To the Editor:
We compliment Bakirci, et al for their article1 on the prevalence of ultrasound (US) elementary lesions at mainly lower extremity enthesis sites in 80 (50 female, 30 male) healthy adults (20–80+ yrs) and their analyses of contributory factors. We also compliment the accompanying editorial by Hánová, et al2. The core set of US enthesitis elementary lesions defined by the Outcomes in Rheumatology (OMERACT) group3 were analyzed in the following categories: (1) inflammation (hypoechogenicity and/or increased thickening of the tendon insertion, and power Doppler activity); (2) damage (enthesophytes, calcification, and erosions); and (3) total US scores, all within a 2-mm distance of the bone cortex1. The Bakirci, et al article1 followed a similar US study by Guldberg-Møller, et al4 of 64 (32 female, 32 male) healthy adults (20–59 yrs) that did not analyze the total US score or its damage and inflammation components. The Bakirci, et al article1 concluded, “These results support the effect of biomechanical forces on the enthesis, not necessarily reflecting a pathology leading to any symptoms.” The accompanying editorial by Hánová, et al2 states, “The innate immune system has been implicated as the main driver of enthesitis….” They2 also raised the consideration that the enthesis may be “hybridly responding to stimuli in health or to mechanoinflammatory stimuli in various diseases.”
In Figure 11, the relation of age (X-axis) to enthesitis scores (Y-axis) is a visual example of a segmental linear regression for inflammation and total US scores with an increase in slope at 50- to 54-year age group. The Bakirci, et al article1 indicated, “The total US scores correlated with age (r = 0.561, P < 0.001; Figure 1)” and that “Supplementary Table 6 [is] available from the authors on request.” Supplementary Table 6 indicates that age significantly correlated with all US categorical lesions at each enthesis site, except for the proximal and distal patellar tendon origin and insertion. Supplementary Table 5 was also requested and showed highly significant differences (P < 0.001) in inflammation, damage, and total scores between the younger (20–39 and 40–59 yrs) and older (60+ yrs) age groups. Assuming that a linear regression was used for age in their multivariate regression (Table 21), the results are likely to be incorrect for age and other analyzed variables.
To confirm our visual interpretation, 5-year interval datapoints were derived from the magnification of Figure 11. Separate linear slopes (95% CI) over 5-year age groups were documented to significantly differ for inflammation in the younger (20–49 yrs; 0.046, 95% CI –0.188 to 0.280) versus older (50–80+ yrs; 1.964, 95% CI 0.910–3.019) groups, as well as total US scores in the younger (0.480, 95% CI –0.012 to 0.972) versus older (3.686; 95 CI 1.953–5.148) age groups. Statistical consultation with an expert in the ADAPTIVEREG procedure for multiple adaptive regression spline (MARS) in SAS could resolve the nonlinear age regression and complex interactions with other variables in Table 21.
Our other observation in the Bakirci, et al study1 is the similarity of inflammation and damage scores in the 5-year age groups (Figure 1) and their similar correlations with other variables (Table 2). Visual examination of Figure 1 shows intertwining inflammation and damage scores over the 13 age groups. In Table 2, negative intercepts are specified for inflammation (–13.2), damage (–12.0), and total (–25.3) scores. Such data are impressive for their additive result, but such a large negative intercept (Y-axis) at zero age (X-axis) is not expected1. The B values in Table 2 for the 5 predictor variables of inflammation, damage, and total scores range from –0.14 to 5.05. The higher B values for sex and physical activity, rather than the usual –1.0 to 1.0 range, may reflect nonlinearity of age and intercorrelation of variables.
The Bakirci, et al1 study raises fundamental questions as to the relative contributions from inflammation vs adaptive (or maladaptive) biomechanical responses to physical stress in causing entheseal changes. Given biomechanical interconnections between multiple adjacent structural tissues, it may be useful to compare pathology in enthesitis to changes in chronic overuse tendon disease/tendinopathy. In tendinopathy—where tendon thickening and localized bone growth are common—biomechanical factors play a significant role in molecular, structural, and functional pathophysiologic contributions and therapy5,6,7,8. Further, while immune cell infiltrate is considered important for early tendon remodeling, impaired homeostasis between anabolic and catabolic cellular repair processes defines chronic tendinopathy pathology in response to repeated mechanical overloading (as opposed to persistent inflammation)5.
Similar to the increase in slope at age 50–54 (Figure 1) in the regression of age with inflammatory and total entheseal scores, the survey prevalence of osteoarthritis (OA) rises sharply after age 50 years9. While the ability of the joint to compensate to stress is preserved in youth, impaired resilience of aged tissues is a hallmark of OA and is connected to cellular senescence and changes in extracellular matrix mechanical properties9.
- Copyright © 2021 by the Journal of Rheumatology