Central noradrenergic deficiency in post-infectious chronic fatigue: neurobehavioral correlates, 2026, Aregawi et al

[forestglip]

Central noradrenergic deficiency in post-infectious chronic fatigue: neurobehavioral correlates

Spoiler: Authors

Aregawi, Lillian; Walitt, Brian; Sullivan, Patti; Norato, Gina; Benjamin, Rohit Ninan; Goldstein, David S

Abstract
Fatigue, brain fog, and post-exertional malaise are major features of post-infectious myalgic encephalomyelitis/chronic fatigue syndrome and post-acute sequelae of severe acute respiratory syndrome coronavirus 2. To date, no specific neurotransmitter abnormalities have been found in either condition.

We examined the possibility of central catecholaminergic involvement and clinical correlates in post-infectious myalgic encephalomyelitis/chronic fatigue syndrome and post-acute sequelae of severe acute respiratory syndrome coronavirus 2 groups compared to healthy volunteers and, as positive controls, Parkinson’s disease patients. In an observational, cross-sectional cohort study conducted at the National Institutes of Health Clinical Center we measured CSF levels of catecholamines and metabolites in the four groups and assessed correlations with neurobehavioral measures.

CSF indices of the central Norepinephrine Pathway (norepinephrine+3,4-dihydroxyphenylglycol+3-methoxy-4-hydroxyphenylglycol) and Dopamine Pathway (dopamine+3,4-dihydroxyphenylacetic acid+homovanillic acid) were measured and related to patient-recorded outcomes and physiological assessments.

Mean values for Norepinephrine Pathway activity (in pmol/mL) were lower in the post-infectious myalgic encephalomyelitis/chronic fatigue syndrome, post-acute sequelae of severe acute respiratory syndrome coronavirus 2, and Parkinson’s disease groups compared to the healthy volunteer cohort (post-infectious myalgic encephalomyelitis/chronic fatigue syndrome (-15.9, 95% confidence interval [-26.1, -5.7], P=0.0006); post-acute sequelae of severe acute respiratory syndrome coronavirus 2 (-9.62, [-17.9, -1.4], P=0.015); Parkinson’s disease (-19.4, [-27.5, -11.3], P<0.0001)). Post-acute sequelae of severe acute respiratory syndrome coronavirus 2 participants with post-exertional malaise had evidence of central noradrenergic deficiency compared to healthy volunteers (-18.3 [-31.3, -5.3], P=0.0018).

The post-infectious myalgic encephalomyelitis/chronic fatigue syndrome and post-acute sequelae of severe acute respiratory syndrome coronavirus 2 groups did not differ from the healthy group in values for the Dopamine Pathway index. Across all participants, Norepinephrine Pathway activity correlated positively with handgrip duration and general health and negatively with fatigue.

We conclude that post-infectious myalgic encephalomyelitis/chronic fatigue syndrome and post-acute sequelae of severe acute respiratory syndrome coronavirus 2 feature a specific central neurotransmitter pattern involving noradrenergic but not dopaminergic deficiency. The noradrenergic abnormality is associated with major symptoms such as post-exertional malaise.

Web | DOI | PDF | Brain Communications | Open Access

 

[SNT Gatchaman]

Presumably the n=16 PI-ME/CFS are from the Intramural cohort under second author Brian Walitt. (The given reference there is off-target it seems.)

On first reading, this seems like an interesting study and findings that does not attempt to presume causation with reasonable hypotheses (around ATP). Limitations in terms of numbers, sex imbalance, drugs etc are noted.

As with the Intramural study they were unable to usefully evaluate OI: at least in terms of orthostatic hypotension and POTS. This applied to the PASC group also. Presumably cerebral blood flow as a potential index was unavailable in both protocols.

https://clinicaltrials.gov/study/NCT04573062 and https://clinicaltrials.gov/study/NCT04564287

Within the PI-ME/CFS and PASC groups, there were too few instances of orthostatic hypotension or excessive orthostatic tachycardia to conduct meaningful calculations of correlation coefficients.

 

[rvallee]

Within the PI-ME/CFS and PASC groups, there were too few instances of orthostatic hypotension or excessive orthostatic tachycardia to conduct meaningful calculations of correlation coefficients.

People with excessive orthostatic tachycardia are extremely unlikely to find themselves into such research programs, they'd have to be sought out, and orthostatic hypotension is such a rare issue in ME/CFS that I don't think I've ever seen more than a handful of comments about it.

