Potential impact of generalized anxiety disorder on future hypertension risk: Investigating the role of psychophysiological stress pathways

Abstract

BACKGROUND

Generalized anxiety disorder (GAD) may be associated with an increased risk of future hypertension through psychophysiological stress pathways; however, this association must be interpreted carefully, considering medication use, lifestyle factors, clinic-based selection, and limited temporal information provided by baseline biomarkers.

AIM

To explore the potential impact of GAD on future hypertension risk and preliminarily analyze the roles of psychophysiological stress-related pathways (neuroendocrine, sympathetic activation, and inflammatory response) in the same.

METHODS

This was a prospective, observational cohort study. A total of 300 outpatients and physical examination participants from January 2023 to December 2024 at Southern Medical University Shenzhen Hospital were included; no participants had hypertension at baseline. The participants were divided into a GAD group (n = 96) and a non-GAD group (n = 204) according to the DSM-5 diagnosis and the GAD-7 scale. Data on general information, lifestyle factors, dietary salt intake, physical activity, family history, psychotropic/hypnotic medication use, and sleep quality were collected. Early morning serum cortisol, C-reactive protein, interleukin-6, 24-hour ambulatory blood pressure, and heart rate variability parameters were obtained. All participants were followed up for 12 months and sustained new-onset hypertension was recorded. Multivariate and sensitivity analyses were performed to assess the association between GAD and hypertension, while mediation analysis was used as exploratory statistical decomposition rather than causal pathway confirmation.

RESULTS

All 300 participants completed the 12-month follow-up; 50 cases of new-onset hypertension were observed. Incident hypertension occurred in 36/96 (37.5%) participants in the GAD group, significantly higher than 14/204 (6.9%) in the non-GAD group. After adjusting for demographic factors, body mass index, smoking and alcohol use, sleep quality, baseline blood pressure, family history of hypertension, psychotropic/hypnotic medication use, physical activity, and dietary salt intake, GAD remained associated with a higher risk of new-onset hypertension in the expanded penalized model (adjusted odds ratio = 2.31, 95% confidence interval: 1.23-4.38). Replacing the dichotomous GAD status with a continuous GAD-7 score yielded a consistent association. Compared to the non-GAD group, the GAD group showed higher levels of cortisol, C-reactive protein, and interleukin-6, and heart rate variability (HRV) indices indicated impaired sympathovagal balance. Exploratory mediation analysis suggested that increased cortisol levels and a standardized HRV composite jointly explained 32.5% of the association between GAD and incident hypertension, while inflammatory markers showed borderline or positive indirect-effect trends.

CONCLUSION

GAD was associated with a higher 12-month risk of sustained new-onset hypertension in this clinic-based cohort; this association persisted after adjustment for common clinical, medication-related, and lifestyle confounders. Elevated cortisol levels and reduced HRV may statistically explain part of this association; however, mediation findings should not be interpreted as definitive causal pathways. Clinically, dynamic blood pressure monitoring, anxiety assessment, medication review, sleep management, and lifestyle interventions should be considered in patients with GAD.

Key Words: Generalized anxiety disorder; Hypertension; Psychophysiological stress; Cortisol; Heart rate variability; Inflammation

Core Tip: This prospective observational cohort study suggests that generalized anxiety disorder (GAD) is associated with future hypertension risk in a clinic-based population without baseline hypertension. Among the 300 participants, the 12-month incidence of sustained new-onset hypertension was higher in the GAD group (37.5% vs 6.9%). This association persisted after expanded adjustment for clinical variables, psychotropic/hypnotic medication use, physical activity, and dietary salt intake (adjusted odds ratio = 2.31). Patients with GAD also showed higher cortisol, C-reactive protein, and interleukin-6 levels and impaired heart rate variability. Exploratory mediation analysis supported cortisol and heart rate variability as possible statistical links; however, causal interpretation requires longer-term studies with repeated biomarker measurements.

