Geriatric assessment in older patients with a hematologic malignancy: a systematic review

Abstract

The aim of this systematic review is to give an update of all currently available evidence on the relevance of a geriatric assessment in the treatment of older patients with hematologic malignancies. A systematic search in MEDLINE and EMBASE was performed to find studies in which a geriatric assessment was used to detect impaired geriatric domains or to address the association between geriatric assessment and survival or clinical outcome measures. The literature search included 4,629 reports, of which 54 publications from 44 studies were included. Seventy-three percent of the studies were published in the last 5 years. The median age of the patients was 73 years (range, 58-86) and 71% had a good World Health Organization (WHO) performance status. The median prevalence of geriatric impairments varied between 17% and 68%, even in patients with a good WHO performance status. Polypharmacy, nutritional status and instrumental activities of daily living were most frequently impaired. Whereas several geriatric impairments and frailty (based on a frailty screening tool or summarized geriatric assessment score) were predictive for a shorter overall survival, WHO performance status lost its predictive value in most studies. The association between geriatric impairments and treatment-related toxicity varied, with a trend towards a higher risk of (non-)hematologic toxicity in frail patients. During the follow-up, frailty seemed to be associated with treatment non-completion, especially when patients were malnourished. Patients with a good physical capacity had a shorter stay in hospital and a lower rate of hospitalization. Geriatric assessment, even in patients with a good performance status, can detect impaired geriatric domains and these impairments may be predictive of mortality. Moreover, geriatric impairments suggest a higher risk of treatment-related toxicity, treatment non-completion and use of healthcare services. A geriatric assessment should be considered before starting treatment in older patients with hematologic malignancies.

Introduction

Given the increasing life expectancy and aging of the population, there is a growing number of older patients with cancer, including patients with a hematologic malignancy. Worldwide, hematologic malignancies account for approximately 9% of all cancers and are the fourth most frequently diagnosed cancer.¹ At present, 60% of these patients are older than 65 years and this proportion will increase in the future.^2,3

Over the last decades, treatment options for hematologic malignancies have progressed. For example, the initial treatment of patients with multiple myeloma changed from cytotoxic chemotherapeutics to better-tolerated agents such as immuno-modulatory drugs or monoclonal antibodies.⁴ Moreover, the proportion of older patients with myelodysplastic syndrome or acute myeloid leukemia undergoing hematopoietic stem cell transplantation has increased, partly due to expansion of age limits.^5,6

However, it can be difficult to deliver optimal cancer treatment tailored to individual needs of an older patient, particularly as older patients are frequently excluded from clinical trials.⁷ Older patients constitute a heterogeneous population due to large differences in comorbidity, functional capacity and psychological and physical reserves. As a result, the benefit of treatment can differ and patients with comorbidity or geriatric impairments are particularly at risk of adverse health outcomes. Choosing the optimal treatment for these patients is a challenge.

It is therefore recommended that the degree of frailty of older patients is assessed.⁸ Frailty is a biological syndrome which can exist alongside age, comorbidity or disease characteristics. Over the years, numerous definitions of frailty have been formulated and there is still no consensus on a definition.⁹ There are two commonly used approaches to define frailty. The first defines frailty based on phe-notypic criteria including reduced grip strength, walking speed, physical capacity, level of energy and weight loss. Patients are considered frail if three or more criteria are present.¹⁰ The second approach proposes a frailty index which is an accumulation of patient’s deficits. These deficits consist of physical or cognitive symptoms, functional impairments, abnormal laboratory values and comorbidities.^11,12 In daily practice, frailty is a dynamic state which needs a multidimensional approach and might have various implications in different scenarios.

An appropriate method to assess the level of frailty of older patients is a geriatric assessment.^8,13 This consists of a systematic assessment of an older patient’s health status focusing on somatic, psychological, functional and social domains. Different tools can be used to detect geriatric impairments in these domains.¹⁴ Moreover, frailty screening tools were developed in order to identify older patients who require a full geriatric assessment.¹⁵ Nowadays, some form of geriatric assessment is increasingly incorporated in hemato-oncologic care to customize hemato-oncologic treatment.¹⁶

In 2014, we published a systematic review on the value of performing a geriatric assessment in older patients with a hematologic malignancy, demonstrating that such an assessment can detect multiple health issues and has predictive value for clinical outcome in older patients with a hematologic malignancy.¹⁷ However, evidence was limited, especially regarding clinical outcomes such as treatment-related toxicity, treatment completion or physical functioning after treatment. Since then, many new studies have been published on this subject. The aim of this present systematic review is, therefore, to give an update of all currently available data on the association between geriatric impairments and hematologic cancer-related outcomes.

