Early Warning and Behavioural Intervention on the Cardiovascular Disease Risk in People Living with HIV: A Multicentre Non-Randomized Controlled Trial Authors: Yuyin Zhou, Chunhui Yan, Zhonghua Xia, Xiaowen Lou, Xiangjie Meng, Yaqin Dai, Qin Hong, Zhujun Lian, Xingliang Zhang, Jiaye Liu, Junfang Xu Published: 2026 ABSTRACT AbstractIntroductionHuman Immunodeficiency Virus (HIV) infection has currently become a manageable chronic disease with the prevalence of antiretroviral therapy (ART). The increase in life expectancy and the rising number of people aged more than 50 years among people living with HIV (PLWH) both indicate an increasing risk of chronic non-communicable diseases, especially cardiovascular disease (CVD). Moreover, the contribution of traditional risk factors to CVD far outweighs that of HIV-related factors, which indicates that interventions for reducing CVD risk factors in PLWH are urgent.MethodsWe designed a multicentre non-randomized controlled trial (nRCT), which incorporated six acquired immunodeficiency syndrome (AIDS)-designated hospitals in Zhejiang Province, and three rounds of surveys were conducted from January 2023 to January 2024. Assessments were performed at baseline, 3 months and 6 months post-intervention, respectively. For the intervention group, PLWH were informed of their 5/10-year CVD risk, on the basis of the assessment in baseline, verbally or online to warn them of the risk of having CVD; at the same time, personalized risky behaviour reinforcement intervention was implemented face-to-face or online for about 15 min. For the control group, PLWH were not informed of the 5/10-year CVD risk and only received routine health education in the clinic without additional intervention. Intervention effects were assessed using 5/10-year CVD risk, health-related behaviours (e.g., smoking, alcohol consumption, physical activities and dietary nutrition), as well as CVD cognitive level. To minimize baseline differences and reduce confounding due to non-treatment factors (e.g., socio-demographic characteristics), propensity score matching (PSM) was applied to construct comparable intervention and control groups, and a generalized estimating equation (GEE) was used to estimate the effect of the intervention. This trial was retrospectively registered with the Chinese Clinical Trial Registry (ChiCTR2500112178) on 11 November 2025.ResultsA total of 948 PLWH were incorporated at baseline, after excluding ineligible participants, including 494 in the intervention group and 454 in the control group. The average age was 42.08 ( ± 11.94) years, with 826 male patients (87.13%). A total of 362 pairs were successfully matched after PSM. After matching, no statistical differences remained between the two groups. After 6 months’ intervention, the intervention group showed significant improvements in CVD cognitive level, with a total increase of 2.30 points compared with 0.41 points in the control group (P = 0.028). Additionally, PLWH who almost never drink and have > 8 h of sedentary time per day, increased and decreased 4.54% and 1.46% respectively, compared with the control group (P < 0.05). The 10-year CVD risk decreased by 0.63% in the intervention group and 0.18% in the control group. Moreover, the smoking rate and drinking rate decreased more in the intervention group compared with the control group, and the physical exercise compliance rate increased more as well, though there was no statistical difference compared with the control group (P > 0.05).ConclusionsRisk warning and behavioural intervention effectively improved CVD cognitive level. Although the risk warning and behavioural intervention could not make PLWH completely quit smoking or using alcohol, the intervention led to notable improvements in reducing the frequency of some health risk behaviours.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s40121-026-01394-5.Keywords: Early warning, Behavioural intervention, Cardiovascular disease risk, People living with HIV, Non-randomized controlled trial Key Summary Points This study evaluated the effectiveness of a combined risk warning plus personalized behavioural intervention strategy for cardiovascular disease (CVD) risk among people living with HIV (PLWH).This multicentre non-randomized controlled trial was conducted across six AIDS-designated hospitals in Zhejiang Province, China, and used propensity score matching and generalized estimating equations to evaluate intervention effects.After 6 months of follow up, the intervention significantly improved CVD cognitive level among PLWH.Although reductions in overall 5/10-year CVD risk were not statistically significant, the intervention significantly improved alcohol consumption frequency and sedentary time.The findings suggest that risk warning combined with behavioural intervention may improve CVD related knowledge and health-related behaviours among PLWH, although longer follow up is needed to determine whether these changes translate into meaningful CVD risk reduction. Introduction The introduction of highly active antiretroviral therapy (HAART) has dramatically reduced mortality rates among people living with HIV (PLWH), transforming HIV infection from a fatal disease into a manageable chronic condition. With standardized treatment, PLWH can now attain a life expectancy approaching that of the general population. However, this increased longevity is accompanied by a rising burden of age-related comorbidities, particularly chronic non-communicable diseases such as cardiovascular disease (CVD) [1, 2]. Emerging evidence suggests that CVD has become a leading cause of death among PLWH [3, 4], with a prevalence surpassing that observed in the general population [5, 6]. The heightened CVD risk in PLWH is attributed not only to the direct effects of HIV or ART-related metabolic disturbances but also to the clustering of traditional cardiometabolic risk factors. Compared with the general population, PLWH exhibit a higher burden of modifiable CVD risk factors, such as smoking, alcohol consumption, poor diet, physical inactivity and being overweight, as well as psychosocial stress [7–10]. Several studies have documented alarming trends. Evidence from a 12,135 treatment-naive PLWH study found that the prevalence of overweight patients among PLWH was high and it rose from 11.4% in 2014 to 17.3% in 2020 and that of obesity from 2.0% to 4.2%, respectively. Overweight and obese PLWH were more likely to exhibit hyperglycaemia, higher triglyceride levels and lower high-density lipoprotein levels [11, 12]. Published studies showed that the rate of dyslipidemia in individuals newly diagnosed with HIV was 49.04% and it continues to increase after receiving HAART [13, 14]. Moreover, the prevalence of hyperglycaemia and diabetes was higher among newly diagnosed PLWH in China and has continued to rise in the past years, from 3.0% in 2013 to 7.0% in 2019 for hyperglycaemia, and 0.5% in 2013 to 2.6% in 2019 for diabetes [11]. These metabolic abnormalities are compounded by the high prevalence of behavioural risk factors. Evidence also showed that among the contribution of modifiable risk factors to non-AIDS-defining cancers, the incidence of CVD was the highest and 34.14% of CVD was attributable to smoking, 7.98% to hypertension and 6.44% to diabetes [15]. Notably, the overall contribution of traditional risk factors to CVD far outweighed that of HIV-related factors, underscoring the critical importance of early prevention, screening and behavioural interventions targeting CVD risk in this population. Although interventions have the potential to reduce CVD risk, relevant services for PLWH remain limited. In China, PLWH usually receive treatment in AIDS designated-hospitals, rather than general hospitals [16–18]. Healthcare services in these settings primarily focus on the administration of HAART and the management of opportunistic infections [17], with limited attention to the prevention and management of CVD-related conditions such as hypertension and diabetes mellitus. Alarmingly, a study conducted in China showed that only 5.6% (10/179) of PLWH with hypertension received antihypertensive treatment, and just 4.6% (4/87) of those with diabetes received blood glucose-lowering therapy [10]. Similar challenges were observed in high-income countries. In the USA, physicians have been shown to underutilize guideline-recommended cardiovascular care and are less likely to prescribe preventive therapies for HIV-infected individual [19]. Despite the establishment of a comprehensive HIV prevention and treatment system in China, CVD risk management is still not routinely integrated into HIV care. Furthermore, few studies worldwide have evaluated effective interventions targeting CVD risk reduction among PLWH, with most existing research limited to descriptive analyses of risk profiles. To address this gap, we conducted a multicentre non-randomized controlled trial to assess whether a combined risk warning and behavioural intervention could improve CVD-related health behaviours and reduce cardiovascular risk among PLWH in China. This trial is registered with the Chinese Clinical Trial Registry, number ChiCTR2500112178. Methods Ethical Approval This study was approved by the Medical Ethics Committee of School of Public Health, Zhejiang University (ZGL202203-5). Also, this trial was retrospectively registered with the Chinese Clinical Trial Registry (ChiCTR2500112178) on 11 November 2025 and was conducted according to the protocol approved by the ethics committee before participant recruitment. Study Design and Participants A multicentre non-randomized controlled trial was conducted to assess the effectiveness of early warning and behavioural intervention on reducing CVD risk in PLWH. The study was implemented across six AIDS-designated hospitals in Zhejiang Province, including sites located in Hangzhou city (Linping District, Fuyang District, Yuhang District and Chun’an County), Lishui city (Qingyuan County) and Jiaxing city (Nanhu District). Participants in the intervention group were recruited from the Linping District site in Hangzhou, while the control group participants were recruited from the remaining five hospitals. Baseline assessments and intervention delivery took place between January 2023 and January 2024, with follow-up assessments conducted at 3 months (T2) and 6 months (T3) after the baseline (T1). Inclusion criteria for PLWH were: (a) confirmed diagnosis of HIV infection; (b) age 18 years or older and (c) no prior diagnosis of CVD. Exclusion criteria included: (a) age over 75 years and (b) current pregnancy or lactation. All participants provided written informed consent prior to study enrolment. The upper age limit was applied because the data collection on adverse effects of anti-HIV drugs (D:A:D) model used for CVD risk estimation has been described as applicable to PLWH aged 18–75 years [20, 21]. Pregnant or lactating participants were excluded because pregnancy-related physiological changes may affect CVD risk estimation and require different clinical counselling and management pathways. Before the formal implementation of the intervention, a pilot test was conducted with 50 PLWH to optimize the questionnaire and intervention strategies. Sample size calculations were performed using PASS 2021 software (PASS, Kaysville, UT, USA), with the 5/10-year CVD risk at T2 selected as the primary outcome. Based on literature review [22] and expert consultation, the sample size was calculated using a two-sided alpha level of 0.05 and a power of 0.80. To account for potential loss to follow-up and the use of propensity score matching (PSM), the required sample size was increased by 20%. Accordingly, a minimum of 340 participants was needed in both the intervention and control groups. Procedures CVD is associated with a range of risk factors, including age, sex, genetic predisposition, obesity, smoking, alcohol consumption, physical inactivity and poor diet. Among these, behavioural risk factors are the most amenable to modification. Therefore, this study focused on four key behavioural domains – smoking, alcohol consumption, physical activity and dietary nutrition – as the primary targets of intervention. Participants in the intervention group received two rounds of intervention, delivered at T1 and T2. At each time point, participants were informed of their individualized 5-year and 10-year CVD risk scores, which had been calculated at baseline, either verbally or online by trained healthcare providers at AIDS-designated hospitals (i.e., risk warning). As a core component of the intervention, personalized behavioural counselling sessions lasting approximately 15 min were delivered either face-to-face or online at both T1 and T2. All medical personnel involved in the intervention received standardized training prior to implementation. The personalized behavioural intervention was tailored according to each participant’s baseline risk profile and health behaviours. Key components included informing participants of the harms associated with risky behaviours (e.g., smoking, excessive alcohol use, inactivity and poor dietary habits), highlighting the potential health benefits of behaviour change, and providing actionable recommendations for improvement. The content and structure of the intervention were primarily based on the European AIDS Clinical Society (EACS) Guidelines (11th edition) [23] for PLWH, the Chinese Guidelines for the Prevention and Treatment of Cardiovascular Diseases in the General Population [24], the Chinese Dietary Guidelines (2022) [25] and the Report on Health Hazards of Smoking in China (2020) [26]. Participants in the control group were not informed of their CVD risk scores and received only standard health education routinely provided in outpatient HIV care settings, without any additional interventions. Primary Outcome: CVD Risk CVD risk was assessed using the reduced data-collection on adverse effects of anti-HIV drugs (D:A:D) model [20], which was developed specifically for PLWH. The predictor variables included age, sex, current smoking status, smoking history, family history of CVD, history of diabetes, CD4+ T-cell count, systolic blood pressure, total cholesterol (TC) and high-density lipoprotein (HDL). Among these variables, age, sex, current smoking status, smoking history, family history of CVD and history of diabetes were collected through questionnaires. Systolic blood pressure was measured with a sphygmomanometer by trained medical staff. Laboratory variables (CD4+ T-cell counts, TC and HDL) were extracted from the HAART