Construction of a predictive scoring system as a guide to screening and confirmation of the diagnosis of primary aldosteronism

Development of a risk-prediction model for primary aldosteronism in veterans with hypertension

BACKGROUND

Rates of screening for primary aldosteronism in patients who meet the criteria are exceedingly low (1%-3%). To help clinicians prioritize screening in patients most likely to benefit, we developed a risk-prediction model.

METHODS

Using national Veterans Health Administration data, we identified patients who met the criteria for primary aldosteronism screening between 2000 and 2019. We performed multivariable logistic regression to identify characteristics associated with positive primary aldosteronism testing before generating a risk-scoring system based on the coefficients (0< β < 0.5 = 1 pt, 0.5 ≤ β < 1 = 2 pts, 1 ≤ β < 1.5 = 3 pts) and then tested the system performance using an internal validation cohort.

RESULTS

We identified 502,190 patients who met primary aldosteronism screening criteria, of whom 1.6% were screened and 15% tested positive. Based on the regression model, we generated a risk-scoring system based on a total of 9 possible points in which age under 50, absence of smoking history, and resistant hypertension each scored 1 point; elevated serum sodium 2 points; and hypokalemia 3 points. Rates of positive screening increased with risk score, with 5.6% to 6.7% of those scoring 0 points testing positive; 7.9% to 9.0% 1 point; 8.6% to 10% 2 points; 13% to 14% 3 points; 21% 4 points; 22% to 38% 5 points; 27% to 38% 6 points; 42% to 49% 7 points; and 50% to 51% ≥8 points.

CONCLUSION

In hypertensive patients who meet the criteria for primary aldosteronism screening, rates of positive screening range from 5.6% to 51%. Use of our risk-predication model incorporating these factors can identify patients most likely to benefit from testing.

A novel clinical-imaging nomogram for predicting primary aldosteronism in patients with hypertension

This study aimed to develop and validate the accuracy of a clinical-imaging index nomogram in predicting primary aldosteronism (PA) in patients with hypertension. This case-control study enrolled 404 hypertension patients in the First Affiliated Hospital of Nanjing Medical University, China, from April 2017 to September 2021. The patients were randomly divided into the training set (n = 283, 70%) and the validation set (n = 121, 30%). Univariate and multivariate logistic regression analyses were performed to identify independent predictors of PA, which were then used construct a nomogram. The receiver operating characteristic (ROC) curve and calibration plot were drawn to assess the predictive value. The accuracies of our nomogram and other known prediction models were compared using decision curve analyses (DCA). Four significant variables (history of hypokalemia [OR = 2.684, 95% CI: 1.281-5.623, P < 0.001], typical imaging feature [OR = 2.316, 95% CI: 1.166-4.601, P = 0.003], 24 h urine potassium [OR = 0.956, 95% CI: 0.932-0.980, P < 0.001], plasma renin activity [PRA] [OR = 1.423, 95% CI: 1.161-1.744, P < 0.001]) in the multivariate logistic regression analysis were sifted out, and used to build the nomogram. The predictive nomogram yielded an AUC of 0.890 (95% CI, 0.853-0.927) in the training set and 0.860 (95% CI, 0.793-0.927) in the validation set. Predicted and actual probability of PA matched well in the nomogram. Moreover, the DCA showed that the nomogram gained a net benefit in clinical practice in predicting PA when the threshold value was set between 0.1 and 1.0. Our four-variable nomogram was accurate in predicting PA patients and might be introduced into clinical management.

Predictive performance of aldosterone-to-renin ratio in the diagnosis of primary aldosteronism in patients with resistant hypertension

Background

The systematic use of confirmatory tests in the diagnosis of primary aldosteronism (PA) increases costs, risks and complexity to the diagnostic work-up. In light of this, some authors proposed aldosterone-to-renin (ARR) cut-offs and/or integrated flow-charts to avoid this step. Patients with resistant hypertension (RH), however, are characterized by a dysregulated renin-angiotensin-aldosterone system, even in the absence of PA. Thus, it is unclear whether these strategies might be applied with the same diagnostic reliability in the setting of RH.

