Unadjusted Plasma Renin Activity as a “First‐Look” Test to Decide Upon Further Investigations for Primary Aldosteronism

Abstract

The authors sought to define the 95th percentile of plasma renin activity (PRA) in a sample of patients with confirmed primary aldosteronism (PA) prior to adjustment of medications as a practical “first‐look” test to identify those with very low ultimate likelihood of having PA. The aldosterone to renin ratio (ARR) was measured without adjustment of antihypertensive medications, with further workup as appropriate. Two groups were defined: patients with surgically “confirmed PA” (n=58) and patients with “high‐probability PA” (n=59), defined as having any of the following: computed tomography–confirmed adrenal adenoma plus lateralizing adrenal vein sampling (AVS) without surgery, high ARR and hypokalemia but nonlateralizing AVS, or ARR more than four times the upper limit of normal. The PRA 95th percentile was 1.0 ng/mL/h. All outliers had hypokalemia and two had adrenal adenomas. There was no difference between the confirmed and high probability groups. In the absence of highly suspicious clinical features, patients with unadjusted PRA >1.0 ng/mL/h do not warrant further investigation for PA.

Primary aldosteronism (PA) is a common cause of resistant hypertension.1 Diagnosis and treatment with surgery or a mineralocorticoid receptor blocker can result in a significant decrease in blood pressure and correction of hypokalemia, a decrease in the number of medications required, and even complete hypertension resolution.2 Compared with age‐ and sex‐matched hypertensive controls, untreated patients with PA have increased cardiovascular morbidity and mortality,3, 4 higher incidence of glomerular hyperfiltration, and potential for bone and metabolic sequelae.5

The standard for screening is the aldosterone‐renin ratio (ARR), which many studies have shown to be a cost‐effective and efficient way to detect PA.6, 7 One Japanese study even recommends routine screening in all hypertension patients.8

In classical PA, there is autonomous secretion of aldosterone in the setting of an appropriately suppressed renin. However, many patients within the normal distribution of “essential hypertension” (EH) can also have low renin measurements,9 resulting in elevated ARR in patients not ultimately diagnosed with PA.10 Attempts to separate the ARR distribution curves of EH and PA are hampered by lack of a readily available and uniformly applicable diagnostic gold standard for PA.11 A study to determine the ideal ARR cutoff would require a very large sample size of hypertension patients in which every individual undergoes a full PA workup, regardless of initial screening results and including adrenal vein sampling (AVS). Stratification according to a gold‐standard PA diagnosis would permit retrospective calculation of the optimal ARR screening cutoffs, which may still differ depending on whether one sought perfect sensitivity or specificity. Such a study could be expensive given a large sample size and the multiple PA diagnostic steps currently recommended. Several large hypertension population‐screening measures for ARR/PA have been reported but the validity of the results is hampered by the selective investigation of just a few patients within each cohort.12 This uncertainty is reflected in the hyperaldosteronism guidelines, which offer a variety of potentially “abnormal” ARR cutoffs for possible clinical use.6

Regardless of the ARR cutoff used, a low renin level is the sine qua non of the PA diagnosis and is the major determinant of the ARR result.13 Most hypertensive patients are already taking medications known to affect the renin–angiotensin–aldosterone system.14, 15 A general recommendation is to stop interfering medications prior to measurement of the ARR,16 and this is the approach taken in the current guidelines.6 However, it is not infrequent in the community that ARR testing is done without these adjustments being made. In addition, this process is cumbersome with reports of complications.17

The challenge lies in the selection of appropriate patients for this level of necessary detailed investigation. A pragmatic approach might be to describe the renin distribution in a sample of carefully investigated and confirmed PA cases where no medication preadjustments are performed at initial screening. One could then conclude that future patients with unadjusted renin measures falling outside (above) the known PA distribution are highly unlikely to be diagnosed with PA after additional detailed medication switches and re‐testing.

Therefore, this study sought to answer the question: When a patient is taking medications that will potentially confound the ARR, how high can the unadjusted plasma renin activity (PRA) be for PA to still be a legitimate diagnosis? It is suggested that by defining the 95th percentile of PRA in a large sample of confirmed cases of PA, the diagnosis can essentially be excluded in most patients with renin higher than this, in the absence of other classic clinical alarm features.

