A biological age evaluation system based on clinical indicators, a construction method and application

Abstract

The present application relates to a biological age evaluation system based on clinical indicators, a construction method and application, belonging to the field of biological aging research. First, the clinical indicators of the subjects are collected, the estimation model of biological age is constructed based on the clinical indicators, and the biological aging rate is further calculated. The present application can be used to identify individuals with accelerated aging or delayed aging, realize the risk prediction of death and aging-related diseases and the risk stratification of the population. The present application considers the nonlinear relationship between clinical indicators and aging degree, so that the biological age estimated by the present method has better prediction effect on the risk of death and new onset of aging-related diseases. The present application quantifies biological age and biological aging rate based on clinical indicators, and the measurement of clinical indicators can be completed by routine physical examination and blood sampling once, which has the advantages of low cost and convenience, and can evaluate the biological aging rate in time after physical examination, and provide health guidance for individuals with accelerated aging.

Description

Technical Field

[0001] This invention relates to the field of biological aging research, and more specifically, to a biological age assessment system based on clinical indicators, its construction method, and its application. Background Technology

[0002] To address age-related diseases and slow or reverse the aging process, it is essential to measure biological age and identify individuals aging at an accelerated or decelerated rate. In existing research, readily available and inexpensive clinical indicators are ideal biomarkers for constructing predictive indicators of biological age. Chen et al. constructed the KDM-BA using 12,377 individuals from the Kadoorie Biobank in China, a method that estimates biological age by assigning higher weights to biomarkers more relevant to age (Chen, Zhang et al. 2023). However, biological age estimators built using supervised models trained on chronological age capture a significant amount of chronological-related variation but provide limited chronological-independent information for disease prediction (Galkin, Mamoshina et al. 2020). Some researchers argue that the survival time of individuals of the same chronological age reflects the relative magnitude of their biological age, thus focusing on developing all-cause mortality risk predictors (Lu, Quach et al. 2019, Pyrkov, Avchaciov et al. 2021). However, existing biological age assessment methods are primarily based on European populations and do not consider the nonlinear association between clinical indicators and mortality risk. Therefore, there is a lack of biological age assessment methods in this field that are based on large prospective cohorts of Chinese populations and take into account the nonlinear association between clinical indicators and mortality risk.

[0003] Chinese invention patent CN 118448041A describes a method for estimating the rate of physiological aging using clinical indicators. This method assumes a linear correlation between each clinical indicator and actual age and mortality risk. Based on this assumption, the patent first uses linear regression to correct the actual age for each clinical indicator, then predicts the mortality risk during a 6-year follow-up period based on the residual values ​​of the corrected clinical indicators, and finally estimates the rate of physiological aging of the sample based on this prediction model. However, previous studies (Nguyen, Colacino et al. 2021) have shown that most clinical indicators exhibit a U-shaped association with mortality risk, meaning that values ​​above or below clinically recommended values ​​lead to a higher mortality risk. Summary of the Invention

[0004] To address the shortcomings of existing biological age estimators, such as the lack of studies based on Chinese population cohorts and the failure to consider the nonlinear relationship between clinical indicators and mortality risk, this invention proposes a biological age assessment system, its construction method, and its application based on clinical indicators. The aim is to provide a biological age assessment method based on a large prospective Chinese population cohort and using readily available clinical indicators such as complete blood count and blood biochemistry.

[0005] According to a first aspect of the present invention, a method for constructing a biological age assessment system is provided, comprising the following steps:

[0006] (1) Collect clinical indicators of subjects;

[0007] (2) Perform nonlinear Cox proportional hazards regression analysis on the clinical indicators described in step (1) against the risk of death;

[0008] (3) For clinical indicators that are not non-linearly associated with mortality risk in step (2), the original values ​​are not converted; for clinical indicators that are U-shapedly associated with mortality risk in step (2), their original measurements are converted based on their optimal values; the optimal value is defined as the value of the clinical variable in the restricted cubic spline (RCS) curve when the relative mortality risk is the lowest.

[0009] (4) Based on actual age, the original clinical indicators obtained in step (3) and the transformed variables, the risk of death is predicted, and the LASSO algorithm is used to screen variables in the Cox proportional hazards regression model.

[0010] (5) Calculate the biological age (PAI) of each sample based on the variables obtained in step (4);

[0011] (6) Based on the biological age PAI calculated in step (5) and the actual age of the sample, calculate the biological aging rate ΔPAI of each sample.

[0012] Preferably, in step (1), the indicators are height, weight, blood pressure, blood lipids, blood glucose, routine blood indicators and blood biochemical indicators, and the skewed distribution indicators among the indicators are normalized.

[0013] Preferably, in step (2), the nonlinear Cox proportional hazards regression model uses the restricted cubic spline method, with actual age and a single clinical indicator as covariates, and the nodes of the clinical indicator are selected as the 10%, 50% and 90% quantiles.

[0014] Preferably, in step (3), the formula for variable transformation of the clinical indicators is:

[0015]

[0016]

[0017] Among them, y i and Let represent the original value and the clinical optimum value of the i-th clinical indicator, respectively.

