Eastern Cooperative Oncology Group Performance Status Scale as a Screening Tool for Sarcopenia in Onco-Hematology Patients

Currently defined by the European Working Group on Sarcopenia in Older People (EWGSOP2) as a progressive and generalized skeletal muscle disorder associated with increased likelihood of adverse outcomes, sarcopenia is a formally recognized muscle disease with an International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM ]M62.84]) diagnosis code.1 A prognostic factor consistently reported in the literature as potentially modifiable for adverse clinical outcomes in solid tumors,2 3 4 5 6 7 some studies also indicate that sarcopenia has an impact on hematologic cancer, showing an association with lower overall survival in patients diagnosed with non-Hodgkin lymphoma and undergoing chemotherapy.8 9 In patients submitted to hematopoietic stem cell transplantation (HSCT), sarcopenia can be associated with lower overall survival and higher non-relapse mortality.10

Computed tomography (CT) of the third lumbar vertebra level is the gold standard for assessing muscle mass and diagnosing sarcopenia in patients with cancer,1 11 although positron-emission tomography-CT (PET-CT) can also be used since it is often performed in cases of hematologic diseases.12 13 However, such tests are not usually requested to assess body composition alone and have a high cost, thus preventing the adequate follow-up of patients in all institutions and on a routine basis.14 Moreover, given the importance of detecting sarcopenia in patients with hematologic cancer, screening tools with adequate sensitivity and specificity that are affordable and feasible in different institutions are essential and can guide the diagnostic process in clinical practice.

Currently, the EWGSOP2 proposes the use of the Strength, Assistance with walking, Rising from a chair, Climbing stairs, and Falls (SARC-F) questionnaire as a screening tool for sarcopenia in clinical practice.1 15 16 Although it has low-to-moderate sensitivity (3–10%) and high specificity (94–99%), this tool can efficiently predict more severe sarcopenia cases and was validated for the elderly population.17

Functional and self-care capacity are potentially compromised in sarcopenia and are assessed in clinical practice using tools such as the functional status scale proposed by the Eastern Cooperative Oncology Group Scale of Performance Status (ECOG-PS), which ranges from 0 (asymptomatic, fully active) to 5 (death).18 19 20 Since previous studies have shown an association of ECOG-PS scores with cancer outcomes,21 22 the feasibility and validity of the ECOG-PS for an early sarcopenia screening have been debated, particularly in cases of anemia, a frequent condition in onco-hematology patients and potentially harmful in those with sarcopenia.23 24 25 26

Considering the above, the objective of the present study was to assess the sensitivity of the ECOG-PS with scores other than 0 as a screening tool for sarcopenia and to characterize the prevalence of sarcopenia as well as the factors associated with it in onco-hematology patients, according to Fig. 1.

The results of the descriptive analysis of the 104 patients are summarized in Table 1.

The ECOG-PS was collected during outpatient follow-up visits and, for patients with more than one appointment within the analyzed period, the score from the most recent visit was considered for the following results. During this period, 71 patients (68.3%) were not undergoing treatment, and, of these, 39.4% were at the time of diagnosis.

The prevalence of sarcopenia was 57.7%, while 71.2% of the patients were classified based on the ECOG-PS as being without functional impairment.

An ECOG-PS score of ≥ 1 was significantly associated (p = 0.016) with a sarcopenia diagnosis when considering a level of significance of 5%.

Table 2 shows the results of the univariate and final regression model.

The univariate analysis showed that age ≥ 60 years, male sex, targeted therapy for cancer, and an ECOG-PS score of ≥ 1 were positively associated with a sarcopenia diagnosis. A higher body mass index (BMI) value was negatively associated with sarcopenia, and lower SMI values were related to sarcopenia.

The multivariate model did not obtain significant associations. The final stepwise model was conducted to evaluate the response of the ECOG-PS variable to the association considering the objectives of this article.

The final model showed that male sex was positively associated with a sarcopenia diagnosis, with male individuals having a 4.11 higher chance of developing muscle disease than female individuals. Similarly, an ECOG-PS score of ≥ 1 was also positively associated with sarcopenia, with a level of significance of 5%.

The variable “anemia” showed a casuistic incompatibility with the statistical model and multicollinearity with the other variables and was, thus, not included. Corticosteroid use (p = 1.000) and disease status (p = 0.830) showed a highly non-significant association and were not included in the models. Based on these results, a Spearman's correlation was conducted between SMI value, which is essential for sarcopenia diagnosis, and the classification on the ECOG-PS scale, obtaining a negative correlation (-0.296; p < 0.01).

Based on the association of an ECOG-PS score of ≥1 with a sarcopenia diagnosis, the accuracy of the scale as a screening tool for sarcopenia was calculated in a sample of mostly onco-hematology patients with sarcopenia, as shown in Table 3.

The ECOG-PS showed a sensitivity and specificity of 40.7% and 83.8%, respectively, as a screening tool for sarcopenia in this sample of onco-hematology patients, showing a greater ability to detect the absence of sarcopenia.

