Primary Aldosteronism: unnecessary complexity in definition and diagnosis as a barrier to wider clinical care

‘Why didn't my doctor make this diagnosis 10 years ago when I first started having this problem?’ This is a troubling question heard all too often in the hypertension clinic, from patients who are diagnosed with primary aldosteronism after many years of medical consultations and failed drug treatment attempts. It is a statement that is impossible to answer in the individual case but perhaps worth considering in the broader context of medicine's approach to hypertension. Primary aldosteronism (PA) has seen a proliferation of research across endocrinology, cardiology and nephrology fields. Once thought to be rare, PA is often claimed to be the most common definable cause of resistant hypertension. With studies showing a higher prevalence than previously estimated, there has been an ongoing debate as to the true prevalence and clinical relevance of this disease entity with some papers describing ‘the renaissance of a syndrome’1 but others decrying ‘the apparent aldosteronism epidemic’.2 I suggest neither characterization is correct; rather, perhaps the ongoing ways in which PA is defined, taught and managed may actually be a barrier for clinicians tasked with controlling the global burden of (resistant) hypertensive disease. Untreated PA is known to be associated with higher rates of cardiac arrhythmia, coronary heart disease, heart failure, stroke, proteinuria and renal impairment compared to essential hypertension.3, 4 Medical therapy of PA can render previously uncontrolled hypertension into a target treatment range, often with fewer medications than at presentation.5 Surgically treated forms of PA may result in complete disappearance of hypertension in up to 40–70%6, 7 with marked improvement in most of the remaining cases. The response to both medical and surgical therapy appears durable over long-term follow-up,8, 9 and limited outcome data suggest that PA-specific therapy substantially reduces the risk of adverse vascular and cardiac outcomes10, 11 and improves quality of life.12 The reported prevalence may depend in part upon the population that is screened. Populations of patients with ‘resistant hypertension’ may have PA prevalence of 10–24%,13, 14 compared to rates of 1–13% in less selected hypertension populations.15, 16 Studies that attempt to define PA prevalence exist from countries including USA (multiracial), Italy, UK, Sweden, Chile, Singapore, Japan, Australia and South Africa.17-24 If it is agreed that PA forms a meaningful subset of hypertension in virtually every human population, the global burden of PA hypertension could theoretically reach into the tens of millions. If true, this global disease prevalence estimate seems at odds with the fact that most studies in PA enrol fewer than 100–300 subjects and the largest PA studies appear to require a decade of multicentre or even multicontinent pooling just to get to 1000 subjects. From a treatment perspective, virtually all PA therapy studies are small, uncontrolled, short duration, often retrospective and involve surrogate or biochemical end-points rather than patient-level clinical outcomes. Perhaps it is the very definition of disease and regimented approach to diagnosis that has turned this apparently common problem into a rigid diagnostic label that is essentially inaccessible to the vast majority of affected persons. Lacking a ‘specialist-verified’ diagnostic label of ‘confirmed PA’, potentially valuable, disease-specific therapy is thus overlooked in many people and outcome-based research will be slow, expensive or even impossible. The origin of the problem might lie with the current paradigm of ‘secondary’ hypertension which is thought of as an entirely separate entity from ‘primary hypertension’, only to be considered in the presence of specific clinical clues or resistant hypertension cases. In turn, ‘secondary hypertension’ has a differential diagnosis of specific diseases that may be sought in a case-finding approach depending on the presence of screening tests like serum potassium that are known to be of very poor sensitivity.25, 26 Basic population health principles would clearly not support an individualized case-finding approach with an insensitive test for a treatable disease with population prevalence of more than 10%. In addition, this case-finding approach leaves the physician with little to no direction in the management of such patients who do not fit strict case definitions. It is universally agreed that the first step for considering the diagnosis requires the measurement of an aldosterone–renin ratio (ARR).27 Notwithstanding interpretation difficulties, simply measuring these two hormones can be intimidating. Many drugs can affect the ARR and current recommendations are that such medications be stopped or replaced.28 The difficulty is that front-line physicians need to know this, need to memorize the entire list of interfering medications and additionally remember which medications can be used in their place along with the duration of time required for specific drug washout. Medication switching requires time, expertise and patient teaching along with new prescriptions/written instructions, all of which have costs. Not surprisingly, this process is usually deferred to the specialist clinic with its accompanying consultation costs and wait times. Switching of the medications is also not without potential patient risk.29 Once measured, interpretation of the ARR can be a formidable task. Plasma aldosterone levels may be reported in either ng/dl or pmol/l. Renin may be