Health improvements after gentle exercise in women with long COVID

An exploratory pre–post non-controlled intervention study was conducted involving women diagnosed with LC, aiming to evaluate the impact of a core-focused plank exercise regimen on body composition and perceived fatigue and to explore associations with physical activity and psychosocial factors.

Recruitment was conducted from September to November 2022 at the Faculty of Health Sciences Clinic, Rey Juan Carlos University. Individuals affiliated with the Madrid Long COVID Association (AMACOP) expressed interest by contacting a dedicated email address managed by the principal investigator, a full professor at Rey Juan Carlos University. Eligible candidates, identified through consecutive non-probability sampling, received detailed study information and documentation via email and were subsequently invited by phone to attend a screening appointment with the Rehabilitation Physician—an attending doctor at Ramón y Cajal University Hospital and part-time faculty member at Rey Juan Carlos University. All individuals contacted for participation agreed to enroll in the study, resulting in an initial sample of 17 women. Compliance was 100%, and no participants dropped out during the 12-week intervention period. Demographic characteristics at baseline are shown in Table 1.

Inclusion criteria required participants to (i) present symptoms consistent with LC for over one year; (ii) be aged 18 years or older; and (iii) have completed the full COVID-19 vaccination schedule as defined by the Spanish Ministry of Health. Exclusion criteria included: (i) severe cardiovascular disease, (ii) abdominal hernia, (iii) pregnancy, (iv) recent musculoskeletal injury or surgery (within the past year), and (v) the presence of neuromuscular disorders.

All participants provided written informed consent. The study protocol was approved by the Ethics Committee of the Hospital Universitario Fundación Alcorcón (approval code: 21/173) and the confidentiality of the participants was ensured, complying with the Declaration of Helsinki 1964 (latest update 2013).

All evaluations were conducted at the Faculty of Health Sciences, Rey Juan Carlos University (Spain). Segmental body composition was measured using bioimpedance analysis (Tanita BC-545N, Tokyo, Japan) before and after the intervention. To ensure measurement reliability, participants were instructed to avoid vigorous physical activity 24 h prior and to urinate before the assessment.

Baseline data collected included age, weight, height, duration of symptoms, admission, pneumonia, emergency, reinfections, and comorbidities. Body mass index (BMI) was calculated as weight (kg) divided by height squared (m²) and categorized according to WHO guidelines [35].

Fatigue was assessed pre- and post-intervention using the Modified Fatigue Impact Scale (MFIS), a validated 21-item instrument commonly used in LC research [28,36]. It evaluates fatigue over the previous four weeks across physical, cognitive, and psychosocial domains, with a total score ranging from 0 to 84. A score ≥ 38 is typically considered indicative of clinically relevant fatigue [37].

Quality of life was assessed by using the EuroQol-5D-5L (EQ-5D-5L) pre- and post-intervention and is a standardized instrument for measuring health-related quality of life, comprising a descriptive system with five dimensions—mobility, self-care, usual activities, pain/discomfort, and anxiety/depression [38]. This instrument allows for the classification of 3125 unique health states, where the state coded as 11,111 represents full health (the best possible state), and 55,555 indicates the worst health state [39].

Self-reported physical activity (PA) was evaluated using the International Physical Activity Questionnaire (IPAQ) at pre- and post-intervention. The IPAQ captured the number of days per week participants engaged in vigorous activity, moderate activity, and walking. Additionally, the typical duration per day spent on each activity was recorded, with activities reported in bouts of at least 10 min [40,41].

The Tampa Scale of Kinesiophobia (TSK-11) was administered at both pre- and post-intervention time points to assess fear of movement and (re)injury. The TSK-11 is a shortened version of the original 17-item instrument, excluding items 4, 8, 9, 12, 14, and 16 [42], Like the full version, the TSK-11 uses a 4-point Likert scale ranging from 1 ("strongly disagree") to 4 ("strongly agree"). The total score is calculated by summing the responses across the 11 items, yielding a possible range from 11 to 44 points. Higher scores reflect greater levels of kinesiophobia, indicating a stronger fear of movement or reinjury [42,43]. Interpretation of the TSK-11 is based on the total score, where a minimum score of 11 denotes negligible or absent kinesiophobia, and a maximum score of 44 indicates a severe fear of movement due to anticipated pain or injury [43]. This scoring framework allows for the quantification of psychological barriers to physical activity, which may influence rehabilitation outcomes and adherence to exercise-based interventions.

