Factors influencing postoperative of hypertension remission in adrenal cavernous hemangioma: A multicenter retrospective cohort study

Abstract

BACKGROUND

Adrenal cavernous hemangioma (ACH) is a rare benign tumor, accounting for less than 1% of adrenal incidentalomas, with marked variability in postoperative hypertension resolution. The lack of reliable predictive tools has long hindered precise clinical decision-making, making it challenging to optimize patient management and prognosis.

AIM

To summarize the clinical characteristics of ACH patients, identify key factors influencing postoperative hypertension remission, and develop a robust predictive model to facilitate individualized preoperative risk stratification and postoperative management.

METHODS

A retrospective cohort study enrolled 102 pathologically confirmed ACH patients with hypertension from 26 tertiary hospitals in China between 2003 and 2023. Standardized data collection was performed, with independent factors identified via univariate and multivariate logistic regression analyses. The model was validated using bootstrap resampling (n = 1000) and 10-fold cross-validation to ensure stability.

RESULTS

Five independent predictors were identified: Tumor diameter, renin-angiotensin-aldosterone system hormone, adrenal medullary hormone, hypertension duration, and surgical approach. The combined model demonstrated good discriminative ability (area under the curve = 0.872, 95% confidence interval: 0.791-0.934). The maximum Youden index (0.616) corresponded to an optimal cutoff value of 12.57 (sensitivity = 0.827, specificity = 0.861). This threshold applies only to hypertensive ACH patients meeting the study’s inclusion criteria. Per Oxford Centre for Evidence-Based Medicine criteria, this threshold is classified as level 2b (retrospective cohort study evidence).

CONCLUSION

This model demonstrates good discriminative performance for predicting postoperative hypertension outcomes in ACH patients in this multicenter cohort. Further prospective and external validation studies are warranted before routine clinical implementation.

Key Words: Adrenal tumor; Cavernous hemangioma; Hypertension; Combined predictor; Predictive model

Core Tip: In this multicenter retrospective cohort of 102 hypertensive patients with adrenal cavernous hemangioma, postoperative hypertension remission at 24 months was independently associated with tumor diameter, hypertension duration, renin-angiotensin-aldosterone system activity, adrenal medullary catecholamines, and surgical approach. A combined model integrating these variables showed good discrimination (area under the curve = 0.872) and provided a practical cutoff (l = 12.57) to stratify risk of persistent hypertension, supporting preoperative risk stratification and individualized perioperative management.

INTRODUCTION

Adrenal cavernous hemangioma (ACH) is a rare benign adrenal tumor, accounting for less than 1% of adrenal incidentalomas, with an annual incidence of approximately 0.001-0.002 per 100000 population[1-4]. It predominantly affects individuals aged 30-60 years, with no gender predilection, and is often detected incidentally because of its insidious clinical manifestations[1-4]. Surgical resection is the standard treatment; laparoscopic adrenalectomy is generally preferred for its minimally invasive advantages, whereas open surgery is reserved for complex or highly adherent lesions[5]. Among ACH patients with a history of hypertension, however, postoperative blood pressure outcomes vary widely. Existing studies have mainly focused on single factors such as tumor diameter or surgical approach and have rarely incorporated renin-angiotensin-aldosterone system (RAAS) hormone levels, adrenal medullary hormone profiles, or hypertension duration into multivariable quantitative models[6-8]. As a result, the determinants of postoperative hypertension remission in ACH remain insufficiently characterized, and evidence-based risk stratification is still lacking.

Previous reports have suggested that most ACHs are hormonally non-functioning[1-4], and hypertension is relatively uncommon in this entity[7-9]. However, individual case reports and small series have described ACH associated with hyperaldosteronism, hypercortisolism, or elevated catecholamines, sometimes leading to secondary hypertension or adrenal crisis. Because these observations are largely anecdotal and often confounded by concomitant adrenal pathology, the true frequency and clinical relevance of hypertension in ACH remain uncertain. Against this background, our multicenter study focuses specifically on ACH patients with documented preoperative hypertension, aiming to delineate the clinical factors associated with postoperative blood pressure outcomes while avoiding overinterpretation of causality.

MATERIALS AND METHODS

Patients and methods

Clinical data: During the study period (June 2003 to June 2023), a total of 240 patients with pathologically confirmed ACH underwent surgery at the 26 participating tertiary hospitals. Among them, only those with documented preoperative hypertension were eligible for the present analysis. Preoperative hypertension was defined as office blood pressure ≥ 140/90 mmHg on at least two separate clinic visits and/or 24-hour ambulatory blood pressure monitoring (ABPM) showing daytime ≥ 135/85 mmHg or nighttime ≥ 120/70 mmHg. Postoperative blood pressure status was assessed at 24 months (± 2 months) after surgery using 24-hour ABPM. Postoperative hypertension remission was defined as daytime blood pressure < 135/85 mmHg and nighttime < 120/70 mmHg on ABPM without any antihypertensive medication. Patients who did not meet these criteria or who required ongoing antihypertensive therapy were classified as having persistent hypertension. Based on these criteria, 102 hypertensive ACH patients were included and followed for postoperative blood pressure outcomes, corresponding to 60.25% of the overall ACH population in these centers. Normotensive ACH patients were not included in this cohort because the primary objective of this study was to identify factors associated with postoperative remission or persistence of hypertension. To address potential selection bias from excluding non-hypertensive ACH patients, we supplemented data from 138 non-hypertensive ACH patients treated at the same 26 centers. We compared baseline characteristics (tumor diameter, RAAS hormone levels, adrenal medullary hormone levels) between hypertensive (n = 102) and non-hypertensive (n = 138) groups using independent samples t-tests or χ² tests. Analysis shows baseline differences between hypertensive and non-hypertensive ACH patients: Tumor diameter (4.1 ± 1.5 cm vs 2.8 ± 1.1 cm, P < 0.001), RAAS hormone levels (156.4 ± 41.3 pg/mL vs 98.7 ± 25.6 pg/mL, P < 0.001), and adrenal medullary hormone levels (108.5 ± 28.7 pg/mL vs 72.3 ± 19.5 pg/mL, P < 0.001). These findings confirm that the key factors identified in this study are specifically associated with hypertension in ACH patients, rather than general tumor characteristics. Patients were divided into the remission group and the non-remission group based on their hypertension status at 24 months postoperatively. All 26 centers adopted unified case screening criteria, and data quality control was performed by a third-party institution to ensure case homogeneity across centers (Figure 1).

Figure 1 Flowchart of patient selection and inclusion. A total of 240 patients with adrenal cavernous hemangioma treated between 2003 and 2023 were screened, and 102 hypertensive adrenal cavernous hemangioma patients meeting the imaging, pathological and follow-up eligibility criteria were included in the final analysis. ACH: Adrenal cavernous hemangioma; RAAS: Renin-angiotensin-aldosterone system; ROC: Receiver operating characteristic.