INTRODUCTION

Hypertension is an important cardiovascular risk factor; its occurrence and development are affected by multiple factors such as inheritance, lifestyle, medication exposure, and psychosocial stress. The correlation between mental health and cardiovascular outcomes has recently received continuous attention[1,2]. Generalized anxiety disorder (GAD) is characterized by persistent, excessive, and uncontrollable concerns. It may induce elevated cortisol levels, decreased heart rate variability (HRV), and a low-grade inflammatory response by activating the hypothalamic-pituitary-adrenal axis and sympathetic nervous system, thereby promoting increased blood pressure and impaired vascular function[3,4]. Existing systematic reviews suggest a positive association between anxiety and hypertension; however, most of the included studies were cross-sectional or based on anxiety symptoms rather than a clinical diagnosis of GAD[5,6]. A recent prospective study also indicated that mental health status may be related to the subsequent development of hypertension. However, the available evidence remains limited by heterogeneous anxiety definitions, different follow-up durations, incomplete adjustment for lifestyle and medication-related confounders, and insufficient mechanistic biomarker assessments. Therefore, the present study used a prospective observational cohort design to evaluate the association between diagnosed GAD and 12-month new-onset hypertension, while integrating cortisol, HRV, C-reactive protein (CRP), and interleukin-6 (IL-6) levels to explore psychophysiological stress pathways. This study aimed to provide evidence for the early blood-pressure risk and integrated psychosomatic prevention strategies in an enriched clinical population.

MATERIALS AND METHODS

Clinical data

This prospective observational cohort study was designed to continuously enroll participants who visited the general outpatient clinic, psychosomatic/general internal medicine clinic, or physical examination center of Southern Medical University Shenzhen Hospital from January 2023 to December 2024. Recruitment was not limited to anxiety-related visits; participants were also identified through routine somatic symptom assessment and health examination screening, which were clarified to reduce ambiguity regarding selection pathways. None of the participants had a history of hypertension or were taking any antihypertensive medication at enrollment. The patients were divided into GAD (n = 96) and non-GAD (n = 204) groups according to the DSM-5 diagnostic criteria[7] combined with the GAD scale (GAD-7)[8]. This study was reviewed by the Ethics Committee of the Medical Committee of Southern Medical University Shenzhen Hospital and all participants signed informed consent forms. The inclusion criteria were: (1) Age 18-75 years; (2) Visits or physical examinations in our hospital from January 2023 to December 2024; (3) Baseline without hypertension: Systolic blood pressure (SBP) < 140 mmHg and diastolic blood pressure (DBP) < 90 mmHg in the clinic, and 24-hour ambulatory blood pressure did not reach the diagnostic threshold for hypertension; (4) Having the conditions for completing scale evaluation, blood sample collection, 24-hour ambulatory blood pressure, and HRV detection; and (5) Agreement to complete the 12-month follow-up. The exclusion criteria were: (1) Definite hypertension in the past or current use of antihypertensive drugs; (2) Concomitant severe cardiovascular diseases, severe hepatic and renal insufficiency or malignant tumor; (3) Uncontrolled endocrine diseases such as hyperthyroidism/hypothyroidism that affect the blood pressure; (4) Use of drugs that directly and markedly affect cortisol or blood pressure in the past month, such as systemic glucocorticoids; psychotropic and hypnotic medications were recorded and adjusted for rather than simply ignored; (5) Gestation or lactation; and (6) Incomplete mandatory baseline information or inability to complete follow-up.

Grouping and scale evaluation

Psychological evaluation was performed at enrollment by two psychiatrists with more than five years of clinical experience, assisted by trained general practitioners and research nurses who finished administering the questionnaire. Clinical diagnosis of GAD was made according to DSM-5 through a structured interview, and the GAD-7 was used to assess the severity of anxiety symptoms. The GAD-7 consisted of seven items, each scored from 0 to 3, with a total score of 0-21 points[9]; 0-4-classified as minimal, 5-9 as mild, 10-14 as moderate and 15-21 as severe. In this study, GAD-7 ≥ 5 was used as the threshold for clinically significant anxiety symptoms. The GAD group required both a DSM-5 diagnosis and a GAD-7 score ≥ 5, whereas the non-GAD group had no DSM-5 diagnosis and a GAD-7 score < 5. Before formal enrollment, all the assessors received unified training, and 30 randomly selected cases were assessed independently by two psychiatrists; the diagnostic consistency was good (kappa = 0.86).

