Cardiorespiratory Fitness Reference Standards and Prognostic Stratification in Chinese Patients With Cardiovascular Disease Authors: Haoning Cui, Xiaojun Wu, Xiang Xie, Zhongyan Li, Huan Ma, Hongyan Shi, Liyue Zhu, Lin Zhong, Wei Jiang, Jiahong Xue, Na Yang, Hu Tan, Fangfang Bai, Yang Zheng, Shiyu Wang, Xianghui Zheng, Xinyu Hou, Tianhui Cao, Qifeng Li, Bo Yu, Jian Wu Published: 2026 ABSTRACT BackgroundCardiorespiratory fitness (CRF) is a key metric for risk stratification and clinical decision-making in cardiovascular disease (CVD). CRF assessment in Chinese patients with CVD currently relies on reference standards derived from other patients, which may lead to prognostic risk misclassification.ObjectivesThis study aimed to establish CRF reference standards for Chinese patients and assess their utility in stratifying major adverse cardiovascular event (MACE).MethodsWe conducted a multicenter retrospective study including 56,495 patients with CVD from 20 hospitals across China from 2018 to 2024. Sex- and age-specific CRF reference values and percentiles were derived using 10-year age groups and compared with those of healthy Chinese adults. Prognostic performance was evaluated in a validation cohort of 6,317 patients (median follow-up: 24 months [IQR: 12-36]) using Kaplan-Meier and Cox regression analyses.ResultsThe established CRF reference standards demonstrated significant differences across sex, age, and CVD subtype (P < 0.001). Compared with age- and sex-matched healthy adults, patients with CVD had lower peak oxygen uptake, with relative CRF levels of 73.3% to 78.7% in men and 75.4% to 84.0% in women. Compared with the high-CRF group, the adjusted MACE risk was higher in the middle-CRF (HR: 1.46; 95% CI: 1.16-1.83; P < 0.001) and low-CRF groups (HR: 2.29; 95% CI: 1.81-2.90; P < 0.001). Kaplan-Meier analysis revealed significantly lower event-free survival with decreasing CRF (log-rank P < 0.001).ConclusionsThis study establishes the first CRF reference standards specific to Chinese patients with CVD and demonstrates their prognostic utility for MACE risk stratification. These findings provide a data-driven framework for functional risk assessment, secondary prevention, and individualized rehabilitation strategies in clinical practice. Cardiorespiratory fitness (CRF), defined as peak oxygen uptake (VO2 peak), is a comprehensive physiological indicator that reflects the integrated capacity of the cardiovascular, respiratory, and skeletal muscle systems to deliver and utilize oxygen under exercise stress.1, 2, 3, 4 In patients with cardiovascular disease (CVD), CRF is considered an important indicator of disease severity and overall health status, and it plays a critical role in prevention, diagnosis, and treatment.4, 5, 6, 7 Compared with traditional risk factors, lower CRF is more strongly associated with increased cardiovascular events and mortality.8, 9, 10 Although the prognostic value of CRF has been widely established and its importance has led to recommendations that it be considered a fifth vital sign in routine cardiovascular risk assessment,11,12 its clinical application remains highly dependent on the accuracy and patient appropriateness of reference standards. With the advancement of precision medicine, accurate interpretation of an individual’s CRF levels has become essential for risk stratification and treatment decision-making in CVD. In recent years, multiple countries and regions worldwide have reported CRF reference standards based on cardiopulmonary exercise testing (CPET). For healthy study group, various CRF reference systems have been established in Western Europe,13, 14, 15 the Americas,16,17 and China.18, 19, 20, 21 In North America, CRF reference standards for patients with CVD have been reported based on the large, multicenter FRIEND (Fitness Registry and Importance of Exercise National Database) registry.22 Although these studies provide important foundations for the clinical assessment of CRF, accumulating evidence indicates substantial patient heterogeneity in CRF.19, 20, 21 Studies consistently show that age- and sex-matched Chinese adults exhibit lower CRF levels than those observed in North American and European adults.17,19, 20, 21 These differences may result from the combined influence of multiple factors, including physiological characteristics and levels of physical activity.23, 24, 25 Notably, disease status may further amplify patient-related differences in CRF. Studies using the FRIEND registry have shown that CRF levels are lower in North American patients with CVD than in their healthy counterparts.22 However, current CRF reference studies in patients with CVD remain largely focused on North American patients, and systematic, multicenter CRF reference standards for Chinese patients with CVD are still lacking. Because of the lack of patient-specific CRF reference standards for Chinese patients with CVD, current clinical practice in China often relies on CRF reference values derived from North American CVD cohorts or healthy Chinese adults.20, 21, 22 The suitability of these standards for accurately assessing functional status and prognostic risk in Chinese patients with CVD has not been systematically evaluated. Moreover, most existing CRF reference studies focus on descriptive distributions and provide limited validation of their prognostic value for risk stratification within the target study group.17,20, 21, 22 This limitation restricts the clinical application of CRF in clinical risk assessment and secondary prevention. Given the large number of patients with CVD and disease burden in China,26 use of non-patient-specific reference standards may lead to systematic misclassification of CRF, thereby compromising prognostic assessment, exercise prescription, rehabilitation strategies, and long-term management. Therefore, there is an urgent need to establish CRF reference standards for Chinese patients with CVD and to validate their value in major adverse cardiovascular event (MACE) risk stratification. To address this critical evidence gap, the present study utilized cycle ergometer–based CPET data from a large multicenter cohort of patients with CVD recruited across 20 centers in China to establish patient-specific CRF reference standards for Chinese patients with CVD. The clinical value of these standards was further evaluated by validating their ability to stratify the risk of MACE. By integrating patient-specific reference development with prognostic validation, this study aims to support data-driven risk assessment, clinical decision-making, and secondary prevention while contributing to a more precise and globally applicable interpretation of CRF. Methods This study included 56,495 patients with CVD from 20 hospitals across China between January 2018 and December 2024. This study is a retrospective observational study; all the data are anonymized historical data and do not involve patient intervention or additional biological sample collection. All personal information was deidentified, and the exemption of informed consent was reviewed and approved by the Ethics Committee of the Second Affiliated Hospital of Harbin Medical University (KY2025-031). The inclusion criteria for participation were as follows: 1) diagnosis of myocardial infarction (MI), heart failure (HF), or stable angina (SA) or having undergone percutaneous coronary intervention (PCI); and 2) completion of a symptom-limited CPET session on a cycle ergometer. The exclusion criteria were as follows: 1) age <30 or >79 years; 2) peak respiratory exchange ratio <1.0; and 3) outliers and extreme values. All CPET data underwent centralized quality control, and CRF values exceeding ±3 SD from the mean were defined as extreme values and excluded from analysis. Suspected outliers were independently reviewed by 2 researchers, and only those confirmed as erroneous or implausible upon consensus were removed. Cohorts for analysis After applying the predefined inclusion and exclusion criteria, a total of 20,224 patients with CVD who underwent cycle ergometer–based CPET