Fax: (011) 81-3-6409-2069
Age disparity between a cancer population and participants in clinical trials submitted as a new drug application of anticancer drugs in Japan
Version of Record online: 3 MAY 2007
Copyright © 2007 American Cancer Society
Volume 109, Issue 12, pages 2541–2546, 15 June 2007
How to Cite
Hori, A., Shibata, T., Kami, M., Kusumi, E., Narimatsu, H., Kishi, Y., Murashige, N. and Mori, K. (2007), Age disparity between a cancer population and participants in clinical trials submitted as a new drug application of anticancer drugs in Japan. Cancer, 109: 2541–2546. doi: 10.1002/cncr.22721
- Issue online: 4 JUN 2007
- Version of Record online: 3 MAY 2007
- Manuscript Accepted: 8 FEB 2007
- Manuscript Revised: 5 FEB 2007
- Manuscript Received: 5 JAN 2007
- elderly patients;
- anticancer drug;
- clinical trial;
- New Drug Application
Underrepresentation of older patients in cancer clinical trials has been reported previously.
To evaluate disparities in age between actual cancer patients and those enrolled in clinical trials, the authors examined all the review reports of the Pharmaceuticals and Medical Devices Agency, Tokyo, Japan, and summaries of data submitted by applicants for the approval of new cancer drugs and that of a partial change in approved cancer drugs.
Information regarding 68 clinical trials was available on the Internet. The median age of trial participants ranged from 33 years to 73 years and was older than 65 years in 13 trials, whereas the estimated median age of patients with all cancers was 69 years, and 64% of these individuals were age ≥65 years. The median age of trial participants was found to be lower than that of the patient population in 60 trials. The median difference in age between the 2 groups was 7 years (range, −16 to +33). With regard to molecular-targeting agents (16 trials) and hormonal agents (10 trials), trial participants were younger than the patient population in 25 of the 26 trials, with a median difference of 6 years (range, −9.5 to +20). The difference was larger for molecular-targeting agents (median, 9.5 years; range, birth-20 years) compared with hormonal agents (median, 2 years; range, −9.5 to +15).
The results of the current study show that participants in cancer clinical trials are younger than the actual Japanese cancer patient population. Cancer 2007. © 2007 American Cancer Society.
Cancer is a significant problem in the developed countries. Approximately 60% of all new cases of cancer and 70% of all cancer-related deaths occur in the elderly population, which is defined as individuals aged ≥65 years.1 However, elderly patients are often underrepresented in clinical trials of new cancer drugs.2–5
Currently, there is no age limitation for clinical trials submitted as a New Drug Application (NDA). In Japan, the Ministry of Health and Welfare issued a guideline that defined older people as those aged ≥65 years and stated that even though age is not the only predictive factor for disease occurring in the elderly, the age distribution of individuals participating in clinical trials should be determined based on their predicted age after study approval and that the enrollment of patients aged ≥75 years must be considered. They added that for the most part it is inappropriate to establish an arbitrary age limitation in clinical trial protocols.6 When elderly patients are treated, one cannot neglect the following factors: the biologic nature of tumors, immunologic and organ functions,7 complications and comorbidity, potential drug interactions with medications for complications, and comorbidity.8 Therefore, trials in nonelderly patients may be insufficient for determining appropriate treatments for the elderly. Therefore, it is necessary to promote clinical trials for older patients.
Attempts to promote clinical trials for older patients are necessary; however, we have little information regarding the age of the participants for anticancer drugs approved in Japan. In the current study, we evaluated the disparity in age between the actual Japanese cancer patient population and patients enrolled in clinical trials submitted as a NDA.