Inclusion criteria: (1) Preoperative enhanced computed tomography (CT) or magnetic resonance imaging (MRI) initially diagnosed as adrenal tumor; (2) Underwent adrenalectomy with postoperative pathological confirmation of ACH; (3) Had a history of hypertension; and (4) Complete clinical data, including preoperative peripheral venous blood samples were collected per standardized procedures: Sampling time: 7:00-8:00 to avoid diurnal hormone variation; posture: 15 minutes of seated rest before sampling to stabilize posture-dependent hormone levels; fasting: ≥ 8 hours of fasting, with no coffee, tea, or alcohol for 24 hours pre-sampling; stress control: Sampling in a quiet environment; rescheduling for patients with obvious stress responses (anxiety, pain). For values exceeding 2 × the reference range, two repeated measurements were performed; the mean was used if the difference between measurements was < 10%. Intra-assay coefficient of variation was controlled < 5%, and inter-assay coefficient of variation < 10%, per national clinical laboratory standards. In all centers, antihypertensive medications were withdrawn according to a standardized protocol during the preoperative phase, which we define as the general pre-surgical period encompassing the hormonal assessment and the time before surgery. Specifically, angiotensin-converting enzyme inhibitors and angiotensin II receptor blockers were discontinued at least 2 weeks before surgery, β-blockers at least 1 week before surgery, diuretics at least 4 weeks before surgery, and calcium channel blockers (CCBs) at least 2 weeks before surgery. For patients in whom complete withdrawal was clinically unsafe, the type and daily dose of antihypertensive drugs were carefully recorded and incorporated into the multivariable model as covariates. Overview to address heterogeneity in hormone assay methods and detection systems across 26 participating centers, we constructed composite indices for RAAS hormone and adrenal medullary hormone. Specifically, we converted individual hormone values to z-scores based on each center’s reference ranges, then averaged these z-scores to generate standardized composite indices. This approach ensures data comparability across centers and enhances reproducibility. All hormone assays adhered to unified quality control standards (inter-assay coefficient of variation < 10%) formulated by a third-party data management institution. For the RAAS, plasma aldosterone concentration was not consistently available across all centers. We therefore constructed a RAAS composite index by averaging the z-scores of renin and angiotensin II, as detailed in the methods overview. For adrenal medullary activity, plasma epinephrine and norepinephrine concentrations were measured via chemiluminescent assay (Siemens Immulite 2000; Siemens Healthcare Diagnostics Inc., NY, United States), and their z-scores were averaged to generate the adrenal medullary hormone composite index. The primary purpose of measuring plasma catecholamines in this study was exploratory: To quantify the degree of sympathetic activation potentially associated with local mechanical compression by the hemangioma. We acknowledge that plasma catecholamines are not always required in routine clinical practice for evaluating catecholamine secretion; therefore, our results are intended to complement, not replace, guideline-recommended diagnostic tests for pheochromocytoma or paraganglioma.

Exclusion criteria: Patients with missing key clinical data (e.g., incomplete imaging data, missing hormone detection results, or interrupted follow-up records).

Surgical methods

Classification and indications of surgical approaches.

Open surgery: Open total adrenalectomy: Indicated for: (1) Tumors with diameter > 6 cm and imaging adhesion grade ≥ III (CT/MRI suggesting adhesion or invasion of large blood vessels); and (2) Suspected malignancy on imaging. The affected adrenal gland was completely resected, and the central vein was ligated. It should be noted that literature reports that ACH ≥ 10 cm can be safely resected laparoscopically[10]; in this study, surgical approach selection strictly followed the modified Adler classification: Laparoscopic surgery was preferred for grade I/II adhesion and tumor diameter ≤ 6 cm, while open surgery was performed for grade III adhesion or tumor diameter > 6 cm to balance minimally invasive benefits and vascular safety. Open partial adrenalectomy: Only applicable to cases with localized tumors (adhesion grade ≤ II) requiring preservation of adrenal function (e.g., bilateral tumors or adrenal cortical insufficiency), with the lesion resected along the tumor capsule while preserving normal tissue.

Laparoscopic surgery: Laparoscopic total adrenalectomy: Indicated for tumors with diameter ≤ 6 cm and adhesion grade ≤ II, with the central vein transected using vascular clips/ultrasonic scalpel before complete resection of the adrenal gland[11]. Laparoscopic partial adrenalectomy: Only recommended for tumors with diameter ≤ 4 cm and adhesion grade ≤ I[12], with delicate laparoscopic techniques to preserve adrenal blood supply and normal structure.

Surgical selection criteria: Surgical selection criteria based on modified Adler classification[13]. Grade I/II adhesion: Laparoscopic surgery was preferred[14-18] (total/partial resection determined by tumor size and location). Grade III adhesion: Open surgery was performed directly to ensure vascular safety. Tumor diameter ≥ 6 cm or suspected malignancy: Open total resection was performed regardless of adhesion degree. Vital signs and hormone levels were routinely monitored during surgery, and hemostasis methods were selected according to intraoperative conditions.

Statistical analysis

All statistical analyses were performed using SPSS 23.0 (IBM Corp., NY, United States) and R 4.2.1 (R Foundation for Statistical Computing) under the supervision of a biostatistician. The primary outcome was persistent hypertension at 24 months (yes/no). Candidate predictors were pre-specified based on clinical relevance and previous literature, and all continuous predictors were standardized to z-scores before entry into the multivariable logistic regression model. Variables with P < 0.10 in univariate analyses were considered for inclusion and were retained in the final model using backward stepwise selection (removal criterion α = 0.05). Model discrimination was assessed using the area under the curve (AUC) and compared using the DeLong test. Calibration was evaluated by the Hosmer-Lemeshow goodness-of-fit test and visually by calibration plots. Internal validation was performed via non-parametric bootstrap resampling (1000 iterations) and k-fold cross-validation, as detailed in the Results. Two-sided P < 0.05 was considered statistically significant[19-21].

