Defining Contralateral Adrenal Suppression in Aldosteronism
Defining Contralateral Adrenal Suppression in Aldosteronism
For this study, four subjects were excluded because of failed AVS and 12 subjects were excluded after they showed lateralization on AVS but ultimately refused surgery (n = 9) or were lost to follow-up (n = 3); these were excluded as their final diagnosis and outcomes could not be definitively verified. Thus, there were 99 subjects available for analysis with follow-up counted to the last visit as of 1 July 2013. The study population demographics, AVS results and blood pressure outcomes are summarized in Table 1 and Fig. 1. At presentation and initial testing, 39% were using beta blockers, 41% using calcium channel blockers, 43% on either ACEI or ARB, 27% using a diuretic and 8% using other agents such as vasodilators or alpha blockers. In 76 patients (64% of whom had hypokalemia at presentation and 64% had concomitant CT adrenal masses), the PA diagnosis was made after a single ARR measurement and 23 subjects had the PA diagnosis made after a second, medication-adjusted ARR was performed.
(Enlarge Image)
Figure 1.
Flowchart of patient adrenal vein sampling and blood pressure outcomes according to unilateral or bilateral PA diagnosis. Lateralized patients who ultimately refused surgery excluded because unilateral PA could not be verified.
The distribution of subject diagnostic categorization according to Definition 1 (proportional suppression of the uninvolved adrenal aldo/cort vs IVC) is shown in Fig. 2. This definition proved to be a strong predictor of lateralization as shown in Fig. 3 (ROC curves). This was true for both baseline and post cosyntropin samples (AUC 0·834 and 0·958, respectively, both P < 0·0001). Comparison of the two ROC curves demonstrated superior performance of the post cosyntropin values (P = 0·011 vs baseline samples). A ROC-optimized post cosyntropin uninvolved adrenalA/C:IVCA/C of <1·4 had a sensitivity of 90% and specificity of 94% with a Youden index J of 0·84. With our local unilateral PA prevalence of 52·5% in those going to AVS, the positive predictive value of an uninvolved adrenalA/C:IVCA/C of <1·4 was 94%.
(Enlarge Image)
Figure 2.
Distribution of results from adrenal vein sampling showing suppressed adrenalA/C: IVCA/C at baseline and post (1–24)-ACTH according to final diagnosis of unilateral aldosteronism. Abbreviations: Lat, lateralized; Nonlat, nonlateralized; Base, baseline; Stim; post cosyntropin IVC, inferior vena cava; A/C, aldosterone-to-cortisol ratio.
(Enlarge Image)
Figure 3.
Receiver-operator characteristics curve analysis of adrenal vein sampling results for the suppressed/lower adrenalA/C: IVCA/C at baseline and post (1–24) ACTH to diagnose unilateral aldosteronism.
For Definition 2 (absolute decrease of the A/C ratio of the uninvolved/lower adrenal post cosyntropin), it may be seen in Fig. 4 that any decrease in the uninvolved A/C ratio did not appear to separate lateralized from nonlateralized PA patients (P = 0·65). However, a more restrictive definition of absolute A/C decrease to <1·0 post cosyntropin did have a specificity of 100% although sensitivity was just 66% (Fisher's exact test P < 0·0001 vs nonlateralized group).
(Enlarge Image)
Figure 4.
Distribution of results from adrenal vein sampling showing the proportion of lateralized vs nonlateralized subjects whose uninvolved/lowest adrenalA/C decreased following (1–24) ACTH bolus.
The distribution of surgical subject outcomes according to Definition 1 and Definition 2 classification is seen in Fig. 5. This demonstrates that baseline suppression of the uninvolved adrenalA/C:IVCA/C did not correlate with medication-free outcome although it is noted that almost 60% of those with medication-free, postoperative normalization of blood pressure were seen in subjects whose post cosyntropin uninvolved adrenal aldo/cortisol ratio was <0·5 vs IVC and this showed a trend towards significance compared to those whose postcortrosyn uninvolved adrenal aldo/cortisol ratio was >1·5 vs IVC (P = 0·08). Definition 2 showed a positive trend towards association with hypertension outcome (Fig. 5b) although there was substantial overlap between the two outcome groups, and ROC curve analyses did not show statistically significant discriminative ability (data not shown).
(Enlarge Image)
Figure 5.
a (top), b (bottom): Adrenal vein sampling results from surgically treated patients according to both definitions of adrenal suppression (see text) relative to post-operative hypertension resolution status.
