Preprint Long-COVID Postural Tachycardia Syndrome: A deep phenotyping study, 2025, Larsen et al.

[SNT Gatchaman]

Long-COVID Postural Tachycardia Syndrome: A deep phenotyping study
Nicholas W Larsen; Jannika V Machnik; Jordan Seliger; Ruba Shaik; Christopher H Gibbons; Paul J Utz; Maarten G Lansberg; Srikanth Muppidi; Safwan Jaradeh; Mitchell G Miglis

BACKGROUND
Postural tachycardia syndrome (POTS) has emerged as one of the most common autonomic complications of Long-COVID (LC). However, disease mechanisms remain incompletely understood.

OBJECTIVES
To evaluate the frequency and severity of autonomic dysfunction in a subset of carefully phenotyped, previously healthy patients with LC-POTS using a detailed protocol of autonomic, cerebrovascular, respiratory, blood, and tissue analyses.

METHODS
Participants in this study completed a battery of autonomic function tests, including measures of sudomotor, cardiovagal, and sympathetic adrenergic function, and head-up tilt (HUT) with transcranial Doppler measures of cerebral blood flow velocity (CBFv), end-tidal CO2 (ETCO2), cerebral and skeletal muscle near-infrared spectroscopy (NIRS) and plasma catecholamines. Skin biopsy was performed at proximal and distal sites and analyzed for intraepidermal nerve fiber density (IENFD) and phosphorylated α-synuclein (P-Syn). Results were compared to healthy controls (HC) ≥ 3 months post-COVID infection with no lasting sequelae.

RESULTS
LC-POTS participants (n=24) exhibited a greater increase in heart rate on HUT (31.1±20.3, p=0.01), and 38% exhibited elevated upright norepinephrine levels consistent with a hyperadrenergic response. CBFv did not significantly differ between LC-POTS and HC (n=10). EtCO2 and NIRS were also similar between groups. Twenty-two percent of LC-POTS and 38% of HC had decreased IENFD on skin biopsy, while 8.7% LC-POTS had dermal P-Syn aggregation on skin biopsy, compared to none of HC.

CONCLUSIONS
LC-POTS was associated with widespread autonomic dysfunction, including orthostatic tachycardia, sympathetic adrenergic hyperactivity, small fiber neuropathy, and dermal P-Syn deposition. Our findings support the concept of multiple pathophysiological mechanisms in most patients with POTS triggered by SARS-CoV-2.

Link | PDF (Preprint: MedRxiv) [Open Access]

 

[Hutan]

intraepidermal nerve fiber density (IENFD)

Twenty-two percent of LC-POTS and 38% of HC had decreased IENFD on skin biopsy

LC-POTS was associated with widespread autonomic dysfunction, including orthostatic tachycardia, sympathetic adrenergic hyperactivity, small fiber neuropathy, and dermal P-Syn deposition. Our findings support the concept of multiple pathophysiological mechanisms in most patients with POTS triggered by SARS-CoV-2.

This claiming of useful finding when it's not even supported in a separate paragraph of the abstract is not going to help people with POTS. To spell it out, a bigger percentage of healthy controls had decreased intraepidermal nerve fibre density than the POTS patients. And the healthy controls don't have symptoms.

No change in cerebral blood flow velocity is disappointing.

The authors are saying that the people selected for the study because they have a bigger heart rate increase than normal when standing showed a bigger heart rate increase in the head-up tilt. And that's not at all surprising. Even then, the change is only 31 beats per minute, which I would have thought was pretty marginal for a tilt test, which is typically thought of as more extreme than a standing test.

The skeletal and cerebral near infra-red spectroscopy didn't find anything different. That's disappointing and a bit surprising. It's probably worth a look to see what they investigated.

8.7% LC-POTS had dermal P-Syn aggregation on skin biopsy, compared to none of HC.

That looks to be the most interesting result. They say that the LC-POTS group had 24 people. 8.7% of 24 is 2.088 people, which is a strange amount of people to have something where you either have it or you don't. It's not many people, but it's worth looking into more.

Dermal phosphorylated alpha-synuclein (p-syn) aggregation refers to the build-up of a misfolded protein, phosphorylated alpha-synuclein, in skin nerve fibers. This aggregation is a hallmark of synucleinopathies, a group of neurodegenerative diseases including Parkinson's disease (PD) and multiple system atrophy (MSA.

 

[EndME]

Seems like a study that one should have a closer look at. Lots of negative results that have sometimes appeared positive elsewhere. @ME/CFS Skeptic, this seems worthy of a deeper dive.

That looks to be the most interesting result. They say that the LC-POTS group had 24 people. 8.7% of 24 is 2.088 people, which is a strange amount of people to have something where you either have it or you don't. It's not many people, but it's worth looking into more.

