Autonomic debate: symptoms are an optimal target for orthostatic hypotension screening and management. An argument in opposition

Introduction

Orthostatic hypotension (OH), a pathological fall in blood pressure (BP) when upright, is a hallmark condition of autonomic failure. Classical OH, arguably the most serious of all OH variants, requires a sustained fall of ≥ 20 mmHg in systolic blood pressure (SBP) and/or ≥ 10 mmHg in diastolic blood pressure (DBP) within 3 min of upright posture [1]. OH is highly prevalent in older individuals and those with significant autonomic risk factors, generally occurring in 20–30% of older adults [2, 3], over 25% of diabetics [4], and 33–80% of those with “central” α-synucleinopathies (Parkinson’s, dementia with Lewy bodies, multiple system atrophy); of note, all individuals with pure autonomic failure have OH [5, 6].

OH is best known for several potential acute consequences, namely disabling symptoms, syncope, falls, and injury. Recurrent “typical” orthostatic intolerance symptoms (e.g., light-headedness, dizziness, dimmed vision, feeling faint) commonly trigger testing of orthostatic vitals that leads to a diagnosis of OH. These symptoms also form the basis of the most commonly used questionnaires probing cardiovascular autonomic function [7, 8]. Similarly, both clinical treatment of OH and inclusion in clinical trials to treat OH rely heavily, often solely, on subjective patientreported symptoms [9, 10]. Much of this reliance stems from the ease of obtaining symptom report and federal regulatory agency emphasis on patient-reported outcomes in drug trials.

This viewpoint will argue against that our reliance on OH symptoms is a suboptimal. Specifically, I will argue that reliance on symptom report is based on three specious assumptions, rebutted below, and that it leaves patients with OH of all etiologies vulnerable to worse outcomes and potentially modifiable risks.

Assumption 1: patients are reliable reporters and physicians are reliable recorders of OH

OH is commonly asymptomatic. Four studies, all from autonomic centers, have specifically examined the prevalence of symptomatic vs asymptomatic (and/or atypically symptomatic) OH during autonomic test procedures [11–14]. These studies, which collectively included over 400 individuals with OH of various severities and etiologies, collectively demonstrated that approximately 55% of all individuals (ranging from 33% to 74%) had “hypotension unawareness,” whereby they denied symptoms despite experiencing OH. Several other studies of community dwelling older adults and individuals in care facilities report similarly high percentages of asymptomatic OH [2, 3, 15–22]. Considering how common asymptomatic OH is, it seems likely that the prevalence of OH may be far higher than is currently estimated.

No study has examined the neural underpinnings of OH symptoms. The inability to reliably perceive OH may therefore be due to several factors. For very mild OH, or in the case of OH superimposed on supine hypertension, there may be sufficient cerebral autoregulatory reserve to appropriately buffer changes in brain blood flow despite systemic falls in BP. Symptomatic OH may therefore reflect greater falls in cerebral blood flow and oxygenation, though this relationship has been minimally examined [23]. Chronic OH may also lead to an expansion of the lower end of cerebral autoregulation in some individuals [24, 25], though how this could occur is not understood. In other scenarios, OH may trigger “atypical” symptoms such as dyspnea, fatigue, confusion or cognitive fluctuations, shoulder and neck (“coat hanger”) pain, or gait dysfunction; importantly, these symptoms may not be recognized by an individual or their treating physician as being related to OH. Amongst their cohort of 205 individuals with severe OH (defined as an SBP fall ≥ 60 mmHg), Arbogast and colleagues found atypical symptoms were reported by 24% of individuals during head-up tilt testing [11]. Many atypical symptoms overlap with common motor and non-motor manifestations of α-synucleinopathies, which could further contribute to a symptom/sign disconnection by the patient and/or physician [26]. An inability to perceive internal events may also be a feature of neurodegenerative processes that lead to autonomic failure. An “interoceptive” lesion could help explain several examples of visceral unawareness seen more broadly in Parkinson’s disease (e.g., dysphagia and constipation unawareness) [27], and diabetes (hypoglycemia unawareness) [28].

Collectively, these factors do NOT support the assumption that patients are reliable reporters of whether or not they are experiencing OH. Physicians are also unlikely to test for OH in the absence of typical OH symptoms [29, 30]. As a result, physicians are often unreliable recorders of OH, particularly when considering time barriers that reduce routine, in-office orthostatic BP measures.

Assumption 2: there is a reliable relationship between orthostatic hemodynamics and OH symptoms

To this author’s knowledge (and time of writing), only two studies have demonstrated a relationship between the lowest recorded BP and OH symptom report. Palma and colleagues [12] reported that, in 105 patients with Parkinson’s disease and OH (67% of whom were asymptomatic), a mean BP < 75 mmHg was sensitive for discriminating between individuals with and without typical OH symptoms. Tipton and Cheshire, in a cohort of 100 individuals with OH (74% of whom were asymptomatic), also found symptomatic individuals had a lower upright BP nadir. However, these findings seem to be the exception, rather than the rule, as a greater number of studies, entailing far more individuals with OH, have found no relationship between hemodynamic responses during OH and symptom report [11, 13, 15, 31]. The aforementioned study by Arbogast and colleagues, for example, found virtually identical upright mean BP nadirs amongst symptomatic, asymptomatic, and atypically symptomatic individuals [11]. A similar finding was demonstrated by Claffey and colleagues [15] in 102 community-dwelling older Irish individuals with OH, whereby asymptomatic individuals (66%) had identical orthostatic BP responses to those with symptomatic OH.

