Difference in risk profiles for pulmonary hypertension between non-elderly and elderly patients undergoing maintenance hemodialysis Authors: Zhen Ai, Qingling Jia, Suiqin Wen, Ying Zhu, DiHua Zhang, Raoping Wang, Jianhua Wu, Yuanwen Xu Published: 2026 ABSTRACT AbstractBackgroundPulmonary hypertension (PH) is highly prevalent and associated with increased mortality in patients undergoing maintenance hemodialysis (MHD). This study aimed to explore the disparities in the prevalence, risk factors, and prognostic impacts of PH between non-elderly and elderly MHD patients.MethodsThis study included 179 MHD patients with complete clinical records. All patients were evaluated using Doppler echocardiography, and PH was defined as a pulmonary artery systolic pressure > 35 mmHg.ResultsThe prevalence rates of PH were 24.8% in non-elderly patients and 33.8% in elderly patients. Serum N-terminal pro-brain natriuretic peptide (NT-proBNP) was a predominant risk factor for PH in hemodialysis patients. A lower ratio of uric acid to high-density lipoprotein (UHR) was associated with PH among non-elderly patients, whereas diabetes mellitus served as a specific risk factor for elderly patients. The areas under the receiver operating characteristic curves of these risk factors identified in non-elderly and elderly patients were 0.86 (bootstrap 95% confidence interval (CI) 0.78–0.93) and 0.90 (bootstrap 95% CI 0.82–0.97), respectively. During a median follow-up duration of 32.90 (4.00-61.50) months, the presence of PH notably elevated the risk of all-cause mortality and cardiovascular hospitalization in both non-elderly and elderly patients. Meanwhile, it significantly augmented the risk of all-cause hospitalization (HR: 2.24, 95% CI 1.26–3.98, P = 0.006) in the non-elderly.ConclusionsDistinct risk profiles for PH were identified between non-elderly and elderly MHD patients, offering valuable clinical data for the development of early detection and prevention strategies based on age stratification.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12882-026-05020-x.Keywords: Pulmonary hypertension, Non-elderly, Elderly, Hemodialysis, Risk profiles Introduction Pulmonary hypertension (PH), characterized by abnormally elevated pulmonary arterial pressure, predominantly affects the small arteries of the pulmonary vasculature [1]. If left untreated, it can progress to severe stages, leading to increased pulmonary vascular resistance, right ventricular failure, and, ultimately, death. The prevalence of PH gradually increases as renal function deteriorates. It affects 14–41% of patients with chronic kidney disease (CKD) at stages 1–5, and up to 60% of patients with end-stage renal disease (ESRD) who are receiving hemodialysis [2–5]. Despite its high prevalence and significant association with increased mortality in patients with kidney failure [6–8], PH remains an under-recognized disorder frequently neglected in clinical practice. Multiple pathophysiological mechanisms, including endothelial dysfunction, oxidative stress, activation of pro-inflammatory factors, anemia, fluid overload, mineral metabolism disorders, and increased cardiac output due to arteriovenous fistula, have been shown to significantly affect the development of PH in patients with CKD [4, 9–11]. The intricate pathophysiological interaction between the kidney and cardiovascular system, variable and dynamic volume state, and highly comorbid patient population pose great challenges to the comprehension, diagnosis, and management of PH in CKD, especially in patients with ESRD undergoing maintenance hemodialysis (MHD). With the global aging population, PH has emerged as a worldwide health concern [12, 13]. Recent studies have revealed that the mean age of patients with PH has increased over the years [14, 15]. Old age has been identified as a potential risk factor for the development of PH in patients with CKD and as an independent risk factor for mortality in patients with PH [12, 16]. Despite the growing prevalence of PH in the elderly population and suboptimal treatment responses observed in these patients, there is a scarcity of research exploring the characteristics of elderly patients with PH. Given that PH is a common comorbidity among patients undergoing maintenance hemodialysis, it remains uncertain whether the pathogenesis and prognosis of PH in elderly patients differ from those in non-elderly patients. To address these issues, this study investigated the