Mechanisms Underlying Long-term Fatigue and Exercise Intolerance Following SARS-CoV-2 Infection (COVID-19)

Persistent symptoms of fatigue and severe exercise intolerance have been observed in some patients following infection with CoV-2. Termed 'Long COVID', these symptoms may persist for months after the acute infection has resolved. The mechanisms underlying exercise intolerance in Long COVID are not fully understood and could be the result of dysfunction at any point along the transfer pathway of oxygen from atmosphere to skeletal muscle (uptake in the lungs, heart and blood vessel function and metabolism within skeletal muscle).

The primary objective of the MEXICO study was to identify underlying mechanisms that contributed to exercise intolerance in the presence of persistent covid-19 symptoms (long COVID). Initially it was planned that this would involve a comparison of people with long COVID with and without exercise intolerance. It became clear that recruitment to these two groups was unlikely to complete as while most people with long COVID had some degree of exercise intolerance, severe exercise intolerance was infrequent. Consequently the study was revised to include a reference group of healthy individuals with the primary analysis focusing on a comparison of people with and without long COVID. A comparison of people with long COVID with or without severe exercise intolerance was retained as a secondary exploratory analysis. Ethical approval for the measurements conducted in healthy control subjects was granted by the UCL research ethics Committee.

Revised sample size calculations for study were based on a two-sample t-test (equal variance for simplicity - although the planned analysis would be using a doubly robust potential outcomes method accounting for potential confounders). Sample size estimates were performed using GPower 3.1.9.7 (alpha = 0.05 (two-tailed) and 80% power) assuming that the minimum clinically important difference for the primary outcomes corresponded to a standardized effect size of 0.88 based on previous studies (typically this represents a 10-20% difference in outcome measure). The study was designed to recruit cases and controls in an approximate proportion of 2:1 to enhance the power of the planned substudy analysis. On this basis 32 and 16 participants were required in the long COVID case and healthy group respectively (48 in total).