The different experiments have different cohort sizes (summarized in the tables in the end appendix). For the skin biopsy it's 23 patients from the LC-POTS group.

This claiming of useful finding when it's not even supported in a separate paragraph of the abstract is not going to help people with POTS. To spell it out, a bigger percentage of healthy controls had decreased intraepidermal nerve fibre density than the POTS patients. And the healthy controls don't have symptoms.

This part I find unusual as well. Seems like they are claming the presence of SFN in LC-POTS when their experiments have seemingly only showed the absence of it.

 

[EndME]

Some of the authors had previously published a small case series (8 patients) on phosphorylated α-synuclein in Long-COVID POTS, titled: A case series of cutaneous phosphorylated α-synuclein in Long-COVID POTS
(funded by Dysautonomia international where one of the authors M. G. Miglis is a board member). Would be interesting to know whether any of those patients were included in this study.

I'm not sure what a percentage of 8% would tell us? In studies for other conditions controls have a rate close to 0% whilst certain conditions (Parkinsons, dementia with Lewy bodies) seem to be close to 100%. I suppose it's possible that some LC patients have an early onset of such a condition without the usual signs yet, but are there studies to suggest that such people also have positive results on such a test?

 

[Utsikt]

 

[EndME]

Regarding increases in heart rate, which apart from the logical observation @Hutan and the authors mention is the reason for them being in the study in the first place, it doesn't seem like there is a mention of BMI or (in)activity which can also drive said results?

In this cohort one wouldn't necessarily have to expect lower activity levels in LC-POTS patients, especially since Dysautonomia international recommend exercise as treatment to this group of patients.

Edit: The authors mention group differences in BMI and that patients report lower physical activity levels due to symptoms, but these weren't quantified.

 

[Utsikt]

@EndME

 

[Yann04]

The SFN null results (or higher rate in controls) is even more striking given that the controls were well age and gender matched.

Twenty-four participants with LC-POTS (32.5 ÷ 8.9 yrs., 87.5% female) and 10 healthy controls (HC, 32.4 $ 8.0 yrs., 70% female) were included in the final analysis

Though to be fair, n=10 for HC leaves a lot of room for randomness to fuzzy up the results. (Although I’m questioning how 38% of HC could have had SFN, given there were 10 people. How do 3.8 people have SFN. I thought having SFN was binary?)

Anyways — this study might bolster the suggestion that HR increase isn’t a useful clinical marker for OI.

 

[EndME]

@EndME
View attachment 25989

Thanks, I did miss that. I suppose those differences are not too large, I guess one maybe would still want to know whether differences are driven by certain individuals (a small portion of higher BMI patients causing the differences) rather than just group differences in general.

 

[ME/CFS Skeptic]

Seems like a study that one should have a closer look at. Lots of negative results that have sometimes appeared positive elsewhere. @ME/CFS Skeptic, this seems worthy of a deeper dive.

Thanks for tagging me. Interesting that cerebral blood flow and small fiber neuropathy showed no significant differences. The sample size was really small though.

This claiming of useful finding when it's not even supported in a separate paragraph of the abstract is not going to help people with POTS

Almost as if the added the conclusions of a different paper, it doesn't match what the study actually found.

 

[EndME]

Anyways — this study might bolster the suggestion that HR increase isn’t a useful clinical marker for OI.

One would think so but on the other hand I wonder whether there is something that makes interpreting this study quite hard: All patients that are part of this study are so because they:
1. Experience a certain increase in heart rate upon standing.
2. Have symptoms consistent with POTS (but without saying what those symptoms might be and the list can be rather long).

The patients are all from the Stanford Autonomic Disorders Clinic based on POTS symptoms that were present after a Covid infection but not before (I think it's a strong part of the study that they chose patients that can be pre-Covid be described as healthy, even if it can't be proven that their ill health was caused by Covid). So they basically say the patients have "autonomic dysfunction" and POTS. Unfortunately, they don't mention any other symptoms as part of the study (although they seem to have recoded them somewhere as they sometimes mention them) so it's not really clear to me what the symptoms here are and if there are large differences in symptom profiles or not? Maybe the authors can add those somewhere since this information seems to have been recorded.

Maybe one can expect that people reporting to such a clinic and being part of such a trial are precisely those that experience problems that can be described as orthostatic intolerance, but it needn't necessarily have to be so since the list of possible POTS and Long-Covid symptoms is very long (for example they might just have an increase in heart rate and continous brain fog, which probably isn't what would be considered OI).