 

[ScoutB]

Interesting to see the ME/CFS patients in comparison to people with parkinsons (PD). Also interesting that, despite a known loss of dopamine-producing neurons, there's still a fair bit of overlap between PD and controls on the dopamine measure.

Making random connections: as this paper mentions, apparently cells make norepinephrine by packaging dopamine and the enzyme that converts dopamine into norepinephrine together into 'Large Dense-Core Vesicles' (which have a slightly different membrane than other vesicles and carry higher molecular weight cargo out from the golgi apparatus). As has been mentioned before, a couple of the DecodeME results have to do with vesicle trafficking from the golgi apparatus:

KLHL20, much like RABGAP1L and ARFGEF2 regulates vesicle trafficking between in the region between golgi and membrane

 

[SNT Gatchaman]

That sounds like a better and more specific suggestion than ATP deficit @ScoutB.

 

[Karen Kirke]

I was wondering when we’d see more about norepinephrine from this team. Interesting that they list the EEfRT as having been assessed, but did not report on it.

While working on our response to the interpretation of the EEfRT task in the NIH intramural study, I got excited about norepinephrine. Thought it would be worth sharing here.

First, here’s what Dr Madian (NIH team, was in charge of the EEfRT in the intramural study) said in the symposium (from about 2:24:35):“A lot of” the process of effort-based decision-making happens in the valuation network, and another region called the locus coeruleus is also involved - it produces norepinephrine for the rest of the brain. He points out

Importantly, the functioning of these brain regions does not appear to be under conscious control.

Importantly, we also found a strong positive correlation between levels of the neurotransmitter norepinephrine and the proportion of hard trials chosen in people with ME/CFS. This suggests a neurobiological underpinning, potentially involving the locus coeruleus and valuation network, for disrupted effort discounting in ME/CFS.

That correlation is shown in the slide as R=0.65, p=0.01, n=14.


Here’s what I wrote 19th August 2024 to @andrewkq and the rest of the team:

"This morning I’ve been thinking about norepinephrine. I feel like I’m onto something here, but it’s entirely possible that I’m barking up a particularly incorrect tree. Interested to see what you think.

They report that norepinephrine levels did not differ between groups. I note that the means are HVs: 146.2pg/ml, ME/CFS: 115.6pg/ml, with a lot of variance, and samples of n=18 and n=16 respectively.

In figs 6d and 6e, they illustrate that the higher the norepinephrine level in patients’ cerebrospinal fluid, the longer they can sustain 50% grip force (R=0.61) and the more hard tasks they chose (PHTC/effort preference) in the EEfRT (R=0.65).

I wondered about the relationship between norepinephrine and how successful people were at completing the hard tasks they chose. [Edit for S4ME: So not just whether they chose a hard task over an easy one, but whether they could actually do the hard task, which was 98 button presses in 21 seconds with the non-dominant little finger.]

I had a look at norepinephrine levels in cerebrospinal fluid of participants who did both lumbar puncture and EEfRT, again dividing the patients into those who patterned with the healthy volunteers on the EEfRT hard task (good completers) and those who successfully completed a lower proportion of hard tasks than every healthy volunteer (poor completers): [Edit: emphasis added for S4ME]

• Healthy volunteers (n=13) NE=157 pg/ml
• Good completers (n=8) NE=134 pg/ml
• Poor completers (n=6) NE=91 pg/ml

… the patients who were poor completers on the EEfRT weren’t choosing anything during the LP. But it looks like their resting norepinephrine levels separate them from people who could do the hard task.

The 8 patients who did the CPET had mean NE=128.6pg/ml, ie closer to the average of the good completers than the poor completers. Only 2 poor completers did the CPET. So while Walitt et al. suggest that CPET didn’t correlate with effort preference because strong verbal encouragement overcame it, maybe it was just that the 8 patients who did the CPET had higher norepinephrine than the average patient…”

I’m cutting it off there as I would want to hear back from @andrewkq about whether he’s comfortable with me posting the next bit.

Hope this is interesting to some.

Thread for use of EEfRT in NIH study
Thread for our response

 

[Karen Kirke]

The sequel.

