Episodic exercise intolerance is a cardinal feature of the postural orthostatic tachycardia syndrome (POTS).1–3 The development of exercise intolerance in patients with POTS is commonly accompanied by symptoms, such as lightheadedness, dizziness, and shortness of breath. However, it is rare for these patients that symptoms and exercise intolerance can be traced back to orthostasis or cardiac ischemia and conduction pathophysiology.1–5 There is also no established clinical phenotype linking peak exercise heart rate (HR) with peak work rate capacity and development of exercise intolerance in POTS.
The recurrent nature of physical stress-induced symptoms coupled with patient concerns about underlying cardiac disease leads individuals to a comprehensive battery of cardiovascular function testing. Routine clinical assessments involved in testing for POTS are the passive head-up tilt test (HUTT) and exercise stress test.1–6 In most cases the absolute change (Δ) in HR (ΔHR) ≥30 bpm during HUTT is predictive of POTS.1,4,6 However, the ΔHR during HUTT has not been shown to be predictive of the development of exercise intolerance or rise in HR during intense physical exertion, such as exercise testing. To date, it also remains unclear how, or if at all, the HR response to submaximal levels of exercise stress relates with exercise intolerance in POTS. Knowing this relationship can have an impact in how submaximal HR training zones are used to guide aerobic exercise therapy prescribed for POTS symptom management.1–3
Accordingly, we sought to study submaximal HR changes during incremental exercise testing and evaluate whether an accentuated and time-sensitive chronotropic response impacts POTS and decrement in peak work rate capacity. To test this aim, we used first-order linear dynamics models to fit the entire exercise test HR response for each patient. This approach is useful because it allows for the quantification of the reaction rate for the HR response to physical stress. Secondarily, because exaggerated submaximal HRs may occur independent of POTS, but instead as a consequence of aging and coronary artery disease risk factors, such as deconditioning, obesity, sedentary lifestyle, or most likely a composite of each, we also evaluated exercise test HRs in a separate group of primary prevention obese and sedentary adults without the diagnosis of POTS and no recent history of regular exercise participation.
This was a cross-sectional and observational study primarily focusing on adults in primary prevention diagnosed with POTS who, from 2015 to 2018, were referred to the Section of Preventive Cardiology and Rehabilitation, Cleveland Clinic, Cleveland, OH, for symptom-limited incremental treadmill exercise testing as part of routine clinical care and guidance on whether they were at risk for exercise-induced cardiac ischemia and/or conduction abnormalities (Table 1).
In order for patients to have been diagnosed with POTS and included in this study, individuals must have met basic criteria outlined in guideline statements of the European Society of Cardiology, American Heart Association, American College of Cardiology, and Heart Rhythm Society.2,4,5 This included an absence of pre-existing cardiac/structural disease, ischemia, and/or conduction abnormalities that would primarily explain signs and symptoms.1,2,6 Prior to performing exercise testing, patients also must have met/exceeded thresholds for the HUTT recommended for the diagnosis of POTS, and not previously undertaken exercise training aimed at symptom management.2,4–6 Patients also must have been able to exercise on a treadmill without orthopedic limitations.4–6
Patients with POTS were compared with adult controls who were obese and self-reported sedentary, but otherwise healthy. Controls were primary prevention patients who were referred to our clinic for symptom-limited incremental treadmill exercise testing as part of standard of care for weight loss management, coronary artery disease risk factor management, and exercise prescription. Patients meeting the following inclusion criteria were considered for this study: ≥ class I obesity (body mass index [BMI] ≥30 kg·m−2), no diagnosis of POTS, no history of diagnostic testing for suspicion of POTS, able to exercise on a treadmill without orthopedic limitations, no evidence of exercise-induced cardiac ischemia and/or conduction abnormalities, not dependent on cardiac device therapy (eg, pacemaker), and self-reported sedentary lifestyle with no recent history of regular structured exercise training for >1 yr (ie, 0-1 sessions/wk; <150 total min/wk). All aspects of this study were reviewed and approved by the Cleveland Clinic Institutional Review Board.