Methods

The Calgary PA‐AVS Database and its analysis was approved by the local ethics review board. The study was conducted in the practice of a single Aldosteronism Interest Group operating a resistant hypertension specialty clinic that conducts aggressive screening for PA in all referred patients. A retrospective review was performed in all patients who were referred to the clinic with a final diagnosis of PA and undergoing AVS between June 2005 and November 2013. The approach to PA diagnosis and definitions, clinical demographics, and long‐term outcomes has been previously published from this cohort.18

All patients referred to the unit have measurement of an ARR prior to their first appointment, irrespective of current medication used, except mineralocorticoid antagonists, which are stopped for 6 weeks. The ARR is drawn with the patient in the seated, upright position prior to 10 am. If the result is >550 (aldosterone in pmol/L, PRA in ng/mL/h), the patient is counseled about a possible diagnosis of PA. If the patient wishes further investigation to determine whether they may have a surgical option, provided they are deemed to be a potential surgical candidate, computed tomography and AVS is performed, using our previously published protocol.18 Patients with resistant hypertension in whom the ARR is normal after first measurement are offered a repeat testing of the ARR after manipulation of their medications to stop any β‐blockers, angiotensin‐converting enzyme inhibitors, angiotensin receptor blockers, and diuretics with use of doxazosin and nifedipine or diltiazem as interim agents where possible. Oral potassium supplements are given as needed to ensure eukalemia prior to repeat testing. Classical PA confirmatory testing is not performed in our unit because we have previously demonstrated successful and accurate diagnosis with our approach in our particular population.19 Patients with normal ARR are managed per usual hypertension guidelines.

For the present analysis, participants were divided into two main groups. The first group was characterized as having “confirmed PA” and included patients who received outcome‐verified surgical intervention (n=58) for either a computed tomography (CT)–confirmed adenoma >8 mm20 and/or lateralizing AVS, defined as an adrenal aldosterone/cortisol (A/C) ratio >3:1 compared with the contralateral adrenal A/C ratio.

The second group included participants designated as having “high‐probability PA” (n=59), defined as one of the following: (1) CT‐confirmed classic adrenal adenoma and lateralizing AVS, but declined surgical intervention (n=14), (2) patients with high ARR and hypokalemia (K <3.5 mmol/L)1, 21, 22 but nonlateralizing AVS (n=27), or (3) patients with an ARR more than four times the upper limit of normal, which corresponds to a very high likelihood of PA11 (n=18).

Exclusions in this study (n=14) included one patient who had concomitant severe renal artery stenosis, patients in whom renin was measured elsewhere using a direct renin assay (n=5), and “lesser‐probability cases” who did not meet the definitions of the two study groups (n=8).

PRA was measured by RIA of angiotensin I in the presence of reagents that inhibit angiotensin I–converting enzyme and angiotensinases. Using GammaCoat Plasma Renin Activity I125RIA Kits (DiaSorin, Stillwater, MN), mean interassay imprecision coefficients of variation were 11.1 %, 7.8 %, and 13.4 % for PRA levels of 0.95 ng/mL/h, 5.17 ng/mL/h, and 21.2 ng/mL/h, respectively. Where PRA was undetectable, the lowest reporting limit of 0.1 ng/mL/h was used to compute the ARR to avoid overinflation of the result.

Standard descriptive statistics were used to describe both groups and all comparisons made using a Mann‐Whitney test for nonparametrically distributed variables. Statistical analysis was performed using MedCalc 12.7 (MedCalc Software, Mariakerke, Belgium).

Results

The demographics for our study participants are shown in the Table. Using the diagnostic stratification of “confirmed” vs “high‐probability” PA, the two groups were significantly different in terms of plasma aldosterone at presentation, initial serum potassium, and age. Note, the PRA was not significantly different between the groups.

Table 1.