Methods

Search strategy and article selection

Our aim was to identify studies concerning patients with a hematologic malignancy in which a geriatric assessment was used to detect geriatric impairments or which addressed the association between baseline geriatric assessment and outcome.

Geriatric assessment was defined as an assessment composed of at least two of the following domains: cognitive function, mood, nutritional status, activities of daily living (ADL), instrumental activities of daily living (IADL), polypharmacy (using five or more drugs), objectively measured physical capacity (for instance, gait speed, hand grip strength or balance tests), social support and frailty (assessed with a frailty screening tool or by summarizing the geriatric assessment). As prior medical history/comorbidity and performance status are routine parts of the hematologic work-up, these were not counted as domains of the geriatric assessment for this particular systematic review. The following items were defined as outcomes: prevalence of geriatric impairments, change in oncologic treatment plan, toxicity of chemotherapy, healthcare utilization, physical functioning after treatment, quality of life after treatment and mortality.

The following search was performed on March 4, 2019 and updated on January 20, 2020, in both MEDLINE and EMBASE: ((("Hematologic Neoplasms"[Mesh] OR "Leukemia"[Mesh] OR "Lymphoma"[Mesh] OR "Multiple Myeloma"[Mesh] OR "Myelodysplastic Syndromes"[Mesh] OR leukemia[tiab] OR leukaemia[tiab] OR lymphoma*[tiab] OR hodgkin*[tiab] OR non-hodgkin*[tiab] OR (multiple myeloma[tiab]) OR myelodysplas*[tiab] OR (haematolog* AND malignan*[tiab]) OR (hematolog* AND malignan*[tiab]) OR (myeloid[tiab] OR lymphoid[tiab] AND neoplas*[tiab]) OR myeloproliferative[tiab] OR (plasma cell neoplas*[tiab]) OR plasma cell dyscrasia*[tiab] OR (myeloid[tiab] AND sarcoma*[tiab]) OR waldenstrom[tiab] OR myelofibrosis[tiab] OR mastocystosis[tiab] OR (polycyth* AND vera[tiab]) OR (essential AND thrombocyt*[tiab]))) AND (("frailty"[All Fields] OR "Geriatric Assessment"[Mesh] OR frail*[tiab] OR vulnerabl*[tiab] OR geriatric assessment*[tiab] OR geriatric*[tiab]))

No age or language limitations were applied. All search results until 2013 were reviewed previously by Hamaker et al.¹⁷ We therefore limited our search to studies published after January 1, 2013. The titles and abstracts of all studies retrieved by the search were assessed by one reviewer (ES) to determine which warranted further examination. The full texts of all potentially relevant articles were subsequently screened. We excluded studies that did not focus exclusively on hematologic malignancies. Finally, references of included studies were cross-referenced to retrieve any additional relevant citations. Eligible studies from all searches (2013, 2019, 2020) were subsequently combined to form the final study selection.

Data extraction

For each eligible study, data regarding study design and results were independently extracted by two authors (ES and AV). Extracted items included the type of study, study population (number of patients, median age, malignancy subtype, stage, treatment) and the content of the geriatric assessment. Only validated tools from the geriatric assessment were included. If multiple tools were used to assess one geriatric domain, the result of the most commonly used tool was noted. We registered the prevalence of geriatric impairments, and the reported results on the association between the geriatric assessment and outcome measures. If necessary, study authors were contacted to obtain additional data.

Quality assessment

The methodological quality of each of the studies was assessed independently by two reviewers (ES and AV), using the Newcastle-Ottawa scale adapted to this subject (Online Supplementary Table S1).¹⁸ As our main focus was on older patients with hematologic malignancies, we classified studies of patients with a me(di)an age less than 68 years old, or with more than one third of the patients younger than 65 years old, as not being fully representative of our target population. Disagreements among the reviewers were discussed during a consensus meeting and in the case of persisting disagreement, the assistance of a third reviewer (MH) was enlisted.

Data synthesis and analysis

Due to the heterogeneity in the populations of patients and in study designs, with a wide variety in content of geriatric assessments, a meta-analysis was not considered feasible. We therefore summarized the study results to describe our main outcomes of interest.