database. In addition, this study adapted the D:A:D(R) model using mean baseline values from this study to calibrate the model. CVD risk was stratified into three categories. For 5-year risk, values of < 5%, 5–7.5% and > 7.5% were defined as low, intermediate and high risk, respectively. For 10-year risk, values of < 5%, 5–10% and > 10% were similarly defined as low, intermediate and high risks [27, 28]. Secondary Outcomes Health-Related Behaviours Health-related behaviours assessed in this study included smoking, alcohol consumption, physical activities and dietary nutrition. Smoking-related variables comprised current smoking status, smoking history, age at smoking initiation and average number of cigarettes smoked per day. Daily cigarette consumption was categorized as 0–5, 6–10 and > 11 cigarettes per day [29]. Alcohol consumption was assessed through three indicators: current drinking status (yes or no), frequency of alcohol intake and age at drinking initiation. Physical activity was evaluated on the basis of whether participants met the recommended standard: performing moderate-intensity physical activity on at least 5 days per week, totalling 150 min or more [25]. Daily sedentary time was also recorded. Dietary nutrition assessment was based on the Chinese Dietary Guidelines (2022) [25]. Participants were asked to report their intake frequency over the past week for six food categories: cereals, potatoes, vegetables, fruits, liquid milk and total intake of fish, poultry, eggs and lean meat. For each food type, intake frequency was scored as follows: 0 for ‘hardly ever,’ 1 for ‘1–3 days/week,’ 2 for ‘4–6 days/week’ and 3 for ‘every day’, yielding a total possible score of 18. Dietary nutrition was considered adequate if all six food categories were consumed daily. CVD Cognitive Level CVD-related knowledge was assessed using a set of self-designed questions on the basis of published literature and the National Free HIV Antiretroviral Therapy Manual [18, 30–32]. The questionnaire covered three dimensions: knowledge of CVD risk factors (such as family history of CVD, psychological stress, obesity and physical inactivity), knowledge of dietary nutrition (including recommended daily salt intake and the nutritional value of fresh vegetables) and knowledge of CVD-related clinical indicators (such as normal ranges for blood pressure, blood lipids, blood glucose and heart rate). Participants received 1 point for each correct response and 0 points for incorrect answers, yielding a maximum total score of 15 points. Higher scores reflected greater knowledge and awareness of cardiovascular health. The internal consistency of the questionnaire, as measured by Cronbach’s alpha, was 0.708, indicating acceptable reliability. Adverse Events This study defined adverse events as any unexpected physical, mental or social problems that arose directly or indirectly from participation in the intervention. Statistical Analysis Descriptive statistics were used to describe socio-demographic characteristics, CVD risk, health-related behaviours and CVD cognitive level among PLWH. Categorical variables were presented as frequencies and percentages, while continuous variables were expressed as means and standard deviations (SDs). Between-group differences in categorical variables were assessed using the chi-square test or Fisher’s exact test, and differences in continuous variables were assessed using independent-samples t-tests. To minimize baseline differences and reduce confounding due to non-treatment factors (e.g., socio-demographic characteristics), propensity score matching (PSM) was applied to construct comparable intervention and control groups. Standardized differences were used as the primary measure of balance because hypothesis testing based on P values is less appropriate for matched samples, which are no longer independent after matching. For categorical variables, standardized differences were calculated for each category, and the maximum absolute standardized mean difference (SMD) was used to summarize the overall imbalance of the variable. Absolute standardized differences less than 0.2 were considered indicative of good balance. We also compared baseline characteristics between participants included in the matched and unmatched samples to assess the potential influence of excluding unmatched participants. Generalized estimating equations (GEE) were used to evaluate the effects of the risk warning and behavioural intervention on outcomes. Baseline values were included as covariates in all models to account for initial differences. Moreover, as a sensitivity analysis to address the potential influence of unmatched participants, we further performed multivariable generalized estimating equation analyses using the full pre-matched samples. All statistical analyses were conducted using SPSS (version 24.0, IBM Corp., Armonk, NY, USA) and R (version 4.2; R Foundation for Statistical Computing, Vienna, Austria). A two-sided P-value of < 0.05 was considered statistically significant. Results Basic Characteristics of PLWH at Baseline A total of 948 PLWH were included in the study after excluding ineligible participants, with 494 individuals assigned to the intervention group and 454 to the control group at baseline. The participant selection process, including inclusion and exclusion criteria, is presented in Fig. 1. The average age was 42.08 (± 11.94) years, with 826 male patients (87.13%). At baseline, there were statistically significant differences between the intervention and control groups in terms of age, sex, marital status, education level, occupation and personal average monthly income (P < 0.05). To minimize these baseline imbalances, propensity score matching (PSM) was performed using these socio-demographic characteristics as matching variables. A total of 362 matched pairs were successfully identified. After matching, no statistically significant differences remained between the two groups in age, sex, marital status, education level, occupation or income (all P > 0.05) (Table 1). Flow diagram of study design. *CVD indicates cardiovascular disease; T1, T2 and T3 represent baseline, 3 months and 6 months after the baseline, respectively Table 1Sociodemographic characteristics of PLWH between intervention and control groupVariablesBefore PSMAfter PSMTotal (%) (n = 948)Intervention group (%) (n = 494)Control group (%) (n = 454)P valueStandardized differenceTotal (%) (n = 724)Intervention group (%) (n = 362)Control group (%) (n = 362)P valueStandardized differenceAge (years), mean (SD)42.08 ± 11.9440.52 ± 10.8343.76 ± 12.85 < 0.0010.27741.57 ± 11.6641.83 ± 11.4341.30 ± 11.900.5370.046Sex < 0.0010.803Male826 (87.13)451 (91.30)375 (82.60)0.260648 (89.50)323 (89.23)325 (89.78)0.018Female122 (12.87)43 (8.70)79 (17.40)0.26076 (10.50)39 (10.77)37 (10.22)0.018Nationality0.7900.251Han909 (95.89)475 (96.15)434 (95.59)0.028695 (95.99)350 (96.69)345 (95.30)0.070Others36 (3.80)18 (3.64)18 (3.96)0.01728 (3.87)11 (3.04)17 (4.70)0.086Marital status0.0050.817Not married591 (62.34)304 (61.54)287 (63.22)0.035461 (63.67)229 (63.26)232 (64.09)0.017Married352 (37.13)188 (38.06)164 (36.12)0.040263 (36.33)133 (36.74)130 (35.91)0.017Education level < 0.0010.947Junior high school or below378 (39.87)164 (33.20)214 (47.14)0.287284 (39.23)143 (39.50)141 (38.95)0.011High school or technical school227 (23.95)115 (23.28)112 (24.67)0.033201 (27.76)97 (26.80)104 (28.73)0.043College or university312 (32.91)197 (39.88)115 (25.33)0.314227 (31.35)116 (32.04)111 (30.66)0.030Graduate school or above16 (1.69)10 (2.02)6 (1.32)0.05512 (1.66)6 (1.66)6 (1.66)0.000Occupation < 0.0010.292Government/public institution26 (2.74)16 (3.24)10 (2.20)0.06417 (2.35)7 (1.93)10 (2.76)0.055Company employee502 (52.95)287 (58.10)215 (47.36)0.216395 (54.56)200 (55.25)195 (53.87)0.028Self-employed194 (20.46)104 (21.05)90 (19.82)0.030153 (21.13)76 (20.99)77 (21.27)0.007Student8 (0.84)3 (0.61)5 (1.10)0.0548 (1.10)3 (0.83)5 (1.38)0.053Retired51 (5.38)24 (4.86)27 (5.95)0.04836 (4.97)24 (6.63)12 (3.31)0.153Unemployed158 (16.67)56 (11.34)102 (22.47)0.300115 (15.88)52 (14.36)63 (17.40)0.083Personal average monthly income (RMB) < 0.0010.587 < 2500171 (18.04)59 (11.94)112 (24.67)0.334130 (17.96)59 (16.30)71 (19.61)0.0862501–5000263 (27.74)119 (24.09)144 (31.72)0.171219 (30.25)108 (29.83)111 (30.66)0.0185001–7500231 (24.37)134 (27.13)97 (21.37)0.135194 (26.80)104 (28.73)90 (24.86)0.0877501–10000139 (14.66)90 (18.22)49 (10.79)0.212102 (14.09)54 (14.92)48 (13.26)0.048 > 10,000132 (13.92)90 (18.22)42 (9.25)0.26379 (10.91)37 (10.22)42 (11.60)0.044Registration place0.6240.914Urban291 (30.70)147 (29.76)144 (31.72)0.043220 (30.39)106 (29.28)114 (31.49)0.048Rural558 (58.86)272 (55.06)286 (63.00)0.162442 (61.05)211 (58.29)231 (63.81)0.113Drug use0.8180.991Yes22 (2.32)12 (2.43)10 (2.20)0.01516 (2.21)8 (2.21)8 (2.21)0.000No920 (97.05)479 (96.96)441 (97.14)0.010704 (97.24)351 (96.96)353 (97.51)0.034Bold values indicate statistical significance (P < 0.05) Sociodemographic characteristics of PLWH between intervention and control group Bold values indicate statistical significance (P < 0.05) We further compared baseline characteristics between participants who were matched and unmatched. Unmatched participants differed from matched participants mainly in socio-demographic characteristics such as age, sex and education level. By contrast, baseline CVD cognitive level, 5/10-year CVD risk and most health-related behaviours were not significantly different between matched and unmatched participants (Supplementary Table S3). Baseline Balance of Outcomes Before and After PSM Before PSM, several outcomes differed significantly between the intervention and control groups. The 5-year and 10-year CVD risks were both higher in the control group than in the intervention group (5-year: 2.65% versus 2.09%, P = 0.010; 10-year: 5.30% versus 4.19%, P = 0.008). Additionally, the prevalence of alcohol consumption was significantly higher in the intervention group (54.25% versus 42.95%, P = 0.005), and the age of drinking initiation was older (21.46 ± 5.06 versus 20.51 ± 4.34, P = 0.042). No significant differences were observed in smoking, physical activity, nutritional compliance or CVD cognitive level at baseline (P > 0.05). After PSM, the standardized differences for the socio-demographic variables included in the matching model were substantially reduced after matching. Most absolute standardized differences were below 0.2, indicating PSM successfully mitigated baseline imbalances and established comparable groups for intervention evaluation (Tables 2, 3 and 4). Table 2CVD risk of PLWH between intervention and control groupVariablesBefore PSMAfter PSMTotal (%) (n = 948)Intervention group (%) (n = 494)Control group (%) (n = 454)P valueStandardized differenceTotal (%) (n = 724)Intervention group (%) (n = 362)Control group (%) (n = 362)P valueStandardized differenceCD4+ T-lymphocyte count, mean (SD)504.85 ± 242.54542.96 ± 253.35462.83 ± 222.87 < 0.0010.336503.66 ± 239.81532.99 ± 248.87473.99 ± 226.810.0010.248Systolic pressure122.03 ± 7.81121.64 ± 7.49122.46 ± 8.150.1110.105121.91 ± 7.27121.64 ± 7.84122.19 ± 6.640.3130.076Diastolic pressure79.04 ± 6.7078.76 ± 5.9579.36 ± 7.440.1720.08978.82 ± 6.6178.69 ± 6.2278.96 ± 6.990.5860.041TC4.78 ± 0.994.92 ± 0.954.62 ± 1.01 < 0.0010.3074.77 ± 0.994.93 ± 0.934.60 ± 1.02 < 0.0010.333HDL1.23 ± 0.301.24 ± 0.291.22 ± 0.300.1980.0871.23 ± 0.301.25 ± 0.301.22 ± 0.300.1320.116Family history of CVD0.6700.835Yes286 (30.17)147 (29.76)308 (67.84)0.019229 (31.63)114 (31.49)115 (31.77)0.006No654 (68.99)346 (70.04)139 (30.62)0.048488 (67.40)247 (68.23)241 (66.57)0.035If suffering from the following diseasesCVD13 (1.37)3 (0.61)10 (2.20)0.0370.13610 (1.38)2 (0.55)8 (2.21)0.0500.142Diabetes mellitus56 (5.91)30 (6.07)26 (5.73)0.9310.01538 (5.25)20 (5.52)18 (4.97)0.8140.025Hypertension93 (9.81)52 (10.53)41 (9.03)0.5540.05063 (8.70)35 (9.67)28 (7.73)0.4260.069Hyperlipidaemia72 (7.59)45 (9.11)27 (5.95)0.0930.12057 (7.87)32 (8.84)25 (6.91)0.3980.072Hypercholesterolemia39 (4.11)17 (3.44)22 (4.85)0.2280.07131 (4.28)12 (3.31)19 (5.25)0.1690.096Self-reported CVD risk0.0020.024No risk151 (15.93)70 (14.17)81 (17.4)0.100108 (14.92)52 (14.36)56 (15.47)0.031Low risk291 (30.70)173 (35.02)118 (25.99)0.197235 (32.46)133 (36.74)102 (28.18)0.184Medium risk286 (30.17)144 (29.15)142 (31.28)0.046218 (30.11)99 (27.35)119 (32.87)0.121High risk127 (13.40)74 (14.98)53 (11.67)0.09798 (13.54)55 (15.19)43 (11.88)0.0975-year CVD risk (%), mean ± SD2.33 ± 3.182.09 ± 2.932.65 ± 3.450.0100.1772.27 ± 2.952.23 ± 2.852.31 ± 3.060.7250.027 < 5774 (81.65)451 (91.30)323 (71.15)0.0060.534595 (82.18)323 (89.27)272 (75.14)0.4250.3755–7.551 (5.38)20 (4.05)31 (6.83)0.12339 (5.39)17 (4.70)22 (6.08)0.061 > 7.552 (5.49)23 (4.66)29 (6.39)0.07640 (5.52)22 (6.08)18 (4.97)0.04810-year CVD risk (%), mean ± SD4.68 ± 6.014.19 ± 5.525.30 ± 6.550.0080.1834.56 ± 5.694.50 ± 5.514.64 ± 5.900.7670.026 < 5632 (66.67)377 (76.32)255 (56.17)0.0010.436491 (67.82)266 (73.48)225 (62.15)0.6240.2445–10 (excluding 10)141 (14.87)74 (14.98)67 (14.76)0.006103 (14.23)57 (15.75)46 (12.71)0.087 > 10104 (10.97)43 (8.70)61 (13.44)0.15180 (11.05)39 (10.77)41 (11.33)0.018Bold values indicate statistical significance (P < 0.05) CVD risk of PLWH between intervention and control group Bold values indicate statistical significance (P < 0.05) Table 3Health-related behaviours of PLWH between intervention and control groupVariablesBefore PSMAfter PSMTotal (%) (n = 948)Intervention group (%) (n = 494)Control group (%) (n = 454)P valueStandardized differenceTotal (%) (n = 724)Intervention group (%) (n = 362)Control group (%) (n = 362)P valueStandardized differenceSmoking0.9650.128Yes279 (29.43)146 (29.55)133 (29.30)0.006210 (29.01)96 (26.52)114 (31.49)0.110No667 (70.36)348 (70.45)319 (70.26)0.004512 (70.72)266 (73.48)246 (67.96)0.122Smoking history0.9630.261Yes376 (39.66)196 (39.68)180 (39.65)0.001288 (39.78)137 (37.85)151 (41.71)0.079No570 (60.13)298 (60.32)272 (59.91)0.008434 (59.94)225 (62.15)209 (57.73)0.090Age of smoking initiation, mean ± SD21.80 ± 5.2022.25 ± 5.3821.33 ± 4.980.0960.17721.42 ± 4.9122.18 ± 5.4820.76 ± 4.250.0180.290Daily smoking quantity, mean ± SD9.60 ± 9.229.15 ± 8.9510.09 ± 9.500.3330.2969.44 ± 9.218.79 ± 9.0910.01 ± 9.300.2670.1330–5158 (16.67)85 (17.21)73 (16.08)0.5770.101123 (16.99)62 (17.13)61 (16.85)0.6370.0076–1080 (8.44)43 (8.70)37 (8.15)0.02063 (8.70)29 (8.01)34 (9.39)0.049 > 11121 (12.76)58 (11.74)63 (13.88)0.14793 (12.85)0.091Alcohol consumption0.431Never485 (51.16)226 (45.75)259 (57.05)0.0050.228365 (50.41)172 (47.51)193 (53.31)0.116Quit consuming