A Simple Tool For Predicting Chemotherapy Toxicity in Older Adults
Abstract & Commentary
By Gary Shapiro, MD, Director of Medical Oncology, Cancer Center of Western Wisconsin. Dr. Shapiro reports no financial relationships relevant to this field of study.
Synopsis: This prospective multicenter study presents an 11-item model for predicting chemotherapy toxicity in older adults with cancer. Its stratification schema identified older adults at low (30%), intermediate (52%), or high (83%) risk for chemotherapy toxicity.
Source: Hurria A, et al. Predicting chemotherapy toxicity in older adults with cancer: A prospective multicenter study. J Clin Oncol 2011;29:3457-3465.
The seven institution Cancer and Aging Research Group recruited 500 cancer patients (age 65 or older) who were about to receive a new chemotherapy regimen in an outpatient setting and had them complete a geriatric assessment before starting treatment. The patients had a mean age of 73, a variety of cancers (29% lung, 27% gastrointestinal, 17% gynecological, 11% breast, genitourinary 10%), and 61% had stage IV disease.
The assessment included functional status, comorbidity, medications, nutrition, psychological state, social support, and function. Most of these domains were evaluated by validated patient self-reported measures, but there also was a short health care provider portion. Tumor characteristics, pretreatment laboratory data, and information regarding the chemotherapy regimen also were noted. The patients were monitored throughout their entire course of chemotherapy, and any chemotherapy-related toxicities were assessed and scored by two physicians using the Common Terminology Criteria for Adverse Events.
Grade 3-5 toxicity occurred in 53% (12% grade 4, 2% grade 5). The most common hematologic toxicities were leukopenia and anemia, and the most common non-hematologic toxicities were fatigue, infection, and dehydration.
Once the data were obtained, the authors used multivariate analysis to identify the risk factors associated with increased risk of grade 3-5 chemotherapy toxicity. The 11-item predictive model included the following risk factors: age > 72 years; GI or GU cancer type; standard chemotherapy dosing; polychemotherapy; Hgb < 11 (male) < 10 (female); creatinine clearance < 34; hearing problems; > 1 fall in last 6 months; requiring help taking medication; limited ability to walk 1 block; and decreased social activity.
Commentary
Although older and younger patients derive similar benefit from chemotherapy, older cancer patients have a greater risk of toxicity. This is due to multiple factors that are not captured by existing predictors, like measurements of performance. Indeed, Karnofsky Performance Status was not correlated with chemotherapy toxicity in this study.
Hurria's predictive model has the advantage of simplicity. Five of its 11 questions (age, cancer type, hemoglobin, number and dosing of chemotherapy drugs) already are part of an oncologist's initial assessment. Though not part of every oncologist's routine, creatinine clearance is easily derived from existing data. The 5 "new" questions (hearing, falls, help with medicine, walking ability, social activity) are short and easy to ask. In addition to the ease with which it can be incorporated into a busy oncology practice, the model is readily adaptable as a simple electronic tool.
That is exactly what Martine Extermann has already done at Moffitt Cancer Center's Senior Adult Oncology Program.1 Like Hurria, Extermann derived her model, which goes by the witty acronym CRASH (Chemotherapy Risk Assessment Scale for High-Age patients), from a multivariate analysis of variables obtained in a prospective study of the standard oncology pre-chemotherapy evaluation and a comprehensive geriatric assessment.2 Her final analysis was based on 518 patients with a median age of 76. Sixty-four percent of the patients experienced severe toxicity (32% grade 4 hematologic toxicity and 56% grade 3-4 non-hematologic toxicity).
As in Hurria's model, all one needs to do to determine the risk of chemotherapy toxicity is add up the points assigned to each risk variable. Interestingly, the seven risk factors that came out in the multivariate analysis that determined the CRASH model are different from those that Hurria found to predict the risk of chemotherapy toxicity. The most important predictors of hematologic toxicity are diastolic blood pressure, instrumental activities of daily living (IADL), and lactate dehydrogenase (LDH); those for non-hematologic risk are performance status (ECOG), Mini Mental Status Exam (Folstein MMS), and Mini-Nutritional Assessment (MNA). Unlike the model developed by Hurria, CRASH incorporates chemotherapy regimen-specific risk data,3 and this was found to be an important risk factor for both hematologic and non-hematologic toxicity.
Both the CRASH and Hurria models are influenced by the investigators' choice of screening tools: different data in gives different data out. For example, despite including regimen-specific risk data, the CRASH model did not include information regarding the dose of chemotherapy that actually was delivered, an important factor in older patients who often get reduced doses of chemotherapy. It is also interesting that cognitive function fell out of the multivariate analysis that produced the Hurria model. It is possible that cognitive issues were accounted for by measures (like social activity), but choice of cognitive assessment tools (MMS for CRASH and Blessed Memory Test for the Hurria model) may have been just as important.
Finally, it is worth noting that unlike the academic medical center consortium led by Hurria, patients for the CRASH model were also recruited from community cancer centers. Though internally validated, Hurria's model has yet to be validated externally and one wonders if it was influenced by the care that her patients received from teams with expertise in geriatric oncology. The CRASH model has the advantage of external validation, but, like the Hurria model, it needs to be studied in specific tumor types and stages.
While we wait for these models to be refined, busy oncologists have, at long last, two tools to help inform the decision-making process as they discuss chemotherapy with their older patients and their families.
References
1. CRASH Score Calculator. Senior Adult Oncology Program, Moffitt Cancer Center. Available at: http://www.moffitt.org/Site.aspx?spid=FB0320F12A6A43D9B179635607FB493F&type=crashScore. Accessed Dec. 15, 2011.
2. Extermann M, et al. Predicting the risk of chemotherapy toxicity in older patients: The Chemotherapy Risk Assessment Scale for High-Age Patients (CRASH) score. Cancer 2011 Nov 9. doi: 10.1002/cncr.26646.
3. Extermann M, et al. MAX2a convenient index to estimate the average per patient risk for chemotherapy toxicity; validation in ECOG trials. Eur J Cancer 2004;40:1193-1198.
This prospective multicenter study presents an 11-item model for predicting chemotherapy toxicity in older adults with cancer. Its stratification schema identified older adults at low (30%), intermediate (52%), or high (83%) risk for chemotherapy toxicity.You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content