Patients arrived at the exercise testing laboratory in the post-absorptive and euhydrated state and had not consumed caffeine for >8 hr. Once in the laboratory patients were fit with a Marquette-12SL electrocardiogram monitor (GE Healthcare) for HR and rhythm data acquisition. American College of Sports Medicine (ACSM) Certified Clinical Exercise Physiologists conducted all exercise testing according to the recommendations of the American Heart Association and the ACSM.7–10 Patients performed symptom-limited incremental exercise testing using clinically validated protocols on a motorized treadmill. Individualized protocol selection was inclusively based on patient characteristics, predicted metabolic equivalents (METs),11 face-to-face patient-clinician consult, and medical history. Patients were continually encouraged to provide a maximal effort throughout testing. Each patient exercised to volitional fatigue.
For each patient, all HR data from exercise onset-to-peak stress were fit using first-order linear dynamics models depicted as, [A]t = [A]0(1−e−kt),12,13 where [A]t is HR at any time (t), [A]0 is HR at t = baseline, and κ is a parameter signifying the HR response rate constant. The HR dynamics half-time (t1/2) parameter was calculated as,
, and can be used to describe the t required for HR to increase from [A]0 to [A]0/2 dependent on κ and initial and final absolute values of HR.
Because of what is already known for the time and physical exercise dependence of ΔHR during HUTT,1,4,6 the first-order linear dynamics model was selected as the logical and preferred approach over the 0th-order differential model in the integrated form as, [A]t = κt + [A]0 (ie, y =mx + b), for fitting the inclusive HR response during exercise testing.
Predicted METs were calculated as, 18 − (0.15 × yr) and 14.7 − (0.13 × yr) for men and women, respectively.11,14 Predicted peak HRs were calculated as, 208 − (0.7 × yr).15 Calculations involving HR reserve (HRR) were based on the Karvonen formula10,16 in standard form as: (exercise HR − resting HR)/(predicted HR − resting HR). Mean arterial pressure (MAP = 1/3 × (systolic blood pressure [SBP] – diastolic blood pressure [DBP]) + DBP) was calculated from measurements acquired via manual sphygmomanometry while either seated upright at rest or at peak exercise.
Continuous data are presented as mean ± SD. Between-group differences for continuous variables were assessed using Welch's variance-weighted 1-way ANOVA tests. χ2 tests were used to evaluate between-group differences for categorical data.
Multivariate logistic regression analyses were performed to determine which variables relating to exercise intolerance were the most important predictors of POTS when compared with controls. Relevant variables considered for inclusion in multivariate models were age, BMI, sex, β-blocker, resting HR, exercise HR dynamics, METs, and/or peak HR. Multiple exercise HR variables were not included in models to lessen the risk for collinearity and false inflation of model significance. The variance inflation factor (VIF) parameter was used to evaluate for each variable the level of collinearity among all predictors for a given logistic regression. A VIF = 1 is ideal, whereas a value ≥10 suggests strong interdependence among variables and high likelihood for type I error.
In order to test the question of which variables were the most important predictors of severity of exercise intolerance among POTS variants, additional and separate multivariate logistic regression analyses were performed using percentage achieved of predicted peak METs as the ordinal response. Patients with POTS achieving ≥85% predicted METs14 were considered exercise tolerant (EX-TL), whereas others were classified as exercise intolerant (EX-INTL). This threshold for exercise tolerance was chosen because an appreciable proportion of patients with POTS are women,2 and separately, it has been shown in a generalizable sample of women that poor prognosis is associated with an inability to exceed 85% predicted METs.14
For multivariate logistic regression comparisons between controls and EX-INTL or EX-TL, the critical probability cut point was set at 0.70 and 0.77, respectively, for determining optimal model sensitivity and specificity levels for predicting either POTS variant. For multivariate comparisons between POTS (whole group) and controls, the critical probability cut point was set at 0.85 for determining optimal model sensitivity and specificity levels for predicting POTS. Two-tailed significance was determined using an α level = .05. Statistical analyses were performed using SAS v.9.4 (SAS Institute).
