- © 2005 by American Society of Clinical Oncology
Influence of Sex on Toxicity and Treatment Outcome in Small-Cell Lung Cancer
- Simron Singh,
- Wendy Parulekar,
- Nevin Murray,
- Ronald Feld,
- William K. Evans,
- Dongsheng Tu and
- Frances A. Shepherd
- From the Division of Medical Oncology, Department of Medicine, Princess Margaret Hospital; University of Toronto; Cancer Care Ontario, Toronto; the National Cancer Institute of Canada Clinical Trials Group; Queens University, Kingston, Ontario; and the British Columbia Cancer Agency, Vancouver, British Columbia, Canada
- Address reprint requests to Frances A. Shepherd, MD, FRCP, Princess Margaret Hospital, Ste 5-104, 610 University Ave, Toronto, Ontario, Canada, M5G 2M9; e-mail: frances.shepherd{at}uhn.on.ca
Abstract
Purpose Female sex has been shown consistently to be a favorable prognostic factor in small-cell lung cancer (SCLC). Studies have shown that women with other tumor types experience greater treatment toxicity, but there have been few studies of sex-related toxicity in SCLC.
Patients and Methods This was a sex-based retrospective analysis of four SCLC trials conducted by the National Cancer Institute of Canada Clinical Trials Group between 1987 and 1999. The 1,006 patients (648 males and 358 females) received similar chemotherapy consisting of cyclophosphamide-doxorubicin-vincristine and etoposide-cisplatin. Toxicities examined included myelosuppression, stomatitis, vomiting, and infection. Other end points included dose reductions and omissions, response, and survival.
Results Women experienced significantly more hematologic toxicity than men (grade 3 and 4 anemia, 16.3% v 7.6%, respectively, P < .001; grade 3 and 4 leukopenia, 80.4% v 69.2%, respectively, P = .0001). However, toxic death rates were similar for men and women (1.5% v 1.1%, respectively, P = .58). Women also had significantly more stomatitis and vomiting of all grades. Despite increased toxicity, 76% of females versus 73.4% of males received all six treatment cycles (P = .38), but 52% of females versus 43.4% of males had treatment delayed for 2 weeks or more (P = .022). Only 31.8% of females and 28.2% of males had at least one cycle of chemotherapy dose reduction (P = .23). The overall response rate was 80.3% for females and 66.9% for males (P < .0001), and the median survival time was 1.31 years for females compared with only 0.91 year for males (P < .0001).
Conclusion Women experience more chemotherapy-related toxicity in the treatment of SCLC, but they also have increased response rates and survival.
INTRODUCTION
Lung cancer continues to be a serious global health problem and is the leading cause of cancer death worldwide. This year, it will account for close to 174,000 and 19,000 deaths in the United States1 and Canada, respectively.2 In the developed world, the incidence of lung cancer is increasing among women but seems to have stabilized among men.3 This may be a result of changing smoking patterns in the female population, but it may also be a result of genetic and molecular mechanisms that make female smokers more susceptible to lung cancer than male smokers.4
Small-cell lung cancer (SCLC) accounts for approximately 15% to 20% of all bronchogenic carcinomas. Demographic studies suggest that women are more likely than men to suffer from SCLC, whereas men are more likely to be diagnosed with non–small-cell lung cancer.3 Interestingly, sex has been found to be an independent prognostic factor for survival for all subtypes of lung cancer, with women surviving longer than men (P < .0001).5 Specifically, with respect to SCLC, prognostic studies of several of the large cooperative group databases have shown in both univariate and multivariate analyses that female sex is associated with a favorable outcome.5-10 However, this has not been observed in all analyses.11
Treatment of both limited- and extensive-stage SCLC involves chemotherapeutic drugs that are associated with significant side effects. In fact, virtually all patients treated will experience some form of toxicity. Patients give informed consent to treatment after a discussion based on the risks and benefits, including an assessment of treatment-related toxicity. Generally this information is applied to the population as a whole, and little, if any, information is given to patients on the role sex can play in treatment toxicity. Traditionally, results from clinical trials and information on drug pharmacokinetics and toxicity have been generalized and applied to both sexes. Only since 1993 has the American Congress required sex-specific analysis of the safety and efficacy of drugs in clinical trials. Legislation now states that clinical trials must be “designed and carried out in a manner sufficient to provide for an analysis of whether the variables being studied in the trial affect women differently than other subjects in the trial.”12