from 20 hospitals in China were included, and CPET was performed by experienced personnel at each center (Supplemental Table 1). All CPETs were conducted either during the inpatient assessment or within one month after discharge. Despite differences in equipment, tests were conducted at all cardiac rehabilitation (CR) centers in accordance with published guidelines.27 These hospitals are located in 7 major regions in China, including 9 hospitals in Northeast China, 3 in Northwest China, 2 in East China, 2 in South China, 2 in North China, 1 in Central China, and 1 in Southwest China. We created CRF reference standards for patients with CVD aged 30 to 79 years based on specific CVD categories. Patients were assigned to 1 of 4 mutually exclusive categories according to a prespecified hierarchy: 1) the HF group, which included all patients diagnosed with HF regardless of MI or PCI history; 2) the MI group, which included patients with a history of MI but no HF; 3) the PCI group, which included patients who underwent percutaneous coronary intervention without a history of MI or HF; and 4) the SA group, which included patients diagnosed solely with SA and no other CVD conditions. External comparison of CRF reference values To contextualize CRF levels in the present Chinese CVD cohort, external comparisons were conducted using published CRF reference data from healthy Chinese adults and North American patients with CVD. Reference values for healthy Chinese adults were obtained from Wang et al,21 the only multicenter study providing age- and sex-specific VO2 peak reference standards based on cycle ergometer CPET, ensuring methodologic comparability. Among available Chinese studies,18, 19, 20, 21 this study was therefore selected as the most appropriate domestic reference. For international comparison, CRF data for North American patients with CVD were derived from the large multicenter cohort reported by Peterman et al,22 which provides standardized CPET-derived VO2 peak values and is widely used as an international reference. Across all comparisons, VO2 peak (mL/kg/min) was used as the CRF metric. When comparing patients with CVD, analyses were further restricted to patients with the same disease category to enhance clinical comparability. All cross-study comparisons were descriptive in nature. External reference values were used as fixed benchmarks and were not re-estimated or recalibrated using the current data set. CRF percentile-based stratification and prognostic validation CRF percentile reference standards were derived from 20,224 cycle ergometer–based CPET examinations with the directly measured VO2 peak. For the overall CVD cohort and for each disease subtype (MI, PCI, SA, and HF), percentiles (P10, P25, P50, P75, and P90) were calculated separately by sex (men, women) and 10-year age group (30-39, 40-49, 50-59, 60-69, and 70-79 years). Within each category, the corresponding sex- and age-specific CRF percentiles served as cut points to classify patients into low (≤25th percentile), middle (25th-75th percentile), or high (≥75th percentile) CRF groups. To evaluate the prognostic significance of CRF, this predefined percentile-based stratification scheme was applied and subsequently validated in a cohort comprising patients from 3 participating centers (n = 6,317), with a median follow-up duration of 24 months (IQR: 12-36 months). Follow-up was conducted at 1, 3, 6, and 12 months after discharge through telephone or outpatient interviews. For patients with follow-up exceeding 1 year, subsequent follow-ups were conducted at 6-month intervals. The primary outcome was MACE, defined as a composite endpoint of all-cause death, recurrent MI, stroke, and cardiac rehospitalization. For the overall CVD cohort and the MI subgroup, associations between CRF categories and clinical outcomes were evaluated for MACE as well as their individual components. For MACE and all-cause death, Kaplan-Meier curves were constructed and compared using the log-rank test, and multivariable Cox proportional hazards models were used to estimate HRs and 95% CIs. The crude model included the CRF group as the sole exposure. Model 1 was additionally adjusted for age, sex, and body mass index (BMI); model 2 was further adjusted for MI, PCI, SA, and HF. The proportional hazard assumption for each Cox model was assessed using scaled Schoenfeld residuals. For nonfatal components of MACE, including recurrent MI, stroke, and cardiac rehospitalization, cumulative incidence function curves were used and compared with Gray’s test, with all-cause death treated as a competing event; associations were further assessed using Fine-Gray subdistribution hazard models. For the PCI, SA, and HF subgroups, analyses were limited to MACE because of limited event numbers, and individual outcome components were not analyzed to avoid model instability. Statistical analyses Statistical analysis was performed using R 4.3.2. (R foundation for statistical computing). For the overall CVD CRF reference standards, a 2-way analysis of variance with a sex × age group interaction term was used to compare differences in CRF between sexes and across age groups and to test their multiplicative interaction. One-way analysis of variance was applied to assess differences in CRF across age groups within each CVD category. No post hoc analysis was conducted because the focus was on general differences and trends. P values <0.05 were considered to indicate statistical significance. Continuous data are reported as the mean ± SD, whereas categorical data are presented as frequencies (%). For the validation cohort, baseline characteristics were compared between patients with and without MACE using Student’s t-test for continuous variables and the chi-square test for categorical variables. Comparisons of event rates across the three CRF groups (low, middle, and high) were performed using the chi-square test or Fisher exact test when expected cell counts were <5. Event rates were estimated with 95% CIs calculated using the Wilson score method. Results This study evaluated a total of 20,224 patients with CVD who underwent cycle ergometer–based CPET (see the enrollment process in Figure 1), including 13,963 men (69%) and 6,261 women (31%). Baseline characteristics stratified by sex and age group are summarized in Table 1. The average VO2 peak differed significantly by sex (F = 687.60, P < 0.001) and age group (F = 643.17, P < 0.001), with a significant sex × age group interaction (F = 5.46, P < 0.001). Overall, compared with women, men exhibited a 1.7 (95% CI: 1.5-1.8) mL/kg/min higher mean VO2 peak. The VO2 peak decreased by 1.4 mL/kg/min per decade in men and 1.6 mL/kg/min per decade in women.Figure 1Study Enrollment Flow ChartThis flow chart depicts the patient selection process of this multicenter study. Among 56,495 patients with cardiovascular disease who underwent cycle ergometer–based cardiopulmonary exercise testing (CPET) at 20 hospitals in China between 2018 and 2024, 20,224 patients were included in the final analytical cohort after applying predefined inclusion and exclusion criteria.Table 1Descriptive Characteristics of the Chinese CVD Cohort for the Cycling Ergometer–Based CPET AnalysisMenAge Group30-39 (n = 1,089)40-49 (n = 2,619)50-59 (n = 5,052)60-69 (n = 4,055)70-79 (n = 1,148)All (N = 13,963)Height, cm174.78 ± 5.43172.63 ± 5.70171.43 ± 5.62170.53 ± 5.76169.41 ± 6.20171.49 ± 5.86Weight, kg82.28 ± 12.6680.34 ± 11.8276.53 ± 10.4573.41 ± 10.3170.80 ± 9.8776.32 ± 11.31BMI, kg/m226.89 ± 3.7426.92 ± 3.4426.04 ± 4.1225.21 ± 3.0124.64 ± 3.0025.92 ± 3.66Age, years36.29 ± 2.9046.09 ± 2.8455.68 ± 2.9065.00 ± 2.9073.72 ± 2.4156.56 ± 10.29VO2 peak, L/min1.76 ± 0.421.60 ± 0.381.46 ± 0.351.28 ± 0.301.13 ± 0.271.43 ± 0.38VO2 peak, mL/kg/min21.66 ± 5.1320.09 ± 4.7619.20 ± 4.3817.62 ± 3.9516.10 ± 3.5218.81 ± 4.56RER peak1.16 ± 0.091.15 ± 0.091.14 ± 0.091.12 ± 0.081.10 ± 0.081.14 ± 0.09HR peak, beats/min147.54 ± 21.63136.42 ± 20.58130.09 ± 19.42124.00 ± 18.12118.61 ± 17.50129.93 ± 20.63Women(n = 384)(n = 898)(n = 2,236)(n = 2,195)(n = 548)(N = 6,261)Height, cm162.66 ± 5.41160.94 ± 5.04160.05 ± 5.25159.69 ± 5.23158.32 ± 5.18160.06 ± 5.29Weight, kg63.22 ± 12.6763.74 ± 9.9063.52 ± 9.6762.17 ± 8.7860.47 ± 8.2662.79 ± 9.54BMI, kg/m223.81 ± 4.2924.61 ± 3.6624.77 ± 3.4624.36 ± 3.1224.10 ± 2.9524.49 ± 3.40Age, y36.29 ± 2.9046.09 ± 2.8455.68 ± 2.9065.11 ± 2.8873.54 ± 2.3958.06 ± 9.94VO2 peak, L/min1.28 ± 0.281.21 ± 0.261.12 ± 0.260.98 ± 0.230.86 ± 0.201.07 ± 0.27VO2 peak, mL/kg/min20.68 ± 4.5919.15 ± 4.1217.74 ± 3.9415.91 ± 3.4814.28 ± 2.8417.18 ± 4.12RER peak1.14 ± 0.091.14 ± 0.091.12 ± 0.081.11 ± 0.081.10 ± 0.081.12 ± 0.08HR peak, beats/min155.51 ± 21.06146.42 ± 22.14137.26 ± 21.75128.11 ± 19.23121.68 ± 19.69135.13 ± 22.54Values are mean ± SD.BMI = body mass index; CPET = cardiopulmonary exercise testing; CVD = cardiovascular disease; HR = heart rate; RER = respiratory exchange ratio; VO2 peak = peak oxygen consumption. Study Enrollment Flow Chart This flow chart depicts the patient selection process of this multicenter study. Among 56,495 patients with cardiovascular disease who underwent cycle ergometer–based cardiopulmonary exercise testing (CPET) at 20 hospitals in China between 2018 and 2024, 20,224 patients were included in the final analytical cohort after applying predefined inclusion and exclusion criteria. Table 1Descriptive Characteristics of the Chinese CVD Cohort for the Cycling Ergometer–Based CPET AnalysisMenAge Group30-39 (n = 1,089)40-49 (n = 2,619)50-59 (n = 5,052)60-69 (n = 4,055)70-79 (n = 1,148)All (N = 13,963)Height, cm174.78 ± 5.43172.63 ± 5.70171.43 ± 5.62170.53 ± 5.76169.41 ± 6.20171.49 ± 5.86Weight, kg82.28 ± 12.6680.34 ± 11.8276.53 ± 10.4573.41 ± 10.3170.80 ± 9.8776.32 ± 11.31BMI, kg/m226.89 ± 3.7426.92 ± 3.4426.04 ± 4.1225.21 ± 3.0124.64 ± 3.0025.92 ± 3.66Age, years36.29 ± 2.9046.09 ± 2.8455.68 ± 2.9065.00 ± 2.9073.72 ± 2.4156.56 ± 10.29VO2 peak, L/min1.76 ± 0.421.60 ± 0.381.46 ± 0.351.28 ± 0.301.13 ± 0.271.43 ± 0.38VO2 peak, mL/kg/min21.66 ± 5.1320.09 ± 4.7619.20 ± 4.3817.62 ± 3.9516.10 ± 3.5218.81 ± 4.56RER peak1.16 ± 0.091.15 ± 0.091.14 ± 0.091.12 ± 0.081.10 ± 0.081.14 ± 0.09HR peak, beats/min147.54 ± 21.63136.42 ± 20.58130.09 ± 19.42124.00 ± 18.12118.61 ± 17.50129.93 ± 20.63Women(n = 384)(n = 898)(n = 2,236)(n = 2,195)(n = 548)(N = 6,261)Height, cm162.66 ± 5.41160.94 ± 5.04160.05 ± 5.25159.69 ± 5.23158.32 ± 5.18160.06 ± 5.29Weight, kg63.22 ± 12.6763.74 ± 9.9063.52 ± 9.6762.17 ± 8.7860.47 ± 8.2662.79 ± 9.54BMI, kg/m223.81 ± 4.2924.61 ± 3.6624.77 ± 3.4624.36 ± 3.1224.10 ± 2.9524.49 ± 3.40Age, y36.29 ± 2.9046.09 ± 2.8455.68 ± 2.9065.11 ± 2.8873.54 ± 2.3958.06 ± 9.94VO2 peak, L/min1.28 ± 0.281.21 ± 0.261.12 ± 0.260.98 ± 0.230.86 ± 0.201.07 ± 0.27VO2 peak, mL/kg/min20.68 ± 4.5919.15 ± 4.1217.74 ± 3.9415.91 ± 3.4814.28 ± 2.8417.18 ± 4.12RER peak1.14 ± 0.091.14 ± 0.091.12 ± 0.081.11 ± 0.081.10 ± 0.081.12 ± 0.08HR peak, beats/min155.51 ± 21.06146.42 ± 22.14137.26 ± 21.75128.11 ± 19.23121.68 ± 19.69135.13 ± 22.54Values are mean ± SD.BMI = body mass index; CPET = cardiopulmonary exercise testing; CVD = cardiovascular disease; HR = heart rate; RER = respiratory exchange ratio; VO2 peak = peak oxygen consumption. Descriptive Characteristics of the Chinese CVD Cohort for the Cycling Ergometer–Based CPET Analysis Values are mean ± SD. BMI = body mass index; CPET = cardiopulmonary exercise testing; CVD = cardiovascular disease; HR = heart rate; RER = respiratory exchange ratio; VO2 peak = peak oxygen consumption. The CVD subgroup characteristics and variations in the VO2 peak are presented in Table 2. Similar patterns were observed across the MI, PCI, SA, and HF subgroups, with consistently higher VO2 peaks in men and an age-related decline in both sexes. Owing to the limited sample sizes of women aged 30 to 39 years in the MI, PCI, and HF groups, analyses in women were restricted to those aged 40 to 79 years. The age- and sex-stratified comparisons of CRF between Chinese patients with CVD and healthy reference adults are shown in Figure 2 and Supplemental Table 2. Across all age groups and both sexes, patients with CVD exhibited substantially lower VO2 peaks than did age- and sex-matched healthy individuals. Although the VO2 peak decreased with age in both study groups, the proportional CRF deficit in patients with CVD persisted across all age groups, indicating a sustained fitness impairment independent of age and sex.Table 2Cycle Ergometer–Based CPET Responses at Peak Effort According to CVD CategoryMIMenWomen30-39 (n= 231)40-49 (n = 618)50-59 (n= 1,175)60-69 (n = 787)70-79 (n = 207)All (N = 3,018)30-39 (n = 9)40-49 (n = 51)50-59 (n = 204)60-69 (n = 251)70-79 (n = 65)All (N = 580)Height, cm174.75 ± 5.48172.28 ± 5.64171.28 ± 5.56170.32 ± 5.57168.21 ± 5.98171.29 ± 5.78155.89 ± 2.86159.45 ± 4.57159.62 ± 5.59159.31 ± 5.45157.37 ± 5.45159.16 ± 5.44Weight, kg84.98 ± 12.3680.49 ± 12.0976.34 ± 10.1472.85 ± 10.0270.13 ± 10.6176.51 ± 11.4255.74 ± 14.2166.19 ± 11.6063.99 ± 9.2161.33 ± 8.6158.01 ± 8.2662.23 ± 9.44BMI, kg/m227.80 ± 3.6227.06 ± 3.4825.99 ± 2.9425.09 ± 3.0024.72 ± 3.1026.02 ± 3.2622.97 ± 5.9425.96 ± 3.8925.10 ± 3.3524.14 ± 3.0523.36 ± 2.6724.53 ± 3.33Age, y36.60 ± 2.6646.12 ± 2.8655.58 ± 2.9464.92 ± 2.8873.79 ± 2.3955.88 ± 9.9734.44 ± 3.5046.84 ± 2.5156.07 ± 2.8565.29 ± 3.0273.37 ± 2.1160.85 ± 8.38VO2, L/min1.55 ± 0.371.44 ± 0.321.34 ± 0.291.20 ± 0.261.07 ± 0.231.32 ± 0.320.84 ± 0.281.05 ± 0.200.97 ± 0.240.89 ± 0.210.76 ± 0.170.92 ± 0.23VO2 peak, mL/kg/min18.48 ± 4.4718.05 ± 3.8417.68 ± 3.6216.54 ± 3.4515.40 ± 3.1417.36 ± 3.7515.09 ± 2.9816.05 ± 3.1815.18 ± 3.2114.54 ± 3.0213.19 ± 2.4514.76 ± 3.12RER peak1.15 ± 0.091.15 ± 0.091.13 ± 0.081.11 ± 0.081.10 ± 0.081.13 ± 0.091.21 ± 0.101.15 ± 0.081.12 ± 0.081.10 ± 0.091.08 ± 0.061.11 ± 0.09HR peak, beats/min137.43 ± 18.50129.41 ± 16.45124.59 ± 16.66120.91 ± 15.79116.31 ± 15.63125.03 ± 17.22146.67 ± 16.37130.29 ± 16.10124.79 ± 15.60120.10 ± 16.33117.43 ± 16.45122.76 ± 16.68PCI(n = 146)(n = 594)(n = 1,318)(n = 1,231)(n = 338)(N = 3,627)(n = 8)(n = 68)(n= 296)(n = 517)(n = 136)(N = 1,025)Height, cm173.91 ± 5.95172.74 ± 5.51171.34 ± 5.57170.61 ± 5.58170.09 ± 6.50171.31 ± 5.75162.38 ± 5.88159.35 ± 4.36158.99 ± 5.00159.25 ± 4.77158.04 ± 5.01159.04 ± 4.87Weight, kg83.30 ± 13.9781.18 ± 11.2075.99 ± 10.3473.43 ± 9.9070.67 ± 10.0175.77 ± 11.0068.25 ± 8.6065.16 ± 8.5363.61 ± 9.3361.40 ± 8.959.91 ± 8.0662.14 ± 8.72BMI, kg/m227.48 ± 4.0027.17 ± 3.2925.85 ± 3.0125.18 ± 2.8424.40 ± 2.9725.77 ± 3.1625.92 ± 3.0925.66 ± 3.3425.14 ± 3.3924.19 ± 2.9423.96 ± 2.9024.54 ± 3.15Age, y36.93 ± 2.9146.34 ± 2.7555.89 ± 2.8165.04 ± 2.8573.46 ± 2.3458.31 ± 9.4136.12 ± 3.0445.87 ± 3.0656.15 ± 2.7665.27 ± 2.8673.46 ± 2.3462.21 ± 7.93VO2, L/min1.67 ± 0.351.55 ± 0.331.42 ± 0.301.27 ± 0.271.14 ± 0.251.38 ± 0.321.01 ± 0.121.03 ± 0.211.02 ± 0.210.92 ± 0.190.83 ± 0.160.94 ± 0.20VO2 peak, mL/kg/min20.40 ± 4.4719.29 ± 4.0118.89 ± 3.9817.41 ± 3.5616.22 ± 3.3218.27 ± 3.9614.81 ± 1.5515.91 ± 2.9416.00 ± 3.2215.12 ± 2.8413.80 ± 2.0015.25 ± 2.94RER peak1.16 ± 0.091.16 ± 0.101.15 ± 0.091.13 ± 0.091.11 ± 0.081.14 ± 0.091.12 ± 0.101.13 ± 0.101.12 ± 0.091.11 ± 0.091.11 ± 0.081.11 ± 0.09HR peak, beats/min146.60 ± 19.09132.71 ± 17.14129.85 ± 18.05124.06 ± 17.51118.95 ± 17.06127.93 ± 18.46141.88 ± 25.80137.97 ± 15.85131.46 ± 18.90127.08 ± 16.77119.12 ± 16.11128.12 ± 17.95SA(n = 655)(n = 1,237)(n = 2,164)(n = 1,579)(n = 427)(N = 6,062)(n = 360)(n = 745)(n = 1,621)(n = 1,253)(n = 263)(N = 4,242)Height, cm175.01 ± 5.31172.83 ± 5.85171.68 ± 5.70170.58 ± 5.95169.69 ± 5.93171.85 ± 5.94162.84 ± 5.31161.24 ± 5.03160.33 ± 5.28159.95 ± 5.40158.84 ± 5.11160.50 ± 5.34Weight, kg81.11 ± 12.3979.61 ± 11.9576.99 ± 10.5973.71 ± 10.6171.47 ± 9.3776.73 ± 11.3763.23 ± 12.4563.41 ± 9.9063.44 ± 9.7262.75 ± 8.9561.56 ± 8.3563.09 ± 9.73BMI, kg/m226.44 ± 3.6726.61 ± 3.4326.18 ± 5.2625.30 ± 3.1324.81 ± 2.9825.97 ± 4.1523.77 ± 4.2224.40 ± 3.6424.65 ± 3.4624.50 ± 3.1924.38 ± 3.0224.47 ± 3.47Age, y36.03 ± 2.9645.89 ± 2.8855.52 ± 2.9164.87 ± 2.9273.73 ± 2.4955.17 ± 10.6935.71 ± 3.1746.28 ± 2.8755.82 ± 2.8664.90 ± 2.8673.56 ± 2.4356.22 ± 10.11VO2, L/min1.88 ± 0.411.73 ± 0.391.59 ± 0.361.39 ± 0.321.22 ± 0.281.57 ± 0.401.31 ± 0.271.24 ± 0.251.17 ± 