[0018] Preferably, in step (4), the Cox proportional hazards regression model is expressed as:

[0019]

[0020] in, h0(t) and β represent the hazard rate function and baseline hazard function at time t, respectively. i Let y represent the regression effect value of the i-th indicator. i· Let represent the i-th variable in all samples, and p be the number of the indicators. The covariates included in the Cox proportional hazards regression model include actual age (CA), the original values ​​of clinical indicators linearly associated with mortality, and the transformed values ​​of clinical indicators non-linearly associated with mortality.

[0021] Preferably, in step (5), the PAI calculation formula for sample j is:

[0022]

[0023] Where k is the number of variables selected for calculating PAI, and y i· Let represent the value of the i-th variable in all samples. This represents the weight of the i-th variable.

[0024] Preferably, in step (5), the variable weights This comes from a multiple linear regression analysis. In this analysis, if a clinical indicator has a U-shaped association with mortality risk, only one variable (y) is transformed. i or y i2 If a y-value is selected into the model by the LASSO algorithm, then y is retained. i and y i2 Used to calculate PAI.

[0025] Preferably, in step (6), the formula for calculating ΔPAI of sample j is:

[0026]

[0027] Where, x j This represents the actual age of the j-th sample. and These represent the slope and intercept terms calculated in the linear regression of PAI on actual age, respectively.

[0028] According to another aspect of the present invention, a biological age assessment system is provided, comprising:

[0029] Clinical indicator collection module: used to collect clinical indicators of subjects;

[0030] Nonlinear effect identification module: Based on nonlinear Cox proportional hazards regression analysis, it identifies clinical indicators that have a U-shaped association with mortality risk; the value of the clinical variable that is at the lowest relative mortality risk in the restricted cubic spline curve is defined as the clinical optimal value;

[0031] Nonlinear effect conversion module: For clinical indicators that have a U-shaped association with mortality risk, the original measurement value is converted based on its optimal value;

[0032] Variable selection module: Based on the candidate variables, a Cox proportional hazards regression model is constructed, and the LASSO algorithm is used for variable selection;

[0033] Biological age estimation module: Calculates biological age (PAI) based on predictors obtained from the nonlinear effect identification module and variable screening module;

[0034] Biological aging rate estimation module: Calculates the biological aging rate ΔPAI based on biological age (PAI) and actual age.

[0035] According to another aspect of the present invention, a biological aging assessment method is provided, which collects the following indicators from samples: chronological age (CA), in years; body mass index (BMI), in kg / m². 2 The following are definitions of blood pressure and blood volume: systolic blood pressure (SBP) in mmHg; fasting blood glucose (FBG) in mmol / L; total cholesterol (TC) in mmol / L; and red blood cell count (RBC) in 10-1. 12 / L; Mean corpuscular volume (MCV), in fL; Mean corpuscular protein (MCH), in pg; Coefficient of variation (RDW), in %; White blood cell count (WBC), in 10-1. 9 / L; Lymphocyte percentage (LYM), in %; Monocyte percentage (MONO), in %; Alkaline phosphatase (ALP), in μ / L; Aspartate aminotransferase (AST), in μ / L; Alanine aminotransferase (ALT), in μ / L; Direct bilirubin (DBIL), in μmol / L; Serum creatinine (SCR), in μmol / L; Blood urea nitrogen (BUN), in μmol / L; The PAI and ΔPAI of the samples are calculated according to the biological aging assessment system described above to assess the degree and rate of biological aging of the samples;

[0036] When ΔPAI > 0, it indicates that the biological aging rate exceeds that of peers; when ΔPAI = 0, it indicates that the biological aging rate is equal to that of peers; when ΔPAI < 0, it indicates that the biological aging rate is lower than that of peers.

[0037] According to another aspect of the invention, the method is provided for use in assessing the degree of aging and predicting the risk of death or disease.

[0038] In summary, compared with the prior art, the above-described technical solutions conceived by this invention mainly possess the following technical advantages:

[0039] (1) A method for assessing biological age applicable to the Chinese population is provided, which is used to assess the biological age and biological aging rate of the Chinese population sample. Compared with existing biological age estimators based on European populations or linear estimators using actual age as the prediction label, this method has higher prediction accuracy for the risk of death and age-related diseases, and is better able to identify individuals who are aging faster or slower.

[0040] (2) Compared with existing Chinese population estimation methods, this method is based on general population data with a larger sample size and longer follow-up time during the model training stage, making the biological age estimation method obtained by this method more robust.

[0041] (3) This invention considers the nonlinear relationship between clinical indicators and mortality risk, providing a biological age estimation method that simultaneously considers different risk effects when values ​​are below and above clinical optimum values. This improves the predictive power of biological age on the risk of death and new-onset age-related diseases. This invention first identifies clinical indicators that exhibit a U-shaped relationship with mortality risk, and then uses a novel method proposed in this invention to transform the original measurements of these clinical indicators. This transformation method enables this invention to model the different risk effects of clinical indicators on mortality when values ​​are above and below clinical recommendations, which is more consistent with the epidemiological characteristics of biological aging. Furthermore, this invention simultaneously provides estimation methods for biological age and biological aging rate, offering more comprehensive scientific advice and guidance for aging prevention and intervention.

