Abstract
Introduction
Treatment toxicities are common in older adults with cancer and consequently, treatment modifications are sometimes considered. We evaluated the prevalence and factors associated with treatment modifications at the first cycle in older patients receiving palliative systemic treatment.
Methods
Patients (n=369) from the GAP 70+ Trial (NCT02054741; PI: Mohile) usual care arm were included. Enrolled patients were aged 70+ with advanced cancer and ≥1 Geriatric Assessment (GA) domain impairment. Treatment modification was defined as any change from National Comprehensive Cancer Network guidelines or published clinical trials. Baseline variables included: 1) sociodemographic factors; 2) clinical variables; 3) GA domains; and 4) physician beliefs about life expectancy. Bivariate analyses and multivariable cluster-weighted generalized estimating equation model were conducted to assess the association of baseline variables with cycle 1 treatment modifications.
Results
Mean age was 77.2 years (range: 70–94); 62% had lung or gastrointestinal cancers, and 35% had treatment modifications at cycle 1. Increasing age by one year (odds ratio (OR) 1.1, 95% confidence interval [CI] 1.0–1.2), receipt of ≥second line of chemotherapy (OR 1.8, CI 1.1–3.0), functional impairment (OR 1.6, CI 1.1–2.3) and income ≤$50,000 (OR 1.7, CI 1.1–2.4) were independently associated with a higher likelihood of cycle 1 treatment modification.
Conclusion
Treatment modifications occurred in 35% of older adults with advanced cancer at cycle 1. Increasing age, receipt of ≥second line of chemotherapy, functional impairment, and lower income were independently associated with treatment modifications. These findings emphasize the need for evidence-based regimens in older adults with cancer and GA impairments.
Keywords: Older adults with cancer, Treatment modification, Chemotherapy dosing
Introduction
Cancer is disease of aging1 and the prevalence of cancer will continue to increase in the context of the current demographic changes.2 Despite the high burden of cancer in older individuals, older patients are underrepresented in cancer clinical trials3. The majority of cancer treatment regimens have been developed through clinical trials enrolling mainly younger and healthier patients and are not representative of “real-world” older adult populations.3
Older adults usually have one or more age-related impairments which can increase the risk of treatment toxicity.4,5 The presence of age-related impairments can be identified using a geriatric assessment (GA), a multidisciplinary systematic process using validated tools to assess the overall health status of older adults.6 The GA can evaluate “physiologic age” of an older adults better than chronologic age alone, and may be used to guide treatment decision-making in this population.7,8 Reliance on chronologic age alone can also result in under-treatment of fit older adults and multiple studies have demonstrated under-treatment of older adults as compared to younger patients in oncology.9–11 Nevertheless, GA remains under-utilized in oncology12, and chronologic age remains an influential decision-making factor for most oncologists.13
In the advanced cancer setting, the goal of treatment is palliative. Oncologists strive to balance the risks and benefits of cancer therapies in order to achieve disease control without compromising physical function or quality of life.4 Given the lack of evidence-based data to guide the optimal treatment regimens and decide whether treatment modifications can lead to worse efficacy or better tolerability outcomes in older patients with age-related issues, oncologists may struggle to maintain this balance.14 In the absence of data, clinicians may modify or adjust cancer treatments in this population due to concerns over toxicity or safety.15 Several groups have advocated for improving the evidence base for chemotherapy dosing in older adults with cancer through clinical trials and other focused research.16–18 However, to date the prevalence of treatment modifications in older adults with advanced cancer remains understudied, and baseline factors associated with treatment modifications at cycle 1 in this population are unknown. In addition, research is needed to examine if chronologic age can modify the relationship between other baseline clinical and geriatric factors and treatment modification at cycle 1.
The aims of this study were 1) to examine the prevalence of treatment modifications at cycle 1 in a cohort of older adults with advanced cancer receiving palliative systemic treatment, and 2) to identify baseline factors, including socio-demographic variables, clinical variables, GA domains, and physician beliefs that are associated with treatment modifications. As an exploratory analysis, we also investigated the possible moderation effect of chronological age on the association of baseline factors with treatment modifications.
Methods
Study Design
This is a secondary analysis of baseline data from a nationwide, multicenter, cluster-randomized study that assessed whether providing information regarding GA plus GA-driven recommendations to community oncologists reduced clinician-rated grade 3–5 chemotherapy toxicity in older patients with advanced cancer starting a new cancer treatment regimen (Geriatric Assessment for Patients [GAP70+] study; University of Rochester Cancer Center [URCC] 13059, PI: Mohile; ClinicalTrials.gov identifier: NCT02054741)19. In the GAP70+ Study, community oncology practices were randomized to GA intervention versus usual care. The current secondary analysis included baseline data from patients in the usual care arm only (GA completed but results not provided to primary oncologists unless the patient screened positive for significant cognitive impairment or depression; no GA-driven interventions were provided) in order to evaluate treatment-modification patterns in daily practice in community settings and to avoid the possible influence of the intervention. The study was conducted by the URCC NCI Community Oncology Research Program and approved by the Institutional Review Boards at participating sites.