I asked Andrew whether there was a correlation between

· norepinephrine x proportion of hard tasks successfully completed
· norepinephrine x grip failure (non-dominant) and norepinephrine x grip failure (dominant)

Andrew ran those analyses and confirmed the next day (20th April 2024) that:

"CSF NE does have a strong correlation with hard task completion rates for the ME subset, rs = 0.72, p = .004, so that could be used as evidence that their proposed neuropharmacological basis for effort preference is actually a basis for the motor dysfunction measured by hard task completion rates...CSF NE is correlated with both dominant and nondominant grip time to failure, stats for nondominant are r = .53, p = .04."
[posted with permission from @andrewkq]

So lower norepinephrine in the cerebrospinal fluid - rather than "effort preference" - could be the "neurobiological underpinning" of patients' difficulty with the hard task in the EEfRT and the grip test.

And differences in CSF norepinephrine levels between the subgroup of patients who did the CPET and the larger groups that did the EEfRT - rather than verbal encouragement - could explain the CPET's lack of correlation with "effort preference".

What do people who know more about neurobiology think? @SNT Gatchaman

 

[Jonathan Edwards]

I find it hard to square low noradrenaline with the claimed link to POTS, which, together with ME/CFS in general has in the past been claimed to indicate upping of sympathetic and downing of parasympathetic I think. But that was likely nonsense anyway.

 

[Karen Kirke]

I find it hard to square low noradrenaline with the claimed link to POTS, which, together with ME/CFS in general has in the past been claimed to indicate upping of sympathetic and downing of parasympathetic I think. But that was likely nonsense anyway.

I wonder if there's a difference between plasma and CSF levels of norepinephrine. I think hyperadrenergic POTS is associated with higher plasma NE levels, but I don't know about CSF.

This seems to be heading towards studies of SNRIs and other norepinephrine modulators:

Our findings have implications for potential treatments of these forms of infection-associated chronic illness, depending on whether future research confirms that central noradrenergic deficiency in these conditions is pathophysiologically significant.

Evaluating causation could be done by examining effects of manipulations of central
norepinephrine release, reuptake, or metabolism on neurobehavioral measures.

That line of inquiry doesn't fill me with enthusiasm, but we'll see.

In psychiatric diseases, this whole area sounds complicated, and treatment messy. These quotes are from https://www.psychiatrist.com/jcp/dysregulation-of-noradrenergic-activity/

today’s clinician is called to think astutely and thoughtfully about the optimum modulation of norepinephrine functioning through a wise combination of pre- and post-synaptic receptor actions

Treatments focusing on maintaining a balance of norepinephrine activity are crucial since symptoms can develop at both high and low levels of activity because of the varying receptor affinity and activation at different levels of norepinephrine activity.

Targeted treatments for conditions with underlying dysregulation of noradrenergic signaling would be able to affect both the “accelerating” and “braking” mechanisms in the neuronal system pharmacologically better to mimic the body’s homeostatic regulation of norepinephrine firing. The most commonly used agents, including norepinephrine reuptake inhibitors, are too nonspecific, preventing norepinephrine reuptake everywhere and flooding the brain with norepinephrine. Conversely, some direct norepinephrine receptor modulators like prazosin, clonidine, and guanfacine, which target only the braking or the accelerating mechanisms, may be too narrow. As such, treatment goals should focus on the ability to modulate the symptoms created both by hypernoradrenergia and hyponoradrenergia safely and appropriately.

 

[SNT Gatchaman]

What do people who know more about neurobiology think? @SNT Gatchaman

I'm well beyond my current knowledge when it comes to neurotransmitters, but we clearly need to chase down synaptic function. As I said a few days ago in another thread

We need more neuroscientists on the forum.

 

[Eleanor]

today’s clinician is called to think astutely and thoughtfully about the optimum modulation of norepinephrine functioning through a wise combination of pre- and post-synaptic receptor actions

So much better than thinking thoughtlessly

 

[Karen Kirke]

As with the Intramural study they were unable to usefully evaluate OI: at least in terms of orthostatic hypotension and POTS.

For OH they used only a 5 minute active stand. This will really only capture those with classical OH, occurring within 3 minutes of standing. I don't understand why they chose to do that.

My understanding was that the OH in ME/CFS is generally delayed. I cannot find a good study that looks at the proportions of POTS, classical OH and delayed OH in ME/CFS that doesn't skew things one way or another.

But since in the wider population referred for assessment of syncope, delayed OH is more common than classical, more common in females, and more common in the age-group most of us are in (see e.g. Torabi et al. 2020), we'd expect more delayed OH than classical OH in ME/CFS.

Why only look for the type you're more unlikely to find, especially when dealing with small samples?