[0042] (4) This invention estimates biological age and biological aging rate simultaneously based on clinical indicators. The measurement of clinical indicators can be completed with only routine physical examination and one blood sample. It has the advantages of being inexpensive and convenient. It can calculate biological age and biological aging rate in a timely manner after physical examination, and provide health guidance for individuals with accelerated aging. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the biological age assessment method provided in the embodiments of the present invention.

[0044] Figure 2 This is a schematic diagram of the biological age assessment system provided in an embodiment of the present invention.

[0045] Figure 3 This is an analysis flowchart provided by an embodiment of the present invention for constructing and validating a biological age index based on a population cohort.

[0046] Figure 4 This is a restrictive cubic spline curve between clinical indicators and mortality risk provided in the embodiments of the present invention.

[0047] Figure 5 It is the joint distribution of actual age and PAI in the training and test sets provided in the embodiments of the present invention.

[0048] Figure 6 This invention provides the predictive effects of PAI and ΔPAI on mortality risk in the "Dongfeng-Tongji" cohort.

[0049] Figure 7 This is the relative risk ratio of ΔPAI for predicting cardiovascular disease in the "Dongfeng-Tongji" cohort provided by the embodiments of the present invention.

[0050] Figure 8 This invention provides the predictive power of PAI and ΔPAI on mortality risk in a UK biobank sample.

[0051] Figure 9 This invention provides the predictive effect of ΔPAI on age-related diseases in a UK biobank sample. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0053] To achieve the above objectives, according to one aspect of the present invention, a method for constructing a biological age assessment system based on clinical indicators is provided, including a data preprocessing stage and a model determination stage, wherein the data preprocessing stage includes S1 and S2, and the model determination stage includes S3, S4, S5 and S6.

[0054] Step S1: Collect information on the age and gender of healthy subjects, and complete measurements of height, weight, blood pressure, blood lipids, blood glucose, complete blood count, and blood biochemistry. Follow up with the subjects for up to 10 years and collect information on the subjects' deaths and new cardiovascular events.

[0055] Preferably, the included clinical indicators (abbreviations, units) include the following: Body Mass Index (BMI, kg / m²) 2 Waist circumference (cm), systolic blood pressure (SBP, mmHg), diastolic blood pressure (DBP, mmHg), pulse pressure (PP, mmHg), fasting blood glucose (FBG, mmol / L), high-density lipoprotein cholesterol (HDL-C, mmol / L), low-density lipoprotein cholesterol (LDL-C, mmol / L), total cholesterol (TC, mmol / L), triglycerides (TG, mmol / L), red blood cell count (RBC, 10 12 Mean corpuscular volume (MCV, fL), red blood cell distribution width (RDW, %), mean corpuscular protein concentration (MCHC, g / L), mean corpuscular protein mass (MCH, pg), hemoglobin (HGB, g / L), hematocrit (HCT, %), white blood cell count (WBC, 10⁻⁶ / L), mean corpuscular hemoglobin (HGB, g / L), hematocrit (HCT, %), white blood cell count (WBC, 10⁻⁶ / L). 9 / L), monocyte percentage (MONO, %), lymphocyte percentage (LYM, %), neutrophil percentage (NEUT, %), basophil percentage (BASO, %), eosinophil percentage (EOS, %), platelet count (PLT, 10 9 The following parameters were measured: mean platelet volume (MPV, fL), platelet distribution width (PDW, fL), plateletcrit percentage (PCT, %), aspartate aminotransferase (AST, μ / L), alkaline phosphatase (ALP, μ / L), alanine aminotransferase (ALT, μ / L), total bilirubin (TBIL, μmol / L), direct bilirubin (DBIL, μmol / L), indirect bilirubin (IDBIL, μmol / L), blood urea nitrogen (BUN, mmol / L), creatinine (SCR, μmol / L), and uric acid (SUA, μmol / L).

[0056] Step S2: Perform quality control and preprocessing on the acquired raw data of sample baseline characteristics and clinical indicators. According to the technical solution of this invention, the data preprocessing flow is as follows:

[0057] A10 used 13,555 participants from the "Dongfeng-Tongji" cohort enrolled at the Central Hospital in 2008 as the training set and 26,990 participants from the "Dongfeng-Tongji" cohort enrolled at the Central Hospital, Huaguo Hospital, Maojian Hospital, and Xiyuan Hospital in 2013 as the test set. Samples and clinical indicators were filtered in both the training and validation sets. Ultimately, 12,769 participants were included in the training set, and 15,904 participants were included in the validation set, along with 36 clinical indicators.

[0058] A20 excludes samples with a baseline of cancer or a number of missing variables > 5.