Participants
Eligible patients for the primary study were 1) aged ≥70 years, 2) diagnosed with incurable stage III/IV solid tumor or lymphoma, 3) impaired in at least one GA domain other than polypharmacy, and 4) planning to start a new cancer treatment regimen with a high risk of grade 3–5 toxicity (Common Terminology Criteria for Adverse Events v4). The planned regimen must include a chemotherapy drug or other agents that have a similar prevalence of toxicity (example: tyrosine kinase inhibitors such as sorafenib and erlotinib). Eligible regimens were determined based on enrolling physicians’ discretion and were reviewed at the primary coordinating site.
Independent Variables
We collected the following baseline information after patients provided informed consent and before they started a new line of palliative cancer treatment. Socio-demographic variables included age (continuous variable), gender, race (White, Black, and others), education (less than high school, high school graduate, and some college or more), and income (≤$50,000 and >$50,000/ declined to answer). Clinical variables included cancer type (gastrointestinal, lung, others), cancer stage (stage III, stage IV), line of palliative treatment (first versus second-line or greater), and physician-reported Karnofsky Performance Score (KPS) (40–60, 70–80, and 90–100). GA domains were captured using validated tools with established cut-offs for impairment including comorbidity, functional status, physical performance, cognition, social support, polypharmacy, psychological health, and nutritional status. These domains have been detailed previously (supplemental Table 1).20 Variables to assess physician beliefs about the prognosis of the disease included answers to the following questions: “How do you think cancer treatment will affect the patient’s quality of life?” (Improve versus no change or decrease), and “What would you estimate the patient’s overall life expectancy to be?” (0–12 months versus >12 months). These questions were adopted from prior studies.21,22
Dependent variable
Treatment modification at cycle 1 (at the beginning of the treatment course) was defined as any change in dose or agents of the planned regimen from the standard treatment guidelines (National Comprehensive Cancer Network [NCCN] guidelines) or published phase II/III clinical trials. Standard treatment “full dose” means regimens that meet the standard guidelines for dose and scheduling according to NCCN or published phase II/III trials. The planned regimens (individual drugs, doses, and schedule) were captured at the beginning of the study from the primary oncology team and reviewed by a blinded research team including at least two oncology clinicians and an information analyst. Modifications were classified as 1) dose reduction (i.e. chemotherapy dose lower than the established regimen); 2) modified schedule to administer chemotherapy less frequently than the established regimen (e.g. patient receives only Day 1 and Day 15 of a regimen with established schedule of Day 1, 8, 15); or 3) modified regimen (e.g. one or more chemotherapy agents are intentionally left out of a polychemotherapy regimen). The composite binary variable of treatment modification (standard versus non-standard) was constructed as the presence of any of these types of treatment modifications and served as the dependent variable in bivariate and multivariate analyses.
Statistical Analysis
Descriptive statistics were performed to describe independent and dependent variables. Subsequently, bivariate analyses were conducted to evaluate the association between each of the baseline variables and treatment modification (standard or non-standard) at cycle 1.We constructed multivariable cluster-weighted generalized estimating equations (GEE) model with a binary distribution, logit link, and robust standard errors to examine the independent associations of baseline covariates with treatment modification at cycle 1 accounting for correlations among patients from the same practice.. The final model included covariates with p<0.1 in bivariate analyses in addition to age, gender, race, and cancer type as these variables have been shown to correlate with treatment disparities in older adults in prior studies15,23,24. Subsequently, we examined the moderation effect of chronological age on the association of baseline variables (functional status, physical performance, physician’s estimated life expectancy for the patient, and comorbidities) with treatment modification at cycle 1. These variables were selected a-priori based on previous literature and clinical relevance.25–27 The moderation effect was assessed by creating an interaction term between each of the baseline variables and age. After that, we created stratified analysis to show the effect estimates for the relationship between these variables and treatment modification across different categories of age (70–74, 75–79, and 80+)
Two-sided p-values of <0.05 were considered statistically significant. All analyses were conducted with SAS 9.4 (SAS Institute Inc.).
Results
Study Population Characteristics and Treatment
The mean age of participants (n=369) was 77.2 years (standard deviation 5.2 years; range 70–94). Overall, 54.7% were males, 94% were non-Hispanic Whites, and the majority (87.8%) had stage IV cancer. The most common cancer types were lung (n=116, 31.4%) and gastrointestinal (n=114, 30.9%). At the time of enrollment, 74.5% (n=275) were scheduled to receive first-line chemotherapy, and 25.5% (n=94) were scheduled to receive second-line or later treatment (table 1). The most commonly used treatment regimens are detailed in table 2.
