To the authors' knowledge, few data currently are available regarding the tolerance to chemotherapy in older cancer patients. This prospective pilot study evaluated the changes in functional, mental, nutritional, and comorbid status, as well as the quality of life (QOL), in geriatric oncology patients receiving chemotherapy.
Sixty patients age ≥ 70 years who were undergoing cancer chemotherapy were recruited in a university-based comprehensive cancer center. Changes in physical function were measured by the Eastern Cooperative Oncology Group performance status (ECOG PS) and Instrumental Activities of Daily Living (IADLs), mental health changes were measured by the Mini-Mental State Examination and the Geriatric Depression Scale (GDS), comorbidity was measured by Charlson's index and the Cumulative Illness Rating Scale-Geriatric, nutrition was measured by the Mini-Nutritional Assessment, and QOL was measured by the Functional Assessment of Cancer Therapy-General (FACT-G). Changes were assessed at baseline and at the end of treatment (EOT). Grade 4 hematologic and Grade 3–4 nonhematologic toxicities were recorded.
Thirty-seven patients (63%) completed both assessments. Older cancer patients demonstrated a significant decline in measurements of physical function after receiving chemotherapy, as indicated by changes in scores on the IADL (P = 0.04) and on the physical (P = 0.01) and functional (P = 0.03) subscales of the FACT-G. They also displayed worse scores on the GDS administered postchemotherapy (P < 0.01). Patients who experienced severe chemotoxicity had more significant declines in ECOG PS (P = 0.03), IADL (P = 0.03), and GDS (P = 0.04), and more gain in the social well-being subscale (P = 0.02) of the FACT-G, than those who did not experience severe chemotoxicity. However, changes in most scores were small in magnitude clinically. No significant change was found between baseline and EOT in nutrition, comorbidity, and other aspects of the FACT-G.
Nearly 60% of all incident cancers and 70% of all cancer mortalities occur in the 65-year-and-older age group.1 Yet, studies of patterns of care in geriatric oncology patients have shown mixed, sometimes discouraging, results. The evidence suggests that the choice of cancer therapy varied with the age of patient.2–7 Older people were less likely to receive treatments that were considered definitive or potentially curative than younger patients, regardless of cancer site or stage at diagnosis.8–10 Whereas chronologic age is not often cited by physicians as a reason for not recommending aggressive therapy, other age-related factors such as the presence of comorbid conditions and functional disability were independent predictors for receiving less standard treatment in older cancer patients.3, 4, 11, 12 Clinicians may be legitimately concerned and thus hesitant to pursue, for instance, chemotherapy because of the existence of these factors that could potentially complicate cancer treatment or recovery in an older patient.13 In addition, treatment decision-making could also be influenced by the oncologist's knowledge about the efficacy of a specific cancer treatment such as chemotherapy among older patients and the uncertainty of the safety of such treatment.14 However, it has been documented that older cancer patients could still benefit from chemotherapy and tolerate the side effects with appropriate dose modification and timely management.15–19 However, few data are available regarding the tolerance to chemotherapy among geriatric oncology patients in terms of changes in functional, mental, nutritional, comorbid status, and quality of life (QOL), which are not traditional end points in cancer treatment research.
The current study was part of a prospective pilot study that evaluated predictors of toxicity from chemotherapy in older cancer patients.20 We evaluated the effect of chemotherapy, in particular chemotherapy-related toxicity, on physical function, mental health, nutrition, comorbidity, and QOL in older cancer patients. Previous studies have shown that functional disability and comorbidity were common among senior adult oncology patients.21, 22 A multidimensional comprehensive geriatric assessment (CGA) provides valuable information regarding the medical and psychosocial aspects of a patient undergoing cancer treatment.23 By documenting the changes in these aspects, we would be better able to understand how chemotherapy may affect the life and tolerance level of elderly patients, who are sometimes undertreated. This information can help us to provide optimal cancer therapy and adequate supportive care to older people who undergo chemotherapy.
