Development and Validation of Criteria for Sparing Confirmatory Tests in Diagnosing Primary Aldosteronism

Abstract

Context

The Endocrine Society Guidelines for the diagnosis of primary aldosteronism (PA) suggest that confirmatory tests (CFT) are not required when the following criteria are met: plasma aldosterone concentration (PAC) is >20 ng/dL, plasma renin is below detection levels, and hypokalemia is present. The evidence for the applicability of the guideline criteria is limited.

Objective

To develop and validate optimized criteria for sparing CFT in the diagnosis of PA.

Design and Setting

The optimized criteria were developed in a Chinese cohort using the captopril challenge test, verified by saline infusion test (SIT) and fludrocortisone suppression test (FST), and validated in an Australian cohort.

Participants

Hypertensive patients who completed PA screening and CFT.

Main Outcome Measure

Diagnostic value of the optimized criteria.

Results

In the development cohort (518 PA and 266 non-PA), hypokalemia, PAC, and plasma renin concentration (PRC) were selected as diagnostic indicators by multivariate logistic analyses. The combination of PAC >20 ng/dL plus PRC <2.5 μIU/mL plus hypokalemia had much higher sensitivity than the guideline criteria (0.36 vs 0.11). The optimized criteria remained superior when the SIT or FST were used as CFT. Non-PA patients were not misdiagnosed by either criteria, but the percentage of patients in whom CFT could be spared was higher with the optimized criteria. In the validation cohort (125 PA and 81 non-PA), the sensitivity of the optimized criteria was also significantly higher (0.12 vs 0.02).

Conclusions

Hypertensive patients with PAC >20 ng/dL, PRC <2.5 μIU/mL, plus hypokalemia can be confidently diagnosed with PA without confirmatory tests.

primary aldosteronism, sparing confirmation, optimized criteria, sensitivity, specificity

Primary aldosteronism (PA) is a major cause of secondary hypertension, with a prevalence of 5% to 10% in hypertensive patients (1-4). PA is characterized by excess autonomous aldosterone production with a suppressed plasma renin. The diagnosis of PA includes screening, confirmation, and subtype classification. The plasma aldosterone:renin ratio (ARR) is commonly used for screening. Patients with positive screening are usually required to undertake at least one of the confirmatory tests: saline infusion test (SIT), captopril challenge test (CCT), or fludrocortisone suppression test (FST) (5-8).

Due to the high cost, time requirement, and inconvenience of confirmatory tests, the Endocrine Society Clinical Practice Guideline on PA recommends that patients who fulfill the following criteria (thereafter termed the guideline criteria) do not require confirmatory tests for the formal diagnosis of PA: spontaneous hypokalemia, plasma renin below detection levels, plus plasma aldosterone concentration (PAC) >20 ng/dL (550 pmol/L) (5). However, the guideline criteria lack a strong evidence base. A study from France suggested that screen-positive subjects who had a PAC above 20 ng/dL (550 pmol/L) could be diagnosed as PA without confirmatory tests; however, not all the patients meeting the criteria were correctly diagnosed by the confirmatory test (9). Recently, a study including 327 hypertensive patients from Japan reported that in patients with a positive ARR and plasma renin activity <1 ng/mL/h, if PAC >30 ng/dL, or PAC between 20 and 30 ng/dL plus spontaneous hypokalemia, CCT could be spared (10). However, these studies are limited by being from single centers, with the diagnosis based on only one confirmatory test and the lack of analysis of a range of renin concentrations. More importantly, the plasma renin activity (PRA), rather than plasma renin concentration (PRC), was measured in the study from Japan. As the automated chemiluminescence–based PRC assay is widely used around the world, developing a PRC-based criteria for sparing confirmatory tests is needed in clinical practice.

Our study aimed to establish optimized criteria to spare confirmatory tests in PA diagnosis and compare it with the guideline criteria. The criteria were developed in a large Chinese cohort using CCT as the confirmatory test, verified in patients from the same cohort who also underwent the SIT or FST, and further validated in an Australian cohort using SIT as the confirmatory test.