Patients classified as EX-INTL were younger, higher proportion men, overweight/obese, and demonstrated increased resting HR compared with EX-TL (Table 1). However, sedentary controls (n = 30; 87% women) were older (52 ± 10 yr) and of greater body weight (100 ± 11 kg) and BMI (36 ± 3 kg·m−2) compared with either POTS variant or the group as a whole (all comparisons P < .0001). Resting SBP (125 ± 15 mm Hg), DBP (82 ± 10 mm Hg), and MAP (96 ± 10 mm Hg) were also increased in controls compared with POTS (all comparisons P < .03). By contrast, resting HR (74 ± 9 bpm) and prevalence of β-blockers (20%) for controls did not differ from POTS (all comparisons P > .05). Outside of obesity and a sedentary lifestyle, the most common coronary artery disease risk factor for controls was essential hypertension (20%) followed by hyperlipidemia (10%).
With the exception METs, on average, peak exercise measurements between EX-INTL and EX-TL did not differ (Table 2). However, HR (156 ± 18 bpm) was decreased and SBP (181 ± 23 mm Hg), DBP (85 ± 12 mm Hg), and MAP (117 ± 13 mm Hg) were increased in controls compared with POTS (all comparisons P < .01). Absolute (8.0 ± 1.6) METs or relative (97% ± 16%) METs for controls also differed from POTS, whereas percent predicted HR (controls 91% ± 9%) did not differ between groups.
For submaximal phases of exercise tests, absolute HRs at equivalent points of 50% and 70% of HRR were decreased in controls (115 ± 12 and 132 ± 14 bpm, respectively) compared with POTS (Table 2). Absolute HR (103 ± 15 bpm) at the time equivalent of HR t1/2 dynamics (159 ± 41 sec) for controls was also decreased compared with POTS (Table 2). By contrast, in Figure 1 the percentage of HRR at the time equivalent of HR t1/2 dynamics was disproportionately exaggerated in EX-INTL compared with both controls and EX-TL.
In multivariate logistic regressions with age, BMI, sex, β-blocker, METs, and resting HR as covariates, peak HR was not a significant predictor of POTS (Table 3). However, the full model was significant, and the VIF value for each variable was <2.2 (Table 3), representing low multicollinearity.
By contrast, for separate multivariate regressions, exaggerated absolute HR and percentage of HRR coinciding with HR t1/2 dynamics were significant predictors of POTS (Table 3). The odds of POTS increased as the size of each submaximal HR response increased. These associations were not due to collinearity since the VIF value for each variable in either multivariate model was <1.8. In Figure 2 (panels A and B), sensitivity and specificity levels and the area under the curve (AUC) for receiver operator characteristic (ROC) curves representing full multivariate models for predicting POTS were robust.
In multivariate comparisons between controls and either EX-TL or EX-INTL, peak HR was again not predictive of either POTS variant despite overall model significance (Table 4). However, the odds of EX-TL or EX-INTL increased alongside exaggerated submaximal HRs linked to HR t1/2 dynamics (Table 4 and Figure 2). These relationships were not strongly influenced by collinearity effects, as the VIF value for each variable in any multivariate model was <3.1. Finally, similar to above, robust sensitivity and specificity levels and AUCs for ROCs representing full multivariate models for the probability of EX-TL or EX-INTL POTS are illustrated in Figure 2 (panels C and D, respectively).
This study documents for the first time in adults diagnosed with POTS, when compared with sedentary, obese, and older adults with no history of POTS, that peak exercise HR is not predictive of POTS: (1) peak work rate capacity; (2) an exaggerated submaximal HR (absolute or percentage of HRR) typically occurring within the first-to-second stages of exercise testing is predictive of POTS; (3) unique and exaggerated submaximal exercise HR responses strongly predict patients achieving <85% predicted METs; (4) decreased absolute peak METs is predictive of POTS and achieving <85% predicted METs; and (5) although physical deconditioning, high BMI, and adult aging can contribute to exercise intolerance, these patient features do not exclusively explain the present dichotomy between POTS achieving ≥85% predicted METs and lesser performing counterparts. Collectively, these observations indicate the type of HR response relating to the size of the rise, timing, and submaximal intensity during exercise testing are important determinants of POTS, peak work rate capacity, and exercise intolerance.
