Selected global HTN guidelines covering 695 pages and 478,537 words recommend pressure-led diagnosis and management practices that result in approximately 25% of treated subjects achieving effective BP control [3]. A basic understanding of the physiology and pathophysiology of HTN identifies central haemodynamics as the source of the circulatory dysfunction in HTN and, therefore, the logical target of interventional therapies. This insight provides a clear opportunity to improve the guidelines, clinical practice and most crucially, patient outcomes for what is, at least as far as effective therapeutics are concerned, arguably an orphan disease.
While the deficiencies of BP-led approaches to HTN management may be rooted in the non-physiologic focus on pressure-dominated concepts, several other significant aspects of BP monitoring may also contribute.
Despite being ubiquitous in clinical practice, automatic office BP (AOBP) monitoring is only modestly accurate. Kallioinen et al. [20] in a meta-analysis of 328 full-text articles identified 29 specific causes of “potential mismeasurement” of BP in adults in clinical settings. The study reported SBP errors ranging from -23.6mmHg to +33mmHg and DBP errors from -14mmHg to +23mmHg across all sources of error, which included eight patient-related causes, two device-related causes, 16 procedure-related causes, and three observer-related causes. The authors concluded that “a single BP value outside the expected range should be interpreted with caution and not taken as a definitive indicator of clinical deterioration.” Importantly the reported errors strongly bias toward the over-measurement of SBP and DBP resulting in the misdiagnosis of subjects as hypertensive, while only 14-17% of errors resulted in under-measurement.
In a meta-analysis of 74 studies comparing invasive central aortic BP to brachial cuff BP measurements in 3,073 subjects Picone et al. [21] found that on average SBP was under-measured by 5.7mmHg and DBP was over-measured by 5.5mmHg by noninvasive brachial measurement. They also estimated that a difference of 5.0mmHg in measured BP may result in misdiagnosis of 48m subjects in the US alone, and concluded that there was “uncertainty as to whether cuff BP accurately measures intra-arterial BP”.
Validation of brachial AOBP devices is by comparison of measurements from proposed devices with those from reference standards, or by proof of equivalence against existing validated devices [22]. While intra-arterial pressure transducer monitoring of the aortic root is the gold standard comparator, the measurement sites are 25-35cm separate and have discrete pressures and flow waveforms [23]. Additionally the flow in the ascending aorta is not laminar but helical with velocity and pressure vortices associated with varying arterial morphology and arterial branching, making positioning of the transducer mid-stream in the aortic flow cross-section critical for accurate referencing [24]. Such technical sources of error combine to make invasive validation challenging, potentially contributing to significant variations between invasive aortic and invasive brachial BP measurements, and invasive and non-invasive brachial values. Nakagami et al. [25] found that invasive aortic SBP was 17.7mmHg greater than non-invasive brachial SBP, and invasive brachial measures of SBP were on average 22.2mmHg greater than non-invasive brachial SBP values. Pelazza and Filho [26] found central aortic catheter transducers measured 3-8% greater values than those measurement by AOBP in the clinically relevant 40 to 80± year olds.
These disagreements between invasive and cuff based brachial measurements, and between invasive central BP measurements and brachial AOBP measurements undermines the reliability of the technologies on which we rely. The recent introduction of brachial AOBPs that measure central BP and waveforms is targeted at reducing this source of error and improving their clinical reliability.
A further challenge in determining the accuracy of BP measurement is that circulatory physiology and its component measures, BP, SV, CO and SVR are dynamic variables with significantly wide normal ranges and standard deviations. These fluxes may reflect normal physiology, such as in respiration, excitement and exercise, or pathophysiologic changes [27]. As SBP and DBP are measured non-contemporaneously during cuff deflation, the accuracy of BP measurements may be influenced by the stage of the respiratory cycle at which the measurements are taken. Inspiration increases the SBP, DBP, PP and the SV, while they fall during expiration [28] at a rate correlating with the depth of respiration and intra-thoracic pressure changes [29]. As the resting adult respiratory cycle is 4-8 seconds and a normal cuff inflation and deflation cycle of an AOBP is in the order of 5-10 seconds, the SBP and DBP measurements will almost certainly be acquired from separate phases of the cardiac cycle, including from the apex and nadir of respiration ensuring maximum inaccuracy. This adds a further ±6% physiologic variability to BP measures even assuming pressures are measured precisely [29].