[0059] A30, due to the significant skewed distribution of some clinical indicators, performed natural logarithmic transformations on triglycerides, fasting blood glucose, monocyte percentage, eosinophil percentage, basophil percentage, red blood cell distribution width, mean hemoglobin concentration, mean platelet volume, aspartate aminotransferase, alkaline phosphatase, alanine aminotransferase, total bilirubin, direct bilirubin, indirect bilirubin, blood urea nitrogen, and creatinine.

[0060] Preferably, samples with extreme outliers in any variable are excluded. Extreme outliers are defined as those that deviate from the median by eight interquartile ranges (IQR).

[0061] A40, based on the multiple imputation (PMM) method, imputes missing variables for each sample.

[0062] Step S3: In the training set, the restricted cubic spline method is used to identify the nonlinear association between clinical indicators and mortality risk. In the Cox proportional hazards model, the actual age is adjusted, and the 10%, 50%, and 90% quantiles are set as nodes for the clinical indicators.

[0063] Step S4: Based on the U-shaped correlation identified in Step S3, determine the optimal value of this clinical indicator. This is the lowest point of the RCS curve. Based on this optimal value... The raw values ​​of clinical indicators were transformed into two new variables:

[0064]

[0065]

[0066] Among them, y i and Let $i$ represent the raw value and the clinical optimum value of the $i$-th clinical indicator, respectively. For clinical indicators that have a U-shaped association with mortality risk, this transformation method can simultaneously convert values ​​above and below $i$. The risk effect is modeled; for clinical indicators that are linearly related to the risk of death, no transformation is made to the original values ​​of the clinical indicators.

[0067] Step S5: Using the Cox proportional hazards regression model and the LASSO method, select clinical indicators with predictive value for mortality risk. Based on the multiple linear regression model and the selected predictive factors, construct the physiological aging index (PAI).

[0068] Preferably, LASSO is used for variable selection in the Cox proportional hazards regression model, and the Cox proportional hazards regression model is expressed as:

[0069]

[0070] in, h0(t) and β represent the hazard rate function and baseline hazard function at time t, respectively. i Let y represent the regression effect value of the i-th indicator. i· Let represent the i-th variable in all samples, and p be the number of the indicators. The covariates included in the Cox proportional hazards regression model include actual age (CA), the original values ​​of clinical indicators linearly associated with mortality, and the transformed values ​​of clinical indicators non-linearly associated with mortality.

[0071] Preferably, k variables with predictive value for mortality risk are selected based on LASSO, and PAI is defined as the weighted sum of the k variables:

[0072]

[0073] Where k is the number of variables selected for calculating PAI, and y i· Let represent the value of the i-th variable in all samples. This represents the estimated regression effect of the i-th variable.

[0074] Preferably, variable weights This comes from a multiple linear regression analysis. In this analysis, if a clinical indicator has a U-shaped association with mortality risk, only one variable (y) is transformed. i or y i2 If a y-value is selected into the model by the LASSO algorithm, then y is retained. i and y i2 Used to calculate PAI.

[0075] Step S6: Based on the PAI calculation formula determined from the training set samples, calculate the PAI and ΔPAI for each sample in the test set. ΔPAI is defined as the residual value of the PAI regressed on actual age using CA. The formula for calculating the ΔPAI of the j-th sample is as follows:

[0076]

[0077] Where, x j This represents the actual age of the j-th sample. and Let represent the slope and intercept terms calculated in the linear regression of PAI on actual age, respectively. The ΔPAI of the test set samples is standardized to have a variance of 1.

[0078] Validating the predictive power of PAI on mortality risk using a Cox regression model:

[0079] h(t,d j )=h0(t)exp(γPAIj )

[0080] Where h(t,d) i ) and h0(t) represent the hazard rate function and baseline hazard function of sample j at time t, respectively, and d i PAI represents the survival state of sample j at time t. j This represents the biological age of sample j at baseline. γ indicates that for every 1 standard deviation increase in PAI, the relative mortality of the sample increases by a factor of “exp(γ)–1”.

[0081] Based on the model determination method provided in steps S3, S4, S5, and S6, an individual's biological age (PAI) and biological rate of aging (ΔPAI) can be quantified. ΔPAI > 0 indicates that the individual is aging rapidly, while ΔPAI < 0 indicates that the individual is aging slowly. This indicator can provide information independent of actual age for predicting the risk of death and new age-related diseases, and has broad application prospects and market potential.

[0082] According to another aspect of the present invention, a biological age assessment system based on the above method is provided, the system comprising the following modules:

[0083] Data collection module: Collects basic information from healthy participants, including gender, age, and baseline medical history. Participants undergo physical examinations and laboratory tests at the hospital after fasting for eight hours to collect clinical indicators required for biological age assessment.