prevalence, risk factors, all-cause mortality, and cardiovascular hospitalizations associated with PH in non-elderly and elderly patients undergoing hemodialysis. Materials and methods Study design and patients This retrospective, single-center cohort study was conducted at The First Affiliated Hospital of Sun Yat-sen University. Patients who underwent echocardiography between October 2019 and November 2020 were recruited by searching the electronic medical records system. The inclusion criteria were as follows: (1) age ≥ 18 years; (2) diagnosis of ESRD who underwent hemodialysis thrice weekly and had a maintenance hemodialysis duration of at least 3 months; (3) complete clinical data on laboratory tests and echocardiography results. The exclusion criteria were as follows: (1) history of congenital heart disease, rheumatic heart disease, valvular heart disease, coronary heart disease, chronic obstructive pulmonary disease, pulmonary embolism, diffuse connective tissue disease, or malignancy; (2) those who underwent combined peritoneal dialysis; and (3) follow-up time of less than 3 months. All patients provided their informed consent. This study was conducted in accordance with the Declaration of Helsinki and approved by the Clinical Research and Laboratory Animal Ethics Committee of The First Affiliated Hospital of Sun Yat-sen University (approval number: [2024]847). Clinical and echocardiographic data collection Patient data regarding demographic characteristics, comorbid conditions, and laboratory parameters including serum high-sensitivity C-reactive protein (hs-CRP), white blood cell count (WBC), hemoglobin (Hb), platelet count (PLT), serum intact parathyroid hormone (iPTH), serum creatinine (Scr), uric acid (UA), total cholesterol (TC), serum triglyceride (TG), high-density lipoprotein-cholesterol (HDL-c), low-density lipoprotein cholesterol (LDL-c), serum albumin, prealbumin, serum globulin, and N-terminal pro-brain natriuretic peptide (NT-proBNP), were retrieved from the patients’ medical records in the hospital. The neutrophil-to-lymphocyte ratio (NLR) was calculated using absolute neutrophil and lymphocyte counts, and the ratio of uric acid to high-density lipoprotein (UHR) was calculated by the following formula: UHR = serum UA (mg/dL) / HDL-c (mg/dL) ×100. All the blood specimens were collected prior to dialysis after the long interdialytic interval, following a period of fasting (> 12 h), up to two weeks before the ultrasonography investigation. All transthoracic Doppler echocardiograms were performed by skilled technicians at the hospital on an interdialytic day in accordance with the recommendations of the American Society of Echocardiography. The following data were subsequently collected: left ventricular end-diastolic diameter (LVEDD); interventricular septum thickness (IVST); left ventricular posterior wall diameter (LVPWD); tricuspid annular plane systolic excursion (TAPSE); left ventricular ejection fraction (LVEF), which reflects cardiac systolic function; the ratio of early mitral inflow velocity to mitral annular early diastolic velocity (E/e’), which reflects cardiac diastolic dysfunction when E/e’ > 14 [17]; and pulmonary arterial systolic pressure (PASP), which is calculated according to Bernoulli’s formula after the measurement of tricuspid regurgitation velocity. Pulmonary hypertension is defined as a value of PASP > 35 mmHg based on echocardiographic criteria [3, 16, 18]. Body composition analysis Bioelectrical impedance analysis (BIA) was conducted before and after the HD sessions following the long interdialytic interval, up to four-weeks before the ultrasonography investigation. This test was conducted using a multi-frequency bioelectrical impedance analyzer (InBody S10). Multiple measures including extracellular water (ECW), intracellular water (ICW), total body water (TBW), fat-free mass (FFM), ECW/TBW, ECW/ICW, ICW/TBW, and TBW/FFM were taken or calculated. Follow-up and outcomes During a median follow-up period of 32.90 (4.00-61.50) months, cardiovascular (CV) hospitalizations, all-cause hospitalizations, and all-cause mortality were recorded. CV hospitalizations resulting from new-onset CV events, including ischemic or hemorrhagic cerebrovascular accidents, acute coronary syndrome, acute heart failure, and aortic dissection, were identified based on the discharge diagnoses. Statistical analyses Values were presented as mean ± standard deviation (SD), median (interquartile