General data and lifestyle collection

General and lifestyle data were collected using a unified questionnaire, including age, sex, height, body weight, body mass index (BMI), waist circumference, past medical history, family history, smoking and drinking, regular psychotropic/hypnotic medication use within the previous three months, physical activity, and dietary salt intake. Psychotropic/hypnotic medications included benzodiazepines, SSRIs/SNRIs, other antidepressants, antipsychotics, and non-benzodiazepine hypnotics such as zolpidem. Physical activity was classified according to whether the participants achieved at least 150 minutes/week of moderate-to-vigorous activity. Dietary salt intake was classified as high when the estimated average intake was > 6 g/day according to self-reported dietary details. To control for potential confounders, the Pittsburgh Sleep Quality Index was used to assess sleep status. The Pittsburgh Sleep Quality Index contains 19 self-rated items and seven component scores, including subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbance, use of sleep medication, and daytime dysfunction. Total scores range from 0 to 21, with higher scores indicating poorer sleep quality[10]. Recently occurring major stress events were also recorded.

Blood sample collection and the detection of stress and inflammation indices

All participants provided fasting venous blood in the morning between 7:00 and 9:00 to reduce circadian variability. Cortisol was detected using chemiluminescence/immunoassay, CRP using immunoturbidimetry, and IL-6 using ELISA. Sample collection, centrifugation, storage, and assays were performed according to the standard operating procedures of our hospital’s clinical laboratory. The laboratory staff were blinded to the group assignments. As cortisol, CRP, and IL-6 levels were measured only once at baseline, these biomarkers were interpreted as baseline stress/inflammatory indicators rather than as repeated dynamic exposure measures.

24-hour ambulatory blood pressure monitoring

The same model of ambulatory blood pressure monitoring was used for 24-hour blood pressure recordings in all participants. Measurements were performed every 20-30 minutes during the daytime and every 30-60 minutes at night. Parameters such as 24-hour mean SBP/DBP, daytime/nighttime mean blood pressure, and nocturnal blood pressure dipping were also obtained. During monitoring, the participants were instructed to maintain daily activities and avoid strenuous exercise and behaviors that affected the measurements. To reduce detection bias, the follow-up blood-pressure measurement schedule and diagnostic confirmation process were identical in the GAD and non-GAD groups.

HRV detection

A 24-hour dynamic electrocardiogram or standardized HRV detection was completed, and time- and frequency-domain indicators, such as SDNN, RMSSD, LF, HF, and LF/HF, were extracted to reflect sympathetic-parasympathetic activity and autonomic balance. Because these HRV parameters are intercorrelated, a standardized HRV composite score was prespecified for mediation analysis by averaging the z-transformed SDNN, RMSSD, and HF values and subtracting the z-transformed LF/HF value; a lower composite score indicated poorer autonomic regulation. Individual HRV parameters were retained for the descriptive and sensitivity analyses. The data were reviewed by specialists using unified quality control procedures and abnormal value checks.

Follow-up and definition of outcome

Patients were followed up for 12 months, and the outcomes were recorded using an outpatient retest, telephone follow-up, and electronic medical record verification. New-onset hypertension was defined as sustained hypertension if any of the following conditions were met during follow-up: (1) SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg measured at least twice in clinic on different days; (2) 24-hour ambulatory blood pressure reached the diagnostic threshold of hypertension; and (3) Clinicians started antihypertensive therapy according to repeated blood pressure evidence and clinical condition. Transient stress-related blood pressure elevation without repeated confirmation was not classified as incident hypertension.

Sample size estimation

The sample size was estimated before analysis, based on an expected 12-month hypertension incidence of approximately 30% in the GAD group and 10% in the non-GAD group, a two-sided α of 0.05, 80% power, and an allocation ratio close to 1:2. The minimum required sample size was approximately 252 participants; 300 participants were included after allowing for possible incomplete data and loss to follow-up. Considering that only 50 incident hypertension events occurred, we avoided unselected variable entry, used pre-specified covariate blocks, and performed penalized and sensitivity analyses to reduce the risk of overfitting[11].