RESULTS

Comparison of general patient data

A total of 102 patients with surgically and pathologically confirmed ACH were included, including 40 males (39.2%) and 62 females (60.8%), with a mean age of 48.2 ± 10.5 years. According to postoperative hypertension outcomes, patients were divided into the positive group (remission, 62 cases, 60.8%) and the negative group (non-remission, 40 cases, 39.2%). There were no significant differences in gender or age between the two groups (P > 0.05), but significant differences were observed in tumor diameter, hypertension duration, RAAS and adrenal medullary hormone levels, and surgical approach (Table 1). The average tumor diameter in the positive group 3.2 ± 1.1 cm was significantly smaller than that in the negative group 4.8 ± 1.5 cm (P = 0.002), and the hypertension duration 3.5 ± 1.2 years was shorter than that in the negative group 7.8 ± 2.3 years (P < 0.001). The remission rate in the laparoscopic total resection group (83.3%) was significantly higher than that in the open surgery group, and for each 1 cm increase in tumor diameter, the risk of uncontrolled hypertension after open surgery increased by 1.8 times [odds ratio (OR) = 1.8, 95% confidence interval (CI): 1.1-2.9]. Among patients with tumor diameter ≤ 6 cm and adhesion ≤ II (n = 68), the remission rate in the laparoscopic group (85.3%, 29/34) was significantly higher than that in the open group (60.6%, 20/33).

Table 1 Baseline characteristics and clinical features of patients in hypertension cure group and non-cure group after adrenal cavernous hemangioma surgery, mean ± SD.

Indicator	Positive group (n = 62)	Negative group (n = 40)	P value
Age (years)	48.1 ± 9.8	50.5 ± 8.9	0.213
Gender (male/female)	24/38	16/24	0.576
Tumor diameter (cm)	3.3 ± 1.2	4.7 ± 1.4	< 0.001
Hypertension duration (years)	3.6 ± 1.3	7.5 ± 1.2	< 0.001
RAAS hormones (pg/mL)	128.5 ± 34.2	182.3 ± 43.8	< 0.001
Adrenal medullary hormones (pg/mL)	85.6 ± 21.3	132.4 ± 30.5	< 0.001
Surgical approach (laparoscopy/open)	52/10	24/16	0.008
Total adrenalectomy (laparoscopy/open)	48/8	16/12	0.012
Partial adrenalectomy (laparoscopy/open)	4/2	8/4	0.659

Measurement data were expressed as mean ± SD, and comparisons between groups were performed using the independent samples t-test. Enumeration data were presented as case numbers, and inter-group comparisons were conducted using the χ² test. Basis for selecting surgical methods: In the laparoscopic group, 93.5% (58/62) of tumors were ≤ 6 cm, while in the open surgery group, 77.5% (31/40) of tumors were > 6 cm. The proportions of adhesion grade ≥ III in the laparoscopic group and the open surgery group were 58% vs 82.5%, respectively (P < 0.001). Positive group: Patients whose hypertension was cured 24 months after surgery (24-hour ambulatory blood pressure monitoring daytime < 135/85 mmHg and nighttime < 120/70 mmHg without antihypertensive medication). Negative group: Patients with persistent hypertension not meeting the above criteria. RAAS: Renin-angiotensin-aldosterone system.

Intraoperative central vein injury (defined as vascular clip dislodgement or ultrasonic scalpel injury) was recorded: The central vein preservation rate in the laparoscopic group was 91.2% (75/82), significantly higher than that in the open group (68.3% 19/28, P < 0.001; Table 2). Correlation analysis correlation analysis showed that surgical approach was strongly positively correlated with tumor diameter (r = 0.68, P < 0.001) and adhesion grade (r = 0.75, P < 0.001). Laparoscopic surgery was primarily used for patients with tumor diameter ≤ 6 cm and adhesion grade ≤ II, while open surgery was reserved for larger tumors (diameter > 6 cm) and higher adhesion grades (≥ III). This confirms that surgical approach selection is heavily influenced by tumor complexity, introducing potential indication bias.

Table 2 Comparison of central vein preservation rate between surgical approach.

Surgical approach	Total cases	Central vein-preserved	Preservation	P value
Laparoscopic	82	75	91.2	< 0.001
Open	28	19	68.3

Central venous injury was defined as dislodgement of vascular clips or accidental injury by ultrasonic scalpel. The laparoscopic group adopted the fine dissection technique with vascular clips/ultrasonic scalpel, while the open surgery group used the ligation + transection method. Fisher’s exact test was applied, and the difference in retention rate between the two groups was of highly statistical significance (P = 0.003), suggesting that the laparoscopic approach provides better vascular protection.

Univariate analysis of factors associated with hypertension remission after ACH surgery

Univariate analysis showed that tumor diameter, hypertension duration, RAAS hormone, adrenal medullary hormone, and surgical approach were significantly associated with postoperative hypertension remission (all P < 0.001), while age and gender were not (both P > 0.1). In the multivariable model using standardized predictors, a 1-SD increase in the RAAS hormone was associated with a significantly higher odds of unresolved postoperative hypertension (OR = 3.63, 95%CI: 1.78-7.40), and a 1-SD increase in the medullary hormone was similarly associated with poorer blood pressure outcomes (OR = 2.98, 95%CI: 1.31-6.76). For clinical interpretation, one SD corresponded approximately to 22 pg/mL for the RAAS hormone and 45 pg/mL for the medullary hormone in this cohort (Table 3). Modified Adler adhesion grade (≤ II = 0, ≥ III = 1) was included as a covariate in the multivariate Logistic regression model to control for tumor complexity. After adjustment for adhesion grade, surgical approach remained an independent risk factor (OR = 16.01, P = 0.002), and laparoscopic surgery increased the remission probability by 2.8 times compared with open surgery (OR = 2.8, 95%CI: 1.1-7.2). In the subgroup of patients with tumor diameter ≥ 6 cm and adhesion grade III (n = 34), laparoscopic surgery showed potential advantages (remission rate 70.6% vs 41.2%), but due to the limited sample size (n = 34), larger-scale studies are needed to verify its generalizability. Statistical power analysis showed a test power of 0.75 for this subgroup comparison (Tables 4 and 5). Diuretics were discontinued ≥ 4 weeks preoperatively, and CCBs ≥ 2 weeks preoperatively; for patients in whom discontinuation of diuretics or CCBs was considered clinically unsafe (n = 12, 11.8%), we added two levels of adjustment. First, the variable ‘inability to discontinue antihypertensive drugs’ was coded as a binary covariate (yes = 1, no = 0) and included in the multivariable logistic regression model. Second, daily doses of diuretics were converted to furosemide-equivalent mg/day and CCB doses to amlodipine-equivalent mg/day according to standard clinical potency tables, and these class-specific equivalent doses were entered into the model as continuous variables. We did not attempt to derive a single pooled “equivalent dose” across different drug classes (Table 6).

Table 3 Univariate logistic regression analysis of hypertension cure after adrenal cavernous hemangioma surgery.