Emerging data suggest that sex-specific pharmacokinetics and toxicity exist for certain chemotherapeutic drugs. Clearance rates and differences in area under the plasma concentration-time curve (AUC) ratios have been reported for doxorubicin and etoposide, which are two drugs that are used frequently in SCLC.13,14 Similarly, Milano et al15 have shown that the capacity to clear fluorouracil is lower in women compared with men, and sex-related differences in drug toxicity profiles and significant variations in both the frequency and severity of toxicity have been found in men and women.16-19 Although few sex-related studies of cisplatin have been undertaken, at least one study has shown that women experience significantly more nausea and vomiting after cisplatin than men.20,21
On the basis of these observations derived from the treatment of other tumors, we hypothesized that women would also experience more frequent and more severe toxicity than men during SCLC chemotherapy. We elected to undertake this sex-based retrospective analysis of treatment delivery and toxicity in four prospectively randomized SCLC trials conducted by the National Cancer Institute of Canada Clinical Trials Group and to correlate these parameters with outcome. The trials were selected because patients in seven of the eight study arms examined received similar chemotherapy regimens with variations only in schedule and radiation treatment. We report here the results of this analysis.
PATIENTS AND METHODS
Studies Included
Table 1 lists the trials included in this review.22-25 All studies were prospectively randomized controlled trials of treatment in limited or extensive SCLC and were performed between 1987 and 1999. Sample sizes were similar in the trials, and all studies had a larger proportion of men enrolled. Eligibility required documentation of SCLC and an Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 to 3. Patients aged 80 years or older were excluded from all studies except BR.8, in which patients older than 68 years were excluded. The staging evaluations and follow-up visits were similar in the studies and are detailed in each of the study reports. Ninety of the 1,206 enrolled patients (7.6%) were found to be ineligible. The 110 patients on the cyclophosphamide-oncovin-doxorubicin-etoposide (CODE) arm of the BR.8 trial were not included because of the marked differences in dose and schedule of drugs in that arm of the trial. The remaining 1,006 patients were included in the analysis.
National Cancer Institute of Canada Clinical Trials Group, SCLC Trials Included
Treatment
The chemotherapy regimens and doses used in the four trials are listed in Table 2. BR.3 compared three cycles of cyclophosphamide, doxorubicin, and vincristine (CAV) followed by three cycles of etoposide and cisplatin (EP) with alternating CAV with EP for three cycles each in limited SCLC. After three courses of chemotherapy, responding patients underwent prophylactic cranial irradiation (PCI) on days 8 through 15 (20 Gy in five fractions). Responding patients were further randomized after the sixth cycle of chemotherapy to receive thoracic radiotherapy consisting of either 37.5 Gy in 15 fractions or 25 Gy in 10 fractions.
Treatment Arms of Trials
BR.4 compared CAV every 3 weeks for six cycles to three cycles of CAV alternating with three cycles of EP in patients with extensive SCLC. Stable and responding patients received PCI on day 8 of cycle 3.
BR.6 compared the timing of thoracic irradiation in limited-stage SCLC. All patients received three cycles of CAV alternating with three cycles of EP and were randomly assigned to receive early thoracic irradiation (40 Gy in 15 fractions) administered concurrently with the first cycle of EP (cycle 2) or with the last cycle of EP (cycle 6).
BR.8 compared three cycles of CAV alternating with three cycles of EP to a dose-intensive weekly chemotherapy regimen in extensive SCLC. Patients with a complete response received PCI (25 Gy in 10 fractions) after chemotherapy. The intensive treatment arm was not included in our analysis.
Assessment
The primary end points of the study consisted of hematologic toxicity (leukocyte count, hemoglobin, and platelets), stomatitis, vomiting, and infection. Toxicity was reported and graded according to the National Cancer Institute of Canada Clinical Trials Group expanded common toxicity criteria from grades 0 to 4. All toxicities (grade 1 and above), as well as severe toxicities (grade 3 and above), were examined.