 

[EndME]

Regarding the seemingly only difference in the groups (phosphorylated α-synuclein) the authors report that these findings might be transient:

we performed a follow-up biopsy set on one of our P-syn positive LC-POTS participants approximately one year after the baseline biopsy, finding that the participant was no longer P-syn positive, which also correlated with a significant improvement in POTS symptoms.

 

[forestglip]

No change in cerebral blood flow velocity is disappointing.

They used transcranial doppler. Maybe extracranial doppler would be more sensitive:

Cerebral blood flow is reduced in ME/CFS during head-up tilt testing even in the absence of hypotension or tachycardia: A quantitative, controlled study using Doppler echography (van Campen et al, 2020, Clinical Neurophysiology Practice)

It is generally assumed that part of the OI symptomatology is related to cerebral hypoperfusion (Freeman et al., 2011, Sheldon et al., 2015, Shen et al., 2017). Indirect cerebral blood flow measurements such as those obtained by transcranial Doppler studies, provide inconsistent support of this hypothesis (Coverdale et al., 1985, Verbree et al., 1985, Park et al., 2017, Novak, 2018).

A limitation of transcranial Doppler is that it measures cerebral flow velocities, not cerebral blood flow. For transcranial Doppler cerebral blood flow velocity measurement to be a valid indicator of total cerebral blood flow, an important assumption is that intracranial vessel diameters would have to remain unchanged during an intervention. Changes in vessel diameters affect flow velocities (Kantos, 1992) independent of the intervention.

During head-up tilt test hypocapnia may develop (Naschitz et al., 2000, Novak, 2018). Previous studies have shown that hypocapnia can reduce intracranial vessel diameters, thereby altering the relation between flow velocity changes and hemodynamic changes (Coverdale et al., 1985, Verbree et al., 1985, Al-Khazraji et al., 2019).

Fifth, our data extend the observation from transcranial Doppler studies that hypocapnia reduces cerebral flow velocities (Immink et al., 1985, Novak, 2018). Using more direct imaging by extracranial Doppler of the internal carotid and vertebral arteries, we showed for the first time that total cerebral blood flow in ME/CFS is also reduced in the presence of hypocapnia.

Cerebral Blood Flow in Orthostatic Intolerance (Khan et al, 2025, Journal of the American Heart Association)

Extracranial CBF ultrasound is another method of measuring CBF with significantly fewer assumptions compared with TCD [transcranial doppler].52 Unlike TCD's approach of trying to measure a single cerebral artery through the cranium, extracranial CBF ultrasound measures all 4 cerebral arteries before they enter the skull, thus avoiding the difficulty of measuring through the skull.52 This allows sonographers to take a carotid duplex scan using simultaneous B‐mode and pulsed‐wave Doppler to capture flow velocities and artery diameters for all the internal carotid and vertebral arteries.61 Flow velocities are multiplied against their respective artery diameter to get a volume flow (mL/min) measurement. Then, volume flow of all 4 cerebral arteries is added together to get total CBF. This method arguably measures true CBF without requiring any of the significant assumptions needed for TCD ultrasound measurements.

In exchange for the higher‐quality CBF measurement, the extracranial CBF technique loses the continuous monitoring nature of TCD ultrasound. Extracranial CBF ultrasound takes snapshots of CBF where the sampling rate is limited to the time it takes to make duplex measurements of all 4 cerebral arteries, which typically takes 1 to 2 minutes for a trained sonographer. It also requires a nontrivial amount of postprocessing including diameter measurements, losing some of the real‐time nature. However, the sampling frequency trade‐off does support a more sensitive, specific, and theoretically sound measurement of true CBF. Extracranial ultrasound has not been studied as frequently and across as many different clinics and laboratories as TCD has in orthostatic populations. However, in the studies in which extracranial CBF ultrasound has been used, it has thus far had more consistently repeatable measurements compared with TCD [transcranial doppler], which has had problems with interlabratory repeatability.35, 46

Edit: What the thread's study said about hypocapnia being involved:

Similarly, we found that participants in both our LC-POTS and HC groups exhibited a significant degree of hypocapnia in the supine and the HUT positions. Possible mechanisms of hypocapnia include hyperventilation, undiagnosed pulmonary disease, and carotid body dysregulation. There is evidence that the carotid chemoreflex is amplified in LC, which may result in hyperventilation and reduced efficiency of breathing during exercise. 26 Hypocapnia is a known cerebral vasoconstrictor, which in turn can result in decreased CBF. 27

One study showed that both LC (POTS not specified) and non-COVID POTS exhibited a significant decrease in CBFv on HUT when compared to healthy, COVID-naïve, pre-pandemic controls. 23 In this study, the decline in CBFv in POTS participants was driven mainly by hypocapnia, while in LC, the decline in CBFv was independent of CO2. It should be emphasized that this particular study used POTS and HC groups recruited prior to the COVID pandemic, which may limit the generalizability of CBFv measurements in those with a history of SARS-CoV-2 infection, as in our study.

