Diagnostic Accuracy of the Aldosterone–to–Active Renin Ratio for Detecting Primary Aldosteronism

Stefan Pilz; Martin H. Keppel; Christian Trummer; Verena Theiler-Schwetz; Marlene Pandis; Valentin Borzan; Matthias Pittrof; Barbara Obermayer-Pietsch; Martin R. Grübler; Nicolas Verheyen; Vinzenz Stepan; Andreas Meinitzer; Jakob Voelkl; Winfried März; Andreas Tomaschitz


J Endo Soc. 2019;3(9):1748-1758. 

In This Article


The participant flow chart of the GECOH study is shown in Figure 1. In brief, of ~440 patients who were approached for study inclusion, 400 agreed to participate in the GECOH study. The SIT was performed regardless of AARR values in 167 of the first 200 study participants; 33 of the first 200 study participants declined to undergo the SIT. Afterward, the SIT was exclusively performed in participants with a positive AARR (i.e., AARR ≥ 3.7 ng/dL/μU/mL), including a PAC ≥9 ng/dL. Because of insufficient funding for this study, a second AARR determination was routinely performed in the first consecutive 268 participants only (resulting in 231 available values) and was thereafter not measured except in participants who had undergone a SIT. In the entire study, two determinations of the AARR with a median (interquartile range) time interval in between of 30 days (22 to 38 days) were available in 241 participants, and the SIT was performed in 181 participants. We excluded 18 participants who declined to undergo further testing to exclude or confirm PA. Therefore, 382 were finally eligible for analyses because they had sufficient data to exclude or confirm the diagnosis of PA with consistently negative (normal) AARR or any positive AARR, plus the result of the SIT. PA was diagnosed in patients with a positive AARR and a positive result of the SIT, whereas PA was excluded if results of only one or none of these two tests was positive. We did not impute for missing data, and no severe adverse events resulted from the index or reference standard test.

Figure 1.

Participant flow chart for the GECOH study.

Baseline characteristics of all eligible study participants and stratified according to the presence or absence of PA are shown in Table 1. PA was diagnosed in 18 of the 382 participants (4.7%). In addition, two patients were diagnosed with pheochromocytoma on the basis of elevated plasma metanephrines and/or normetanephrines and two were diagnosed with hypercortisolism.[26,27] Hypercortisolism was diagnosed according to the Endocrine Society guideline (i.e., pathologic result on a 1-mg dexamethasone suppression test plus elevated midnight salivary cortisol concentration); one patient had ACTH-dependent and one had ACTH-independent hypercortisolism.[27] These patients were not excluded from the ROC analyses.

The contingency table for the diagnostic accuracy of a positive AARR at the first study visit (at baseline), and for a positive SIT result (i.e. PAC ≥ 10 ng/dL after the SIT) is shown in Table 2. The area under the ROC curve for the AARR at the first study visit in detecting PA was 0.973 (95% CI, 0.956 to 0.990) (Figure 2). The point on the ROC curve with the minimum distance from the left-upper corner of the unit square was at an AARR of 4.33 ng/dL/μU/mL with a sensitivity of 100% and a specificity of 91.8% (Figure 2). The sensitivity, specificity, PPV, NPV, and positive and negative likelihood ratios (with 95% CIs) of a positive AARR in detecting PA were 100% (81.5% to 100.0%), 89.6% (86.0% to 92.5%), 32.1% (26.0% to 39.0%), 100%, 9.58 (7.09 to 12.94), and 0, respectively. The respective area under the ROC curve for the PAC concentration after the SIT was 0.998 (95% CI, 0.994 to 1.000). The sensitivity, specificity, PPV, NPV, and positive and negative likelihood ratios (with 95% CIs) of a positive SIT result for detecting PA were 100% (81.5% to 100.0%), 96.1% (90.6% to 97.9%), 69.2% (53.4% to 81.6%), 100%, 20.4 (10.4 to 40.1), and 0, respectively. Characteristics of patients without PA who had a positive AARR at the first study visit or a positive SIT result are shown in Table 3. Of note, among participants without PA, the PAC after the SIT (median with interquartile ranges) did not significantly different between those with (n = 38) and without (n = 125) a positive baseline AARR [5.7 (4.6 to 7.4) vs 5.2 (3.8 to 7.0) ng/dL; P = 0.071].

Figure 2.

ROC curve for the AARR in detecting PA. The point on the ROC curve with the minimum distance from the left-upper corner of the unit square (marked with a circle) was at an AARR of 4.33 ng/dL/μU/mL, with a sensitivity of 100% and a specificity of 91.8%.

Some of the prespecified secondary outcomes could not be analyzed because we did not have funding to measure PAC by liquid chromatography mass spectrometry or 24-hour urine aldosterone concentrations with subsequent comparisons of the ROC curves using these parameters with the ROC curve of the AARR in detecting PA. Furthermore, we did not compare the ROC curves before (first AARR determination) and after (second AARR determination) discontinuation of β-blockers because the study participants were largely unwilling to alter their antihypertensive treatment; only 10 patients receiving β-blocker therapy stopped this treatment before the second visit. Of note, in all GECOH study participants with two available AARR measurements (n = 241), values at the first measurement did not significantly differ from those at the second measurement [1.58 (0.67 to 3.51) vs 1.54 (0.66 to 3.38) ng/dL/μU/mL; P = 0.685; Spearman correlation coefficient, 0.841; P < 0.001].

Although this was not prespecified in the study protocol, according to previously defined criteria, the six patients in the GECOH study with an aldosterone-producing adenoma who were subsequently treated by unilateral adrenalectomy showed a complete biochemical success; four of these patients had a complete clinical success and two had a partial clinical success (data not shown).[28] Follow-up data for the remaining 12 patients with PA who were treated with mineralocorticoid receptor–blocker therapy are not reported because there are no clear criteria for medical treatment success of PA; thus, we did not perform a prespecified follow-up of the patients.[29]