Influencing factor comparison group	Comparison group	OR	95%CI	P value
Tumor diameter (cm)	Per 1-cm increase	0.68	0.51-0.90	< 0.01
Hypertension duration (year)	Per 1-year increase	0.59	0.43-0.81	< 0.01
RAAS hormones (pg/mL)	Per 1-SD increase	3.63	1.78-7.40	< 0.01
Adrenal medullary hormones (pg/mL)	Per 1-SD increase	2.98	1.31-6.76	< 0.001
Surgical approach (laparoscopy/open)	Laparoscopy vs open	2.8	1.1-7.2	< 0.01
Age (years)	> 50 years/< 50 years	1.05	0.48-2.30	0.90
Gender (male/female)	Male/female	1.12	0.49-2.55	0.79

Group comparisons for surgical approach used both χ² test (Table 2) and logistic regression. P values varied across models, but both methods indicated significant associations. RAAS: Renin-angiotensin-aldosterone system; OR: Odds ratio; CI: Confidence interval.

Table 4 Subgroup analysis of patients with tumor diameter ≥ 6 cm and adhesion grade III, n (%)/mean ± SD.

Variable	Laparoscopic group (n = 17)	Open group (n = 17)	OR (95%CI)	P value
Hypertension cure	12 (70.6)	7 (41.2)	2.68 (1.15-6.23)	0.024
Tumor diameter (cm)	7.2 ± 1.3	7.5 ± 1.1	0.89 (0.56-1.41)	0.62
RAAS hormones (pg/mL)	195.3 ± 42.7	198.6 ± 40.5	0.97 (0.71-1.33)	0.86
Adrenal medullary hormones (pg/mL)	142.5 ± 31.8	145.8 ± 29.7	0.96 (0.68-1.35)	0.81
Hypertension duration (years)	8.2 ± 2.5	8.5 ± 2.3	0.94 (0.61-1.45)	0.77

Subgroup inclusion criteria: Tumor diameter ≥ 6 cm and modified Adler adhesion grade III, with a total of 34 cases (accounting for 33.3% of the total sample). The baseline characteristics (tumor diameter, hormone levels, disease duration) of the two groups showed no significant differences by t-test or χ² test (all P > 0.05), indicating comparability. The cure rate in the laparoscopic group was 29.4% higher than that in the open surgery group. Logistic regression analysis showed that the surgical approach was an independent influencing factor (odds ratio = 2.68), suggesting that laparoscopic surgery remains advantageous in complex cases. RAAS: Renin-angiotensin-aldosterone system; OR: Odds ratio; CI: Confidence interval.

Table 5 Statistical power analysis for subgroup comparisons.

Subgroup	Sample size	Cure rate difference (%)	Statistical power (1-β)	P value
Open surgery	28	33.3 (vs laparoscopic)	0.72	< 0.008
Laparoscopic partial	18	33.4 (vs total)	0.58	< 0.659
Tumor > 6 cm + adhesio III	34	29.4	0.75	< 0.024

Cure rate difference refers to the difference in cure rates between the subgroup and the control group (e.g., the open surgery group vs the total laparoscopic resection group). Statistical power calculation was based on Cohen’s h formula (α = 0.05). In the subgroup with tumor ≥ 6 cm and grade III adhesions, the statistical power (1-β) is 0.75 (with α = 0.05), which necessitates cautious interpretation of the conclusion regarding the advantages of laparoscopy.

Table 6 Baseline characteristics of patients with and without preoperative drug withdrawal.

Indicator with	Withdrawal group (n = 90)	Non-withdrawal group (n = 12)	P value
Age (years)	48.5 ± 10.2	50.1 ± 9.8	0.612
Gender (male/female)	36/54	4/8	0.735
Tumor diameter (cm)	3.5 ± 1.3	3.8 ± 1.2	0.478
Hypertension duration (years)	4.2 ± 1.8	4.5 ± 1.9	0.631
RAAS hormones (pg/mL)	146.3 ± 38.5	152.7 ± 40.2	0.524
Adrenal medullary hormones (pg/mL)	96.8 ± 24.1	101.5 ± 25.3	0.497
Surgical approach (laparoscopy/open)	72/18	8/4	0.689

Patients unable to discontinue medications were defined as those who failed to stop taking diuretics or calcium channel blockers within 4 weeks before surgery (n = 12, accounting for 11.8%), among whom 8 cases were diuretic-dependent and 4 cases were calcium channel blocker-dependent. Baseline characteristics of the two groups showed no significant differences by independent samples t-test or χ² test (all P > 0.05), indicating comparability. RAAS: Renin-angiotensin-aldosterone system.

Multivariate logistic regression analysis of factors associated with hypertension remission after ACH surgery

Variables with P < 0.1 in univariate analysis were included in the multivariate logistic regression model, with variables screened using the backward stepwise method. The results showed that tumor diameter (OR = 3.54, 95%CI: 1.82-6.91, P = 0.001), RAAS hormone (OR = 3.63, 95%CI: 1.78-7.40, P = 0.001), adrenal medullary hormone (OR = 2.98, 95%CI: 1.31-6.76, P = 0.009), hypertension duration (OR = 20.23, 95%CI: 4.15-98.27, P < 0.001), and surgical approach (OR = 16.01, 95%CI: 2.78-91.82, P = 0.002) were independent risk factors. Interaction analysis showed that each 1 cm increase in tumor diameter increased the risk of uncontrolled hypertension after open surgery by 1.8 times (OR = 1.8, 95%CI: 1.1-2.9, P = 0.031), suggesting that tumor size may moderate the effect of surgical approach on prognosis. The equivalent dose of β-blockers (standardized to metoprolol mg/day) was included as a covariate in the model, and the results showed no significant effect on RAAS hormone (P = 0.17). In a sensitivity analysis excluding the 12 patients who could not discontinue diuretics or CCBs, the effect estimates for tumor diameter, RAAS hormone, adrenal medullary hormone, hypertension duration and surgical approach were very similar to those in the main analysis, with overlapping 95%CIs (Table 7). The direction of all associations remained unchanged, suggesting that the inability to discontinue these medications did not materially bias the primary findings. Nevertheless, we acknowledge that residual confounding by antihypertensive treatment cannot be completely eliminated (Table 7). Based on this, a combined predictive model was constructed: L = adrenal medullary hormone + 1.263 × tumor diameter + 1.290 × RAAS hormone + 2.768 × surgical approach + 5.611 × hypertension duration (surgical approach coded as 0 for laparoscopic, 1 for open). To further control the confounding effect of tumor complexity on surgical approach, 1:1 nearest neighbor propensity score matching was performed using tumor diameter, modified Adler adhesion grade, and RAAS hormone levels as covariates (caliper = 0.03). After matching, baseline characteristics of the two groups were well-balanced, with standardized differences of tumor diameter, adhesion grade, and RAAS hormone levels all < 0.1 (Table 8). The remission rate in the laparoscopic group (n = 30) remained significantly higher than that in the open group (n = 18; 83.3% vs 50.0%), and logistic regression analysis showed that open surgery was an independent risk factor for uncontrolled hypertension, suggesting that the advantage of laparoscopic surgery in postoperative hypertension remission independent of baseline tumor characteristics.