This is from Torabi et al. 2020 (not an ME/CFS study):

In healthy subjects, plasma norepinephrine concentration doubles within 5 min of standing (22). Orthostatic hypotension is associated with impaired norepinephrine release (27, 28) whereas in dOH norepinephrine levels are normal or increased (4).

Consistent with previous studies, we found that orthostatic norepinephrine concentration was higher in dOH and that it increased more on HUT compared with cOH. These data would indicate dOH as a milder form of sympathetic autonomic dysfunction compared with cOH.

So perhaps there are differences in CSF norepinephrine too.

 

[EndME]

But since in the wider population referred for assessment of syncope, delayed OH is more common than classical, more common in females, and more common in the age-group most of us are in (see e.g. Torabi et al. 2020), we'd expect more delayed OH than classical OH in ME/CFS.

In the study you cite dOH (delayed OH) was defined as a SBP/DBP fall ≥20/10 mmHg occurring first after 3 min of tilt testing.

For OH they used only a 5 minute active stand.

Did they not also look at the data from the 40 minute HUT? I don't have data access anymore but orthostatic hypotension is mentioned under "negative results"-tilt table test.

 

[Karen Kirke]

In the study you cite dOH (delayed OH) was defined as a SBP/DBP fall ≥20/10 mmHg occurring first after 3 min of tilt testing.

Not sure I follow. Are you pointing out that Torabi used tilt test rather than active stand? Tilt is more sensitive than active stand because the person's ability to compensate with their legs is removed. So Aregawi et al.'s choice of active stand reduced their chances of finding anyone with hypotension. Let me know if I have missed your point!

Did they not also look at the data from the 40 minute HUT? I don't have data access anymore but orthostatic hypotension is mentioned under "negative results"-tilt table test.

In the paper this thread is about, Aregawi et al. write:

Orthostatic blood pressure changes were assessed through measurements of systolic and
diastolic blood pressure by a brachial cuff with the patient lying supine and then standing for 5
minutes. Orthostatic hypotension was defined by a decrease in systolic pressure ≥ 20 mmHg or in
diastolic pressure ≥ 10 mmHg.

The occurrence of excessive orthostatic tachycardia as is found in postural tachycardia
syndrome was examined using a motorized tilt table...
An increase in heart rate of ≥ 30 BPM or to ≥ 120 BPM at 10 minutes of tilting was considered indicative of postural tachycardia syndrome.

So according to that, Aregawi et al. defined OH by a 5 minute active stand alone, and POTS by whether it was present by 10 mins on a tilt.

Delayed OH can take 20, 30, 40 mins to show on a tilt. And according to the above, they wouldn't have counted it as OH in this study even if documented.

 

[EndME]

Let me know if I have missed your point!

I was attempting to suggest that the delayed OH data on a tilt should still exist somewhere and allegedly was analysed for a subset before the PASC study was conducted, which could have provided grounds for their current assessment. Since for the ME/CFS portion of patients this was allegedly assessed as part of the ME/CFS study with no differences between patients and controls ( at least there is a reference to such a suggestion in "Supplementary Data 22" of the ME/CFS study, but I wouldn't count on that being accurate).

I don't think we've had the full protocol for the deep phenotyping of PASC study published (with things like https://clinicaltrials.gov/study/NCT04573062?cond=NCT04573062&rank=1 not providing full details) and they also seemingly haven't published the full data anywhere, but I would think that this data should exist somewhere and should be analysable, if it hasn't already been analysed.

It looks like they've decided to publish the Long-Covid data bit by bit. I hope they also make the full data available. I remember hearing that there would be a flurry of follow-on publications on ME/CFS coming from the intramural ME/CFS study and if I remember correctly they ended up only publishing the deep phenotyping study and then 2 studies on PEM. My guess would be that PASC patients will be left equally disappointed.

 

[Karen Kirke]

I was attempting to suggest that the delayed OH data on a tilt should still exist somewhere and allegedly was analysed for a subset before the PASC study was conducted, which could have provided grounds for their current assessment.

OK, I understand. Yes, in the ME/CFS intramural study Walitt et al. 2024 wrote:

Head-up tilt table testing for up to 40 min showed no group differences in frequency of orthostatic hypotension, excessive orthostatic tachycardia, or tilt-related symptoms requiring test cessation.

In the Supplementary Information to the intramural ME/CFS study they provided more info [spaces added to make it easier to read, you can download the document from the main study page]:

Head-up tilt table testing at 70 degrees from horizontal for up to 40 minutes was performed during which finger blood pressure (BP) was monitored continuously and upper arm blood pressure measured with a cuff every four minutes.