MATERIALS AND METHODS
Sixty patients age ≥ 70 years who were treated for cancer in the H. Lee Moffitt Cancer Center and Research Institute at the University of South Florida (Tampa, FL) participated in this prospective study between 1998 and 1999. They received a standard chemotherapy regimen for a solid tumor or a nonleukemic hematologic malignancy, without concomitant radiation therapy. Any stage of disease and any line of treatment were accepted. Patients in Phase I studies and those receiving myelosuppressive chemotherapy (e.g., for acute leukemia) were not eligible. Patients were evaluated by a research nurse before initiation of chemotherapy (baseline) and at the end of treatment (EOT). The EOT assessment was administered one cycle after the patient finished the last chemotherapy cycle. When the treatment was longer than 6 months, EOT assessment was truncated at 6 months to minimize the influence of outliers. The study was approved by the institutional review board of the University of South Florida. Informed consent was obtained from all study participants.
Patients' demographic and clinical characteristics were collected from their medical charts. Chemotherapy toxicities were recorded using the National Cancer Institute/World Health Organization toxicity scale. Severe side effects included Grade 4 hematologic and Grade 3–4 nonhematologic toxicities. In most protocols, these are dose-limiting toxicities that prompt a delay in treatment, a reduction in dosage, or a cessation of treatment. Alopecia was not counted. Tumor responses to the chemotherapy/adjuvant treatment were recorded as disease progression, or as a stable, partial, or complete response.
The CGA evaluates the functional, mental, nutritional, comorbid status, and the QOL of a patient. Physical function was measured by the Eastern Cooperative Oncology Group performance status (ECOG PS) and by the Instrumental Activities of Daily Living (IADL). Scores on the ECOG PS range from 0 (fully independent) to 4 (totally bedridden)24 and scores on the IADL range from 10 to 29 (the higher the score, the better the functional status).25 Mental health was measured by the Mini-Mental State Examination (MMSE) and by the Geriatric Depression Scale (GDS) short form with scores ranging from 0 to 30 (the higher the score on the MMSE, the better the cognition)26 and from 0 to 15 (a score of 5 or above on the GDS indicates a high risk of depression).27 Nutrition was measured by the Mini-Nutritional Assessment (MNA) with scores ranging from 0 to 30 (a score of 23 or below indicates a high risk of malnutrition).28 Comorbidity was measured by Charlson's index and the Cumulative Illness Rating Scale-Geriatric (CIRS-G). Scores on the Charlson's index range from 0 (no comorbidity) to 3729 and possible scores on the CIRS-G range from 0 (no comorbidity) to 56.30 Patient's QOL was measured by the 34-item Functional Assessment of Cancer Therapy-General (FACT-G, Version 3.0),31 which contains five subscales: physical, social/family, emotional and functional well-being, and relationship with doctor. Total scores on the FACT-G range from 0 to 136 (highest QOL).
Descriptive data on patients' demographic and clinical characteristics were assessed. Student t tests for paired data were used to evaluate the changes in CGA measurements between baseline and EOT. In addition, changes in CGA scores between patients who experienced severe chemotoxicities (Grade 4 hematologic and/or Grade 3–4 nonhematologic toxicities) and those who did not experience such chemotoxicities were compared using repeated measures of analysis of covariance (ANCOVA) with time as a repeated measure at baseline and EOT. The analysis was adjusted for patients' age and the dose intensity and toxicity potential of the chemotherapy regimens as covariates. Dose intensity was calculated as the percentage of the planned dosage of chemotherapy.32 The toxicity potential of each chemotherapy regimen was calculated empirically with the MAX2 index at the beginning of this study (i.e., before the analysis of the data) as a means to compare the published toxicity of various chemotherapy regimens with low or moderate toxicity. The MAX2 index was used as a covariate in the ANCOVA and its construction has been published in detail elsewhere.20 Finally, changes in scores of CGA measurements among patients with different tumor responses were compared using the analysis of variance. All analyses were performed using SPSS software (version 11.0.1) (SPSS, Chicago, IL).