reported according to a plasma renin activity (PRA) in ng/l/s, ng/ml/h or pmol/l/min or alternately as a direct renin concentration in mU/l or ng/l. Depending on the assays in use, the case-defining ARR can be numerically highly variable from ‘1·6’ to ‘1000’.30 Not surprisingly, there can be many misinterpreted ARR results in referred patients whose physicians ‘looked up’ the ARR interpretation cut-offs without understanding the details in various laboratory assays, reporting methods and measurement units. The exact ARR level that warrants further attention is not clear; even within a standardized methodology and unit reporting (aldosterone in pmol/l and PRA in ng/ml/h), different centres around the world have advocated for different cut-offs ranging from 550 to 2800.31-33 According to the PA guidelines, an abnormal ARR should be followed by some form of dynamic confirmatory testing,30 a procedure that almost assuredly takes this diagnosis out of the hands of primary care and even most specialists. Fludrocortisone suppression testing34 requires admission to hospital and may be unsafe in the hypokalaemic patient. Saline suppression tests require IV access and the administration of 2000 cc of 0·9% saline along with a nurse to draw the timed serum aldosterone samples.35 IV salt loading (especially with fludrocortisone) is unadvised in situations where cardiac or renal function is uncertain.36 Oral NaCl-suppressed 24-h urine aldosterone or postcaptopril testing may be more accessible approaches but are less well studied and the accuracy of the captopril test may be particularly limited.37 Regardless of the test used, there continues to be confusion in the medical literature regarding interpretation of the results.38-41 One of the major limiting factors in the ‘confirmatory test’ literature is the inability to test procedures against a uniformly applied gold standard as response to spironolactone does not necessarily diagnose PA42 and it is ethically impossible to obtain adrenal tissue from an entire prospective study population including control subjects in order to verify the test characteristics. This explains why there continues to be uncertainty as to the very definition of a PA case30 and debate as to the proper interpretation of so-called screening and confirmatory tests.43 Finally, if a patient is found to have PA (by whatever definition is being used), current practice would dictate that adrenal imaging should follow and there are a multitude of papers now claiming the necessity of adrenal vein sampling (AVS) to determine the patient's surgical candidacy.44 The costs and accessibility of these tests, particularly AVS which requires a high degree of technical expertise, may be a major limiting factor even in advanced healthcare settings. In summary, a PA diagnosis in 2015 usually starts with limited, targeted screening. It uses a test that requires detailed patient preparation unknown to most physicians. The interpretation of the test is not standardized across laboratories and may generate a vast number of potential figures that will be unfamiliar to most clinicians outside of highly specialized hypertension units. It requires access to a funded metabolic testing unit or hospital to perform a test that has unknown sensitivity for results other than what has been reported in small studies. Reflex ordering of CT abdomen and AVS then add to the cost and time if ordered inappropriately for incorrect diagnoses or nonsurgical candidates. It should thus be no surprise that this cumbersome process has made a ‘strict’ PA diagnosis almost impossible outside of highly experienced and dedicated hypertension units which in turn can practically serve only a tiny fraction of the potentially affected population. It is now time for PA to be reconsidered in the global healthcare context in order to permit simplified patient access to the diagnosis and treatment of this apparently common disease. A starting point might be to recognize that PA may not even exist as a wholly independent entity within the spectrum of hypertensive disease. It has already been shown that, at least according to the population distribution of the ARR, ‘PA’ appears to blend seamlessly into a general hypertension population and is highly heritable.45 Functional and genetic studies have further confirmed that ‘PA’ is not a single disease but in fact arises by multiple different processes including various genes,46 both germline and somatic,47 ectopic adrenal receptors48 and adipokines.49 Aldosterone levels predict hypertension development and not necessarily in the form of classical PA.50 Thus, unilateral, surgically amenable, classical PA may simply represent the extreme spectrum of a much larger population of aldosterone (plus sodium)-mediated hypertensives. Perhaps it is time to revisit a simplified, mechanistic approach to resistant hypertension that focuses less on detailed case definition and more upon effective blood pressure control. A simplified approach to PA consideration for resistant hypertension would permit front-line healthcare workers and community hypertension specialists a much better opportunity to diagnose and appropriately intervene in a manner commensurate with local healthcare resources. In communities with adequate resources, surgical forms of PA could still be diagnosed and surgically treated in accordance with patient suitability and willingness. Resource-restricted health communities would likely see an increase in mineralocorticoid receptor antagonist (MRA) use among such patients, a practice for which there are already ample, published supportive data.51-53 In fact, failure to consider