carboplatin. In this study, hematologic toxicity was increased on the cisplatin relative to the carboplatin arm. In non small cell lung cancer, a comparison of cisplatin versus carboplatin in a multi-institutional trial was reported in 1990 [46]. Patients received etoposide 300 mg/m 2 in daily divided doses over three days plus either cisplatin 120 mg/m2 on day one or carboplatin 325 mg/m 2 (fixed dose) on day one. Two hundred and two patients were evaluable for response with a trend (P = 0.07) favoring cisplatin (response rate 27% vs. 16%) but no significant difference in survival was observed. Toxicity, especially renal and hematologic, were greater on the cisplatin arm.
A dose response effect for cisplatin has been reported in one study [47] but the updated analysis in 1996 demonstrated a reduced benefit [48] and a second study completely failed to show a clinical difference by doubling the dose [49]. The first study in ovarian cancer compared two doses of cisplatin (50 mg/m2 vs. 100 mg/m2) with both groups receiving cyclophosphamide [47]. With 165 patients randomized, a significant survival difference was identified for the higher cisplatin dose 69 weeks vs. 114 weeks (P = 0.0008) mandating interruption of the trial. The update however reported relative death benebit of 0.68 at four years compared to 0.52 at two years. The second trial compared cisplatin at 50 vs. 100 mg/m2 with cyclophosphamide and epidoxorubicin in 145 patients with ovarian cancer [49] and showed no difference in response rates or survival times. Thigpen in an editiorial addressing dose intensity for platinum analogues suggested that increasing cisplatin dose beyond a dose of intensity of 25 mg/m2/week or carboplatin beyond an AUC of 4 does not increase the benefit [50]. A dose response study has also been carried out for carboplatin [51]. In a phase III trial, advanced ovarian cancer patients were randomized between carboplatin at an AUC dose of 4 versus carboplatin at an AUC dose of 8. The frequency of complete pathologic response was similar for both doses (32% versus 30%) and there was no significant difference with regard to survival. The study conclusion is that a doubling of the carboplatin dose intensity does not result in a significant increase in therapeutic benefit thereby challenging the dose response hypothesis for carboplatin administration.
Carboplatin AUC dosing A major concern relative to the analysis of clinical trials comparing cisplatin and carboplatin has been the issue of the dose of carboplatin and potential underdosing resulting in decreased therapeutic efficacy. Older trials routinely employed a fixed arbitrary carboplatin dose based on body surface area. In contrast, contemporary studies most commonly use AUC dose determinations in recognition of the variability of renal excretion which influences the optimal dose for carboplatin. The relationship of carboplatin to cisplatin dose was initially based upon a 4:1 ratio with all of the single agent comparative trials in ovarian cancer maintaining such a ratio while in combination chemotherapy regimens for ovarian cancer as well as other tumors, the carboplatin dose employed was variable. The development of the concept and application of carboplatin dosing based upon a projected AUC permits an optimization of carboplatin dose in relationship to toxicity with the presumption that the maximum dose will result in maximal anti-tumor effect. Since AUC was used for the carboplatin dose in only two of the comparative trials the possibility exists that some patients were potentially underdosed by carboplatin. This criticism presumes that a dose response effect is an important determinant of the efficacy of carboplatin.
Cost comparisons The actual cost of cisplatin and carboplatin as measured in dollars per milligram and calculated using the usual 4 :1 ratio for therapeutic dose provides a dollar number of 1.36/mg (($1088/course) for carboplatin versus 2.84/ mg ($568/course) for cisplatin. However, this simplistic method for comparing cost does not factor in the costs of drug administration; use of additional supportive resources and therapies such as anti-emetics, cytokines, transfusions and the cost of hospitalization. Hospitalization cost contributes substantially to the economic burden and platinum treatment programs that require hospitalization for drug administration or induce toxicity or complications mandating hospitalization are destined to be more costly than low toxicity outpatient programs. One report has addressed an economic analysis of cisplatin versus carboplatin in a retrospective review of a randomized trial comparing cis- and carboplatin in ovarian cancer [52]. In 94 patients receiving 450 courses of therapy, carboplatin resulted in lower costs for hydration and anti-emetics (S31.04 vs. S66.81) and total toxicity related costs (S 1074.53 vs. S2882.40). Furthermore, toxicity related hospitalizations were reduced with carboplatin
19 (7 of 246 courses vs. 22 of 214). This trial did not utilize ondansetron as an anti-emetic which would have substantially increased cost for cisplatin administration in terms of the drug cost but would also have potentially decreased hospitalization for adverse gastrointestinal effects and dehydration. On the other side of the ledger, the total drug cost for six cycles of carboplatin was S4140 vs. $2886 for cisplatin or a savings of S1254 favoring the use of cisplatin. The actual cost savings for carboplatin relative to cisplatin in toxicity cost was $290/cycle to which may be added S100 savings for preparation. Therefore, the actual costs savings per cycle by using carboplatin is relatively modest and if one considers the use of cytokines (more likely with the marrow suppressing analogue), carboplatin may not have an economic advantage as outlined in this single report.
Discussion Antineoplastic drug analogue development has generally been directed at altering patterns of toxicity but tumor responsiveness may also be effected by the change in chemical structure. A number of analogues have been developed within the different classes of cytotoxic compounds and with some exceptions, the general toxicity pattern of the original compound is altered in terms of intensity or frequency while in some instances the spectrum of anti-tumor efficacy is expanded. The periwinkle alkaloids, for example, represent a group of four analogues (including vindesine and vinorelbine) with which the two original analogues (vincristine and vinblastine) bear a striking similarity to the cis- and carboplatin story in that neuropathy is common with vincristine and marrow suppression is unusual (as with cisplatin) and vice versa for vinblastine (as with carboplatin). Carboplatin was being developed as a platinum analogue during the time cisplatin came into common clinical usage. The toxicity profile for cisplatin was problematic because of the major gastrointestinal effects of the drug as well as the renal toxicity and neurotoxicity which were common and dose limiting. The problems of gastrointestinal toxicity of cisplatin has subsequently been largely overcome with the use of serotonin antagonists in conjunction with dexamethasone while the renal toxicity can be obviated for the most part with the use of optimal hydration and/or fractionation of the therapeutic dose over multiple days. One important aspect of cisplatin use is the rarity of clinically important neutropenia or thrombocytopenia permitting the use of full doses in combination with agents which do produce bone marrow suppression. The introduction of carboplatin to clinical trials was heralded because it simplified the use of outpatient therapy minimizing the need for hydration or the use of expensive anti-emetics and with a decreased incidence of nephrotoxicity, neurotoxicity and ototoxicity. Bone marrow suppression, however, is the major dose limiting toxicity for this agent with a delayed
nadir pattern similar to that seen with nitrosoureas and mitomycin C. This effect results in the possible need for substantial dose adjustment when used in conjunction with combination chemotherapy with resultant delays in treatment cycles. Because of the substantial differences in toxicity profile and presumed relative equivalency of activity, comparative trials of the two analogues have been relatively limited with the majority being carried out in ovarian cancer. The general sense has been that carboplatin is therapeutically equivalent to cisplatin and is accompanied by a toxicity advantage in terms of lesser nephrotoxicity and neuro-toxicity translating into an improved therapeutic index compared to cisplatin. However, the present detailed analysis of the literature offers another perspective. In two of the randomized trials in ovarian cancer, cisplatin was shown to be superior to carboplatin although the trials have some flaws as indicated previously. The other three trials in ovarian cancer demonstrate parity for the two platinum analogues in terms of therapeutic efficacy. In the trials in germ cell tumors and in head and neck cancer, cisplatin is the superior analogue. The comparative trials in lung cancer and bladder cancer similarly favor cisplatin although the numbers of patients accrued in the single bladder cancer trial is small and survival is generally comparable. In terms of toxicity, these randomized trials generally favor carboplatin with reduced renal and neurotoxicity and in two trials hematologic toxicity was also greater with cisplatin. A major criticism of the comparative trials of cis- and carboplatin may be directed at the inconsistency of carboplatin dose and the fact that the use of AUC dose determinations instead of the fixed dose per m2 schema was not applied in most of the trials. However, a recent large randomized trial comparing a carboplatin AUC of 4 to an AUC of 8, failed to demonstrate a difference in response rate. Therefore, it is unlikely that the dose of carboplatin employed in the randomized trials played a role in underestimating the therapeutic benefit of this analogue. Selection of the optimal platinum analogue to be utilized may be dependent on many factors including a) treatment goals (palliative vs. cure); b) companion drugs in combination; and c) disease to be treated. If the goal is palliation, carboplatin may be the preferable agent because it is associated with has less symptom producing toxicity. For drug combinations that involve neuropathic agents (for example paclitaxel) or may be influenced by renal function (for example topotecan), carboplatin may also be preferred. For some specific tumors such as germ cell malignancy, cisplatin is unequivocably superior and for the four other major tumors, cisplatin may be the preferable agent but the data is not compelling or definitive. Although definitive conclusions may not be acceptable to all clinicians with regard to the probable superiority of cisplatin over carboplatin, it is reasonable to interpret the existing phase III comparative trials of cis- and carboplatin as at least suggesting
20 that cisplatin could be the preferred agent in some, if not most, clinical settings and for most solid tumors. Acknowledgement Supported in part by The Cancer Center Research Foundation.