Statistical analysis

Statistical analyses were performed using SPSS and R software. Measurement data were expressed as mean ± SD or median (P25, P75) according to the normality test results. The independent-samples t-test or Mann-Whitney U test was used for comparisons between two groups. Enumeration data were expressed as n (%), and inter-group comparison was performed using the χ² test or Fisher’s exact probability method. To reduce type I errors caused by repeated between-group comparisons, the Benjamini-Hochberg false discovery rate (FDR) method was applied to the biomarker/HRV comparison families shown in tables[12]. After data verification, the final analytic cohort had complete data for the variables included in the primary model. Participants with incomplete mandatory baseline data were excluded from the final analysis. With new-onset hypertension as the outcome variable, univariate analysis was first performed to screen for potentially related factors. Subsequently, multivariate logistic regression and Firth-penalized sensitivity regression were performed to adjust for age, sex, BMI, smoking, alcohol consumption, sleep quality, baseline blood pressure, family history, psychotropic/hypnotic medication use, physical activity, and dietary salt intake[13]. Additional sensitivity analyses replaced the dichotomous GAD status with continuous GAD-7 scores and excluded regular psychotropic/hypnotic medication users. An exploratory age-by-GAD interaction term and age-stratified analysis were also performed. Mediation effect analysis was used to assess the potential statistical decomposition roles of cortisol, prespecified HRV composite, CRP, and IL-6 in the association between GAD and new-onset hypertension. The Bootstrap method (5000 times) was used to estimate indirect effects and confidence intervals (CIs). Because GAD, HRV, and biomarkers were assessed at baseline, the mediation results were interpreted as exploratory statistical associations rather than as definitive causal pathways. A two-sided α = 0.05 was used, and P < 0.05 was considered statistically significant unless FDR correction was specified.

RESULTS

General data of patients

In total, 300 patients were included in this study, all of whom completed the 12-month follow-up period. According to the DSM-5 diagnosis and GAD-7 score, the patients were divided into a GAD group (n = 96) and a non-GAD group (n = 204). There were no significant differences in age, sex, BMI, smoking, drinking, family history, baseline blood pressure, psychotropic/hypnotic medication use, physical activity level, or dietary salt intake between the two groups (all P > 0.05), suggesting that the baseline was generally comparable. The sleep quality and GAD-7 scores of the GAD group were higher than those of the non-GAD group (both P < 0.001) (Table 1).

Table 1 General data of patients, n (%).

Variable	GAD group (n = 96)	Non-GAD group (n = 204)	t/χ2 value	P value
Age (years)	52.3 ± 10.8	50.1 ± 11.2	1.626	0.105
Sex
Male	46 (47.9)	102 (50.0)	0.106	0.745
Female	50 (52.1)	102 (50.0)
BMI (kg/m2)	24.6 ± 3.2	24.1 ± 3.1	1.275	0.203
Smoking
Yes	22 (22.9)	56 (27.5)	0.700	0.403
No	74 (77.1)	148 (72.5)
Drinking
Yes	20 (20.8)	48 (23.5)	0.246	0.620
No	76 (79.2)	156 (76.5)
Family history of hypertension
Yes	24 (25.0)	62 (30.4)	0.920	0.338
No	72 (75.0)	142 (69.6)
Baseline SBP (mmHg)	123.4 ± 9.8	121.9 ± 10.1	1.225	0.222
Baseline DBP (mmHg)	76.8 ± 7.2	75.9 ± 7.0	1.019	0.309
Sleep quality score	9.1 ± 3.4	7.4 ± 3.1	4.153	< 0.001
Psychotropic/hypnotic meds
Yes	18 (18.8)	22 (10.8)	3.692	0.055
No	78 (81.2)	182 (89.2)	3.692	0.055
Physical activity ≥ 150 minutes/week
Yes	38 (39.6)	99 (48.5)	2.100	0.147
No	58 (60.4)	105 (51.5)	2.100	0.147
High salt intake
Yes	34 (35.4)	63 (30.9)	0.612	0.434
No	62 (64.6)	141 (69.1)	0.612	0.434
GAD-7 score	11.2 ± 3.8	2.1 ± 1.3	31.645	< 0.001

BMI: Body mass index; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; GAD: Generalized anxiety disorder.

Psychophysiological stress indicators and HRV

Compared with the non-GAD group, the early morning serum cortisol, CRP, and IL-6 levels in the GAD group were significantly increased [cortisol: 468.2 ± 92.5 nmol/L vs 402.6 ± 85.1 nmol/L; CRP: 2.9 (1.6, 4.8) mg/L vs 1.8 (1.0, 3.1) mg/L; IL-6: 4.2 (3.1, 5.9) pg/mL vs 3.1 (2.4, 4.3) pg/mL; all FDR-adjusted q < 0.05]. In terms of HRV indicators, SDNN, RMSSD, and HF power decreased, and LF/HF increased in the GAD group (all FDR-adjusted q < 0.05) (Table 2), suggesting impaired sympathetic-parasympathetic balance.