A key aspect of how exercise HRs are documented in our study differing from others3,17 is the inclusive exercise test data analyses confirming the HR response is dynamic as opposed to linear (ie, y =kx + b) as patients with POTS transition across submaximal phases of exercise intensity toward peak exertion. This means the rate constant (ie, slope—the k parameter)12,13 for HR does not independently drive the chronotropic response to exercise in POTS. Thus, the following should be considered when interpreting exercise test data while also using this information to guide therapeutic recommendations, such as exercise training for symptom management: (1) peak HR cannot be reliably predicted from a straight-line equation (ie, y =kx + b), (2) an increase of 1 MET should not be assumed to approximate an increase in HR by 10 bpm,10 and (3) achieving age-predicted peak HR is not unusual for patients, whereas backward extrapolation of age-predicted peak HR (ie, without performing exercise testing) cannot be reliably used to estimate submaximal exercise HRs. By contrast, all HR data throughout exercise testing can be fit using the equation, [A]t = [A]0(1–e–kt), where the absolute HR or percentage of HRR response corresponding with the time point of the t1/2 parameter can be used to evaluate the severity of exercise intolerance and provide guidance to patients on what constitutes an exaggerated submaximal exercise HR level.
The present observations suggest a HR response caused by physical exertion that is dynamic and demonstrates a steep trajectory at the early- to mid-exercise transition during exercise testing represents exercise chronotropic instability not detectable by independently assessing peak responses. Up until now, this is something that may have been assumed given the nature of POTS, although it has never been demonstrated. By contrast, it has been reported elsewhere in adolescents with POTS during cycle ergometry that the ratio for the absolute change in HR relative to oxygen uptake occurring from baseline to peak exercise is linear and constant while not differing across patients.17 The lack of similar observations and interpretations for the present study may be attributable to basic study features found herein, such as adult age and/or treadmill exercise modality. Irrespective of the reason, these data do not endorse observations reported on the basis of adolescent patients17 or cycle ergometry are translational to the present adults.
This study also demonstrates for the first time in patients achieving <85% predicted METs, the percentage of HRR at work rates consistent with modest-to-moderate intensity rises disproportionately high. Whether the present observations for EX-INTL patients are due to the combination of heightened sympathetic drive and an early and atypically lengthened parasympathetic withdrawal window is unclear based on available data, whereas submaximal exercise performed at a disproportionately high percentage of HRR is unlikely to have been a random phenomenon simply due to deconditioning and obesity. An exaggerated percentage of HRR stood out as a significant predictor of patients with POTS achieving <85% predicted METs when compared with sedentary, obese, and older adults without the diagnosis of POTS.
A final practical implication of these data is that, in contrast to the limited clinical availability of the HUTT and expertise of electrophysiologists, this study leverages the widespread accessibility of exercise testing to illustrate how substantial information extending beyond its traditional clinical use can be extracted to test patients suspected of having POTS or optimize care for patients already diagnosed with POTS. Contemporary studies examining the efficacy of exercise training for POTS symptom management already suggest this therapy provides benefit when prescribed based on HR training zones.2,3 The utility of exercise therapy is also clear since exercise test HRs are easy to acquire and commercial HR monitoring systems are readily available. The findings of this study should aid the interpretation of HR data in the setting of exercise and POTS.
A limitation of this study is that patients did not perform cardiopulmonary exercise testing with gas exchange and ventilation data acquisition. This additional information may have provided confirmation that patients provided true maximal efforts during testing. Accordingly, between-group differences could have been attributable to submaximal effort on the part of EX-INTL patients. While this is possible, it is an unlikely scenario, given for example, there were no differences in peak HRs between POTS variants, and yet the EX-INTL group demonstrated appreciably higher submaximal HRs than EX-TL counterparts as well as controls. Symptom-limited exercise testing is also meant for patients to exercise to volitional fatigue, selected protocols have been validated and designed to provoke maximal exercise in order to induce cardiac ischemia and/or conduction abnormalities, and no test was terminated early due to clinically indicated reasons (eg, symptomatic and hypotensive response to exercise).10 Lastly, given the cross-sectional nature of this study, the logical future direction is to conduct a training study using the present approach for analyzing exercise test HRs for guiding exercise prescription aimed at symptom management.