Methods

Development study

The development study conducted at the First Affiliated Hospital of Chongqing Medical University in China used the database of the Chongqing Primary Aldosteronism Study (CONPASS) (ClinicalTrials.gov: NCT03224312). Hypertensive patients who had completed ARR screening and at least one confirmatory test for PA from November 2013 to November 2019 were included. CONPASS enrolled hypertensive patients who qualified for PA screening according to the Endocrine Society Guidelines as previously described (11-14). The Ethics Committee of the First Affiliated Hospital of Chongqing Medical University approved the protocol. Informed written consent was obtained from each participant.

For PA screening, treatment with diuretics, including mineralocorticoid receptor antagonists, was withdrawn for at least 4 weeks, and β-blockers, angiotensin-converting enzyme inhibitors, and angiotensin-II receptor blockers were stopped for at least 2 weeks. Only non-dihydropyridine calcium channel blockers and/or α-adrenergic blockers were allowed for uncontrolled hypertension. Samples for PRC and PAC were collected in the morning after participants were out of bed for at least 2 hours and after they have been seated for 5 to 15 minutes. The screening test was considered positive when the ARR was ≥2.0 ng·dL⁻¹/μIU·mL⁻¹ (27 pmol·L⁻¹/μIU·mL⁻¹) (12).

Patients who tested positive proceeded to the confirmatory tests as previously described (11). For patients who tested negative (ARR <2.0 ng·dL⁻¹/μIU·mL⁻¹), if PA was strongly suspected based on young age, hypokalemia, or resistant hypertension, they also proceeded to the confirmatory test. PA was confirmed if any of the following criteria was met: (1) The PAC remained >11 ng/dL (305 pmol/L) 2 hours after administration of 50 mg captopril (11); (2) the PAC remained >8 ng/dL (221 pmol/L) after the infusion of 2L normal saline in the recumbent position (11) or >8.5 ng/dL (235 pmol/L) in the seated position (based on unpublished data from our prospective study of 196 PA patients and 73 essential hypertension patients); or (3) the PAC remained >6 ng/dL (221 pmol/L) on the fourth day of fludrocortisone administration. Spontaneous hypokalemia was defined as a serum potassium <3.5 mmol/L before or at the diagnosis of PA, or if a patient was taking an oral potassium supplement.

Patients with a positive result in any confirmatory test underwent an enhanced adrenal computed tomaography (CT) scan. Patients also underwent adrenal vein sampling (AVS) to determine the lateralization of aldosterone hypersecretion if surgery was desired. The diagnosis of unilateral PA was made when patients who underwent surgery achieved complete biochemical resolution based on the Primary Aldosteronism Surgery Outcome (PASO) criteria at the 6- to 12-month post-adrenalectomy follow-up assessment (15) The criteria used for lateralization of aldosterone hypersecretion were established according to the Endocrine Society Guidelines (5). The lateralization index (LI) was defined as aldosterone-cortisol ratio (ACR) (dominant adrenal vein)/ACR (nondominant adrenal vein) and the diagnosis of unilateral PA was made if LI > 2 pre-ACTH and LI > 4 post-ACTH. The selectivity index (SI) was defined as cortisol (adrenal vein) / cortisol (peripheral vein) and successful cannulation of the adrenal veins were defined as SI > 2 pre-ACTH or SI > 3 post-ACTH (16, 17).

Verification study

Among the patients who completed CCT within the development cohort, those who also completed SIT or FST were used to set up 2 additional datasets (SIT-dataset and FST-dataset) to evaluate the applicability of both the guideline criteria and optimized criteria in the context of different confirmatory tests.

Validation study

The validation study was retrospectively conducted in Monash Health in Australia. From January 2010, a standardized PA Diagnostic Protocol was introduced at Monash Health. The database of hypertensive patients referred to Monash Health for PA screening and confirmatory tests from January 2011 to January 2019 was searched, with only patients who had completed recumbent SIT being included. Screening was considered positive when the ARR was >2.5 ng·dL⁻¹/μIU·mL⁻¹ (70 pmol/L: mU/L). Patients who screened positive, or those in whom PA was strongly suspected, proceeded to the recumbent SIT. In this retrospective analysis, PA was diagnosed if the PAC remained >8 ng/dL (221 pmol/L) following saline infusion in the recumbent position. The criteria for unilateral PA diagnosis were consistent with that of the First Affiliated Hospital of Chongqing Medical University in China.