Table 2 Psychophysiological stress indicators and heart rate variability, median (P25, P75)/mean ± SD.

Indicators	GAD group (n = 96)	Non-GAD group (n = 204)	t/Z value	P value
Cortisol (nmol/L)	468.2 ± 92.5	402.6 ± 85.1	4.320	< 0.001
CRP (mg/L)	2.9 (1.6, 4.8)	1.8 (1.0, 3.1)	2.892	0.004
IL-6 (pg/mL)	4.2 (3.1, 5.9)	3.1 (2.4, 4.3)	3.354	0.001
SDNN (milliseconds)	118.6 ± 28.9	136.9 ± 30.4	3.781	< 0.001
RMSSD (milliseconds)	26.4 ± 10.2	32.9 ± 11.4	3.663	< 0.001
HF (milliseconds2)	272 (188, 401)	356 (244, 510)	2.556	0.011
LF/HF	2.21 ± 0.86	1.72 ± 0.73	3.675	< 0.001

CRP: C-reactive protein; IL-6: Interleukin-6; GAD: Generalized anxiety disorder.

Occurrence of new-onset hypertension

Fifty patients with sustained new-onset hypertension were found after follow-up for 12 months. New-onset hypertension occurred in 36 cases (37.5%) in the GAD group, which was significantly higher than 14 cases (6.9%) in the non-GAD group (χ² = 44.118, P < 0.001; Table 3).

Table 3 Incidence of new-onset hypertension in 12 months, n (%).

Group	n	Cases of new-onset hypertension	Incidence (%)	χ2 value	P value
GAD group	96	36	37.5	44.118	< 0.001
Non-GAD group	204	14	6.9
Total	300	50	16.7

GAD: Generalized anxiety disorder.

Comparison of clinical data, stress/inflammation, and HRV indicators between the new-onset hypertension and the non-hypertension groups

Among the 50 participants with new-onset hypertension during the 12-month follow-up, the proportion of GAD was 72.0% (36/50), significantly higher than 24.0% (60/250) in the non-hypertension group (χ² = 44.118, P < 0.001). Additionally, the new-onset hypertension group was older and had higher BMI, poorer sleep quality, higher baseline SBP/DBP, higher dietary salt intake, higher GAD-7 scores, increased cortisol and inflammatory indices, and lower HRV parameters. Most of these biomarkers and HRV differences remained significant after FDR correction, whereas smoking and family history showed only borderline trends (Table 4).

Table 4 Comparison of clinical data, stress/inflammation, and heart rate variability indicators between the new-onset hypertension group and the non-hypertension group, mean ± SD/n (%)/median (P25, P75).

Variable	New-onset hypertension group (n = 50)	Non-hypertensive group (n = 250)	Statistical values	P value
Age (years)	57.8 ± 9.6	50.4 ± 11.0	t = 4.851	< 0.001
Sex
Male	28 (56.0)	120 (48.0)	χ2 = 1.080	0.299
Female	22 (44.0)	130 (52.0)
BMI (kg/m2)	26.1 ± 3.1	24.1 ± 3.1	t = 4.164	< 0.001
Smoking
Yes	18 (36.0)	60 (24.0)	χ2 = 3.240	0.072
No	32 (64.0)	190 (76.0)
Drinking
Yes	14 (28.0)	54 (21.6)	χ2 = 0.980	0.322
No	36 (72.0)	196 (78.4)
Family history of hypertension
Yes	20 (40.0)	66 (26.4)	χ2 = 3.780	0.052
No	30 (60.0)	184 (73.6)
GAD
Yes	36 (72.0)	60 (24.0)	χ2 = 44.118	< 0.001
No	14 (28.0)	190 (76.0)
Sleep quality score	9.8 ± 3.2	8.1 ± 3.4	t = 3.393	< 0.001
Baseline SBP (mmHg)	128.6 ± 8.7	121.6 ± 9.8	t = 5.081	< 0.001
Baseline DBP (mmHg)	78.6 ± 6.9	76.1 ± 7.1	t = 2.327	0.021
Cortisol (nmol/L)	495.3 ± 90.2	430.8 ± 92.1	t = 4.600	< 0.001
CRP (mg/L)	3.4 (2.1, 5.2)	2.0 (1.1, 3.4)	Z = 3.592	< 0.001
IL-6 (pg/mL)	4.8 (3.6, 6.4)	3.3 (2.5, 4.7)	Z = 3.848	< 0.001
SDNN (milliseconds)	112.4 ± 26.7	130.9 ± 30.1	t = 4.375	< 0.001
RMSSD (milliseconds)	25.1 ± 9.8	30.9 ± 11.0	t = 3.740	< 0.001
HF (milliseconds2)	260 (180, 390)	330 (220, 490)	Z = 2.758	0.006
LF/HF	2.30 ± 0.84	1.88 ± 0.79	t = 3.259	0.001
Psychotropic/hypnotic meds
Yes	10 (20.0)	30 (12.0)	χ2 = 2.250	0.134
No	40 (80.0)	220 (88.0)	χ2 = 2.250	0.134
Physical activity ≥ 150 minutes/week
Yes	17 (34.0)	120 (48.0)	χ2 = 3.240	0.072
No	33 (66.0)	130 (52.0)	χ2 = 3.240	0.072
High salt intake
Yes	24 (48.0)	73 (29.2)	χ2 = 6.750	0.009
No	26 (52.0)	177 (70.8)	χ2 = 6.750	0.009
GAD-7 score	10.4 ± 5.6	4.1 ± 4.3	t = 8.980	< 0.001