One of the 60 patients who entered the study died before the initiation of chemotherapy. Of the remaining 59 patients, 37 patients (62.7%) completed both the baseline and EOT assessments and were eligible for evaluation of changes in CGA measurements (except for the FACT-G, which was completed by 34 of these 37 patients at EOT). All patients completed the chemotherapy regimens within 6 months, except for one patient who continued to receive capecitabine beyond 6 months. Of the 22 patients who did not complete the EOT assessment, 2 refused, 1 had surgery, 2 had radiation therapy, 5 were transferred to another hospital or hospice, 3 died after starting chemotherapy, and 9 did not return the questionnaires or lost contact. Baseline CGA measurements and other aspects were not significantly different between the 37 patients who completed the EOT assessment and the 23 patients who did not, except that the former scored higher on the emotional well-being subscale of FACT-G than the latter (16.5 vs. 14.6, P = 0.02).
Table 1 shows the characteristics of the study patients. The average age of the 37 evaluable patients was 75.6 years (range, 70–87). Fourteen percent of them were 80 years and older. All but one patient were white, greater than one-half (56.8%) were females, and the majority (78.4%) were married. At baseline, 24 patients (65%) had an ECOG PS score of 0 and 10 (27%) scored 1 point. Patients were diagnosed with a variety of cancers, with 21 (56.8%) of them at a metastatic stage. Thirty patients (81%) had received cancer treatment before they started chemotherapy in this study. The chemotherapy regimens prescribed for the patients reflected the wide variety of cancers treated. The median dose intensity received was 95% (range, 55–114%) of the planned dose. Greater than one-half of the patients (n = 19) experienced severe side effects from the chemotherapy. Fifteen patients (40.5%) had Grade 3–4 nonhematologic and nine (24.3%) had Grade 4 hematologic toxicities.
Table 1. Patients' Characteristics (n = 37)
No. of patients
CHOP: cyclophosphamide, doxorubin, vincristine, and prednisone; 5-FU: 5-flourouracil; VX 710; a P-glycoprotein inhibitor.
One patient had concurrent lung and prostate carcinoma.
Doxorubicin and cyclophosphamide with/without paclitaxel
CHOP with/without rituximab
5-FU and leucovorin
Gemcitabine and vinorelbine
Mitoxantrone and prednisone with/without VX 710
Cisplatin and paclitaxel
Carboplatin and etoposide
Table 2 shows the CGA scores before and after the patients received chemotherapy. Function-wise, a significant decline in IADL scores, an average of 1.44 points, was found. This corresponds to a significant decline in the subscale scores of physical and functional well-being in the FACT-G measurement of QOL. No significant change was observed in scores on the ECOG PS, MMSE, nutrition, comorbidity, or FACT-G total and other subscales. The distribution of the changes in CGA scores is illustrated in Figure 1. Greater than one-half of the patients had lower scores on the IADL, GDS, MNA, and the physical and functional subscales of FACT-G after receiving chemotherapy treatment. Most patients had no score changes on the ECOG PS and relationship with physicians' scores. For example, 90% of patients scored 0–1 on the ECOG PS at baseline. At EOT, that number decreased slightly to 86%. It is noteworthy that a substantial number of patients even had score improvement in the social and emotional subscales of the FACT-G.
Table 2. Changes in CGA Scores after Receiving Chemotherapy (n = 37)
CGA: comprehensive geriatric assessment; CIRS-G; Cumulative Illness Rating Scale-Geriatric; ECOG PS: Eastern Cooperative Oncology Group performance status; EOT: end of treatment; FACT-G: Functional Assessment of Cancer Therapy-General; GDS: Geriatric Depression Scale; IADL: Instrumental Activities of Daily Living; MMSE: Mini-Mental State Examination; MNA: Mini-Nutritional Assessment.
Functional Assessment of Cancer Therapy-General (FACT-G) scores were available from 34 patients at both end points.
