Unilateral Primary Aldosteronism Lacking KCNJ5 Somatic Mutations Is Associated With an Elevated Risk of New-Onset Diabetes Free

Primary aldosteronism (PA) is characterized by autonomous overproduction of aldosterone and is the most common form of secondary hypertension (HTN) (1,2). It affects ∼5% of individuals with HTN (3), but its prevalence increases to 15–20% in patients with refractory HTN (2). Patients with PA have a higher risk of cardiac dysfunction, impaired kidney function, and cerebrovascular events than those with matched HTN (4–7). Moreover, aldosterone overproduction in patients with PA may be linked to impaired glucose metabolism, insulin resistance, and glucocorticoid excess (8–10). Unilateral PA (uPA), which includes unilateral aldosterone-producing adenoma (APA), aldosterone-producing micronodules, and adrenal cortical hyperplasia, is typically treated by unilateral adrenalectomy of the adrenal gland responsible for oversecretion. In contrast, bilateral forms are usually treated with mineralocorticoid receptor antagonists (11).

Previous studies demonstrated that patients with PA had a higher prevalence of diabetes and metabolic syndrome than the general population and patients with essential HTN (12). Adrenalectomy for patients with APA could reduce the risk of new-onset diabetes (NOD) compared with matched controls with essential HTN; however, treatment with MRAs in patients with APA could not attenuate the risk of NOD during the long-term follow-up period (13). Although adrenalectomy could decrease the comorbidities of PA, only ∼37–50% of patients uPA achieve complete clinical success (defined as normal blood pressure [BP] without antihypertensive medications) after adrenalectomy (14). It is important to evaluate who may benefit more from adrenalectomy in patients with uPA to reduce the risk of NOD.

The most frequently observed somatic mutation in uPA affects the KCNJ5 gene (15), which encodes for a G-protein–activated inward rectifier potassium channel that causes a calcium-mediated increase in aldosterone secretion (16). In patients with uPA with KCNJ5 mutations, clinical features include younger age, female predominance, higher plasma aldosterone levels, and larger adrenal tumors (17). In addition, our previous research showed that patients with APA with KCNJ5 mutations displayed lower CYP11B1 expression and were less likely to have cortisol cosecretion (17). Chen et al. (18) also reported that patients with APA with KCNJ5 mutations had a lower incidence of metabolic syndrome and abdominal obesity than those without KCNJ5 mutations (17,18). However, the risk of NOD in patients with uPA harboring KCNJ5 somatic mutations during long-term follow-up is unknown.

In this study, we aimed to investigate the long-term NOD outcomes in patients with uPA harboring KCNJ5 mutations. We used the Taiwan Primary Aldosteronism Investigation (TAIPAI) database and validated it using a population cohort to conduct a large longitudinal population study to examine the association between KCNJ5 mutations and NOD.

We enrolled 360 patients with uPA who had undergone adrenalectomy from January 2012 to December 2017 and were observed until Decemmber 2021. All patients provided written informed consent and were registered in the TAIPAI database (19,20). The institutional review board of the National Taiwan University Hospital approved the study (21).

Before adrenalectomy at the index date, we collected data on various demographic and medical factors, including age, sex, body weight, waist circumference, BMI, medical history of hyperlipidemia, BP, categories of antihypertensive medications, HTN duration, serum creatinine, estimated glomerular filtration rate (eGFR) (22), lowest serum potassium, plasma aldosterone concentration (PAC), and plasma renin activity (PRA). We also assessed factors related to metabolic syndrome, such as fasting glucose, lipid profile (including cholesterol, triglycerides [TGs], HDL, and LDL), waistline, fasting insulin levels, and HOMA2 for insulin resistance (HOMA2-IR; calculated using the calculator supplied by the University of Oxford website) at the index date. Diagnosis criteria for metabolic syndrome were defined as having three or more of the following conditions: 1) waist circumference ≥102 cm for men or ≥88 cm for women (measured across the abdomen), 2) BP ≥130/85 mmHg or taking BP medication, 3) TG level >150 mg/dL or currently taking medication for elevated TGs, 4) Fasting blood glucose level >100 mg/dL or taking glucose-lowering medications, and 5) HDL level <40 mg/dL for men or <50 mg/dL for women or currently taking medication for reduced HDL. In addition, we collected data on body weight, BMI, waist circumference, serum potassium, BP, categories of antihypertensive medications, renin levels, and aldosterone levels 6 months to 1 year after adrenalectomy. This was done to evaluate clinical and biochemical outcomes after adrenalectomy. To ensure accuracy, patients were withdrawn from antihypertensive medications for at least 21 days before the index date, with the exception of nondihydropyridine calcium channel blockers and α-blockers.