Participant Characteristics

Parameter		Surgical (Median–IQR)	High Probability (Median–IQR)	P Value
No.		58	59
Age		48 (40–55)	51 (46–60)	.036
BMI, kg/m2		28.3 (25–34)	30.9 (25–33)	.533
Initial systolic BP, mm Hg		146 (134–162)	145 (133–154)	.285
Initial diastolic BP, mm Hg		90 (84–100)	90 (80–100)	.615
Medications at initial presentation, No. (%)	Total medications, No.	2.0 (1–3)	3.0 (2–4)	.160
	None	2 (3)	0
	β‐Blocker	24 (44)	33 (52)
	CCB	43 (78)	40 (63)
	ACE inhibitor	13 (24)	18 (29)
	ARB	21 (38)	25 (40)
	Direct renin inhibitor	3 (5)	11 (17)
	Diuretic	11 (20)	26 (41)
	Amiloride	3 (5)	3 (5)
	Vasodilator	1 (2)	3 (5)
	Central‐acting agent	3 (5)	2 (3)
	α‐Blocker	9 (16)	3 (5)
CT abnormality (adenoma >8 mm)		48 (87.3)	28 (44.4)
Initial K, mmol/L		2.8 (2.5–3.3)	3.4 (3.1–3.8)	< .001
Initial PRA, ng/mL/h		0.16 (0.1–0.4)	0.11 (0.1–0.29)	.198
Initial aldosterone level, pmol/L		619 (397–942)	384 (249–465)	< .001
ARR on presentation		3580 (753–7360)	4377 (2211–10,295)	.965
Base AVS ratio		18.6 (5.3–28.9)
Highest stimulation ratio on AVS		15.6 (6.5–31.7)
AVS failed		1	6
Postoperative PRA, ng/mL/h		0.7 (0.47–1.55) (n=40)
Postoperative aldosterone level, pmol/L		110 (70–189) (n=40)
Postoperative ARR		161 (70–342)
BP outcome, No. (%)	Hypertension resolution, No. (%)	25 (45.5)
	BP <140/90 mm Hg and taking fewer medications	26 (47.3)	50 (79.4)
	<10% BP change	1 (1.8)	4 (6.3)
	Loss to follow‐up	1 (1.8)	3 (4.8)
	Awaiting outcome	2 (3.6)	6 (9.5)

Abbreviations: ACE, angiotensin‐converting enzyme; ARB, angiotensin receptor blocker; ARR, aldosterone to renin ratio; AVS, adrenal vein sampling; BMI, body mass index; BP, blood pressure; CCB, calcium channel blocker; CT, computed tomography; IQR, interquartile range; PRA, plasma renin activity. Bold values indicate significance.

Postoperative outcomes are summarized at the bottom of the Table. A total of 45.5% of participants who underwent surgery had complete resolution of their hypertension, no longer requiring any medications. In the high‐probability PA group, 79.4% of participants were taking fewer medications and had blood pressures below 140/90 mm Hg after appropriate workup and initiation of a mineralocorticoid receptor antagonist or potassium‐sparing diuretic. Note that 9.5% of these patients were still awaiting final outcome determination at the time of database closure.

Mineralocorticoid receptor blockers were excluded from the count of medications at initial presentation as shown in the Table as these medications were stopped prior to measurement of ARR and other biochemical parameters.

Figure 1 shows the distribution of initial, unadjusted PRA values of all of our study participants. The 95th percentile of PRA distribution in the entire cohort falls at 1.0 ng/mL/h, indicating that 95% of patients had a value equal or lower than this without or prior to any medication adjustment. There were only five outliers who had values above this, all of whom had other factors highly suggestive of PA. These are described in detail below.

Figure 1.

The distribution of initial, unadjusted plasma renin activity values of all study participants.

Figure 2 illustrates distributions of PRA values from the entire study population, stratified by factors that are easily defined in clinical scenarios: hypokalemia, CT abnormality >8 mm, and ages below and above 40. The distribution of PRA in patients younger than 40 was even more uniform, with no patient having a PRA higher than 0.76 ng/mL/h.

Figure 2.

Distributions of plasma renin activity values from the entire study population, stratified by select factors. CT indicates computed tomography.