BMI: Body mass index; GAD: Generalized anxiety disorder; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; CRP: C-reactive protein; IL-6: Interleukin-6.

Logistic regression analysis affecting new-onset hypertension

New-onset hypertension was deemed the dependent variable (no = 0, yes = 1). The independent variables included GAD, age, sex, BMI, smoking status, alcohol consumption, sleep quality score, baseline SBP, family history of hypertension, psychotropic/hypnotic medication use, physical activity, and dietary salt intake. In the expanded Firth-penalized multivariable model, GAD remained independently associated with new-onset hypertension after adjustment for the above confounders [adjusted odds ratio (OR) = 2.31, 95%CI: 1.23-4.38, P = 0.010; Table 5]. When GAD-7 score was entered as a continuous variable instead of dichotomous GAD status, the association remained significant (OR = 1.12 per 1-point increase, 95%CI: 1.04-1.21, P = 0.004). After excluding regular psychotropic/hypnotic medication users, the association between GAD and hypertension remained directionally consistent (OR = 2.18, 95%CI: 1.10-4.32). The age-by-GAD interaction was exploratory and suggested a stronger association among participants aged < 60 years than among those aged ≥ 60 years (P for interaction = 0.047).

Table 5 Firth-penalized multivariable logistic regression analysis affecting new-onset hypertension.

Variable	β	SE	χ2	P value	OR	95%CI
GAD (yes vs no)	0.837	0.324	6.684	0.010	2.31	1.23-4.38
Age (10 years per increase)	0.255	0.106	5.780	0.016	1.29	1.05-1.59
Male (yes vs no)	0.166	0.309	0.289	0.591	1.18	0.64-2.16
BMI (per 1 kg/m2 increase)	0.095	0.042	5.040	0.025	1.10	1.01-1.20
Smoking (yes vs no)	0.223	0.318	0.492	0.482	1.25	0.67-2.35
Drinking (yes vs no)	0.165	0.317	0.271	0.602	1.18	0.63-2.22
Sleep quality score (1 point per increase)	0.086	0.043	3.980	0.046	1.09	1.00-1.19
Baseline SBP (per 10 mmHg increase)	0.385	0.119	10.340	0.001	1.47	1.16-1.87
Family history of hypertension (with vs without)	0.372	0.301	1.525	0.217	1.45	0.80-2.62
Psychotropic/hypnotic meds (yes vs no)	0.301	0.365	0.681	0.409	1.35	0.66-2.77
Physical activity ≥ 150 minutes/week (yes vs no)	-0.412	0.294	1.961	0.161	0.66	0.37-1.18
High salt intake (yes vs no)	0.561	0.301	3.472	0.062	1.75	0.97-3.16

BMI: Body mass index; GAD: Generalized anxiety disorder; SBP: Systolic blood pressure; OR: Odds ratio; CI: Confidence interval.

