Effects of Health Action Process Approach (HAPA)-Guided Exercise Rehabilitation on Cardiovascular Risk Factors in Perimenopausal Women with CHD: A Randomized Controlled Trial

Introduction

Menopause is a critical physiological transition characterized by widespread metabolic disruption due to declining ovarian function and estrogen secretion, leading to adverse changes in lipid profiles, blood glucose, and blood pressure.¹ Women in the perimenopausal phase typically experience more than a year of menstrual irregularity and face a significantly elevated risk of developing coronary heart disease (CHD) during and after the menopausal transition.^2,3 For women already diagnosed with CHD, the perimenopausal period presents a particularly vulnerable window during which cardiovascular risk accelerates beyond what would be expected from aging alone, with each year of advancing menopause associated with progressive worsening of arterial stiffness and endothelial dysfunction.^1,4 Consequently, perimenopausal women with established CHD may represent a clinically important subgroup that requires targeted secondary prevention strategies.

Exercise-based cardiac rehabilitation (CR) is a cornerstone of secondary prevention for patients with established CHD, with systematic reviews confirming that exercise-based CR significantly reduces all-cause and cardiovascular mortality, myocardial infarction, and hospitalization rates when compared to usual care.^5,6 However, suboptimal adherence to rehabilitation programs remains a major barrier, substantially limiting their effectiveness.^7,8 This adherence challenge may be particularly pronounced in perimenopausal women with CHD, who face unique barriers including menopausal symptoms (eg, vasomotor disturbances, sleep disruption, mood changes), multiple-role demands, and lower rates of CR referral and participation compared to men.^9,10 Indeed, despite having comparable or greater cardiovascular risk burden, women consistently demonstrate lower enrollment and adherence to exercise-based CR programs.¹¹

The cardiometabolic benefits of exercise in postmenopausal women have been well documented. A recent network meta-analysis of 32 randomized controlled trials demonstrated that various exercise modalities—including continuous endurance training, interval training (INT), resistance training, and aerobic combined with resistance training—significantly improve flow-mediated dilation and reduce pulse wave velocity in postmenopausal women, with INT showing the greatest efficacy for improving endothelial function.¹² Furthermore, in postmenopausal women with hypertension, continuous aerobic exercise was most effective in reducing systolic and diastolic blood pressure, whereas high-intensity interval training was most beneficial for lowering blood pressure and increasing flow-mediated dilation in those without hypertension.¹³ These findings suggest that tailored exercise prescriptions can effectively modify key cardiovascular risk factors in postmenopausal populations, including those with established CHD.

Given the adherence challenges inherent in this population, effective interventions must not only prescribe appropriate exercise but also address the behavioral determinants of sustained engagement. The Health Action Process Approach (HAPA) provides a structured framework for health behavior change that distinguishes between motivational and volitional phases of behavior change (Figure 1).^14,15 Unlike traditional continuous models such as the Theory of Planned Behavior (TPB), which assumes that intention directly translates into behavior, HAPA explicitly addresses the well-documented intention–behavior gap by incorporating post-intentional volitional factors—including action planning, coping planning, and maintenance self-efficacy—that mediate the transition from intention to sustained action.^16,17 This distinction is particularly relevant for perimenopausal women with CHD, who may form strong intentions to engage in exercise but face substantial barriers (menopausal symptoms, fatigue, time constraints, competing responsibilities) that impede the translation of intention into consistent action.¹⁸ A meta-analysis of 95 studies applying HAPA across various health behavior contexts confirmed that action and maintenance self-efficacy had small-to-medium sized effects on health behavior, with effects mediated by intentions and planning, and highlighted the model’s utility for designing interventions that target both motivational and volitional processes.¹⁷ Moreover, compared to stage-based models such as the Transtheoretical Model (TTM), which describes readiness for change but provides limited guidance on the specific mechanisms driving stage transitions, HAPA offers a more parsimonious and testable framework that specifies distinct social-cognitive predictors for the motivational and volitional phases.¹⁹

Figure 1 Theory of HAPA.

