The problems with POTS

The diagnostic criteria for postural orthostatic tachycardia syndrome (POTS) have been questioned by several research groups. Healthy people often have orthostatic tachycardia without any symptoms and most patients with orthostatic symptoms do not have tachycardia when standing upright. Are the POTS criteria invalid? This blog post takes a closer look at the evidence.

Do you spot errors or find data that supplements or contradicts some of the statements in this blog? Let us know in the comment section or contact us at mecfsskeptic@yahoo.com. Thanks in advance!

Introduction

Orthostatic intolerance (OI) is one of the most debilitating aspects of myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). OI refers to symptoms such as lightheadedness, palpitations, dizziness, and fatigue that develop or worsen in an upright position. Most ME/CFS patients experience OI and for some, it is their most disabling symptom and the reason why they are bedbound or unable to care for themselves.

Although the pathology of OI remains unclear, it is often viewed through the lens of POTS, a syndrome that emphasizes rapid heart rate increases (tachycardia) upon standing. Definitions vary but POTS usually refers to OI with a sustained heart rate increase of 30 beats per minute (bpm) or more within 10 minutes of standing. Low blood pressure in an upright position excludes people from POTS diagnosis while in individuals aged 12 to 19 years, the required increment is 40 bpm. Doctors can use a active standing test to assess POTS, but many prefer a head-up tilt (HUT) where a table is tilted to an upright position, typically at an angle between 60 and 80 degrees.

The POTS criteria consist of two main parts:

• orthostatic postural tachycardia (POT): a large increase in heart rate when changing posture from lying down to standing up.
• the symptoms that make up OI such as lightheadedness, palpitations, weakness, fainting, nausea, etc.

The POTS diagnostic criteria imply that there is a relationship between the two. While tachycardia may not directly cause OI symptoms, there is an implicit suggestion that it reflects their underlying pathology. Some believe that the increased heart rate is a compensatory mechanism for low blood volume or blood pooling in the legs. The latter could be due to nerve damage, such as small fiber neuropathy, that prevents blood vessels from constricting. Others think tachycardia reflects an overactivation of the sympathetic nervous system responsible for the ‘fight or flight’ response. The evidence supporting these theories, however, is weak. A recent review concluded that “the volume of adult POTS literature is small and the validity and reliability of the research lacks rigour.”

This blog post takes a step back from potential explanations to examine if there is a significant relationship between POT and OI in the first place. We found four problems with the validity of current POTS criteria:

• Low specificity: a heart rate increase of 30 bpm (or 40 bpm in 12–19-year-olds) is far from abnormal. Many young and healthy people without OI symptoms also have this. This suggests that POT is not pathologic and not a good marker of underlying disease processes.
• Low sensitivity: most people with OI do not have POT. Only a minority of patients with OI symptoms has tachycardia upon standing.
• Low separation: the relationship between POT and OI is weak. Symptom severity often correlates poorly with the magnitude of postural tachycardia. Increasing the threshold to a value higher than 30 bpm is therefore unlikely to solve the problem.
• Low reproducibility: POT measurements vary and are influenced by many factors including water intake, time of day, deconditioning, etc.

Where does 30 bpm come from?

Let’s start with the threshold of 30 bpm. It is not clear where this figure comes from. POTS definitions do not cite any data to justify this value.

Reference values

Medical tests usually have reference values: measurements taken in a large and representative sample of the general population. These indicate which values are abnormal. Blood test results, for example, highlight values that are outside the normal range for a sex and age group. The 97.5 percentile is often used as the upper limit*. Values higher than this only appear in 2.5 percent of the population and are therefore considered ‘abnormal’. Such values are not pathologic per se, but they are unusual enough to be highlighted. Doctors can then take a closer look and examine the patient for an underlying disease.

Dr. David Streeten

It appears, however, that the 30 bpm criterium for POT was not correctly validated with reference values. The most likely source for it is a study performed almost 40 years ago. The researcher leading the project was Dr. David Streeten, a pioneer and authority in the field of autonomic disorders. He published the data in his 1987 book “Orthostatic Disorders of the Circulation”. In 92 healthy people aged 17 to 61, the 97.5 percentile was a heart rate increase upon standing of 27 bpm. Therefore, Streeten suggested that an increase of 28 bpm or higher should be considered abnormal.