alcohol107 (11.29)59 (11.94)48 (9.72)0.04380 (11.05)42 (11.60)38 (10.50)0.035Consume alcohol occasionally319 (33.65)189 (38.26)130 (28.63)0.205249 (34.39)131 (36.19)118 (32.60)0.076Consume alcohol everyday30 (3.16)16 (3.24)14 (3.08)0.00924 (3.31)14 (3.87)10 (2.76)0.062Age of alcohol consumption initiation, mean (SD)21.04 ± 4.7721.46 ± 5.0620.51 ± 4.340.20221.11 ± 4.9021.83 ± 5.3920.37 ± 4.230.0080.301Frequency of alcohol consumption0.0420.221Hardly ever107 (11.29)59 (11.94)48 (10.57)0.0780.04380 (11.05)42 (11.60)38 (10.50)0.035 < 1 time per week246 (25.95)152 (30.77)94 (20.70)0.260189 (26.10)104 (28.73)85 (23.48)0.1201–3 times per week53 (5.59)25 (5.06)28 (6.17)0.09742 (5.80)16 (4.42)26 (7.18)0.1184–7 times per week30 (3.16)2 (0.40)6 (1.32)0.07226 (3.59)2 (0.55)5 (1.38)0.000 > 7 times per week4 (0.42)0.0602 (0.28)12 (3.31)9 (2.49)0.000Physical activity up to standard0.997Yes285 (30.06)158 (31.98)127 (27.97)0.3510.088217 (29.97)111 (30.66)106 (29.28)0.030No618 (65.19)322 (65.18)296 (65.20)0.000471 (65.06)241 (66.57)230 (63.54)0.064Sedentary time0.037 < 3 h per day375 (39.56)195 (39.47)180 (39.65)0.0540.004285 (39.36)150 (41.44)135 (37.29)0.0853–8 h per day428 (45.15)214 (43.32)214 (47.14)0.077324 (44.75)148 (40.88)176 (48.62)0.156 > 8 h per day0.15785 (11.74)51 (14.09)34 (9.39)0.146Dietary nutrition score11.27 ± 3.4811.03 ± 3.4811.53 ± 3.480.0320.14211.34 ± 3.5111.03 ± 3.4711.64 ± 3.530.0200.175Dietary nutrition up to standardYes43 (4.54)25 (5.06)18 (3.96)0.3810.05333 (4.56)18 (4.97)15 (4.14)0.5640.040No881 (92.93)452 (91.50)429 (94.49)0.118672 (92.82)332 (91.71)340 (93.92)0.086Dietary nutritionCereals2.22 ± 0.992.06 ± 0.932.39 ± 1.03 < 0.0010.3392.21 ± 0.992.06 ± 0.932.36 ± 1.03 < 0.0010.306Potatoes1.27 ± 0.861.40 ± 0.721.12 ± 0.97 < 0.0010.3241.26 ± 0.871.38 ± 0.721.15 ± 0.97 < 0.0010.267Vegetables2.43 ± 0.822.29 ± 0.872.57 ± 0.74 < 0.0010.3492.43 ± 0.822.31 ± 0.872.55 ± 0.750.0580.305Fruits1.95 ± 0.951.89 ± 0.912.00 ± 0.980.0790.1141.97 ± 0.941.91 ± 0.922.04 ± 0.96 < 0.0010.141Liquid milk1.41 ± 0.941.57 ± 0.841.24 ± 1.02 < 0.0010.3521.41 ± 0.941.54 ± 0.841.28 ± 1.02 < 0.0010.282Fish/poultry/eggs/lean meat2.04 ± 0.901.88 ± 0.922.21 ± 0.84 < 0.0010.3782.08 ± 0.891.88 ± 0.922.27 ± 0.82 < 0.0010.440Bold values indicate statistical significance (P < 0.05) Health-related behaviours of PLWH between intervention and control group Bold values indicate statistical significance (P < 0.05) Table 4CVD cognitive level of PLWH between intervention and control groupVariablesBefore PSMAfter PSMTotal (%) (n = 948)Intervention group (%) (n = 494)Control group (%) (n = 454)P valueStandardized differenceTotal (%) (n = 724)Intervention group (%) (n = 362)Control group (%) (n = 362)P valueStandardized differenceCVD cognitive level score, mean ± SD6.43 ± 3.596.44 ± 3.756.43 ± 3.400.9450.0036.50 ± 3.556.25 ± 3.796.73 ± 3.290.0750.135CVD risk factorsFamily history of CVD502 (52.95)226 (45.75)276 (60.79) < 0.0010.305392 (54.14)161 (44.48)231 (63.81) < 0.0010.396Mental stress300 (31.65)137 (27.73)163 (35.90)0.0140.176229 (31.63)91 (25.14)138 (38.12) < 0.0010.282High salt intake404 (42.62)192 (38.87)212 (46.70)0.0320.159323 (44.61)140 (38.67)183 (50.55)0.0040.241Obesity546 (57.59)276 (55.87)270 (59.47)0.4780.073427 (58.98)195 (53.87)232 (64.09)0.0170.209Physical inactivity388 (40.93)211 (42.71)177 (38.99)0.1380.076306 (42.27)153 (42.27)153 (42.27)0.7400.000Heavy drinking462 (48.73)258 (52.23)204 (44.93)0.0080.146358 (49.45)187 (51.66)171 (47.24)0.1110.088Smoking349 (36.81)188 (38.06)161 (35.46)0.2650.054273 (37.71)136 (37.57)137 (37.85)0.8230.006Second-hand smoking181 (19.09)98 (19.84)83 (18.28)0.4310.040140 (19.34)69 (19.06)71 (19.61)1.0000.014Bold values indicate statistical significance (P < 0.05) CVD cognitive level of PLWH between intervention and control group Bold values indicate statistical significance (P < 0.05) The Effect of the Interventions Primary Outcome: CVD Risk From T1 to T3, although the 5-year CVD risk of PLWH in the intervention group decreased by 0.32%, which was more than in the control group (0.09%) (Fig. 2A), there was no statistical difference in the interaction effect of group × time (P = 0.825) (Table 5). The 5/10-year CVD risk (A: 5-year CVD risk; B: 10-year CVD risk) and CVD cognitive level score (C) of PLWH. *CVD indicates cardiovascular disease; T1, T2 and T3 represents baseline, 3 months and 6 months after the baseline, respectively Table 5The effect evaluation of CVD cognitive level and 10-year CVD riskVariablesaβSE95%CIP value5-year CVD risk (T1)b0.8740.0330.809, 0.939 < 0.001GroupIntervention group −0.0510.083 −0.213, 0.1120.539Control group––––TimeT2––––T3−0.2500.093 − 0.433, −0.0680.007Interaction effect[Intervention group] × [T3]0.0260.116−0.201, 0.2530.825[Intervention group] × [T2]––––10-year CVD risk (T1)c0.8780.0310.818, 0.939 < 0.001GroupIntervention group−0.0910.157−0.400, 0.2170.562Control group––––TimeT2––––T3−0.4990.169−0.830, −0.1680.003Interaction effect[Intervention group] × [T3]0.0590.214−0.360, 0.4790.781[Intervention group] × [T2]––––CVD cognitive level score (T1)d0.6350.0260.585, 0.685 < 0.001GroupIntervention group1.0880.2100.677, 1.499 < 0.001Control group––––TimeT2––––T3−0.0270.147−0.316, 0.2620.856Interaction effect[Intervention group] × [T3]0.5840.2660.063, 1.1050.028[Intervention group] × [T2]––––aAfter PSM, none of the other variables that might be associated with the outcomes were statistically different between the intervention and control group; therefore, covariates were not included these variables. bThe outcome was the absolute value of 10-year CVD risk. cThe outcome was the absolute value of 10-year CVD risk. dThe outcome was the absolute value of the CVD cognitive level score. SE, standard errorBold values indicate statistical significance (P < 0.05) The effect evaluation of CVD cognitive level and 10-year CVD risk aAfter PSM, none of the other variables that might be associated with the outcomes were statistically different between the intervention and control group; therefore, covariates were not included these variables. bThe outcome was the absolute value of 10-year CVD risk. cThe outcome was the absolute value of 10-year CVD risk. dThe outcome was the absolute value of the CVD cognitive level score. SE, standard error Bold values indicate statistical significance (P < 0.05) From T1 to T3, although the 10-year CVD risk of PLWH in the intervention group decreased (0.63%) more than in the control group (0.18%) (Fig. 2B), there was no statistical difference in the interaction effect of group × time (P = 0.781) (Table 5). Secondary Outcomes: Health-Related Behaviours Smoking From T1 to T3, the intervention group saw a significant decrease in smoking rate compared with the control group, with a total decrease of 2.66% versus 1.67% in the control group (Fig. 3A). However, there was no statistical difference in smoking rate of group × time (P = 0.284) (Table 6). In addition, from T1 to T3, although PLWH who smoked > 11 cigarettes in the intervention group (1.51%) decreased more than in the control group (1.08%) (Fig. 3A), there was no statistical difference in the interaction effect of group × time (P = 0.276) (Table 6). Health-related behaviours of PLWH (A: Smoking; B: Physical activity; C: Alcohol consumption; D: Dietary nutrition) * T1, T2 and T3 represents baseline, 