Table 7 Effect of non-withdrawal status as covariate in multivariate model.

Variable	β	OR	95%CI	Ward value	P value
Tumor diameter (cm)	1.258	3.5	1.82-6.91	10.48	0.001
RAAS hormones (pg/mL)	1.285	3.63	1.78-7.40	10.62	0.001
Adrenal medullary hormones (pg/mL)	1.089	2.98	1.31-6.76	6.74	0.009
Hypertension duration (years)	3.012	20.23	4.15-98.27	16.53	< 0.001
Surgical approach (laparoscopy/open)	2.769	16.01	2.78-91.82	9.36	0.002
Tumor × surgery	0.587	1.80	1.12-2.90	4.52	0.031
Diuretic/CCB non-withdrawal	0.173	1.19	0.58-2.45	2.56	0.672

In the multivariate model, “inability to discontinue medications” was included as a covariate (1 = inability to discontinue, 0 = discontinued). The results showed that it had no significant effect on the postoperative cure of hypertension (P = 0.672). The equivalent dose of diuretics (standardized to furosemide mg/day) and calcium channel blocker dose (standardized to amlodipine mg/day) were included in the model as continuous variables, which did not significantly alter the odds ratio values of other variables (fluctuation range < 3%). RAAS: Renin-angiotensin-aldosterone system; CCB: Calcium channel blocker; OR: Odds ratio; CI: Confidence interval.

Table 8 Baseline characteristics comparison between laparoscopic and open surgery groups after propensity score matching, n (%)/mean ± SD.

Variable	Group	n	Mean/proportion	SD	Statistical test (post-matching)
Tumor diameter (cm)	Laparoscopic	30	3.5 ± 1.2	0.06	t = 0.58, P = 0.56
	Open	30	3.8 ± 1.3
RAAS hormones (pg/mL)	Laparoscopic	30	145.2 ± 36.1	0.07	t = 0.69, P = 0.49
	Open	30	149.8 ± 37.5
Adhesion grade (≥ III)	Laparoscopic	30	3 (10.0)	0.05	χ2 = 0.42, P = 0.52
	Open	30	4 (13.3)
Hypertension duration (years)	Laparoscopic	30	4.3 ± 1.4	0.08	t = 0.74, P = 0.46
	Open	30	4.5 ± 1.5
Adrenal medullary hormones (pg/mL)	Laparoscopic	30	94.7 ± 23.1	0.09	t = 0.87, P = 0.39
	Open	30	98.5 ± 24.3

After matching, the standardized mean differences of all variables were < 0.1, indicating that the baseline characteristics were balanced. The cure rate in the laparoscopic group (80.0%, 24/30) was significantly higher than that in the open surgery group (53.3%, 16/30). Logistic regression analysis showed that the surgical method was an independent risk factor (odds ratio = 2.65, 95% confidence interval: 1.17-6.03, P = 0.021). RAAS: Renin-angiotensin-aldosterone system.

Diagnostic value of combined predictors and other independent factors for hypertension remission after ACH surgery

Evaluation of the diagnostic efficacy (sensitivity, specificity, positive predictive value/positive predictive value, negative predictive value/negative predictive value) of the combined predictor and other independent factors showed: The combined predictor had a sensitivity of 82.7%, specificity of 86.1%, positive predictive value of 85.7%, and negative predictive value of 83.3%; hypertension duration had a sensitivity of 78.8% and specificity of 79.4%; tumor diameter had a sensitivity of 76.9% and specificity of 77.6%; surgical approach had a sensitivity of 73.1% and specificity of 75.3%; RAAS hormone had a sensitivity of 65.4% and specificity of 68.2%; adrenal medullary hormone had a sensitivity of 61.5% and specificity of 64.7% (Table 9).

Table 9 Diagnostic efficacy comparison between combined predictor and independent factors.

Variable	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
Combined predictor	82.7	86.1	85.7	83.3
Hypertension duration (years)	78.8	79.4	76.5	81.2
Tumor diameter (cm)	76.9	77.6	74.3	79.8
Surgical approach (laparoscopy/open)	73.1	75.3	71.1	77.5
RAAS hormones (pg/mL)	65.4	68.2	63.5	70.1
Adrenal medullary hormones (pg/mL)	61.5	64.7	60.2	66.3

Sensitivity, specificity, positive predictive value and negative predictive value are expressed as percentages. The combined predictor corresponds to the multivariable model integrating tumor diameter, hypertension duration, renin-angiotensin-aldosterone system hormone, adrenal medullary hormone and surgical approach. RAAS: Renin-angiotensin-aldosterone system; PPV: Positive predictive value; NPV: Negative predictive value.

Receiver operating characteristic curves of combined predictors and other independent factors for predicting hypertension remission after ACH surgery

Receiver operating characteristic curve analysis showed that the AUC of the combined predictor was 0.872 (95%CI: 0.791-0.934), significantly higher than that of individual factors: Hypertension duration AUC = 0.813 (95%CI: 0.712-0.895), tumor diameter AUC = 0.796 (95%CI: 0.688-0.882), surgical approach AUC = 0.725 (95%CI: 0.598-0.831), RAAS hormone AUC = 0.650 (95%CI: 0.513-0.768), adrenal medullary hormone AUC = 0.632 (95%CI: 0.487-0.763); after excluding the adrenal medullary hormone variable, the model AUC decreased to 0.854 (95%CI: 0.768-0.915), suggesting a high contribution of this variable to the model (Table 10). The maximum Youden index of the combined predictor was 0.616, corresponding to a sensitivity of 0.827 and specificity of 0.861, with the optimal cutoff value determined as 12.57, i.e., an l value > 12.57 indicates high accuracy in predicting uncontrolled postoperative hypertension (Figure 2).

Figure 2 Comparison of area under the curve values between the combined predictor and individual factors. The area under the curve of the combined predictor was 0.872 (95% confidence interval: 0.791-0.934), which was significantly higher than that of each individual factor (P < 0.05, Delong test). ROC: Receiver operating characteristic; ACH: Adrenal cavernous hemangioma; RAAS: Renin-angiotensin-aldosterone system.

Table 10 Receiver operating characteristic curve analysis of combined predictor and independent factors.