Few studies have examined the relationship between OH symptoms, orthostatic BP, and cerebral hemodynamics. We recently showed, in a meta-analysis of 17 transcranial Doppler ultrasound studies of individuals with neurogenic OH, that orthostatic changes in BP were strongly correlated with measured changes in cerebral blood flow [23]. Only three of these studies stratified patients by symptom presence. Amongst these studies, a lower BP nadir and greater fall in cerebral blood flow was demonstrated in those with symptomatic OH. However, there was significant overlap in the data between symptomatic and asymptomatic patients. For example, some study groups of asymptomatic patients actually experienced greater falls in BP and cerebral blood flow than those who reported feeling symptomatic in other studies.

Collectively, the evidence suggests that, even for measures of cerebral perfusion, the assumption of a relationship between OH symptoms and hemodynamics is neither reliable nor straightforward.

Assumption 3: symptoms are a useful proxy for OH-related outcomes

Here I will focus on three relevant outcomes with respect to OH: overall function, gait and falls, and cognition. OH symptoms are indeed a significant detriment to independent function, as rated by both patients and caregivers [32]. However, at least two studies [16, 19] demonstrated that, in patients with Parkinson’s disease, OH is associated with reduced independence in instrumental and general activities of daily living regardless of symptom presence. The severity of OH tends to be a greater predictor of worse overall function, rather than symptoms [16]. Similarly, several studies of OH coincident with parkinsonism [16, 19] or in community-dwelling older people [15] indicate that fall risk is inherently greater in OH, and paradoxically may be greatest in asymptomatic OH. Claffey and colleagues [15] found that only asymptomatic OH in community-dwelling older individuals was associated with unexplained falls—a two-fold higher risk than those without OH over 6 years. These findings are especially striking given that falls are one of the most associated, and feared, risks of OH. Indeed, it is likely that OH causes independent detrimental effects on gait superimposed on effects of frequently concomittant sensory neuropathy, parkinsonism, or cerebellar ataxia. Studies examining gait in patients with OH and α-synucleinopathies and the general population have demonstrated that (1) OH is associated with impairment in several aspects of gait (e.g., slower speed, greater postural sway) [16, 17, 20, 33, 34], and (2) these impairments occur irrespective of OH symptom presence [16, 17, 20].

Several studies have demonstrated associations between OH and cognitive decline/dementia in both general populations, those with diabetes, and those with α-synucleinopathies [21, 35–37]. This association has been attributed to recurrent bouts of cerebral hypoperfusion in the setting of OH, which may explain MRI-based findings of greater temporal lobar cortical atrophy in individuals with OH from Lewy body disease [38]. It is presumed that symptomatic OH reflects critically low cerebral perfusion and that, contrastingly, asymptomatic OH is indicative of a sufficiently perfused brain. If this were the case, then logic would have it that individuals with asymptomatic OH should have a reduced risk of poor cognitive outcomes compared to those who are symptomatic. This is not borne out by the evidence. In the α-synucleinopathies, worse cross-sectional and longitudinal cognitive outcomes in those with OH from Parkinson’s disease or multiple system atrophy were similar irrespective of symptom presence [21, 22]. Rather, it was the duration in which patients had OH that best correlated with time to onset of cognitive impairment [21]. Three separate studies [39–41] have also demonstrated that acute, reversible cognitive deficits are detectable in asymptomatic patients with Parkinson’s disease during OH.

Collectively, a significant amount of evidence supports the notion that, in OH, impaired patient function, greater fall risk, and cognitive decline all occur irrespective of OH symptom presence. These findings show the lack of utility for symptoms serving as a useful proxy of important OH- related outcomes.

Conclusion

To be clear, this position is not one in support of discarding the use of symptom report in clinical practice or clinical trials. Symptoms remain a debilitating aspect of OH for many individuals and are required outcomes in clinical trials of anti-OH agents by regulatory agencies [42]. It is also a significant achievement to now have minimally clinically important differences established for at least one OH-related symptom scale, the Orthostatic Hypotension Questionnaire [8]. These are also not arguments in favor of solely using BP changes as the primary measure of treatment response in OH, particularly given how labile BP often is in patients with OH [43]. It could be argued that the above arguments were framed around studies of OH resulting from different conditions. This approach was taken because the definition of OH and most screening and treatment guidelines are agnostic to OH etiology. Further, the general prevalence of asymptomatic OH appears conserved across etiologies and the negative consequences seen in symptomatic and asymptomatic patients alike are also conserved in that respect. These factors suggest that separately discussing the topic based on etiology may not be necessary.

In summary, the arguments above support the notion that the supremacy of OH symptoms in screening for and managing OH is not optimal: patients are unreliable reporters and physicians are unreliable recorders, likely leading to underdiagnosis of OH; there is a poor correlation between symptom presence and orthostatic hemodynamics; and symptoms are a poor proxy for many of the most important negative outcomes associated with OH. Last, our reliance on symptom presence leads to exclusion of asymptomatic or minimally symptomatic patients in clinical trials, leading to a lack of clarity on how these patients may be best treated [44]. Future work in the field should move toward the development and validation of objective markers that optimally capture OH-related outcome risks and response to treatments. Integrating these objective markers with established symptom-based assessments would create a holistic and inclusive framework meant to improve detection of OH in those at risk and better clinical outcomes for all patients with OH.