PAC and PRA were evaluated using commercial radioimmunoassay kits (ALDO-RIACT RIA Kit; Cisbio Bioassays, Codolet, France, and Angiotensin I RIA Kit; Beckman Coulter/Immunotech, Prague, Czech Republic, respectively) (17).

The primary outcome of this study was NOD. To minimize misclassification, we defined having diabetes as using any diabetes drugs (including oral antihyperglycemic agents or insulin) in consecutive visits >30 days apart. Additionally, we validated the TAIPAI Study Group records by comparing them with the Taiwan National Health Insurance Research Database (NHIRD) to confirm the occurrence of long-term outcome events. We also evaluated the outcome, which was defined as achieving complete clinical success 6 months to 1 year after adrenalectomy according to the Primary Aldosteronism Surgical Outcome consensus (Supplementary Table 1) (14), indicating normal BP without the need for anti-HTN medication. In addition to clinical outcome evaluations, we recorded biochemical outcomes, defined as potassium, PAC, and PRA levels, 6 months to 1 year after adrenalectomy.

We conducted an analysis to determine whether any baseline characteristics were associated with KCNJ5 mutations in patients with uPA who underwent adrenalectomy. To assess the specificity of our findings, we identified episodes of gastrointestinal (GI) bleeding as a negative outcome control (as listed in Supplementary Table 2), which are not known to be associated with KCNJ5 mutations.

Baseline characteristics are presented as mean ± SD for continuous variables and frequency with percentage for categorical variables. The Mann-Whitney U test was used to compare continuous variables, whereas categorical variables were analyzed using the χ² or Fisher exact test. Cox proportional hazards models were used to calculate the hazard ratios (HRs) and 95% CIs for NOD and GI bleeding incidence in patients with uPA with or without KCNJ5 mutations. Univariate analysis was performed to identify known predictors or variables with P value <0.05, which were then subjected to multivariate regression analysis to determine significance. In the Cox regression analysis, the effect of unilateral adrenalectomy was considered a time-varying covariate. Model 1 included KCNJ5 mutations, BMI, eGFR, age ≥35 years, systolic BP, and male sex. Model 2 added log PAC (postoperation), log PRA (postoperation), and metabolic syndrome to the variables in model 1. Model 3 included the variables in model 2 and incorporated unilateral adrenalectomy as a time-varying covariate. To evaluate the potential influence of unmeasured confounders, E value (23) was used to determine the strength of the association between KCNJ5 mutations and NOD.

A generalized additive model (GAM) was plotted to determine the age cutoff point to predict the incidence of NOD. The model incorporated the subject-specific (longitudinal) random effects, expressed as the logarithm of the odds (logit). The optimal cutoff value was defined as a log odds value of zero (24).

We used R software (version 3.4.4; Free Software Foundation, Inc., Boston, MA) and IBM SPSS statistics software (version 24; IBM Corp., Armonk, NY). A two-sided P value <0.05 was considered statistically significant.

This study enrolled patients who were referred to the TAIPAI Study Group, including two tertiary medical centers, three affiliated hospitals, and two regional hospitals in different cities in Taiwan. Ethical approval was obtained from the institutional review board of National Taiwan University Hospital. Written informed consent for clinical data collection and research use was obtained from all participants before enrollment in the study (21). The authors declare that all supporting data are available within the article and the Supplementary Materials.