Measurements and assay methods

In both the Chinese and Australian centers, PRC and PAC were measured with an automated chemiluminescence immunoassay (LIAISON; DiaSorin, Italy). For aldosterone, the within-run coefficients of variation were 3.5% at 6.8 ng/dL and 1.8% at 28.8 ng/dL. The total coefficients of variation were 9.5% at 6.8 ng/dL and 5.6% at 28.8 ng/dL. The analytical sensitivity (defined as the minimum detectable concentration that could be distinguished from zero) for the PAC was 2.2 ng/dL. For renin, the within-run coefficients of variation were 12.4% at 13.2 µIU/mL and 4.7% at 260.3 µIU/mL. The total coefficients of variation were 0.6% at 13.2 µIU/mL and 1.7% at 260.3 µIU/mL. The analytical sensitivity for the PRC was 0.53 µIU/mL (the lowest detectable level). Quality control was performed every day in the laboratory.

Statistical analysis

SPSS 23.0 (SPSS Inc., Chicago, IL, USA) and MedCalc 15.2 (MedCalc Software, Ostend, Belgium) was used for statistical analysis. The distribution of the data was analyzed with the Kolmogorov-Smirnov test. Normally distributed variables were expressed as the mean and standard deviation (SD) analyzed by Student’s t test; variables with a skewed distribution were expressed as median (interquartile range [IQR]) analyzed by the Mann-Whitney U test; categorical variables were described as percentages analyzed by the χ² test. Multiple logistic regression analyses was used to identify prediction factors related to diagnosis of PA. To assess the diagnostic accuracy of the criteria, parameters including the sensitivity and specificity were calculated. The bypassing rate was defined as the number of patients who fulfilled the bypassing criteria (optimized or guideline criteria) divided by the number of patients who were diagnosed as PA based on confirmatory tests. P values < 0.05 (2-tailed) were considered statistically significant.

Results

Development of optimized criteria

In total, 784 hypertensive patients with a positive screening test (ARR > 2ng•dL⁻¹/μIU•mL⁻¹) or clinical features consistent with probable PA (hypokalemia and/or resistant hypertension) who had undergone CCT confirmation in the CONPASS database were included in the development cohort. Among them, 518 were diagnosed as PA and 266 as non-PA. Demographic and clinical characteristics of the development cohort were shown in Table 1.

Table 1.

Demographic and Clinical Characteristics of the Patient Cohorts

	Development Cohort			Validation Cohort
Characteristic	Non-PA (n = 266)	PA (n = 518)	P value	Non-PA (n = 81)	PA (n = 125)	P value
Age (years)	52 ± 12	49 ± 12	0.002	55 ± 12	53 ± 12	0.559
Sex (% women)	72%	59%	<0.001	51%	43%	0.319
SBP (mmHg)	148 ± 18	154 ± 20	<0.001	150 (140-160)	154 (140-170)	0.025
DBP (mmHg)	89 ± 14	94 ± 14	<0.001	90 (80-97)	93 (85-100)	0.005
eGFR (ml/min/1.73 m2)	93 (81-109)	89 (75-108)	0.135	94 ± 27	85 ± 24	0.059
Hypokalemia (%)	39%	77%	<0.001	9%	47%	<0.001
Serum K+ (mmol/L)	3.9 (3.5-4.1)	3.3 (2.9-3.8)	<0.001	4.2 (4.0-4.5)	3.8 (3.4-4.2)	<0.001
PAC (ng/dL)	15.3 (11.1-20.5)	28.1 (19.9-43.9)	<0.001	14.33 (10.29-18.47)	20.95 (15.93-34.58)	<0.001
PRC (μIU/mL)	3.6 (2.0-9.0)	2.3 (0.9-5.7)	<0.001	2.7 (2.0-5.0)	3.3 (2.0-7. 0)	0.255
ARR (ng•dL−1/μIU•mL−1)	3.6 (2.0-6.7)	11.5 (4.3-41.3)	<0.001	4.6 (3.4-7.0)	5.9 (3.2-13.6)	0.026
PAC posttest (ng/dL)	8.0 (6.4-9.6)	22.9 (15.4-38.6)	<0.001	5.6 (4.4-6.5)	11.9 (9.3-16.2)	<0.001