0.241.06 ± 0.220.96 ± 0.191.15 ± 0.26VO2 peak, mL/kg/min23.37 ± 4.8421.99 ± 4.8420.79 ± 4.4819.04 ± 4.0917.13 ± 3.6020.60 ± 4.7421.07 ± 4.3919.85 ± 3.9318.68 ± 3.7616.99 ± 3.3715.61 ± 2.7218.40 ± 3.95RER peak1.16 ± 0.091.15 ± 0.091.15 ± 0.091.12 ± 0.081.11 ± 0.071.14 ± 0.091.14 ± 0.091.14 ± 0.091.13 ± 0.081.12 ± 0.081.10 ± 0.071.13 ± 0.08HR peak, beats/min152.71 ± 21.51142.91 ± 22.00134.51 ± 20.44126.82 ± 18.84121.84 ± 18.32135.30 ± 22.14156.36 ± 20.84149.01 ± 21.99140.50 ± 21.96131.59 ± 19.82125.09 ± 20.64139.76 ± 22.80HF(n = 57)(n = 170)(n = 395)(n = 458)(n = 176)(N = 1,256)(n = 7)(n = 34)(n = 115)(n = 174)(n = 84)(N = 414)Height, cm174.53 ± 5.27172.12 ± 5.39170.85 ± 5.45170.45 ± 5.86168.87 ± 6.29170.77 ± 5.82162.00 ± 7.68159.88 ± 5.99159.54 ± 4.35159.69 ± 4.89157.87 ± 5.36159.33 ± 5.04Weight, kg82.13 ± 11.9682.24 ± 11.5376.32 ± 10.9473.30 ± 10.7870.20 ± 9.8575.43 ± 11.4866.57 ± 22.0164.50 ± 9.2663.64 ± 10.5361.57 ± 8.9159.90 ± 7.8962.13 ± 9.63BMI, kg/m226.95 ± 3.5727.75 ± 3.6326.09 ± 3.1125.17 ± 3.0624.59 ± 2.9825.81 ± 3.3125.03 ± 6.2525.12 ± 2.8024.98 ± 3.8024.13 ± 3.2624.03 ± 2.9224.44 ± 3.40Age, y36.35 ± 2.8646.59 ± 2.6456.13 ± 2.9065.47 ± 2.9574.10 ± 2.3759.87 ± 10.0634.86 ± 3.7246.88 ± 3.3456.18 ± 2.7365.87 ± 2.7073.73 ± 2.5962.69 ± 9.06VO2, L/min1.49 ± 0.321.34 ± 0.301.21 ± 0.281.10 ± 0.260.99 ± 0.251.17 ± 0.300.92 ± 0.280.96 ± 0.260.85 ± 0.190.76 ± 0.200.70 ± 0.130.79 ± 0.21VO2 peak, mL/kg/min18.15 ± 3.1616.47 ± 3.5415.99 ± 3.5015.11 ± 3.2414.19 ± 3.1315.58 ± 3.4614.58 ± 5.7115.06 ± 4.0913.50 ± 2.9012.50 ± 3.0411.76 ± 2.3112.87 ± 3.15RER peak1.15 ± 0.091.14 ± 0.091.13 ± 0.081.10 ± 0.081.10 ± 0.071.12 ± 0.081.14 ± 0.111.10 ± 0.081.11 ± 0.081.10 ± 0.071.09 ± 0.081.10 ± 0.08HR peak, beats/min136.56 ± 21.52127.68 ± 19.77123.06 ± 19.40119.37 ± 19.13112.81 ± 16.63121.52 ± 19.79138.71 ± 20.81130.71 ± 23.87128.69 ± 21.66117.73 ± 18.21118.44 ± 22.52122.34 ± 21.35Values are mean ± SD.BMI = body mass index; CPET = cardiopulmonary exercise testing; CVD = cardiovascular disease; HF = heart failure; HR = heart rate; MI = myocardial infarction; PCI = percutaneous coronary intervention; RER = respiratory exchange ratio; SA = stable angina; VO2 peak = peak oxygen uptake.Figure 2Age- and Sex-Stratified Cardiorespiratory Fitness (CRF) in Chinese Patients With Cardiovascular Disease (CVD) and Healthy AdultsAge- and sex-specific peak oxygen uptake between Chinese patients with cardiovascular disease and age- and sex-matched healthy Chinese adults. Bars represent mean peak oxygen uptake (VO2 peak) values, whereas the overlaid line denotes relative CRF expressed as a percentage of healthy reference values. Across all age groups and in both sexes, patients with CVD exhibited substantially lower CRF than healthy adults, indicating a persistent functional deficit independent of age and sex. Table 2Cycle Ergometer–Based CPET Responses at Peak Effort According to CVD CategoryMIMenWomen30-39 (n= 231)40-49 (n = 618)50-59 (n= 1,175)60-69 (n = 787)70-79 (n = 207)All (N = 3,018)30-39 (n = 9)40-49 (n = 51)50-59 (n = 204)60-69 (n = 251)70-79 (n = 65)All (N = 580)Height, cm174.75 ± 5.48172.28 ± 5.64171.28 ± 5.56170.32 ± 5.57168.21 ± 5.98171.29 ± 5.78155.89 ± 2.86159.45 ± 4.57159.62 ± 5.59159.31 ± 5.45157.37 ± 5.45159.16 ± 5.44Weight, kg84.98 ± 12.3680.49 ± 12.0976.34 ± 10.1472.85 ± 10.0270.13 ± 10.6176.51 ± 11.4255.74 ± 14.2166.19 ± 11.6063.99 ± 9.2161.33 ± 8.6158.01 ± 8.2662.23 ± 9.44BMI, kg/m227.80 ± 3.6227.06 ± 3.4825.99 ± 2.9425.09 ± 3.0024.72 ± 3.1026.02 ± 3.2622.97 ± 5.9425.96 ± 3.8925.10 ± 3.3524.14 ± 3.0523.36 ± 2.6724.53 ± 3.33Age, y36.60 ± 2.6646.12 ± 2.8655.58 ± 2.9464.92 ± 2.8873.79 ± 2.3955.88 ± 9.9734.44 ± 3.5046.84 ± 2.5156.07 ± 2.8565.29 ± 3.0273.37 ± 2.1160.85 ± 8.38VO2, L/min1.55 ± 0.371.44 ± 0.321.34 ± 0.291.20 ± 0.261.07 ± 0.231.32 ± 0.320.84 ± 0.281.05 ± 0.200.97 ± 0.240.89 ± 0.210.76 ± 0.170.92 ± 0.23VO2 peak, mL/kg/min18.48 ± 4.4718.05 ± 3.8417.68 ± 3.6216.54 ± 3.4515.40 ± 3.1417.36 ± 3.7515.09 ± 2.9816.05 ± 3.1815.18 ± 3.2114.54 ± 3.0213.19 ± 2.4514.76 ± 3.12RER peak1.15 ± 0.091.15 ± 0.091.13 ± 0.081.11 ± 0.081.10 ± 0.081.13 ± 0.091.21 ± 0.101.15 ± 0.081.12 ± 0.081.10 ± 0.091.08 ± 0.061.11 ± 0.09HR peak, beats/min137.43 ± 18.50129.41 ± 16.45124.59 ± 16.66120.91 ± 15.79116.31 ± 15.63125.03 ± 17.22146.67 ± 16.37130.29 ± 16.10124.79 ± 15.60120.10 ± 16.33117.43 ± 16.45122.76 ± 16.68PCI(n = 146)(n = 594)(n = 1,318)(n = 1,231)(n = 338)(N = 3,627)(n = 8)(n = 68)(n= 296)(n = 517)(n = 136)(N = 1,025)Height, cm173.91 ± 5.95172.74 ± 5.51171.34 ± 5.57170.61 ± 5.58170.09 ± 6.50171.31 ± 5.75162.38 ± 5.88159.35 ± 4.36158.99 ± 5.00159.25 ± 4.77158.04 ± 5.01159.04 ± 4.87Weight, kg83.30 ± 13.9781.18 ± 11.2075.99 ± 10.3473.43 ± 9.9070.67 ± 10.0175.77 ± 11.0068.25 ± 8.6065.16 ± 8.5363.61 ± 9.3361.40 ± 8.959.91 ± 8.0662.14 ± 8.72BMI, kg/m227.48 ± 4.0027.17 ± 3.2925.85 ± 3.0125.18 ± 2.8424.40 ± 2.9725.77 ± 3.1625.92 ± 3.0925.66 ± 3.3425.14 ± 3.3924.19 ± 2.9423.96 ± 2.9024.54 ± 3.15Age, y36.93 ± 2.9146.34 ± 2.7555.89 ± 2.8165.04 ± 2.8573.46 ± 2.3458.31 ± 9.4136.12 ± 3.0445.87 ± 3.0656.15 ± 2.7665.27 ± 2.8673.46 ± 2.3462.21 ± 7.93VO2, L/min1.67 ± 0.351.55 ± 0.331.42 ± 0.301.27 ± 0.271.14 ± 0.251.38 ± 0.321.01 ± 0.121.03 ± 0.211.02 ± 0.210.92 ± 0.190.83 ± 0.160.94 ± 0.20VO2 peak, mL/kg/min20.40 ± 4.4719.29 ± 4.0118.89 ± 3.9817.41 ± 3.5616.22 ± 3.3218.27 ± 3.9614.81 ± 1.5515.91 ± 2.9416.00 ± 3.2215.12 ± 2.8413.80 ± 2.0015.25 ± 2.94RER peak1.16 ± 0.091.16 ± 0.101.15 ± 0.091.13 ± 0.091.11 ± 0.081.14 ± 0.091.12 ± 0.101.13 ± 0.101.12 ± 0.091.11 ± 0.091.11 ± 0.081.11 ± 0.09HR peak, beats/min146.60 ± 19.09132.71 ± 17.14129.85 ± 18.05124.06 ± 17.51118.95 ± 17.06127.93 ± 18.46141.88 ± 25.80137.97 ± 15.85131.46 ± 18.90127.08 ± 16.77119.12 ± 16.11128.12 ± 17.95SA(n = 655)(n = 1,237)(n = 2,164)(n = 1,579)(n = 427)(N = 6,062)(n = 360)(n = 745)(n = 1,621)(n = 1,253)(n = 263)(N = 4,242)Height, cm175.01 ± 5.31172.83 ± 5.85171.68 ± 5.70170.58 ± 5.95169.69 ± 5.93171.85 ± 5.94162.84 ± 5.31161.24 ± 5.03160.33 ± 5.28159.95 ± 5.40158.84 ± 5.11160.50 ± 5.34Weight, kg81.11 ± 12.3979.61 ± 11.9576.99 ± 10.5973.71 ± 10.6171.47 ± 9.3776.73 ± 11.3763.23 ± 12.4563.41 ± 9.9063.44 ± 9.7262.75 ± 8.9561.56 ± 8.3563.09 ± 9.73BMI, kg/m226.44 ± 3.6726.61 ± 3.4326.18 ± 5.2625.30 ± 3.1324.81 ± 2.9825.97 ± 4.1523.77 ± 4.2224.40 ± 3.6424.65 ± 3.4624.50 ± 3.1924.38 ± 3.0224.47 ± 3.47Age, y36.03 ± 2.9645.89 ± 2.8855.52 ± 2.9164.87 ± 2.9273.73 ± 2.4955.17 ± 10.6935.71 ± 3.1746.28 ± 2.8755.82 ± 2.8664.90 ± 2.8673.56 ± 2.4356.22 ± 10.11VO2, L/min1.88 ± 0.411.73 ± 0.391.59 ± 0.361.39 ± 0.321.22 ± 0.281.57 ± 0.401.31 ± 0.271.24 ± 0.251.17 ± 0.241.06 ± 0.220.96 ± 0.191.15 ± 0.26VO2 peak, mL/kg/min23.37 ± 4.8421.99 ± 4.8420.79 ± 4.4819.04 ± 4.0917.13 ± 3.6020.60 ± 4.7421.07 ± 4.3919.85 ± 3.9318.68 ± 3.7616.99 ± 3.3715.61 ± 2.7218.40 ± 3.95RER peak1.16 ± 0.091.15 ± 0.091.15 ± 0.091.12 ± 0.081.11 ± 0.071.14 ± 0.091.14 ± 0.091.14 ± 0.091.13 ± 0.081.12 ± 0.081.10 ± 0.071.13 ± 0.08HR peak, beats/min152.71 ± 21.51142.91 ± 22.00134.51 ± 20.44126.82 ± 18.84121.84 ± 18.32135.30 ± 22.14156.36 ± 20.84149.01 ± 21.99140.50 ± 21.96131.59 ± 19.82125.09 ± 20.64139.76 ± 22.80HF(n = 57)(n = 170)(n = 395)(n = 458)(n = 176)(N = 1,256)(n = 7)(n = 34)(n = 115)(n = 174)(n = 84)(N = 414)Height, cm174.53 ± 5.27172.12 ± 5.39170.85 ± 5.45170.45 ± 5.86168.87 ± 6.29170.77 ± 5.82162.00 ± 7.68159.88 ± 5.99159.54 ± 4.35159.69 ± 4.89157.87 ± 5.36159.33 ± 5.04Weight, kg82.13 ± 11.9682.24 ± 11.5376.32 ± 10.9473.30 ± 10.7870.20 ± 9.8575.43 ± 11.4866.57 ± 22.0164.50 ± 9.2663.64 ± 10.5361.57 ± 8.9159.90 ± 7.8962.13 ± 9.63BMI, kg/m226.95 ± 3.5727.75 ± 3.6326.09 ± 3.1125.17 ± 3.0624.59 ± 2.9825.81 ± 3.3125.03 ± 6.2525.12 ± 2.8024.98 ± 3.8024.13 ± 3.2624.03 ± 2.9224.44 ± 3.40Age, y36.35 ± 2.8646.59 ± 2.6456.13 ± 2.9065.47 ± 2.9574.10 ± 2.3759.87 ± 10.0634.86 ± 3.7246.88 ± 3.3456.18 ± 2.7365.87 ± 2.7073.73 ± 2.5962.69 ± 9.06VO2, L/min1.49 ± 0.321.34 ± 0.301.21 ± 0.281.10 ± 0.260.99 ± 0.251.17 ± 0.300.92 ± 0.280.96 ± 0.260.85 ± 0.190.76 ± 0.200.70 ± 0.130.79 ± 0.21VO2 peak, mL/kg/min18.15 ± 3.1616.47 ± 3.5415.99 ± 3.5015.11 ± 3.2414.19 ± 3.1315.58 ± 3.4614.58 ± 5.7115.06 ± 4.0913.50 ± 2.9012.50 ± 3.0411.76 ± 2.3112.87 ± 3.15RER peak1.15 ± 0.091.14 ± 0.091.13 ± 0.081.10 ± 0.081.10 ± 0.071.12 ± 0.081.14 ± 0.111.10 ± 0.081.11 ± 0.081.10 ± 0.071.09 ± 0.081.10 ± 0.08HR peak, beats/min136.56 ± 21.52127.68 ± 19.77123.06 ± 19.40119.37 ± 19.13112.81 ± 16.63121.52 ± 19.79138.71 ± 20.81130.71 ± 23.87128.69 ± 21.66117.73 ± 18.21118.44 ± 22.52122.34 ± 21.35Values are mean ± SD.BMI = body mass index; CPET = cardiopulmonary exercise testing; CVD = cardiovascular disease; HF = heart failure; HR = heart rate; MI = myocardial infarction; PCI = percutaneous coronary intervention; RER = respiratory exchange ratio; SA = stable angina; VO2 peak = peak oxygen uptake. Cycle Ergometer–Based CPET Responses at Peak Effort According to CVD Category Values are mean ± SD. BMI = body mass index; CPET = cardiopulmonary exercise testing; CVD = cardiovascular disease; HF = heart failure; HR = heart rate; MI = myocardial infarction; PCI = percutaneous coronary intervention; RER = respiratory exchange ratio; SA = stable angina; VO2 peak = peak oxygen uptake. Age- and Sex-Stratified Cardiorespiratory Fitness (CRF) in Chinese Patients With Cardiovascular Disease (CVD) and Healthy Adults Age- and sex-specific peak oxygen uptake between Chinese patients with cardiovascular disease and age- and sex-matched healthy Chinese adults. Bars represent mean peak oxygen uptake (VO2 peak) values, whereas the overlaid line denotes relative CRF expressed as a percentage of healthy reference values. Across all age groups and in both sexes, patients with CVD exhibited substantially lower CRF than healthy adults, indicating a persistent functional deficit independent of age and sex. Age- and sex-specific differences in CRF between healthy adults and CVD subtypes are shown in Supplemental Figure 1, with detailed values in Supplemental Table 2. Distinct patterns of CRF impairment were observed across disease categories. Patients with MI and HF exhibited the greatest reductions in CRF across age groups and both sexes, whereas patients with SA had relatively preserved CRF, and patients with PCI showed intermediate impairment. Although the absolute VO2 peak decreased with age in both sexes across all subtypes, the magnitude of CRF reduction relative to that in healthy adults differed markedly by disease category. Age-matched, sex-matched, and disease subtype–matched comparisons between Chinese patients with CVD and North American patients with CVD from the FRIEND registry are shown in Supplemental Figure 2.22 In the MI and PCI subgroups, Chinese patients generally exhibited lower VO2 peak values than patients from the FRIEND registry across most age and sex strata. In contrast, among patients with HF, relative CRF in Chinese patients approached or exceeded 100% of FRIEND registry reference values in several male age groups and remained largely comparable among women. Overall, these external comparisons suggest disease-specific heterogeneity in CRF across patients. Using directly measured VO2 peaks from 20,224 cycle ergometer–based CPET examinations, patient-specific CRF percentile reference standards were established for the overall CVD cohort and individual subtypes. As shown in Figure 3A and Supplemental Table 3, the VO2 peak percentiles differed across disease categories, with higher values in the SA group and lower values in the HF group. For clinical interpretability, the VO2 peak was additionally converted to metabolic equivalents (METs; 1 MET = 3.5 mL/kg/min), with the MET-based percentile distributions presented in Supplemental Figure 3.Figure 3Percentile-Based Classification of Cardiorespiratory Fitness (CRF) Levels in Chinese Patients With Cardiovascular Disease (CVD)Patient-specific CRF reference standards derived from 20,224 cycle ergometer–based cardiopulmonary exercise testing examinations in Chinese patients with CVD. (A) Shows the percentile distribution of peak oxygen consumption for the overall CVD cohort and major CVD subtypes, including myocardial infarction (MI), percutaneous coronary intervention (PCI), stable angina (SA), and heart failure (HF). (B) Displays age- and sex-specific CRF percentile matrices across disease categories. Color gradients indicate low, middle, and high CRF levels based on patient-specific percentiles. Percentile-Based Classification of Cardiorespiratory Fitness (CRF) Levels in Chinese Patients With Cardiovascular Disease (CVD) Patient-specific CRF reference standards derived from 20,224 cycle ergometer–based cardiopulmonary exercise testing examinations in Chinese patients with CVD. (A) Shows the percentile distribution of peak oxygen consumption for the overall CVD cohort and major CVD subtypes, including myocardial infarction (MI), percutaneous coronary intervention (PCI), stable angina (SA), and heart failure (HF). (B) Displays age- and sex-specific CRF percentile matrices across disease categories. Color gradients indicate low, middle, and high CRF levels based on patient-specific percentiles. Figure 3B shows the age- and sex-specific CRF percentile matrices for the overall CVD cohort and each disease subtype, with numeric values summarized in Supplemental Tables 4 and 5. VO2 peak percentiles decreased progressively with age in both sexes across all disease categories. MI and PCI showed intermediate CRF levels within the age and sex strata, whereas HF consistently showed the lowest CRF and SA maintained relatively high CRF levels. Although absolute VO2 peak values were lower in women than in men, the overall patterns were similar across sexes. For visualization, CRF was categorized into low (≤25th percentile), middle (25th-75th percentile), and high (≥75th percentile) levels based on patient-specific percentiles, providing a standardized representation of the CRF distribution across age, sex, and disease subgroups. The baseline characteristics of the validation group are summarized in Supplemental Table 6. The mean age was 58.5 ± 10.0 years; 80.0% were men, the median follow-up duration was 24 months (IQR: 12-36), and the mean VO2 peak was 16.4 ± 4.1 mL/kg/min. Patients who experienced MACE were older and had lower VO2 peak values than those without events, whereas BMI did not differ between groups. The incidence of MACE and individual components across CRF categories is summarized in Supplemental Table 7. The CRF groups were defined a priori using the established percentile-based reference standards. In the overall cohort (n = 6,317), the incidence of MACE differed significantly across CRF levels (P < 0.001), with rates of 27.3%, 19.4%, and 11.8% in the low-, middle-, and high-CRF groups, respectively (see Supplemental Table 7 for detailed event counts and 95% CIs). Similar graded associations were observed for all-cause death (P < 0.001), recurrent MI (P = 0.009), and cardiac rehospitalization (P = 0.018), whereas the incidence of stroke did not significantly differ across the CRF strata (P = 0.104). In the MI subgroup, the incidence of MACE and all individual components decreased significantly across increasing CRF categories (all P < 0.001). Similar associations were observed in the PCI subgroup for MACE and all components (all P < 0.05) except stroke (P = 0.056). In contrast, differences in MACE and individual components across CRF categories were not statistically significant in the SA and HF subgroups, although event rates were numerically higher in the low-CRF groups, likely reflecting limited sample sizes and event numbers. Table 3 summarizes the associations between CRF levels and the risk of MACE and their individual components among patients with CVD. The proportional hazard assumption was examined using scaled Schoenfeld residuals for all Cox models. The CRF category satisfied the assumption in every model (all P > 0.05), supporting the validity of the reported HR. Compared with the high-CRF group, the low-CRF group was associated with progressively higher MACE risk. In model 2, the middle-CRF group had a 46% higher risk of MACE (HR: 1.46, 95% CI: 1.16-1.83; P = 0.001) and the low-CRF group had a >2-fold higher risk (HR: 2.29; 95% CI: 1.81-2.90; P < 0.001). Low