Novel prediction equation for appendicular skeletal muscle mass estimation in patients with heart failure: Potential application in daily clinical practice

Abstract

Sarcopenia, reduction in muscle mass and function, is frequently observed in patients with chronic heart failure (CHF).1,2 Appendicular skeletal muscle mass (ASM), which is a diagnostic criterion for sarcopenia, is quantitatively measured by using dual-energy X-ray absorptiometry (DEXA).3 However, DEXA is a costly and hospital-based modality that is not suitable for a daily clinical setting. For this reason, the prevalence and impact of sarcopenia in CHF patients have not been investigated in large population-based and epidemiological studies. On the other hand, anthropometric measurements such as calf circumference (CC) and mid arm circumference (MAC) are inexpensive, repeatable and quantitative methods for assessing nutritional status and body composition.4,5 However, prediction equations derived from gold standard techniques are needed to validate anthropometric assessment of ASM.6,7 The aim of this study was to develop and cross-validate an equation to predict ASM from anthropometric measurements in Japanese patients with CHF by the use of DEXA-measured ASM as a gold standard. We retrospectively analysed data for 196 consecutive CHF patients (73 ± 14 years of age; women, 52%) who received a DEXA scan (Horizon A DXA System; HOLOGIC, Waltham, MA, USA) and measurements of CC and MAC. ASM was calculated as the sum of bone-free lean masses in the arms and legs. The appendicular skeletal muscle mass index (ASMI) was defined as ASM/height2. CC and MAC were assessed using a plastic tape measure as previously described.4,5 The cut-off values of ASMI for low muscle mass (LM) were less than 7.00 kg/m2 and less than 5.40 kg/m2, less than 7.00 kg/m2 and less than 6.00 kg/m2, and less than 6.87 kg/m2 and less than 5.46 kg/m2 for men and women, respectively, according to the criteria of the Asian Working Group for Sarcopenia (AWGS), the European Working Group on Sarcopenia in Older People 2 (EWGSOP2) and the criteria for Japanese developed by Sanada and colleagues.8,9–10 Detailed methods and baseline characteristics of patients are shown in the Supplementary Appendix. This study was conducted in strict adherence with the principles of the Declaration of Helsinki and was approved by the clinical investigation ethics committee of Sapporo Medical University Hospital (number 302–104). Patients were randomly divided into a development group (n = 98) and a cross-validation group (n = 98). In the development group, multiple regression analysis using parameters correlated with DEXA-measured ASM in single linear regression analysis indicated that gender, weight, CC and MAC were independent determinants of DEXA-measured ASM. The prediction equation for anthropometric estimation of ASM, calculated by a multiple linear regression model, was as follows: ASM (kg) = 0.214 × weight (kg) + 0.217 × CC (cm) –0.189 × MAC (cm) + 1.098 (men = 1, women = –1) + 0.576. Application of the prediction equation for ASM estimation to the cross-validation group indicated a significant correlation between predicted ASM and DEXA-measured ASM (R2 = 0.87) as shown in Figure 1(a). Bland–Altman analysis showed that there was no significant difference between predicted ASM (14.1 ± 3.8 kg) and DEXA-measured ASM (13.9 ± 3.8 kg) with a bias of 0.2 kg (95% limits of agreement (LoA) –2.7 kg to 2.9 kg, Figure 1(b)) and 1.7% (95% LoA –18.5% to 21.9%, Figure 1(c)). The correlation of DEXA-measured ASM with predicted ASM was significantly stronger than that of DEXA-measured ASM with CC alone (Pearson’s r 0.93 vs. 0.70, 95% confidence interval (CI) for the difference of the coefficient 0.15 to 0.34, z = 7.34, P < 0.01) or MAC alone (Pearson’s r 0.93 vs. 0.56, 95% CI for the difference of the coefficient 0.26 to 0.52, z = 8.40, P < 0.01). Development of a prediction equation for ASM estimation and accuracy of predicted ASMI in the diagnosis of low muscle mass. (a–c) Scatter plot (a) and Bland–Altman plots (b, c) of predicted ASM versus DEXA-measured ASM in a cross-validation group, expressed as kilograms (b) and percentages (c) of the values on the vertical axis. Open circles and crisscrosses indicate men and women, respectively. (d) Accuracy of predicted ASMI in the diagnosis of low muscle mass. LoA: limit of agreement; DEXA: dual energy X-ray absorption; ASM: appendicular skeletal muscle mass; ASMI: appendicular skeletal muscle mass index; RMSE: root mean squared error; CI: confidence interval; PPV: positive predictive value; NPV: negative predictive value; LR+: positive likelihood ratio; LR–: negative likelihood ratio; TP: true positive; TN: true negative; FP: false positive; FN: false negative; AWGS: Asian Working Group for Sarcopenia; EWGSOP2: European Working Group on Sarcopenia in Older People 2. By assigning DEXA-measured ASMI to the criteria of LM defined by AWGS and EWGSOP2, 137 (70%) and 162 (83%) patients were classified as having LM, respectively. On the other hand, application of predicted ASMI to the criteria of LM showed that 150 (77%) and 168 (86%) patients were classified as having LM. Using the diagnosis of LM by DEXA-measured ASMI as the reference standard, the diagnosis of LM by the use of the predicted ASMI had accuracies of 83% and 91%, sensitivities of 93% and 96%, and specificities of 61% and 65%, respectively (Figure 1(d)). In addition, the diagnostic accuracy of LM was 87% when the predicted ASMI was applied to the criteria for the diagnosis of LM for Japanese men and women developed by Sanada and colleagues (Figure 1(d)). These results suggested that our equation for anthropometric ASM estimation is applicable to the assessment of sarcopenia in a daily clinical setting. However, anthropometric parameters are markers of total mass in the segments measured, which is undoubtedly affected by fat mass and the levels of extra and intracellular water.5 In fact, non-negligible differences between DEXA-measured ASM and predicted ASM were observed in patients with a high tertile of percentage body fat and body mass index (BMI) of 25.0 kg/m2 or greater (Table 1), indicating that the impact of fat mass on predicted ASM by the use of CC and MAC remains. Furthermore, a significant difference between DEXA-measured ASM and predicted ASM was observed in patients with New York Heart Association (NYHA) grade III (Table 1), possibly due to differences between hydrated levels in patients with NYHA I–II and patients with NYHA III. Importantly, our prediction equation for ASM estimation may not necessarily be applicable to other races with CHF and non-CHF patients because there is race/disease-dependent variation in body composition (Supplementary Appendix Table 3). Therefore, further cross-validation analysis among subgroups of patients will be needed to demonstrate its accuracy in a larger population. Differences between DEXA-measured ASM and predicted ASM among subgroups of patients. The 1st, 2nd and 3rd tertiles of percentage body fat are as follows: 1st tertile, < 23.6% in men and < 28.3% in women; 2nd tertile, 23.6 to < 27.4% in men and 28.3 to < 35.4% in womens; 3rd tertile, ≥ 27.4% in men and ≥ 35.4% in women. ASM: appendicular skeletal muscle mass; CI: confidence interval; BMI: body mass index; DEXA: dual-energy X-ray absorptiometry; NYHA-FC: New York Heart Association functional class; LVEF: left ventricular ejection fraction; HT: hypertension; DM: diabetes mellitus; DL: dyslipidemia; eGFR: estimated glomerular filtration rate. Differences between DEXA-measured ASM and predicted ASM among subgroups of patients. The 1st, 2nd and 3rd tertiles of percentage body fat are as follows: 1st tertile, < 23.6% in men and < 28.3% in women; 2nd tertile, 23.6 to < 27.4% in men and 28.3 to < 35.4% in womens; 3rd tertile, ≥ 27.4% in men and ≥ 35.4% in women. ASM: appendicular skeletal muscle mass; CI: confidence interval; BMI: body mass index; DEXA: dual-energy X-ray absorptiometry; NYHA-FC: New York Heart Association functional class; LVEF: left ventricular ejection fraction; HT: hypertension; DM: diabetes mellitus; DL: dyslipidemia; eGFR: estimated glomerular filtration rate. Supplementary material is available at European Journal of Preventive Cardiology online. SK, TY, KO and TM designed the study. SK, TY, KO, NN, SH, KS, TI, AW, RH, TF, MKa, and HK collected the patients’ data. SK, TY, AH, MKa and TM analysed and discussed data. SK and TY performed statistical analyses. SK, TY, and TM drafted the manuscript. The author(s) would like to thank all participants and the heart failure team at Sapporo Medical University. The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by grant-in-aid for young scientists (Katano S) from the Japan Society for the Promotion of Science KAKENHI grant number JP18K17677, Tokyo, Japan.

Topics & Concepts