References 1. Horwich A. Carboplatin in the treatment of testicular cancer: The Royal Marsden Hospital Testicular Tumour Unit Experience. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 53-62. 2. Kavanagh JJ. Carboplatin in refractory epithelial ovarian cancer. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 141-6. 3. Trump DL, Elson P, Madajewicz S. A phase II trial of carboplatin in advanced transitional cell carcinoma of the urothelium. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 93-7. 4. Smith IE. Carboplatin in small-cell lung cancer: The Royal Marsden Hospital experience. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 243-7. 5. Green M, Kretsman H, Lutkart S et al. Carboplatin in nonsmall-cell lung cancer: The Cancer and Leukemia Group B experience. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 301-6. 6. Volling P, Schroeder M, Achterrath W. Phase II trial with carboplatin and 5-fluorouracil in previously untreated patients with advanced squamous cell carcinoma of the head and neck. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 213-9 7. Raghavan D, Gianoutsos P, Bishop J et al. Phase II trial of carboplatin in the treatment of advanced malignant mesothelioma. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 341-4. 8. Martinez JA, Morton G, Sanz GF et al. A phase II clinical trial of carboplatin infusion in high risk acute non lymphoblastic leukemia. J Clin Oncol 1991; 9: 39. 9. Lee E and Van Echo DA. The potential of carboplatin in human leukemia. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 361-5. 10. Ruckdeschel JC. The future role of carboplatin. Semin Oncol 1994; 21: 114-8. 11. Comis RL. Cisplatin: The future. Semin Oncol 1994; 21:109-13. 12. Weed DL. Methodologic guidelines for review papers. J Natl Cancer Inst 1997; 89(1). 13. Knox RJ, Friedlos F, Lydall DA et al. Mechanism of cytotoxicity of anticancer platinum drugs: Evidence that cis-diam-minedichloroplatinum (II) and cis-diammine(l,l-cyclobutane-diacarboxylate) platinum (II) differ only in kinetics of their interaction with DNA. Cancer Res 1986; 46: 1972-9. 14. Reed E, Dabholkar M, Chabner BA. In Chabner A and Longo D (eds): Platinum Analogues in Cancer Chemotherapy and Biotherapy. 2nd Edition. Philadelphia: Lippincott-Raven Publishers 1996. 15. Calvert AH, Newell DR, Grumbrell LA et al. Carboplatin dosage: Prospective evaluation of a simple formula based on renal function. J Clin Oncol 1989; 7: 1748-56. 16. Jodrell DI, Egorin MJ, Canetta RM et al. Relationships between carboplatin exposure and tumor response and toxicity in patients with ovarian cancer. J Clin Oncol 1992; 10: 520-8.
17. Van der Vijgh VJF. Clinical pharmacokinetics of carboplatin. Clin Pharmacokinet 1991; 21: 242-61. 18. Trump DL, Grem JL, Tutsch KD et al. Platinum analogue combination chemotherapy: Cisplatin and carboplatin - a phase I trial with pharmacokinetic assessment of the effect of cisplatin administration on carboplatin excretion. J Clin Oncol 1987; 5: 1281-9. 19. Patton JF, Powell BL, White DR et al. Combination cisplatin and carboplatin in advanced squamous cell carcinoma of the head and neck. Cancer Invest 1996; 14(2): 98-102. 20. Piccart MJ, Nogaret JM, Marcelis L et al. Cisplatin combined with carboplatin: A new way of intensification of platinum dose in the treatment of advanced ovarian cancer. J Natl Cancer Inst 1990; 82(8): 703-7. 21. Taylor AE, Wiltshaw E, Gore ME et al. Long-term follow up of the first randomized study of cisplatin versus carboplatin for advanced epithelial ovarian cancer. J Clin Oncol 1994; 12: 2066-70. 22. Adams M, Kerby U, Rocker I et al. A comparison of the toxicity and efficacy of cisplatin and carboplatin in advanced ovarian cancer. The Swons Gynacological Cancer Group. Acta Oncol 1989; 28(1): 57-60. 23. Mangioni C, Bolis G, Pecorelli S et al. Randomized trial in advanced ovarian cancer comparing cisplatin and carboplatin. J Natl Cancer Inst 1989; 81: 1464-71. 24. Swenerton K, Jeffrey J, Stuart G et al. Cisplatin-cyclophosphamide versus carboplatin-cyclophosphamide in advanced ovarian cancer: A randomized phase III study of the National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol 1992; 10: 718-26. 25. Ten Bokkel Huinink WW, Van der Burg MEL, Van Oosterom AT et al. Carboplatin combination therapy for ovarian cancer. Cancer Treat Rep 1988; 15(8): 9-15. 26. Alberts DS, Green S, Hannigan EV et al. Improved therapeutic index of carboplatin plus cyclophosphamide versus cisplatin plus cyclophosphamide. Final report by the Southwest Oncology Group of a phase III randomized trial in stages III and IV ovarian cancer. J Clin Oncol 1992; 10: 706-17. 27. Conte PF, Bruzzone M, Carmino F et al. Carboplatin, doxorubicin and cyclophosphamide versus cisplatin, doxorubicin and cyclophosphamide. A randomized trial in stage III—IV epithelial ovarian carcinoma. J Clin Oncol 1991; 9: 658-63. 28. Belpomme D, Bugat R, Rives M et al. Carboplatin versus cisplatin in association with cyclophosphamide and doxorubicin as first line therapy in stage III—IV ovarian carcinoma: Results of an ARTAC phase III trial. Proc ASCO 1992; 11: 227. 29. Edmonson JH, McCormack GM, Wieand HS et al. Cyclophosphamide-cisplatin versus cyclophosphamide carboplatin in myelosuppressive regimens. J Natl Cancer Inst 1989; 81: 1500-4. 30. Neijt JP, Hansen M, Hansen PG et al. Randomized phase III study in previously untreated epithelial ovarian cancer FIGO stage IIB, IIC, III, IV, comparing paclitaxel-cisplatin and paclitaxel-carboplatin. Proc ASCO 1997; 16: 352a. 31. Advanced Ovarian Cancer Trialists Group. Chemotherapy in advanced ovarian cancer: An overview of randomised clinical trials. BMJ 1991; 303: 884-93. 32. Vermorken JB, ten Bokkel Huinink WW, Eisenhauer EA et al. Advanced ovarian cancer. Carboplatin versus cisplatin. Ann Oncol 1993; 4 (Suppl 4): 41-8 (Review). 33. Gore ME, Fryatt I, Wiltshaw E, Dawson T. Treatment of relapsed carcinoma of the ovary with cisplatin or carboplatin following initial treatment with these compounds. Gynecol Oncol 1990; 36: 207-11. 34. Repetto L, Chiara S, Mammoliti S et al. Crossover study with cisplatin or carboplatin in advanced ovarian cancer. Gynecol Oncol 1990; 39: 146-9. 35. Alberts DS, Liu PY, Hannigan EV et al. Intraperitoneal cisplatin plus intravenous cyclophosphamide versus intravenous cisplatin plus intravenous cyclophosphamide for stage III ovarian cancer. N Engl J Med 1996; 335: 1950-5.