PA was diagnosed by PAC posttest (post-CCT-PAC >11 ng/dL and PAC post-SIT >8 ng/dL in Development Cohort and Validation Cohort, respectively. Data were presented as mean ± SD, %, or median (interquartile range).

Abbreviations: ARR, plasma aldosterone/renin ratio; CCT, captopril challenge test; DBP, diastolic blood pressure; eGFR, creatinine based estimated glomerular filtration rate (MDRD); PAC, plasma aldosterone concentration (1 ng/dL = 27.7 pmol/L); PRC, plasma renin concentration; SBP, systolic blood pressure; SIT, saline infusion test.

Multivariate logistic analysis identified 3 parameters as key indicators of confirmed PA: spontaneous hypokalemia (odds ratio [OR] 3.21 [95% CI, 2.23-4.63], P < 0.001); PAC (OR 1.11 [1.09-1.14] for 1-unit increment, P < 0.001), and PRC (OR 0.96 [0.95-0.97] for 1-unit decrement, P < 0.001).

To develop optimized criteria for sparing confirmatory testing, the above 3 parameters were categorically defined. Spontaneous hypokalemia was set as criterion A. PACs were categorically defined as criterion B, which comprised 6 levels of >5 ng/dL (B1), >10 ng/dL (B2), >15 ng/dL (B3), >20 ng/dL (B4), >25 ng/dL (B5) and >30 ng/dL (B6). Suppressed PRCs were categorically defined as criterion C, which comprised 4 levels of <0.53 μIU/mL (C1), <1.0 μIU/mL (C2), <2.5 μIU/mL (C3) and <5.0 μIU/mL (C4). The sensitivity and specificity of the combination of 2 parameters (A + B, A + C, and B + C) or 3 parameters (A + B + C) for PA diagnosis were shown in Table 2 and Appendix Table1 (18). In order to select criteria with optimal clinical relevance, we reasoned that the optimized criteria for sparing confirmatory testing should have 100% specificity, that is, never misdiagnose patients who do not have PA with PA. The sensitivity of the criteria should be high enough to allow more patients to bypass the confirmatory test. Applying these considerations, we found that the optimal combination was A + B4 + C1 (PAC >20 ng/dL, PRC <2.5 μIU/mL and spontaneous hypokalemia), thereafter termed optimized criteria. A total of 185 patients from the development cohort met the optimized criteria and none of them would have been misdiagnosed if confirmatory testing was bypassed, resulting in a specificity of 1.00 (0.97-1.00). The original guideline criteria did not provide a specific value for PRC, likely due to the need to accommodate nonstandardized assay methods for renin concentration or plasma renin activity. In the context of this analysis, however, they equate most closely to criteria A + B + C3 (PAC >20 ng/dL, PRC <0.53 μIU/mL, and spontaneous hypokalemia) where PRC is <0.53μIU/mL. The guideline criteria also resulted in a specificity of 1.00; however, only 58 patients met it. Using the optimized criteria, 36% (185/518) of confirmed PA patients could have bypassed CCT, compared with 11% (58/518) based on the guideline criteria (Table 3). Furthermore, the optimized criteria allowed 48% (69/145) patients with unilateral PA to bypass the CCT compared with only 17% (25/145) using the guideline criteria.

Table 2.