We demonstrate peak work rate capacity and the development and severity of exercise intolerance in adults with POTS cannot be primarily and exclusively explained by effects from being overweight/obese, aging, and/or on the basis of deconditioning and a sedentary lifestyle. The size of the peak HR response also cannot be counted on as a strong indicator of peak work rate capacity and exercise intolerance unique to POTS. Alternatively, we demonstrate that a disproportionately high absolute HR or percentage of HRR within the first-to-second stages of exercise testing represents a chronotropic response phenotype that predicts POTS and identifies a subgroup of patients at a higher odds for worsened symptoms associated with physical exertion extending beyond moderate-intensity energy requirements (eg, >6 METs).
1. Wells R, Elliott AD, Mahajan R, et al Efficacy of therapies for postural tachycardia syndrome: a systematic review and meta-analysis. Mayo Clin Proc. 2018;93(8):1043–1053.
2. George SA, Bivens TB, Howden EJ, et al The international POTS
registry: evaluating the efficacy of an exercise training intervention in a community setting. Heart Rhythm. 2016;13(4):943–950.
3. Shibata S, Fu Q, Bivens TB, Hastings JL, Wang W, Levine BD. Short-term exercise training improves the cardiovascular response to exercise in the postural orthostatic tachycardia
syndrome. J Physiol. 2012;590(15):3495–3505.
4. Brignole M, Moya A, de Lange FJ, et al 2018 ESC Guidelines for the diagnosis and management of syncope. Eur Heart J. 2018;39(21):1883–1948.
5. Shen WK, Sheldon RS, Benditt DG, et al 2017 ACC/AHA/HRS Guideline for the Evaluation and Management of Patients With Syncope: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society. J Am Coll Cardiol. 2017;70(5):e39–e110.
6. Plash WB, Diedrich A, Biaggioni I, et al Diagnosing postural tachycardia syndrome: comparison of tilt testing compared with standing haemodynamics. Clin Sci (Lond). 2013;124(2):109–114.
7. Myers J, Buchanan N, Walsh D, et al Comparison of the ramp versus standard exercise protocols. J Am Coll Cardiol. 1991;17(6):1334–1342.
8. Pollock ML, Bohannon RL, Cooper KH, et al A comparative analysis of four protocols for maximal treadmill stress testing. Am Heart J. 1976;92(1):39–46.
9. Buchfuhrer MJ, Hansen JE, Robinson TE, Sue DY, Wasserman K, Whipp BJ. Optimizing the exercise protocol for cardiopulmonary assessment. J Appl Physiol Respir Environ Exerc Physiol. 1983;55(5):1558–1564.
10. Fletcher GF, Ades PA, Kligfield P, et al Exercise standards for testing and training: a scientific statement from the American Heart Association. Circulation. 2013;128(8):873–934.
11. Kim ES, Ishwaran H, Blackstone E, Lauer MS. External prognostic validations and comparisons of age- and gender-adjusted exercise capacity predictions. J Am Coll Cardiol. 2007;50(19):1867–1875.
12. Wigertz O. Dynamics of ventilation and heart rate in response to sinusoidal work load in man. J Appl Physiol. 1970;29(2):208–218.
13. Whipp BJ, Davis JA, Torres F, Wasserman K. A test to determine parameters of aerobic function during exercise. J Appl Physiol Respir Environ Exerc Physiol. 1981;50(1):217–221.
14. Gulati M, Black HR, Shaw LJ, et al The prognostic value of a nomogram for exercise capacity in women. N Engl J Med. 2005;353(5):468–475.
15. Tanaka H, Monahan KD, Seals DR. Age-predicted maximal heart rate revisited. J Am Coll Cardiol. 2001;37(1):153–156.
16. Davis JA, Convertino VA. A comparison of heart rate methods for predicting endurance training intensity. Med Sci Sports. 1975;7(4):295–298.
17. Pianosi PT, Schroeder DR, Fischer PR. Cardiac responses to exercise distinguish postural orthostatic tachycardia
syndrome variants. Physiol Rep. 2016;4(22):e13040.