21 nomatous germ cell cancer: A multiinstitutional medical research 36. McGuire WP, Hopkins WJ, Brady ME et al. Cyclosphosphamide council/European organization for research and treatment of and cisplatin compared with paclitaxel and cisplatin in patients cancer trial. J Clin Oncol 1997; 15: 1844-52. with stage III and stage IV ovarian cancer. N Engl J Med 1996; 334:1-6. 46. Skalos DV, Samantas E, Kosmidis P et al. Randomized compari37. Kohn EC, Sarosy GA, Davis P et al. A phase I/II study of dosesion of etoposide-cisplatin vs. etoposide-carboplatin and irradiation in small-cell lung cancer. Ann Oncol 1994; 5: 601-7. intense paclitaxel with cisplatin and cyclophosphamide as initial therapy of poor-prognosis advanced-stage epithelial ovarian can- 47. Klastersky J, Sculier JP, Lacroix G et al. A randomized study cer. Gynecol Oncol 1996; 62: 181-91. comparing cisplatin or carboplatin with etoposide in patients with advanced non-small-cell lung cancer: European Organization for 38. Markman M. Carboplatin and cisplatin: Are they equivalent in Research and Treatment of Cancer Protocol 07861. J Clin Oncol efficacy in 'optimal residual' advanced ovarian cancer? J Cancer 1990; 8: 1556-62. Res Clin Oncol 1996; 122: 443-4. 39. De Andres L, Brunet J, Lopez-Pousa A et al. Randomized trial of 48. Kaye SB, Lewis CR, Paul J et al. Randomised study of two doses neoadjuvant cisplatin and fluorouracil versus carboplatin and of cisplatin with cyclophosphamide in epithelial ovarian cancer. fluorouracil in patients with stage IV-MO Head and Neck cancer. Lancet 1992; 340: 329-33. J Clin Oncol 1995; 13: 1493-500. 49. Kaye SB, Paul J, Cassidy J et al. Mature results of a randomized trial of two doses of cisplatin for the treatment of ovarian cancer. 40. Forestiere AA, Metch B, Schuller DE et al. Randomized comparison of cisplatin plus fluorouracil and carboplatin plus fluJClin Oncol 1996; 14: 2113-9. orouracil versus methotrexate in advanced squamous-cell carci- 50. Conte PF, Bruzzone M, Carnino F et al. High-dose versus lownoma of the head and neck: A southwest Oncology Group Study. dose cisplatin in combination with cyclophosphamide and epiJ Clin Oncol 1992; 10: 1245-251. doxorubicin in suboptimal ovarian cancer: A randomized study of the Gruppo Oncologico Nord-Ovest. J Clin Oncol 1996; 14: 41. Petrioli R, Frediani B, Manganelli A et al. Comparison between a 351-6. cisplatin-containing regimen and a carboplatin-containing regimen for recurrent or metastatic bladder cancer patients. A ran51. Thigpen JT. Dose-intensity in ovarian carcinoma: Hold, enough? domized phase II study. Cancer 1996; 77(2): 344-51. J Clin Oncol 1997; 15:1291-3 (Editorial). 42. Bajorin DF, Sarosdy MF, Pfister DG et al. Randomized trial of 52. Bertelsen JK, Anderson JE, Havsteen H et al. Dose-effect study etoposide and cisplatin versus etoposide and carboplatin in paof carboplatin in ovarian cancer: A Danish ovarian cancer group tients with good-risk germ cell tumors: A multiinstitutional study. study. J Clin Oncol 1997; 15: 193-8. J Clin Oncol 1993; 11: 598-606. 53. Alberts D. Cisplatin versus carboplatin in advanced ovarian cancer: An economic analysis. PharTher July 1994. 43. Bokemeyer C, Kohrmann O, Tischler L et al. A randomized trial of cisplatin, etoposide and bleomycin (PEB) versus carboplatin, etoposide and bleomycin (CEB) for patients with 'good-risk' Received 23 April 1997; accepted 31 July 1997. metastatic non-seminomatous germ cell tumors. Ann Oncol 1996; 7: 1015-21. 44. Bliss J, Horwich A. Carboplatin in the combination chemo- Correspondence to: therapy of non-seminomatous germ cell tumours. Ann Oncol Jacob Lokich, MD 1996; 7: 989-91 (Editorial). The Cancer Center 125 Parker Hill Avenue 45. Horwich A, Sleijfer DT, Fossa SD et al. Randomized trial of bleomycin, etoposide and cisplatin compared with bleomycin, Boston, MA 02120 etoposide and carboplatin in good-prognosis metastatic nonsemi- USA
Review Carboplatin versus cisplatin in solid tumors: An analysis of the literature J. Lokich & N. Anderson The Cancer Center of Boston in Boston, Plymouth and Framingham, MA, USA
or equivalent to carboplatin in therapeutic efficacy in all five tumors but was associated with an increased toxicity profile for Background: Introduced into clinical usage in 1992 as a plati- gastrointestinal, renal and neurologic effects. Conclusions: For some tumors, cisplatin appears to be num analogue with a distinctively different toxicity profile from cisplatin, carboplatin has become a commonly preferred superior to carboplatin in terms of therapeutic effectiveness agent over cisplatin. The comparative therapeutic efficacy of (germ cell tumors, bladder cancer, head and neck cancer), the two agents remains controversial however, prompting an while for others, effectiveness is comparable (lung cancer, analysis of the phase III trials in ovarian cancer and other ovarian cancer). Toxicity profiles are distinctly different for the two analogues however, generally favoring carboplatin. The tumors in which the two were compared. Methods: Clinical trials comparing carboplatin with cis- issue of potential carboplatin underdosing related to the lack platin, both as single agents and in combination with other of physiologic dose calculations (utilizing the AUC [area under agents, were analyzed within the tumors for which platinum the curve] method) in the comparative trials of cis- versus has become a standard or commonly employed agent. A Med- carboplatin is probably not clinically important since a dose line search identifying the randomized trials and references response effect has not been established for carboplatin or for cisplatin. The selection of the optimal platinum analogue to be from these reports were collated for analysis. Results: The original clinical comparative trials as well as employed is dependent on the type of tumor, the treatment literature reviews and commentaries were reviewed. Five solid intention (palliative vs. curative) and the other component tumors were identified within which comparative trials had drugs being used in combination. been conducted: ovary, 10 trials; lung, 2 trials; head and neck, 2 trials; germ cell tumors, 3 trials and 1 trial in bladder cancer. Depending upon the end point selected, cisplatin was superior Key words: carboplatin, cisplatin, randomized trials Summary
Introduction
Platinum analogues have become the mainstay of treatment for many tumors including ovarian cancer, lung cancer (both non-small-cell and small-cell), germ cell tumors, head and neck cancer, bladder cancer and to a lesser degree breast cancer and gastric cancer. Cisplatin was the initial platinum analogue that was introduced into clinical practice with a toxicity profile characterized by nausea and vomiting, renal dysfunction and neuroand ototoxicity. Carboplatin was the second clinically important platinum analogue developed with a substantially different toxicity profile which included bone marrow suppression (particularly thrombocytopenia) and a lesser incidence of gastrointestinal, renal and neurotoxicity. Carboplatin is commonly considered to be the therapeutic equivalent of cisplatin while offering the additional benefit of reduced gastrointestinal effects permitting an ease of outpatient administration and obviating the requirement for hydration and anti-emetic therapy. Therapeutic activity for carboplatin has been identified in all the tumors for which cisplatin has had a role with 'comparable' therapeutic effects reported in germ cell tumors [1], ovary [2], bladder cancer [3], small-
cell and non-small-cell lung cancer [5], head and neck cancer [6] and mesothelioma [7]. In addition, antitumor activity for carboplatin has been established in resistant acute leukemia [8]. In 1994, Ruckdeschel stated that "carboplatin (rather than cisplatin) is recommended for chemotherapy for palliative or non curative intent" in a non protocol clinical arena due to its "more favorable toxicity profile and lower overall cost" [9]. On the other hand, Comis has stated that "cisplatin currently (1994) remains the recommended agent of choice for treatment of several malignancies" [10]. Cisplatin is being replaced in common clinical practice by carboplatin and the toxicologic and pharmacologic differences between the two drugs may justify such a substitution provided therapeutic efficacy is not compromised. The initial studies comparing cisplatin and carboplatin in ovarian cancer suggested that the agents were comparable therapeutically although the observation was not consistent across all trials. In contrast, randomized trials comparing carboplatin and cisplatin in combination regimens in nonsmall-cell lung cancer, bladder cancer, head and neck cancer and other tumors have indicated that cisplatin is superior to carboplatin in some instances.