Table 1:
Baseline characteristics and bivariate analyses of baseline variables and treatment modification at cycle 1
| Variable | category | All patients | Standard dose | Treatment modification | p-value |
|---|---|---|---|---|---|
| N | 369 (100%) | 240 (65%) | 129 (35%) | ||
| Socio-demographic variables | |||||
| Agea, mean (SD) | 77.2 (5.2) | 76.6 (4.8) | 78.6 (5.7) | <0.01** | |
| Gendera | Male | 202 (54.7%) | 134 (55.8%) | 68 (53.1%) | 0.62 |
| Female | 166 (45.3%) | 106 (44.2) | 60 (46.9%) | ||
| Racea | White | 350 (95.1%) | 228 (95.0%) | 122 (95.3%) | 0.57 |
| Black | 12 (3.3%) | 9 (3.8%) | 3 (2.3%) | ||
| Others | 6 (1.6%) | 3 (1.3%) | 3 (2.3%) | ||
| Educationa | Less than High school | 53 (14.4%) | 34 (14.2%) | 19 (14.8%) | 0.90 |
| High school | 125 (33.9%) | 80 (33.3%) | 45 (35.2%) | ||
| College or above | 190 (51.5%) | 126 (52.5%) | 64 (50.0%) | ||
| Incomea | ≤$50,000 | 182 (49.3%) | 108 (45.0%) | 74 (57.8%) | 0.02** |
| >50,000 or declined to answer | 186 (50.4%) | 132 (55.0%) | 54 (42.2%) | ||
| Baseline clinical variables | |||||
| Cancer type | GI | 114 (30.9%) | 69 (28.8%) | 45 (34.9%) | 0.36 |
| Lung | 116 (31.4%) | 75 (31.3%) | 41 (31.8%) | ||
| Others | 86 (23.3%) | 96 (40.0%) | 43 (33.3%) | ||
| Cancer stage | Stage 3 | 35 (9.5%) | 24 (10%) | 11 (8.5%) | 0.52 |
| Stage 4 | 324 (87.8%) | 208 (86.7%) | 116 (89.9%) | ||
| others | 10 (2.7%) | 8 (3.3%) | 2 (1.6%) | ||
| Karnofsky Performance Scale | 40–60 | 55 (14.9%) | 32 (13.3%) | 23 (17.8%) | 0.20 |
| 70–80 | 193 (52.3%) | 122 (50.8%) | 71 (55.0%) | ||
| 90–100 | 121 (32.8%) | 86 (35.8%) | 35 (27.1%) | ||
| Line of treatment | First line | 275 (74.5%) | 188 (78.3%) | 87 (67.4%) | 0.02** |
| Second or greater | 94 (24.5%) | 52 (21.6%) | 42 (32.6%) | ||
| physician beliefs’ about the prognosis of the disease | |||||
| Life expectancy according to physicianb | <12 months | 134 (36.3%) | 78 (33.1%) | 56 (43.4%) | 0.04** |
| >= 12 months | 231 (62.6%) | 158 (67.0%) | 73 (56.6%) | ||
| Impact of chemo on quality of life according to physicianc | Decreased or no change | 170 (46.1%) | 109 (46.6%) | 61 (47.3%) | 0.90 |
| Improve | 193 (52.3%) | 125 (53.4%) | 68 (52.7%) | ||
| Geriatric domains | |||||
| Functional statusa | Non-Impaired | 156 (42.3%) | 113 (47.1%) | 43 (33.6%) | 0.01** |
| impaired | 212 (57.6%) | 127 (52.9%) | 85 (66.4%) | ||
| Physical performance | Non-impaired | 14 (3.8%) | 10 (4.2%) | 4 (3.1%) | 0.61 |
| Impaired | 355 (96.2%) | 230 (95.8%) | 125 (96.9%) | ||
| Comorbiditya | Non-Impaired | 120 (32.8%) | 82 (34.2%) | 38 (29.7%) | 0.38 |
| impaired | 248 (67.2%) | 158 (65.8%) | 90 (70.3%) | ||
| Polypharmacy | Non-Impaired | 72 (19.5%) | 50 (20.8%) | 22 (17.1%) | 0.38 |
| impaired | 297 (80.5%) | 190 (79.2%) | 107 (83.0%) | ||
| Nutritional statusa | Non-Impaired | 141 (38.2%) | 94 (39.2%) | 47 (36.4%) | 0.61 |
| impaired | 228 (61.8%) | 146 (60.8%) | 82 (63.6%) | ||
| Social supporta | Non-Impaired | 285 (77.2%) | 185 (77.1%) | 100 (78.1%) | 0.82 |
| impaired | 83 (22.5%) | 55 (22.9%) | 28 (21.9%) | ||
| Psychological health | Non-Impaired | 271 (73.5%) | 179 (74.6%) | 92 (71.3%) | 0.50 |
| Impaired | 98 (26.5%) | 61 (25.4%) | 37 (28.7%) | ||
| Cognition | Non-impaired | 248 (67.2%) | 164 (68.3%) | 84 (65.1%) | 0.53 |
| Impaired | 121 (32.8%) | 76 (31.7%) | 45 (34.9%) | ||
1 patient had missing data
4 had missing data
6 had missing data
P value <0.0
Table 2:
Common treatment regimens received at cycle 1
| Treatment regimen | |
|---|---|
| Lung cancer regimens | N= 116 |
| Pemetrexed- carboplatin +/− pembrolizumab | 53 (46%) |
| Paclitaxel- carboplatin +/− monoclonal antibody | 16 (14%) |
| Carboplatin- etoposide | 15 (13%) |
| Carboplatin- nab paclitaxel | 10 (9%) |
| Gastro-intestinal cancers regimens | N= 114 |
| FOLFOX +/− bevacizumab | 40 (35.1%) |
| Gemcitabine- nab paclitaxel | 20 (17.5%) |
| FOLFIRI +/− bevacizumab | 6 (5%) |
| FOLFIRINOX +/− bevacizumab | 6 (5%) |
| Genito-urinary cancers regimens | N= 53 |
| Docetaxel +/− prednisone | 13 (24.5%) |
| Abiraterone +/− prednisone | 13 (24.5%) |
| Enzalutamide +/− prednisone | 10 (18.9%) |
| Gemcitabine carboplatin | 8 (15.1%) |
| Breast Cancer regimens | N= 37 |
| Palbociclib+ AI | 12 (32%) |
| Paclitaxel single agent | 8 (22%) |
| Capecitabine | 4 (11%) |
| Lymphoma Regimens | N= 23 |
| BR | 11 (48%) |