MATERIALS AND METHODS
Data Source Regarding Clinical Trials
To investigate the age of patients enrolled in clinical trials submitted as NDA that were evaluated primarily in the review process in each application, we examined all the review reports of the Pharmaceuticals and Medical Devices Agency (PMDA) and summaries of the data submitted by applicants for the approval of new cancer drugs and those data for a partial change in approved cancer drugs (eg, indication and dosages). Since September 1999, all the reports and summaries have been disclosed on the websites of the PMDA,9 the National Institute of Health Sciences,10 and the Japan Pharmacists Education Center.11 As of April 2005, information concerning 45 applications of anticancer agents was available. Of these, we analyzed 43 applications containing 68 clinical trials that were approved for their antitumor effects (Table 1). Review of a NDA is within the jurisdiction of the PMDA, which is authorized by the Ministry of Health, Labor, and Welfare of Japan.12
|Generic name||Indication||Types of clinical trial||Median age of trial participants, y (A)||Median age of the cancer patient population (B)||(B) – (A), y||Age limitation in clinical trial protocol, y|
|Tamibarotene||Acute promyelocytic leukemia||Phase 2||47.4||66||18.6||15–75|
|Oxaliplatin||Colorectal cancer||Phase 2||61||70||9||20–75|
|Arsenic trioxide||Acute promyelocytic leukemia||Phase 1/2||33||66||33||None|
|Long-term (Phase 1/2)||40||66||26||Unknown|
|Long-term (Phase 3)||44.5||66||21.5||Unknown|
|Cisplatin (arterial injection)||Hepatocellular carcinoma||Phase 1/2I||64||70||6||75|
|Tegafur, 5-fluorouracil, levofolinate calcium||Colorectal cancer||Phase 3||64||70||6||None|
|Capecitabine||Breast cancer||Phase 2||57||57||0||20–75|
|Amrubicin hydrochloride||Lung cancer||Phase 2||71||73||2||79|
|Cladribine||Hairy cell leukemia||Phase 2||59.5||66||6.5||85|
|Topotecan||Small cell lung cancer||Phase 2||63.8||73||9.2||80|
|Gemcitabine hydrochloride||Pancreatic cancer||Phase 2||62||73||11||None|
|Paclitaxel||Uterine endometrial cancer||Phase 2||60||54||−6||Unknown|
|Vinorelbine ditartate||Breast cancer||Phase 2||55||57||2||75|
|Tegafur, gimeracil, oteracil, potassium||Nonsmall cell lung cancer||Phase 2||67.5||73||5.5||Unknown|
|Docetaxel||Esophageal cancer||Phase 2||64||69||5||Unknown|
|Cladribine||Malignant lymphoma||Phase 2||59.5||66||6.5||85|
|Fludarabine||Chronic lymphocytic leukemia||Phase 1/2||62.1||66||3.9||None|
|Etoposide||Cervical cancer||Phase 2||70||54||−16||80|
|Cytarabine||Acute leukemia||Phase 2||37||66||29||70|
|Molecular-targeting agents and hormonal agents|
|Gefitinib||Nonsmall cell lung cancer||Phase 2||60.5||73||12.5||None|
|Leuprolide acetate||Prostate cancer||Phase 2||73.1||75||1.9||90|
|Goserelin acetate||Prostate cancer||Phase 2||60||75||15||None|
|Anastrozole||Breast cancer||Phase 2||62.2||57||−5.2||51–80|
|Imatinib mesilate||Chronic myelogenous leukemia||Phase 2||57||66||9||None|
|Trastuzumab||Breast cancer||Phase 2||48||57||9||18–75|
|Repeated dose open trial||50||57||7||None|
|Rituximab||Malignant lymphoma||Phase 2||52||69||17||15–75|
We excluded those applications based on procedures for the approval of off-label usages with the exception of arsenic trioxide because clinical trials submitted as NDA were not performed for this agent. We included the data regarding arsenic trioxide for the treatment of acute promyelocytic leukemia because the PMDA reviewed the clinical trial data, which were submitted to the European Medicines Agency. We also excluded phase 1 clinical trials, given the nature of these trials.
Comparison of the Median Age Between the Japanese Cancer Patient Population and Participants in Clinical Trials
For the studies in which multiple treatment groups were compared,1 we used the median age of the entire group of enrolled patients2; when this information was unavailable, we used the mean of the median ages of each group. We used mean ages in cases in which it was the only available information.