[0084] Nonlinear effect identification module: Based on nonlinear Cox proportional hazards regression analysis, it identifies clinical indicators that have a U-shaped association with mortality risk; the value of the clinical variable that is at the lowest relative mortality risk in the restricted cubic spline curve is defined as the clinical optimal value;

[0085] Nonlinear Effects Transformation Module: For clinical indicators showing a U-shaped association with mortality risk, the module transforms their original measurements based on their optimal values. The variable transformation calculation formula is as follows:

[0086]

[0087]

[0088] Among them, y i and Let $i$ represent the raw value and the clinical optimum value of the $i$-th clinical indicator, respectively. For clinical indicators that have a U-shaped association with mortality risk, this transformation method can simultaneously convert values ​​above and below $i$. The risk effect is modeled; for clinical indicators that are linearly related to the risk of death, no transformation is made to the original values ​​of the clinical indicators.

[0089] Preferably, a U-shaped association exists between 25 clinical indicators and mortality risk, including: aspartate aminotransferase (AST), serum blood urea nitrogen (BUN), mean corpuscular volume (MCV), mean corpuscular protein (MCH), alanine aminotransferase (ALT), red blood cell count (RBC), serum creatinine (SCR), hematocrit (HCT), platelet count (PLT), high-density lipoprotein cholesterol (HDL-C), hemoglobin (HGB), direct bilirubin (DBIL), fasting blood glucose (FGB), plateletcrit percentage (PCT), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), lymphocyte percentage (LYM), neutrophil percentage (NEUT), alkaline phosphatase (ALP), serum uric acid (SUA), total bilirubin (TBIL), white blood cell count (WBC), body mass index (BMI), monocyte percentage (MONO), and waist circumference (Waist).

[0090] Variable selection module: Based on the candidate variables, a Cox proportional hazards regression model is constructed, and the LASSO algorithm is used for variable selection;

[0091] Preferably, according to the biological age assessment system provided by the present invention, 33 mortality risk-related variables are screened from 61 variables (11 original values ​​and 50 transformed values) of actual age (CA) and 36 clinical indicators using LASSO-based Cox proportional hazards regression. The effect size of each predictor is derived from multivariate Cox regression.

[0092] Biological age calculation module: Based on the predictive factors and their weights obtained from the nonlinear effect identification module and the variable screening module, calculate the biological age PAI and the biological aging rate ΔPAI.

[0093] Preferably, the formula for calculating PAI is:

[0094]

[0095] Wherein, each predictor variable y i· and their weights As shown in Table 1.

[0096] Table 1

[0097]

[0098] Biological aging rate estimation module: Calculates the biological aging rate ΔPAI based on biological age (PAI) and actual age.

[0099] Preferably, the formula for calculating ΔPAI is:

[0100] ΔPAI j =PAI j -0.12×x j+7.4.

[0101] Where, x j This represents the actual age of the j-th sample.

[0102] Example 1

[0103] Figure 1 This is a schematic diagram of the biological age assessment method provided by the present invention, see reference. Figure 1 This embodiment explains the biological age assessment method based on data from the "Dongfeng-Tongji" cohort. Figure 2 This is a schematic diagram of the biological age assessment system provided by the present invention. (See attached diagram) Figure 2 This embodiment further explains the biological age assessment system based on individual data from the UK Biobank. All analysis procedures for this embodiment are described in [link to documentation]. Figure 3 .

[0104] Operation S1 collects data and divides samples from the "Dongfeng-Tongji" queue.

[0105] In the large prospective cohort established at the applicant's institution—the "Dongfeng-Tongji" cohort—all study participants underwent questionnaire surveys, physical examinations, and laboratory tests. The questionnaires, administered by qualified, uniformly trained investigators using a semi-structured format, collected basic information from participants, including gender, age, date of birth, individual health behaviors, medical history, and medication history. Physical examinations were conducted by clinicians, measuring participants' height, weight, blood pressure, electrocardiogram, waist circumference, and hip circumference. Laboratory tests required participants to fast for at least 8 hours before drawing blood from their elbow vein for complete blood count, blood biochemistry, blood lipids, and blood glucose levels. The follow-up team conducted final follow-ups with participants using electronic medical records from the Dongfeng Motor Corporation health insurance network and categorized mortality events according to the International Classification of Diseases, 10th Revision (ICD-10). The follow-up cutoff date for mortality events was defined as May 20, 2019. Accidental deaths (ICD-10 code: V01-Y98) were not included as outcomes in this study. Based on the participants' enrollment time, the samples were divided into a training set and a validation set. The training set contained 12,769 samples, and the validation set contained 15,904 samples.

[0106] Operation S2 performs sample and variable quality control on the training and validation sets. This step corresponds to... Figure 1Step 1 of the biological age assessment method proposed in this invention: In the training set sample: ① Exclude samples with cancer at enrollment in 2008; ② Exclude samples with missing values ​​greater than 5 and those lost to follow-up; ③ Perform natural logarithmic transformation on right-skewed variables; ④ Exclude outliers based on a deviation of 8 times the interquartile range from the mean; ⑤ Use the multiple imputation (PMM) algorithm to impute missing values. In the test set sample: ① Exclude samples with cancer at enrollment in 2013; ② Exclude samples with missing values ​​greater than 5 and those lost to follow-up; ③ Perform natural logarithmic transformation on right-skewed variables; ④ Exclude outliers based on a deviation of 8 times the interquartile range from the mean; ⑤ Use the multiple imputation (PMM) algorithm to impute missing values; ⑥ Exclude samples that conflict with the training set.