Methods

We enrolled 129 consecutive patients diagnosed with RH and no other causes of secondary hypertension. All patients underwent full biochemical assessment for PA, encompassing both basal measurements and a saline infusion test.

Results

34/129 patients (26.4%) were diagnosed with PA. ARR alone provided a moderate-to-high accuracy in predicting the diagnosis of PA (AUC=0.908). Among normokalemic patients, the ARR value that maximized the diagnostic accuracy, as identified by the Youden index, was equal to 41.8 (ng/dL)/(ng/mL/h), and was characterized by a sensitivity and a specificity of 100% and 67%, respectively (AUC=0.882); an ARR > 179.6 (ng/dL)/(ng/mL/h) provided a 100% specificity for the diagnosis of PA, but was associated with a very low sensitivity of 20%. Among hypokalemic patients, the ARR value that maximized the diagnostic accuracy, as identified by the Youden index, was equal to 49.2 (ng/dL)/(ng/mL/h), and was characterized by a sensitivity and a specificity of 100% and 83%, respectively (AUC=0.941); an ARR > 104.0 (ng/dL)/(ng/mL/h) provided a 100% specificity for the diagnosis of PA, with a sensitivity of 64%.

Conclusions

Among normokalemic patients, there was a wide overlap in ARR values between those with PA and those with essential RH; the possibility to skip a confirmatory test should thus be considered with caution in this setting. A better discriminating ability could be seen in the presence of hypokalemia; in this case, ARR alone may be sufficient to skip confirmatory tests in a suitable percentage of patients.

Subtyping primary aldosteronism by inconclusive adrenal vein sampling: A derivation and validation study in a tertiary centre

OBJECTIVE

Indices based on aldosterone/cortisol (A/C) concentration in the successfully cannulated adrenal vein (AV) and in the inferior vena cava (IVC) (AV/IVC) appear to be possible markers to verify the subtype of primary aldosteronism (PA) in the case of inconclusive results of adrenal vein sampling (AVS). The variability of results in previous studies encouraged us to calculate AV/IVC and adrenal A/C cutoff values that could predict the aetiology of PA.

METHODS

This retrospective study included 96 patients who underwent AVS due to PA between 2015 and 2020. The derivation cohort ultimately consisted of 60 patients with bilaterally successful AVS and a clear diagnosis of unilateral or bilateral disease. Receiver operating characteristic analysis was used to find the optimal A/C and AV/IVC cutoff values predicting the subtype of PA. The validation cohort consisted of 11 patients with either unsuccessful cannulation or a borderline lateralization index (LI), those patients underwent adrenalectomy because their indices were suggestive of unilateral disease based on the derivation cohort data.

RESULTS

The cutoff values of A/C ≤ 0.63 or AV/IVC ≤ 0.37 identified unaffected glands with a sensitivity of 91.2% and 97.1%, respectively, and a specificity of 90.7% and 88.4%, respectively. Unilateral ipsilateral gland involvement was characterized by A/C ≥ 3.5 or AV/IVC ≥ 3.4 with a corresponding specificity of 100%. All patients in the validation cohort achieved biochemical remission postoperatively.

CONCLUSIONS

A/C and AV/IVC cutoff values could be a useful tool to determine the subtype of PA in patients with unilateral successful AVS as well as in patients with a borderline LI.

Clinical and biochemical predictors and predictive model of primary aldosteronism

Background

Primary aldosteronism (PA) is the most common cause of secondary hypertension. The diagnosis of PA currently requires multiple complicated measures. The aims of this study were to identify easy-to-obtain clinical and biochemical predictors, and to create predictive model to facilitate the identification of a patient at high risk of having PA.

Materials and methods

This 2-year retrospective cohort study was conducted at a tertiary care medical center. A total of 305 patients who had been tested for plasma aldosterone concentration (PAC) and plasma renin activity (PRA) were identified. Patients with incomplete results of PAC and PRA and those who had an established diagnosis of Cushing’s syndrome or pheochromocytoma were excluded. Logistic regression analysis was used to identify significant predictors and to create predictive model of PA.