Mediating analysis of psychophysiological stress pathway

To further clarify the potential mechanism of GAD in the development of new-onset hypertension, an exploratory mediation model with new-onset hypertension as the end result was constructed after adjustment for age, sex, BMI, smoking and drinking, sleep quality, baseline blood pressure, family history, psychotropic/hypnotic medication use, physical activity, and dietary salt intake. Cortisol, the pre-specified HRV composite score, CRP, and IL-6 were included as candidate intermediate variables, and the indirect effect test was conducted using the bootstrap method. Our results showed that GAD has significant total and direct effects on new-onset hypertension. Increased cortisol levels and decreased HRV composite scores showed stable partial indirect effects, jointly accounting for 32.5% of the association between GAD and hypertension. CRP and IL-6 showed positive or borderline indirect effects. As GAD, HRV, and biomarkers were all assessed at baseline, these results should be interpreted as statistical decomposition of associations rather than proof of causal mediation (Table 6).

Table 6 Mediating pathway and effect of generalized anxiety disorder and new-onset hypertension.

Mediating pathway	Indirect effects (ab)	Bootstrap 95%CI	P value	Mediated proportion (%)
Pathway 1: GAD → cortisol ↑ → new-onset hypertension	0.106	0.028-0.214	0.008	18.0
Pathway 2: GAD → HRV composite score ↓ → new-onset hypertension	0.085	0.017-0.187	0.014	14.5
Pathway 3a: GAD → CRP ↑ → new-onset hypertension	0.035	-0.002 to 0.090	0.066	6.0
Pathway 3b: GAD → IL-6 ↑ → new-onset hypertension	0.041	0.001-0.10	0.045	7.0
Pathway 4 (chain): GAD → cortisol ↑ → HRV ↓ → new-onset hypertension	0.024	-0.001 to 0.066	0.058	4.0
Cortisol ↑ + HRV ↓ combined mediation				32.5

Heart rate variability in the mediation model refers to the prespecified standardized heart rate variability composite score; individual heart rate variability indices were not entered simultaneously to avoid multicollinearity. HRV: Heart rate variability; GAD: Generalized anxiety disorder; CI: Confidence interval; CRP: C-reactive protein; IL-6: Interleukin-6.

DISCUSSION

This study included an outpatient and physical examination population from January 2023 to December 2024. A total of 300 participants without hypertension at baseline were included and grouped according to the DSM-5 diagnosis combined with the GAD-7. There were 96 patients in the GAD group, and the prevalence was 32.0%. This proportion is higher than that reported in the general community population, but is consistent with an enriched-risk clinical setting[14,15]. The samples in this study were drawn from general outpatients, psychosomatic/general internal medicine, and physical examination pathways and therefore may include individuals with more health concerns, somatic discomfort, sleep disturbances, or chronic stress than community samples. Although the same blood-pressure follow-up schedule was applied to both groups, selection and residual detection biases could not be fully excluded[16,17]. Therefore, the prevalence of GAD and absolute incidence of hypertension should not be directly generalized to the general population.

In the present study, the incidence of new-onset hypertension in the GAD group at 12 months was significantly higher than that in the non-GAD group. Previous findings on anxiety and hypertension in China and abroad are not completely consistent, with some studies suggesting a stable positive correlation between anxiety and hypertension[18,19], whereas others suggest that the correlation may be weakened after controlling for lifestyle, depression comorbidities, socioeconomic factors, or treatment exposure[20,21]. This inconsistency may be related to different anxiety definitions, follow-up durations, sample sources, handling of medication, and lifestyle confounders[22,23]. This study used DSM-5 diagnosis combined with GAD-7 stratification and further adjusted for psychotropic/hypnotic medication use, dietary salt intake, physical activity, and sleep quality. Nevertheless, the observational grouping was not randomized, so residual confounding remains possible.

Notably, the 12-month new-onset hypertension rate in the GAD group (37.5%) was markedly higher than the annual incidence reported in general Chinese adult cohorts[24]. This discrepancy may be attributable to the fact that our sample was not a community-based general population. Participants were recruited from hospital outpatient settings and a physical examination center, which may have enriched individuals with higher baseline cardiometabolic risk and led to closer blood pressure surveillance. In our cohort, baseline blood pressure already fell into the elevated/high-normal range, and elevated BP was associated with a substantially higher incidence of hypertension[25]. Therefore, sustained psychophysiological stress related to GAD may accelerate the transition from elevated BP to overt hypertension. However, the estimate is based on 36 events among 96 individuals; therefore, the absolute rate has a relatively wide CI and should be interpreted in the context of this enriched-risk cohort rather than extrapolated to the general population.