[0107] Operation S3, based on the Restricted Cubic Spline (RCS) method, identifies clinical indicators that exhibit a U-shaped association with mortality risk; in this Cox proportional hazards regression model, chronological age is adjusted, and the 10%, 50%, and 90% quantiles are set as nodes for the clinical indicators. This step corresponds to... Figure 1 Step 2 of the biological age assessment method proposed in this invention.

[0108] In step S4, for the clinical indicators identified in step S3 that show a U-shaped association with mortality risk, variable transformation is performed based on the clinically optimal values ​​determined by the RCS curve. This step corresponds to... Figure 1 Step 3 of the biological age assessment method proposed in this invention. The formula for variable transformation based on the clinical optimum is expressed as:

[0109]

[0110]

[0111] Among them, y i and Let $i$ represent the raw value and the clinical optimum value of the $i$-th clinical indicator, respectively. For clinical indicators that have a U-shaped association with mortality risk, this transformation method can simultaneously convert values ​​above and below $i$. The risk effect is modeled; for clinical indicators that are linearly related to the risk of death, no transformation is made to the original values ​​of the clinical indicators.

[0112] Furthermore, this step revealed a linear association between 11 clinical indicators and mortality risk, and a U-shaped association between 25 clinical indicators and mortality risk. Therefore, a total of 11 original clinical indicators and 50 transformed variables were identified. The RCS curves for the 11 clinical indicators that showed a U-shaped association with mortality risk but no significant linear effect are shown below. Figure 4 As shown.

[0113] Operation S5, this step corresponds to Figure 1Step 4 of the biological age assessment method proposed in this invention. In the training set, 33 predictors related to mortality risk were screened from 61 variables generated in step S4 and chronological age using LASSO-based Cox regression. These included the original variables of chronological age (CA), systolic blood pressure (SBP), and red blood cell distribution width (RDW), as well as 30 transformed variables from 15 clinical indicators.

[0114] Furthermore, multivariate Cox regression was used to estimate the weight of each predictor, with the weight in units of the clinical indicator. The multivariate Cox regression included raw measurements of actual age (CA), systolic blood pressure (SBP), and red blood cell distribution width (RDW), as well as the transformed values ​​(y) of 15 clinical indicators. i1 and y i2 This weight is used for PAI, and its formula is expressed as:

[0115]

[0116] The mortality predictors and their estimated weights selected according to this embodiment are shown in Table 1.

[0117] Operation S6 calculates PAI and ΔPAI on the test set samples and verifies their predictive ability for mortality and new disease risk. This step corresponds to... Figure 1 Steps 5 and 6 of the biological age assessment method proposed in this invention are described below. The specific analysis process is as follows:

[0118] Analysis of A1: In the test set sample, the clinical indicators that showed a U-shaped association with mortality risk were transformed using the parameters pre-provided by the biological age assessment system provided by the present invention (Table 1).

[0119] Analysis A2, based on the predefined formula of the biological age assessment system provided by this invention, calculates the PAI and ΔPAI of the sample. The relationship between PAI and ΔPAI is as follows: Figure 5 As shown.

[0120] In analysis of A3, a Cox proportional hazards regression model was used to evaluate the discrimination (C-index and time-dependent AUC index), calibration (slope of the calibration curve and outcome), and clinical decision value of PAI in predicting mortality risk in the test set. The results are shown in Table 2. Compared with chronological age (CA) and the established biological age (PhenoAge), PAI showed superior discrimination and calibration in predicting mortality risk.

[0121] Table 2

[0122]

[0123]

[0124] Furthermore, this study assesses whether the C-index and time-dependent AUC index can improve the predictive power of chronological age and a single clinical biomarker on mortality risk. Three Cox regression models are compared: ① including chronological age as a covariate; ② including chronological age and a single clinical biomarker as covariates; ③ including chronological age and ΔPAI as covariates. Results are as follows... Figure 6 As shown, ΔPAI can improve the predictive power of actual age and sex on mortality risk, and this predictive performance is superior to any single clinical indicator.

[0125] In analysis of A4, Cox proportional hazards regression analysis was used in the test set to evaluate the predictive effect of ΔPAI on new-onset cardiovascular disease. Samples with abnormal electrocardiograms and existing cardiovascular disease were excluded first. ΔPAI was transformed to a standard normal distribution, and PAI was analyzed as both a continuous and categorical variable. As a categorical variable, the samples were divided into four groups using PAI quartiles. Multivariate Cox regression was performed with cardiovascular disease, coronary artery disease, stroke, and acute coronary syndrome as outcomes, comparing the risk of cardiovascular disease and its subtypes in the three groups with higher PAI compared to the group with the lowest PAI. This analysis compared two models: the baseline model included only ΔPAI and chronological age as covariates, while the full model additionally included 11 CVD risk factors as covariates, including gender, hypertension, hyperlipidemia, diabetes, obesity, smoking, alcohol consumption, diet, sleep duration, exercise frequency, and family history of cardiovascular disease.