There are, however, big problems with this study. This was not a representative sample of the general population but “mainly physicians, nurses, secretaries, and laboratory technicians” as Streeted admitted.The standing test only took 3 minutes rather than the 10 minutes used by most POTS definitions. This is important because the heart rate often continues to rise after 3 minutes of standing. If Streeten and colleagues had taken this test for 10 minutes, the heart rates, and thus the normative values, would likely have been substantially higher. It is also common practice to rest in a supine position for 15-20 minutes before standing up to get an accurate measurement of the heart rate value at rest. In the study of Streeten, participants were recumbent for only 3 to 4 minutes which resulted in a relatively high heart rate of 80 bpm at baseline.  

Philip Low

The other likely source for the 30-bpm threshold is Dr. Philip Low from the Mayo Clinic, also a big name in the field. In a seminal 1993 paper, Low and his colleague Ronald Schondorf published the first definition of POTS. They used a heart rate threshold two standard deviations above the mean of an age-matched control group. This resulted in a value of  31 bpm for women and 37.9 bpm for men. This control group, however, only consisted of 20 men and 21 women, a sample size that is too small to provide reliable reference values. Participants in the study were also tilted upright for only 2-3 minutes (rather than 10 minutes).

A couple of years later, in 1997, Low and colleagues published a bigger study with normative values of 557 people aged 10 to 83. Unfortunately, they only published the heart rate increases after just one minute of tilt testing. For participants aged 10 to 29 years (the age group where POTS is often diagnosed) the 97.5 percentile was 40 bpm for men and 34 bpm for women. For participants between 30 and 39 years, the values for men and women were 38 bpm and 32 bpm respectively. Even though Low subsequently used a POT threshold of 30 bpm, his data indicated this was problematic in the age group of young adults, especially given that it eventually became used for 10 (rather than 1 or 3) minutes of standing upright.

Healthy people also have orthostatic tachycardia

Most researchers have adopted the 30-bpm threshold without questioning it. Only a couple took the effort to check if it was appropriate. They almost unanimously found that the threshold was far too low.

The most striking results came from the Canadian research team of Kurt Kimpinski. They used posters and newspaper advertisements across Ontario to collect reference values from the general population. 252 people participated and took a 5-minute tilt test. As in Low’s dataset, the heart rate differed strongly by age: younger adults had far greater increases upon standing than older people. In the age group of 18–29 years, the median heart rate increase was 33.7 bpm. That means more than half of these healthy participants had a heart rate increase above the 30-bpm threshold! The 97.5 percentile was 52.7 bpm. Even in the older group between 30 and 60 years of age, the 97.5 percentile was 49.4 bpm, much higher than the POT-criterium requires. The authors concluded that “among young healthy adults the current HR [heart rate] increment of 30 bpm is a normal response.”

In an earlier study from 2013, the Canadian team had found similar results. The median heart rate increase was 28 bpm in 120 participants aged 14 to 76 years. This again suggests that almost half of the participants would have exceeded the POT threshold for adults.

Interestingly, Kimpinski and colleagues also followed up on healthy people with a heart rate increase above 30 bpm for about a year. One might suspect that their tachycardia is a precursor to OI and POTS, but this was not the case. They were still healthy, and their tachycardia was benign. “These data further argue for the re-evaluation of the heart rate criteria for diagnosing POTS in young populations” the authors write.

Kimpinski’s results are supported by a German study from the 1990s that tested 137 healthy volunteers between 18 and 85 years. The mean heart rate increase after 2 minutes of active standing was 29 bpm. In the age group of 18- to 40-year-olds, the mean was 35 bpm. This once again suggests that it is common for healthy adults to exceed the POT threshold. The increases after tilt testing were (surprisingly) lower but with a mean of 20.98 and a standard deviation of 8.99, these data also suggest that a large proportion had increases higher than 30 bpm. The authors concluded: “in comparison with the previously published diagnostic limits for POTS, the reference values based on our population appeared to be […]  higher regarding the differences to baseline.”

Others found similar results in children and adolescents. These studies are the reason why the heart rate threshold for this age group was increased from 30 to 40 bpm. The studies themselves, however, indicate that this is still too low. Skinner and colleagues, for example, examined 307 public high school students aged 15 to 17. The heart rate response was measured after 5 minutes of standing. The 97.5 percentile was 48 bpm, substantially higher than the threshold of 40 bpm.

The research team of Low studied 106 normal children and adolescents from Minnesota between 8 and 19 years. They all received a tilt table test at a 70-degree angle for 5 minutes. Here the 97.5 percentile was 52.7 bpm. 5% of healthy participants had a heart rate increase of 42.9 or higher and 10% had an increase of 39 or higher. These results indicate that many healthy adolescents have HR increases above the POT threshold of 40 bpm.