Changes in CGA scores were compared between patients who experienced severe side effects from chemotherapy and those who did not. Table 3 lists CGA variables that showed significant Group × Time differences using repeated measures of ANCOVA. Patients who had severe chemotoxicities had lower physical function scores, as indicated by changes in the ECOG PS and IADL scores, compared with patients who did not have severe chemotoxicities during the treatment. In addition, patients with severe side effects also showed a more significant decline in GDS scores than those without severe side effects. However, their scores on the social well-being subscale of the FACT-G achieved a significant improvement at EOT. Patients with severe chemotoxicities scored 4 points higher at EOT than at baseline on the FACT-G social/family subscale, compared with a 0.84 point decline in that score among patients who did not experience severe chemotoxicities. Although patients without severe side effects had more declines in changes of FACT-G total scores compared with those who experienced severe side effects (−6.48 vs. −0.74), the difference was not statistically significant (P = 0.32).
Table 3. Significant Changes in CGA Scores as a Function of Severe Chemotoxicity Occurrence and Time of Measurement (Adjusted for Age; Dose Intensity, and Toxicity Levels of Different Chemotherapy Regimens)
With chemotoxicity (n = 18)
Without chemotoxicity (n = 19)
P value (interaction)
CGA: comprehensive geriatric assessment; ECOG PS: Eastern Cooperative Oncology Group performance status; EOT: end of treatment; FACT-G: Functional Assessment of Cancer Therapy-General; GDS: Geriatric Depression Scale; IADL: Instrumental Activities of Daily Living.
FACT-G Functional Assessment of Cancer Therapy - General (FACT-G) scores were available from 34 patients (16 with chemotoxicity and 18 without chemotoxicity).
Of the 37 patients who received various chemotherapy regimens, 13 (35%) achieved partial or complete responses, 9 had stable disease, and 7 had cancer progression, whereas 7 (19%) patients received adjuvant chemotherapy. Changes in CGA scores were also compared among patients with different tumor responses to chemotherapy treatment. For the FACT-G total score and the emotional and functional subscale scores, patients with progression of tumor or stable disease showed significant declines. Conversely, patients with partial or complete responses showed improved scores at EOT than at baseline (Table 4). For instance, patients who had partial or complete responses to chemotherapy scored at least 10 points better on the FACT-G total score at EOT than at baseline, compared with an 8 and 13-point decline for patients who were stable or had cancer progression, respectively (P = 0.003). No significant difference was found in terms of changes on the ECOG PS, IADL, MMSE, GDS, MNA, Charlson's index, CIRS-G, or other FACT-G subscale scores.
Table 4. Significant Changes in FACT-G Scores among Patients with Different Tumor Responses
Progression (n = 7)
Stable (n = 9)
Partial response (n = 8)
Complete response (n = 3)
Adjuvant (n = 7)
FACT-G: Functional Assessment of Cancer Therapy-General.
−13.98 ± 15.81
−8.30 ± 14.26
10.05 ± 7.60
11.23 ± 14.07
−10.00 ± 11.21
−2.00 ± 4.55
−2.67 ± 3.04
2.38 ± 2.72
2.67 ± 1.53
0.29 ± 2.14
−6.86 ± 4.95
−3.11 ± 6.51
2.59 ± 3.17
1.27 ± 7.34
−3.54 ± 3.49
This study was limited by the small sample size and the short follow-up period (the median follow-up for 37 study patients was 130 days) due to its pilot nature. To our knowledge, this study is the first to prospectively and empirically document the changes in physical function, mental health, and QOL among older patients undergoing cancer chemotherapy. The results of the current study revealed that cancer patients aged 70 years and older who were treated at one academic tertiary cancer center experienced deteriorations in functional and emotional measurements after receiving chemotherapy. The deterioration was greater among patients who experienced severe side effects from chemotherapy than those who did not, and among patients with tumor progression or stable disease than among those with partial or complete treatment responses. However, other aspects of function and mental health, as well as comorbidity and nutrition, were generally preserved. In addition, there was evidence of improved QOL in social well-being among patients who experienced severe chemotoxicities and in function and emotion among patients who had partial or complete treatment responses.