Of the participants in our study with an initial PRA >1 ng/mL/h, all five had hypokalemia in the setting of hypertension. The two such individuals from the confirmed cohort had a unilateral adrenal mass on CT. One had an ARR of 221 initially and 1442 on repeat examination, and experienced hypertension resolution after surgery. The other had an ARR of 470 initially and 2358 on repeat examination, and experienced marked improvement in blood pressure after surgery. Both had histologic evidence of adrenocortical adenoma on pathology and normalization of the ARR postoperatively.

The individual in the confirmed cohort with a PRA of precisely 1 ng/mL/h had hypokalemia, unilateral mass on CT, an initial ARR of 424, and a repeat ARR of 1210, and experienced hypertension resolution following surgery for a documented adrenal adenoma with normal ARR post‐surgery. Of the three individuals with initial high PRA in the high‐probability cohort, all had resistant hypertension and hypokalemia, two were CT negative, and had initial an ARR of 255 and 244 and a repeat ARR (after adjustment of medications) of 2281 and 549, respectively. The third had a unilateral CT mass, an initial ARR of 193, and a repeat ARR of 847. None of the patients lateralized on AVS but all three had marked improvement of blood pressure on spironolactone. Most importantly, because of the hypokalemia and known adrenal masses, none of these patients would have been missed due to the use of initial PRA as a “first‐look” test.

Discussion

The greatest challenge in any study of PA is defining the disease and population. At present, the gold standard of diagnosis is tissue confirmation in a lateralized case, in combination with postoperative normalization of ARR and potassium (if low), along with hypertension resolution. Future use of novel immunohistochemical or genetic markers may assist in case definitions among supposed PA patients but such tests are not yet widely available or broadly validated.23, 24 All of the case definition markers that are commonly determined in the preoperative or nonoperative state such as blood pressure, renin, aldosterone, potassium, and presence of CT findings, may fall into a spectrum of PA, which overlaps that of essential hypertension.25 Thus, a busy clinician or over‐burdened health system needs a simple and pragmatic way of making decisions about whether to spend time and money investigating for a relatively uncommon yet highly useful diagnosis.26 Current guidelines require every patient suspected of having PA to undergo medication adjustment before the first ARR measurement.6 If the prevalence of true PA is just 10% to 15%,27 then up to 90% of patients with resistant hypertension could be forced to stop or change critically important drugs to undergo testing for a disease they are ultimately unlikely to have.

We recognize that the current guidelines use the best available evidence to guide clinicians in consideration of a diagnosis of PA. What our study adds is a first step to aid clinical decision‐making, given that every patient with resistant hypertension does not currently undergo full investigation for this condition. This new information helps the clinician to confidently rule out the diagnosis prior to making any of the adjustments described above.

The present analysis of patients with well‐documented PA suggests that while the ARR can be highly variable even within confirmed cases, a PRA >1.0 mg/mL/h, measured prior to any adjustment of medications (save aldosterone antagonists) will exclude up to 95% of PA cases that would otherwise be of sufficient diagnostic certainty to come to AVS. This cutoff is probably even lower in the hypertensive population younger than 40 years.

Other investigators have reported the distribution of PRA in PA cohorts. Tanabe and colleagues28 described 385 PRA measurements performed in various conditions in a cohort of 71 patients with confirmed PA. Their results suggest that just 11 of 385 measurements showed a PRA >1.0 ng/mL/h, which is nearly identical to the results presented here. Although this is apparent in analysis of their presented data, it is not discussed explicitly in the paper. Hirohara and colleagues29 published the distribution of PRA in PA cases following furosemide and upright posture. In that study, the patients with aldosterone‐producing adenoma had a mean stimulated renin of 0.39±0.39 ng/mL/h. Thus, this also supports our findings by showing that virtually all true PA patients will not have a medication‐induced PRA >1.0 ng/mL/h. Previous smaller series of PA patients diagnosed without medication adjustment also report unadjusted PRA <0.4 ng/mL/h in all patients.30 However, there is at least one case series of confirmed PA with nonsuppressed renin caused by renal injury in which PRA was up to 1.8 ng/mL/h,31 which suggests that the presence of chronic kidney disease might be an important confounder to PRA interpretation.

