Predicting post-exertional malaise in Gulf War Illness based on acute exercise responses

Abstract

Aims: Post-exertional malaise (PEM) is poorly understood in Gulf War Illness (GWI). Exercise challenges have emerged as stimuli to study PEM; however, little attention has been paid to unique cardiorespiratory and perceptual responses during exercise. This study tested whether select exercise parameters explained variability in PEM responses.

Main methods: Visual analog scale (0-100) versions of the Kansas questionnaire were used for daily symptom measurements one week before and one week after 30-min of cycling at 70% heart rate reserve in 43 Veterans with GWI and 31 Veteran controls (CON). Cardiopulmonary exercise testing (CPET) methods were used to measure oxygen (VO₂), carbon dioxide (VCO₂), ventilation (VE), heart rate, work rate, and leg muscle pain. Symptom changes and CPET parameters were compared between groups with independent samples t-tests. Linear regression (GLM) with VE/VCO₂, cumulative work, leg muscle pain, and self-reported physical function treated as independent variables and peak symptom response as the dependent variable tested whether exercise responses predicted PEM.

Key findings: Compared to CON, Veterans with GWI had greater ventilatory equivalent for oxygen (VE/VO₂), peak leg muscle pain, fatigue, and lower VCO₂, VO₂, power, and cumulative work during exercise (p < 0.05), and greater peak symptom responses (GWI = 38.90 ± 29.06, CON = 17.84 ± 28.26, g = 0.70, p < 0.01). The final GLM did not explain significant variance in PEM (Pooled R² = 0.15, Adjusted R² = 0.03, p = 0.34).

Significance: The PEM response was not related to the selected combination of cardiorespiratory and perceptual responses to exercise.

Keywords: Cardiopulmonary; Chronic; Exercise tests; Fatigue; Fatigue syndrome; Pain; Persian Gulf War.

Figure 1.. Flow chart of study procedures

Note. Visit 1 consists of consenting and screening, demographic and physical/mental health questionnaires, a blood draw, autonomic testing, and a magnetic resonance imaging (MRI) scan. Seven days later, participants return for Visit 2 and complete a second blood draw, a submaximal exercise test, and a post-exercise blood draw. Participants return to the lab on the following day for Visit 3 to complete an additional blood draw, autonomic test, and MRI scan. Symptom measurements are obtained on a daily basis via at-home questionnaires for seven days prior to and following exercise testing.

Figure 2.. Illustration of exercise challenge protocol

Note. The exercise test began with a 2-minute period of resting data collection, followed by a 5-minute gradual ascension into the prescribed 70% HRR zone, starting at 50 W. Participants completed 30 minutes of steady-state exercise, ending with a 3-minute active recovery period at 0 W. Cardiopulmonary exercise test variables were collected from three different 3-minute steady state periods during the exercise test.

Figure 3.. Median (IQR) symptom severity for Veterans with Gulf War Illness (GWI; n=43) and healthy control Veterans (CON; n = 31) pre and post-exercise

Note. Symptom severity for GWI and CON groups pre-exercise and post-exercise. The GWI group experienced larger symptom exacerbation from pre-exercise (median = 19, IQR = 6.25, 40) to post-exercise (median = 60, IQR= 40, 75) compared to the CON group (pre-exercise: median=3.33, IQR= 0, 8.5); post-exercise: median=10, IQR= 5, 40), indicating a post-exertional malaise response in the GWI group.

Figure 4.. Days post-exercise that PEM occurred in Veterans with Gulf War Illness (GWI; n=43) and healthy control Veterans (CON; n = 31)

Note. Each column corresponds to the frequency of participants whose peak symptom response to exercise occurred that day. About 60% (44/74) of peak symptom responses were observed within 72 hours post-exercise.

Figure 5.. Scatter plots and bivariate correlations between symptom responses and predictors

Note. Cumulative Work (kJ), Peak Leg Pain, VR-36 Physical Component Score (PCS), and $\stackrel{.}{\mathsf{V}}$E/$\stackrel{.}{\mathsf{V}}$CO₂ were included as independent variables in the final general linear model. Both $\stackrel{.}{\mathsf{V}}$E/$\stackrel{.}{\mathsf{V}}$CO₂ (Pearson r = 0.31, 95% CI = 0.01, 0.56, p = 0.05) and leg muscle pain (Pearson r = 0.31, 95% CI = 0.01, 0.55, p < 0.05) were significantly correlated with peak symptom responses. Cumulative work (Pearson r = 0.17, 95% CI = −0.14, 0.45, p = 0.27) and VR-36 PCS (Pearson r = −0.15, 95% CI = −0.43, 0.16, p = 0.49) were not significantly correlated with peak symptom responses.