Diagnostic Accuracy of Different Criteria in the Development Cohort

	Criteria	Specificity	Sensitivity	Criteria	Specificity	Sensitivity	Criteria*	Specificity	Sensitivity
A	Hypokalemia	0.61 (0.55-0.67)	0.77 (0.73-0.80)
B	PAC (ng/dL)			A+B1+C1	0.99 (0.96-1.00)	0.13 (0.10-0.16)	A+B4+C1	1.00 (0.98-1.00)	0.11 (0.09-0.14)
B1	>5	0.02 (0.01-0.04)	1.00 (0.99-1.00)	A+B1+C2	0.97 (0.94-0.99)	0.25 (0.21-0.29)	A+B4+C2	1.00 (0.98-1.00)	0.21 (0.18-0.25)
B2	>10	0.18 (0.14-0.23)	0.98 (0.97-0.99)	A+B1+C3	0.89 (0.85-0.93)	0.43 (0.39-0.47)	A+B4+C3	1.00 (0.97-1.00)	0.36 (0.32-0.40)
B3	>15	0.50 (0.44-0.56)	0.88 (0.85-0.91)	A+B1+C4	0.81 (0.76-0.86)	0.56 (0.52-0.61)	A+B4+C4	0.97 (0.94-0.99)	0.46 (0.41-0.50)
B4	>20	0.71 (0.66-0.77)	0.75 (0.71-0.78)	A+B2+C1	0.99 (0.97-1.00)	0.13 (0.10-0.16)	A+B5+C1	1.00 (0.98-1.00)	0.09 (0.07-0.12)
B5	>25	0.88 (0.83-0.91)	0.58 (0.53-0.62)	A+B2+C2	0.97 (0.95-0.99)	0.25 (0.21-0.29)	A+B5+C2	1.00 (0.98-1.00)	0.18 (0.15-0.22)
B6	>30	0.92 (0.88-0.95)	0.45 (0.40-0.49)	A+B2+C3	0.91 (0.87-0.94)	0.42 (0.38-0.47)	A+B5+C3	0.98 (0.97-1.00)	0.30 (0.26-0.34)
C	PRC (μIU/mL)			A+B2+C4	0.84 (0.79-0.88)	0.56 (0.51-0.60)	A+B5+C4	0.98 (0.96-1.00)	0.37 (0.33-0.41)
C1	<0.53	0.95 (0.92-0.98)	0.15 (0.12-0.18)	A+B3+C1	0.99 (0.97-1.00)	0.12 (0.09-0.15)	A+B6+C1	1.00 (0.98-1.00)	0.07 (0.05-0.10)
C2	<1	0.89 (0.85-0.93)	0.29 (0.25-0.33)	A+B3+C2	0.98 (0.96-0.99)	0.23 (0.19-0.27)	A+B6+C2	1.00 (0.98-1.00)	0.15 (0.12-0.18)
C3	<2.5	0.66 (0.60-0.72)	0.53 (0.49-0.57)	A+B3+C3	0.95 (0.92-0.97)	0.39 (0.35-0.44)	A+B6+C3	0.99 (0.97-1.00)	0.25 (0.21-0.29)
C4	<5	0.38 (0.33-0.45)	0.71 (0.67-0.75)	A+B3+C4	0.92 (0.88-0.95)	0.52 (0.48-0.56)	A+B6+C4	0.98 (0.96-1.00)	0.31 (0.27-0.36)

Abbreviations: PAC, plasma aldosterone concentration; PRC, plasma renin concentration.

* Criteria in the bold font were further validated.

Table 3.

Diagnostic Accuracy of Candidate Criteria in Development and Validation Cohort

		Optimized Criteria	Guideline Criteria
Development Cohort	Specificity	1.00 (0.97-1.00)	1.00 (0.98-1.00)
	Sensitivity	0.36 (0.32-0.40)	0.11 (0.09-0.14)
	Bypassing rate*	36% (185/518)	11% (58/518)
Validation Cohort	Specificity	1.00 (0.96-1.00)	1.00 (0.96-1.00)
	Sensitivity	0.12 (0.07-0.19)	0.02 (0.01-0.07)
	Bypassing rate*	12% (15/125)	2% (3/125)

Optimized criteria: PAC>20ng/dL + PRC<2.5 μIU/mL + spontaneous hypokalemia.

Guideline criteria: PAC>20ng/dL + PRC<0.53 μIU/mL + spontaneous hypokalemia.