Results

Study characteristics

The literature search yielded 4,629 citations (832 from MEDLINE and 3,797 from EMBASE), of which 403 were duplicates and 4,184 were excluded for other reasons (Online Supplementary Figure S1). This resulted in 42 eligible publications from 34 studies. Cross-referencing yielded four additional publications. Eight publications from the 2014 review by Hamaker et al.¹⁷ were also eligible. Thus, we ultimately included 54 publications from 44 studies in this review.^19–72

The characteristics of these 44 studies are summarized in Table 1. Seventy-three percent were published in the last 5 years. The median sample size of the studies was 100 (range, 25-869), and the me(di)an age of included patients ranged from 58 to 86 years. Eight studies focused on acute myeloid leukemia and/or myelodysplastic syndromes,^19–25,27 two on chronic lymphocytic leukemia,^28,29 13 on lymphoma,^30–42 seven on multiple myeloma,^42–48 and 15 studies included various hematologic malignancies.^49–63

Table 1.

Characteristics of studies on the association between the geriatric assessment and outcome measures.

The median number of domains addressed in the geriatric assessment was four (range, 2-9). These included ADL in 30 studies (68%), IADL in 37 (84%), cognition in 29 (66%), mood in 24 (55%) and objectively measured physical capacity in 20 studies (46%). Domains less commonly assessed were nutritional status (11 studies; 25%), social support (8 studies; 18%), polypharmacy (13 studies; 30%) and frailty (8 studies assessed with a frailty screening tool and 17 studies by summarizing the geriatric assessment; 18% and 39%, respectively).

The prevalence of geriatric impairments was assessed in all studies (100%). The association between geriatric impairments and mortality was addressed in 33 studies (75%), treatment-related toxicity in ten studies (23%), treatment completion in five (11%) and healthcare utilization in seven studies (16%). No studies assessed the association of geriatric impairments with physical functioning or quality of life after treatment.

Quality assessment

The results of the quality assessment are shown in Figure 1. Detailed results per study are listed in Online Supplementary Table S1. The overall quality of the studies was good. Nine studies included a significant proportion of younger patients (i.e. median age less than 68 years old, or more than one third of the patients younger than 65 years old);^{22,27,41,43,46,48,50,58,59} these studies were assessed as not being fully representative of the target cohort of the average older patients with a hematologic malignancy. Similarly, eight studies focused on a very specific treatment^{20,23,24,31,51,55,56,60} which we considered as not fully representative of our target population. Overall, the duration of follow-up was sufficient but in nine studies the follow-up rate was less than 90%^24,30,46 or the adequacy of follow-up was not reported.^{27,32,33,56,57,62} There were no other quality concerns.

Figure 1.

Outcome of the quality assessment. Details are reported in Online Supplementary Table S1A (quality assessment questionnaire) and Online Supplementary Table S1B (assessment per study).

Prevalence of geriatric impairments

The prevalence of geriatric impairments is shown in Table 2. The most commonly reported issues were polypharmacy (in a median of 51% of patients; range, 17-80%), risk of malnutrition (median 44%; range, 27-82%) and IADL impairments (median 37%; range, 3-85%). Impaired physical capacity (median 27%; range, 3-80%), ADL impairments (median 18%; range, 4-67%), symptoms of depression (median 25%; range, 10-94%), and cognitive impairment (median 17%; range, 0-44%) were less common. Four studies that addressed social support showed impairment in a median of 20% (range, 7-54%). The median proportion of patients seen as frail based on a frailty screening tool was 68% (range, 25-76%). The median proportion of patients screened as frail based on a summarized geriatric assessment score was 45% (range, 10-88%).

Table 2.

Comparison of impaired performance status with impairments in geriatric domains.