These multiple errors of AOBP measures vary within and across subjects, operators and devices, and result in summed errors in clinical measurement of at least ±15mmHg [20]. This scale of mismeasurement may result in significant misclassification of hypertensive status, and an insensitivity for monitoring the effectiveness of therapy, thus seriously limiting the implementation of current BP-led HTN guidelines.
International ISO 81060-2:2018(E) guidelines for the clinical investigation of intermittent automated measurement of non-invasive sphygmomanometers recommends equivalence validation values of ±5mmHg with an SD of <8mmHg between AOBP’s for accurate assessment of clinical hypertension. Given the above findings, this comparison may be practically unattainable.
So not only is BP a physiologically inappropriate target for monitoring circulatory dysregulation and guiding its treatment, but its clinical resolution is inadequate for the implementation of precision therapeutics, both of which may conspire to impair outcomes from current HTN management and explain inadequacies of current outcomes.
The potential to adopt central haemodynamic-led approaches to improve management of HTN depends on the availability of a non-invasive technology with clinical resolution at least as accurate as that of AOBPs. Continuous wave Doppler (CW) is well validated with a sensitivity of ±3% in string phantoms [30,31]. An anthropometrically calibrated CW (USCOM 1A, Uscom Ltd, Australia) has been validated from 0.12l/min to 18.7l/min across multiple clinical applications [33-38], is non-invasive, simple to use and interpret, with an examination time of less than five minutes [39], and can effectively guide therapy in free breathing subjects and those in sinus rhythm and with atrial fibrillation [36]. The technology generates multiple parameters of cardiovascular performance beat to beat, including SV, CO and SVR [40], and can display values from each stroke, or as an average output value from a 10-second window, thus negating the effects of respiratory variation and increasing the reliability of measurements and its sensitivity to change. Kusomoto estimated a single measure to measure sensitivity of CW of 11% [41], and using generalizability theory predicted the sensitivity from 10 averaged SV measures repeated as being approximately 3.5%. In vivo sensitivity of ±5% was reported in sheep studies against surgically implanted ultrasonic flow probes over a six-fold range of cardiac outputs varied using inotropes and vasoactive therapies [33]. These high sensitivities to therapeutic change, which are much greater than AOBP measurements, are essential to the early detection and management of early central haemodynamic changes in occult HTN, prehypertension and HTN. The integration of contemporaneous central haemodynamic monitoring with central BP also allows reconstruction of noninvasive pressure volume loops which can simplify and improve the representation of dynamic ventriculo-arterial coupling in physiology, pathophysiology and therapy not achieved by BP monitoring alone.
A pilot validation of central haemodynamics-led diagnosis and management of HTN is in the field of hypertensive disease in pregnancy (HDP), a common and dangerous complication of pregnancies worldwide [42]. Despite widespread publicity and public health spending, HDP remains the second most common cause of maternal and foetal morbidity and mortality worldwide [43], affecting 1 in 7 hospital deliveries [44]. HDP Guidelines from ACC/AHA [45,48], ACOG [47], ESC [15] and Nice [48] are all BP-led, show little consensus regarding therapy, with none mentioning central haemodynamic targets.
