📄 medicine

Development, Temporal External Validation, and Updating of a Nomogram for Predicting Hypertension Cure After Unilateral Adrenalectomy in Primary Aldosteronism

This study developed and temporally validated a recalibrated four-variable nomogram using preoperative clinical data to accurately predict hypertension cure rates following unilateral adrenalectomy in Chinese patients with primary aldosteronism, addressing the limitations of existing models in Asian populations.

Original authors: Yuchao Huangfu, Mingchuan Li, Yiqing Du, Zixiong Huang, Ning Zhang, Tao xu

Published 2026-10-01

📖 3 min read☕ Coffee break read

CC BY 4.0

Original authors: Yuchao Huangfu, Mingchuan Li, Yiqing Du, Zixiong Huang, Ning Zhang, Tao xu

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

High blood pressure is a common condition, often managed with daily medication, but for some people, the root cause lies in a small, overactive gland sitting atop the kidney. This gland, called the adrenal gland, can sometimes produce too much of a hormone known as aldosterone. When this happens in excess, it forces the body to hold onto salt and water, driving blood pressure up and putting extra strain on the heart and blood vessels. This specific condition is called primary aldosteronism. While medication can control the symptoms, the most effective way to fix the problem is often to remove the single, rogue gland causing the trouble through a surgical procedure. However, surgery is not a guaranteed fix for the high blood pressure itself. Even after the gland is gone, many patients find they still need to take pills to keep their blood pressure in check, while others see their numbers return to normal without any help at all. For doctors and patients standing at the crossroads of deciding whether to operate, knowing who will actually be cured is a difficult puzzle.

A team of researchers from two major hospitals in Beijing set out to solve this puzzle by creating a new tool to predict the outcome of this surgery. They looked back at the medical records of 112 patients who had undergone the procedure between 2018 and 2021. By studying these cases, they identified four specific factors that were strongly linked to whether a patient would be cured of their high blood pressure: the patient's age, the measurement of their waist, the highest blood pressure reading they had ever recorded, and whether they had a history of heart or stroke-related issues. Using these four pieces of information, they built a visual guide, known as a nomogram, which allows a doctor to calculate the specific chance of a successful cure for an individual patient before the surgery even begins.

The researchers then tested this new guide on a fresh group of 75 patients who had the surgery in 2023. This step was crucial because tools that work perfectly on the group used to create them often fail when applied to new people. When they first tried their original guide on this new group, it did not perform as well as hoped; it tended to be too optimistic, predicting cures that did not happen. The researchers realized that one of the factors they had included, the length of time a patient had suffered from high blood pressure, was not actually helping to predict the outcome in this new group. They removed this factor and adjusted the guide to fit the new data more accurately. The result was a streamlined, four-factor model that proved much more reliable.

This updated tool showed a significant improvement over older methods that had been used for years, particularly those developed for Western populations. The new guide, which relies only on information a doctor can easily gather during a standard pre-operative visit, offers a clearer picture of what to expect. It does not promise a cure for everyone, but it provides a realistic estimate based on the patient's specific history and physical traits. The researchers emphasize that while this tool is a strong step forward for patients in China, it still needs to be tested in different hospitals and regions to ensure it works just as well for everyone. Until then, it stands as a refined, practical method for helping doctors and patients make informed decisions about whether surgery is the right path to a life without high blood pressure medication.

Technical Summary: Development, Temporal External Validation, and Updating of a Nomogram for Predicting Hypertension Cure After Unilateral Adrenalectomy in Primary Aldosteronism

Problem Statement
Primary Aldosteronism (PA) is the leading cause of secondary hypertension. While unilateral adrenalectomy is the standard treatment for unilateral PA and achieves biochemical cure, the resolution of hypertension post-surgery is highly variable, with clinical cure rates ranging from 30% to over 70% across studies. Existing predictive tools, such as the Aldosteronoma Resolution Score (ARS), demonstrate inconsistent performance, particularly in Asian populations. Furthermore, clinical prediction models often suffer from "optimism bias," performing well in derivation cohorts but failing to generalize to new patient populations due to temporal drift and case-mix differences. There is a lack of models that have undergone the full cycle of development, temporal external validation, and formal updating within a Chinese population.