Participants engaged in a therapeutic exercise program designed to promote correct body alignment and optimal biomechanics, while minimizing compensatory movements (MORETA program) as previously showed [44]. Throughout the intervention period, all participants maintained their habitual dietary patterns. The program consisted of two 60 min sessions per week, conducted on non-consecutive days, totaling 24 sessions. The intervention focused on trunk stabilization through plank-based exercises, aiming to enhance muscular strength and neuromuscular control of the core musculature [45].

Each session was structured into three phases: a 10 min warm-up, a 40 min core training segment, and a 10 min cool-down. The warm-up included 3 min of specific exercises (e.g., wall sits), 1 min of anterior plank on elbows and feet, 50 sit-ups, and 1 min of sustained sit-ups. The core training phase focused on trunk musculature and incorporated exercises targeting the abdominal muscles, gluteus maximus and medius, pelvic bridge, and various plank positions (anterior plank on elbows or hands with extended arms, and side plank). The cool-down phase consisted of breathing exercises, stretching, and muscle relaxation techniques.

Participants progressively increased their workload throughout the intervention. In the initial sessions, they performed between 150 and 200 sit-ups per class, reaching 300 to 400 repetitions by the end of the program. Each week, more complex exercises were introduced, combining leg or arm movements with abdominal exercises to enhance muscle tone, proprioception, and the acquisition of new motor patterns involved in trunk stabilization.

Progressive workload was applied throughout the intervention. In the pelvic bridge exercise, participants increased their maintenance time from 5 to 9 min, while alternating with sets ranging from 10 to 25 repetitions. For gluteus medius activation, maintenance time progressed from 2 to 4 min, and repetitions increased from 20 to 60 per side, all while maintaining correct postural alignment. During plank exercises, participants extended their hold duration from 15–30 s to up to 60 s.

Gradually, more complex exercises were introduced, incorporating external loads (not exceeding 1 kg) and the use of a fit ball to increase exercise difficulty. These modifications aimed to enhance muscle tone, proprioception, and the involuntary activation of trunk-stabilizing musculature.

Depending on the session's focus, exercises targeting the trunk and gluteal muscles were alternated with routines emphasizing hip and lower limb strength. For instance, quadriceps split exercises progressed from 10 to 50 repetitions. Similarly, shoulder girdle training evolved from 10 plank push-ups (performed on knees and hands) to 50 repetitions executed on feet and hands, while maintaining all biomechanical parameters required for proper execution. In these exercises, correct activation of the serratus anterior and latissimus dorsi was essential to prevent scapular movement or detachment from the thoracic wall.

Exercise intensity was maintained at a moderately high level throughout the program, as assessed using the Perceived Exertion Rating Category Scale [46].

Given the heterogeneous clinical presentation of LC, the exercise program was fully individualized, considering each participant's symptoms and baseline characteristics at the start of the intervention. The individualized approach was guided by symptom severity and tolerance reported at baseline and during each session. For example, participants presenting higher levels of fatigue or musculoskeletal pain performed shorter plank holds and fewer repetitions, while those with better tolerance progressed to more complex exercises, such as plank combined with limb movements. Adjustments were made to ensure safety and adherence, and no participant experienced adverse effects or required discontinuation of the program.

Analyses were performed using IBM SPSS version 29.0 software (IBM Corp, Armonk, NY, USA). Baseline characteristics (age, weight, and other initial variables) were summarized using means and standard deviations for continuous variables, and frequencies and percentages for categorical variables.

Given the quasi-experimental design with pre–post measurements in a single group of women with LC, the following statistical procedures were applied: The Shapiro–Wilk test was used to assess the normality of the distribution of continuous variables.