3 months and 6 months after the baseline, respectively Table 6The effect evaluation of health-related behavioursVariablesaβSE95%CIP valueSmoking now (T1)bYes193.6940.304106.685, 351.667 < 0.001No––––GroupIntervention group0.7320.3160.394, 1.3600.323Control group––––TimeT2––––T30.6520.1890.450, 0.9440.023Interaction effect[Intervention group] × [T3]1.4660.3570.728, 2.9520.284[Intervention group] × [T2]––––Number of cigarettes smoked (T1)c0–5––––6–102.5490.3481.868, 3.231 < 0.001 > 115.0330.4964.026, 6.005 < 0.001GroupIntervention group−0.0800.280−0.628, 0.4690.776Control group––––TimeT2––––T30.2490.1630.070, 0.5670.126Interaction effect[Intervention group] × [T3]0.4660.428 −0.373, 1.3050.276[Intervention group] × [T2]––––Drinking currently (T1)dYes16.6940.18011.732, 23.755 < 0.001No––––GroupIntervention group1.0110.2040.679, 1.5080.955Control group––––TimeT2––––T30.7230.1280.562, 0.9300.012Interaction effect[Intervention group] × [T3]0.8970.2330.568, 1.4160.640[Intervention group] × [T2]––––Drinking frequency (T1)eAlmost never–––– < 1 time per week1.9700.3271.329, 2.612 < 0.0011–3 time(s) per week4.3280.5513.248, 5.408 < 0.0014–7 times per week7.0110.7165.607, 8.414 < 0.001 > 7 times per week4.0610.4343.209, 4.912 < 0.001GroupIntervention group0.2870.245 −0.194, 0.7680.242Control group––––TimeT2––––T30.0020.179 −0.349, 0.3520.993Interaction effect[Intervention group] × [T3]−0.6620.335−1.319, −0.0060.048[Intervention group] × [T2]––––Physical activities up to standard (T1)fYes16.6500.18911.508, 24.090 < 0.001No––––GroupIntervention group1.1640.2110.770, 1.7590.471Control group––––TimeT2––––T30.7720.1030.631, 0.9450.012Interaction effect[Intervention group] × [T3]1.1120.2400.695, 1.7780.658[Intervention group] × [T2]––––The sedentary time (T1)g0–3 h per day––––3–8 h per day1.8320.1611.516, 2.148 < 0.001 > 8 h per day3.8950.3103.287, 4.503 < 0.001GroupIntervention group0.3350.1620.017, 0.6530.039Control group––––TimeT2––––T30.1310.117−0.098, 0.3600.261Interaction effect[Intervention group] × [T3]−0.4040.203−0.801, −0.0060.047[Intervention group] × [T2]––––If the dietary nutrition up to standard (T1)hYes6.3330.4192.787, 14.393 < 0.001No––––GroupIntervention group0.8170.3680.397, 1.6810.582Control group––––TimeT2––––T30.4100.3610.202, 0.8330.014Interaction effect[Intervention group] × [T3]2.4220.5350.848, 6.9120.098[Intervention group] × [T2]––––aAfter PSM, none of the other variables that might be associated with the outcomes were statistically different between the intervention and control group; therefore, covariates were not included these variables. bThe outcome was if smoke now. cThe outcome was the absolute value of the number of cigarettes smoked. dThe outcome was if drinking currently. eThe outcome was the drinking frequency. fThe outcome was if the physical activities up to standard. gThe outcome was the absolute value of sedentary time. hThe outcome was if the dietary nutrition up to standardBold values indicate statistical significance (P < 0.05) The effect evaluation of health-related behaviours aAfter PSM, none of the other variables that might be associated with the outcomes were statistically different between the intervention and control group; therefore, covariates were not included these variables. bThe outcome was if smoke now. cThe outcome was the absolute value of the number of cigarettes smoked. dThe outcome was if drinking currently. eThe outcome was the drinking frequency. fThe outcome was if the physical activities up to standard. gThe outcome was the absolute value of sedentary time. hThe outcome was if the dietary nutrition up to standard Bold values indicate statistical significance (P < 0.05) Alcohol Consumption From T1 to T3, the intervention group saw a significant decrease in drinking rate compared with the control group, with a total decrease of 8.83% versus 4.48% in the control group (Fig. 3B); however, there was no statistical difference in the interaction effect of group × time (P = 0.640) (Table 6). From T1 to T3, the intervention group saw a significant increase in PLWH who were almost never drinking compared with the control group, with a total gain of 4.54% versus 2.78% in the control group (Fig. 3B) with a statistical difference in the interaction effect of group × time (P = 0.048), indicating that the drinking frequency of PLWH in the intervention group decreased significantly more than in the control group after T3 (Table 6). Physical Activity From T1 to T3, although the intervention group and the control group both saw an increase in the standard rate of physical activities (Fig. 3C), there was no statistical difference in the interaction effect of group × time (P = 0.658) (Table 6). From T1 to T3, the intervention group saw a significant decrease in PLWH whose sedentary time was over 8 h per day compared with the control group, with a total decrease of 1.46% versus 0.97% in the control group (Fig. 3C). There was a statistical difference in the interaction effect of group × time (P = 0.047), indicating that the sedentary time in the intervention group decreased significantly more than in the control group after T3 (Table 6). Dietary Nutrition From T1 to T3, the intervention group saw a stable trend in the standard rate of dietary nutrition compared with a decrease of 1.68% in the control group (Fig. 3D). Although the standard rate of dietary nutrition of PLWH in the intervention group was more stable than in the control group, there was no statistical difference in the interaction effect of group × time (P = 0.098) (Table 6). Secondary Outcome: CVD Cognitive Level From T1 to T3, the intervention group showed a significant increase in CVD cognitive level compared with the control group, with a total gain of 2.30 points versus 0.41 in the control group (Fig. 2C). There was a statistical difference in the interaction effect of group × time (P = 0.028), indicating that the CVD cognitive level scores of PLWH in the intervention group increased significantly more than in the control group after T3 (Table 5). Sensitivity Analysis with Full Pre-Matched Samples To examine whether the findings were sensitive to the exclusion of unmatched participants, we further conducted multivariable GEE analyses using the full pre-matched samples. From T1 to T3, for 5-/10-year CVD risk, the results were consistent with the matched analysis with no statistical difference in the interaction effect of group × time. For CVD cognitive level score, the intervention group had a significantly higher adjusted score than the control group at T2, but there was also no statistical difference in the interaction effect of group × time. For health-related behaviours, the drinking frequency remained statistically significant in the full pre-matched samples (P = 0.033). Other behavioural outcomes did not show statistically significant group-by-time interaction effects in the full pre-matched samples. The full results are shown in Supplementary Tables S1 and S2. Discussion This study is one of the first multicentre trials in China targeting reducing CVD risk among PLWH. We found a high CVD risk burden and unhealthy behaviours in this population. While overall risk reduction was not statistically significant in this short-term period, the intervention effectively improved knowledge and showed trends toward