HAPA-based interventions have demonstrated efficacy in cardiovascular populations. A recent randomized controlled trial found that an 8-week home-based virtual exercise rehabilitation program based on HAPA significantly improved physical activity levels and key psychological constructs (perceived threat, outcome expectations, self-efficacy) in patients with myocardial infarction, with benefits sustained at 6-month follow-up.²⁰ Another randomized controlled trial protocol describes a HAPA-guided, motivated action-based intervention targeting physical activity, exercise self-efficacy, and cardiovascular risk factors in patients with CHD.²¹ However, despite this accumulating evidence, the efficacy of HAPA-guided exercise-based CR specifically in perimenopausal women with established CHD has not been adequately examined. Furthermore, existing CR studies have predominantly enrolled male participants, with limited evidence on how behavioral theory-guided exercise interventions should be tailored to address the unique physiological and psychosocial context of perimenopausal women with CHD.

This study therefore investigates the effects of a HAPA theory-guided, exercise-based cardiac rehabilitation program on risk factor intervention in perimenopausal patients with CHD. Specifically, Body Mass Index (BMI) was selected as the primary outcome measure for evaluating intervention efficacy and determining sample size. This is because BMI serves as a reliable, standardized macroscopic indicator of energy balance and actual behavioral change, responding directly to increased physical activity.²² Furthermore, in the physiological cascade of exercise rehabilitation, improvements in overall body composition typically precede and strongly mediate the subsequent modifications of downstream metabolic risk factors, such as dyslipidemia, hypertension, and insulin resistance, particularly in women transitioning through menopause.^1,23

Participants and Methods

Participants

The participants in the study included perimenopausal patients with CHD hospitalized between April and December 2024 at a grade III level A hospital in Jiaxing City, Zhejiang Province, China. A convenience sampling method combined with competitive recruitment was employed. Specifically, daily screening was conducted via the Hospital Information System (HIS) to identify female inpatients aged 45–60 years with an admission diagnosis of coronary heart disease. For eligible candidates, attending physicians introduced the study overview post-consultation. Subsequently, research nurses provided detailed face-to-face explanations and distributed recruitment brochures during routine health education sessions. The complete participant selection, randomization, and follow-up processes are detailed in the CONSORT flow diagram (Figure 2). The patients in accordance with the following criteria were included: (1) females in the perimenopausal period, defined as women aged 45 years or older who have experienced a variation of more than 7 days in menstrual cycle length for at least two cycles within the past 10 months, or who have had at least two intervals of amenorrhea lasting 60 days or longer;²⁴ (2) those who capable of using a smartphone; (3) those without language or communication barriers, and capable of independently providing informed consent and complying with the study procedures; (4) those who meet the diagnostic criteria for CHD.²⁵ Patients with the following diseases or conditions were excluded: (1) untreated severe coronary artery disease; (2) acute myocardial infarction; (3) heart failure; (4) menstrual irregularities or amenorrhea caused by organic diseases (eg, polycystic ovary syndrome, endometrial lesions, or thyroid dysfunction) rather than the natural perimenopausal transition; (5) application of estrogen replacement therapy; (6) a history of severe psychiatric disorders (eg, schizophrenia, bipolar disorder, or major depressive disorder with suicidal risk) that would impair the capacity to comply with the behavioral intervention or accurately complete psychological self-evaluations; (7) presence of cancer or systemic malignant diseases; (8) impaired liver or kidney function, or other organ dysfunctions; (9) limitations in physical mobility.

Figure 2 CONSORT flow diagram of the study participants.