A total of 360 patients with uPA who underwent adrenalectomy during the study period were included in this analysis. The cohort had a mean age of 49.9 ± 11.4 years, and 44.4% of the participants were male. Among the patients, 53.1% had adenomas with KCNJ5 mutations, whereas 46.9% did not carry the mutation (Table 1). The distribution of known non-KCNJ5 somatic mutations of uPA in our cohort is shown in Supplementary Fig. 1.

Compared with patients with uPA without KCNJ5 mutations, patients with uPA harboring KCNJ5 mutations were younger (46.9 ± 10.3 vs. 53.4 ± 11.7 years; P < 0.01) and had a lower BMI (24.8 ± 3.9 vs. 25.6 ± 4.1 kg/m²; P = 0.03) and waist circumference (81.2 ± 10.2 vs. 85.9 ± 12.1 cm; P < 0.01), as well as lower PRA (0.39 ± 0.60 vs. 1.00 ± 3.82 ng/mL/h; P = 0.01) and serum potassium levels (3.2 ± 0.6 vs. 3.8 ± 0.5 mEq/L; P < 0.01), but higher diastolic BP (94 ± 15 vs. 91 ± 14 mmHg; P = 0.03), PAC (63.4 ± 42.5 vs. 47.5 ± 33.4 ng/dL; P < 0.01), and eGFR (93.1 ± 20.7 vs. 86.8 ± 21.5 mL/min/1.73 m²; P < 0.01). Furthermore, patients with uPA harboring KCNJ5 mutations had a lower prevalence of metabolic syndrome (30.4% vs. 44.8%; P = 0.01) and lower fasting glucose levels (95.17 ± 15.17 vs. 99.01 ± 14.69 mg/dL; P < 0.01), fasting insulin levels (8.6 ± 8.3 vs. 11.8 ± 11.1 μU/mL; P < 0.01), HOMA2-IR (1.12 ± 1.04 vs. 1.54 ± 1.39; P < 0.01), cholesterol (180.40 ± 32.16 vs. 191.19 ± 39.83 mg/dL; P = 0.01), and TGs (99.08 ± 49.54 vs. 138.87 ± 84.82 mg/dL; P < 0.01) at the index date. At 6 months to 1 year after adrenalectomy, patients with uPA with KCNJ5 mutations had a smaller waist circumference (82.4 ± 11.3 vs. 86.2 ± 12.3 cm; P = 0.02) and fewer antihypertensive medications (0.33 ± 0.70 vs. 0.83 ± 1.08; P < 0.01) than those without KCNJ5 mutations. There were no significant differences between the two groups regarding sex, history of hyperlipidemia, body weight, number of antihypertensive medications, HTN duration, systolic BP, or HDL or LDL levels at the index date. We also found that patients with uPA harboring KCNJ5 mutations had a higher rate of complete clinical success (69.5% vs. 43.8%; P < 0.01) and complete biochemical success (93.8% vs. 86.6%; P = 0.04) compared with those without KCNJ5 mutations at 6 months to 1 year after adrenalectomy (Table 1). In addition, we observed that patients with uPA in our cohort who developed NOD after adrenalectomy had lower eGFR, longer HTN duration, and higher BMI, fasting glucose levels, HOMA2-IR, and TGs, as well as lower HDL levels, at the index date (Supplementary Table 3). We also created a GAM plot to identify the optimal cutoff point for age, which was found to be associated with the timing of NOD diagnosis (Supplementary Fig. 2).

Baseline cortisol levels in our cohort showed no statistically significant differences between the KCNJ5 mutation group and the group without KCNJ5 mutations (baseline cortisol levels 12.1 ± 6.0 vs. 12.4 ± 5.7 μg/dL; P = 0.63). In our cohort, 134 patients underwent a 1-mg dexamethasone suppression test (DST). The prevalence of autonomous cortisol secretion (ACS) was 30% in KCNJ5 carriers and 42.3% in those without KCNJ5 mutations (P = 0.08).