To evaluate changes in body composition, MFIS, EQ-5D-5L, IPAQ and TSK-11 pre- and post-intervention paired t-tests were used. Cohen's D effect size (ES) was reported for paired comparisons to quantify the magnitude of the intervention effects. Given the number of paired comparisons and correlation analyses, significance levels were adjusted for multiple testing using the Benjamini–Hochberg False Discovery Rate (FDR) procedure. For transparency, original and corrected p-values are reported. Additionally, all t-tests include the t-statistics and degrees of freedom.

To assess the influence of age, weight, and baseline characteristics on the magnitude of change in the dependent variables, multiple linear regression models were constructed. Change scores (post–pre) in fat mass and fatigue served as dependent variables, while covariates were entered as predictors.

A significant level of p < 0.05 was set for all statistical tests.

The intervention produced significant reductions in body fat across all measured regions (p = 0.003), with large ES observed in the trunk (p = 0.002, ES = 1.186), right arm (p = 0.003, ES = 1.076), and left arm (p = 0.031, ES = 0.974). However, fat reduction was not evident in the lower limbs, both in the right leg (p = 0.089, ES = 0.574) and left leg (p = 0.069, ES = 0.617). Additionally, body water percentage increased significantly from 46.08% to 47.02% (p = 0.011, ES = −1.524), suggesting improved hydration status or a shift in body composition favoring lean tissue (Table 2).

In contrast, changes in muscle mass were not statistically significant. Although slight increases were observed in total body muscle (from 39.74% to 40.11%) and in specific regions such as the arms, legs, and trunk, none of these differences reached significance (all p > 0.05). These findings indicate that the intervention primarily influenced fat reduction and water balance, with minimal impact on muscle hypertrophy during the study period (Table 2).

The intervention led to a significant reduction in perceived fatigue across multiple domains. The physical sub-scale showed the most pronounced improvement, with scores decreasing from 29.71 to 21.06, corresponding to a large effect size (p = 0.001, ES = 1.199). Psychosocial fatigue also declined significantly, from 6.00 to 4.29 (p = 0.015, ES = 0.970). Overall fatigue scores dropped from 66.59 to 52.59, reflecting a substantial improvement (p = 0.012, ES = 0.936).

Although the cognitive sub-scale showed a reduction from 30.88 to 27.24, this change did not reach statistical significance (p = 0.142, ES = 0.465). These findings suggest that the intervention was particularly effective in alleviating physical and psychosocial aspects of fatigue, while cognitive fatigue may require additional or targeted strategies to achieve meaningful improvements (Table 3).

Following the intervention, changes in health-related quality of life were assessed using the EuroQoL-5D-5L instrument. Among the five dimensions evaluated, the most notable improvement was observed in the domain of pain/discomfort, which showed a substantial reduction in mean score from 2.86 to 1.79, with a large effect size (ES = 1.293) and a statistically significant result (p = 0.001). This suggests a clinically meaningful alleviation of pain symptoms post-intervention. Self-care also demonstrated a moderate improvement (p = 0.107, ES = 0.564), with scores decreasing from 1.57 to 1.21. The remaining dimensions—mobility, usual activities, and anxiety/depression—showed small effect sizes and non-significant changes (all p > 0.05), indicating limited impact of the intervention on these aspects of daily functioning and mental health.

In addition to the dimensional analysis, the total EQ-5D-5L score, likely reflecting a composite or utility-based index, decreased slightly from 24,030.86 to 22,934.21, with no statistical significance and a small effect size (p = 0.725, ES = 0.148). This suggests that while specific domains such as pain/discomfort improved markedly, the overall health status as captured by the EQ-5D-5L remained relatively stable. These findings highlight the sensitivity of the pain/discomfort dimension to the intervention and underscore the importance of domain-specific analysis when interpreting EQ-5D-5L outcomes (Table 4).

Following the intervention, several dimensions of physical activity assessed by the IPAQ showed meaningful changes. Notably, vigorous physical activity (IPAQ-1) significantly increased from 0.77 to 1.85 days per week (p = 0.048, ES = −0.698), indicating a substantial improvement in high-intensity exercise engagement. Similarly, time spent walking (IPAQ-4.2) rose markedly from 27.50 to 63.75 min per day (p = 0.041, ES = −1.415), and the number of days walking at least 10 min (IPAQ-4.1) decreased significantly from 1.80 to 1.03 (p = 0.020, ES = 1.789). Additionally, time spent sitting on weekdays (IPAQ-6) decreased significantly from 1.57 to 1.13 h (p = 0.036, ES = 1.069), suggesting a reduction in sedentary behavior.