healthier behaviours. These findings suggest that risk warning and behavioural intervention may be a useful strategy for improving CVD risk awareness among PLWH, but longer follow-up and more intensive behavioural support may be needed to achieve sustained behavioural change and meaningful CVD risk reduction. CVD Risk of PLWH PLWH tend to exhibit a higher prevalence of traditional CVD risk factors compared with the general population. In our study, the smoking rate among PLWH was 29.43%, which is similar to 33.9% (95% CI: 31.4%–36.5%) reported among HIV-positive men in a previous study [33]. By contrast, data from the 2018 China Adult Tobacco Survey, which included 19,376 participants aged > 15 years, showed a lower smoking rate of 26.6% in the general population [34]. The alcohol consumption rate among PLWH in our study was 36.81%, slightly higher than the national average of 32.5% [35]. Previous evidence has suggested that HIV-related stigma may contribute to increased rates of smoking and alcohol use among PLWH [36]. In terms of physical activity, only 30.06% of PLWH in our study met the recommended standards. Although limited data are available on physical activity levels among PLWH in China, existing research has shown that aerobic exercise can improve vascular endothelial function, myocardial performance and mental well-being in this population [37, 38]. Furthermore, dietary nutrition was also suboptimal among our participants, particularly in the intake of potatoes, fruits and fluid milk. Although some studies have reported that a diagnosis of HIV can motivate individuals to improve health behaviours, with up to 59% reportedly improving their diet after diagnosis [39]. A large cohort study showed that almost half of PLWH had poor dietary nutrition quality [40]. Research from China has consistently shown that PLWH tend to have unbalanced dietary patterns, characterized by insufficient intake of vegetables, fruits, aquatic products, eggs and milk, potentially owing to factors such as poor appetite or limited attention to dietary health [41, 42]. The 5-year CVD risk in our study population was 2.33%, which is comparable to the 2.42% reported in the original D:A:D study [20]. However, the proportion of PLWH with a 10-year CVD risk > 10% reached 10.97%, indicating a substantial long-term risk. This finding is consistent with previous studies conducted in China using the D:A:D(R) model: in 2017 and 2020, two studies among treatment-naive PLWH found that 3.3% (32/973) and 17.26% (65/365) of participants had 5-year and 10-year CVD risk > 10%, respectively [10, 43]. Moreover, a 2022 study evaluating 601 PLWH receiving long-term ART found that 10-year CVD risk > 10% was observed in 4.3% of those on long-term ART and 6.3% of treatment-naive individuals [44]. Overall, our findings reveal that PLWH experience a disproportionately high burden of modifiable CVD risk factors – such as smoking, alcohol use, physical inactivity and poor dietary habits – compared with the general population. This highlights a critical gap in existing HIV care, where CVD prevention remains insufficiently addressed. These results underscore the urgent need for targeted interventions that not only improve knowledge but also promote sustainable behaviour change. The Effect Evaluation of the Intervention The intervention group demonstrated a significantly higher CVD cognitive level compared with the control group, indicating that the intervention effectively improved CVD-related knowledge among PLWH. Although research on CVD cognitive interventions in PLWH remains limited, existing evidence underscores their importance. For instance, a study in the USA found that 76% of PLWH were aware of CVD risk factors, attributing this heightened awareness to targeted public health interventions aimed at reducing CVD burden in this population [22]. Conversely, in settings lacking such interventions, PLWH often exhibit poor CVD awareness. A cross-sectional study in Kenya revealed that most PLWH had limited knowledge of CVD risk factors and underestimated their personal susceptibility to CVD in 2015 [45]. These findings collectively emphasize the critical role and demonstrated efficacy of cognitive interventions in enhancing CVD awareness among PLWH. After the intervention, the intervention group exhibited a greater reduction in both smoking and drinking rates compared with the control group. Additionally, while the proportion of nutrition meeting recommended standards remained stable in the intervention group, it declined in the control group; however, these differences were not statistically significant. Notably, the intervention group demonstrated significant improvements in two key areas: a greater reduction in alcohol consumption frequency and sedentary time, both of which showed statistically significant differences. These findings suggest that, although the short-term intervention did not lead to complete smoking or alcohol cessation among PLWH, it effectively reduced the frequency of health risk behaviours. In the future, longer follow-up may be needed to determine whether these intermediate changes translate into clinically meaningful reductions in CVD risk. Several explanations may account for the absence of statistically significant reductions in CVD risk in this study. First, the participants in this study were relatively young, and only about one-fifth of the participants were aged more than 50 years after PSM. Since the elderly tended to suffer from CVD, younger PLWH may perceive that they were at lower risk of having CVD and ignore health risk behaviours. Second, some PLWH do not focus on health risk behaviours in their daily life because they believe that they will die from HIV in the future and tend to smoke or use alcohol to reduce pain and stress [46, 47], which may also be the reasons for no statistical differences in smoking and alcohol consumption in this study. Third, the intensity and format of the intervention may not have been sufficient to produce measurable short-term changes in clinical CVD risk. For example, the Swiss HIV Cohort Study randomized trial showed that providing computerized CVD risk profiles to physicians, in addition to risk-factor management guidelines, did not significantly improve CVD risk score, suggesting that risk assessment or feedback alone may be insufficient to change CVD risk factors[48]. Similarly, the Australian HealthMap program highlighted that CVD risk management in HIV care may require broader self-management support, such as interactive shared health records, access to personal health information, health coaching and repeated patient–provider engagement [49]. Moreover, owing to the short follow-up period of this study, there may be a delayed effect of behavioural change. Although the CVD cognitive level of PLWH in the intervention group improved, it may be difficult to translate cognition into behaviours in a short period of time, which makes it difficult to show an obvious effectiveness. Indeed, an intervention among rural drinking population also showed that after the intervention, the adjusted abstinence rate of the intervention group was 1.13 times higher than the control group without statistical difference; however, the frequency of alcohol use in the intervention group was less than the control