CRF was also independently associated with all-cause death (HR: 4.72; 95% CI: 1.84-12.09; P = 0.001). In Fine-Gray competing risk analyses treating all-cause death as a competing event, low CRF remained significantly associated with recurrent MI (sHR: 4.07; 95% CI: 1.73-9.55; P = 0.001) and cardiac rehospitalization (sHR: 2.07; 95% CI: 1.62-2.64; P < 0.001). However, associations for the middle-CRF group were weaker and not statistically significant for all-cause death, recurrent MI, or stroke. In contrast, the middle-CRF group showed a significantly elevated risk of cardiac rehospitalization (sHR: 1.45; 95% CI: 1.15-1.83; P = 0.002). The crude association between low CRF and stroke (sHR: 2.47; 95% CI: 1.19-5.14; P = 0.016) was attenuated and no longer significant after adjustment (sHR: 1.93; 95% CI: 0.86-4.34; P = 0.110).Table 3Association Between the Risk of MACE and Different CRF Groups in Patients With CVDEventsCrude ModelModel 1Model 2HR (95% CI)P ValueHR (95% CI)P ValueHR (95% CI)P ValueMACE High-CRF groupReferenceReferenceReference Middle-CRF group1.49 (1.19-1.86)<0.0011.45 (1.16-1.82)0.0011.46 (1.16-1.83)0.001 Low-CRF group2.37 (1.91-2.96)<0.0012.24 (1.78-2.83)<0.0012.29 (1.81-2.90)<0.001All-cause death High-CRF groupReferenceReferenceReference Middle-CRF group1.68 (0.66-4.24)0.2761.75 (0.68-4.46)0.2431.57 (0.61-4.02)0.351 Low-CRF group4.34 (1.76-10.7)0.0014.96 (1.95-12.61)<0.0014.72 (1.84-12.09)0.001Recurrent MI High-CRF groupReferenceReferenceReference Middle-CRF group2.11 (0.91-4.90)0.0812.21 (0.95-5.13)0.0641.94 (0.84-4.47)0.120 Low-CRF group4.15 (1.81-9.51)<0.0014.44 (1.88-10.47)<0.0014.07 (1.73-9.55)0.001Stroke High-CRF groupReferenceReferenceReference Middle-CRF group1.43 (0.68-3.01)0.3401.25 (0.58-2.70)0.5601.15 (0.53-2.46)0.730 Low-CRF group2.47 (1.19-5.14)0.0161.94 (0.86-4.36)0.1101.93 (0.86-4.34)0.110Cardiac rehospitalization High-CRF groupReferenceReferenceReference Middle-CRF group1.48 (1.18-1.87)<0.0011.44 (1.14-1.82)0.0021.45 (1.15-1.83)0.002 Low-CRF group2.20 (1.75-2.77)<0.0012.04 (1.60-2.61)<0.0012.07 (1.62-2.64)<0.001Model 1: Adjusted for age, sex, and BMI.Model 2: Adjusted for age, sex, BMI, PCI, MI, SA, and HF.HRs for MACE and all-cause death were estimated using Cox proportional hazards models. For recurrent MI, stroke, and cardiac rehospitalization, subdistribution HRs were estimated using Fine-Gray models, treating all-cause death as a competing event.BMI = body mass index; CRF = cardiorespiratory fitness; CVD = cardiovascular disease; HF = heart failure; MACE = major adverse cardiovascular event; MI = myocardial infarction; PCI = percutaneous coronary intervention; SA = stable angina. Table 3Association Between the Risk of MACE and Different CRF Groups in Patients With CVDEventsCrude ModelModel 1Model 2HR (95% CI)P ValueHR (95% CI)P ValueHR (95% CI)P ValueMACE High-CRF groupReferenceReferenceReference Middle-CRF group1.49 (1.19-1.86)<0.0011.45 (1.16-1.82)0.0011.46 (1.16-1.83)0.001 Low-CRF group2.37 (1.91-2.96)<0.0012.24 (1.78-2.83)<0.0012.29 (1.81-2.90)<0.001All-cause death High-CRF groupReferenceReferenceReference Middle-CRF group1.68 (0.66-4.24)0.2761.75 (0.68-4.46)0.2431.57 (0.61-4.02)0.351 Low-CRF group4.34 (1.76-10.7)0.0014.96 (1.95-12.61)<0.0014.72 (1.84-12.09)0.001Recurrent MI High-CRF groupReferenceReferenceReference Middle-CRF group2.11 (0.91-4.90)0.0812.21 (0.95-5.13)0.0641.94 (0.84-4.47)0.120 Low-CRF group4.15 (1.81-9.51)<0.0014.44 (1.88-10.47)<0.0014.07 (1.73-9.55)0.001Stroke High-CRF groupReferenceReferenceReference Middle-CRF group1.43 (0.68-3.01)0.3401.25 (0.58-2.70)0.5601.15 (0.53-2.46)0.730 Low-CRF group2.47 (1.19-5.14)0.0161.94 (0.86-4.36)0.1101.93 (0.86-4.34)0.110Cardiac rehospitalization High-CRF groupReferenceReferenceReference Middle-CRF group1.48 (1.18-1.87)<0.0011.44 (1.14-1.82)0.0021.45 (1.15-1.83)0.002 Low-CRF group2.20 (1.75-2.77)<0.0012.04 (1.60-2.61)<0.0012.07 (1.62-2.64)<0.001Model 1: Adjusted for age, sex, and BMI.Model 2: Adjusted for age, sex, BMI, PCI, MI, SA, and HF.HRs for MACE and all-cause death were estimated using Cox proportional hazards models. For recurrent MI, stroke, and cardiac rehospitalization, subdistribution HRs were estimated using Fine-Gray models, treating all-cause death as a competing event.BMI = body mass index; CRF = cardiorespiratory fitness; CVD = cardiovascular disease; HF = heart failure; MACE = major adverse cardiovascular event; MI = myocardial infarction; PCI = percutaneous coronary intervention; SA = stable angina. Association Between the Risk of MACE and Different CRF Groups in Patients With CVD Model 1: Adjusted for age, sex, and BMI. Model 2: Adjusted for age, sex, BMI, PCI, MI, SA, and HF. HRs for MACE and all-cause death were estimated using Cox proportional hazards models. For recurrent MI, stroke, and cardiac rehospitalization, subdistribution HRs were estimated using Fine-Gray models, treating all-cause death as a competing event. BMI = body mass index; CRF = cardiorespiratory fitness; CVD = cardiovascular disease; HF = heart failure; MACE = major adverse cardiovascular event; MI = myocardial infarction; PCI = percutaneous coronary intervention; SA = stable angina. The associations between CRF levels and the risk of MACE and their individual components among patients with MI are summarized in Supplemental Table 8. Compared with the high-CRF group, the low-CRF group was associated with a significantly higher risk of MACE, whereas the association for the middle-CRF group was not significant. In model 1, low CRF was associated with a more than 2-fold increased risk of MACE (HR: 2.07; 95% CI: 1.63-2.63; P < 0.001). Low CRF was also independently associated with a markedly increased risk of all-cause death (HR: 7.90; 95% CI: 2.70-23.13; P < 0.001). In Fine-Gray competing risk analyses, low CRF was significantly associated with recurrent MI (sHR: 9.17; 95% CI: 2.18-38.60; P = 0.002) and cardiac rehospitalization (sHR: 2.49; 95% CI: 1.75-3.54; P < 0.001). Notably, the middle-CRF group also showed a significantly elevated subdistribution hazard of cardiac rehospitalization (sHR: 1.62; 95% CI: 1.15-2.27; P = 0.006). The crude association between low CRF and stroke (sHR: 2.81; 95% CI: 1.01-7.87; P = 0.049) was attenuated and no longer significant after adjustment (sHR: 2.46; 95% CI: 0.81-7.51; P = 0.110). Figure 4 presents Kaplan-Meier curves for cumulative event-free survival across CRF groups. Clear graded separation was observed, with lower CRF associated with worse event-free survival for MACE (Figure 4A) and all-cause death (Figure 4B). The middle-CRF group consistently demonstrated intermediate survival, and differences across groups were significant for both endpoints (all log-rank P < 0.001). For outcomes subject to competing risks (recurrent MI, stroke, and cardiac rehospitalization), cumulative incidence functions accounting for all-cause death as a competing event are shown in Supplemental Figure 4. Low CRF was associated with the highest cumulative incidence of recurrent MI (Gray's test, P < 0.001; Supplemental Figure 4A), stroke (P = 0.002; Supplemental Figure 4B), and cardiac rehospitalization (P < 0.001; Supplemental Figure 4C). Across all 3 endpoints, the high-CRF group exhibited the lowest cumulative incidence, with the middle-CRF group falling in between.Figure 4Kaplan-Meier Curves for MACE and All-Cause Death by CRF GroupKaplan–Meier survival curves illustrate the association between CRF (cardiorespiratory fitness) categories and clinical outcomes in Chinese CVD (cardiovascular disease) patients. Panels show event-free survival for (A) major adverse cardiovascular events (MACE) and (B) all-cause death. Patients were classified into low, middle, and high CRF groups based on population-specific percentiles. Lower CRF levels were consistently associated with poorer outcomes. CRF = cardiorespiratory fitness; CVD = cardiovascular disease; MACE = major adverse cardiovascular events. Kaplan-Meier Curves for MACE and All-Cause Death by CRF Group Kaplan–Meier survival curves illustrate the association between CRF (cardiorespiratory fitness) categories and clinical