Sampling Methodology and Sample Size Estimation

The included patients were assigned to control group and experimental group using a random number table method. To ensure allocation concealment, the sequentially numbered, opaque, sealed envelopes (SNOSE) method was employed. An independent research assistant, not involved in participant recruitment or intervention delivery, prepared the allocation sequence and envelopes. Each envelope contained a card indicating the group assignment, which was only opened after the participant had completed baseline assessments and was formally enrolled in the trial. Initially, a pilot study was conducted with 20 perimenopausal patients with CHD to collect body mass index three months post-discharge for both groups. Subsequently, the sample size was determined using the following formula: n= (uα/2+uβ) 2 (σ12 +σ22) /δ2. In the formula n represents sample size, uα/2=1.96 and uβ =1.28 were obtained according to the value of α=0.05, β=0.01. δ denotes the mean difference of two groups. Based on these calculations, the required sample size was determined to be 48 cases. Specifically, in our 20-patient pilot study, the mean BMI at the 3-month follow-up was 24.8 ± 2.1 kg/m² in the control group and 23.5 ± 2.0 kg/m² in the experimental group, yielding a mean difference (δ) of 1.3 kg/m². Using these variance parameters with α=0.05 and a power (1-β) of 0.99, the required sample size was determined to be 48 cases per group. Considering a potential 10% loss during the follow-up, 53 cases per group was determined as the final sample size. Ultimately, 2 patients in the control group were lost during follow-up (due to acute renal failure secondary to severe infection in one case, and onset of hematological malignancy in another case). Three patients in the experimental group were lost (with 2 withdrawing due to personal reasons and 1 failed to continuously undergo cardiac rehabilitation post-surgery due to breast disease).

Ethics Approval

All enrolled participants voluntarily participate in the research and signed informed consent forms. The study has been approved by the ethics committee (Approval number: JXEY2024-092-01).

Study Methods

The patients in the control group received conventional exercise rehabilitation, including distribution of health education manuals involving CHD knowledge in perimenopausal period, and precautions before discharge. Besides, offering exercise rehabilitation approaches developed based on individual assessments. The experimental group additionally implemented an exercise intervention program targeting risk factors for CHD based on the HAPA theory.

Establishment of Intervention Team

The team is composed of seven specialized nurses (three cardiovascular nurses, two cardiac rehabilitation nurses, two research nurses), one cardiologist, and one physiotherapist. All the members have a work experience of more than five years, with a bachelor’s degree or above, and possessing intermediate or higher professional titles.

Development of the Intervention Program

We searched the literatures involving CHD in perimenopausal women published between 2013 and 2023 in both Chinese and foreign databases, such as China National Knowledge Infrastructure (CNKI), Wanfang, PubMed, and Web of Science. Due to the notable scarcity of literature specifically targeting cardiac rehabilitation in perimenopausal women with CHD, we adapted evidence from general women’s cardiovascular disease primary prevention guidelines the Chinese Expert Consensus on Cardiovascular Disease Prevention for Women²⁶ and the American Heart Association Guidelines for Preventing Heart Disease in Women² —to a secondary prevention context. Based on this adaptation, we formulated a targeted intervention program. The team then further refined the strategies targeting CHD risk factors through an evidence-based Delphi consensus and integrated them into the HAPA framework (Table 1). The detailed methodology, construction, and validation of this tailored program have been previously published by our research team.²⁷ To ensure standardized implementation and optimize patient adherence, all team members underwent comprehensive training to master the entire intervention process.

Table 1 Exercise Intervention Program for Risk Factors in Perimenopausal Patients with CHD Based on the HAPA Theory