| R-CHOP | 4 (17.4%) |
| Gynecological cancers regimens | N= 15 |
| Paclitaxel carboplatin | 9 (60%) |
This table only included commonly received regimens at cycle 1
Abbreviations: AI, aromatase inhibitors; BR, bendamustine/rituximab; FOLOFOX, 5-fluorouracil/ leucovorin/ oxaliplatin; FOLFIRI, 5-fluorouracil/ leucovorin/ irinotecan; FOLFIRINOX/ 5-fluorouracil/ leucovorin/ oxaliplatin/ irinotecan; R-CHOP, rituximab/ cyclophosphamide/ doxorubicin/ prednisone/ vincristine
Of 369 patients, 240 (65.0%) received standard regimens at cycle 1 according to NCCN guidelines or published clinical trials, and 129 (35.0%) received a modified treatment regimen. Of those patients with treatment modification, 79% (n=292) received a regimen with a dose reduction, 12% (n=44) received a different combination of agents (modified regimen), and 9% (n=33) received a modified schedule.
Bivariate Analyses
The mean age of patients who received a modified regimen was 78.6 years, compared to 76.6 years in patients who received standard regimens (p<0.01). Patients who received treatment modifications were more likely to have functional impairments (66.4% versus 52.9%, p=0.01) and had physician-estimated life expectancies of <12 months (43.4% versus 33.1%, p=0.04). Treatment modifications at cycle 1 were also more likely in patients reporting lower incomes (57.8% of patients with yearly incomes ≤$50,000, compared to 42.2% of patients with incomes >$50,000 and those who declined to answer [p = 0.02]). In addition, patients who received a modified treatment regimen were more likely to be receiving second line or later treatment (32.6% versus 21.7%, p=0.02).
There were no significant differences between groups (i.e. patients received standard vs treatment modification) across race, gender, education, cancer type, KPS, other impaired GA domains, or physician estimated impact of treatment on quality of life (Table 1).
Multivariable analysis to examine the association of baseline variables with treatment modifications at cycle 1
In multivariable GEE regression, we found that each additional year of age was associated with 10% increased odds of treatment modification at cycle 1. (OR 1.1, 95% confidence interval [CI] 1.0–1.2). Receipt of second or greater line of chemotherapy (OR 1.8, CI 1.1–3.0), functional impairment (OR 1.6, CI 1.1–2.3), and lower income (OR 1.7, CI 1.1–2.4) were also associated with increased odds of treatment modification at cycle 1 (Table 3).
Table 3:
Cluster-weighted multivariable generalized estimating equation model to examine the association between baseline variables and treatment modification at cycle 1
| Variable | Odds ratio | 95% confidence interval |
|---|---|---|
| Age (continuous) | 1.08*** | 1.02–1.15 |
| Female (ref. male) | 0.97 | 0.72–1.32 |
| Cancer type (ref. Gastrointestinal) | ||
| Lung cancer | 0.87 | 0.47–1.62 |
| Other cancer types | 0.67 | 0.30–1.51 |
| Race (ref. non-Hispanic white) | ||
| African American | 0.61 | 0.19–1.96 |
| Others | 1.39 | 0.44–4.39 |
| Impaired functional status (ref. non-impaired) | 1.57** | 1.08–2.27 |
| Second or more line of chemo (ref. first line) | 1.79** | 1.08–2.95 |
| Income =<$50,000 (ref. >$50,000 or declined to answer) | 1.67*** | 1.07–2.43 |
| Life expectancy> 1 year (ref.=< 1 year) | 0.81 | 0.54–1.21 |
p<0.05
p<0.01
Moderation effect of increasing age
A significant moderation effect was found between patients’ age and physician-estimated life expectancy for cycle 1 treatment modification (p=0.02). In adjusted models and after stratification of the study population by age group (70–74 (n=136), 75–79 (n=114), and ≥80 (n=118)), we found that among patients aged 70–74 years, those who had an estimated life expectancy of more than one year had 2.5 times the odds of receiving standard of care doses compared to those who had a life expectancy of less than one year (CI 1.1–5.4). Among patients aged 75–79 years and patients ≥80 years, those who had life expectancies greater than one year had 1.3 (CI 0.6–3.1) and 1.2 (CI 0.5–2.6) times the odds of receiving standard of care doses compared to those had life expectancies of less than one year. Increasing age did not moderate the associations of functional status, physical performance, or cognition with treatment modification at cycle 1 (p>0.05) (Table 4).