14 We undertook a review of randomized trials comparing carboplatin and cisplatin published through 1997 in a spectrum of solid tumors using the methodological guidelines suggested by Weed [11] for a literature review paper. Our specific purpose was to determine if the literature supported a superiority of one analogue over the other or an equivalency of clinical efficacy of the two analogues. We used a Medline search to identify all English language published papers describing a clinical trial comparison of cisplatin and carboplatin as well as reference lists from those papers. Only papers which represented randomized trials were selected for analysis with the criteria for evaluating the quality of studies based upon sufficient patient numbers to validate a statistical analysis of the data base. The numbers of published clinical trials within each tumor category selected were limited in some tumors eliminating the possibility of a meta analysis and necessitating a narrative analysis and summary. Pharmacologic and toxicologic comparison of cisplatin and carboplatin Cisplatin and carboplatin demonstrate the identical mechanism of cell kill, functioning as non classical alkylating agents by binding to cellular DNA to form cross links. The quantitative level of platinum DNA adduct formation correlates with the degree of cell killing. Knox et al. demonstrated that equivalent tumor cytotoxicity is dependent upon achieving an equivalent level of DNA binding [12]. Carboplatin has a greater degree of chemical stability than cisplatin resulting in a lesser reactivity with DNA; as a consequence, a higher dose of carbobplatin is necessary to obtain comparable anti-tumor effects in experimental systems. In in vivo experimental anti tumor systems, carboplatin is 8 to 45 times less potent than cisplatin although the spectrum of experimental tumor system activity is similar for the two agents [13]. The therapeutic dose of carboplatin relative to the therapeutic dose of cisplatin has generally been described as a ratio of 4:1 (400-500 mg/m2 versus 100 mg/m 2 ) based on clinical studies in ovarian cancer in which these doses achieved a relative equivalency. The toxic effects of carboplatin are more precisely predicted utilizing the formula developed by Calvert et al. [14] and subsequently modified by Jodrell et al. [15]. In this formula, a specific area under the curve (AUC) is established as a goal for the carboplatin plasma concentration related to the creatinine clearance since 90% of a carboplatin dose is excreted by the kidneys. An optimal single agent carboplatin AUC of 5-7 mg/ml x minutes has been suggested as a basis for optimal dose calculation. Most prospective randomized trials comparing carboplatin and cisplatin have employed a fixed dose for carboplatin delivery and have not used the Calvert formula. This represents an important consideration in comparing the two agents within such
Table 1 Comparison of pharmacology and toxicity patterns of carboplatin and cisplatin.
Pharmacology Metabolism Excretion Kinetics Tl/2Beta Tl/2 Gamma Toxicity3
a b
Cisplatin
Carboplatin
Inactivated by sulfhydryl groups 25% excreted; 9:1 urine: bile
No significant metabolism 90% renal recovery
60 minutes 24 hours Gastrointestinal Renal Insufficiency Neuropathy
80-100 minutes 22-40 hours Myelosupressionb
In order of severity and dose limiting effect (see text). Characteristically delayed with nadir at day 21 to 28.
trials and is discussed in a subsequent section. The pharmacology and toxicity patterns of carboplatin and cisplatin are described in Table 1. A major difference in the two analogues is in their metabolism and excretion. In vivo, carboplatin demonstrates no significant catabolism while cisplatin is inactivated by sulfhydryl groups. Generally, 90% of an injected dose of carboplatin is recovered in the urine while only 25% of cisplatin is excreted unaltered. The general pharmacokinetic pattern as reported by Reed et al. [13] indicates that cisplatin and carboplatin are relatively similar. However, there is less protein binding of carboplatin and more sophisticated pharmacologic studies separating ultra filterable platinum (UF) and total platinum reveal that carboplatin has a longer half-life of UF platinum with a T 1/2 beta of 120 minutes versus 36 minutes for cisplatin [16]. Nonetheless, the terminal half-life is comparable for the two drugs at 5.8 and 5.4 days. With regard to toxicity, carboplatin is less nephrotoxic and less emetogenic than cisplatin and neurotoxicity and ototoxicity are virtually absent. Myelosuppression is the major toxic effect of carboplatin and combining carboplatin with other cytotoxic agents may be complicated since the pattern of myelosuppression, particularly thrombocytopenia may be delayed to a nadir at day 21 to 28. In contrast, the major toxicity for cisplatin has been nausea and vomiting and generalized gastrointestinal effects including post platinum diarrhea. For the most part, this pattern of toxicity is minimized with the use of serotonin antagonists and with dose fractionation of the cisplatin. Hydration and dose fractionation mitigate most of the nephrotoxic effect of the cisplatin. The neuropathic effects for cisplatin are relatively common and are related to the cumulative dose administered. The combination of simultaneous administration of carboplatin and cisplatin has been employed in a number of clinical trials [17-19], based upon the fact that the toxicity profiles are dissimilar for the two agents and since tumor cell killing is dependent upon maximizing DNA adduct formation, higher total platinum doses
15 have been postulated to yield higher levels of cell killing. Three studies combining cis- and carboplatin delivered a maximum carboplatin dose of 280, 300 and 350 mg/ m2 with the paired cisplatin dose of 75,100 and 150 mg/ m2 respectively. No significant increase in tumor response was observed in spite of the substantial increase in UF platinum in studies in ovarian cancer [19] and head and neck cancer [18]. The logistical complexities of these double platinum regimens combined with the lack of an important therapeutic advantage observed has discouraged further study of the analogue combination in spite of the interesting if not compelling rationale for its use. Clinical trials The tumors for which platinum analogues have become standard chemotherapy broaches a wide spectrum. Platinum is part of the first line chemotherapy regimen for five solid tumors including: ovarian cancer, germ cell tumors, head and neck cancer, both small-cell and nonsmall-cell lung cancer and bladder cancer. Additionally, selected tumors of the gastrointestinal tract (specifically esophageal, gastric and anal cancers) are commonly treated with platinum regimens although the role of 5-fluorouracil is arguably more important. Cisplatin also is active in breast cancer although it is infrequently utilized as first or second line therapy. Randomized trials comparing cis- and carboplatin have been reported for the five common tumors for which a platinum analogue is routinely administered as part of first line therapy.
Table 2. A comparison of response rates to cis- or carboplatin single agent or combination chemotherapy regimens as first line therapy in ovarian cancer. Study [reference]
No. patients
Response rate %
Dose mg/m Carbo
Cis
Carbo
Cis
Single agent Taylor [20] Adams et al. [21] Mangioni et al. [22]
131 88 173
400 400 400
100 100 100
38.4 67.5 57.3
53.8 47.5 71.6
Combination studies NCI Canada 1992 [23] EORTC 1988 [24] SWOG 1992 [25] Italy 1991 [26] ARTAC 1992 [27] Mayo Clinic 1989 [28] Netherlands 1997 [29]
417 342 291 164 144 103 190
300 350 300 200 300 150 AUC5
75 100 100 50 75 60 75
59 36.2 61 62 47 No datsi Equal
57 45.9 52 66 73 a
' P = 0.008.
