Comprehensive Geriatric Assessment and Clinical Trials

The key consideration in developing trials for elderly patients is the effect of treatment on the patient’s overall function and well-being. Comprehensive geriatric assessment (CGA) provides as assessment of key domains related to quality of life and survival, including functional status, cognition, social support, psychological state (especially evaluation for anxiety and depression), nutritional status, and medication use. The CGA is a set of validated instruments that include evaluation of the activities of daily living (ADL): eating, bathing, dressing, toileting, and getting in and out of bed; and instrumental activities of daily living (IADL): managing finances, cooking, shopping, taking medications, performing housework, traveling, and communicating with the telephone. These data can help identify vulnerable patients—those most likely to experience toxicity—and shorter versions of the assessment that can be partially self-administered are now being tested in clinical trials. Preliminary data show that they are feasible in the cooperative group setting, and studies are underway to determine what components of these instruments can help predict which patients are at greatest risk for side effects. Moreover, a recent trial showed that geriatric assessment in older breast cancer survivors was not only predictive of poor tolerance of treatment as self-reported by patients, but also of mortality at 7 years. Further studies of this type are needed. Studies using the CGA as a research tool in elderly cancer patients have shown that it can independently predict survival, toxicity to chemotherapy, morbidity, and mortality. Also, incorporating quality-of-life assessment tools such as the Functional Assessment of Chronic Illness Therapy questionnaires ( www.facit.org ) into research trials in the elderly will help measure the impact of treatment on quality of life.

Although when one hears the term clinical trials one thinks of treatment trials, important trials now in progress are testing whether geriatric assessment can help predict treatment toxicity. Two recent trials showed that geriatric assessment when added to standard clinical variables (for example performance status and hemoglobin) can help accurately predict toxicity for older patients receiving chemotherapy both in the adjuvant and advanced setting. Extermann and colleagues assessed 518 patients 70 years and older who were initiating chemotherapy for both early and late-stage cancer. A score based on clinical and geriatric assessment data clearly predicted significant differences in hematologic and nonhematologic toxicity among cancer patients. A similar study by Hurria and colleagues in 500 patients also showed the added value of geriatric assessment data in predicting moderate and severe chemotherapy-related toxicity. These trials, and others in progress like these, are important. For example, several large cooperative group trials are now incorporating geriatric assessment prior to treatment and may allow for more accurate prediction of treatment-related toxicity for older patients treated with newer state-of-the-art regimens.

Eligibility Criteria

Eligibility criteria must be carefully considered when designing trials for older patients, and in most instances should be as broad as possible ; that is, “let doctors be doctors” and let doctors and patients together decide on what level of risk is appropriate. There should be no upper age restrictions. Instead, for adjuvant trials, older patients who are otherwise healthy and have life expectancies greater than 5 to 10 years should be offered participation. Using life expectancy makes much more sense and can be reasonably estimated. For trials where improving the probability of cure is not the goal, eligibility criteria should not exclude elders on the basis of arbitrary criteria such as organ dysfunction unless the specific treatment being studied is metabolized by or has an toxic effect on the particular organ. For instance, creatinine clearance decreases linearly with increasing age; arbitrarily adding criteria with a threshold for renal function to a trial that does not include treatment that is renally excreted should be avoided. Hematologic, hepatic, and cardiac function thresholds should also be omitted when not related to the treatment being evaluated. It is estimated that by appropriately relaxing eligibility criteria, participation of the elderly in clinical trials can be increased by 60%. Unless convincing data exist that support adding restrictive eligibility criteria on the basis of function, eligibility criteria should be as flexible as possible, allowing patients and their physicians flexibility in making decisions on trial participation.

Statistical Considerations

For state-of-the-art clinical trials, statistical considerations and getting an adequate sample size of older patients are also major concerns so as to make the outcome, and especially the toxicity data, generalizable to older patients. One strategy is to keep the elderly cohort open after the trial has met its major accrual goals so that one might reasonably determine that the major risks and benefits of any new treatments are similar for older and younger patients. A larger sample of older patients would be especially important in testing novel agents or procedures, as it would allow for adequate toxicity data to be gathered in this more vulnerable older population. Another strategy would be to require that a specific number of older patients be required in all phase II and III trials. This strategy, although tempting, might hamper completion of the trial, as extensive data show older patients are less likely to be offered clinical trials participation compared to younger patients. For these reasons, the strategy of leaving an elderly cohort open to evaluate a possible age-related treatment interaction appears a more practical and potentially more successful approach for future trials. Such a strategy could use an adaptive design based on the number of elders accrued to the trial and how many more elders should be accrued to better characterize any major toxicity among all the trial participants. For instance, if in a trial of 1000 patients, neutropenic fever was seen in 10% of the entire sample, and only 30 patients in the trial were 70 and older, one could leave the trial open for patients 70 and older to better determine a narrow confidence interval for this toxicity in the older age group.

Designing clinical trials specifically for the older cancer population should also be considered when there are potential differences in tumor biology with age (for example acute myelogenous leukemia, where the natural history of disease is different than in younger age groups), and where older patients—especially the frail and vulnerable—are not good candidates or are excluded from trials of regimens likely to be associated with major toxicity and loss of function. An example of a successful trial designed specifically for older patients was performed by the Cancer and Leukemia Group B and restricted entry to women 65 years and older with early-stage breast cancer. The plan made certain that an adequate sample of patients 70 and older would be accrued and used a novel adaptive Bayesian design to optimize the sample size. In addition, two companion trials, one assessing compliance with oral chemotherapy and another evaluating the effect of the different treatments on quality of life, were made optional but highly recommended parts of the trial; both successfully met their accrual goals. The advantage of this approach is that such trials can focus on effective but potentially less-toxic treatments, can include specific assessments such as CGA instruments for identifying patients likely to be the most vulnerable to side effects, and can include or be restricted to the frail elderly. For trials focused on vulnerable or frail populations, it is important that clear and reproducible definitions be used to define the population at risk.

Translational Research

Opportunities to further understand the effects of cancer treatment in the elderly may lie in evaluating biomarkers of aging, cytokine regulation, and the molecular interactions of cancer and age. For example, there is evidence that interleukin-6 (IL-6 , an inflammatory cytokine that promotes differentiation of T cells and B cells, activation of T cells and macrophages, and secretion of immunoglobulin) increases during aging. Increased IL-6 expression has also been found in certain cancers, such as multiple myeloma, lymphoma, Hodgkin lymphoma, renal cell carcinoma, chronic lymphocytic leukemia, and breast cancer. Older patients with cancer and high IL-6 levels might be considered for clinical trials to determine their safety and efficacy in targeting cancer. In addition, IL-6 might serve as a marker of physiologic reserve and add to information obtained by geriatric assessment in predicting toxicity. Another exciting molecular marker of aging is p16 gene expression, which increases tenfold between ages 20 and 80 years. Increased p16 expression is associated with cell senescence and may possibly prove to predict organ-related toxicity from radiation and chemotherapy.

Future studies might also address host factors related to drug activation and metabolism as related to clinical outcomes. Although these issues are not specific for older patients, they are of major importance. For example, the cytochrome P450 (CYP450) metabolic enzyme CYP2D6 has a major role in tamoxifen metabolism, activating tamoxifen to endoxifen, its most active metabolite. The CYP2D6 gene is polymorphic, but even the wild-type variant is affected by many antidepressants, medications that are commonly used in older patients, and which cause a decrease in the conversion of tamoxifen to endoxifen. These data point out the importance of trials that measure pharmacokinetic and pharmacogenomic parameters in older patients, especially in those taking medications that might interact with enzymes important in drug activation and metabolism.