Overall, the median proportion of patients with at least one geriatric impairment was 51% (range, 9-82%). By comparison, the median proportion of patients with a World Health Organization (WHO) performance status of 2 or higher was only 29% (range, 1-91%). Even in studies in which the median age of patients was ≤65 years old, or a small proportion of patients had a poor WHO performance status, geriatric impairments were quite common. For example, in one study, 93% of included patients had a WHO performance status of 0-1; nonetheless, 45% of patients had impairments in IADL, 39% in physical capacity and 25% were frail based on a frailty screening tool (Table 2).⁴⁹

Association between geriatric impairments and mortality

The association of geriatric impairments with mortality was addressed in 33 studies (Table 3). In univariate analysis, 27 out of 29 studies (93%) showed a significant association between at least one geriatric impairment and mortality. The association between a specific geriatric domain and mortality varied between 0 and 74%. Polypharmacy was assessed in only two studies and showed no association. For all other geriatric domains except mood, nutritional status and social support, at least 50% of the studies reported a univariate association between impairment and mortality. IADL, ADL, impaired physical capacity and cognition were most frequently associated with mortality (in 74%, 67%, 63% and 55% of the studies, respectively). In multivariable analyses, ADL, IADL, impaired physical capacity and cognition remained associated with mortality (in 40%, 62%, 50% and 50% of the studies, respectively). Moreover, at least 75% of all studies that assessed frailty (with a frailty screening tool or by summarizing the geriatric assessment), demonstrated that this was associated with mortality in multivariable analyses.

Table 3.

The association of geriatric assessment, age, performance status, and comorbidity with mortality.

Risk factors for mortality commonly used in hematooncology such as age, WHO performance status and comorbidity were also associated with mortality in univariate analysis (in 79%, 63% and 64% of the studies, respectively). However, in multivariable analyses, this association was no longer present for WHO performance status; age and comorbidity retained their association with mortality in 43% and 47% of the studies, respectively.

Association of geriatric impairments with treatment-related toxicity

Ten studies assessed geriatric impairments in relation to treatment-related toxicity.^{28,29,32–34,37,45,48,55,59} Four out of six studies in which frailty was assessed (based on a summarized geriatric assessment score) reported an association between frailty and treatment-related toxicity.^33,34,45,48 This included hematologic toxicity in one study,³³ non-hematologic toxicity in two studies^45,48 and overall toxicity in one study.³⁴ One study showed an association specifically between impaired IADL and treatment-related infections in patients with chronic lymphocytic leukemia.²⁸ In studies in which patients with various hematologic malignancies were included, associations between physical capacity⁵⁵ or cognition⁵⁹ and treatment-related toxicity were demonstrated. No other associations between frailty (based on a summarized geriatric assessment score) or individual geriatric domains and treatment-related toxicity were found in these ten studies.

Association of geriatric impairments with treatment completion

The association of geriatric impairments with the ability to complete the proposed treatment was studied in five studies.^{25,32,36,38,45} Four out of five studies found an association between geriatric impairments and treatment completion. The risk of treatment non-completion was significantly higher in frail patients (based on a summarized geriatric assessment score or frailty screening tool) than in fit patients.^25,36,38,45 Three studies showed a significant association between a specifically geriatric domain and treatment non-completion: in two studies that included patients with non-Hodgkin lymphoma, malnutrition was associated with treatment non-completion.^36,38 Another study, in which patients with acute myeloid leukemia or myelodysplastic syndrome were included, showed an association between impaired IADL, impaired physical capacity or cognitive impairment and treatment non-completion. In this study, no other geriatric impairments or clinical characteristics such as comorbidity or WHO performance status were associated with treatment non-completion.²⁵

Association of geriatric impairments with healthcare utilization

The association of geriatric impairments and health care utilization was addressed in seven studies.^{32,46,53,55,57,59,62} Six out of these studies showed an association between geriatric impairments and health care utilization. In four studies impaired physical capacity was associated with increased use of health care.^46,55,57,62 In patients with various hematologic malignancies, other geriatric impairments, such as ADL,⁶² IADL,⁵³ cognition⁵⁹ and mood⁴⁶ were also associated with health care utilization. In one study with patients with diffuse large B-cell lymphoma, no association was found between frailty (assessed by a summarized geriatric assessment score) and unplanned admissions.³²

Discussion

This systematic review of 44 studies shows that impairment in geriatric domains is common among older patients with a hematologic malignancy, even in those with a good performance status. The most relevant impairment is frailty (assessed with a frailty screening tool or by summarizing the geriatric assessment), which showed an association with mortality, treatment-related toxicity and treatment non-completion. Other relevant geriatric impairments were IADL functioning, nutritional status and polypharmacy. Impaired physical capacity was mainly associated with healthcare utilization.