In the present retrospective cohort study, we showed a considerable prevalence of sarcopenia among onco-hematology patients, a condition also associated with physical dysfunction on the ECOG-PS scale. A sarcopenia diagnosis in onco-hematology patients has been previously reported. A systematic review that included 6,894 patients diagnosed with solid and hematologic tumors found a prevalence of pretherapeutic sarcopenia of 38.6% (95% confidence interval [CI] 37.4–39.8) based on CT scans, associating the condition with an increased mortality risk and impact on antineoplastic treatment,3 while a meta-analysis of 1,578 patients found a prevalence of 39.1% in onco-hematology patients based on CT scans and association with lower overall survival of individuals diagnosed with non-Hodgkin lymphoma.33 Regarding the 104 onco-hematology patients included in the current study, most were diagnosed with lymphoma (76.0%), were not undergoing treatment at the time of the PET-CT examination (68.3%), and had sarcopenia (57.7%), with a prevalence higher than that reported in the literature. Although caution should be taken when comparing results obtained by different assessment methods, prior studies diagnosing sarcopenia based on PET-CT are scarce. However, the prevalence of sarcopenia in the patients included in the present study was significant.

Considering sarcopenia as a multifactorial diagnosis potentially associated with hematologic cancer and its related inflammatory and metabolic changes1 3 and the importance of early detection, screening instruments that facilitate the diagnosis of this muscle disease are crucial; such tools help to reveal the patient population that would benefit from a more detailed sarcopenia assessment.

Our study found an association of sarcopenia diagnosis with male sex (p = 0.027). A study of 122 individuals aged ≥ 70 years living in a nursing home detected sarcopenia in 32.8% of the population, with the condition most often found in men (68%) rather than in women (21%) (p < 0.001).34 Another study investigated the prevalence of sarcopenia diagnosed by bioelectrical impedance analysis in 164 patients who had hematologic diseases and were scheduled for HSCT. Sarcopenia was found in 50.6% of patients and was associated with the lower BMI (odds ratio 0.70; p < 0.01) and sex (odds ratio 3.09; p < 0.01); this led to the conclusion that male patients can be more susceptible to sarcopenia,35 with a potential reduction in daily life activity compared with female patients.

Furthermore, an ECOG-PS score of ≥ 1 proved to be a good predictor of a sarcopenia diagnosis based on a significant association between the 2 variables (p = 0.048). Other studies have also investigated the association of the ECOG-PS scores with a sarcopenia diagnosis. A retrospective study of 42 patients, who were diagnosed with advanced non-small cell lung cancer and had received previous treatment, detected sarcopenia in 52.4%; however, the ECOG-PS scores were not significantly different in patients with or without sarcopenia (p = 0.13), possibly due to the reduced sample size.36 However, although not significant, patients with worse PS tended to have reduced muscle mass compared with those with good PS, thus reinforcing the need for further studies on this subject.36 Another retrospective study of 100 patients with advanced cancer treated with immunotherapy found a significant association between the SMI determined in the sarcopenia assessment and ECOG-PS (p = 0.0324). The majority of individuals had non-small cell lung cancer (NSCLC), melanoma, or kidney as their primary tumor.37 Although the hypothesis of predicting sarcopenia using the ECOG-PS is promising, studies assessing the relationship between these variables are scarce. As previously mentioned, the ECOG-PS also considers functional and self-care capacity, which are potentially compromised in sarcopenia, thus resulting in an increased risk of falls and fractures, as well as motility disorders, and greater restrictions on the individual's daily living activities.1 38

In this context, a cut-off score of ≥ 1 was used when analyzing the ECOG-PS as a screening tool for sarcopenia, as, according to the scale, any score other than 0 already indicates some degree of functional impairment.

Questionnaires, such as the SARC-F, were created to screen patients at risk of sarcopenia, and were previously characterized with low-to-moderate sensitivity (3–10%) and high specificity (94–99%) in an elderly Chinese population17 has since then been used in the elderly population.1 15 In a cross-sectional study of 179 non-institutionalized elderly individuals, the SARC-F showed a sensitivity of 33.3% and a specificity of 84.2% and was associated with the anthropometric measurement of the calf circumference; then, these results were compared with sensitivity and specificity associated with a diagnosis of sarcopenia by dual-energy X-ray absorptiometry.39 An improved performance and sensitivity (66.7%) of the SARC-F as a screening tool for sarcopenia was observed when associating it with calf circumference measurements (area under the curve: 0.736 [95% CI 0.575–0.897]; comparing with SARC-F alone: p = 0.027), without compromising its specificity (82.9%).39 However, similar to the SARC-F, this tool is also aimed at the elderly population.

Our study has several limitations, including its retrospective design with simultaneous data collection, and the use of the ECOG-PS variable taken from electronic medical records, whose classification permeates the evaluator's subjectivity, although following intrinsic instructions to the scale.

In conclusion, the ECOG-PS tool showed a sensitivity of 40.7% and, consequently, an increased negative predictive value (51.4%), indicating a high probability of false negatives. Although our results do not demonstrate adequate accuracy for using the scale as a screening tool for sarcopenia in onco-hematology patients, the study hypothesis remains relevant considering the association found between the variables and a sarcopenia diagnosis. Furthermore, the intrinsic characteristics of the ECOG-PS, such as simplicity, speed of use, and low cost, support its use in diverse health centers and by different members of a multidisciplinary team, but we reinforce the need for further studies.