Table 4:
Multivariable analyses between baseline variables and standard dose at cycle 1 stratified by age group
| Variable | Standard dose at cycle 1 | ||
|---|---|---|---|
| Odds ratio | 95% CI | P-value for interaction | |
| Functional status (ref: Impaired) | 0.24 | ||
| age 70–74 | 1.54 | 0.46–5.14 | |
| age 75–79 | 5.13 | 1.11–23.61 | |
| Age 80+ | 2.32 | 0.43–12.51 | |
| Comorbidities (ref: Impaired) | 0.57 | ||
| age 70–74 | 0.68 | 0.29–1.58 | |
| age 75–79 | 1.06 | 0.43–2.58 | |
| Age 80+ | 1.43 | 0.62–3.31 | |
| physical performance (ref: Impaired) | 0.66 | ||
| age 70–74 | 2.98 | 0.63–14.13 | |
| age 75–79 | 0.68 | 0.10–4.72 | |
| Age 80+ | 0.69 | 1.18–2.70 | |
| Life expectancy according to physician (ref: < 1 year) | 0.02 | ||
| age 70–74 | 2.45 | 1.12–5.41 | |
| age 75–79 | 1.32 | 0.56–3.14 | |
| Age 80+ | 1.17 | 0.54–2.56 | |
P values measure the significance of interaction term of age and each of the examined variables based on GEE model
Discussion
In this secondary analysis of a large, nationwide clinical trial of older adults with advanced cancer from community oncology practices, we identified a high prevalence of treatment modifications at cycle 1. More than one-third of enrolled patients received modified treatment regimens, as compared to NCCN guidelines or published phase II/ III trials. After adjusting for covariates, increasing age, prior lines of chemotherapy, functional impairment, and income ≤$50,000 per year were independently associated with increased odds of treatment modification at cycle 1. In addition, we demonstrated that chronological age moderated the association between physician-estimated life expectancy and treatment modification at cycle 1.
We observed chronological age was independently associated with cycle 1 treatment modifications, consistent with prior studies.15,28 Gajra et al, demonstrated almost 25% of older adults with cancer receiving palliative chemotherapy had dose reductions at first cycle (n=321) and chronologic age was the only independent factor associated with dose reduction.15 Similar to the current analysis, this study included baseline GA assessment for their participants. However, their population included patients treated at academic medical centers, which may explain the difference in prevalence of treatment modification compared to our analysis. In another retrospective study of 529 patients (191 patients >70 years) with localized colorectal cancer receiving chemotherapy, 13% of patients received primary dose reduction and this was more common in patients aged >65.28 In this study, the majority of the population were younger adults and were receiving curative intent treatment, which may explain the difference in prevalence of treatment modification from our analysis. Chronological age alone is an insufficient measure of the fitness of older adults and thus should not independently guide treatment decisions in oncology.29 Reliance on chronologic age alone, as opposed to a more comprehensive view of overall health status (e.g. geriatric assessment) can contribute to both overtreatment of frail patients, as well as undertreatment of fit older adults.30–32 The current analysis reinforces that oncologists do use chronological age alone in determining dosing strategy.
Patients with a history of prior treatment also had a higher likelihood of dose modification. Presumably, toxicity to a first-line therapy can play a role in the decision to modify subsequent treatments. This is consistent with prior literature which discussed the impact of prior treatments on dose modifications in older adults. For example, in a study examining chemotherapy dosing patterns for metastatic pancreatic cancer, treatments were more likely to be discontinued in patients receiving third line therapy versus first or second line therapy;33 this was most commonly related to disease progression, although treatment-related toxicity and disease-related symptoms were also reported. In the current analysis, patients undergoing second- or third-line therapy could have had treatment modifications at cycle 1 based on their existing symptom burden and the knowledge of the clinical course with previous treatment.
In the current analysis, we observed an association between lower income and treatment modification. Several studies have shown a relationship between lower socio-economic status (SES) and treatment modification.34,35 For example, patients receiving adjuvant breast cancer therapy, women of lower SES background were more likely to receive dose reductions.34 However, to our knowledge, the current study is the first to demonstrate the impact of income on dosing strategy in advanced cancer setting and specifically for the older adult population.