After adrenalectomy, the incidence of NOD was lower in patients with uPA with KCNJ5 mutations compared with that in patients without (5.8% vs. 13.6%; P = 0.01) during a mean follow-up period of 8.5 ± 3.7 years (Table 1 and Fig. 1). The adjusted HR (95% CI) for NOD in patients with uPA with KCNJ5 mutations compared with those without was 0.39 (0.16–0.97; P = 0.04) in model 1 and 0.41 (0.17–0.996; P = 0.049) in model 2. In addition, the lower incidence of NOD in KCNJ5 mutation carriers compared with in those without KCNJ5 mutations was also observed in model 3, which included unilateral adrenalectomy as a time-varying covariate in the multivariate Cox regression analysis (adjusted HR 0.41; 95% CI 0.17–0.995; P = 0.049) (Table 2).

The univariate analysis showed that patients with uPA harboring KCNJ5 mutations had a lower risk (HR 0.38; 95% CI 0.19–0.78; P = 0.01) of developing NOD compared with those without KCNJ5 mutations. In contrast, high BMI (HR 1.19; 95% CI 1.11–1.27; P < 0.01), lower eGFR (HR 0.98; 95% CI 0.97–1.00; P = 0.03), and metabolic syndrome at baseline (HR 3.10; 95% CI 1.58–6.18; P < 0.01) were found to be positively associated with the incidence of NOD (Table 3 [variables included in Table 3 are the same as those included in model 2]). In the multivariate analysis, after adjusting for other variables, uPA in patients harboring KCNJ5 mutations remained independently predictive of a lower incidence of NOD (HR 0.41; 95% CI 0.17–0.996; P = 0.049). Furthermore, high BMI (HR 1.23; 95% CI 1.11–1.37; P < 0.01) and lower eGFR (HR 0.98; 95% CI 0.97–0.99; P = 0.01) were also found to be independently predictive of the incidence of NOD after adrenalectomy in patients with uPA.

Age is indeed an important factor in assessing diabetes risk. The 2022 American Diabetes Association Standards of Medical Care in Diabetes recommends annual diabetes screening for individuals aged ≥35 years because of cost and impact (25). Based on this recommendation, diabetes screening would most likely occur in clinical practice for individuals older than 35 years of age. In addition, we used GAM plots to identify appropriate cutoff points of the continuous parameters to predict which patients with uPA were at higher risk of NOD. We found that the baseline characteristic of age ≥35 years was associated with the incidence of NOD in our cohort during the follow-up periods. Therefore, we chose 35 years of age as the cutoff point to predict NOD in our cohort.

The E value for point estimates (upper limit of the CI) for NOD was 4.31 (1), greater than the HR for established NOD factors in patients with uPA. This analysis indicated no substantial unmeasured confounding in our cohort.

Our subgroup analyses further delineated the effect of KCNJ5 mutations, consistent with our main findings. The association between patients with uPA without KCNJ5 mutations and higher risk of NOD was attenuated in subgroups with younger age (<35 years), higher BMI (≥25 kg/m²), higher eGFR (≥90 mL/min/1.73 m²), and lower potassium levels (<3.3 mEq/L) (Fig. 2). However, the P value for the interaction test in our subgroup analysis did not show statistical significance, which may attenuate the differences in long-term outcomes in our subgroups. We also conducted a specificity analysis and found no association between KCNJ5 mutations and the incidence of GI bleeding (7.5% of enrolled patients had a GI bleeding episode; HR 0.95; 95% CI 0.41–2.19; P = 0.90), indicating a low possibility of health indication biases or unobserved confounders (Supplementary Table 4).

This study marks the inaugural comprehensive longitudinal cohort analysis exploring the relationship between somatic KCNJ5 mutations and metabolic syndrome in patients with uPA over an average follow-up period of 8.5 years. Our observations indicate that patients with uPA without KCNJ5 mutations may experience an increased likelihood of developing NOD compared with those with KCNJ5 mutations. These findings provide valuable insights into the different follow-up strategies for different somatic mutations that lead to NOD in patients with uPA. Additional investigations are essential to validate and build on these findings.