Implementation of Intervention Program

(1) During the preintention stage, the members of the team focused on the knowledge of perimenopausal issues and CHD for enrolled patients, along with the consequences of unhealthy behaviors. Besides, we shared and presented the successful and typical cases that were conductive to promotion of their confidence and actions for rehabilitation. After learning and deep reflection, patients gradually recognized that scientific and rational self-management can indeed prevent and reduce the occurrence of cardiovascular events, thereby fostering a sense of self-efficacy for taking action. (2) In intention stage, the rehabilitation team conducted comprehensive assessments for patients encompassing their medical history, physical examination, electrocardiogram, echocardiogram, current medication, and risk factors of cardiovascular disease after identifying their willingness in behavioral improvement. Based on consequences of assessment, personalized exercise programs consisting of aerobic and resistance exercises were conducted for them.^28,29 The aerobic exercises of choice include cycling, walking, jogging, ect, with an intensity achieving 70–80% of reserve heart rate.³⁰ Target heart rate was calculated using the Karvonen formula: Target HR = [(220 - age) - resting HR] × target intensity (%) + resting HR. During exercise, real-time HR was continuously monitored using wearable devices synced to a nurse terminal. If the HR deviated beyond ±5 beats per minute from the target zone, the responsible nurse provided immediate verbal guidance to adjust exercise speed. To ensure progressive overload, exercise intensity was dynamically re-assessed. The 6-minute walk distance test was repeated every 4 weeks to recalculate maximal functional capacity. Furthermore, Borg Rating of Perceived Exertion (RPE) was monitored weekly; if a patient reported an RPE < 11 (“fairly light”) for two consecutive sessions, the aerobic target intensity was increased by 5–10% for the subsequent cycle. It is recommended to perform 5 to 7 days per week for 30 to 60 minutes each day. Following two weeks of aerobic training, some moderate-intensity resistance training protocols using elastic bands and dumbbell exercises is additionally incorporated, which was considered to be helpful and more available to strengthen the efficacy. The resistance training intensity requires below 11 and not over 14 at its peak as recommended by the Rating of Perceived Exertion (RPE),^31,32 with a frequency of three times a week on alternate days and lasting 15 to 30 minutes per time. Emphasis is placed on proper breathing coordination during exercise to avoid the Valsalva maneuver, targeting muscle groups such as the upper arms, back, abdomen, and legs. During the training, patients were guided to exhale and inhale in the correct way to avoid the Valsalva maneuver,³³ focusing on strengthening the upper arms, back, abdomen, and leg muscles. For beginners, patients are proper to start with exercises via elastic band and dumbbell in low resistance and gradually increase to higher load within allowable limits.^33–35 Each rehabilitation session begins with a 5- or 10-minute warm-up, followed by aerobic exercises, then resistance training, and concludes with a 5- or 10-minute cool-down. Besides, we provided them a set of appropriate response measures and techniques to help them coping with special or emergency situations. (3) In behavior stage, the team established WeChat group to share health-related information, and conducted telephone follow-up every two weeks and monthly outpatient clinic visits. Besides, the patients utilized heart rate monitors to record their pre- and post-exercise blood pressure, pulse rates, duration of exercise, modes of exercise, and personal perceptions of well-being during the training. Adherence was rigorously quantified. Good adherence was defined a priori as achieving an exercise execution rate of ≥80% (actual sessions completed/prescribed sessions × 100%) and maintaining a complete weekly log (recording HR, duration, and Borg RPE at least three times weekly). For those who persistently adhered to the protocols, the team members offered incentives, while for those who exhibited signs of withdrawal, provided encouragement, and strengthened monitoring. Both groups adhere to medication regimens outlined in the Consensus on the Management Pathway for Pharmacotherapy in Patients with Coronary Artery Disease,³⁶ with the observation period spanning from the time of hospital admission to three months post-discharge.

Evaluation Indicators

Due to the nature of the behavioral intervention, participants and the intervention team (physicians, nurses, physiotherapists) were not blinded to group assignment. However, to minimize assessment bias, the research staff responsible for collecting and analyzing the outcome data (including the 6MWD assessors and the laboratory technicians processing blood samples) were blinded to the group allocation of the participants throughout the study period.

Objective Indicators

Objective indicators were collected upon admission and three months post-discharge. For blood biomarkers, participants were required to undergo a strict overnight fast (12 hours without food and 8 hours without water). Fasting venous blood samples (5 mL) were drawn from the cubital vein in the early morning. Plasma Total Cholesterol (TC), Triglycerides (TG), Low-Density Lipoprotein Cholesterol (LDL-C), High-Density Lipoprotein Cholesterol (HDL-C), and Fasting Plasma Glucose (FPG) were analyzed using the enzymatic method on a Hitachi 7600 automated biochemical analyzer. For blood pressure assessment, participants rested quietly in a seated position for at least 15 minutes. Systolic and diastolic blood pressure (SBP and DBP) were then measured on the right upper arm using an Omron HEM-7136 electronic sphygmomanometer by uniformly trained research nurses. Measurements were taken three times with a 2-minute interval between each, and the average value was recorded for analysis. In the 6MWD test, patients were instructed to walk at their fastest tolerable speed along a straight corridor, and the distance in six minutes was measured. BMI was also calculated based on standardized height and weight measurements.

