Primary Aldosteronism: Clinical Lateralization and Costs

In this issue of the JCEM, Küpers et al. (1) make a truly groundbreaking contribution to the management of primary aldosteronism (PA). In a retrospective study on consecutive patients investigated for PA over an 18-month period, they show that in 87 subjects with confirmed PA, adrenal venous sampling (AVS) lateralized hypersecretion in 49 of them. In this latter group, the adrenal oversecreting aldosterone could be predicted with 100% accuracy by the combination of a unilateral adrenal mass of at least 10 mm on imaging, plus a plasma [K+] on presentation of less than 3.5, and/or an effective glomerular filtration rate (eGFR) of at least 100 ml/min. In this way, 30% of patients with confirmed PA can confidently forego AVS and proceed directly to laparoscopic adrenalectomy. As the authors themselves note, this initial study needs to be validated on an independent sample. This should pose no great difficulty, given the potential availability of similarly curated (and importantly, recently derived) cohorts from other European countries and elsewhere. This accompanying editorial commentary will thus focus on two areas. The first is what follow-up studies might address, to extend as well as to validate the initial study by Küpers et al. (1). This is not meant to be picky and is a tribute to the study, rather than a criticism; the additional issues stem largely from questions raised by the Küpers et al. (1) paper and may prompt subsequent investigators to build on, rather than merely repeat, the initial study. The second—quite different—area is to address the concerns about access to and cost of AVS voiced by Küpers et al. (1). The question there is whether, in addition to the substantial advance they have demonstrated, we need a radical reassessment of the management of PA as a whole and a truly fresh approach, on public health grounds, to what needs to be done. There are a number of questions raised by the initial study. First, of those confirmed with PA, 49 of 87 subjects were lateralized, pointing to a prevalence of 56% of aldosterone-producing adenoma (APA) as responsible for aldosterone hypersecretion—a percentage higher than in most other reports, with the exception of some from Japan. What this may reflect are the relatively strict criteria applied for a positive aldosterone to renin ratio (ARR). One is the requirement of a baseline sitting/standing plasma aldosterone concentration of 500–550 pmol/liter (17–19 ng/dl). This is in addition to a positive ARR on at least two occasions, where the lower level of active renin concentration is taken as 5 mU/ml, and a positive ARR set at 64 pmol/mU. Taken together, those cutoffs may exclude a group of patients with bilateral adrenal hyperplasia (BAH), who are included on perhaps more relaxed criteria elsewhere, and contribute to the more commonly reported 2:1 ratio of BAH to APA. Second, when a unilateral mass—regardless of patient age—is used for lateralization, the authors report six “false positives.” Five of those six were when the lateralization index (LI) fell short of the cutoff of the required 4-fold difference in aldosterone/cortisol ratio. It would be interesting to know whether these LI values were in the 2–4 range, and in particular whether these subjects—any or all—had plasma [K+] below 3.5 mmol/liter, and/or eGFR of at least 100 ml/min. In addition, it would be comforting to know that the remaining false-positive (who lateralized to the contralateral side) was biochemically cured and had an improved blood pressure after laparoscopic adrenalectomy on the basis of the AVS. Third, further refinement of the diagnostic nomogram [see Table 3 in Küpers et al. (1)] may be possible. As it stands, the presence of a characteristic at least 10-mm mass on imaging scores 3 points, plasma [K+] of less than 3.5 mmol/liter scores 2 points, and eGFR of 80–100 ml/min scores 1 point and 100 ml/min or greater scores 2 points. A total score of 5 is 100% predictive. From the authors' own data and in studies to follow, it would be interesting to know whether hypokalemia could also be usefully calibrated—for example, a score of 1 for a plasma [K+] of 3.5–3.8 mmol/liter. If this were the case, then a [K+] very low in the accepted normal range, together with an eGFR in the 80–100 ml/min range, might extend the percentage of confirmed PA patients who would not require AVS. Finally, there are a host of relatively minor questions that might be asked. In the authors' sample, approximately 25% of the PA patients were unable to be completely weaned off their medication before investigation; it would have been interesting to know whether this group—who presumably had higher levels of baseline blood pressure—differed in any other way from the remainder. The 26 subjects able to be lateralized equally well by AVS and by a combination of imaging and hypokalemia/eGFR had LI values ranging from 6.8–117; it would similarly be of interest to know three things. The first is the LI for the remaining 23 patients who lateralized, on the presumption that they would show a lower range, and possibly significantly lower mean values, given that plasma [K+] and eGFR both in large part reflect plasma aldosterone concentration levels. The second is whether any of the 14 patients in whom bilateral AVS was unsuccessful had the combination of mass on imaging, hypokalemia, and elevated eGFR, and if so underwent laparoscopic adrenalectomy. The third is whether the authors contemplated using unilaterally successful AVS as a guide to therapy if the successful AVS sample showed suppressed aldosterone production. The answers to all of those questions do not affect the central finding of the paper but point to the sorts of issues raised by a study such as this. In their introductory section, Küpers et al. (1) write: “Regardless of its diagnostic value, AVS has several drawbacks.” They cite complication and failure rates and nonstandardized methodology and interpretation, before coming to the question of costs. The final drawback cited is that “…AVS is not available everywhere and (is) costly. Referral centers are able to perform only a few AVS at most each year. They will be unable to meet the demand induced by broad screening for PA even in high-income countries, unless AVS is avoided in a significant number of cases.” Quite so, and although my fervent hope is that the Küpers et al. (1) study is unequivocally validated, a 30% fall in the necessity of AVS for all patients confirmed with PA is actually only the tip of the iceberg in terms of costs, capacity, and most importantly, access. What we need is a radical rethink, recognizing PA as a substantial public health issue, and squarely facing how best to address it. The Endocrine Society guidelines and those of the European Society for Endocrinology and