Bidirectional Associations Between Autonomic Function, Sleep Quality, and Daily Fatigue and Energy Symptoms in Long COVID: Longitudinal Digital Health Study.
Aboagye N, R Baker M, Baker K, Del Din S • JMIR cardio • 2026
Sleep duration and REM sleep showed day-to-day within-person associations with next-day fatigue in long COVID, while HRV-RMSSD distinguished individuals with better versus worse average symptom burden rather than tracking day-to-day fluctuations.
Key Findings
Results
Sleep duration and REM sleep were negatively associated with next-day fatigue within individuals with long COVID.
Sleep duration correlated with next-day fatigue: r=-0.276 (95% CI -0.392 to -0.152; P<.001)
REM sleep correlated with next-day fatigue: r=-0.215 (95% CI -0.342 to -0.079; P=.002)
Both associations survived Bonferroni correction (α=.0031)
Analysis based on n=260 observations across 14 participants over a median of 28 days
Negative correlations indicate that longer sleep duration and more REM sleep were associated with higher (worse) next-day fatigue ratings, though the scale direction should be considered when interpreting
Results
Sleep efficiency showed a counterintuitive negative association with next-day energy levels.
Sleep efficiency correlated with next-day energy: r=-0.196 (95% CI -0.317 to -0.068; P=.003)
This association survived Bonferroni correction (α=.0031)
The authors interpret this as most plausibly reflecting chance, given the null sleep efficiency-fatigue association and the multiple-testing context
No significant association was found between sleep efficiency and next-day fatigue
Results
Nocturnal HRV-RMSSD showed no significant within-person day-to-day associations with fatigue or energy symptoms.
HRV-RMSSD to next-day fatigue: r=-0.064, P=.36 — not statistically significant
High-frequency power and low-frequency/high-frequency ratio also showed no significant within-person day-to-day associations with symptoms
HRV metrics did not survive Bonferroni correction for within-person prospective analyses
This finding indicates HRV does not track daily symptom fluctuations within individuals
Results
Individuals with chronically higher HRV-RMSSD reported consistently lower fatigue and higher energy levels, reflecting a between-person rather than within-person effect.
Mundlak (person-mean-centered) cumulative link mixed models disaggregated within-person from between-person effects
Higher between-person HRV-RMSSD was associated with lower fatigue: β=-1.260 (95% CI -2.231 to -0.290; P=.01)
Higher between-person HRV-RMSSD was associated with higher energy: β=0.539 (95% CI 0.056 to 1.022; P=.03)
This indicates that HRV distinguishes individuals with better versus worse average symptom burden rather than predicting daily fluctuations
Results
A significant Sleep × Steps interaction indicated that the protective association of longer sleep on fatigue was attenuated on more physically active days, consistent with a postexertional malaise pattern.
Interaction term β=0.091 per SD of steps per 1000; P=.02
On more active days, longer sleep duration was less strongly associated with reduced next-day fatigue
Authors interpret this as consistent with a postexertional malaise pattern characteristic of long COVID
Physical activity (step count) was included as a moderating variable in Mundlak mixed models
Results
In an exploratory reverse-direction analysis, higher evening energy was nominally associated with higher nocturnal HRV-RMSSD that same night, but this did not survive Bonferroni correction.
Repeated-measures correlation: r=0.327 (95% CI 0.080 to 0.537; P=.01), but α=.0050 threshold not met after correction
Log-linear mixed model supported the association: β=0.0069; P=.02
Evening fatigue (as opposed to energy) did not show a significant reverse-direction association with subsequent nocturnal HRV
Authors note this preliminary finding 'warrants replication' in larger samples with experimental designs
Methods
The study used continuous wearable monitoring in 14 long COVID individuals over a median of 28 days, yielding 678,057 minute-level observations.
Device used: Fitbit Inspire 3 for continuous monitoring
Symptom reporting via smartphone app (FatigueSense) using a 4-level fatigue scale and 5-level energy scale (0-100)
Monitoring range: 12-73 days per participant (median 28 days)
Two temporal linking strategies applied: prospective (prior night sleep/HRV to next-day symptoms) and reverse-direction (evening symptoms to subsequent night sleep/HRV)
Repeated-measures correlation (rmcorr) was used to account for within-person clustering
What This Means
This research suggests that in people with long COVID who experience persistent fatigue, the amount of sleep and the amount of REM (dream-stage) sleep they get on a given night are associated with how fatigued they feel the next day. Interestingly, a measure of heart rate variability (HRV) — a marker of the nervous system's ability to regulate the heart — did not predict day-to-day changes in fatigue within a person, but people who had consistently higher HRV across the study tended to report lower overall fatigue and higher energy. This means HRV may reflect a person's general health status rather than serve as a daily predictor of how they will feel.
The study also found evidence consistent with 'postexertional malaise,' a hallmark symptom of long COVID in which physical activity worsens fatigue. On days when participants were more physically active, the usual benefit of longer sleep on next-day fatigue appeared to be reduced. Additionally, a preliminary and uncorrected finding suggested that feeling more energetic in the evening was associated with higher heart rate variability that same night, hinting that daily energy levels might influence the body's nighttime recovery — though this reverse-direction finding needs confirmation in larger studies.
This research was exploratory and involved only 14 participants monitored with consumer-grade wearables and a symptom-tracking app, so the findings should be interpreted cautiously. Nonetheless, the study demonstrates the feasibility of using wearable technology for continuous, real-world monitoring of long COVID symptoms and autonomic function, and highlights the importance of distinguishing between traits that differ between people (like average HRV) and day-to-day fluctuations within a single person when studying complex conditions like long COVID.
Check Your Own Numbers
Upload your bloodwork. We'll cross-reference your results against this study and 4,700 others.
Aboagye N, R Baker M, Baker K, Del Din S. (2026). Bidirectional Associations Between Autonomic Function, Sleep Quality, and Daily Fatigue and Energy Symptoms in Long COVID: Longitudinal Digital Health Study.. JMIR cardio. https://doi.org/10.2196/99630