Variable	AUC	95%CI (lower limit)	95%CI (upper limit)	P value	Sensitivity	Specificity	Youden’s index	Cutoff value
Combined predictor	0.872	0.791	0.934	< 0.001	0.827	0.861	0.616	12.57
Tumor diameter (cm)	0.813	0.712	0.895	0.003	0.788	0.794
Surgical approach (laparoscopy/open)	0.796	0.688	0.882	0.005	0.769	0.776
RAAS hormones (pg/mL)	0.725	0.598	0.831	0.021	0.731	0.753
Adrenal medullary hormones (pg/mL)	0.65	0.513	0.768	0.045	0.654	0.682
Tumor diameter (cm)	0.632	0.487	0.763	0.057	0.615	0.647
Combined predictor (excluding AMH)	0.854	0.768	0.915	0.036	0.805	0.842	0.647	11.89

Surgical approach was coded as laparoscopy = 0 and open = 1 in the combined predictor. Delong test showed the combined predictor had a significantly larger area under the curve than any single factor (P < 0.05). RAAS: Renin-angiotensin-aldosterone system; AMH: Anti-Mullerian hormone; CI: Confidence interval; AUC: Area under the curve.

Multivariate analysis of 5 independent variables may have a risk of overfitting due to insufficient sample size. Validation via 1000 bootstrap resampling showed that the fluctuation range of OR values for each variable was < 5%, and the model AUC fluctuation range under 95%CI was 0.856-0.889, consistent with the original AUC = 0.872, suggesting acceptable model stability (Table 11). In addition, to further verify model stability, the data were divided using 10-fold cross-validation, with 10% of the samples as the test set and the rest as the training set each time, and the average AUC was calculated after 10 repetitions. The results showed that the fluctuation range of variable OR values was < 3%, and the model AUC fluctuation range under 95%CI was 0.858-0.895, indicating significantly improved stability. However, some variables had relatively wide CIs, likely due to the limited sample size.

Table 11 Stability of odds ratio values for each variable verified by bootstrap resampling (1000 iterations).

Variable	Original OR	Bootstrap mean	95%CI lower	95%CI upper	Fluctuation range (%)
Tumor diameter (cm)	3.54	3.48	3.36	3.72	1.7
RAAS hormones (pg/mL)	3.63	3.59	3.42	3.78	1.1
Adrenal medullary hormones (pg/mL)	2.98	2.95	2.83	3.07	1.0
Hypertension duration (years)	20.23	20.01	19.25	21.08	1.1
Surgical approach (laparoscopy/open)	116	15.89	15.23	16.57	0.7

The non-parametric bootstrap (1000 resamples) was used to assess the stability of odds ratios for each variable. Results showed all odds ratios fluctuated by < 5%, indicating stable model parameters. RAAS: Renin-angiotensin-aldosterone system; OR: Odds ratio; CI: Confidence interval.

Diagnostic efficacy of the combined predictor

The AUC of the combined predictor was 0.872, significantly higher than that of hypertension duration (0.813, P = 0.02), tumor diameter (0.796, P = 0.01), and other independent factors. The calibration curve showed good consistency between the predicted probability and the actual remission rate, with a slope of 0.98 ± 0.03 (95%CI 0.92-1.04), intercept of 0.02 (95%CI: -0.05 to 0.09), and Brier score of 0.12. 5-fold cross-validation showed an AUC fluctuation range of 0.851-0.893, indicating good model stability. The maximum Youden index of 0.616 corresponded to a cutoff value of 12.57, and when the l value > 12.57, the accuracy of predicting uncontrolled postoperative hypertension was 85.7% (Figure 3).

Figure 3 Calibration curve of the combined predictor. The curve showed good agreement between predicted probabilities and actual remission rates (slope = 0.98, intercept = 0.02), indicating no significant over fitting or under fitting. The critical value l = 12.57 achieved 85.7% accuracy in predicting failure of postoperative hypertension remission. AUC: Area under the curve.

Clinical benefit of decision curve analysis

Decision curve analysis showed that when the threshold probability was 0.2-0.8, the net benefit of the combined predictor was significantly higher than that of all single factors (tumor diameter, RAAS hormone, hypertension duration, surgical approach), suggesting that its utility in clinical decision-making is superior to traditional indicators (Figure 4). When the threshold probability > 0.3, the net benefit of the combined factor prediction model was 15% higher than that of single factors; at a threshold probability of 0.5, the net benefit of the combined predictor was 0.37 (95%CI: 0.31-0.43), 37% higher than that of tumor diameter (0.27), 68% higher than that of RAAS hormone (0.22), 16% higher than that of hypertension duration (0.32), and 48% higher than that of surgical approach (0.25; Table 12). This indicates that using the combined predictor to guide treatment decisions can maximize the net benefit to patients within the moderate risk threshold range.

Figure 4 Decision curve analysis showing the net benefit of the combined predictor vs individual factors for postoperative hypertension remission. The black solid line with gray shading represents the combined predictor (95% confidence interval), while colored dashed lines denote individual factors. The red and gray dashed lines indicate treat-all and treat-none strategies. RAAS: Renin-angiotensin-aldosterone system.

Table 12 Net benefit of combined predictor vs individual factors at different threshold probabilities.

Threshold probability	Combined predictor		Tumor diameter		RAAS hormones		Hypertension duration		Surgical approach
	Net benefit	95%CI	Net benefit	95%CI	Net benefit	95%CI	Net benefit	95%CI	Net benefit	95%CI
0.2	0.18	0.12-0.24	0.10	0.05-0.15	0.08	0.03-0.13	0.12	0.07-0.17	0.09	0.04-0.14
0.3	0.25	0.19-0.31	0.18	0.12-0.24	0.15	0.09-0.21	0.20	0.14-0.26	0.16	0.10-0.22
0.4	0.32	0.26-0.38	0.23	0.17-0.29	0.19	0.13-0.25	0.27	0.21-0.33	0.21	0.15-0.27
0.5	0.37	0.31-0.43	0.27	0.21-0.33	0.22	0.16-0.28	0.32	0.26-0.38	0.25	0.19-0.31
0.6	0.40	0.34-0.46	0.30	0.24-0.36	0.24	0.18-0.30	0.35	0.29-0.41	0.28	0.22-0.34
0.7	0.41	0.35-0.47	0.32	0.26-0.38	0.25	0.19-0.31	0.36	0.30-0.42	0.30	0.24-0.36
0.8	0.39	0.33-0.45	0.31	0.25-0.37	0.24	0.18-0.30	0.34	0.28-0.40	0.29	0.23-0.35

Net benefit calculation: Net Benefit = true positive rate - [false positive rate × (threshold probability/(1 - threshold probability)]. The net benefit is 0 when all patients receive treatment and also 0 when none receive treatment. Within the threshold probability range of 0.2-0.8, the net benefit of the combined predictor was significantly higher than that of any single factor (all P < 0.05, verified by Bootstrap method). CI: Confidence interval; RAAS: Renin-angiotensin-aldosterone system.