For each type of cancer, the median age of the patient population were estimated as follows. We obtained the age-specific incidence from Cancer Statistics in Japan 200313 and the population by age group from the Annual Report on Health, Labour and Welfare 2003–2004,14 and multiplied them to estimate the number of age-specific, newly diagnosed cancer patients. The median age of the cancer patients was computed based on the number of age-specific, newly diagnosed patients. For each type of cancer, we compared the estimated median age of the cancer patient population with the median age of participants in clinical trials submitted as NDA.
The age-specific incidence rates in the Japanese cancer patient population were available for leukemia; lymphoma; and cancers of esophagus, stomach, intestine, liver, pancreas, lung, breast, prostate, head and neck, and uterus (Table 2). The median age of the patient population with all cancers was 69 years (range, 54–75 years) and 64% of these patients were aged ≥65 years. Patients with cancers of the breast and uterus were younger compared with those with other cancers (median age of 57 years vs a median age of 54 years).
|Cancer site||Estimated no. of newly diagnosed patients each year (×1000)*||Median age, y|
|Head and neck (mouth, pharynx, and larynx)||13||67|
The median ages of both the patient population and the clinical trial participants were available for 68 clinical trials of 26 applications. The median age of the clinical trial participants ranged from 33 years to 73 years, and was older than 65 years in 13 trials (19%). Thirty of the 57 evaluable trials had age-related inclusion/exclusion criteria (Table 3).
|Type of drug||No. of applications|
|Molecular-targeting agents and hormonal agents||27|
The median difference in age between the clinical trial participants and the cancer patient population was 7 years in 68 trials (range, −16 to +33 years) (Table 4). The median age of the clinical trial participants was lower than that of the patient population in 60 of the 68 clinical trials, whereas the clinical trial participants were older than the cancer patient population in 5 trials. These included 3 trials of anastrozole for breast cancer, 1 trial of paclitaxel for endometrial cancer, and 1 trial of etoposide for cervical cancer. As shown in Table 4, the age differences were minimal in the clinical trials of breast and uterine cancer, whereas they were large in clinical trials of leukemia. The median age of clinical trial participants was 33 years younger than that of the cancer patient population in the clinical trial of arsenic trioxide for the treatment of acute promyelocytic leukemia.
|Cancer site||Median age of cancer patient population, y||Total no. of trials||Age differences: median ages of the clinical trial participants minus those of the cancer patient population|
|<0 y||<5 y||5–9 y||10–14 y||15–19 y||>20 y|
|Head and neck||67||2||0||0||2||0||0||0|
The current study included applications of 4 molecular-targeting agents (17 trials) and 3 hormonal agents (10 trials), which are expected to have low toxicities. Nine of the 27 trials regarding hormonal and molecular-targeting agents had age restrictions in their eligibility criteria, whereas 21 of the 30 evaluable trials regarding cytotoxic agents had age-related eligibility criteria (P = .0081). Differences in the median age between the cancer patient population and the clinical trial participants for molecular-targeting and hormonal agents are shown in Figure 1. With the exception of a clinical trial of anastrozole, clinical trial participants were younger than the cancer patient population. The median age difference between the cancer patient population and the clinical trial participants was 6 years (range, −9.5 to +20 years). The age difference was found to be larger in trials of molecular-targeting agents (median, 9.5 years [range, birth-20 years]) compared with hormonal agents (median, 2 years [range, −9.5 to +15 years]).
The results of the current study demonstrated that the median age of participants in clinical trials is younger than that of actual cancer patients in the majority of clinical trials submitted as NDAs in Japan. The differences in the median age between the participants in cancer clinical trials and the cancer patient population were all >20 years in 6 trials, all of which examined the treatment of acute leukemia. The low rates of enrollment of older cancer patients have been attributed to physicians' judgment that the risk of toxicity is high in elderly patients due to their reduced physical function and comorbidity, and to the difficulty of older patients in meeting the eligibility criteria.4, 15 Moreover, the higher number of participants and the higher cost are necessary due to high numbers of dropouts and adverse events if elderly patients are enrolled more frequently. Considering that, at the current time, the median age of Japanese cancer patients is 69 years, and that as many as 64.4% of these patients are aged ≥65 years, we should be careful in interpreting the results of the clinical trials submitted as NDAs.