comparable for the two analogues. No difference in median or progression free survival was observed between the arms in any of the three trials but the data are complicated by the fact that crossover to the alternative platinum analogue was permitted in all trials. In the seven combination chemotherapy studies involving cis- versus carboplatin, the alkylating agent cyclophosphamide was employed in six in which two trials added an anthracycline. Paclitaxel was the companion agent in one trial. The dose of carboplatin was 150 mg/m2 in the Mayo Clinic study (although dose escalation was employed as part of the design) but was 300 mg/m2 or more in four studies and one utilized the Ovarian cancer trials AUC method. The seven studies involved almost 1700 Ten randomized trials comparing cis-and carboplatin in patients, predominantly with suboptimal residual tuthe treatment of ovarian cancer have been reported mor. In two of the studies [26, 27], the pathologic between 1989 and 1997 [20-29] (Table 2) with the most complete response rate for cisplatin was superior to that recent review and analysis of these trials appearing in of carboplatin but was not statistically significant. Over1993 [29]. A meta analysis of chemotherapy in ovarian all survival and response rates were statistically equivacancer [30] reported in 1991 and a general review of the lent for cis- and carboplatin in all trials with two excepsame topic in 1993 [31] identified 11 trials comparing cis- tions. The Mayo Clinic trial which was closed to patient and carboplatin but many of the trials had limited follow entry after an interim analysis demonstrated a statistiup at that time and two of the eleven were either cally significant disease free survival advantage for the unpublished or had limited patient numbers. cisplatin arm [28]. This trial also employed a strategy Of the 10 now mature trials, three involved a com- of dose equivalency based on toxicity for the cis- and parison of single agent carbo and cisplatin in previously carboplatin arms. In this trial, 103 patients were ranuntreated patients one of which was updated in 1994. domized and the interim analysis revealed a median Almost 400 patients were entered in the three single progression free survival of 17 months for the cisplatin agent studies which were carried out predominantly in containing arm versus 12 months for the carboplatin patients with suboptimal residual disease. The carbopla- containing arm (P = 0.04). Median overall survival was tin dose was fixed in all three studies with a ratio of 4:1 also superior for the cisplatin arm (27 vs. 20 months) to the cisplatin dose. Although not statistically signifi- but was not statistically significant (P - 0.14). The cantly different, the overall response rate was higher in carboplatin patients on therapy at time of study closure the cisplatin arms in two of the three studies and in the were offered crossover to cisplatin therapy and within third study the overall response rate was higher in the this group of 21 patients, the survival pattern reverted to carboplatin arm. Pathologic complete response rates that of patients receiving cisplatin as their initial therwere similar and the median response durations were apy.
16 The other trial demonstrating a significant therapeutic difference between the analogues and favoring cisplatin was reported by Belpomme et al. in abstract form for the French ARTAC ovarian cancer study group [27]. The pathologic CR and overall response rate was higher for the cisplatin containing arm 33% CR and 73% PR versus 15% and 47% (P - 0.008) respectively for the carboplatin arm. In addition, median survival was different and statistically significant 27.9 months vs. 20.6 months (P - 0.028) favoring cisplatin. This trial has not as yet been reported in a peer reviewed journal. Three of the remaining five trials of combination chemotherapy comparing cis- and carboplatin are remarkably similar across the trials in terms of response rates (59%, 61%, 62% RR for carboplatin and 57%, 52%, 66% RR for cisplatin) with apparent equivalency for the analogues. One early trial [24] resulted in lower overall response rates (36% and 45%) and cisplatin yielded the higher respone rate but the difference was not statistically significant. A preliminary report from the Netherlands has reported on a comparison of paclitaxel 175 mg/m2 with either cisplatin 75 mg/m2 or carboplatin to an AUC of 5 [29]. This trial does not have the experimental design flaws or problems of the ongoing GOG trial (vide infra) and is reported with an interim analysis only but efficacy was not significantly different between the two arms. Another method of comparing cis- and carboplatin is through an analysis of data from trials in which patients are treated in sequence with the alternative platinum analogue after developing refractory or resistant disease to one analogue. Thus, if a patient is treated first with cisplatin and develops resistance (or becomes refractory), the second line or crossover treatment is carboplatin and vice versa. In fact, crossover was allowed in most of the randomized trials reviewed. Two reports have appeared in which carboplatin or cisplatin was either the rechallenging agent or the crossover agent [32, 33]. In the study from the Royal Marsden Hospital, 15 of 43 crossover patients responded (34%) as opposed to a 9% response rate in rechallenged patients i.e., the same platinum analogue was used upon relapse [32]. Ten patients achieved a complete response to carboplatin after cisplatin but there was no statistically significant difference in response rate according to which analogue was administered first. In the other crossover study, 57 patients were analyzed; 24 received cisplatin and 33 received carboplatin as the crossover platinum therapy [33]. Response rates were higher with cisplatin (25% vs. 9% for carboplatin) and crossover responses were observed only with cisplatin as second line in patients not responding to first line platinum therapy (3 of 12 for cisplatin vs. 0 of 11 for carboplatin). Most of these randomized trials of single agent and combination chemotherapy were completed prior to the advent of new agents and effective second-line therapy for ovarian cancer including paclitaxel and topotecan. Thus, the state-of-the-art for the treatment of ovarian cancer in 1997 has changed substantially with more
recent trials indicating that intraperitoneal cisplatin (with i.v. cyclophosphamide) is superior to intravenous cisplatin in terms of survival [34] and that paclitaxel plus cisplatin is superior to cyclophosphamide plus cisplatin [35]. In fact, the future direction of chemotherapy for ovarian cancer is being guided by phase I and phase II trials investigating the three drug combination of paclitaxel, cyclophosphamide and cisplatin [36] with cisplatin perferred over carboplatin in such trials because of its reduced marrow suppression but up front combinations including topotecan may preferentially use carboplatin because of the renal effect of cisplatin. A definitive conclusion with regard to the question of cis- vs. carboplatin in ovarian cancer may still be elusive. Markman has suggested that carboplatin and cisplatin well may be therapeuticaly equivalent in .^optimal residual disease but that does not necessarily extend to optimal residual disease [37]. In the randomized trials listed in Table 2, the majority of patients had suboptimal residual tumor. The ongoing Gynecologic Oncolcogy Group (GOG) trial in optimal residual stage III ovarian cancer compares cisplatin with carboplatin, the latter utilizing an AUC of 7.5. Paclitaxel is added to both arms with different doses and infusion times 135 mg/m2 over 24 hours versus 175 mg/m2 over three hours. This trial may permit a definitive comparison of cis- and carboplatin but the use of different paclitaxel doses and schedules may create problems for the final analysis.
Non-ovarian cancer trials
In the four solid tumors for which platinum has become a standard component of combination chemotherapy, eight studies have been devoted to comparing cisplatin and carboplatin in a randomized trial design (Table 3). Carboplatin dosage varied from 250 mg/m2 to 500 mg/ m2 as a fixed dose depending upon the type of tumor or the combination of drugs employed but two trials did use an AUC guided carboplatin dose. Head and neck cancer
Two prospective randomized trials address the issue of cisplatin versus carboplatin in head and neck cancer [38, 39]. The study from the Southwest Oncology Group (SWOG) involved 277 patients with the companion drug being 5-fluorouracil (5-FU) administered as a 96-hour infusion. Single agent methotrexate was employed as a control arm. Ototoxicity and renal toxicity were substantially greater in the cisplatin arm compared to the carboplatin arm and uncharacteristically, hematologic toxicity was also greater in the cisplatin group. In spite of the greater response rate in the cisplatin group (32% vs. 21%), the response duration and median survival was similar for both cisplatin and carboplatin arms as well as the single agent methotrexate arm. The second study in head and neck cancer comparing cisplatin and carbo-
17 Table 3. Carboplatin versus cisplatin in lung, head and neck, bladder and germ cell tumors and lung cancer. Study [reference]
Head and neck cancer Andres et al. [38] Forestiere et al. [39]
No. patients
Dose mg/m 2
Response rate %
Carbo
Cis
Carbo
Cis
95 277
400 300
100 100
96 32
76 21
m2; however, it would seem unlikely that in the context of this particular combination of drugs that carboplatin dose escalation would be possible in view of the 58% incidence of leukopenia. Germ cell tumors