Metabolic syndrome is a multifaceted condition characterized by insulin resistance and atypical adipose function. This cluster of conditions includes hypertension, hyperglycemia, abdominal obesity, and dyslipidemia, which can elevate the risk of cardiovascular disease, stroke, and type 2 diabetes (26,27). Past research has demonstrated a link between PA and metabolic syndrome, with several observational studies reporting a prevalence of metabolic syndrome ranging from 41 to 45% among patients with PA (28,29). These findings highlight the importance of monitoring and managing metabolic syndrome in patients with PA to reduce the risk of associated cardiovascular and metabolic complications. The main risk factors for metabolic dysfunction in patients with PA may be related to aldosterone overproduction and insulin resistance, which contribute to abnormal glucose metabolism (30).

Freel et al. (31) conducted clinical studies demonstrating an association between aldosterone excess and insulin resistance. Additionally, a cohort study in Japan found that patients with PA had a higher prevalence of diabetes than the general population, even after adjusting for age and sex (12). We reported that patients with PA who underwent adrenalectomy had a reduced risk of developing incident diabetes compared with control participants with HTN who were matched for other factors (13). Based on this evidence, it can be concluded that there is a link between PA and a higher incidence of NOD.

Several possible pathophysiologic links exist between PA and various aspects of glucose metabolism. First, previous studies (32) have demonstrated an association between a specific single nucleotide polymorphism in the promoter region of the CYP11B2 gene and higher fasting plasma glucose levels, as well as an increased incidence of diabetes and metabolic syndrome. Second, aldosterone may contribute to insulin resistance by altering vascular smooth muscle (33) and skeletal muscle function (34). Third, aldosterone can affect the insulin receptor and intracellular glucose pathways in adipocytes (35). Fourth, cortisol cosecretion in PA, present in 25–30% of patients with unilateral PA (10), may also contribute to an increased metabolic risk. Finally, a recent study showed that insulin levels are reduced in PA because of decreased β-cell function and increased insulin clearance, which could be reversed by targeted treatment (36). In addition, previous research has shown that aldosterone directly impairs glucose-stimulated insulin secretion through the mediation of reactive oxygen species (37), and hypokalemia has been implicated in the impairment of insulin secretion (38). However, further research is needed to clarify the relationship between PA and glucose metabolism.

The association between patients with PA harboring KCNJ5 mutations and metabolic syndrome has been reported in few studies. In a 1-year follow-up study by Chen et al. (18), it was demonstrated that patients with APA harboring KCNJ5 mutations had a lower prevalence of metabolic syndrome and abdominal obesity than the APA group without KCNJ5 mutations. Our study also found that patients with unilateral PA harboring KCNJ5 mutations had lower BMI, fasting glucose, TG, HOMA2-IR, and cholesterol levels at the time of diagnosis.

Patients with NOD have been shown to have a significantly higher amount of visceral adipose tissue (VAT) compared with healthy controls. VAT is also an important indicator of insulin resistance (39). Previous studies have shown that patients with APA have smaller VAT areas than patients with idiopathic hyperaldosteronism (IHA) and essential HTN (40,41). Chen et al. (18) also found that patients with APA harboring KCNJ5 mutations had smaller subcutaneous adipose tissue and VAT areas than those without KCNJ5 mutations. Among patients with uPA, KCNJ5 mutation carriers had higher aldosterone levels and lower BMI than those without KCNJ5 mutations, suggesting that in our cohort they may have had lower VAT. These results suggest that the association between KCNJ5 mutations and lower incidence of NOD may be related to the relatively small amount of VAT, which may surpass the excess aldosterone oversecretion. Additionally, recent studies have shown better clinical outcomes after adrenalectomy in KCNJ5 mutation carriers (42,43), which may contribute to the lower incidence of NOD after adrenalectomy.