Older cancer patients in this study showed a significant decline in functional measurement after receiving chemotherapy, as indicated by changes in IADL scores and in the physical and functional subscale scores of the FACT-G QOL instrument between baseline and EOT. Meanwhile, this group of patients also displayed a worse score on postchemotherapy GDS. Overall, greater than one-half of the patients experienced some level of decline in terms of functional and/or emotional measurements after treatment. Previous literature that evaluated treatment effect on these aspects of older cancer patients is sparse and few used standard instruments.17, 18 A recent study comprising younger patients also found that women who received chemotherapy for breast carcinoma reported lower physical well-being on the FACT-G scale than those who did not receive chemotherapy.33 In the current study, ECOG PS scores also declined but did not achieve a significant level as other functional instruments. This most likely was because the ECOG PS, which measures a more basic function, was not as sensitive as the IADL and other scales. The IADL and other scales measure higher levels of functional activities (e.g., transportation and meal preparation) that could be affected by chemotherapy or its toxicity at a much earlier stage of disease treatment. The worsening depressive mood after chemotherapy has not been reported before. For the most part, the GDS scores at EOT were still below the usual cutoff score of 5 on the GDS scale, which measures clinically significant depression. However, similar items on the emotional well-being subscale of the FACT-G showed no significant change after chemotherapy. In fact, the items that patients most often cited on the GDS after treatment were relevant to decreasing activities and low energy level, most likely due to the treatment itself.
It is not clear to what extent this kind of functional and mood status change was due to the cancer itself or to the impact of the side effects experienced by patients during treatment. Further analysis revealed that the declines in function and depression measurements were not significantly different among patients who had different tumor responses to the chemotherapy treatment. However, for the FACT-G total score and the emotional and functional subscale scores, patients with stable disease or progression of tumor showed worse scores at EOT than at baseline. Conversely, patients with partial or complete responses achieved better scores at EOT than at baseline. Other investigators have demonstrated the impact of various side effects that older cancer patients experienced when they received chemotherapy for different cancers, such as nausea, emesis, neutropenia, and neurotoxicity.17, 18 Modestly increased hospitalization rates with increasing age due to complications of adjuvant chemotherapy were observed in Stage III colon carcinoma patients age ≥ 65 years during a 6-month interval.6 In the current study, an incidence rate of greater than 50% was observed for severe treatment side effects, defined as Grade 4 hematologic and Grade 3–4 nonhematologic toxicities. This study was not designed to compare the CGA changes among older cancer patients who received chemotherapy with a control group of patients who did not receive treatment. Instead, the available data allowed us to compare the differences in CGA changes between patients who experienced severe chemotoxicities and those who did not. Not surprisingly, the occurrence of treatment-related severe toxicities contributed to the deterioration in measurements of functional and mood status in this group of patients. Although patients without severe chemotoxicities showed a slight improvement in ECOG PS scores, those experiencing chemotoxicities became more dependent. This was more evident when comparing the IADL scores between these two groups, i.e., there was a 17-fold difference in the score changes. A similar trend was also noted for GDS scores between the these two groups.