We also found an older study from 1992 that tested 100 adolescents, aged 12 to 19 years, at a military clinic in Hawaii. The mean orthostatic heart rate change after 2 minutes upright was 21.5 bpm with a standard deviation of 10.6. The authors concluded that “a healthy adolescent may have an orthostatic heart rate increase of 40 to 50 beats per minute […] These changes in normal individuals must be considered when interpreting the significance of a tilt test result.”

Notice that all the studies mentioned above used a standing or tilt test that lasted 5 minutes or less. We have found no studies that justified the 30-bpm threshold in adults or the 40-bpm threshold in adolescents after 10 minutes of standing or tilt testing. There was one study that examined participants for 40 minutes using a 60-degree tilt and showed low mean heart rates at the 10-minute mark. Unfortunately, the data is only visible on a graph; the exact values and spread are not reported. The authors report that the 97.5 quartile for their protocol was 45 bpm for the group under 35 years and 40 bpm for those between 35 and 69 years.

Overall, the current evidence indicates that the orthostatic heart rate increases used in POTS criteria are not abnormal and are also seen in many healthy people. The ME/CFS studies that also tested healthy controls (see table below) support this view as approximately 10-15% of the controls met the POT threshold.

Most OI patients do not have POT

Studies also indicate that most patients with OI do not have POTS. One research group, for example, found that only 40 (16%) out of 246 patients seen at their clinic with symptoms of OI, met the criteria for POTS. A South Korean study examined 464 patients with orthostatic symptoms. Only 87 (19%) were found to have POTS. While another 28% had orthostatic hypotension: an abnormal drop in blood pressure, the majority of OI patients had normal responses during head-up tilt testing. Other studies have found similar results (examples here, here, and here).  

One can see it in the ME/CFS literature as well. While 67–92% of ME/CFS patients report OI symptoms, only a minority meets the POTS criteria. Pooling the results of ME/CFS studies that tested for POTS (see table below), we found a prevalence of approximately 21%. In other words, the majority of ME/CFS patients with OI do not have POTS.

We see a similar picture in Long Covid research. The LOCOMOTION consortium found that only 10% of Long Covid patients with OI symptoms met criteria for POTS. Other studies concluded that “the most commonly encountered clinical scenario was symptoms of orthostatic intolerance without demonstrable orthostatic tachycardia or orthostatic hypotension” or “most long-COVID patients presenting to our laboratory with Ol had no significant HUTT [Head Up Tilt Testing] abnormalities”. More examples can be found here and here.

TABLE: POTS prevalence in ME/CFS**

Study	ME/CFS patients	Controls subjects
Lamanca 1999	9/39 23%	3/31 10%
Schondorf 1999	9/75 12%
Stewart 1999	16/23 70%
Rowe 2001	46/171 27%
Timmers 2002	4/36 11%	3/36 8%
Razumovsky 2003	9/26 35%	7/23 30%
Winkler 2004	2/22 9%
Jones 2005	3/10 30%	4/25 16%
Galland 2008	7/26 27%	0/26 0%
Hoad 2008	16/59 27%	5/52 10%
Hollingsworth 2010	20/64 31%	4/64 6.3%
Stewart 2012	25/25 100%
Lewis 2013	24/179 13%
Reynolds 2014	33/306 11%
Roerink 2016	35/419 8%
Van Campen 2018	155/627 25%
Miwa 2018	13/44 30%
Garner 2019	6/39 15%	1/25 4%
Van Campen 2020a	49/128 38%
Van Campen 2020b	120/429 28%
Van Campen 2021	26/60 43%
Lee 2020	60/150 40%	25/75 33%
Miwa 2020	16/72 22%
Azue 2022	15/50 30%
Natelson 2022	10/63 16%
Vernon 2022	12/26 46%	5/20 20%
Van Campen 2023	13/41 32%
Ryabkova 2023	12/24 50%	1/9 11%
Domingo 2024	4/31 13%	1/31 3.2%
Walitt 2024	6/16 38%	3/17 18%
TOTAL	775/3280 23%	62/434 14%

A weak correlation between tachycardia and OI

Our third point was summarized by a 2019 NIH Expert Consensus: “Symptom severity correlates poorly with the magnitude of postural tachycardia”. Although there appears to be a relationship between POT and OI, it is surprisingly weak effect.

In the general population

You can see it in different settings. First, there are studies of the general population. Kimpinski and colleagues report that ‘there was no significant correlation between […] heart rate increment on HUT [Head-Up Tilt] and symptoms of orthostatic intolerance.” The estimated correlation was only 0.15 [-0.05, 0.33] which is generally considered a weak effect. It suggests that only about 2% of the variance in OI symptoms is explained by heart rate increases on standing. Other studies found similar results and concluded: “one cannot predict which seemingly healthy adolescents will have high heart rate changes based on orthostatic symptoms.”