Results
For this study, four subjects were excluded because of failed AVS and 12 subjects were excluded after they showed lateralization on AVS but ultimately refused surgery (n = 9) or were lost to follow-up (n = 3); these were excluded as their final diagnosis and outcomes could not be definitively verified. Thus, there were 99 subjects available for analysis with follow-up counted to the last visit as of 1 July 2013. The study population demographics, AVS results and blood pressure outcomes are summarized in Table 1 and Fig. 1. At presentation and initial testing, 39% were using beta blockers, 41% using calcium channel blockers, 43% on either ACEI or ARB, 27% using a diuretic and 8% using other agents such as vasodilators or alpha blockers. In 76 patients (64% of whom had hypokalemia at presentation and 64% had concomitant CT adrenal masses), the PA diagnosis was made after a single ARR measurement and 23 subjects had the PA diagnosis made after a second, medication-adjusted ARR was performed.
(Enlarge Image)
Figure 1.
Flowchart of patient adrenal vein sampling and blood pressure outcomes according to unilateral or bilateral PA diagnosis. Lateralized patients who ultimately refused surgery excluded because unilateral PA could not be verified.
Prediction of Lateralization
The distribution of subject diagnostic categorization according to Definition 1 (proportional suppression of the uninvolved adrenal aldo/cort vs IVC) is shown in Fig. 2. This definition proved to be a strong predictor of lateralization as shown in Fig. 3 (ROC curves). This was true for both baseline and post cosyntropin samples (AUC 0·834 and 0·958, respectively, both P < 0·0001). Comparison of the two ROC curves demonstrated superior performance of the post cosyntropin values (P = 0·011 vs baseline samples). A ROC-optimized post cosyntropin uninvolved adrenalA/C:IVCA/C of <1·4 had a sensitivity of 90% and specificity of 94% with a Youden index J of 0·84. With our local unilateral PA prevalence of 52·5% in those going to AVS, the positive predictive value of an uninvolved adrenalA/C:IVCA/C of <1·4 was 94%.
(Enlarge Image)
Figure 2.
Distribution of results from adrenal vein sampling showing suppressed adrenalA/C: IVCA/C at baseline and post (1–24)-ACTH according to final diagnosis of unilateral aldosteronism. Abbreviations: Lat, lateralized; Nonlat, nonlateralized; Base, baseline; Stim; post cosyntropin IVC, inferior vena cava; A/C, aldosterone-to-cortisol ratio.
(Enlarge Image)
Figure 3.
Receiver-operator characteristics curve analysis of adrenal vein sampling results for the suppressed/lower adrenalA/C: IVCA/C at baseline and post (1–24) ACTH to diagnose unilateral aldosteronism.
For Definition 2 (absolute decrease of the A/C ratio of the uninvolved/lower adrenal post cosyntropin), it may be seen in Fig. 4 that any decrease in the uninvolved A/C ratio did not appear to separate lateralized from nonlateralized PA patients (P = 0·65). However, a more restrictive definition of absolute A/C decrease to <1·0 post cosyntropin did have a specificity of 100% although sensitivity was just 66% (Fisher's exact test P < 0·0001 vs nonlateralized group).
(Enlarge Image)
Figure 4.
Distribution of results from adrenal vein sampling showing the proportion of lateralized vs nonlateralized subjects whose uninvolved/lowest adrenalA/C decreased following (1–24) ACTH bolus.
Prediction of Hypertension Outcome
The distribution of surgical subject outcomes according to Definition 1 and Definition 2 classification is seen in Fig. 5. This demonstrates that baseline suppression of the uninvolved adrenalA/C:IVCA/C did not correlate with medication-free outcome although it is noted that almost 60% of those with medication-free, postoperative normalization of blood pressure were seen in subjects whose post cosyntropin uninvolved adrenal aldo/cortisol ratio was <0·5 vs IVC and this showed a trend towards significance compared to those whose postcortrosyn uninvolved adrenal aldo/cortisol ratio was >1·5 vs IVC (P = 0·08). Definition 2 showed a positive trend towards association with hypertension outcome (Fig. 5b) although there was substantial overlap between the two outcome groups, and ROC curve analyses did not show statistically significant discriminative ability (data not shown).
(Enlarge Image)
Figure 5.
a (top), b (bottom): Adrenal vein sampling results from surgically treated patients according to both definitions of adrenal suppression (see text) relative to post-operative hypertension resolution status.
Source...