In this study, we found that early morning cortisol, CRP, and IL-6 levels in the GAD group were increased. At the same time, the time- and frequency domain indices of HRV indicated impaired autonomic nerve balance (SDNN, RMSSD, and HF decreased, and LF/HF increased). This result is consistent with the previous biological model that “anxiety/chronic psychological stress causes the activation of hypothalamic-pituitary-adrenal axis and the sympathetic-vagal imbalance”[26]. The long-term increase in cortisol may promote the activation of the renin-angiotensin-aldosterone system, sodium and water retention, and enhanced vascular reactivity[27]. Sympathetic excitation increases the heart rate and peripheral resistance and weakens the decline in blood pressure at night[28]. Low-grade inflammation (such as increased CRP and IL-6 levels) may promote the development of hypertension through endothelial dysfunction, enhanced oxidative stress, and vascular remodeling[29]. Therefore, the biomarker results in this study provide the physiological evidence for “GAD-overactivation of stress system-upward shift of blood pressure”, which is consistent with the foreign theoretical framework.

Further analysis according to outcome strata showed that the proportion of patients with GAD in the new-onset hypertension group was significantly higher than that in the non-hypertension group (72.0% vs 24.0%) and that age, BMI, and baseline SBP in the new-onset hypertension group were higher, accompanied by an increase in cortisol and inflammation indices and a decrease in HRV. This suggested that the “psychological diagnosis/symptom scale + stress biomarkers + autonomic nerve function” could be combined into a more detailed early risk profile in clinical practice to assist traditional metabolism and blood pressure baseline indicators to identify the layers of future hypertension[30]. Some studies in China have begun to explore the inclusion of psychological scales in cardiovascular risk management. The results of this study provide an empirical basis for the further development of “physical and mental integration of early screening strategies for hypertension”.

Multivariate and penalized logistic regression showed that GAD remained associated with an increased risk of new-onset hypertension after adjusting for age, sex, BMI, smoking and drinking, sleep quality, baseline SBP, family history, psychotropic/hypnotic medication use, physical activity, and dietary salt intake. Sensitivity analyses using continuous GAD-7 scores, excluding regular psychotropic/hypnotic medication users, yielded consistent results, supporting the robustness of the association. The exploratory mediation model showed that increased cortisol and decreased HRV composite scores partially explained the association, whereas CRP and IL-6 showed positive or borderline indirect effect trends. This finding is compatible with the comprehensive paths of chronic worry, hypothalamic-pituitary-adrenal axis activation, autonomic imbalance, inflammatory activation, and changes in vascular tension[31]. However, because levels of cortisol and inflammation markers, as well as HRV were measured only once at baseline, and because the mediation analysis used baseline mediators with follow-up outcomes, the results should be interpreted as a statistical decomposition of associations rather than confirmation of a causal chain.

This study indicated that: (1) GAD may be included in the assessment process as an identifiable psychological risk marker for the primary prevention of hypertension, especially among outpatient and physical examination populations with high-normal blood pressure; (2) Intervention strategies may integrate anxiety management, sleep intervention, medication review, dietary salt reduction, physical activity promotion, and ambulatory blood pressure monitoring; and (3) At the level of hospital management and community screening, short scales such as the GAD-7 could be combined with basic physiological indicators to achieve low-cost stratification. The main limitations of this study include the single-center clinic-based recruitment, short follow-up, limited number of events, non-randomized grouping, single-time-point cortisol/inflammatory measurements, possible residual confounding factors, and exploratory mediation modelling. Future studies should adopt multicenter community-based cohorts, extend the follow-up to 3-5 years, repeat blood-pressure and biomarker measurements, and incorporate dynamic salivary cortisol or circadian rhythm monitoring to verify the stability and causality of these findings.

CONCLUSION

This study suggests that the population with GAD in a clinic-based cohort had a significantly increased 12-month risk of sustained new-onset hypertension, and this association remained after adjusting for common clinical, medication-related, and lifestyle confounders. Elevated cortisol levels and reduced HRV composite scores statistically explained part of this association, suggesting that neuroendocrine stress and autonomic imbalance are important links. However, this was a single-center prospective observational study with a relatively limited sample size, short follow-up period, non-randomized grouping, and single-time-point biomarker measurement. Future multicenter, large-sample, long-term follow-up studies with repeated biomarkers and ambulatory blood-pressure measurements are needed to verify the causal chain and evaluate the antihypertensive benefits of psychosomatic integration interventions.