Do the above data suggest that older cancer patients are not able to tolerate chemotherapy? Does chemotherapy do more harm than good to this patient population, who can still benefit from the treatment but who are usually undertreated? Although greater than 80% of the study participants had received previous cancer treatment (surgery, chemotherapy, radiation, and other therapy), the baseline CGA scores showed that they were, in general, functionally independent and mentally sound with good nutrition and QOL. Declines notwithstanding, the changes in CGA scores observed in this study were small in magnitude clinically. Caution should be taken when extrapolating statistically significant findings to clinical meaningful interpretations. In addition, these results represented patients' status at the end of chemotherapy, probably the highest level of repercussion of that treatment. It is likely that if they were followed for a longer period of time after the treatment, some aspects of physical and mental health could eventually return to prechemotherapy baseline levels. Equally important is the fact that patients' nutrition, comorbidity, and other QOL levels were largely preserved and that the majority of patients were retained a fairly normal level of basic function (ECOG PS) and cognition (MMSE) during treatment. Comorbidity, as measured by the Charlson index and the CIRS-G scale in this study, was not uncommon among the study patients. However, it did not seem to be affected significantly by the treatment. The reasons for the little change in comorbidity scales could be that the treatment per se did not lead to additional comorbid conditions or the conditions, if any, did not develop within the follow-up period. Furthermore, although patients with severe chemotoxicities encountered more declines in terms of functional and emotional measurements than other patients, they also achieved more gains in social/family well-being and QOL. A reasonable explanation could be that these patients received more social support from family and friends when they were suffering from more severe and detrimental toxic effects.
More data from studies like this one are still needed to provide definite answers to the above questions. Nevertheless, the current study does not support the notion that older patients are not able to tolerate cancer chemotherapy. Patients in this study received a fairly high level of the planned doses (95% dose intensity on average). Older age does not necessarily lead to more severe toxic effects from chemotherapy, compared with younger patients, nor does age-associated comorbidity.17, 19, 20 In our detailed analysis of the predictive power of patients' characteristics on toxicity from chemotherapy, baseline functional, emotional, and mental status did not correlate significantly with the occurrence of severe chemotherapy-related toxicity in older cancer patients.20 Future research in this field should use prospective designs to follow a larger sample and a more diverse geriatric oncology population receiving various chemotherapeutic regimens. The outcome measurements need to be comprehensive enough to capture clinical, psychological, social, and QOL dimensions with valid instruments. They also need to take into account potential confounding factors (e.g., potential interaction effects between chemotherapy and other therapies) that could affect the treatment efficacy of older people.
A strength of this study is that it encompasses nontraditional end points in outcome research of cancer treatment. Recognizing these treatment-related changes can help oncologists and other health care providers to carefully monitor possible deteriorations in functional and mood status among older cancer patients as a result of chemotherapy treatment. Efforts should be made to identify older cancer patients who may be at increased risk of developing treatment toxicities and to establish patient profiles that can predict and eventually prevent such incidents.20 This information can also help older patients and their families to anticipate and prepare for the changes and the inconvenience that chemotherapy may cause. It is necessary to pay more attention to providing adequate supportive care to this patient group through antidotes to drug toxicity such as hematopoietic growth factors, prophylactic antibiotics, timely management of complications14 and, if necessary, early psychiatric intervention for depression.
Another important issue is to increase physicians' knowledge about the efficacy of treating older cancer patients with chemotherapy and their tolerability to such treatment. As a result, patterns of cancer care in older patients would become more optimal through better utilization of the adequate treatment options that are currently available. A recent survey revealed that oncologists' perceptions of barriers to accrual of older patients with breast carcinoma to clinical trials were multidimensional.13 Some of the most important barriers include concerns of treatment toxicity, and the significant comorbid conditions that older patients have, which are not excluded by the protocol but may affect how they would respond to treatment. Providing physicians with educational materials concerning treatment toxicity in older people was suggested by oncologists as one of the most effective interventions for improving the accrual of older patients to trials. Finally, it is equally important to educate older patients and their families of the myths regarding chemotherapy because they may choose not to accept chemotherapy based on the misperception that they would experience greater toxic effects than younger patients or that they would tolerate the treatment poorly.
Older cancer patients undergoing chemotherapy may experience toxicity but generally are able to tolerate it with limited impact on independence, comorbidity and QOL levels. Comprehensive geriatric asssessment is very useful in detecting treatment-related changes in geriatric oncology patients and should be incorporated into clinical outcome analysis.22, 34 It is important to recognize and monitor these changes during cancer treatment. Future research needs to identify older people who are most susceptible to treatment toxicity and to subsequent functional and emotional status change. In addition, we should find ways to provide optimal cancer treatment with adequate supportive care as well as treatment modification strategies.