These data should, however, be interpreted with care. The included studies are heterogeneous in study population, design, treatment regimens, content of geriatric assessment and reported outcomes. Various hematologic malignancies can have very different disease courses and require very different intensities of treatment; geriatric impairments that were associated with outcome in one setting may not retain their predictive value in another disease entity. In addition, the content of geriatric assessments, including the definition of frailty (assessed by summarizing the geriatric assessment), was not consistent. Moreover, geriatric impairments were mainly assessed with screening tools (for example, the Mini-Mental State Examination for cognition), and it should be realized that the ensuing results are not the same as an actual diagnosis made by a comprehensive geriatric assessment. Given this heterogeneity, a meta-analysis or a meaningful subgroup analysis (for example, by type of malignancy) could not be performed; and interpretation and extrapolation of results should be done with caution. Another limitation of this review is the procedure used to select the literature. We decided to select only those studies for which a full text is available and which performed a geriatric assessment with validated tools covering at least two geriatric domains. Studies which focused on a single impairment and its relation to outcome were not included, meaning some information on individual associations may have been missed.

Despite these limitations, this review provides a thorough update and overview of all currently available evidence on the relevance of a geriatric assessment for older patients with a hematologic malignancy. At the time of the previous systematic review, by Hamaker et al.,¹⁷ the evidence was limited because of a lack of published studies. In the last 5 years, the number of publications concerning the association of geriatric assessment with outcomes in patients with hematologic malignancies has increased greatly, enabling a useful update on the available data.

Performing a geriatric assessment could have an additive value to clinical judgment, treatment allocation and the implementation of non-oncological interventions.

In daily practice, oncologists are able to detect obviously frail patients by clinical judgment. However, estimating the reserve capacity and resilience of the remaining older patients by clinical judgment is difficult, as demonstrated by the discrepancy between performance status and geriatric assessment. In addition, it can be challenging to distinguish whether the detected vulnerabilities are disease-related or patient-related. This may require a more thorough evaluation of the patient’s overall health status, including consultation of a geriatrician.

The impact of performing a geriatric assessment on treatment allocation has already been demonstrated in older patients with solid malignancies.^73,74 In a systematic review, the oncological treatment plan was altered in 28% of patients after geriatric assessment, primarily resulting in a less intensive treatment option. This review showed that using a geriatric assessment to guide treatment decisions appeared to have a positive effect on clinical outcome, resulting in less treatment-related toxicity, fewer complications, and increased treatment completion.⁷⁵ For example, in patients with cognitive impairments, treatment decisions should be made with great care because of the higher risk of chemotherapy-related progression of cognitive dysfunction, treatment non-compliance and death.^52,71

In order to tailor cancer treatment to individual needs, it could be interesting to incorporate patient-reported outcome measures (PROMS) into the treatment decision-making process. PROMS, such as physical functioning and quality of life during and after treatment, were hardly assessed in the studies included in this review, despite quality of life being of primary importance to many older patients.⁷⁶ It is, therefore, highly relevant that future studies address the association between geriatric impairments and PROMS.⁷⁷

In addition to clinical judgment and treatment allocation, a geriatric assessment can be used to introduce non-oncological interventions before and during treatment in the hope of improving the patient’s health status, resilience and treatment tolerance. However, evidence concerning the effectiveness of such non-oncological interventions is limited. Previous research suggests that perhaps physiotherapy^78,79 as well as nutritional counseling^80–82 can improve survival, physical functioning and quality of life. Non-oncological interventions in older patients undergoing chemotherapy can improve treatment completion and treatment modifications.⁸³ The process by which a patient’s condition can be enhanced before starting treatment is called prehabilitation. Although results of the first studies assessing the effectiveness of prehabilitation in patients with solid malignancies are promising,^84,85 the level of evidence is weak, making it too early to draw definitive conclusions. Currently, according to clinicaltrials.gov (searched February 5, 2020), there are 29 ongoing trials in which the effect of non-oncological interventions on clinical outcome measures in older cancer patients is being assessed; six out of these 29 trials focus on hematologic malignancies.⁸⁶ Based on these numbers, further data will follow in the coming years.

In conclusion, this review demonstrates the relevance of performing a geriatric assessment in older patients with a hematologic malignancy. Although the results should be interpreted and extrapolated carefully, our review shows that even in patients with a good performance status, a geriatric assessment can detect geriatric impairments that might be predictive of mortality. Moreover, geriatric impairments seem to be associated with a higher risk of treatment-related toxicity, treatment non-completion and utilization of healthcare services. Future research is needed to extend these findings with a focus on reserve capacity, resilience, quality of life and the effectiveness of non-oncological interventions.