outcomes in Chinese CVD (cardiovascular disease) patients. Panels show event-free survival for (A) major adverse cardiovascular events (MACE) and (B) all-cause death. Patients were classified into low, middle, and high CRF groups based on population-specific percentiles. Lower CRF levels were consistently associated with poorer outcomes. CRF = cardiorespiratory fitness; CVD = cardiovascular disease; MACE = major adverse cardiovascular events. Supplemental Figure 5 presents Kaplan-Meier survival curves for MACE and all-cause death in patients with MI. Compared with the middle- and high-CRF groups, the low-CRF group consistently had lower event-free survival for both MACE (Supplemental Figure 5A) and all-cause death (Supplemental Figure 5B), with the middle-CRF group demonstrating intermediate survival (both log-rank P < 0.001). Cumulative incidence functions accounting for all-cause death as a competing event are shown in Supplemental Figure 6. Low CRF was associated with the highest cumulative incidence of recurrent MI (Gray's test, P < 0.001; Supplemental Figure 6A), stroke (P = 0.002; Supplemental Figure 6B), and cardiac rehospitalization (P < 0.001; Supplemental Figure 6C), with clear graded separation across CRF groups. The results of the Kaplan-Meier analyses of MACE stratified by CRF levels in the PCI (Supplemental Figure 7A), SA (Supplemental Figure 7B), and HF subgroups (Supplemental Figure 7C) are shown in Supplemental Figure 7. Across these disease-specific subgroups, patients with lower CRF levels were associated with poorer MACE-free survival than those with higher CRF levels; the difference was significant for PCI (log-rank P < 0.001) but not for SA or HF, likely reflecting the limited number of events and the reduced statistical power. Discussion This study is the first to establish CRF reference standards specifically for Chinese patients with CVD using large-scale cycle ergometer–based CPET data from 20 centers in China. To our knowledge, it represents the largest CRF reference study in the Chinese adults to date and includes the largest number of female participants undergoing CPET. By comparing CRF levels in Chinese patients with CVD with those in healthy Chinese adults and North American CVD cohorts, this study extends current understanding of patient- and disease-specific CRF differences and fills a critical evidence gap. Moreover, the established percentile-based CRF reference standards were validated for MACE risk stratification in a validation cohort, confirming that lower CRF levels are independently associated with a higher risk of MACE in Chinese patients with CVD. These findings provide a practical framework for interpreting CRF levels, risk stratification, and prognostic assessment in these patients and further support the clinical value of CRF in secondary prevention (Central Illustration). Collectively, our results underscore the importance of integrating CRF assessment into routine clinical evaluation of patients with CVD.Central IllustrationCardiorespiratory Fitness (CRF) Reference Standards and Risk Stratification in Chinese Patients With Cardiovascular Disease (CVD)For the first time, CRF reference standards were established for Chinese patients with CVD using cycle ergometer–based cardiopulmonary exercise testing (CPET) data from 20 centers. CRF levels in Chinese patients with CVD were only 73.3% to 84.0% of those in age- and sex-matched healthy adults. Based on patient-specific CRF percentiles, patients were classified into low-, middle-, and high-CRF categories. This percentile-based framework enables simple, clinically actionable risk assessment, with progressively higher risks of MACE observed at lower CRF percentiles. MACE = major adverse cardiovascular events. Cardiorespiratory Fitness (CRF) Reference Standards and Risk Stratification in Chinese Patients With Cardiovascular Disease (CVD) For the first time, CRF reference standards were established for Chinese patients with CVD using cycle ergometer–based cardiopulmonary exercise testing (CPET) data from 20 centers. CRF levels in Chinese patients with CVD were only 73.3% to 84.0% of those in age- and sex-matched healthy adults. Based on patient-specific CRF percentiles, patients were classified into low-, middle-, and high-CRF categories. This percentile-based framework enables simple, clinically actionable risk assessment, with progressively higher risks of MACE observed at lower CRF percentiles. MACE = major adverse cardiovascular events. The CRF patterns observed in this study are consistent with previous reports, showing a progressive decline in the VO2 peak with age and significantly lower values in women than in men.21,22 This age- and sex-related pattern has been well documented in both healthy adults and patients with CVD,17,20,22 reflecting physiological aging processes and sex-specific differences in functional reserve. Beyond confirming these established trends, the present study further contextualized CRF levels in Chinese patients with CVD by direct comparison with age- and sex-matched healthy Chinese adults, thereby providing a patient-specific reference framework for disease-related interpretation. In recent years, 4 studies have reported CRF reference values in healthy Chinese adults.18, 19, 20, 21 Among these, we selected the only study that derived CRF reference standards from a multicenter study group using cycle ergometer–based CPET as the comparator to maximize consistency in both patient characteristics and testing modality.21 Compared with this healthy reference study group, we found that, within the same age and sex strata, CRF levels in Chinese patients with CVD were markedly reduced, amounting to only 73.3% to 84.0% of the corresponding healthy reference values. This difference directly reflects the substantial impairment in CRF associated with CVD. In addition, the present study established disease-specific CRF reference standards for different types of CVD in the Chinese study group. Consistent with previous CRF studies in patients with CVD, clear differences were observed across disease categories.22 Among these subtypes, patients with SA had the highest CRF levels, whereas patients with HF had the lowest. Such differences are likely closely related to the extent of myocardial injury associated with different disease types, as well as the accompanying systemic metabolic abnormalities and skeletal muscle dysfunction.22 For example, patients with HF often demonstrate pronounced exercise intolerance owing to severely impaired cardiac pump function, together with systemic metabolic dysregulation and skeletal muscle deterioration, resulting in markedly reduced CRF.28 These findings underscore the importance of interpreting CRF levels within a disease-specific context for an accurate assessment of functional status and prognostic risk. In addition to delineating differences in CRF levels between Chinese patients with CVD and the local healthy adults, we further explored whether CRF exhibits regional specificity. At present, the only study that has established CRF reference standards based on a large, multicenter cohort of patients with CVD is derived from the FRIEND registry, which predominantly represents North American patients (the United States and Canada) and includes 15,045 CPET records, of which 6,966 were obtained using cycle ergometer–based CPET.22 When comparing the CRF data derived from the cycle ergometer–based CPET in the FRIEND study with those from the present cohort, we observed that, within the same age, sex, and disease category strata, the CRF levels were not fully concordant between Chinese and North American patients with CVD. Specifically, CRF levels in Chinese patients with MI and PCI were generally lower than those in their North American counterparts, whereas CRF levels in patients with HF were broadly comparable to those observed in North American patients with HF and were even slightly higher in some strata.22 These differences are likely attributable to the combined effects of multiple factors. From a physiological perspective, Asian