Chronic OI

Other studies have looked at patients with more severe OI symptoms and divided these into two groups: those that met the heart rate threshold of POTS and those that didn’t. Remarkably few differences were found between the two groups. The research team of Dr. Jeffrey Boris, a pediatric POTS expert, reported that “there are […] no clinical differences between patients as a function of HR increase during standing.”  Of 28 recorded OI symptoms measured in Boris’ study, only insomnia showed a significant difference as it was more common in the group with lower heart rates. In a similar study design, Parsaik and colleagues divided OI patients into a group with POTS and a group without and concluded: “clinical presentation, autonomic parameters, laboratory findings, and degree of deconditioning were overall similar between the two groups.” One study even found that after 10 minutes of head-up tilt, the 22 patients with OI symptoms had a lower heart rate increase than the 106 healthy controls in their study. Such results indicate a weak relationship between heart rate increases and chronic OI symptoms.

In ME/CFS and Long Covid

Some studies in Long Covid and ME/CFS patients have also failed to find a relationship between OI and POT. The research team of James Baraniuk at Georgetown University found that “the magnitudes of OI symptoms were not different between POTS and non-POTS subjects in either the CFS or control groups.” The authors conclude that “orthostatic tachycardia did not account for OI symptoms in CFS”. Another study that measured acute OI symptoms in 72 ME/CFS patients, reported that “POT was not associated with OI to an appreciable extent.” A long Covid study reported that “there was no correlation between the raw dizziness or palpitation severity scores and degree of haemodynamic abnormality (heart rate increase or blood pressure decrease).”  While other studies did find more acute OI symptoms in patients with POTS (examples here and here), the number of studies that failed to find a relationship suggests that it’s not as strong as one might suspect.

Moving the threshold

Because of the weak relationship between POT and OI, moving the threshold above 30 bpm might not solve the problem. One study found that using the current 30 bpm norm, 60% of the control subjects would have been diagnosed with orthostatic tachycardia (false positives). But moving the threshold up did not help much. The optimal value that maximized sensitivity and specificity was 38 bpm. But even with this higher threshold, 27% of the healthy controls would still exceed it. Conversely, 20% of patients who had POTS would no longer meet this higher threshold. The authors argue that “diagnosis of POTS should consider orthostatic intolerance criteria and not be based solely on orthostatic tachycardia regardless of test used.”

The reproducibility of standing tests

The fourth problem is that heart rate increases after standing are influenced by many factors including sleep, salt and fluid intake, and deconditioning. One study, for example, found that drinking half a liter of water decreased orthostatic tachycardia from 123 bpm to 108 bpm. Some POTS definitions argue that the diagnosis should not be made if a patient is dehydrated or severely deconditioned but that only provides guidance for the most extreme cases.

If a tilt test was repeated several times in the same individual, would the results be the same? Unfortunately, very few studies have tried to answer this question. One German study repeated the tilt table test 10 times in 40 people and found that “all subjects showed considerable variations in heart rate response”. An ME/CFS study tested patients on a good and bad day. Of the 7 patients who had POTS on a good day, 4 (55%) did not have it on a bad day. Of the 17 patients who did not have POTS on a good day, 8 (47%) had it on a bad day. Another ME/CFS group repeated a standing test when patients revisited their clinic 8 to 15 months later. Of 9 patients with POTS at the first visit, 7 (77%) patients did not have POTS at the second visit. Of 23 patients without POTS at the first visit, 3 (13%) had POTS at a later visit.

Some researchers argue that a POTS diagnosis should not be withdrawn if the heart rate decreases below 30 bpm in a new measurement. That’s a reasonable suggestion, but it might also lead to more false positives if patients are tested multiple times. Others have argued that a high heart rate increase on a single test should not be a concern. They make the analogy with blood pressure measurements which can also fluctuate strongly over time or during a day. The authors argue: “there is no reason to assume that a solitary or any tilt-test is valid or represents the “gold standard” to secure a diagnosis – or not. […] it is important to evaluate at least several orthostatic heart rate measurements over time and at various times of the day; it may even be necessary to repeat testing on the same day.”

Lastly, there is a smaller issue that needs to be mentioned because it has led to a lot of confusion among researchers and patients. Many POTS patients do not have OI symptoms during a standing test. A patient is not required to have these to receive a POTS diagnosis. Having frequent and chronic OI symptoms is sufficient, one does not need to experience them during the test. After all, a 10-minute test or doctor’s visit only provides a snapshot of your condition, and symptoms may fluctuate or change day by day. What is surprising, however, is how many POTS patients do not experience acute symptoms during the standing test. It is not a negligible minority. In one large study, half of the POTS patients were asymptomatic during a 10-minute tilt table test. Other studies reported roughly similar results (examples here and here). This also raises questions about the relevance of tilt table measurements to the chronic OI that patients experience.