Platinum chemotherapy represented a major advance in the treatment and curability of germ cell tumors from 57 71 70 41 250 the time of its initial introduction into clinical use. Germ cell tumor Platinum based combination chemotherapy regimens Bajorinset al. [41] 500 100 90 265 for this tumor have included VPB (adding vinblastine Bokemeyer et al. [42] 76 81 54 AUC 5 100 and bleomycin), VIP (adding etoposide and ifosfamide) Horwich et al. [43] 87 44 598 AUC5 100 and BEP (adding bleomycin and etoposide). In the 1990s Lung cancer etoposide and cisplatin forms the basis of virtually all Skalos et al. [45] 143 300 100 57 58 regimens for the treatment of germ cell cancer. Klatersky et al. [46] 120 27 202 16 325 A large multi-center study in germ cell tumors compared etoposide with either cisplatin or carboplatin in a randomized design [41]. The dose for carboplatin in this platin also employed 5-FU infusion but for a 120-hour study was the highest of any of the randomized trials period and the dose of carboplatin was increased by 1/3 comparing carboplatin at 500 mg/m2 with cisplatin at to 400 mg/m2. Cycles were repeated at 21 days in 100 mg/m2. The response rates, as would be expected in contrast to the SWOG study in which the carboplatin this responsive tumor, were extremely high with 88% of arm was repeated at 28 days. The study was interrupted patients achieving a CR on the cisplatin arm and 90% with only 95 patients accrued because of the significant achieving a complete response on the carboplatin arm. superiority of the cisplatin arm compared to the carbo- Importantly, however, the group receiving carboplatin platin arm in response rate, disease free survival and had a 12% incidence of relapse compared to only 3% in overall survival. The contrasting results of the two trials the group receiving cisplatin. Although no difference in may be related to the fact that the smaller study did not overall survival was established, the event free and include patients with metastatic or recurrent cancer but relapse free survival was inferior for the patients treated with carboplatin with a P-value equal to 0.02 and 0.005 both studies demonstate an advantage for cisplatin. respectively. Furthermore, hematologic toxicity was substantially greater in the patients receiving carboplatin and although neuropathy and ototoxicity were increased Bladder cancer in the cisplatin arm no grade 3 or grade 4 toxicity was seen. The authors conclude that cisplatin remains as the Platinum has become a standard part of the MCV standard platinum analogue for germ cell tumors. regimen (methotrexate, cisplatin and vinblastine) in bladder cancer which is a outgrowth of the original A more recent randomized trial compared cisplatin M-VAC program that included the anthracycline doxo- versus carboplatin in a three-drug combination with rubicin. In a small study carried out by Petrioli et al., etoposide and bleomycin in a group of 54 patients [42]. the European version of M-VAC designated as MVEC This study utilized an AUC of 5 mg/ml x min for (epirubicin replacing doxorubicin), was employed in a carboplatin dose calculation. Response rates were comrandomized comparison of cisplatin versus carboplatin parable for both arms (81% vs. 76%); however, relapse [40]. Of potentially critical importance is the fact that was higher in the carboplatin arm at 32% (vs. 13% for the cisplatin arm employed a recognized therapeutic cisplatin) and tumor related mortality was higher for the single agent dose (70 mg/m2) while the carboplatin carboplatin arm, four patients vs. one patient. In the dose was only slightly over one-half of the single agent companion editorial by Horwich and Bliss [43], the dose at 250 mg/m2 maintaining a carbo to cisplatin authors describe their original phase II trial of carboratio of 3.6:1. Toxicities were substantially different be- platin with etoposide and bleomycin as well as their tween the two arms with nephrotoxicity and neurotox- report of a phase III trial in which carboplatin is comicity at 54% and 58% respectively in the cisplatin arm by pared to cisplatin in the same three drug combination cycle six compared to only 16% and 10% in the carbo- [44]. In this large study involving 598 patients, the platin arm. In contrast, in the cisplatin arm, only 37% response rate with cisplatin was 94% vs. 87% for carbohad grade 2 to 4 leukopenia by cycle six compared to platin (P = 0.009) and failure free rates at one year of 58% in the carboplatin arm. The difference in response 91% vs. 77% respectively. The authors concluded that rates (71% cisplatin vs. 41% carboplatin) was statisti- carboplatin in this combination chemotherapy regimen cally significant. One may challenge the difference as was inferior to cisplatin and remains the standard first being related to the low dose of carboplatin at 250 mg/ line platinum analogue for metastatic germ cell tumors. Bladder cancer Petrioli et al. [40]
18 Lung cancer The combination of etoposide and cisplatin has become a widely employed standard chemotherapy regimen in both small cell and non-small-cell lung cancer. Carboplatin is active in both small cell carcinoma as a single agent as well as in non small cell carcinoma as part of some combination regimens including ICE (with ifosfamide and etoposide) and paclitaxel plus carboplatin. In a recent large study, Skalos et al. reported on results in small-cell lung cancer in patients with limited or extensive stage disease who received etoposide at a dose of 300 mg/m 2 and cisplatin or carboplatin at 100 and 300 mg/m2 respectively [45]. Response rates were similar in the two arms at 57% and 58%. Median survival was essentially the same for both groups at 12.5 months for cisplatin versus 11.8 months for carboplatin. In this study, hematologic toxicity was increased on the cisplatin relative to the carboplatin arm. In non small cell lung cancer, a comparison of cisplatin versus carboplatin in a multi-institutional trial was reported in 1990 [46]. Patients received etoposide 300 mg/m 2 in daily divided doses over three days plus either cisplatin 120 mg/m2 on day one or carboplatin 325 mg/m 2 (fixed dose) on day one. Two hundred and two patients were evaluable for response with a trend (P = 0.07) favoring cisplatin (response rate 27% vs. 16%) but no significant difference in survival was observed. Toxicity, especially renal and hematologic, were greater on the cisplatin arm.
A dose response effect for cisplatin has been reported in one study [47] but the updated analysis in 1996 demonstrated a reduced benefit [48] and a second study completely failed to show a clinical difference by doubling the dose [49]. The first study in ovarian cancer compared two doses of cisplatin (50 mg/m2 vs. 100 mg/m2) with both groups receiving cyclophosphamide [47]. With 165 patients randomized, a significant survival difference was identified for the higher cisplatin dose 69 weeks vs. 114 weeks (P = 0.0008) mandating interruption of the trial. The update however reported relative death benebit of 0.68 at four years compared to 0.52 at two years. The second trial compared cisplatin at 50 vs. 100 mg/m2 with cyclophosphamide and epidoxorubicin in 145 patients with ovarian cancer [49] and showed no difference in response rates or survival times. Thigpen in an editiorial addressing dose intensity for platinum analogues suggested that increasing cisplatin dose beyond a dose of intensity of 25 mg/m2/week or carboplatin beyond an AUC of 4 does not increase the benefit [50]. A dose response study has also been carried out for carboplatin [51]. In a phase III trial, advanced ovarian cancer patients were randomized between carboplatin at an AUC dose of 4 versus carboplatin at an AUC dose of 8. The frequency of complete pathologic response was similar for both doses (32% versus 30%) and there was no significant difference with regard to survival. The study conclusion is that a doubling of the carboplatin dose intensity does not result in a significant increase in therapeutic benefit thereby challenging the dose response hypothesis for carboplatin administration.
Carboplatin AUC dosing A major concern relative to the analysis of clinical trials comparing cisplatin and carboplatin has been the issue of the dose of carboplatin and potential underdosing resulting in decreased therapeutic efficacy. Older trials routinely employed a fixed arbitrary carboplatin dose based on body surface area. In contrast, contemporary studies most commonly use AUC dose determinations in recognition of the variability of renal excretion which influences the optimal dose for carboplatin. The relationship of carboplatin to cisplatin dose was initially based upon a 4:1 ratio with all of the single agent comparative trials in ovarian cancer maintaining such a ratio while in combination chemotherapy regimens for ovarian cancer as well as other tumors, the carboplatin dose employed was variable. The development of the concept and application of carboplatin dosing based upon a projected AUC permits an optimization of carboplatin dose in relationship to toxicity with the presumption that the maximum dose will result in maximal anti-tumor effect. Since AUC was used for the carboplatin dose in only two of the comparative trials the possibility exists that some patients were potentially underdosed by carboplatin. This criticism presumes that a dose response effect is an important determinant of the efficacy of carboplatin.
Cost comparisons The actual cost of cisplatin and carboplatin as measured in dollars per milligram and calculated using the usual 4 :1 ratio for therapeutic dose provides a dollar number of 1.36/mg (($1088/course) for carboplatin versus 2.84/ mg ($568/course) for cisplatin. However, this simplistic method for comparing cost does not factor in the costs of drug administration; use of additional supportive resources and therapies such as anti-emetics, cytokines, transfusions and the cost of hospitalization. Hospitalization cost contributes substantially to the economic burden and platinum treatment programs that require hospitalization for drug administration or induce toxicity or complications mandating hospitalization are destined to be more costly than low toxicity outpatient programs. One report has addressed an economic analysis of cisplatin versus carboplatin in a retrospective review of a randomized trial comparing cis- and carboplatin in ovarian cancer [52]. In 94 patients receiving 450 courses of therapy, carboplatin resulted in lower costs for hydration and anti-emetics (S31.04 vs. S66.81) and total toxicity related costs (S 1074.53 vs. S2882.40). Furthermore, toxicity related hospitalizations were reduced with carboplatin
19 (7 of 246 courses vs. 22 of 214). This trial did not utilize ondansetron as an anti-emetic which would have substantially increased cost for cisplatin administration in terms of the drug cost but would also have potentially decreased hospitalization for adverse gastrointestinal effects and dehydration. On the other side of the ledger, the total drug cost for six cycles of carboplatin was S4140 vs. $2886 for cisplatin or a savings of S1254 favoring the use of cisplatin. The actual cost savings for carboplatin relative to cisplatin in toxicity cost was $290/cycle to which may be added S100 savings for preparation. Therefore, the actual costs savings per cycle by using carboplatin is relatively modest and if one considers the use of cytokines (more likely with the marrow suppressing analogue), carboplatin may not have an economic advantage as outlined in this single report.