We demonstrated a significant association between functional impairment and treatment modification at cycle 1. The GA functional domain included assessment of activities of daily living (ADL; self-care tasks such as bathing and dressing) and instrumental ADLs (IADL; measures of independence such as managing medications and finances). Interestingly, KPS, the measure oncologists traditionally use to assess physical fitness,36 was not associated with treatment modifications. Although GA results were not introduced to the treating physicians before starting treatment, this finding suggests that ADL and IADL questionnaires may better reflect the true functional status of the patients. This is consistent with recent studies demonstrating that GA can provide more information on the functional status of older patients with cancer compared to KPS, even among those with normal KPS scores.37,38 In addition, models including GA variables are superior to KPS in predicting chemotherapy toxicity risk and overall survival in older adults with cancer.39,40
We also demonstrated that chronological age acted as a moderating factor in the relationship between estimated life expectancy and cycle 1 treatment modification. This observation suggests that physicians incorporate life expectancy into treatment decisions differently for patients < 75 years than for older patients (75–79 and ≥80). This result might be explained by age-based cognitive bias whereby physicians apply mental shortcuts regarding age groups to simplify their decision making. This concept has previously been demonstrated in other treatment recommendations.41 Use of a GA and validated life expectancy prediction tools may also help to mitigate the effects of age-based cognitive bias by providing a more comprehensive view of a patient’s overall health status and life expectancy.
Our study has clinical and public health implications. The high prevalence of treatment modifications for older adults with advanced cancer suggests the need for better evidence-based approaches for treatment recommendations in this population. Recent NCCN and American Society of Clinical Oncology guidelines support the utilization of GA to evaluate older adults prior to starting a new cancer treatment, and this approach may help to overcome the influence of chronologic age on treatment recommendations that we observed.4 Extensive data indicate that a GA influences treatment decision-making and supportive care recommendations.13,20,42 In palliative settings, very few studies have investigated the efficacy and safety of modified treatment regimens in frail older patients.43,44 Clinical trials that incorporate GA as a component of treatment decision making may help mitigate the strong effect that chronologic age has on treatment recommendations by oncologists.4 Additionally, designing clinical trials that actively test less toxic regimens, either by altering dosing or scheduling of established regimens or by utilizing novel targeted therapies, can offer oncologists evidence-based options for older patients who may be vulnerable.
A major strength of this study is its inclusion of a population that is historically underrepresented in clinical trials – older adults with advanced cancer receiving care in community oncology clinics. Additionally, we evaluated a comprehensive set of variables, including an extensive evaluation of baseline health status through the GA, to identify associations with treatment modifications at cycle 1. The study also has several limitations. First, as patients in this analysis were enrolled as a part of a GA intervention clinical trial, this may have introduced bias based upon the population selecting to participate, which possibly limited the generalizability of the study. Second, patients were mostly non-Hispanic White and well educated, and therefore our results may not be generalizable to patients of other races and with lower education levels. Third, as this was an exploratory analysis, we did not control for multiple comparisons when we examined the moderation effect of chronologic age45. Fourth, we were not able to examine cumulative relative dose intensity as a dependent variable in this study, however as this analysis was focused on treatment decision-making at cycle 1, future analyses will consider this aspect. Fifth, we have considered “treatment modifications” as any treatment that deviates from what is considered as a “standard” based on the NCCN guidelines or other published phase II/III trials. However, previous research has demonstrated that the majority of cancer treatment regimens have been developed through clinical trials enrolling mainly younger and healthier patients which might not be representative of older individuals with cancer, thus the “standard” regimen is not truly defined in many cases for this population.46
In conclusion, in this large, nationwide study of older adults with advanced cancer receiving palliative treatment, we observed that 35% received a modified treatment regimen at cycle 1. Increasing age, lower income, functional impairment, and receipt of prior lines of treatment were associated with increased odds of treatment modification. Additionally, we observed chronologic age as a moderator in the relationship between life expectancy and treatment recommendations, suggesting that life expectancy is weighted more or less heavily in treatment decisions depending on chronologic age. The rate of treatment modification in this population emphasizes the importance of developing evidence-based treatment regimens for older adults with advanced cancer and GA impairments. Future studies should evaluate the longitudinal effect of cycle 1 treatment modifications in the advanced cancer setting on both cancer and patient-reported outcomes.
Supplementary Material
Funding Statement:
Study was funded through R01 CA177592 (Mohile), K24 AG056589 (Mohile), U01CA233167 (Mohile), (NIA) R21/R33AG059206 (Mohile), NCI UG1CA189961 (Mohile, Flannery - PI is Morrow/ Mustian), NCI K99 CA237744 (Loh), Wilmot Research Fellowship Award (Loh), and NIA K76 AG064394 (Magnuson).
Authors’ Disclosures of Potential Conflicts of Interest:
Dr. Mohile received research funding from Carevive for other projects. Dr. Loh serves as a consultant to Pfizer and Seattle Genetics. Dr. Dunne received honoraria for consulting for Exelixis Inc.