It is important to note that the lack of significant findings in the younger, higher BMI and eGFR, and lower potassium level subgroups of our cohort does not necessarily rule out the potential association between KCNJ5 mutations and NOD. Age and BMI are well-established risk factors for NOD, and the effect of KCNJ5 mutations on NOD may have been masked or diluted in these subgroups because of the overwhelming influence of these risk factors (44–46). Additionally, although lower potassium levels and glomerular hyperfiltration may indicate PA and lead to earlier diagnosis, it is also possible that the severity of PA in patients with lower potassium levels and hyperfiltration may be more significant. Therefore, the association between KCNJ5 mutations and a lower incidence of NOD may be outweighed by the increased severity of PA (47). Further research is required to fully understand the relationship between KCNJ5 mutations and NOD in different subgroups of patients with uPA.

The role of KCNJ5 mutations in patients with uPA after adrenalectomy is still uncertain. Indeed, once adenomas harboring KCNJ5 mutations are surgically removed, the direct effects of the mutations on aldosterone production are eliminated. However, it is important to consider the potential long-term metabolic changes that may persist even after adenoma removal. First, it is well-established that baseline characteristics and preoperative conditions can significantly influence postoperative outcomes in patients with PA (14). For instance, our findings suggest that patients with uPA without KCNJ5 mutations exhibit characteristics such as older age, obesity, lower eGFR, higher potassium levels, hyperlipidemia, and insulin impairment, which may require more vigilant monitoring for NOD. Additionally, arterial stiffness and cardiovascular function have been implicated as factors affecting outcomes after adrenalectomy (48). Furthermore, KCNJ5 mutations have been linked to aldosterone production, broader metabolic effects, and cardiovascular outcomes. Studies have suggested that patients with KCNJ5 mutations may experience distinct cardiovascular function benefits from adrenalectomy (42). These findings indicate that KCNJ5 mutations could potentially influence metabolic abnormalities beyond aldosterone secretion, which may persist even after tumor removal. Although our study focused on the direct association between KCNJ5 mutations and NOD, our research could explore the long-term metabolic effects and legacy effects of KCNJ5 mutations after adrenalectomy. This would help elucidate how preoperative mutations may predispose patients to metabolic risks that persist or evolve postoperatively, affecting overall patient outcomes. Additional studies may help to clarify the extent to which metabolic characteristics developed in response to the mutations may persist after surgery and influence patient outcomes.

In 2017, Arlt et al. (10) raised the important issue of the additional glucocorticoid effect in patients with PA, which at least partially contributes to associated adverse metabolic risks in PA. However, previous diagnostic workups in primary aldosteronism did not include an assessment of hypercortisolism (11). Our recent study showed that patients with PA with concomitant higher cortisol levels after 1-mg DST, independent of aldosterone levels, had a higher incidence of NOD (49). In this study, the relatively higher prevalence of ACS in patients with uPA without KCNJ5 mutations may be compatible with previous pathologic features (17) and associated with a higher rate of metabolic syndrome at enrollment. Future studies should include assessment of cortisol secretion to provide a more complete understanding of the metabolic consequences of PA, particularly in subgroups defined by specific genetic mutations.

Our study proved the association between patients with uPA harboring KCNJ5 mutations and NOD during long-term follow-up. This allowed a more comprehensive evaluation of the association between KCNJ5 mutations and NOD. However, given the complex interaction of aldosterone and metabolic syndrome and the fact that few studies have discussed the impact of KCNJ5 mutations on metabolic syndrome, further investigation is needed. Although our previous study showed that adrenalectomy reduces the risk of diabetes compared with controls with essential HTN, the current study is a separate project and does not include a comparison group from the general population or the population with HTN. In addition, the inclusion of population-level data or controls with essential HTN would be valuable to determine whether patients with uPA without KCNJ5 mutations retain the benefit of adrenalectomy in terms of diabetes risk in future research.