Secondary end points included assessment of the number of dose reductions required and the number of omitted cycles. Response rates, survival, and toxic deaths were also examined. Standard response criteria were used, and survival was measured from the date of randomization until the date of death or last follow-up examination.
Statistical Analyses
Fisher’s exact test was used to compare toxicity between the sexes. Logistic regression was performed to examine for potentially confounding covariates, including age, body surface area (BSA), PS, pretreatment lactate dehydrogenase (LDH), and individual trial. Survival curves were estimated by the Kaplan-Meier method.26 The total dose and dose reductions of chemotherapy were calculated using all included patients from the start to completion of the last cycle of chemotherapy.
RESULTS
Patient Demographics
The baseline pretreatment patient characteristics are listed in Table 3. There were more male patients over age 65 years and with an ECOG PS of 2 (P = .0064 and P = .026, respectively). As expected, very few patients had a BSA greater than 2.0 m2.
Baseline Patient Characteristics
Toxicity
Hematologic toxicities of anemia (55.9% of females v 38.6% of males, P < .0001) and leukopenia (92.7% of females v 84.7% of males, P = .0002) were significantly higher in female patients (Table 4). This was consistent when examined both for all toxicity levels and for grade 3 and 4 toxicity. Despite greater leukopenia, there was no significant difference in infection rates or toxic death rates (10 deaths in males, 1.5%; and four deaths in females, 1.1%; P = .58).
Comparison of Toxicity by Patient Sex
Females also experienced more nonhematologic toxicities than men (Table 4), including stomatitis (16.8% v 8.3%, respectively, P < .0001) and vomiting (76.8% v 67.2%, respectively, P = .0014). Logistic regression analysis was performed to adjust the comparisons between sexes by the following baseline characteristics: age (≤ 65 v > 65 years), baseline BSA (< 2 v ≥ 2 m2), PS (ECOG 0 to 1 v 2 to 3), baseline LDH, and trial (Table 5). After adjusting for the effect of these variables, female sex was still a significant predictor for increased anemia, leukopenia, stomatitis, and vomiting.
Logistic Regression: Grade 1 or Higher Toxicity
Chemotherapy Treatment Delivery
Treatment received and dose reductions required are listed by sex in Table 6. Despite experiencing more toxicity, 76% of female and 73.4% of male patients received all six cycles (P = .38). Only 31.8% of females and 28.2% of males had one or more cycles of chemotherapy reduced (P = .23). However, more females than males had treatment delays of 2 weeks or more (52% v 43.4%, respectively, P = .022), likely because of their increased toxicity.
Comparison of Treatment Delivery, by Sex
Response and Survival
Females had a superior overall response rate (Table 7) compared with males (80.3% v 66.9%, respectively, P < .0001), and they also had more complete responses. As shown in Figure 1, women had significantly better survival (P < .0001). After adjusting for age, baseline BSA, PS, baseline LDH, and trial, female sex was still a highly significant predictor for survival.
A comparison of survival for male and female patients. HR, hazard ratio.
Comparison of Response and Survival, by Sex
DISCUSSION
The results of this study provide strong evidence to support the hypothesis that women experience more chemotherapy-related toxicity than men during the treatment of SCLC. Furthermore, the differences identified were highly significant and remained significant after adjusting for other important prognostic variables including age, PS, and pretreatment LDH.
Sex-related toxicity has been reported previously for fluorouracil in the treatment of gastrointestinal cancers, where it has been shown that females experienced greater hematologic toxicity and nonhematologic toxicity including stomatitis, alopecia, nausea, vomiting, and diarrhea.16-19
Patients in our study received two regimens, CAV and EP. Few sex-related studies of cisplatin have been undertaken, but one study has shown that women experience significantly more nausea and vomiting after cisplatin than men.20 This is consistent with our results. Another study showed that female sex was a negative predictor for complete antiemetic control with 5-hydroxytryptamine-3 antagonists in patients receiving doxorubicin or cyclophosphamide.21 However, sex differences have not been observed with the new class of antiemetics, neurokinin-1 antagonists.27 Pharmacokinetic studies of the chemotherapeutic agent etoposide have been shown to vary slightly in women, although significant differences in AUC ratios have not been reported.14 However, it is possible that small differences in etoposide AUC may have contributed to the excess toxicity experienced by the women in our analysis.