Methodology
The study employed a retrospective cohort design involving two distinct groups from a single tertiary referral center in China:

Derivation Cohort: 112 consecutive patients with unilateral PA undergoing adrenalectomy between 2018 and 2021.
Temporal Validation Cohort: 75 consecutive patients undergoing the same procedure in 2023, with a minimum of 6 months postoperative follow-up.

Data and Variables:
Clinical data were extracted from electronic medical records, including demographics, hypertension history, comorbidities (specifically cerebro-cardiovascular disease [CCD]), and laboratory parameters. Postoperative outcomes were classified according to the Primary Aldosteronism Surgical Outcome (PASO) criteria, dichotomized into "cure" (complete clinical remission) and "non-cure."

Model Development and Updating:

Initial Development: A multivariable logistic regression with forward stepwise selection was performed on the derivation cohort. Independent predictors were combined with hypertension duration (retained for clinical relevance) to construct a five-variable nomogram.
Temporal Validation: The derivation model was applied to the 2023 cohort without refitting to assess discrimination (C-index), calibration (slope and intercept), and clinical utility (Decision Curve Analysis).
Model Updating: Upon observing overfitting and poor discrimination of specific variables in the validation set, the authors performed:
- Parsimony Reduction: Removing variables that lost significance (specifically hypertension duration) to create a four-variable model.
- Logistic Recalibration: Re-estimating the intercept and slope using the validation data to correct for overfitting.

Key Results

Derivation Performance: The initial five-variable nomogram (incorporating age, waist circumference, highest systolic blood pressure [SBP], hypertension duration, and CCD history) showed excellent apparent discrimination (C-index 0.907) in the derivation cohort, significantly outperforming the ARS (C-index 0.743).
Validation Performance: When applied to the 2023 temporal validation cohort, the original model's performance dropped significantly, revealing overfitting. The C-index decreased to 0.734, and the calibration slope was 0.32, indicating systematic over-prediction of cure probabilities. Notably, hypertension duration was not a discriminative factor in the validation cohort ( $P=0.354$ ).
Updated Model Performance:
- Parsimony: Excluding hypertension duration yielded a four-variable model (Age, Waist Circumference, Highest SBP, CCD History) with improved discrimination in the validation cohort (C-index 0.812).
- Recalibration: Logistic recalibration further improved calibration (slope adjusted to 1.00) while maintaining the C-index of 0.812.
- Comparison: The updated, recalibrated four-variable model significantly outperformed both the original five-variable nomogram and the ARS (C-index 0.638) in the validation cohort.
- Clinical Utility: Decision Curve Analysis confirmed that the updated model provided a superior net benefit across clinically relevant threshold probabilities (0.20–0.80) compared to the original models and treat-all strategies.

Significance and Claims
The authors claim that this study provides a rigorous example of the necessity of temporal external validation and model updating in clinical prediction research. The primary significance lies in demonstrating that a model with excellent apparent performance can be substantially "tempered" when tested on a temporally distinct cohort.

The study proposes that the recalibrated four-variable model offers the best transportability for clinical use. Its significance is threefold:

Practicality: It relies entirely on routinely available preoperative variables (age, waist circumference, highest SBP, and CCD history), requiring no specialized testing.
Superiority: It demonstrates better discrimination and calibration than the widely used ARS and the initial derivation-stage nomogram in a Chinese population.
Clinical Application: The authors propose this model for individualized preoperative counseling to manage patient expectations regarding hypertension cure following unilateral adrenalectomy.

The authors explicitly note limitations, including the single-center nature of the study, the lack of geographic external validation, and the retrospective design. They conclude that while the updated model is promising, prospective multi-center validation is required before routine clinical adoption.

Technical Summary: Development, Temporal External Validation, and Updating of a Nomogram for Predicting Hypertension Cure After Unilateral Adrenalectomy in Primary Aldosteronism

More like this