While a formal receiver operating characteristic analysis is clearly more accurate and exact in defining any test sensitivity and specificity, it is only as valid as the stratification measures used to discriminate between the populations in question. In PA, such an analysis would be greatly confounded as a result of the large biochemical overlap between classically recognized PA and other low‐renin hypertension.32

In the absence of large cohort receiver operating characteristic analysis, our simplified and pragmatic approach is completely unable to inform the practitioner as to the specificity of any given unadjusted PRA measurement. Therefore, in the absence of additional clinical clues such as hypokalemia, adrenal masses, or extremely high ARR, an unadjusted PRA <1.0 ng/mL/h does not confirm a PA diagnosis but rather alerts the clinician to the possibility that further investigation for PA may well be fruitful. Similarly, in the hypertensive patient with unadjusted PRA >1.0 ng/mL/h, unless there are compelling additional clinical clues, the clinician can be at least 95% confident that cumbersome and potentially risky medication adjustment is unlikely to lead to a formal PA diagnosis and AVS requirement.

Study Strengths

Strengths of this study include the relatively large sample size, all assessed and treated at a single center, which ensured uniform investigation and treatment of participants, as well as access to particular information pertaining to outliers in our study. Use of a single clinical laboratory service minimizes potential for errors in data collection and combining results of different assays.

By including only surgically confirmed and high‐probability cases, and showing their biochemical similarities with surgical, clinical, pathologic, and biochemical outcomes, we can confidently say that our study describes a population in whom most would agree about the diagnosis of PA.

Study Limitations

The greatest weakness of our study is the potential for spectrum bias. Many of the patients who were referred to our resistant hypertension clinic had easily recognizable PA and, therefore, one cannot be certain that our results would seamlessly translate to a population of purely undifferentiated, uninvestigated hypertension as might be seen in a primary care clinic. Lack of comparison to a hypertensive but definitively non‐PA cohort precludes us from making any statement about the optimal cutoff that provides the best combination of both sensitivity and specificity for either ARR or PRA in the initial screening steps. There are other publications that have attempted to determine such a figure33, 34 and it may be that the sophisticated diagnostician might use both “optimized ARR” and our PRA figure since it might represent close to 100% sensitivity for medically relevant PA diagnoses.

In addition to this, it is difficult to measure the external validity of our study, which assumes that our population of PA patients is representative of all PA patients. We attempted to minimize this problem through our rigorous inclusion criteria, which sought to select only patients with high‐probability or confirmed disease. Other centers may use different laboratory assays, different disease definitions, and diagnostic algorithms, which may limit the immediate use of our findings in clinical practice elsewhere. Further validation with results from other research groups, ideally in a larger multinational study, would be highly useful to confirm the overall diagnostic principle and suggested cutoffs. As for all endocrine tests, each specialized hypertension center should ideally perform a similar analysis after accrual of sufficient numbers to ensure local validation.

An argument could be made limiting the generalization of our study, given that confirmatory testing is not done in our unit. We have previously published the rationale for our clinic's nonuse of traditional confirmatory tests,19 wherein we demonstrated no meaningful increase in false‐positive PA diagnosis compared with other centers. It would nonetheless be valuable if the present data could be further validated in centers using traditional confirmatory testing.

Conclusions

We propose that the absolute value of PRA be used with a cutoff of PRA >1.0 ng/mL/h to exclude clinically relevant PA in patients in the absence of CT findings or hypokalemia, prior to the adjustment of medications. Patients with an initial unadjusted PRA <1.0 ng/mL/h will not necessarily have a final diagnosis of PA but might at least warrant further diagnostic consideration. Patients with initial unadjusted PRA >1.0 ng/mL/h might still be considered for further “medication‐adjusted screening” if they have known hypokalemia or an adrenal mass. We are not advocating any change to the current PA diagnostic guidelines, but rather contributing practical information to be used in initial consideration of the diagnosis. This addition to the initial approach may prove to be safer, easier, and less expensive than uniform application of detailed medication adjustment protocols for all resistant hypertension patients.

Disclosures

The authors have no conflicts of interest to declare. Funding source: not applicable.

J Clin Hypertens (Greenwich). 2015;17:541–546. DOI: 10.1111/jch.12523 © 2015 Wiley Periodicals, Inc.