Subjective Indicators

Upon admission and at three months post-discharge, patients completed self-assessment by Self-rating Anxiety Scale (SAS) and Self-rating Depression Scale (SDS). SAS, a widespread used screener to measure anxiety disorders comprising 20 items, reflects anxiety symptoms of subjective feelings. It employs a four-point scoring system (scoring 1, 2, 3, 4), with a score range of 0 to 100. According to scoring outcomes, the anxiety is categorized as varied degrees of severity: severe anxiety (scoring≥70), moderate anxiety (scoring 60 to 69), and mild anxiety (scoring 50 to 59). SDS, including 20 items and a total score ranging 0 to 80, is classified as severe depression (scoring≥73), moderate depression (scoring 63 to 72), and mild depression (53 to 62 points)³⁷ based on assessment of severity grading. Higher scores indicate more severe conditions, and vice versa. At admission and three months post-discharge, patients fulfilled the China Cardiovascular Quality of Life Questionnaire (CCQQ) to evaluate their quality of life. The questionnaire encompasses six dimensions: life function, psychosocial function, physical strength, disease condition, medical status, and work status, with 24 items in total and a maximum score of 154. Higher total scores demonstrate better quality of life.

Outcome Indicators

Outcome indicators were collected at three months post-discharge. (1) All-cause mortality: including both cardiac and non-cardiac deaths occurring during the follow-up period, with death information collected from death certificates, hospital records, and their families; (2) non-fatal myocardial infarction: cases in accordance with the diagnostic criteria for acute myocardial infarction, with no occurrence of death; (3) coronary revascularization: reoperation for interventional treatment or coronary artery bypass grafting surgery on coronary arteries; (4) recurrent angina pectoris: reoccurrence of angina pectoris after discharge, including stable and unstable angina, with a severity of at least CCS class II; (5) heart failure: diagnosed conforming to the criteria outlined in the National Heart Failure Guidelines 2023.³¹

Statistical Analyses

Statistical analyses were performed using SPSS software (version 22.0). Prior to analysis, the normality of all continuous data (eg, BMI, lipid profiles, 6MWD, SAS, SDS) was evaluated and confirmed using the Kolmogorov–Smirnov test (all P > 0.05); therefore, parametric tests were deemed appropriate and applied. The primary efficacy analyses were conducted strictly according to the intention-to-treat (ITT) principle, encompassing all 106 randomized participants. To address missing data at the 3-month follow-up (3 dropouts in the experimental group and 2 in the control group), the Last Observation Carried Forward (LOCF) method was employed for imputation. Additionally, a per-protocol (PP) analysis was performed as a sensitivity analysis to verify the robustness of the findings. Continuous variables with normal distribution were expressed as mean ± standard deviation (SD), and categorical variables were presented as frequencies (percentages). Baseline characteristics were compared between groups using independent-samples t-tests for continuous variables and χ²-tests or Fisher’s exact test for categorical variables. To rigorously evaluate the intervention effects at the 3-month follow-up, an analysis of covariance (ANCOVA) was employed as the primary statistical test, with baseline values entered as covariates. This approach effectively controls for baseline discrepancies and robustly assesses between-group differences. Consequently, redundant statistical tests (eg, paired t-tests and one-way ANOVA) were omitted to minimize the inflation of Type I error and improve interpretability. To further address the risk of Type I error arising from multiple comparisons across various clinical endpoints, the Benjamini-Hochberg False Discovery Rate (FDR) procedure was applied to the ANCOVA P-values. Findings were considered statistically significant and robust if the FDR-adjusted P-value (q-value) remained < 0.05, while unadjusted endpoints were strictly interpreted as exploratory.