Discussion Antineoplastic drug analogue development has generally been directed at altering patterns of toxicity but tumor responsiveness may also be effected by the change in chemical structure. A number of analogues have been developed within the different classes of cytotoxic compounds and with some exceptions, the general toxicity pattern of the original compound is altered in terms of intensity or frequency while in some instances the spectrum of anti-tumor efficacy is expanded. The periwinkle alkaloids, for example, represent a group of four analogues (including vindesine and vinorelbine) with which the two original analogues (vincristine and vinblastine) bear a striking similarity to the cis- and carboplatin story in that neuropathy is common with vincristine and marrow suppression is unusual (as with cisplatin) and vice versa for vinblastine (as with carboplatin). Carboplatin was being developed as a platinum analogue during the time cisplatin came into common clinical usage. The toxicity profile for cisplatin was problematic because of the major gastrointestinal effects of the drug as well as the renal toxicity and neurotoxicity which were common and dose limiting. The problems of gastrointestinal toxicity of cisplatin has subsequently been largely overcome with the use of serotonin antagonists in conjunction with dexamethasone while the renal toxicity can be obviated for the most part with the use of optimal hydration and/or fractionation of the therapeutic dose over multiple days. One important aspect of cisplatin use is the rarity of clinically important neutropenia or thrombocytopenia permitting the use of full doses in combination with agents which do produce bone marrow suppression. The introduction of carboplatin to clinical trials was heralded because it simplified the use of outpatient therapy minimizing the need for hydration or the use of expensive anti-emetics and with a decreased incidence of nephrotoxicity, neurotoxicity and ototoxicity. Bone marrow suppression, however, is the major dose limiting toxicity for this agent with a delayed
nadir pattern similar to that seen with nitrosoureas and mitomycin C. This effect results in the possible need for substantial dose adjustment when used in conjunction with combination chemotherapy with resultant delays in treatment cycles. Because of the substantial differences in toxicity profile and presumed relative equivalency of activity, comparative trials of the two analogues have been relatively limited with the majority being carried out in ovarian cancer. The general sense has been that carboplatin is therapeutically equivalent to cisplatin and is accompanied by a toxicity advantage in terms of lesser nephrotoxicity and neuro-toxicity translating into an improved therapeutic index compared to cisplatin. However, the present detailed analysis of the literature offers another perspective. In two of the randomized trials in ovarian cancer, cisplatin was shown to be superior to carboplatin although the trials have some flaws as indicated previously. The other three trials in ovarian cancer demonstrate parity for the two platinum analogues in terms of therapeutic efficacy. In the trials in germ cell tumors and in head and neck cancer, cisplatin is the superior analogue. The comparative trials in lung cancer and bladder cancer similarly favor cisplatin although the numbers of patients accrued in the single bladder cancer trial is small and survival is generally comparable. In terms of toxicity, these randomized trials generally favor carboplatin with reduced renal and neurotoxicity and in two trials hematologic toxicity was also greater with cisplatin. A major criticism of the comparative trials of cis- and carboplatin may be directed at the inconsistency of carboplatin dose and the fact that the use of AUC dose determinations instead of the fixed dose per m2 schema was not applied in most of the trials. However, a recent large randomized trial comparing a carboplatin AUC of 4 to an AUC of 8, failed to demonstrate a difference in response rate. Therefore, it is unlikely that the dose of carboplatin employed in the randomized trials played a role in underestimating the therapeutic benefit of this analogue. Selection of the optimal platinum analogue to be utilized may be dependent on many factors including a) treatment goals (palliative vs. cure); b) companion drugs in combination; and c) disease to be treated. If the goal is palliation, carboplatin may be the preferable agent because it is associated with has less symptom producing toxicity. For drug combinations that involve neuropathic agents (for example paclitaxel) or may be influenced by renal function (for example topotecan), carboplatin may also be preferred. For some specific tumors such as germ cell malignancy, cisplatin is unequivocably superior and for the four other major tumors, cisplatin may be the preferable agent but the data is not compelling or definitive. Although definitive conclusions may not be acceptable to all clinicians with regard to the probable superiority of cisplatin over carboplatin, it is reasonable to interpret the existing phase III comparative trials of cis- and carboplatin as at least suggesting
20 that cisplatin could be the preferred agent in some, if not most, clinical settings and for most solid tumors. Acknowledgement Supported in part by The Cancer Center Research Foundation.
References 1. Horwich A. Carboplatin in the treatment of testicular cancer: The Royal Marsden Hospital Testicular Tumour Unit Experience. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 53-62. 2. Kavanagh JJ. Carboplatin in refractory epithelial ovarian cancer. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 141-6. 3. Trump DL, Elson P, Madajewicz S. A phase II trial of carboplatin in advanced transitional cell carcinoma of the urothelium. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 93-7. 4. Smith IE. Carboplatin in small-cell lung cancer: The Royal Marsden Hospital experience. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 243-7. 5. Green M, Kretsman H, Lutkart S et al. Carboplatin in nonsmall-cell lung cancer: The Cancer and Leukemia Group B experience. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 301-6. 6. Volling P, Schroeder M, Achterrath W. Phase II trial with carboplatin and 5-fluorouracil in previously untreated patients with advanced squamous cell carcinoma of the head and neck. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 213-9 7. Raghavan D, Gianoutsos P, Bishop J et al. Phase II trial of carboplatin in the treatment of advanced malignant mesothelioma. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 341-4. 8. Martinez JA, Morton G, Sanz GF et al. A phase II clinical trial of carboplatin infusion in high risk acute non lymphoblastic leukemia. J Clin Oncol 1991; 9: 39. 9. Lee E and Van Echo DA. The potential of carboplatin in human leukemia. Carboplatin (JM-8) Current Perspectives and Future Directions. Philadelphia: W.B. Saunders Company 1990; 361-5. 10. Ruckdeschel JC. The future role of carboplatin. Semin Oncol 1994; 21: 114-8. 11. Comis RL. Cisplatin: The future. Semin Oncol 1994; 21:109-13. 12. Weed DL. Methodologic guidelines for review papers. J Natl Cancer Inst 1997; 89(1). 13. Knox RJ, Friedlos F, Lydall DA et al. Mechanism of cytotoxicity of anticancer platinum drugs: Evidence that cis-diam-minedichloroplatinum (II) and cis-diammine(l,l-cyclobutane-diacarboxylate) platinum (II) differ only in kinetics of their interaction with DNA. Cancer Res 1986; 46: 1972-9. 14. Reed E, Dabholkar M, Chabner BA. In Chabner A and Longo D (eds): Platinum Analogues in Cancer Chemotherapy and Biotherapy. 2nd Edition. Philadelphia: Lippincott-Raven Publishers 1996. 15. Calvert AH, Newell DR, Grumbrell LA et al. Carboplatin dosage: Prospective evaluation of a simple formula based on renal function. J Clin Oncol 1989; 7: 1748-56. 16. Jodrell DI, Egorin MJ, Canetta RM et al. Relationships between carboplatin exposure and tumor response and toxicity in patients with ovarian cancer. J Clin Oncol 1992; 10: 520-8.