MRA use in resistant hypertension may be a result of a lingering belief that such agents are only for the rare, specialist-confirmed cases of classical PA. Finally, it may be worth considering the global contributions that may be realized if future hypertension guidelines endorsed a simplified approach to PA consideration. While guideline-writing bodies may take care to specify that their guidelines are written with a specific population or health system in mind, the reliance upon current complex PA case definitions for medical school teaching means that many resource-restricted healthcare environments around the world will have absolutely no starting point for the consideration of PA in resistant hypertension despite the often widespread availability of spironolactone. Figures 1-3 present examples of a potential theoretical framework to the consideration of PA in resistant hypertension with diagnostic and treatment algorithms that may be applied to a wide range of health resource environments. This is modelled after the IDF Global Diabetes Guidelines54 which recognize the need for an approach in Limited Care environments, Recommended Care environments (for many parts of the world) and Comprehensive Care environments (most up-to-date and complete range of options). The figures should not be interpreted as a rigid and complete algorithm; the numerical ‘cut-offs’ and spironolactone doses may be changed based upon local evidence or more comprehensive study. Rather, these are simply meant to represent examples of what such a framework may encompass after broader input from the PA research community and primary care physicians. The desired result is to reduce the risk of premature CV disease and stroke, in the most feasible manner for the most number of people. The Limited Care algorithm (Fig. 1) acknowledges that a classic biochemical diagnosis of PA cannot be made. However, as PA is probably the most common specifically treatable cause of resistant hypertension and can be treated with an inexpensive and often widely available medication, this algorithm focuses more upon safety of MRA use and patient follow-up. A formal label of ‘PA’ is not emphasized but still gives affected individuals the opportunity to be treated with a specific potential diagnosis in mind. The Recommended Care algorithm (Fig. 2) represents a ‘shared care’ approach between primary care (shaded part of figure) and a medical consultant (nonshaded part of figure). It presumes access to at least one measurement of aldosterone and renin, computed tomography and adrenalectomy. The algorithm pragmatically redirects nonsurgical patients at an early stage and reserves further testing only for those in whom surgery is considered a legitimate possibility. In doing so, it makes use of several recently reported clinical and simple biochemical variables that appear to have good value for predicting unilateral forms of PA that might be amenable to surgery.55-58 Confirmatory testing is explicitly reserved for surgical candidates in whom the certainty of PA is considered to be lower (eukalaemic or older patients with marginal ARR elevations and/or confounding medications). Overtly abnormal findings on adrenal CT are a requisite finding, often in concert with predictive variables, to warrant surgery and presume the nonavailability of adrenal vein sampling. For the most part, the Comprehensive Care algorithm (Fig. 3) reflects the Endocrine Society PA Guidelines except that it envisions a ‘shared care’ approach with primary care based upon an initial, largely unadjusted ARR measurement in the community. Patients with clearly elevated renin at presentation (i.e. plasma renin activity >1·0 ng/ml/h) need not undergo extensive medication switching and repeat testing unless classic features (hypokalaemia, adrenal mass) are already known to be present.59, 60 Confirmatory testing is restricted to surgical candidate patients who lack gross biochemical abnormalities and adrenal vein sampling is performed in most if not all surgical candidates to confirm unilateral PA prior to surgery. Medical control options are considered early and repeatedly for those not likely to proceed to surgery. It would be ideal if these suggested algorithms could be further refined with the help of a panel of hypertension/PA experts and adopted across both primary and specialty care. The Comprehensive Care algorithm in particular is not intended to supplant the evidence-based PA guidelines already in existence but merely to clarify their potential implementation in a broader population setting and to highlight the role of primary care workers in bringing potentially affected patients forward for improved case detection. Additionally, this algorithm gives practical advice for management of patients who may never be formally diagnosed with PA but in whom their resistant hypertension is still probably in the mineralocorticoid spectrum. Rigorous analyses of the cost-effectiveness of PA case-finding and treatment approach have yet to be published but it is hoped that the proposed simplified approach could offer effective targeted therapy to a much wider population while limiting diagnostic costs to situations that have the highest expected yield in treatment decisions. As ‘Conn's syndrome’ prepares to celebrate its 60th birthday in 2015, it is time to move past the midlife crisis of restricted diagnostic labelling and commence a new era that concentrates on maximizing the population health benefits of treating this important contributor to cardio-renal disease across the globe. None. Nothing to disclose.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.