Orthostatic BP decreases of ≥ 20 mmHg were similar for both groups (PI-ME/CFS=9/16, HV=7/17).

The frequencies of excessive orthostatic tachycardia at 10 minutes also did not differ (PI-ME/CFS=6/16, HV=3/17)2.

The occurrences of symptoms by 40 minutes did not differ between groups (PI-ME/CFS=7/16, HV=7/17).

So 56% of 16 patients and 41% in 17 controls had BP decreases of ≥ 20 mmHg at some point during the 40 minutes. A type II error seems possible. I'd want to see that confirmed in bigger samples where the study burden was not as high, so that people with (delayed) OH and POTS would not be selected out, as @rvallee alluded to above.

We don't know when those BP decreases happened. If they all happened within the first 5 minutes, then yes, I could see that they might have thought that there was no point looking further. I'd be surprised, given that it's delayed OH that gets talked about in ME/CFS. But even if they did all happen within the first 5 minutes, requiring it to happen in an active stand rather than a tilt would further reduce the likelihood of picking it up.

Editing to add: the BP decreases would also need to have been detected by the brachial cuff measurements rather than the continuous monitoring to justify the switch to the methods of Aregawi et al., where only brachial cuff was used.

And 37.5% of patients had tachycardia vs 17.6% of controls. Again, that's something that could be significant in a bigger sample.

Relating to norepinephrine, I just noticed how striking graph S4B in the Supplementary Information for the ME/CFS intramural study is. NB this is plasma, not cerebrospinal fluid.

Here's the caption [formatted for easier reading:

Figure S4: Serial plasma catecholamine levels measures during an orthostatic challenge:

A. Plasma epinephrine levels normalized to resting value on the y axis and elapsed time on the x
axis in HV (blue; n = 16 independent participants) and PI-ME/CFS (red; n = 16 independent
participants).

B. Plasma norepinephrine levels normalized to resting value on the y axis and
elapsed time on the x axis in HV (blue; n = 16 independent participants) and PI-ME/CFS (red; n =
16 independent participants).

C. Ratio of the change in Epinephrine:Norepinephrine levels on
the y axis and elapsed time on the x axis in HV (blue; n = 16 independent participants) and PIME/
CFS (red; n = 16 independent participants) volunteers. No difference in sympathoadrenal
balance was noted between the groups.

Abbreviations. Epi: epinephrine; NE: norepinephrine.
Source data are provided as a Source Data file.

As the test goes on, it looks like plasma norepinephrine levels rise a lot more relative to supine levels for at least 6 patients than for controls.

Thinking back to Torabi et al. 2020 who wrote:

In healthy subjects, plasma norepinephrine concentration doubles within 5 min of standing (22). Orthostatic hypotension is associated with impaired norepinephrine release (27, 28) whereas in dOH norepinephrine levels are normal or increased (4).

I would love someone like Julia Newton to explain to us how they would interpret those graphs.

 

[EndME]

OK, I understand. Yes, in the ME/CFS intramural study Walitt et al. 2024 wrote:


In the Supplementary Information to the intramural ME/CFS study they provided more info [spaces added to make it easier to read, you can download the document from the main study page]:


So 56% of 16 patients and 41% in 17 controls had BP decreases of ≥ 20 mmHg at some point during the 40 minutes. A type II error seems possible. I'd want to see that confirmed in bigger samples where the study burden was not as high, so that people with (delayed) OH and POTS would not be selected out, as @rvallee alluded to above.

We don't know when those BP decreases happened. If they all happened within the first 5 minutes, then yes, I could see that they might have thought that there was no point looking further. I'd be surprised, given that it's delayed OH that gets talked about in ME/CFS. But even if they did all happen within the first 5 minutes, requiring it to happen in an active stand rather than a tilt would further reduce the likelihood of picking it up.

And 37.5% of patients had tachycardia vs 17.6% of controls. Again, that's something that could be significant in a bigger sample.

Relating to norepinephrine, I just noticed how striking graph S4B in the Supplementary Information for the ME/CFS intramural study is. NB this is plasma, not cerebrospinal fluid.

Here's the caption [formatted for easier reading:


View attachment 32452
As the test goes on, it looks like plasma norepinephrine levels rise a lot more relative to supine levels for at least 6 patients than for controls.

Thinking back to Torabi et al. 2020 who wrote:


I would love someone like Julia Newton to explain to us how they would interpret those graphs.