Abbreviations: PAC, plasma aldosterone concentration; PRC, plasma renin concentration.

* The bypassing rate was defined as the number of patients who met the bypassing criteria (optimized or guideline criteria) divided by the number of patients who were diagnosed as PA based on confirmatory tests.

Verification of optimized criteria

Among the 784 patients who completed CCT within the development cohort, 687 patients also underwent SIT and 539 patients underwent FST. Therefore, patients who completed SIT or FST were used to set up 2 additional datasets (SIT-dataset and FST-dataset) to further evaluate how many patients could be spared SIT and FST if the optimized criteria were used. Based on SIT, 453 were diagnosed as PA and 234 as non-PA. Based on FST, 451 were diagnosed as PA and 88 as non-PA. Demographic and clinical factors of the subjects in the SIT-dataset and FST-dataset were shown in Appendix Table 2 (18).

In the SIT-dataset, significantly more PA patients fulfilled the optimized criteria than the guideline criteria (38% vs 12%) with the same specificity of 1.00 (0.98-1.00). In the FST-dataset, the optimized criteria also applied to more PA patients than the guideline criteria (30% vs 8%) (Appendix Table 3) (18).

Validation of optimized criteria

To evaluate the applicability of the optimized criteria in a different population, we performed validation in an Australian cohort. In total, 206 hypertensive patients who completed ARR screening and SIT confirmation were included in the validation cohort. Among them, 125 were diagnosed as PA and 81 as non-PA. Demographic and clinical factors in the validation cohort were shown in Table 1.

Neither the optimized criteria nor the guideline criteria would have led to the misdiagnosis of non-PA as PA in the validation cohort. However, among the 125 PA patients, 15 met the optimized criteria while only 3 met the guideline criteria (12% vs 2%, P < 0.001) (Table 3). In addition, the optimized criteria allowed 18% (10/55) of patients with unilateral PA to bypass confirmatory tests, compared with only 4% (2/55) using the guideline criteria.

Discussion

We developed a set of optimized criteria for patients who do not require confirmatory testing for the diagnosis of PA and can instead proceed directly to subtyping investigations after initial screening. Setting the optimized criteria as PAC >20 ng/dL, PRC <2.5μIU/mL, and spontaneous hypokalemia, more patients with PA would be allowed to bypass the CCT, SIT, or FST compared with the current Endocrine Society guideline criteria. In the validation study, these optimized criteria were further verified in an independent and geographically distinct patient cohort. The optimized criteria has clear advantages over existing guideline criteria in terms of sensitivity and feasibility with applicability to different populations.

PA is common, but the complicated diagnostic process deters many patients from achieving a diagnosis (10, 19, 20). Screening is the first and easiest step in the diagnosis of PA and provides important information about the key parameters: aldosterone, renin, and potassium. Using the baseline information to confirm PA could significantly simplify the diagnostic pathway and possibly increase the detection of PA. The Endocrine Society clinical practice guidelines recommend the sparing of confirmatory tests if specific criteria are met in patients with typical PA (5). The guideline criteria included increased PAC > 20 ng/dL, an undetectable renin, and a history of spontaneous hypokalemia, all of which are clinical features of florid PA which were verified in our development and validation cohorts. However, in the Chinese cohort, only 8% to 11% of PA patients could have fulfilled the guideline criteria in contrast to 30% to 36% using the optimized criteria. In the Australian cohort, the optimized criteria were also superior to the guideline criteria and would have allowed 12% of PA patients to bypass the SST in comparison to 2%. The impact of these criteria is lower in the Australia cohort compared with the Chinese cohort, due to fewer of the patients having florid hyperaldosteronism; PA patients from China had significantly higher PAC, lower PRC, and lower serum potassium than those from Australia. Importantly, the specificity of the optimized criteria remained at 100% irrespective of the confirmatory test used and the cohort involved, such that none of the non-PA patients would have been misdiagnosed with PA if confirmatory testing was bypassed. Our analysis demonstrates that the optimized criteria for sparing confirmatory tests will impact a greater proportion of patients than the current guideline criteria.