the Japanese Society for Hypertension recommend screening for a substantial proportion of hypertensives; since their publication, a convincing case can be made for patients with atrial fibrillation and obstructive sleep apnea to be added to the list. The Japanese Society for Endocrinology recommends that all hypertensive patients be screened. Norman Kaplan, a celebrated contrarian in the area, would confine screening to three groups: patients with an adrenal mass on imaging, those with spontaneous hypokalemia, and subjects with resistant hypertension (2); even this rather straitened list, however, includes approximately 30% of hypertensives. Now is the time for a little math. In the United States, where there is a normally spread age demographic, the prevalence of hypertension is approximately 30% on a population basis. In Japan and in many European countries with a relatively low birthrate, the prevalence is closer to 40%, given the increasing median age of the population. We were taught in medical school that PA was a rare (<1% of hypertension, relatively benign) condition, which we now know is not the case. A consensus figure for prevalence is approximately 10%, and the reported risk factor profile (stroke, nonfatal myocardial infarction, atrial fibrillation) is four to 12 times higher than in age-, sex-, and blood pressure-matched essential hypertensives (3). More math. If 30% of the U.S. population is hypertensive, there are at least 10 million subjects with PA. The normal “return” from screening hypertensives, in general practice or in referred patients (4), is that approximately 20% test positive, of whom about half are confirmed with PA. In most series, a unilateral APA is responsible for 30–40% of the cases, with the remainder largely BAH. Norman Kaplan (2) cites an annual figure in the United States of 11,000 unilateral adrenalectomies, the majority presumably for PA. If this is the case, then by extrapolation approximately 20,000 cases of PA are diagnosed and treated medically or by laparoscopic adrenalectomy each year. By extrapolation, this represents approximately 200,000 subjects screened each year. This is very possibly an overestimate because many patients diagnosed with PA at present are at the more florid end of the spectrum of severity. We may see a stochastic rise in screening rate if three conditions can be met. First, awareness of PA and the public health problems it poses needs to increase markedly among family doctors, internists, cardiologists, and nephrologists, as well as endocrinologists; this is no easy task. Second, we need cheap, reproducible point of care (finger-prick) assays for renin and aldosterone, like those currently existing for glucose and D-dimer. Finally, we need extended guidelines on how to interpret an ARR (5) in the presence of antihypertensives (excluding mineralocorticoid receptor antagonists and epithelial sodium channel blockers), so that point-of-care screening can become routine. It is clearly impractical even to consider screening 30, 60, or 100% of hypertensives in a single year, not only for the obvious reasons of cost and laboratory capacity, but also for a rather less obvious reason: what do we do with the millions of hypertensive subjects who will then test positive on ARR? Currently, if approximately 200,000 hypertensives are screened in the United States in any year, the other 99.8% go unscreened; the overall numbers of unscreened hypertensives will actually grow with population increase and aging. If the screening rate increased 10-fold, we would be marking time. Without going into actuarial gymnastics, even screening 10% of hypertensives per year—a 100-fold increase, and starting now—would take decades to reduce the unscreened hypertensives to a figure of 20 million, rather than the current 100 million. Again more math. At our current screening rate, far fewer than 1% of hypertensives in any jurisdiction are screened, let alone diagnosed and appropriately managed. Even if the screening rate were to jump 10-fold overnight, just to mark time, there is no way in which the capacity to do confirmatory/exclusion testing, imaging, AVS, and laparoscopic surgery could hope to cope. The outcome of extended and, even more so, universal screening would be to tell positive subjects “You have an approximately 50% chance of having primary aldosteronism, so here is a prescription to add a mineralocorticoid receptor antagonist to your medication regime.” We are not going to see a 10-fold increase in screening overnight, and it would be a bold prediction that it will be the case in 2020. In the meantime, what should we do? The answer is to include a low-dose mineralocorticoid receptor antagonist (spironolactone 12.5/25 mg daily, or equivalent doses of another antagonist) as part of first-line therapy for all hypertensive subjects. In treating all for the benefit of a minority, the guiding principle has to be “primum non nocere.” In uncomplicated essential hypertension, mineralocorticoid receptor antagonists are safe, efficacious in lowering blood pressure (6), and uniquely vasoprotective (7). For patients with resistant hypertension (three conventional antihypertensives including a diuretic), who constitute approximately 20% of hypertensives, they are selectively blood pressure lowering (8). For the more than 99% of patients with occult PA—representing essentially 1 in 10 hypertensives—low-dose mineralocorticoid receptor antagonists would be game changing: specific for BAH, (the majority), and much better than no targeted medical treatment in APA. The cost can be minimal and can be dwarfed by the personal and societal costs in terms of morbidity, mortality, loss of productivity, and medical care if the condition goes untreated. There are things we know, things we don't know, and some things we may never know. One thing I know is that there are more than 10 times more cardiologists than endocrinologists in most countries, and although the endocrinologists write the guidelines for PA, the cardiologists write them for hypertension. Something I may never know, given the disconnect between who writes the guidelines, is why we pay lip service to evidence-based medicine but do not include low-dose spironolactone in first-line treatment for hypertension. We need to rework the 2008 Endocrine Society Guidelines (9), for the case detection, diagnosis, and management of PA—to take into account the little-recognized public health imperatives that the prevalence of untreated PA poses. This work was supported by the Victorian Government's Operational Infrastructure Support Program. Disclosure Summary: The author has nothing to disclose. Aldosterone-producing adenoma aldosterone to renin ratio adrenal venous sampling bilateral adrenal hyperplasia effective glomerular filtration rate lateralization index primary aldosteronism.