Our study had several limitations. First, the lower NOD incidence makes the study susceptible to α errors, and unmeasured confounders may have biased the association between KCNJ5 mutations and NOD incidence. Our sample size was relatively small, which may have limited the power of our analysis and subgroup analysis. In our sensitivity analysis, the E values were greater than the known relative risk factors in patients with uPA, and the risk of incident GI bleeding was similar between the groups. Therefore, the observed effects on NOD were unlikely to be explained by unmeasured confounding or systemic bias. Second, it should be noted that we did not evaluate the existence of subclinical Cushing syndrome as a variable in our cohort. ACS in patients with PA provides new insight into the association between PA and metabolic syndrome. A large Japanese cohort study reported that the higher prevalence of diabetes in patients with PA was associated with subclinical hypercortisolism (12). Moreover, the presence of KCNJ5 mutations was negatively associated with cortisol cosecretion in patients with PA (17). The relatively higher prevalence of ACS in patients with uPA without KCNJ5 in our cohort may have been associated with a higher rate of metabolic syndrome at enrollment. However, additional studies should be performed to clarify the association between ACS and NOD in different somatic genetic mutation subgroups. Third, our study only included patients with uPA, which may limit the generalizability of our findings to other patient populations. Fourth, the included patients in our cohort were of a similar ethnic background; therefore, our findings must be confirmed in other ethnic groups. Fifth, we did not have data on potentially important variables, such as dietary and lifestyle factors, which may have influenced our results. Sixth, the group without KCNJ5 mutations demonstrated features that might overlap with IHA. This could have had an impact on the observed outcomes. To clarify the subgroup of PA, all patients included in our study underwent adrenal venous sampling to confirm unilateral aldosterone hypersecretion, ensuring accurate lateralization. This procedure effectively excludes most cases of bilateral adrenal hyperplasia. Consequently, the likelihood of misclassifying IHA as unilateral disease in our cohort was decreased. Seventh, there remains a risk of undiagnosed diabetes in our cohort because of missed follow-up visits or other factors. A large proportion of the patients in our study were part of the TAIPAI Study Group, which requires routine follow-up visits after adrenalectomy. These visits include comprehensive metabolic assessments to ensure consistent diabetes screening. In addition, we used the NHIRD to confirm the incidence of NOD, which provides comprehensive nationwide data on health care use, including outpatient visits, hospitalizations, and prescriptions. The high coverage of the NHIRD (>99% of Taiwan’s population) ensures that almost all health care interactions are captured, minimizing the likelihood of missed diagnoses because of a lack of follow-up. Lastly, although the use of age as a continuous variable in multivariable models is statistically valid, it does not adequately capture the threshold effect observed in our cohort. To better illustrate this relationship, we used a GAM plot, and it showed a nonlinear relationship between age and NOD risk, identifying 35 years as the optimal threshold at which the risk begins to increase significantly. This result suggested that using age as a categorical variable (cutoff at 35 years) better captures the biologic and clinical interactions underlying our findings.

Our study uncovered a significant finding that patients with uPA without KCNJ5 mutations exhibit a higher incidence of NOD during long-term follow-up compared with those with KCNJ5 mutations. Patients with uPA with KCNJ5 mutations have higher rates of both complete clinical and biochemical success. These results imply that individuals without KCNJ5 mutations may require more frequent monitoring for NOD. These insights underscore the importance of genetic testing and personalized follow-up strategies in the clinical management of uPA.

Acknowledgements. The authors thank the staff of the Second Core Lab, Department of Medical Research, National Taiwan University Hospital, for technical assistance and all Taiwan Clinical Trial Consortium staff and the TAIPAI Study Group.

Funding. This study was supported by grants from the Ministry of Science and Technology of Taiwan (106-2321-B-182-002, 108-2314-B-002-058, 109-2314-B-002-174-MY3, 112-2628-B-002-026-MY3) and the National Taiwan University Hospital (107-A141).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. C.-K.C. performed the experiments. C.-K.C., W.-S.Y., and Y.-H.L. analyzed data. C.-K.C. and V.-C.W. wrote the manuscript. C.-K.C., V.-C.W., and J.S.C. conceived and designed the experiments. V.-C.W. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.