A high PAC is expected in the setting of florid hyperaldosteronism; however, its specificity for PA varies in different studies, ranging from 0.77 to 0.95 (21-23). In our study, the multivariate logistic analyses showed that high PAC was an indicator of confirmed PA; however, PAC alone had limited accuracy. Even when PAC was as high as 30 ng/dL, it only showed a specificity of 0.92 and a sensitivity of 0.45. The addition of a renin threshold improved the specificity for PA; however, an “undetectable renin” as recommended by the guideline criteria lacked sensitivity. Furthermore, although the chemiluminescent immunoassay-based PRC measurement is now widely used around the world (24-28), studies on PRC thresholds for the diagnosis of PA are limited. In one study, when the PRC is <5.2 μIU/mL, a sensitivity of 1.00 and a specificity of 0.90 were obtained for PA diagnosis (29). However, 100% specificity is desired if confirmation testing is to be bypassed. Our study found PRC <2.5 μIU/mL to be more sensitive than an undetectable renin (ie, PRC < 0.50 μIU/mL) but more specific than PRC >5.0 μIU/mL for the accurate diagnosis of PA.

The prevalence of hypokalemia in patients with PA is reported to be 9% to 37% (1, 30). Thus, the presence of hypokalemia alone has low accuracy for the diagnosis of PA. Recently, Kawashima et al published an analysis indicating that confirmatory tests could be spared in patients with PRA <0.6 ng/mL/h (equivalent to PRC <5 μIU/mL) and PAC >30.85 ng/dL (31). When applied to our development cohort, these criteria achieved a lower sensitivity (0.33) and specificity (0.97) compared with our optimized criteria (Appendix Table 1, criteria B6 + C4) (18). However, the Kawashima criteria may be useful for patients who lack a history of hypokalemia and have a detectable baseline renin concentration up to 5 μIU/mL. To achieve a slightly higher specificity of 0.98, with a sensitivity of 0.32, the thresholds of PRC <2.5 μIU/mL and PAC >25 ng/dL can also be used to identify patients who can bypass confirmatory tests (Appendix Table 1, criteria B5 + C3) (18). However, when hypokalemia is combined with PRC and/or PAC, it can be highly specific for the diagnosis of PA, as demonstrated both by our study and the recent Japanese study (10).

To make full use of the clinical data and increase the diagnostic efficacy, a range of combination criteria were developed and evaluated to determine the optimal thresholds for PAC and PRC. Our study found that, in addition to the optimized criteria and guideline criteria, a range of combinations, including hypokalemia plus PRC <0.53 μIU/mL, or PRC <1 μIU/mL plus PAC >25 ng/dL or PAC > 30 ng/dL, were also 100% specific for the diagnosis of PA. However, these combinations displayed low sensitivity. Our analyses demonstrated that the optimized criteria with PAC >20 ng/dL, PRC <2.5 μIU/mL, and spontaneous hypokalemia offered the best trade-off between improved sensitivity and high specificity.

The main strength of our study lies in the large number of subjects who completed a range of different confirmatory tests and the use of a geographically distinct cohort to validate the findings. In addition, our study not only included hypertensive patients with positive screening results, but also patients with negative results (15.3%, 120/784 in the Chinese cohort and 14.6%, 30/206 in the Australian cohort). Such design is in line with clinical practice, as some patients with marked excess aldosterone or spontaneous hypokalemia might have negative screening results but be finally diagnosed as unilateral PA after AVS or surgery (32-34). Furthermore, in our study, the plasma renin level is measured by PRC, which has several advantages over PRA in specimen handling, reproducibility, and standardization. The optimized criteria for sparing confirmatory tests are therefore widely applicable to most centers where the PRC is measured. However, the criteria are limited to medical centers where PAC and PRC are measured by DiaSorin Liaison direct aldosterone and renin chemiluminescent immunoassays rather than liquid chromatography and mass spectrometry or other assays.

Conclusions

In conclusion, we developed the optimized criteria of PAC >20 ng/dL (550 pmol/L), PRC < 2.5 μIU/mL, and spontaneous hypokalemia to identify patients who can bypass confirmatory testing and proceed directly to subtyping investigations. The criteria were validated in Chinese and Australian cohorts. The wider use of the criteria is expected to improve diagnostic efficiency, cost-effectiveness, patient convenience, and ultimately, patient outcomes in the management of primary aldosteronism.