Results

Comparison of General Demographics

Baseline characteristics, including residence, age, marital status, education level, disease duration, degree of coronary artery lesion, treatment status, and medication use, were comparable between the control and experimental groups (all P > 0.05), indicating no significant differences at baseline and ensuring the validity of subsequent between-group comparisons (Table 2).

Table 2 Comparison of Baseline Characteristics Between Groups

Comparison of Objective Outcome Measures

At the 3-month follow-up, ANCOVA (adjusting for baseline values as covariates) revealed significant between-group differences favoring the experimental group across all objective measures. After applying the Benjamini-Hochberg FDR correction for multiple comparisons, the improvements in the experimental group remained robustly significant: BMI (F = 7.352, adjusted P = 0.009), FPG (F = 12.408, adjusted P = 0.003), TC (F = 14.215, adjusted P = 0.002), TG (F = 9.845, adjusted P = 0.004), LDL-C (F = 8.624, adjusted P = 0.006), SBP (F = 8.112, adjusted P = 0.007), DBP (F = 15.340, adjusted P = 0.002), and 6MWD (F = 6.458, adjusted P = 0.013). The detailed estimated marginal means and statistical values are presented in Table 3.

Table 3 Comparison of Objective Outcome Measures Between Groups at 3 Months Post-Discharge (ANCOVA Adjusted for Baseline)

Comparison of Subjective Psychological and Quality of Life Scores

At 3 months post-discharge, ANCOVA (adjusting for baseline scores as covariates) demonstrated that the experimental group achieved significantly better psychological and quality of life outcomes than the control group. After applying the Benjamini-Hochberg FDR correction, the between-group differences remained statistically significant for SAS (F = 6.114, adjusted P = 0.021), SDS (F = 5.804, adjusted P = 0.024), and CCQQ scores (F = 13.963, adjusted P < 0.001). The detailed estimated marginal means and statistical values are presented in Table 4.

Table 4 Comparison of Subjective Psychological and Quality of Life Scores (Mean ± SD)

Clinical Outcomes at Three-month Post-Discharge

No incidents of all-cause mortality, non-fatal myocardial infarction, coronary revascularization, or heart failure were recorded in either group during the follow-up period. Recurrent angina occurred in one patient (2.0%) in the experimental group and five patients (9.8%) in the control group.

Discussions

This study demonstrated that exercise-based cardiac rehabilitation guided by the HAPA theory significantly improved objective risk factors, psychological well-being, and quality of life in perimenopausal women with CHD without increasing adverse clinical events. This discussion interprets these findings by exploring how behavioral theory addresses key rehabilitation challenges and the specific benefits for this patient population.

The acute stress response induced by exercise training contributes to chronic adaptive changes that improve blood pressure regulation, lipid metabolism, endothelial function, and anti-thrombosis, benefiting patients with CHD.^38,39 However, the transition to menopause significantly elevates CHD risk in women due to ovarian function decline and estrogen deficiency.^1,40 While cardiac rehabilitation is crucial, its effectiveness is often limited by poor patient adherence and the absence of sustained behavioral changes after discharge.⁴¹ The HAPA theory addresses this gap by providing a structured framework to translate behavioral intentions into maintained action.⁴² In this intervention, the HAPA model was operationalized across stages: the pre-intention stage focused on risk awareness and outcome expectations, particularly regarding estrogen-related metabolic changes; the intention stage involved creating personalized plans and building coping self-efficacy through skill training; and the action/maintenance stage utilized follow-up, social support (eg, WeChat groups), and non-material incentives to sustain rehabilitation behaviors.⁴³ While the HAPA framework provided the structural basis for our intervention, the observed improvements in the experimental group should be interpreted as associations rather than definitive causal evidence of the theory’s efficacy. As we did not quantitatively assess key psychological mediators—such as self-efficacy, action planning, and coping planning—we cannot confirm the precise theoretical mechanisms through which the intervention exerted its effects. Consequently, the HAPA model serves as a plausible organizational framework for our intervention rather than a verified causal pathway in this specific cohort.