The epidemic of tachycardia

“In 1996, a colleague referring a patient with orthostatic intolerance to me asked rhetorically ‘Where is this epidemic of orthostatic tachycardia coming from?’” This quote is taken from a 1999 editorial by Dr. David Robertson, one of the most influential POTS researchers who sadly passed away earlier this year. Robertson and his colleague assumed that the many new cases of orthostatic tachycardia came from better recognition and testing. But there was likely another big factor: the normative values that were used were too low and misleading. Normal heart rates were incorrectly seen as abnormal.

Where did things go wrong? The data from Streeten was likely not from a representative sample. And just like Low, he used a much shorter standing test than the 10 minutes that POTS definitions use. The most important error, however, might have been the lack of adjustment for age. When defining POTS, both Streeten and Low underestimated the influence that age has on reference values. Adolescents and young adults – which not surprisingly make up the majority of POTS cases – have heart rate increases upon standing that are much higher than initially assumed.

Other researchers have mostly been concerned with avoiding false negatives. They have argued, for example, that a 5-minute test is inappropriate as it might miss POTS patients whose heart rate will continue to rise after the 5-minute mark. Others highlighted that an active standing test is less sensitive than a passive head-up tilt test because the heart rate increase is usually lower with the former (because of the muscle activity needed to stand up). In reality, the least sensitive method was probably the most accurate one. Based on the reference values discussed above, a 10-minute tilt test will simply overdiagnose and result in more false positives.

Moving forward

So, what does this all mean? Are the POTS criteria invalid? Should they no longer be used? We wouldn’t go that far but it is clear there are big concerns that need to be addressed, starting with more epidemiological studies and reference data. It is striking that in the 30 years since Low and Schondorf published the first definition of POTS, not a single prevalence study has been conducted. Most of the POTS literature consists of single-center autonomic clinics that performed several tests or experiments on their patients. These studies suffer from selection bias and have limited scientific value.

POTS diagnoses have also evolved to encompass evermore symptoms that have little relationship to tachycardia and OI. As one research group explained: “the original unified and meaningful initial POTS description of a select group of patients with orthostatic intolerance and sinus tachycardia has evolved to encompass a multitude of patients with a constellation of complaints. In many instances the presenting symptoms are nonspecific in nature, encompass multiple body systems and often are not even associated with orthostatic intolerance.”

We suspect that POTS became a popular diagnosis for syndromes and unexplained symptoms (that include OI) because it offered an objective marker and a plausible explanation. But this may not be the most useful way to approach these problems. The current evidence suggests that we need to move away from viewing OI in terms of orthostatic tachycardia as the relation between the two is weak. By focusing on POTS, researchers have overlooked the majority of patients with (unexplained) OI.

In reviews and online discussions OI, POTS, and dysautonomia are used almost interchangeably even though the relationship between them is muddled. Most patients with POTS have only mild autonomic dysfunction and there is a lack of correlation between OI symptoms and objective measures of autonomic dysfunction.

Perhaps it is time to look for other explanations and measurements. One interesting avenue might be cerebral blood flow, which seems to capture a larger proportion of the OI population than tachycardia. One large ME/CFS study found that OI symptoms are correlated (R = 0.4) to cerebral blood flow reduction and that most patients (also those without POTS) show an abnormal cerebral flow reduction compared to healthy controls.

Notes

Many thanks to Nightson and SNT Gatchaman at the S4ME forum for helping me find some of the papers discussed above.

* The exact calculation is more complicated as one needs to correct for the fact that the mean and standard deviation of the sample will deviate from the true population mean.

** We followed the authors’ definition of POTS as closely as possible even though many of these are problematic. Some used the NASA lean test rather than active standing or tilt testing, or a test duration shorter or longer than 10 minutes, some did not exclude patients with orthostatic hypotension, and some used acute symptoms during the test as a requirement for POTS instead of chronic OI symptoms, etc. We also suspect that the Dutch studies by Van Campen & Visser overlap with the some participants’ results appearing in multiple papers. However, excluding their papers and retaining only van Campen 2018, lowered the total percentage by only a small amount: from 23.6% to 21.6%. Lastly, the studies listed here may overestimate the POTS prevalence in ME/CFS as most data comes from single-center autonomic clinics that suffer from selection bias.