References
- 1.White MC, Holman DM, Boehm JE, et al. : Age and cancer risk: a potentially modifiable relationship. American journal of preventive medicine 46:S7–S15, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bluethmann SM, Mariotto AB, Rowland JH: Anticipating the “Silver Tsunami”: Prevalence Trajectories and Comorbidity Burden among Older Cancer Survivors in the United States. Cancer Epidemiol Biomarkers Prev 25:1029–36, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Soto-Perez-De-Celis E, Lichtman SM: Considerations for clinical trial design in older adults with cancer. Expert opinion on investigational drugs 26:1099–1102, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mohile SG, Dale W, Somerfield MR, et al. : Practical Assessment and Management of Vulnerabilities in Older Patients Receiving Chemotherapy: ASCO Guideline for Geriatric Oncology. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 36:2326–2347, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Jayani RV, Magnuson AM, Sun CL, et al. : Association between a cognitive screening test and severe chemotherapy toxicity in older adults with cancer. J Geriatr Oncol 11:284–289, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Extermann M, Aapro M, Bernabei R, et al. : Use of comprehensive geriatric assessment in older cancer patients: recommendations from the task force on CGA of the International Society of Geriatric Oncology (SIOG). Crit Rev Oncol Hematol 55:241–52, 2005 [DOI] [PubMed] [Google Scholar]
- 7.Loh KP, Duberstein P, Zittel J, et al. : Relationships of self-perceived age with geriatric assessment domains in older adults with cancer. Journal of Geriatric Oncology 11:1006–1010, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Quinten C, Kenis C, Decoster L, et al. : The prognostic value of patient-reported Health-Related Quality of Life and Geriatric Assessment in predicting early death in 6769 older (≥70 years) patients with different cancer tumors. J Geriatr Oncol 11:926–936, 2020 [DOI] [PubMed] [Google Scholar]
- 9.Quipourt V, Jooste V, Cottet V, et al. : Comorbidities alone do not explain the undertreatment of colorectal cancer in older adults: a French population‐based study. Journal of the American Geriatrics Society 59:694–698, 2011 [DOI] [PubMed] [Google Scholar]
- 10.Fourcadier E, Trétarre B, Gras-Aygon C, et al. : Under-treatment of elderly patients with ovarian cancer: a population based study. BMC cancer 15:1–10, 2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Meresse M, Bouhnik AD, Bendiane MK, et al. : Chemotherapy in old women with breast cancer: is age still a predictor for under treatment? The breast journal 23:256–266, 2017 [DOI] [PubMed] [Google Scholar]
- 12.Magnuson A, Allore H, Cohen HJ, et al. : Geriatric assessment with management in cancer care: Current evidence and potential mechanisms for future research. Journal of geriatric oncology 7:242–248, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Mohile SG, Magnuson A, Pandya C, et al. : Community Oncologists’ Decision-Making for Treatment of Older Patients With Cancer. Journal of the National Comprehensive Cancer Network : JNCCN 16:301–309, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.DuMontier C, Loh KP, Bain PA, et al. : Defining Undertreatment and Overtreatment in Older Adults With Cancer: A Scoping Literature Review. J Clin Oncol:Jco 1902809, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gajra A, Klepin HD, Feng T, et al. : Predictors of chemotherapy dose reduction at first cycle in patients age 65 years and older with solid tumors. Journal of geriatric oncology 6:133–140, 2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hurria A, Dale W, Mooney M, et al. : Designing therapeutic clinical trials for older and frail adults with cancer: U13 conference recommendations. Journal of Clinical Oncology 32:2587, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hurria A, Levit LA, Dale W, et al. : Improving the evidence base for treating older adults with cancer: American Society of Clinical Oncology statement. J Clin Oncol 33:3826–3833, 2015 [DOI] [PubMed] [Google Scholar]
- 18.Hurria A, Lichtman S: Clinical pharmacology of cancer therapies in older adults. British journal of cancer 98:517–522, 2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Mohile SG, Mohamed MR, Culakova E, et al. : A geriatric assessment (GA) intervention to reduce treatment toxicity in older patients with advanced cancer: A University of Rochester Cancer Center NCI community oncology research program cluster randomized clinical trial (CRCT). Journal of Clinical Oncology 38:12009–12009, 2020 [Google Scholar]
- 20.Mohile SG, Epstein RM, Hurria A, et al. : Communication With Older Patients With Cancer Using Geriatric Assessment: A Cluster-Randomized Clinical Trial From the National Cancer Institute Community Oncology Research Program. JAMA Oncology 6:196–204, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Fried TR, Bradley EH, O’Leary J: Changes in prognostic awareness among seriously ill older persons and their caregivers. J Palliat Med 9:61–9, 2006 [DOI] [PubMed] [Google Scholar]