Another study documented a decrease in CBFv during HUT in normocapnic POTS patients, suggesting that CO2 is not the only driver of altered CBFv in POTS. 24

Our cohort is the first to our knowledge that compared these measures in both contemporaneous LC-POTS and COVID-recovered controls, perhaps accounting for the disparity between our CBFv findings and those of pre-pandemic cohorts.

While it is possible that hypocapnea contributed to the decreases in CBFv seen in our cohort, there were no differences in CO2 between LC-POTS participants and HC at baseline or on HUT. Furthermore, we did not find any significant correlations between CO2 and CBFv within either the LC-POTS group or HC.

Taken together, these findings suggest that although hypocapnia is common in LC-POTS, hypocapnia is not the only driver of CBFv changes during HUT.

The high rates of hypocapnia seen in the HC group could have also been due to incomplete recovery from asymptomatic post-viral changes.

 

[Utsikt]

The primary aim of our study was to perform deep phenotyping across cardiovascular, cerebrovascular, respiratory, and neurocutaneous domains to better understand the potential mechanisms of autonomic dysfunction in LC-POTS, with the aim of extrapolating these findings to post-infectious POTS in general. In our carefully selected cohort of previously healthy LC-POTS participants, we found evidence of widespread autonomic dysfunction spanning sudomotor, cardiovagal, and sympathetic adrenergic domains, with high rates of hyperadrenergic responses on HUT, reduced IENFD, and low but surprising rates of cutaneous P-Syn deposition. However, except for P-Syn deposition, we also found similar rates of some of these same abnormalities in our HC group, despite these participants having no pre-existing conditions or persistent symptoms after SARS-CoV-2 infection.

I don’t understand the bolded sentence. Why do they believe that the findings are abnormal? Compared to what?

 

[Kronos]

1. If 38% of your HC have "reduced" IENFD on skin biopsy, doesn't this make the definition quite useless? n=10 but still...
Does anyone know how it is in the Fibromyalgia studies with roughly 50% positive in meta reviews? Unfortunately, I can't find a positive rate for HC there.

2. +31.1±20.3 bpm at HUT (end-tilt to supine) and 42.7+-16.7 (max-tilt to supine) for LC-POTS seems...weirdly low? Here the precise definition of POTS comes into play, is it >30bpm *once* within the 10min HUT? Or did they previously fulfill POTS criteria (>30 average over HUT?) and for the study measurement a significant portion did not?

Could only skim the paper though!

If these points and other issues already mentioned here persist, it makes it hard to derive anything of value.

 

[Utsikt]

Here the precise definition of POTS comes into play, is it >30bpm *once* within the 10min HUT? Or did they previously fulfill POTS criteria (>30 average over HUT?) and for the study measurement a significant portion did not?

Especially in younger adults, the HR will vary naturally. So you probably need a sustained >30 BPM increase. Although the increase can take many minutes to appear, so I don’t think it’s based on an average.

And many will not be able to stand for 10 minutes, so their average would be lower than those that can stand for 10 minutes.

A bit like how the average pace of a 200 meter sprint is higher than for the 100 meter sprint because the acceleration phase is a smaller part of the total duration.

 

[Kronos]

Especially in younger adults, the HR will vary naturally. So you probably need a sustained >30 BPM increase. Although the increase can take many minutes to appear, so I don’t think it’s based on an average.

And many will not be able to stand for 10 minutes, so their average would be lower than those that can stand for 10 minutes.

A bit like how the average pace of a 200 meter sprint is higher than for the 100 meter sprint because the acceleration phase is a smaller part of the total duration.

In that case it should be even more weird that the negative SD of LC-POTS significantly goes into the non-POTS regime (<30bpm), right? And that's only 1 sigma.

For me it was actually quite simple at TTT: it nearly instantly went to a value of >40bpm at tilt and was stable at +-2 bpm over 10 mins (max difference, not SD). The curve looked very satisfying

---

Looking at the not-possible percentage values my trust in the study is a bit damaged.
EDIT: Only 8 of 10 HC were tested for SFN, so need to retract this (and the other related comment from above), 3/8 were positive then, i.e. 37.5%

 

[Utsikt]

Looking at those not-possible percentage values my trust in the study is a bit damaged.

There’s some data missing for certain measurements, which might explain why the percentages don’t match up with the original cohorts.

In that case it should be even more weird that the negative SD of LC-POTS significantly goes into the non-POTS regime (<30bpm), right? And that's only 1 sigma.

Good point!