The clinical manifestations of post-COVID fatigue are comparable with chronic fatigue
syndrome or myalgic encephalomyelitis (CFS/ME) whose clinical features comprise
substantial decrease in function persisting for more than 6 months, post-exertional
malaise, unrefreshing sleep, cognitive impairment or orthostatic intolerance (Blomberg,
2021; Logue, 2021). From the clinical perspective, there is high plausibility of
abnormalities of energy production or mitochondrial function as the etiology of CFS/ME
(Tomas, 2017; Filler, 2014). The similar clinical features were observed in other
infections such as EBV, Ebola, lime disease, etc., which led to coining a terminology,
post-infectious syndrome. Considering the recent study showing SARS-CoV-2 reduces
mitochondrial proteins uniquely, known as Complex One, to quiet the cell's metabolic
output (Miller, 2021), the etiological relevance of post-COVID fatigue and CFS/ME should
be scrutinized. By the way, fatigue is prevalent in cancer survivors even after achieving
the complete remission, which is described as "paralyzing" and ensues physical or mental
exhaustion that may not reversed by taking rest or sleep. This cancer related fatigue may
compound the effect of chemo brain or chemo fog which denotes cancer-related cognitive
dysfunction lingering for months or years after cancer treatment. As such, there are many
analogous clinical features shared by post-COVID fatigue and cancer fatigue. The
comparison of the metabolic changes between post-COVID fatigue and cancer fatigue may
provide hints to understand the mechanism of fatigue better.

It has been challenging technically to verify the role of mitochondria in CFS/ME. A
commercially available test, e mitochondrial energy score protocol, failed to show the
reliability and reproducibility required of a diagnostic test (Thomas, 2019), which
implies the practical limitations of in vitro test to evaluate mitochondrial function
given the intrinsic and extrinsic factors affecting the test result of target
metabolites. That said, proton Magnetic Resonance spectroscopy (MRS) can be a useful tool
which detects the major brain metabolites such as N-acetylaspartate (NAA), creatinine
(Cr), myo-inositol (MI) and lactate (Kantarci, 2003). The MRS spectra including abnormal
positive lactate peaks correlate well with other clinical markers in mitochondriopathies
(Lin, 2003). The level of organic acid (including 3-methylglutacoric acid) in urine is
associated with compromised mitochondrial energy metabolism (Ikon N, 2016), which will be
used as an adjunct index.

To date, little is understood regarding post-COVID fatigue or cancer fatigue though it is
known to affect a large proportion of patients (10-70% depending on the population). This
study aims to investigate potential mitochondrial function and metabolic changes in brain
to provide further information regarding the etiology of these changes leading to
fatigue.

This study aims:

1) to determine whether there are changes in the NAA/MI indicating perturbed
bioenergetic role in neuronal mitochondrial function in post-COVID fatigue and 2) to
prove changes of cerebral blood flow (CBF) assessed by 3D Arterial Spin Labelling
(ASL) sequence for brain in post-COVID fatigue.

This study hypothesized that Post-COVID fatigue is ensued by perturbations in metabolism
and mitochondrial function in the brain.

This study will include 30 patients (experimental group) complaining of persistent
fatigue lasting longer than 4 weeks after recovering from SARS-CoV-2 infection and the
age/gender-matched control of 30 healthy subjects (control group 1) and 30 patients
suffering from cancer-related fatigue patients longer than 4 weeks after remission
(control group 2). Both the experimental group (post-COVID fatigue) and control group 2
(cancer fatigue) will be recruited from NUH outpatient clinic. The healthy control will
be recruited from NUH employees from various work groups utilizing a recruitment flyer
and word of mouth.

Screening Visits and Procedures:

(Visit 1) The study details will be explained clearly to the prospective participants and
consent will be taken from the participants prior to commencing study procedures which
includes screening, accessing and extracting data from their medical records,
administering the study questionnaires (Chalder fatigue scale) and arranging for the
study participants to undergo MRI, 3D Arterial Spin Labelling (ASL) and 1H magnetic
resonance spectroscopy (MRS) at the NUS-Center for transitional magnetic resonance (TMR).
Eligible subjects will then proceed to the following visits.

Visit 2 which will take place within 28 days of the first visit and will involve
MRI/MRS/ASL measurements to assess metabolic and mitochondrial perturbations in
post-COVID fatigue or cancer fatigue (details below). The urine sample (10ml) will be
collected from all subjects to measure changes of organic acid level against normal range
which may reflect mitochondrial dysfunction.

Final Study Visit (3rd visit or phone call). For the participants in the post-COVID
fatigue group and the cancer fatigue group, the study team will apply the same
questionnaires including Chalder Fatigue Scale, Health Questionnaire (EQ-5D-5L) and
Hamilton Depression Rating Scale through a phone call or in person, which will coincide
with their follow up visit with their treating physician. For healthy normal subjects
without fatigue symptoms, the questionnaires will be administered via a phone call and do
not need to attend Visit 3 in person, which is set to monitor the clinical course of
post-COVID fatigue and cancer fatigue.

This is a case-control study. With a sample size of 30 per group (90 in total for 3
groups), the study will have 90% power to detect a standard difference of 0.95 and the
mean difference in the primary endpoints equal to 95% of the SD of the endpoint. As such,
the minimum sample size will be at least 27 per group. Since the two primary endpoints of
the study are NAA/MI measured by proton MRS and cerebral blood flow (CBF) assessed by 3D
Arterial Spin Labelling (ASL) sequence for brain, the significant level will be adjusted
to 0.0125 (for 2 primary endpoints and 2 controls, the significant level will be 0.05/4).

Those numerical variables measured in MRI/MRS will be compared with either healthy
control or disease control (cancer fatigue patients). 2 sample T test or Mann-Whitney U
test, whichever is more appropriate for the hypothesis testing will be used. Linear
regression can be applied to adjust for other confounders or covariates.

The Null hypothesis: the mean NAA/MI and CBF of study group equals to the mean NAA/MI and
CBF of disease control group (cancer fatigue) and healthy control group. The alternative
hypothesis: the mean NAA/MI and CBF of study group is different from the mean NAA/MI and
CBF of disease control group and healthy control group.

Type I error: 0.0125 Type II error: 0.1