group with statistical difference [50], which was consistent with the results of this study. These studies suggest that future interventions may need to incorporate more continuous and interactive support to achieve clinically meaningful CVD risk reduction. There are some limitations in this study. First, owing to the large scale of PLWH, only part of the AIDS-designated hospitals in Zhejiang province were included in this study as study sites, which may influence the representation of the findings. Second, owing to the inconsistent time interval between the biochemical index tests of PLWH in each study site, the TC and HDL values of T2 and T1 were assumed to be the same among some PLWH, which may also affect the CVD risk. Third, some participants were deceased, discontinued and lost to follow-up, which may have overestimated or underestimated CVD risk and related outcomes. Fourth, because this was a non-randomized controlled trial, residual confounding could not be fully excluded. Although PSM improved baseline comparability between groups, it also excluded a non-negligible number of participants with less comparable baseline covariate profiles, which may limit the generalizability of the matched-sample findings. Therefore, we additionally performed multivariable GEE analyses using the full pre-matched samples. The sensitivity analysis generally supported the main findings, although the statistical significance of some secondary behavioural outcomes varied across analytic approaches. In addition, although the GEE models accounted for within-participant correlation owing to repeated measurements over time, the participants within each PSM pair may not be fully independent, this may have influenced the estimated standard errors and confidence intervals. Therefore, the statistical significance of some matched-sample findings, particularly behavioural outcomes, should be interpreted cautiously. Policy Implications Macro Perspective The development of CVD prevention and treatment guidelines for PLWH and the enhancement of their healthcare services are of critical importance in China. These guidelines would offer healthcare professionals evidence-based recommendations for CVD prevention, diagnosis, treatment and rehabilitation, ensuring that PLWH receive scientifically validated and standardized care. Such measures are expected to significantly reduce CVD incidence and mortality rates among this population. Given that comprehensive HIV care programs were established earlier in high-income countries, it is essential to strengthen international collaboration and knowledge exchange with global health organizations and other nations in the field of HIV and CVD co-management. This cooperative approach would facilitate the implementation of best practices and contribute to more effective CVD risk reduction strategies for PLWH worldwide. Meso Perspective Improving the comprehensive care capacity of medical workers in designated hospitals of AIDS treatment and reducing the prevalence of CVD among PLWH is important. For example, integrating comprehensive CVD services into the daily clinical process for PLWH, and promoting the screening and assessment of CVD risk factors as the clinical routine for PLWH especially those who have not initiated ART [51]. For PLWH with 10-year CVD risk > 10%, it is urgent to strengthen their self-management, regulate follow-up and consider modifying the HAART regimen; for PLWH with low CD4+ T-lymphocyte counts and pre-existing associated underlying diseases, their follow-up testing should be more frequent, with multidisciplinary combined management, electrocardiograms every 3 months and echocardiograms every 6 months [52]. Moreover, special attention needs to be paid to PLWH in vulnerable populations, especially the elderly. Micro Perspective It is pivotal to provide health education-based interventions to PLWH to improve their health literacy. Although only cognition does not improve CVD health, the CVD cognitive level is a key prerequisite for people to make behavioural changes that will improve their CVD health. Therefore, when conducting interventions to reduce the CVD risk in PLWH, it is necessary to link knowledge, risk cognition and health consequences, and adopt effective communication approaches such as motivational interviewing techniques to promote behavioural changes. In addition, improving the mental health of PLWH and providing timely mental health interventions were also significant, considering the sensitivity of HIV and influences of stigma on unhealthy behaviours. Adverse Events No adverse events were observed. Conclusions Risk warning and behavioural intervention can effectively improve CVD cognitive level among PLWH. While this study found no significant differences in smoking rates, alcohol consumption rates or overall CVD risk between the intervention and control groups, the intervention successfully reduced both alcohol consumption frequency and sedentary time. These findings underscore the importance of sustained, long-term CVD risk interventions for PLWH. Supplementary Information Below is the link to the electronic supplementary material. Supplementary file1 (PDF 211 KB) Author Contributions All authors were responsible for the structure of this paper. Yuyin Zhou conducted the literature review, participant recruitment and data collection, data analysis and drafted the paper. Junfang Xu contributed to the study’s conception and design, interpretation of the data and critical revisions of the paper. Jiaye Liu contributed to critical revision of the manuscript. Chunhui Yan, Zhonghua Xia, Xiaowen Lou, Xiangjie Meng, Yaqin Dai, Qin Hong, Zhujun Lian and Xingliang Zhang contributed to participant recruitment and data collection. All authors approved the final version for submission. Funding We are grateful to all study participants for their valuable participation. This work was supported in part by the China Medical Board (No. 20–391). The journal’s Rapid Service Fee was funded by the authors. Data Availability The datasets generated or analysed during the current study are not publicly available because they contained sensitive personal and health-related information of the participants. Declarations Conflict of Interest Yuyin Zhou, Chunhui Yan, Zhonghua Xia, Xiaowen Lou, Xiangjie Meng, Yaqin Dai, Qin Hong, Zhujun Lian, Xingliang Zhang, Jiaye Liu and Junfang Xu confirm that no competing interests exist. Ethical Approval This study was approved by the Medical Ethics Committee of School of Public Health, Zhejiang University (ZGL202203-5). Also, this trial was retrospectively registered with the Chinese Clinical Trial Registry (ChiCTR2500112178) on 11 November 2025 and was conducted according to the protocol approved by the ethics committee before participant recruitment. Footnotes Junfang Xu and Jiaye Liu are joint corresponding authors for this work. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Jiaye Liu, Email: liujiaye1984@163.com. Junfang Xu, Email: xujf2019@zju.edu.cn. References Associated Data Supplementary Materials Supplementary file1 (PDF 211 KB) Data Availability Statement The datasets generated or analysed during the current study are not publicly available because they contained sensitive personal and health-related information of the participants.