That sounds very sensible!

 

[Karen Kirke]

In the Supplementary Information for the ME/CFS intramural study, the reference listed for the definitions of OH and POTS is:

Freeman, R. et al. Consensus statement on the definition of orthostatic hypotension, neurally mediated syncope and the postural tachycardia syndrome. Clin Auton Res 21, 916 69-72 (2011). https://doi.org:10.1007/s10286-011-0119-5

Freeman et al. 2011 define OH as:

2.1. Definition
Orthostatic hypotension is a sustained reduction of systolic blood pressure of at least 20 mm Hg or diastolic blood pressure of10 mm Hg within 3 min of standing or head-up tilt to at least 60° on a tilt table.

but go on to describe two variants - Initial and Delayed. Note the timeframe for initial OH given is much shorter than in the definition above:

Initial orthostatic hypotension is defined as a transient blood pressure decrease (40 mmHg systolic blood pressure and/or 20 mm Hg diastolic blood pressure) within 15 s of standing. This blood pressure fall, which occurs in both old and young subjects, is observed with continuous beat-to-beat blood pressure monitoring. It may occur during active standing and to a lesser degree with passive tilting.

2.6.2. Delayed orthostatic hypotension
Some patients present with symptomatic orthostatic hypotension that occurs beyond three minutes of standing. The clinical significance of delayed orthostatic hypotension is unknown. These delayed falls in blood pressure may be a mild or early form of sympathetic adrenergic failure. This disorder may be revealed in patients with suspected orthostatic hypotension by extending the period of orthostatic stress(head-up tilt or stand) beyond 3 min.

They also define neurally mediated (reflex) syncope, but do not give a timeframe or mention testing.

I find that ME/CFS doctors don't generally distinguish clearly between delayed orthostatic hypotension and neurally mediated hypotension - some use one term, others use another - but for the most part they talk about symptoms happening in pwME/CFS with prolonged standing.

 

[Hutan]

A few thoughts.

One is that levels of neurotransmitters perhaps don't necessarily correlate with how quickly they can cycle through the system, and therefore how effectively the system is working.

Another is that I've commented before how novelty or a particularly strong need to function seems to, for me at least, give a much better chance of good function. I've said it is like adrenalin can stop the feeling of fatigue and extend functioning time. Possibly it delays PEM, possibly it even prevents it. But, it's not something that can be sustained.

Another is the need for our usual task of checking the demographics of the cohorts in studies like these. i.e. male/female, ages, pre-menopausal/post-menopausal, time of the menstruation cycle, time of day of sample collection, sedentary versus more active. The latter could possibly explain the correlation between hand grip and button clicking performances with norepinephrine levels. I haven't checked yet whether any of those possibilities are likely to apply to these findings. But the samples are small and I think someone made a comment about poor cohort matching, so, it's worth a look.

 

[Hutan]

One is that levels of neurotransmitters perhaps don't necessarily correlate with how quickly they can cycle through the system, and therefore how effectively the system is working.

The authors seem to have a reply for that:

One may ask, would it not be sufficient to report on CSF norepinephrine alone as the most direct measure of central noradrenergic activity, since circulating norepinephrine does not pass through the blood-brain barrier (BBB)? Although it is correct that there is an efficient BBB for circulating catecholamines,6 this does not mean that CSF norepinephrine straightforwardly reflect “activity” in the sense of exocytotic release in response to pathway traffic. This is because most of released norepinephrine is recycled by reuptake into the nerve terminals and metabolized before the norepinephrine reaches the CSF space. For instance, if there were decreased activity of the cell membrane norepinephrine transporter (NET), CSF norepinephrine would overestimate the rate of exocytosis. Simultaneous measurement of DHPG (the main neuronal metabolite of NE) avoids this sort of misinterpretation.

So, they have measured two other molecules that they say together with NE give a better view of the functioning of the system than just the pool of norepinephrine in the CSF:

NE+DHPG+MHPG

In particular, central norepinephrine deficiency is reflected by low CSF levels of 3,4-dihydroxyphenylglycol (DHPG), the main neuronal metabolite of norepinephrine, and of 3-methoxy-4-hydroxyphenylglycol (MHPG), the main end-product of central norepinephrine metabolism

We used the sum of CSF norepinephrine+DHPG+MHPG (referred to here as the Norepinephrine Pathway) as an index of central noradrenergic function

I don't know how correct their assumptions are, but they do seem to have thought about the problem I raised, and made measurements that they believe addresses it.