Other IPAQ items did not show statistically significant changes. Moderate physical activity (IPAQ-2) slightly decreased, while walking frequency (IPAQ-3), time spent sitting on weekends (IPAQ-5), and total physical activity time (IPAQ-7) showed minor, non-significant variations (all p > 0.05). These results suggest that the intervention was particularly effective in increasing vigorous activity and walking behavior, while its impact on moderate activity and overall sedentary time was more limited (Table 5).

Changes in kinesiophobia were evaluated using the TSK-11. The total score showed a slight reduction from 24.29 pre-intervention to 23.29 post-intervention, with no statistically significant difference and a small effect size (p = 0.863, ES = 0.172). This suggests that the intervention had a limited impact on overall fear of movement or reinjury. Among individual items, the greatest improvements were observed in the statements "I'm afraid that I might injure myself if I exercise" and "Pain always means I have injured my body," with moderate effect sizes although neither reached statistical significance (p = 0.112, ES = 0.473 and p = 0.219, ES = 0.387, respectively). These trends may indicate a partial shift in pain-related beliefs and attitudes toward physical activity.

Conversely, some items showed minimal or even negative changes, such as "My body is telling me I have something dangerously wrong" and "My accident has put my body at risk for the rest of my life," with negligible effect sizes and non-significant p-values. Interestingly, the item "I wouldn't have this much pain if there weren't something potentially dangerous going on in my body" increased slightly post-intervention, suggesting a potential reinforcement of maladaptive beliefs in some participants. Overall, while the intervention did not significantly reduce kinesiophobia, certain cognitive aspects related to fear of movement showed promising trends that warrant further investigation in larger samples or with more targeted psychological strategies (Table 6).

To further explore the relationships between changes in body composition and psychosocial outcomes, Pearson correlation analyses were conducted between fat mass reduction and changes in MFIS, EQ-5D-5L, IPAQ, and TSK-11. No significant correlations were found between fat mass reduction and changes in MFIS (r = −0.84, p = 0.765), overall EQ-5D-5L scores (r = −0.121, p = 0.694), IPAQ scores (r = −0.241, p = 0.532) or total TSK-11 scores (r = 0.014, p = 0.961).

To identify predictors of fat mass reduction, a multiple linear regression model was constructed using changes in fat mass as dependent variant and MFIS, EQ-5D-5L, IPAQ, and TSK-11 as independent variables using age, weight, and baseline characteristics as covariables. The model was statistically significant (F = 4.12, p = 0.022), explaining 52% of the variance in fat mass change (adjusted R² = 0.52). Among the predictors, physical fatigue improvement (β = −0.41, p = 0.014) and increased walking time (β = −0.36, p = 0.032) emerged as significant contributors to fat mass reduction. These findings underscore the interplay between physical function and body composition in women with LC and highlight the relevance of targeting fatigue and mobility in rehabilitation programs.

Pearson correlation analyses revealed that changes in total muscle mass were positively associated with improvements in IPAQ walking time (r = 0.753, p < 0.001) and EQ-5D-5L (r = 0.643, p = 0.003). Participants who reported greater increases in walking time and better health status tended to show more favorable gains in muscle mass following the intervention. In contrast, no significant correlations were observed between muscle mass change and MFIS or TSK-11, suggesting that psychological factors may have a limited direct influence on muscle hypertrophy in this context.

To further explore predictors of muscle mass change, a multiple linear regression model was constructed including MFIS, EQ-5D-5L, IPAQ, and TSK-11 change scores, as well as age, baseline weight, and demographic variables such as hospitalization, pneumonia history, and reinfection status. The model explained a moderate proportion of the variance in muscle mass change, although no individual predictors reached statistical significance (all p > 0.05). These findings suggest that while physical activity and perceived health improvements may contribute to muscle mass gains, the effects are likely multifactorial and may require larger sample sizes to detect robust predictive patterns.