Results

PA was diagnosed in 128 of the patients (41.96%). Significant predictive factors for PA were age >60 years (OR 2.12, p = 0.045), female (OR 1.65, p<0.001), smoking (OR 2.79, p<0.001), coronary artery disease (OR 2.29, p<0.001), obstructive sleep apnea (OR 1.50, p = 0.017), systolic blood pressure >160 mmHg (OR 1.15, P<0.001), serum potassium <3 mEq/L (OR 3.72, p = 0.030), fasting blood glucose >126 mg/dL (OR 0.48, p = 0.001) and estimated glomerular filtration rate (eGFR) <60 mL/min/1.73m2 (OR 1.79, p = 0.001). Predictive model was created with a total score ranged from 0 to 42. A score above 7.5 indicated a higher probability of having PA with a sensitivity of 72% and a specificity of 70%. The diagnostic performance of the predictive model based on area under the curve was 71%.

Conclusions

The clinical and biochemical predictive factors including predictive model identified in this study can be employed as an additional tool to help identify patients at risk of having PA and could help reduce the number of screening and confirmation tests required for PA.

InCASEOf scoring system for distinction between pachychoroid-associated macular neovascularization and neovascular age-related macular degeneration in patients older than 50 years

To develop a novel scoring system aiming at guiding the differential diagnosis between macular neovascularization secondary to pachychoroid disease (pMNV) and neovascular age-related macular degeneration (AMD) in patients aged 50 years and older. In this retrospective study performed at University Vita-Salute San Raffaele (Milan, Italy) and Créteil University Eye Clinic (Créteil, France), we enrolled patients 50 years of age and older, visited between January 2017 and January 2019, who were diagnosed with either treatment-naïve pMNV or neovascular AMD. At the time of diagnosis, all patients underwent a comprehensive ophthalmologic evaluation, spectral-domain optical coherence tomography, fluorescein angiography, indocyanine green angiography, and optical coherence tomography angiography. Univariate comparison between pMNV and neovascular AMD groups was performed to identify the main clinical predictors for pMNV. The selected predictors were taken into a binomial logistic regression and eventually served as the basis for the development of InCASEOf scoring system. Receiver operating characteristic (ROC) curves were used to study the model performance. Forty-eight right eyes from 48 patients with pMNV and 39 right eyes from 39 patients with neovascular AMD were considered in this study. Age (+ 2 points), sex (+ 2 points), choroidal thickness (+ 2 points), early pachyvessels (+ 2 points), and evidence of MNV at OCTA (+ 3 points) turned out to be predictors for pMNV. Four additional factors significant at univariate analysis were considered: type 2 and type 3 MNVs and presence of intraretinal fluid (− 0.5 points each), and presence of subretinal fluid (+ 0.5 points). InCASEOf scoring system was built with a high score of 11.5 points. The cutoff value of 6.5 showed good accuracy in separating pMNVs from neovascular AMDs. InCASEOf is a straightforward clinical scoring system, accessible to comprehensive ophthalmologists, with the purpose of enabling easy distinction and expert-like diagnosis of pMNV and neovascular AMD in patients aged 50 years or older.

Clinical Score and Machine Learning-Based Model to Predict Diagnosis of Primary Aldosteronism in Arterial Hypertension

[Figure: see text].

Development and validation of a novel diagnostic nomogram model to predict primary aldosteronism in patients with hypertension

PURPOSE

Primary aldosteronism (PA) remains, to a large extent, an under-diagnosed disease. We aimed to develop and validate a novel clinical nomogram to predict PA based on routine biochemical variables including new ones, calcium-phosphorus product.

METHODS

Records from 806 patients with hypertension were randomly divided into 70% (n = 564) as the training set and the remaining 30% (n = 242) as the validation set. Predictors for PA were extracted to construct a nomogram model based on regression analysis of the training set. An internal validation was performed to assess the nomogram model's discrimination and consistency using the area under the curve for receiver operating characteristic curves and calibration plots. The diagnostic accuracy was compared between nomogram and other known prediction models, using receiver operating characteristics (ROC) and decision curve analyses (DCA).