17. Van der Vijgh VJF. Clinical pharmacokinetics of carboplatin. Clin Pharmacokinet 1991; 21: 242-61. 18. Trump DL, Grem JL, Tutsch KD et al. Platinum analogue combination chemotherapy: Cisplatin and carboplatin - a phase I trial with pharmacokinetic assessment of the effect of cisplatin administration on carboplatin excretion. J Clin Oncol 1987; 5: 1281-9. 19. Patton JF, Powell BL, White DR et al. Combination cisplatin and carboplatin in advanced squamous cell carcinoma of the head and neck. Cancer Invest 1996; 14(2): 98-102. 20. Piccart MJ, Nogaret JM, Marcelis L et al. Cisplatin combined with carboplatin: A new way of intensification of platinum dose in the treatment of advanced ovarian cancer. J Natl Cancer Inst 1990; 82(8): 703-7. 21. Taylor AE, Wiltshaw E, Gore ME et al. Long-term follow up of the first randomized study of cisplatin versus carboplatin for advanced epithelial ovarian cancer. J Clin Oncol 1994; 12: 2066-70. 22. Adams M, Kerby U, Rocker I et al. A comparison of the toxicity and efficacy of cisplatin and carboplatin in advanced ovarian cancer. The Swons Gynacological Cancer Group. Acta Oncol 1989; 28(1): 57-60. 23. Mangioni C, Bolis G, Pecorelli S et al. Randomized trial in advanced ovarian cancer comparing cisplatin and carboplatin. J Natl Cancer Inst 1989; 81: 1464-71. 24. Swenerton K, Jeffrey J, Stuart G et al. Cisplatin-cyclophosphamide versus carboplatin-cyclophosphamide in advanced ovarian cancer: A randomized phase III study of the National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol 1992; 10: 718-26. 25. Ten Bokkel Huinink WW, Van der Burg MEL, Van Oosterom AT et al. Carboplatin combination therapy for ovarian cancer. Cancer Treat Rep 1988; 15(8): 9-15. 26. Alberts DS, Green S, Hannigan EV et al. Improved therapeutic index of carboplatin plus cyclophosphamide versus cisplatin plus cyclophosphamide. Final report by the Southwest Oncology Group of a phase III randomized trial in stages III and IV ovarian cancer. J Clin Oncol 1992; 10: 706-17. 27. Conte PF, Bruzzone M, Carmino F et al. Carboplatin, doxorubicin and cyclophosphamide versus cisplatin, doxorubicin and cyclophosphamide. A randomized trial in stage III—IV epithelial ovarian carcinoma. J Clin Oncol 1991; 9: 658-63. 28. Belpomme D, Bugat R, Rives M et al. Carboplatin versus cisplatin in association with cyclophosphamide and doxorubicin as first line therapy in stage III—IV ovarian carcinoma: Results of an ARTAC phase III trial. Proc ASCO 1992; 11: 227. 29. Edmonson JH, McCormack GM, Wieand HS et al. Cyclophosphamide-cisplatin versus cyclophosphamide carboplatin in myelosuppressive regimens. J Natl Cancer Inst 1989; 81: 1500-4. 30. Neijt JP, Hansen M, Hansen PG et al. Randomized phase III study in previously untreated epithelial ovarian cancer FIGO stage IIB, IIC, III, IV, comparing paclitaxel-cisplatin and paclitaxel-carboplatin. Proc ASCO 1997; 16: 352a. 31. Advanced Ovarian Cancer Trialists Group. Chemotherapy in advanced ovarian cancer: An overview of randomised clinical trials. BMJ 1991; 303: 884-93. 32. Vermorken JB, ten Bokkel Huinink WW, Eisenhauer EA et al. Advanced ovarian cancer. Carboplatin versus cisplatin. Ann Oncol 1993; 4 (Suppl 4): 41-8 (Review). 33. Gore ME, Fryatt I, Wiltshaw E, Dawson T. Treatment of relapsed carcinoma of the ovary with cisplatin or carboplatin following initial treatment with these compounds. Gynecol Oncol 1990; 36: 207-11. 34. Repetto L, Chiara S, Mammoliti S et al. Crossover study with cisplatin or carboplatin in advanced ovarian cancer. Gynecol Oncol 1990; 39: 146-9. 35. Alberts DS, Liu PY, Hannigan EV et al. Intraperitoneal cisplatin plus intravenous cyclophosphamide versus intravenous cisplatin plus intravenous cyclophosphamide for stage III ovarian cancer. N Engl J Med 1996; 335: 1950-5.
21 nomatous germ cell cancer: A multiinstitutional medical research 36. McGuire WP, Hopkins WJ, Brady ME et al. Cyclosphosphamide council/European organization for research and treatment of and cisplatin compared with paclitaxel and cisplatin in patients cancer trial. J Clin Oncol 1997; 15: 1844-52. with stage III and stage IV ovarian cancer. N Engl J Med 1996; 334:1-6. 46. Skalos DV, Samantas E, Kosmidis P et al. Randomized compari37. Kohn EC, Sarosy GA, Davis P et al. A phase I/II study of dosesion of etoposide-cisplatin vs. etoposide-carboplatin and irradiation in small-cell lung cancer. Ann Oncol 1994; 5: 601-7. intense paclitaxel with cisplatin and cyclophosphamide as initial therapy of poor-prognosis advanced-stage epithelial ovarian can- 47. Klastersky J, Sculier JP, Lacroix G et al. A randomized study cer. Gynecol Oncol 1996; 62: 181-91. comparing cisplatin or carboplatin with etoposide in patients with advanced non-small-cell lung cancer: European Organization for 38. Markman M. Carboplatin and cisplatin: Are they equivalent in Research and Treatment of Cancer Protocol 07861. J Clin Oncol efficacy in 'optimal residual' advanced ovarian cancer? J Cancer 1990; 8: 1556-62. Res Clin Oncol 1996; 122: 443-4. 39. De Andres L, Brunet J, Lopez-Pousa A et al. Randomized trial of 48. Kaye SB, Lewis CR, Paul J et al. Randomised study of two doses neoadjuvant cisplatin and fluorouracil versus carboplatin and of cisplatin with cyclophosphamide in epithelial ovarian cancer. fluorouracil in patients with stage IV-MO Head and Neck cancer. Lancet 1992; 340: 329-33. J Clin Oncol 1995; 13: 1493-500. 49. Kaye SB, Paul J, Cassidy J et al. Mature results of a randomized trial of two doses of cisplatin for the treatment of ovarian cancer. 40. Forestiere AA, Metch B, Schuller DE et al. Randomized comparison of cisplatin plus fluorouracil and carboplatin plus fluJClin Oncol 1996; 14: 2113-9. orouracil versus methotrexate in advanced squamous-cell carci- 50. Conte PF, Bruzzone M, Carnino F et al. High-dose versus lownoma of the head and neck: A southwest Oncology Group Study. dose cisplatin in combination with cyclophosphamide and epiJ Clin Oncol 1992; 10: 1245-251. doxorubicin in suboptimal ovarian cancer: A randomized study of the Gruppo Oncologico Nord-Ovest. J Clin Oncol 1996; 14: 41. Petrioli R, Frediani B, Manganelli A et al. Comparison between a 351-6. cisplatin-containing regimen and a carboplatin-containing regimen for recurrent or metastatic bladder cancer patients. A ran51. Thigpen JT. Dose-intensity in ovarian carcinoma: Hold, enough? domized phase II study. Cancer 1996; 77(2): 344-51. J Clin Oncol 1997; 15:1291-3 (Editorial). 42. Bajorin DF, Sarosdy MF, Pfister DG et al. Randomized trial of 52. Bertelsen JK, Anderson JE, Havsteen H et al. Dose-effect study etoposide and cisplatin versus etoposide and carboplatin in paof carboplatin in ovarian cancer: A Danish ovarian cancer group tients with good-risk germ cell tumors: A multiinstitutional study. study. J Clin Oncol 1997; 15: 193-8. J Clin Oncol 1993; 11: 598-606. 53. Alberts D. Cisplatin versus carboplatin in advanced ovarian cancer: An economic analysis. PharTher July 1994. 43. Bokemeyer C, Kohrmann O, Tischler L et al. A randomized trial of cisplatin, etoposide and bleomycin (PEB) versus carboplatin, etoposide and bleomycin (CEB) for patients with 'good-risk' Received 23 April 1997; accepted 31 July 1997. metastatic non-seminomatous germ cell tumors. Ann Oncol 1996; 7: 1015-21. 44. Bliss J, Horwich A. Carboplatin in the combination chemo- Correspondence to: therapy of non-seminomatous germ cell tumours. Ann Oncol Jacob Lokich, MD 1996; 7: 989-91 (Editorial). The Cancer Center 125 Parker Hill Avenue 45. Horwich A, Sleijfer DT, Fossa SD et al. Randomized trial of bleomycin, etoposide and cisplatin compared with bleomycin, Boston, MA 02120 etoposide and carboplatin in good-prognosis metastatic nonsemi- USA