There are fewer sex-related studies for the agents in the CAV regimen. Dobbs et al13 showed that in patients with normal liver biochemistry, men had a higher rate of clearance of doxorubicin than women. Similarly, older women have been shown to have a lower rate of clearance of another anthracycline, epirubicin, than do men of any age.28 As with etoposide, lower clearance and a higher doxorubicin AUC might be expected to result in excess toxicity for women treated with the CAV regimen. Gemcitabine, a drug that is used extensively in non–small-cell lung cancer, although not in SCLC, has been reported to have a higher clearance rate in men when compared with women of the same age.29 There have been no published sex-based studies of vincristine. However, rates of neuropathy were not significantly different in men and women in our study.
Another possible reason for the increased toxicity seen in women may have to do with influence of body mass index (BMI). Women generally have a higher baseline BMI than men because of increased body fat,30 and this may affect drug distribution and potentially increase toxicity. Obesity and high BMI have been shown to be associated with a decrease in drug clearance and a resultant increase in the elimination half-life for both doxorubicin and cyclophosphamide.31,32 A study of 262 patients treated for limited and extensive SCLC found no effect of BMI levels on toxicity.33 However, this study was not designed to evaluate sex differences, but rather, it evaluated obesity.
It has been suggested that women overreport morbidity and disability compared with men, resulting in a reporting bias with respect to toxicity.34 However, this has been shown not to be the case when studied and when other possible confounding variables are examined.35,36 Clearly, this was not the case in our study because women had higher toxicity rates and grades not only in subjective areas, such as nausea, but also in objective areas, including leukopenia and anemia, in which self-reporting of symptoms is not subject to patient bias.
Our results show that women not only experience more frequent toxicity, but they also suffer from more severe toxicity. Despite this, the overall doses of chemotherapy delivered to females were not significantly different from those received by their male counterparts. Although more female patients required dose delays because of toxicity, treatment compliance remained high, with equal proportions of men and women completing six cycles of chemotherapy. Despite similar treatment delivery, females in our study had significantly better outcomes with respect to both response rates and overall survival. These results are consistent with the previous studies7-9 that have shown that female sex is one of the strongest prognostic indicators in SCLC. These findings raise the possibility of sex-based differences in drug pharmacodynamics for agents used in SCLC. Sex-based differences in outcome may also be a result of altered tumor biology between the sexes, although this remains unproven. Women smokers have been shown to have a higher rate of some forms of lung cancer, such as adenocarcinoma, thus raising the possibility of estrogens playing a role in carcinogenesis.37 Women smokers also seem to be more susceptible to the carcinogenic effects of smoke than men.3 Finally, men have been shown to have a greater allele loss for marker HSAS on chromosome 11p(LOH11B), which has been associated with a poorer survival in non–small-cell lung cancer.38 Thus, it is possible that inherent biologic differences may also play a role in explaining toxicity and efficacy differences between the sexes, and further study is warranted.
Further investigation into the mechanism for increased toxicity in women is warranted to determine whether differences between the sexes may be seen in pharmacokinetics and pharmacogenomic profiles. Our current system of dosing by BSA, although well established, has never been validated in adults29 and is not sex specific. It may be a suboptimal tool for maximizing drug activity and decreasing side effects. Recent attempts have been made to develop individualized dosing of drugs, such as amonafide, based on a pharmacodynamic model involving phenotype, sex, and pretreatment WBC levels.39 Carboplatin dosing is also often adjusted for sex because of the inclusion of glomerular filtration rate calculations in the Calvert formula.40 It is possible that such individualized dosing protocols for the other drugs used in the treatment of SCLC might minimize the potential for toxicity in women.
In conclusion, women experience more toxicity when treated for SCLC, although they do not receive fewer chemotherapy cycles, and they have an improved outcome compared with men. The cause for this is likely multifactorial, and further studies are needed to identify the mechanisms that lead to these sex-specific effects.
Authors’ Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Footnotes
-
Authors’ disclosures of potential conflicts of interest are found at the end of this article.
- Received March 27, 2004.
- Accepted November 2, 2004.