[0126] The results are as follows Figure 7 As shown, when ΔPAI is used as a continuous variable to predict risk factors for new-onset cardiovascular disease and its subtypes, the relative hazard ratios (HR per SD) for ΔPAI in predicting new-onset cardiovascular disease, coronary artery disease, acute coronary syndrome, and stroke were 1.18 (95% confidence interval: 1.13, 1.24), 1.13 (1.07, 1.18), 1.22 (1.12, 1.34), and 1.40 (1.28, 1.52), respectively. When the samples were grouped by quartiles of PAI, the risk of cardiovascular disease was significantly higher in the highest quartile of ΔPAI than in the lowest quartile, especially for stroke (HR = 2.35 [95% confidence interval: 1.76, 3.13]). When we further adjusted for 11 traditional risk factors, the effect size of ΔPAI decreased, but it remained statistically significant in CVD, ACS, and stroke.

[0127] Furthermore, the biological age assessment system provided by this invention is used to perform the second aspect of this embodiment. In this embodiment, PAI is estimated in a UK biobank sample and its predictive value for mortality is assessed, performing operations S7 to S9. Wherein, S7 corresponds to... Figure 2The provided biological age assessment system includes a data collection module, a nonlinear effect identification module, and a biological age assessment module. Operations S8 and S9 describe the application of the aging rate assessment system.

[0128] To ensure data quality for the UKB sample, in step S7, samples lost to follow-up, those without blood samples collected on the enrollment day, those with missing death times, those with baseline cancer, or those lacking clinical indicators were first excluded. Next, the right-skewed variables were transformed using the natural logarithm. Following the previously described method, the clinical indicators showing a U-shaped association with mortality risk were transformed using the clinically optimal values ​​estimated from the training set of the "Dongfeng-Tongji" cohort (see Table 1). Finally, the sex-specific PAI for 298,284 samples was calculated using model weights (see Table 1).

[0129] Operation S8 was performed, using a Cox proportional hazards regression model to evaluate the discrimination (C-index and time-dependent AUC index), calibration (slope of the calibration curve and outcome), and clinical decision value of PAI in predicting mortality risk. The results are shown in Table 3. Compared to chronological age (CA) and the established biological age (PhenoAge), PAI demonstrated superior discrimination and calibration in predicting mortality risk.

[0130] Table 3

[0131]

[0132]

[0133] Furthermore, this study assesses whether the C-index and time-dependent AUC index can improve the predictive power of chronological age and a single clinical biomarker on mortality risk. Three Cox regression models are compared: ① including chronological age as a covariate; ② including chronological age and a single clinical biomarker as covariates; ③ including chronological age and ΔPAI as covariates. Results are as follows... Figure 8 As shown, ΔPAI can improve the predictive power of actual age and sex on mortality risk, and this predictive performance is superior to any single clinical indicator.

[0134] Operation S9 assessed the predictive value of ΔPAI for age-related diseases using the UK Biobank. The predictive value of ΔPAI for the incidence of nine age-related diseases was evaluated in the UK Biobank, including cancer, ischemic heart disease, stroke, chronic obstructive pulmonary disease, diabetes, dementia, chronic kidney disease, hearing loss, and chronic liver disease. For each specific disease, cases diagnosed prior to enrollment were excluded. Subsequently, a sex-stratified Cox model was used, adjusted for age, center of gravity, smoking, alcohol consumption, and Thomson Scale index. Results are as follows: Figure 9The study indicated that ΔPAI was significantly associated with all nine diseases, with the strongest associations with chronic kidney disease (relative hazard ratio HR = 1.59 [95% confidence interval: 1.58, 1.61]) and chronic liver disease (relative hazard ratio HR = 1.48 [95% confidence interval: 1.45, 1.50]). Notably, both ΔPAI and ΔPhenoAge, as comprehensive aging indices, outperformed any single biomarker in predicting the nine diseases excluding diabetes. Compared to ΔPhenoAge, ΔPAI performed better in predicting eight diseases excluding cancer, showing a significant improvement in predicting stroke and dementia.

[0135] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method of constructing a biological age assessment system, characterized by, The method comprises the following steps: (1) collecting clinical indicators of subjects; (2) performing non-linear Cox proportional hazards regression analysis on the death risk of the clinical indicators in step (1); (3) for the clinical indicators obtained in step (2) that are not non-linearly associated with the death risk, the original values are not converted; for the clinical indicators obtained in step (2) that are U-shapedly associated with the death risk, the original measurement values thereof are converted based on the optimal values thereof; the optimal value is defined as the value of the clinical variable when the relative death risk is the lowest in the restricted cubic spline curve; The formula for converting the clinical indicators into variables is: wherein, and denote the raw value and the optimal value of the i th clinical indicator, respectively; (4) predicting the death risk based on the actual age, the original clinical indicators obtained in step (3) and the converted variables, and performing variable screening in a Cox proportional hazards regression model by using a LASSO algorithm; (5) calculating the biological age PAI of each sample based on the variables screened in step (4); Sample j The PAI is calculated by the formula: in, k The number of variables selected for calculating PAI. Represents the first of all samples i The possible values ​​of each variable Indicates the first i Weights of each variable; (6) calculating the biological aging speed ΔPAI of each sample based on the biological age PAI calculated in step (5) and the actual age of the sample.