DISCUSSION

ACH is a rare benign vascular tumor that is typically hormonally non-functioning[1-4], and most reported cases are either normotensive or hypertensive due to coexisting conditions such as essential hypertension[7,8]. Isolated reports have described ACH accompanied by hormone excess, but these remain exceptional. Therefore, our multicenter study does not attempt to establish causality between ACH and hypertension; instead, we focus on ACH patients with documented preoperative hypertension and explore which clinical characteristics are associated with postoperative blood pressure remission vs persistence[22]. This study conducted long-term follow-up of 102 ACH patients who underwent surgical treatment and found that 41.2% of patients had unresolved hypertension postoperatively, indicating that although surgical resection is the preferred treatment, a considerable proportion of patients still have poor postoperative blood pressure control. Unlike functional adrenal tumors such as pheochromocytoma, the pathophysiological mechanism of ACH-related hypertension is significantly complex.

From the perspective of pathophysiological mechanisms, the mechanical compression effect of the tumor and the activation of the neurohumoral system form a synergistic pathogenic network: (1) Activation of the RAAS: Direct compression of the renal artery by tumors > 6 cm leads to excessive RAAS activation. For example, animal models have shown that adrenal tumor compression of the renal artery can increase renin secretion by 2.3 times[23]; even for tumors < 6 cm, the abnormal hemodynamics of cavernous vascular lacunae may activate RAAS through renal artery stretch receptors. In this study, 3 non-remission patients with tumors < 4 cm had a 2.1-fold increase in RAAS hormone levels, supporting this mechanism. In addition, immunohistochemical analysis showed[24] that the vascular endothelial growth factor expression intensity in tumor tissues of the non-remission group was significantly higher than that of the remission group, and P = 0.015 after Bonferroni multiple test correction, supporting the vascular endothelial growth factor-mediated local ischemia-RAAS activation vicious cycle; (2) Abnormality of the sympathetic-catecholamine axis: Tumor compression of the adrenal medulla causes local ischemia or mechanical stimulation, activating sympathetic nerve endings to release catecholamines, forming a “mechanical compression-hormonal disorder” vicious cycle; and (3) Long-term effect of vascular remodeling: Each 1-year extension of hypertension duration increases the risk of unresolved postoperative blood pressure, which may be related to target organ damage (such as renal arteriosclerosis and vascular endothelial dysfunction) caused by long-term hypertension. Even after tumor resection, the impaired blood pressure regulation mechanism is difficult to recover[25].

Accurate preoperative differential diagnosis of ACH is crucial for formulating treatment strategies, and it differs significantly from pheochromocytoma in the following dimensions. Biochemical characteristics and hormonal regulation mechanisms: (1) Differences in medullary hormone levels: The preoperative and 6-month postoperative follow-up levels of medullary hormone in the non-remission group were significantly higher than those in the remission group, suggesting that elevated medullary hormone may act as a promoting factor for persistent hypertension. It is noteworthy that the decrease in hormone levels in the non-remission group after surgery was significantly lower than that in the remission group, reflecting irreversible vascular remodeling caused by long-term sympathetic activation; and (2) Differentiation of core pathogenic pathways: ACH is characterized by RAAS activation as the main pathological link (RAAS hormone level is an independent risk factor)[26], while pheochromocytoma relies on tumor-autonomous secretion of catecholamines to induce hypertension. There are essential differences in their hormonal regulation axes, and RAAS index detection can be used as a specific biochemical clue for ACH differential diagnosis. Clinical phenotype and optimization of diagnostic pathways: ACH-related hypertension is mainly persistent, lacking the typical paroxysmal headache and palpitations seen in pheochromocytoma[27-29]. Its diagnosis requires combining imaging features (CT/MRI showing cavernous vascular lacunae and non-enhancing nodules in the tumor) and pathological histological evidence (single layer of vascular endothelial cells, dilated lumens with thrombosis)[30-33]. In clinical practice, it is recommended to construct a “three-dimensional differential system”: (1) Biochemical detection: Simultaneously detect RAAS hormone (renin, angiotensin II) and 24-hour urinary catecholamines to distinguish RAAS-dominant and catecholamine-dominant hypertension; (2) Imaging assessment: Identify tumor vascular structural features using enhanced CT/MRI[34,35]; and (3) Dynamic follow-up: Monitor the changing trend of medullary hormone after surgery to assess the degree of vascular remodeling[36,37]. ACH is characterized by RAAS activation (OR = 3.63), while pheochromocytoma relies on tumor-autonomous catecholamine secretion, with essential differences in their hormonal regulation axes. Clinically, differentiation can be achieved by simultaneous detection of RAAS hormone and 24-hour urinary catecholamines, which can improve the preoperative identification accuracy from 68% to 89%[26,35].

This study is the first to construct a predictive model for hypertension cure after ACH surgery. We found that hypertension duration is the strongest independent predictor, with a 41% decrease in cure risk for each 1-year extension. This may be related to target organ damage from long-term hypertension, such as renal arteriosclerosis and vascular endothelial dysfunction. Even after tumor removal, the impaired blood pressure regulation mechanism is difficult to recover. The choice of surgical approach also significantly affects the prognosis: The hypertension remission rate in the laparoscopic surgery group was significantly better than that in the open surgery group when technically feasible, which may be related to the following mechanisms: (1) The magnified view in laparoscopy allows precise preservation of the adrenal central vein, avoiding venous injury caused by traction in open surgery and reducing postoperative ischemic hormonal disorders[36-38]. Interaction term (tumor × surgery) analysis showed that each 1 cm increase in tumor diameter increased the risk of uncontrolled hypertension after open surgery by 1.8 times, suggesting that the prognostic disadvantage of open surgery is more significant for larger tumors. In addition to tumor complexity (large diameter, strong adhesion), it may be related to extensive damage to the adrenal central vein and surrounding tissues in open surgery[39]. However, literature confirms that ACH ≥ 10 cm can be resected laparoscopically, suggesting that the model needs dynamic adjustment based on tumor size and adhesion grade; and (2) The retroperitoneal approach avoids abdominal adhesions and reduces mechanical damage to residual adrenal tissue. Despite adjustment for adhesion grade and propensity score matching, this retrospective study cannot fully exclude residual confounding by surgeon experience or center surgical volume. Future prospective multicenter studies are needed to further validate the advantages of laparoscopic surgery for ACH-related hypertension. In this retrospective cohort, the combined predictor L showed good discriminative ability for postoperative hypertension outcomes, and a cutoff of 12.57 derived from the Youden index allowed us to distinguish patients at higher risk vs lower risk of persistent hypertension. At present, however, the l score should be regarded solely as a prognostic tool. The cutoff value of 12.57 has not been prospectively validated, and there is no biological or interventional evidence to support using this threshold to mandate specific antihypertensive regimens. Therefore, our model may be useful for risk estimation and for generating hypotheses, but decisions regarding pharmacologic management should continue to follow established hypertension guidelines and individualized clinical judgment[20]. In a prospective study of ACH patients, perindopril (4 mg/day) combined with metoprolol (47.5 mg/day) can reduce the risk of postoperative hypertension by 31%, supporting the application of this intervention strategy in high-risk patients, with blood pressure and hormone levels monitored every 3 months after surgery. In this study, the decrease in RAAS hormone levels in the laparoscopic group after surgery was significantly greater than that in the open group, confirming the protective effect of minimally invasive surgery on endocrine function. In addition, elevated adrenal medullary hormone suggest that tumor compression may activate medullary cells, maintaining hypertension through the sympathetic excitation-catecholamine secretion pathway. The remission rate was significantly lower in patients with tumor diameter > 6 cm, suggesting that early surgical intervention is crucial for preserving blood pressure regulation function. To control confounding factors, tumor diameter and tumor adhesion degree were adjusted in univariate analysis, and surgical approach was further included as an independent variable in the multivariate model. The results showed that after controlling for tumor diameter, hormone levels, and other variables, open surgery remained an independent risk factor for unresolved postoperative hypertension, suggesting that the low remission rate in the open surgery group is not only related to tumor complexity but also may be due to more extensive damage to adrenal function in open surgery.

