Two continuously distributed numerical definitions were considered as possibly indicating ZG suppression opposite a unilateral PA source.
Definition 1: Comparison vs IVC Measures
Similar to what has been used elsewhere, this definition was based upon the idea that adrenal vein aldosterone levels (normalized to adrenal vein cortisol levels) from a suppressed ZG would be less than the general circulating normalized aldosterone level as measured in the inferior vena cava (IVC). The actual calculation determined the ratio of the Uninvolved Adrenalaldo/cortisol compared to IVCaldo/cortisol with calculations done both pre- and post (1–24)-ACTH infusion.
Definition 2: Relative Suppression of Aldosterone Response to (1–24) ACTH
Adrenocorticotropic hormone is known to stimulate aldosterone production from a normal adrenal in addition to cortisol. Thus, it was hypothesized that if the ZG was truly suppressed relative to the zona fasciculata, there would be an absolute decrease of the Uninvolved Adrenalaldo/cortisol ratio following (1–24) ACTH infusion. The proportional decrease in this aldo/cortisol ratio might then be a quantitative measure of the degree of ZG suppression.
The two definitions of adrenal suppression were retrospectively tested in the subjects of the Calgary Adrenal Vein Sampling Database comprising all patients diagnosed with PA and submitted to AVS from June 2005 to July 2013 (n = 115). This database and analysis were approved by the regional institutional review board. The population demographics, diagnostic criteria, lab methodologies and patient outcomes have been previously published. Briefly, all patients had measurement of the aldosterone–renin ratio (ARR) in upright position, before 1000 h. Antihypertensive medications are not routinely stopped with the exception of mineralocorticoid antagonists which are stopped for 6 weeks prior to testing. Hypokalemia is corrected with oral potassium supplements where necessary. At our institution, an ARR >550 (aldosterone in pmol/l and renin activity in ng/ml/h) has been used to identify patients with high ARR. Depending on the presenting features, patient safety and clinical suspicion, patients with initially normal ARR may undergo repeat ARR measurement after stopping all drugs except calcium channel blockers and alpha blockers. Adrenal imaging and AVS are sequentially performed to characterize the aldosteronism as either unilateral or bilateral with ultimate classification based upon AVS findings. At our institution, AVS is only performed on PA patients who are considered to be surgical candidates. Thus, PA patients who refuse AVS or who have complex, serious medical comorbidities are excluded from the database. Adrenal vein sampling and interpretation was performed according to our previously published protocol. Based upon our previous analysis of within-subject diagnostic consistency of AVS results, successful catheterization is defined as a Cortisoladrenal: CortisolIVC >2:1 prior to cosyntropin and >3:1 post cosyntropin administration. Lateralization is defined as Aldo/cortisol ratio >3:1 between the adrenals at either pre- or post cosyntropin sampling. Criteria defining suppressed contralateral adrenals have not been used as part of the AVS interpretation or clinical decision making. Following the diagnosis, medical or surgical therapy is offered as per AVS results with clinic follow-up at 4–6 week intervals and patient disposition may take the following courses: (i) lost to follow-up, (ii) discharged to primary care physician with BP consistently <140/90 mmHg, (iii) ongoing follow-up because of BP above target or medication difficulties, (iv) re-referred after discharged to primary care with target BP because of recurrent BP or medication issue. Hypertension resolution post adrenalectomy was defined as a blood pressure <140/90 off of all medications.
The criteria for adrenal suppression per Definition 1 and Definition 2 were retrospectively applied to the AVS results of the study cohort, and ROC curve analysis performed to determine the ability of the two definitions to accurately characterize PA as unilateral or bilateral. A comparison of ROC curves was then performed to determine whether there was a significant difference in ROC curves generated by pre- or post (1–24)-ACTH sampled values. The sensitivity and specificity of the ROC curve generated optimal cut-points, and the categorical data from Definition 2 were then calculated. An exploratory analysis was then performed on the surgically treated subgroup of patients (n = 52) to determine whether the definitions of suppression correlated with ultimate medication-free normotension vs persistent hypertension medication requirement. A multiple logistic regression analysis was not attempted given the small size of the database and subsequent high probability of type 2 error.
Standard descriptive statistics were used to compare the population demographics; nonparametrically distributed values were reported as medians with IQR and compared by the Mann–Whitney U-test. The Wilcoxon signed rank test was used to compare the paired repeat measures pre- and post cosyntropin stimulation, and a McNemar's test compared the categorical paired measures pre- and post cosyntropin. Fisher's exact test was used for comparison of categorical diagnostic classifications prior to calculation of sensitivities and specificities. Standard ROC curve analysis was performed using the deLong method. All comparisons were two-sided with α = 0·05 set as the definition of significance. Computations were performed using MedCalc 12.7.0 (MedCalc Software, Ostend, Belgium).