Anthropometrically calibrated CW monitoring of central haemodynamics has recently been adopted in HDP to better understand the normal evolution of the maternal circulation and its pathophysiology [49]. While the aetiology of HDP is unique, the circulatory dysregulation and therapeutic approach are analogous to non-maternal HTN. Eighty five percent of normal haemodynamic adaptation in pregnancy occurs before the 16th gestational week [50], while early dysregulation of maternal SV, CO and SVR leading to HDP can be detected at 5-11 weeks gestational age. However BP changes of HTN are not detected until 20-25 weeks [50], thus providing a therapeutic window for early and physiologically precise therapy during the period of ANS compensated normotension. Importantly the concept of cardiogenic and vasogenic aetiology has been established allowing for a simplified approach to treatment [51] using vasodilators, negative inotropes and diuretics implemented before the decompensated phase of frank HTN is established by the re-set of baroreceptor set-points. Precision management according to objective measures of SV, CO and SVR is consistent with circulatory physiology and pathophysiology and is improving the understanding of HDP [52-55], with an SVR value greater than 1069 dynes.s.cm-5 significantly associated with increased maternal-foetal complications [52]. This strategy is producing promising therapeutic results, and improving maternal-foetal outcomes [56,57] with principles that may also translate to non-maternal HTN and prehypertension.
Anthropometrically calibrated CW has also established a utility in adult and paediatric critical care for the guidance of fluid, inotropes and vasoactive agents in sepsis, where BP can be either pathologically high or low, and labile, and where circulatory performance is characterised by shifting intravascular volumes, and intermittent vasoplegia and cardioplegia [35,58-61].
The Framingham Heart Study has provided a cornerstone of our understanding of HTN with 74 years of study of a database of ~15,000 participants covering three generations resulting in 10 and 30-year risk prediction algorithms for heart disease, including HTN [62]. However, this monumental study is referenced substantially from BP measurements, thus embedding in the data all the limitations addressed in the preceding discussion. While the Framingham data has provided unequalled epidemiological evidence, the benefit with regard to therapeutic precision is less established, a limitation partly explained by its dependence on BP measurements. The predominance of BP data in this foundation study has also ensured that the evidence bar for the adoption of new technologies is exceptionally high. This limits the adoption of complimentary technologies which may expand our understanding and management of HTN, and ensures the persistence of the “treatment of all with the therapy to benefit the majority” strategy; a strategy that contributes to the troublingly substantial “BP treated but uncontrolled” cohort in HTN outcomes data.
The clinical adoption of central haemodynamics-led HTN management may be most effectively delivered by bespoke hypertensive clinics, with specialised clinicians and technicians and an expanded data monitoring network which may generate additional public health cost. However any additional spending will be more than offset by even a small improvement in effectiveness of care and improved clinical outcomes, particularly in early detection and treatment of prehypertension.
The limitations of this study include the discretionary choice of practice guidelines for review based on geographic and reputational criteria. While additional guidelines may have been included, it is unlikely they would have significantly impacted the study’s findings.
A further limitation is that the presence or absence of recommendations in practice guidelines, while correlating with clinical outcomes, does not prove causation, nor prove the hypothesis that the inclusion of central haemodynamic parameters in guidelines will ensure an improvement in clinical outcomes. However the physiologic argument is compelling and the pilot applications in HDP, and management of HTN and hypotension in sepsis and paediatrics are established, suggesting further studies to investigate the potential incremental value of central haemodynamics in HTN are warranted.
The poor outcomes of HTN management can be attributed to multiple causes including pressure-led management, inaccuracy of BP technologies, the complexity of the underlying pathophysiology, and inadequacy of examination and monitoring techniques. Global HTN practice guidelines, founded on BP-led monitoring, are widely adopted, yet substantially ineffective, while central haemodynamics which are core to the aetiology and therapy of HTN remain neglected. In the absence of significant changes it is difficult to envisage improvements in current clinical outcomes. The adoption of flow-led practices is physiologically rational and may compliment current practice and improve our understanding of HTN, facilitating the delivery of earlier and more precise anti-hypertensive therapy and improved outcomes in this serious and increasingly common disease.