2. The method of constructing a biological age assessment system according to claim 1, wherein, In step (1), the clinical indicators are height, weight, blood pressure, blood lipids, blood glucose, blood routine indicators and blood biochemical indicators, and the skew distribution indicators among the clinical indicators are normally converted.

3. The method of constructing a biological age assessment system according to claim 1, wherein, In step (2), the model used in the non-linear Cox proportional hazards regression analysis adopts a restricted cubic spline method, and the actual age and a single clinical indicator are used as the covariates, and the node selection of the clinical indicators is 10%, 50% and 90% quantiles.

4. The method of constructing a biological age assessment system according to claim 1, wherein, In step (4), the Cox proportional hazards regression model is expressed as: wherein, and respectively represent t the hazard rate function at time point t and the baseline hazard function, represents the regression effect value of the jth i indicator, represents the jth i variable of all samples, p is the number of clinical indicators; the covariates included in the Cox proportional hazards regression model include the actual age, the variable original value of the clinical indicators linearly associated with death, and the transformed value based on the clinical indicators nonlinearly associated with death.

5. The method of constructing a biological age assessment system according to claim 4, wherein, Variable weights from the multivariate linear regression analysis; in the multivariate linear regression analysis, if a clinical indicator with a U-shaped association with mortality risk, only one variable transformation value was selected into the model by the LASSO algorithm, i.e. or was selected into the model by the LASSO algorithm, then and were simultaneously retained for the calculation of PAI.

6. The method of constructing a biological age assessment system according to claim 1, wherein, In step (6), the sample j ΔPAI is calculated according to the following formula: wherein, represents the actual age of the j th sample, and represents the calculated slope and intercept terms in the linear regression of the PAI on the actual age, respectively.

7. A biological age assessment system, characterized by, The method comprises: a clinical indicator collection module for collecting clinical indicators of subjects; a non-linear effect identification module for identifying clinical indicators that are U-shapedly associated with the death risk based on non-linear Cox proportional hazards regression analysis, and defining the value of the clinical variable when the relative death risk is the lowest in the restricted cubic spline curve as the clinical optimal value; a non-linear effect conversion module for converting the original measurement values of the clinical indicators that are U-shapedly associated with the death risk based on the optimal values thereof; the formula for converting the clinical indicators into variables is: wherein, and denote the raw value and the optimal value of the i th clinical indicator, respectively; a variable screening module for constructing a Cox proportional hazards regression model based on candidate variables, and performing variable screening by using a LASSO algorithm; a biological age estimation module for calculating the biological age PAI based on the prediction factors obtained by the non-linear effect identification module and the variable screening module; Sample j The PAI is calculated by the formula: in, k The number of variables selected for calculating PAI. Represents the first of all samples i The possible values ​​of each variable Indicates the first i Weights of each variable; a biological aging speed estimation module for calculating the biological aging speed ΔPAI according to the biological age PAI and the actual age.

8. A method for assessing biological aging, characterized in that, The following indicators were collected for the sample: actual age CA in years; body mass index BMI in kg / m 2 , defined as weight in kg divided by height in m squared. Systolic blood pressure, SBP, in mmHg; fasting blood glucose, FBG, in mmol / L; total cholesterol, TC, in mmol / L; red blood cell count, RBC, in 10 12 / L; mean corpuscular volume, MCV, in fL; mean corpuscular hemoglobin, MCH, in pg; red cell distribution width, RDW, in %; white blood cell count, WBC, in 10 9 / L; lymphocyte percentage, LYMP, in %; The percentage of monocytes MONO is in %; alkaline phosphatase ALP is in μ / L; glutamic-oxaloacetic transaminase AST is in μ / L, glutamic-pyruvic transaminase ALT is in μ / L, direct bilirubin DBIL is in μmol / L, serum creatinine SCR is in μmol / L, urea nitrogen BUN is in μmol / L, and the skew distribution indicators are naturally logarithmically converted; The biological age assessment system according to claim 7 calculates the PAI and the ΔPAI of the sample, thereby assessing the biological age and the biological aging speed of the sample; When the ΔPAI > 0, it indicates that the biological aging speed is faster than that of the same age group; when the ΔPAI = 0, it indicates that the biological aging speed is equal to that of the same age group; when the ΔPAI < 0, it indicates that the biological aging speed is slower than that of the same age group.

9. Use of the method according to claim 8 in assessing the aging degree, predicting the risk of death or disease.

Images