Novelty and Significance

What Is New?

For the diagnosis of primary aldosteronism (PA), we developed a set of optimized criteria to allow patients to bypass the confirmatory tests.

What Is Relevant?

The study results provide valuable information for the diagnosis of PA which is a common cause of endocrine hypertension.

Summary

We developed the optimized criteria of PAC >20 ng/dL (550 pmol/L), PRC <2.5 μIU/mL, and spontaneous hypokalemia to identify patients who can bypass confirmatory testing based in a large Chinese cohort using CCT as the confirmatory test, and verified in patients from the same cohort who also underwent the SIT or FST, and further validated in an Australian cohort using SIT as the confirmatory test. The optimized criteria has clear advantages over existing guideline criteria in terms of sensitivity and feasibility with applicability to different populations.

Abbreviations

Abbreviations
 
• ACTH

  adrenocorticotrophin

 
• ARR

  aldosterone:renin ratio

 
• AVS

  adrenal vein sampling

 
• CCT

  captopril challenge test

 
• CFT

  confirmatory test

 
• FST

  fludrocortisone suppression test

 
• LI

  lateralization index

 
• PA

  primary aldosteronism

 
• PAC

  plasma aldosterone concentration

 
• PRA

  plasma renin activity

 
• PRC

  plasma renin concentration

 
• SI

  selectivity index

 
• SIT

  saline infusion test

Acknowledgments

Membership of the Chongqing Primary Aldosteronism Study (CONPASS) Group: Qifu Li, MD, PhD, Ying Song, MD, Shumin Yang, MD, PhD, Wenwen He, MD, Mei Mei, MD, PhD, Jinbo Hu, PhD, Suxin Luo, MD, PhD, Kangla Liao, MD, Yao Zhang, MD, PhD, Yunfeng He, MD, PhD, Yihong He, MD, Ming Xiao, PhD, Bin Peng, PhD. The authors thank Laboratory of Endocrine and Laboratory of Lipid and Glucose Metabolism, the First Affiliated Hospital of Chongqing Medical University as well.

Financial Support: Provided by the National Natural Science Foundation of China (81670785, 81800701, 81870567, 81800731 and 81970720). Joint Medical Research Project of Chongqing Science and Technology Commission & Chongqing Health and Family Planning Commission (Youth Project, 2018QNXM001). Outstanding Talents of the First Affiliated Hospital of Chongqing Medical University 2019 (2019-4-22). Chongqing Outstanding Youth Funds (cstc2019jcyjjq0006). Victorian Government’s Operational Infrastructure Scheme.

Clinical Trial Information: ClinicalTrials.gov registration no. NCT03224312.

Author Contributions: Conception and design: Q.F. Li, S.M. Yang, J.B. Hu, and J. Yang. Analysis and interpretation of the data: J.B. Hu and K.R. Wang. Drafting of the article: J. Yang, J.B. Hu, S.M. Yang, and K.R. Wang. Critical revision of the article for important intellectual content: P. Fuller, J. Yang, Q.F. Li, and S.M. Yang. Statistical expertise: J.B. Hu. Obtaining of funding: J. Yang, Q.F. Li, S.M. Yang, Y. Song, and J.B. Hu. Administrative, technical, or logistic support: Y. Song and W.W. He. Collection and assembly of data: H. Hashimura, W.W. He, Z.P. Feng, Q.F. Cheng, Z.P. Du, Z.H. Wang, and L.Q. Ma. All authors have read the journal’s authorship agreement and that the manuscript has been reviewed by and approved by all named authors.

Additional Information

Disclosure Summary: Authors have disclosed no conflicts of interest.

Data Availability: Available to approved persons through written agreements with the authors and the data partner. Study protocol and statistical code are available from the authors upon reasonable request.

Author notes

These authors contributed equally to this article as co-first authors.

Shumin Yang and Qifu Li should be considered joint senior authors.