people generally have lower skeletal muscle mass and strength than Western people,29 as well as relatively smaller lung volumes,30 both of which may limit CRF. Lifestyle and environmental factors are also important considerations because previous studies have shown that a relatively high proportion of Asian people are physically inactive, with a relatively small percentage meeting recommended levels of physical activity.23,29 In addition, differences at the health care system level may play a substantial role. Although CR has been shown to significantly improve CRF and clinical outcomes,31 its coverage and participation rates in China remain markedly lower than those in North America.32 Globally, approximately 54.7% of regions have access to CR services,33 whereas in China, only approximately 24% of large medical centers offer CR programs, with even more limited availability in rural areas.34 The combined influence of physiological characteristics, lifestyle factors, and health care system differences may therefore help explain the observed disparities in CRF levels between Chinese and North American patients with CVD. These findings further suggest that the direct application of CRF reference standards established in other regions may lead to systematic misclassification of functional status and prognostic risk in Chinese patients with CVD. Consequently, they underscore the necessity of developing patient-specific CRF reference standards while also highlighting that CRF reference values should be used as supportive decision-making tools and interpreted in conjunction with individual patient characteristics in clinical practice. Although previous studies have highlighted the prognostic value of CRF,7 most CRF reference standard studies have remained at the level of descriptive characterization of CRF distributions,21,22 and direct evidence demonstrating whether the defined reference values can achieve effective clinical risk stratification in the corresponding patient is lacking. Building on this gap, the present study applied CRF percentile-based reference thresholds derived from a large, multicenter Chinese CVD cohort to a validation cohort and systematically evaluated their associations with the risk of MACE. In doing so, this study achieved a critical transition of CRF reference standards from purely descriptive constructs to practical tools for clinical risk stratification. The results demonstrated that low, middle, and high CRF levels defined by these reference standards were able to significantly discriminate the risk of MACE, with the associations remaining robust after adjustment for potential confounding factors. Kaplan-Meier analyses further confirmed a clear dose-response relationship between CRF and adverse outcomes, whereby lower CRF levels were associated with poorer event-free survival. These findings not only validate the independent prognostic value of CRF in Chinese patients with CVD but also address a key evidence gap in prior CRF reference research, which has largely lacked prognostic validation. Collectively, the results underscore the central role of CRF as a functional risk marker in the long-term management of patients with CVD. From a clinical practice perspective, the transition achieved in this study—from the establishment of CRF percentile thresholds to their prognostic validation—provides clinicians with an operational, interpretable, and prognostically meaningful framework for CRF assessment and interpretation. Specifically, following completion of CPET, an individual’s VO2 peak or METs can be mapped to age-specific, sex-specific, and disease category–specific CRF percentiles, thereby assisting in risk appraisal and informing risk stratification and secondary prevention decisions. In addition, the present study demonstrated that CRF is a significant metric for predicting cardiac-related rehospitalization, suggesting that the clinical relevance of CRF extends beyond survival outcomes and may also reflect patients’ disease management needs and health care resource utilization. Overall, these findings further support the incorporation of CRF as a quantifiable functional metric into clinical assessment and risk stratification frameworks for patients with CVD to inform long-term management and secondary prevention strategies. Study Strengths and limitations This study has several strengths. To our knowledge, this is the largest multicenter investigation to date to establish CRF reference standards specifically for Chinese patients with CVD based on cycle ergometer–derived CPET data from 20 centers. The large sample size and broad geographic coverage in China increase the robustness and representativeness of the derived reference standards. Importantly, beyond descriptive reference construction, this study validated the prognostic relevance of percentile-based CRF stratification in a validation cohort, thereby enabling the translation of CRF reference values into a clinically meaningful risk stratification tool. However, several limitations should be acknowledged. First, this study included only CPET results obtained with a cycle ergometer. The CRF performance of patients with CVD assessed via treadmill-based CPET should be further considered. Second, there are fewer data from women in the age group of 30 to 39 years, especially after grouping according to specific CVD categories, which may lead to potential deviations; therefore, more samples need to be included in the future. Third, this reference standard does not consider other comorbidities that may affect CRF. Fourth, this study is a retrospective study, which may be affected by selection and information bias. Fifth, this study established reference standards according to clinical disease categories; however, CRF heterogeneity may still be present within each category because of various potential influencing factors. Future studies should build upon this framework by incorporating relevant clinical and functional indicators to achieve more refined stratification within different disease categories, thereby further improving the precision and clinical applicability of the reference standards. Finally, although the CPET data are all from hospitals with complete qualifications and rich experience, different test equipment and protocols may cause deviations in the CRF results. Conclusions In conclusion, based on the largest available data set of cycle ergometer–based CPET in Chinese patients with CVD, this study is the first to establish CRF reference standards for this study group. In doing so, it fills a long-standing gap arising from the absence of patient-specific reference standards and expands the global CRF data landscape. These findings further demonstrate that CRF levels differ substantially across patient and disease contexts, underscoring the necessity of interpreting CRF within disease- and patient-specific frameworks. By integrating CRF percentile-based reference standards with prognostic validation, this study provides a function-oriented risk stratification framework for Chinese patients with CVD while simultaneously offering data-driven support for clinical decision-making and personalized rehabilitation interventions. Collectively, these results contribute to advancing CVD management from a disease-centered approach toward a precision prevention and rehabilitation model guided by functional assessment, thereby supporting improved patient outcomes and more efficient allocation of health care resources. Funding Support and Author Disclosures This study was supported by the Heilongjiang Provincial Medical and Health Scientific Research Project (grant No. 20240303010021, Harbin, Heilongjiang, China), the Key Research and Development Program of Heilongjiang Province (grant No. 2022ZX01A28, Harbin, Heilongjiang, China), the National Natural Science Foundation of China (grant Nos. 82172537 and 82372565, Beijing, China), and the Natural Science Foundation of Heilongjiang Province (grant No. ZL2024H007, Harbin, Heilongjiang, China). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.