Compared to typical CHD, perimenopausal women face a marked elevation in risk primarily driven by metabolic disruptions following estrogen decline.^1,40 Although estrogen supplementation has been explored, evidence suggests it fails to confer expected cardiovascular benefits and may increase risks of thromboembolism and certain cancers.^44,45 Therefore, interventions must prioritize modifiable risk factors like dyslipidemia, hypertension, hyperglycemia, obesity, and adverse emotions. The significant improvements in BMI, lipid profiles, blood pressure, fasting glucose, and 6-minute walk distance in the experimental group align with this focus. The potential mechanisms are multifaceted. First, the supervised, combined aerobic and resistance exercise program directly improved metabolic and cardiorespiratory fitness. Second, previous research suggests that exercise may enhance central β-endorphin levels, which can positively influence the hypothalamic-pituitary-ovarian axis, potentially mitigating perimenopausal endocrine dysregulation and its associated risks.⁴⁶ In contrast, the control group, despite receiving standard education and individualized plans, lacked the structured behavioral support necessary to overcome barriers to sustained engagement, leading to diminishing adherence over time—a finding consistent with our prior work.⁴⁷

Furthermore, perimenopausal women with CHD are particularly vulnerable to anxiety and depression due to hormonal fluctuations and illness burden.^48,49 These negative emotions are not only independent risk factors for CHD but also significantly undermine motivation for rehabilitation. Our findings, supported by existing literature, confirm that structured exercise combined with psychosocial support can effectively alleviate these psychological symptoms.⁴⁹ The significantly greater reductions in SAS and SDS scores, alongside the superior improvement in CCQQ quality of life scores and lower angina recurrence rate in the experimental group, underscore the value of integrating psychological and behavioral strategies into cardiac rehabilitation. Compared to interventions focusing solely on physical training, this HAPA-based comprehensive approach may offer a unique advantage in simultaneously addressing the physiological and psychosocial complexities of perimenopausal CHD.

Several limitations of this study must be acknowledged. First, its single-center design, moderate sample size, and relatively short 3-month follow-up period limit the generalizability of the findings and the ability to assess long-term clinical outcomes such as mortality. Second, we did not monitor changes in estrogen levels, preventing a direct examination of the proposed endocrine pathway through which exercise may confer benefits. Most importantly, as noted earlier, the lack of measurement of HAPA-specific theoretical mediators (eg, self-efficacy, action planning) restricts our understanding of the precise mechanisms of behavioral change. Furthermore, it is important to note that the sample size calculation was strictly powered based on the primary outcome, BMI. Consequently, the study may lack sufficient statistical power for the evaluation of secondary endpoints, including lipid profiles (TC, TG, LDL-C), fasting glucose, blood pressure, and psychological scores (SAS/SDS). These results should be considered exploratory and hypothesis-generating. Furthermore, while we utilized the Benjamini-Hochberg FDR procedure to mitigate Type I error, the high number of statistical tests performed across multiple physiological and psychological domains remains a limitation that may affect the overall interpretability and robustness of the results. Future multi-center studies with larger samples, longer follow-up, and incorporation of hormonal assays and theory-based mediator analyses are warranted to confirm the long-term efficacy and elucidate the causal pathways of this theory-guided rehabilitation model.

Conclusions

In conclusion, our findings suggest that exercise-based cardiac rehabilitation structured around the HAPA framework may improve short-term modifiable risk factors, psychological status, and quality of life in perimenopausal women with CHD. While the model demonstrates a favorable safety profile and shows promise in addressing the challenge of rehabilitation adherence, the lack of data on theoretical mediators precludes definitive conclusions regarding the underlying mechanisms. These preliminary results highlight the necessity for larger, mechanistic, and longitudinal research to confirm these benefits and explore the causal pathways.