COMPARISON OF CT-BASED FAT CONTENT IN THIGH MUSCLES BETWEEN KNEE OSTEARTHRITIS PATIENTS WITH AND WITHOUT NEUROPATHIC-LIKE PAIN PHENOTYPE

Abstract

INTRODUCTION

Lower muscle volume and higher intra- and inter-muscular fat content has previously been associated with structural outcomes of knee osteoarthritis (OA) and physical performance (extensor strength). Neuropathic-like pain (NP-L) was associated with less structural damage of knee OA but greater physical function impairment compared to patients without. Measurements of fat have not been studied among patients with knee OA and different pain phenotypes (NP-L versus non-NP-L).

OBJECTIVE

To compare CT-based measures of thigh muscle fat content between knee OA participants with and without an NP-L phenotype.

METHODS

We studied participants from the IMI-APPROACH study who had CT and pain measures assessed. The pain-DETECT questionnaire was used to define different NP pain categories: NP-L was defined as a score ≥19, and non-NP-L as a score ≤12. For these analyses, we focused on the group with NP-L versus non-NP-L pain. For each participant, an index knee with OA was selected based on ACR clinical criteria as applied in the IMI-APPROACH cohort. NP-L patients were matched for SF-36 and KOOS scores with non-NP-L participants. We used a manual segmentation method for bilateral thigh muscle segmentations from whole-body CTs. The axial slice corresponding to 33% distal length of the femur bone was selected. The cross-sectional area (CSA) of the thigh muscles (quadriceps, flexors, adductors, and sartorius) were calculated and summed to represent the total thigh muscle CSA, as shown in figure 1. Hounsfield unit thresholding was applied to estimate the adipose component within the muscle. CT-based markers of muscle quality included CSA of intramuscular adipose tissue (Intra-MAT), intermuscular adipose tissue (Inter-MAT), subcutaneous adipose tissue (SAT), and the total thigh muscle (Muscle). Normality was determined by Shapiro-Wilk test. We evaluated the difference in CSAs between NP-L and matched non-NP-L using the Wilcoxon rank sum test (if at least one of the distributions was non-normal), or the Student's t-test (if both distributions were normal).

RESULTS

Twenty-one participants with NP-L and 22 participants with non-NP-L phenotypes were included. Mean age was 65.1 (NP-L) versus 66.2 (non-NP-L). Among patients with NP-L, 86% were female versus 68% females for non-NP-L. Mean BMI was 31 for those with NP-L and 28.5 for those with non-NP-L. The KOOS score for NP-L was 51.3 and 52.8 for those with non-NP-L pain. As shown in Figure 2, participants with NP-L had a statistically higher Inter-MAT CSA (p-value = 0.03) and SAT CSA (p-value <0.001), when compared to patients with non-NP-L. Intra-MAT CSA was higher among those with NP-L versus non-NP-L however without reaching statistical significance. The ratios of Inter-MAT CSA / Muscle CSA and (Intra-MAT + Inter-MAT) CSA / Muscle CSA, were statistically higher among those with NP-L in comparison with non-NP (p-value = 0.03).

CONCLUSION

The Inter-MAT CSA and the ratios Inter-MAT CSA / Muscle CSA and (Intra-MAT + Inter-MAT) CSA / Muscle CSA were statistically significantly higher among those with NP-L. These results require further investigation since sex and BMI imbalance may be contributing to this observation.