Compared with the postoperative hypertension prediction models for pheochromocytoma and primary aldosteronism[40,41], the combined model in this study, which integrates tumor physical factors (diameter), biochemical indicators (RAAS/medullary hormone), and surgical approach, has better predictive efficacy (Table 13). In addition, this study is consistent with the conclusion of Tetti et al[42] that multi-factor models improve prediction accuracy and aligns with the multi-omics research direction advocated by the MIH-SPARK consortium[43]. Compared with single factors, the combined model significantly improves predictive efficacy, which is particularly important for preoperative assessment: For patients with l value > 12.57 (high-risk group), future clinical research could explore whether preoperative blood pressure intervention strategies improve outcomes - this serves as a hypothesis for prospective studies. Clinical treatment should follow current hypertension guidelines and be tailored to individual patient circumstances[44-46]. A prospective study by Li and Wan[41] in ACH patients confirmed that perindopril (4 mg/day) combined with metoprolol (47.5 mg/day) can reduce the risk of postoperative hypertension by 31%, supporting the application of this strategy in high-risk patients. It is noteworthy that the model’s ability to capture both RAAS activation and medullary hormone release gives it an advantage over the primary aldosteronism model, which relies solely on aldosterone indicators, in the differential diagnosis of ACH and pheochromocytoma. For example, when the preoperative differential diagnosis threshold probability is set to 0.5, the net benefit of this model reaches 0.37, 27.6% higher than that of the primary aldosteronism model, suggesting higher clinical decision-making value in the diagnosis and treatment of complex adrenal tumors.

Table 13 Comparisons of area under the curve for predictive model of hypertension cure after adrenal cavernous hemangioma surgery and predictive models for other related diseases.

Predictive model	AUC (95%CI)	Predictor	Research cohort	P value
This study	0.872 (0.791-0.934)	Tumor diameter/hormone/surgical approach/disease course	ACH (n = 102)
Ge et al[40]	0.996 (0.990-1.000)	Age/duration of hypertension/preoperative systolic blood pressure/preoperative 24-hour urinary catecholamine level/maximum tumor diameter/combined metabolic syndrome/long-term use of multiple antihypertensive drugs before surgery	PHEO (n = 259)	0.0211
Li and Wan[41]	Not available	Duration of hypertension/body mass index/serum potassium level/estimated glomerular filtration rate	PA (n = 1190)	0.0031

¹Delong test, bootstrap = 1000.

AUC: Area under the curve; CI: Confidence interval; ACH: Adrenal cavernous hemangioma; PHEO: Pheochromocytoma; PA: Primary aldosteronism.

This study constructed a combined predictive model based on a multicenter retrospective cohort from 26 tertiary hospitals in China. Although stability was verified by bootstrap resampling, there are still limitations such as small sample size and geographical restrictions. This study initially included only hypertensive ACH patients, introducing potential selection bias. Future studies are recommended to: (1) Conduct prospective multicenter randomized controlled clinical studies, including clinical data from international centers in different regions and with different medical levels for large-sample validation; (2) Hormone-related predictive factors were based on single preoperative measurements. Despite standardized sampling and repeatability controls, hormone indices exhibit inherent physiological variability, which may affect predictive factor stability. Future studies should adopt multiple preoperative measurements (e.g., 3 measurements within 1 week) and use the mean value to reduce variability, improving model reproducibility; and (3) Although antihypertensive drugs were discontinued according to a standardized protocol supplement molecular biological indicators such as RAAS system gene polymorphisms (e.g., RAAS activation in patients with angiotensin-converting enzyme insertion/deletion gene DD type is 2.1 times higher than that in II type, P = 0.015)[26,41,45,46]. Differences in baseline characteristics of patients in different regions (such as hypertension control and comorbidity incidence) and disparities in medical technology (such as the proportion of laparoscopic surgery) may affect the generalizability of the model. It is necessary to conduct multicenter large-sample studies, extend the follow-up period, combine transcriptomics and genetic testing to reveal pathological mechanisms, and update variable weights to improve model adaptability. Second, although patients with overt Cushing’s syndrome or cortisol-producing adenomas were excluded based on clinical and biochemical evaluation, a standardized 1-mg overnight dexamethasone suppression test was not systematically performed for all patients across all centers. Therefore, mild autonomous cortisol secretion cannot be completely excluded and may have contributed to persistent hypertension in a subset of cases. Third, although antihypertensive drugs were discontinued according to a standardized protocol, the washout period for β-blockers (≥ 1 week) may not have been sufficient to completely eliminate their renin-suppressive effects, so a certain degree of residual bias in RAAS measurements cannot be ruled out.

CONCLUSION

Hypertension outcomes after ACH surgery were associated with tumor diameter, RAAS activity, adrenal medullary catecholamines (including epinephrine and norepinephrine), duration of hypertension, and surgical approach. In this multicenter retrospective cohort, laparoscopic total adrenalectomy was associated with a higher rate of postoperative blood pressure normalization than open surgery, suggesting a potential advantage of minimally invasive surgery when technically feasible; however, this observation requires confirmation in larger prospective studies. In this multicenter retrospective cohort, the combined predictive model demonstrated good discriminative ability for postoperative hypertension outcomes in ACH patients. Nevertheless, the model and its cutoff value require external validation and prospective evaluation before they can be generalized to broader populations or used to guide therapeutic decisions.

ACKNOWLEDGEMENTS

We thank all participating centers, funding bodies, patients and statisticians for their invaluable support to this study.