RESULTS

Female gender, serum potassium, serum calcium-phosphorus product, and urine pH were adopted as predictors in the nomogram. The nomogram resulted in an area under the curve of 0.73 (95% confidence interval: 0.68-0.78) in the training set and an area under the curve of 0.68 (0.59-0.75) in the validation set. Predicted probability and actual probability matched well in the nomogram (p > 0.05). Based on ROC and DCA, 21-70% threshold to predict PA in the nomogram model was clinically useful.

CONCLUSIONS

We have developed a novel nomogram to predict PA in hypertensive individuals based on routine biochemical variables. External validation is needed to further demonstrate its predictive ability in primary care settings.

Development of a Prediction Score to Avoid Confirmatory Testing in Patients With Suspected Primary Aldosteronism

CONTEXT

The diagnostic work-up of primary aldosteronism (PA) includes screening and confirmation steps. Case confirmation is time-consuming, expensive, and there is no consensus on tests and thresholds to be used. Diagnostic algorithms to avoid confirmatory testing may be useful for the management of patients with PA.

OBJECTIVE

Development and validation of diagnostic models to confirm or exclude PA diagnosis in patients with a positive screening test.

DESIGN, PATIENTS, AND SETTING

We evaluated 1024 patients who underwent confirmatory testing for PA. The diagnostic models were developed in a training cohort (n = 522), and then tested on an internal validation cohort (n = 174) and on an independent external prospective cohort (n = 328).

MAIN OUTCOME MEASURE

Different diagnostic models and a 16-point score were developed by machine learning and regression analysis to discriminate patients with a confirmed diagnosis of PA.

RESULTS

Male sex, antihypertensive medication, plasma renin activity, aldosterone, potassium levels, and the presence of organ damage were associated with a confirmed diagnosis of PA. Machine learning-based models displayed an accuracy of 72.9%-83.9%. The Primary Aldosteronism Confirmatory Testing (PACT) score correctly classified 84.1% at training and 83.9% or 81.1% at internal and external validation, respectively. A flow chart employing the PACT score to select patients for confirmatory testing correctly managed all patients and resulted in a 22.8% reduction in the number of confirmatory tests.

CONCLUSIONS

The integration of diagnostic modeling algorithms in clinical practice may improve the management of patients with PA by circumventing unnecessary confirmatory testing.

Performance of the unilateral AV/IVC index in primary hyperaldosteronism subtype prediction: A validation study in a single tertiary centre

BACKGROUND

Failure of cannulation of the right adrenal vein is frequent during AVS for investigation of primary aldosteronism (PA). The aldosterone:cortisol ratio of either adrenal vein compared with the inferior vena cava (AV/IVC index) has been proposed to differentiate between unilateral and bilateral disease, and aid in lateralization of unilateral disease.

METHODS

Sixty-two patients with unilateral or bilateral PA identified by either successful bilateral (45 patients) or unilateral (17 patients) adrenal vein cannulation, and with biochemical remission following surgery were enrolled into the analysis. The diagnostic performances of the previously identified AV/IVC index cut-offs of ≥5.5 to predict ipsilateral disease and ≤0.5 to predict contralateral disease were validated using data from the entire cohort.

RESULTS

Fifty-three patients had unilateral PA and 9 patients bilateral PA. The area under ROC curve (AUROC) of the AV/IVC cut-off ≤0.5 for identifying unilateral aldosterone secretion from the contralateral adrenal was 0.95 (95% CI; 0.88-0.99), whereas the AUROC of the AV/IVC cut-off ≥5.5 for identifying unilateral aldosterone secretion from ipsilateral adrenal was 0.96 (95% CI; 0.92-0.99). The AV/IVC index cut-off value of 0.5 had 93% sensitivity and 91% specificity, and the AV/IVC index cut-off value of 5.5 had 21% sensitivity and 100% specificity. The optimal AV/IVC cut-offs to achieve 100% specificity for our cohort were >2.4 and <0.1 to predict ipsilateral and contralateral disease.

CONCLUSION

Our data confirm that the AV/IVC index is a potential tool for subtype classification and lateralization in patients with PA in the setting of failed bilateral, but successful unilateral, adrenal vein cannulation during AVS.