Cancer treatment has recently shifted from toxic drugs to less toxic treatment, including molecular-targeting agents and hormonal agents; however, to our knowledge, there are few data available regarding the clinical trial participants for these agents. The current study included 10 trials of hormonal agents (eg, leuprorelin and goserelin for prostate cancer and anastrozole for breast cancer). The disparity in the median age between Japanese cancer patients and the trial participants was small with regard to hormonal treatment (median, 2 years [range, −9.5–15 years]) (Fig. 1). Because the major adverse effects of hormonal agents are usually mild,16, 17 appropriate representatives of the patient population may be enrolled in clinical trials of these agents.
However, this is in contrast to the clinical trials of molecular-targeting agents, which the majority of clinicians believe are more specific and less toxic than conventional cytotoxic agents. With regard to these drugs, the median difference in age between the trial participants and the cancer patient population was 9.5 years (range, birth to 20 years). These disparities are observed in the U.S. as well as in Japan. For example, the median age of patients who were enrolled in pivotal trials of rituximab submitted to the U.S. Food and Drug Administration (FDA)18–20 ranged from 55 years to 59 years, whereas the median age of patients with malignant lymphoma was 66 years.21 The mean age of patients who were enrolled in the pivotal trial of trastuzumab for breast cancer submitted to the FDA was 53 years,22 whereas the median age of patients with breast cancer was 61 years. However, previous studies have demonstrated that molecular-targeting agents have substantial toxicities.23–25 These toxicities were underestimated due to the age disparities. Further clinical trials are required to evaluate the safety of these agents for elderly patients.
Several hypotheses can be put forward for these disparities. First, because the target antigens of molecular-targeting agents are frequently expressed in normal tissues, clinicians might have a concern that these drugs may damage the healthy tissue. Several reports have been published regarding the adverse events of molecular-targeting agents, including the cardiotoxicity of trastuzumab,23 pulmonary fibrosis resulting from gefitinib,24 and multiorgan failure occurring after the administration of TGN1412.25 Second, eligibility criteria for molecular-targeting agents might be more strict than those for hormonal agents; however, there was no significant difference noted with regard to age restrictions between clinical trials of molecular-targeting agents and those of hormonal agents. It is unlikely that the age disparities between the clinical trial participants and the cancer patient population in clinical trials of molecular-targeting agents were attributable to their eligibility criteria. Finally, we have little experience regarding the clinical use of molecular-targeting agents, and to our knowledge little information is available concerning their safety. These situations are in contrast to hormonal agents, which have been utilized in many clinical settings other than cancer chemotherapy. Many physicians may be concerned about unexpected toxicities and therefore might avoid the enrollment of elderly patients into clinical trials of molecular-targeting agents.
In conclusion, the results of the current study found that participants in clinical trials of anticancer drugs are younger than the actual Japanese cancer population. Approved drugs may possibly be administered to older patients who have not been evaluated in clinical trials submitted as NDA. The development of a methodology of clinical trials to evaluate the efficacy and safety of anticancer drugs specifically in older patients is awaited.
- 6Health Policy Bureau. Studies in support of special populations: geriatrics, 1993. PAB dated Dec. 2 (ICH-E7 Guideline): Notification No. 104 of the New Drugs Division.
- 13Cancer Statistics in Japan-2003. Tokyo: Foundation for Promotion Cancer Research; 2003.
- 14Ministry of Health Labor and Welfare. Annual Report on Health, Labour and Welfare 2003–2004. Tokyo: Gyosei; 2004.
- 21National Cancer Institute Surveillance Epidemiology and End Result. Available at: http://seer.cancer.gov/statfacts/html/nhl.html. Accessed December 25, 2006.