- 22.Epstein RM, Duberstein PR, Fenton JJ, et al. : Effect of a Patient-Centered Communication Intervention on Oncologist-Patient Communication, Quality of Life, and Health Care Utilization in Advanced Cancer: The VOICE Randomized Clinical Trial. JAMA oncology 3:92–100, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Green AK, Aviki EM, Matsoukas K, et al. : Racial disparities in chemotherapy administration for early-stage breast cancer: a systematic review and meta-analysis. Breast cancer research and treatment 172:247–263, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Glanz K, Croyle RT, Chollette VY, et al. : Cancer-related health disparities in women. American journal of public health 93:292–298, 2003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Tuna HD, Edeer AO, Malkoc M, et al. : Effect of age and physical activity level on functional fitness in older adults. European Review of Aging and Physical Activity 6:99, 2009 [Google Scholar]
- 26.Takagi D, Nishida Y, Fujita D: Age-associated changes in the level of physical activity in elderly adults. Journal of physical therapy science 27:3685–3687, 2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Incalzi RA, Capparella O, Gemma A, et al. : The interaction between age and comorbidity contributes to predicting the mortality of geriatric patients in the acute-care hospital. Journal of internal medicine 242:291–298, 1997 [DOI] [PubMed] [Google Scholar]
- 28.Lund CM, Nielsen D, Dehlendorff C, et al. : Efficacy and toxicity of adjuvant chemotherapy in elderly patients with colorectal cancer: the ACCORE study. ESMO Open 1:e000087, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Shahrokni A, Kim SJ, Bosl GJ, et al. : How We Care for an Older Patient With Cancer. Journal of Oncology Practice 13:95–102, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Soto-Perez-de-Celis E, Li D, Yuan Y, et al. : Functional versus chronological age: geriatric assessments to guide decision making in older patients with cancer. The Lancet Oncology 19:e305–e316, 2018 [DOI] [PubMed] [Google Scholar]
- 31.Hurria A, Wong FL, Villaluna D, et al. : Role of age and health in treatment recommendations for older adults with breast cancer: the perspective of oncologists and primary care providers. Journal of clinical oncology 26:5386, 2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Feliu J, Espinosa E, Basterretxea L, et al. : Undertreatment and overtreatment in older patients treated with chemotherapy. Journal of geriatric oncology 12:381–387, 2021 [DOI] [PubMed] [Google Scholar]
- 33.Barzi A, Miksad R, Surinach A, et al. : Real-World Dosing Patterns and Outcomes of Patients With Metastatic Pancreatic Cancer Treated With a Liposomal Irinotecan Regimen in the United States. Pancreas 49:193, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Griggs JJ, Culakova E, Sorbero ME, et al. : Effect of patient socioeconomic status and body mass index on the quality of breast cancer adjuvant chemotherapy. J Clin Oncol 25:277–84, 2007 [DOI] [PubMed] [Google Scholar]
- 35.Neuner JM, Kong A, Blaes A, et al. : The association of socioeconomic status with receipt of neoadjuvant chemotherapy. Breast cancer research and treatment 173:179–188, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Kelly CM, Shahrokni A: Moving beyond Karnofsky and ECOG Performance Status Assessments with New Technologies. Journal of oncology 2016:6186543–6186543, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Jolly TA, Deal AM, Nyrop KA, et al. : Geriatric assessment-identified deficits in older cancer patients with normal performance status. 20:379, 2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Repetto L, Fratino L, Audisio RA, et al. : Comprehensive Geriatric Assessment Adds Information to Eastern Cooperative Oncology Group Performance Status in Elderly Cancer Patients: An Italian Group for Geriatric Oncology Study. 20:494–502, 2002 [DOI] [PubMed] [Google Scholar]
- 39.Hurria A, Mohile S, Gajra A, et al. : Validation of a Prediction Tool for Chemotherapy Toxicity in Older Adults With Cancer. J Clin Oncol 34:2366–71, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Ghosn M, Ibrahim T, El Rassy E, et al. : Abridged geriatric assessment is a better predictor of overall survival than the Karnofsky Performance Scale and Physical Performance Test in elderly patients with cancer. J Geriatr Oncol 8:128–132, 2017 [DOI] [PubMed] [Google Scholar]
- 41.Olenski AR, Zimerman A, Coussens S, et al. : Behavioral Heuristics in Coronary-Artery Bypass Graft Surgery. New England Journal of Medicine 382:778–779, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Hamaker ME, Te Molder M, Thielen N, et al. : The effect of a geriatric evaluation on treatment decisions and outcome for older cancer patients - A systematic review. J Geriatr Oncol 9:430–440, 2018 [DOI] [PubMed] [Google Scholar]
- 43.Seymour MT, Thompson LC, Wasan HS, et al. : Chemotherapy options in elderly and frail patients with metastatic colorectal cancer (MRC FOCUS2): an open-label, randomised factorial trial. The Lancet 377:1749–1759, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Ibusuki M, Inoue T, Kitano R, et al. : Gemcitabine plus nab-paclitaxel with initial dose reduction for older patients with advanced pancreatic cancer. Journal of Geriatric Oncology, 2020 [DOI] [PubMed] [Google Scholar]
- 45.Bender R, Lange S: Adjusting for multiple testing--when and how? J Clin Epidemiol 54:343–9, 2001 [DOI] [PubMed] [Google Scholar]
- 46.Battisti NML, Sehovic M, Extermann M: Assessment of the External Validity of the National Comprehensive Cancer Network and European Society for Medical Oncology Guidelines for Non–Small-Cell Lung Cancer in a Population of Patients Aged 80 Years and Older. Clinical Lung Cancer 18:460–471, 2017 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.