range, IQR), or absolute numbers (n) and percentages (%), as appropriate. Qualitative variables were compared using the chi-square test, while quantitative variables were compared using the Student’s t-test or Mann-Whitney U test. Univariate and multivariate analyses were performed using logistic regression to assess the factors associated with PH. Receiver operating characteristic (ROC) curves were created, and the apparent area under the curve (AUC) was calculated to assess the discriminative ability of the variables in predicting events. Internal validation was performed through bootstrap resampling with 1000 iterations to estimate model stability and mitigate overfitting. The prognostic value of PH in non-elderly and elderly MHD patients for predicting study outcomes was evaluated using Kaplan-Meier survival analysis with a log-rank test and Cox regression analysis. Cox proportional hazards regression models were used to estimate hazard ratios (HR) and their corresponding 95% confidence intervals (CIs) for relevant outcomes. Fine-Gray competing risk regression analysis was conducted to evaluate the cumulative incidence for cardiovascular hospitalization in the presence of competing risks (all-cause mortality) in the elderly patients. Results were reported as subdistribution hazard ratios (SHR) with 95% CI. Statistical analyses were conducted using the SPSS statistical software (v.26.0; IBM Corp) and R software (version 4.4.2). All tests were 2 sided, and P < 0.05 was considered statistically significant. Graphs were plotted using GraphPad Prism (version 5.01; GraphPad Software, San Diego, CA, USA). Results Baseline characteristics in distinct age-stratified populations This study included 179 patients, comprising 105 non-elderly (< 60 years) and 74 elderly (≥ 60 years) patients. The baseline characteristics of the patients were shown in Table 1. The median age of the patients was 55.00 (45.00, 67.00) years, and the median HD vintage was 74.90 (37.97, 138.00) months. The prevalence of PH (PASP > 35 mmHg) in non-elderly and elderly patients was 24.8% (26/105) and 33.8% (25/74), respectively. Compared with non-elderly patients, elderly patients had a lower arteriovenous fistulas usage rate, lower serum albumin, pre-albumin, serum creatinine, uric acid, iPTH, and UHR levels, higher diabetes mellitus (DM), LVEF, and average E/e’ ratio (P < 0.05). The sex distribution, HD vintage, hs-CRP, hemoglobin, platelets, serum globulin, serum electrolytes, serum lipids, NT-proBNP, NLR, urea clearance ratio (URR), and LVEDD were similar between the non-elderly and elderly patients. Table 1Baseline characteristics of patients stratified according to age groupVariablesTotal (n = 179)Age < 60 years (n = 105)Age ≥ 60 years (n = 74)P value Demographic Male, n (%)102 (57.0)65 (61.9)37 (50.0)0.11Age, years55.00 (45.00, 67.00)47.00 (42.50, 53.00)69.50 (64.00, 74.25)< 0.001AV fistula, n (%)171 (95.5)104 (99.0)67 (90.5)0.007Diabetes mellitus, yes33 (18.4)12 (11.4)21 (28.4)0.004HD vintage, months74.90 (37.97, 138.00)82.63 (45.92, 139.50)67.38 (35.75, 126.80)0.21 Laboratory Data Hs-CRP, mg/L2.25 (0.95, 5.48)1.99 (0.86, 5.56)2.61 (1.17, 5.58)0.24Hemoglobin, g/L110.94 ± 16.94110.25 ± 17.91111.89 ± 15.550.56Neutrophils, ×109/L4.10 (3.22, 5.20)3.87 (3.20, 5.10)4.25 (3.20, 5.43)0.28Lymphocyte, ×109/L1.21 (0.94, 1.57)1.25 (0.93, 1.61)1.18 (0.95, 1.49)0.66Platelet, ×103/ul182.00 (147.50, 218.50)182.00 (145.00, 223.00)182.50 (151.50, 218.25)0.64Albumin, g/L37.85 (35.90, 39.50)38.50 (36.68, 40.10)36.90 (34.65, 38.55)< 0.001Globulin, g/L27.60 (25.53, 30.98)27.10 (25.40, 29.93)28.60 (25.58, 32.00)0.05Pre-albumin, mg/dL30.00 (25.90, 34.80)31.20 (27.23, 35.83)28.50 (24.53, 33.20)0.001Calcium, mmol/L2.30 (2.20, 2.50)2.30 (2.20, 2.50)2.30 (2.10, 2.50)0.87Phosphorus, mmol/L1.79 (1.43, 2.16)1.77 (1.47, 2.12)1.80 (1.32, 2.19)0.38Intact PTH, pg/mL278.10 (150.40, 550.05)298.10 (159.90, 617.70)250.55 (130.10, 410.45)0.03Total cholesterol, mg/dL158.55 (15.35, 185.62)154.68 (135.35, 177.88)158.55 (131.48, 193.35)0.60TG, mg/dL115.99 (8.81, 170.89)115.99 (76.59, 168.24)117.76 (80.35, 184.84)0.66HDL-c, mg/dL42.15 (35.58, 49.50)42.15 (35.77, 48.53)40.80 (34.90, 51.91)0.90LDL-c, mg/dL95.51 (79.27, 114.46)93.97 (79.27, 113.50)97.84 (79.18, 122.49)0.58Serum creatinine, mg/dL11.92 ± 3.0313.13 ± 2.7110.20 ± 2.63< 0.001Uric acid, mg/dL8.22 (7.24, 9.11)8.53 (7.74, 9.42)7.70 (6.65, 8.67)< 0.001NT-proBNP, ×103 pg/mL4.19 (1.81, 13.62)4.19 (1.70, 11.91)4.17 (1.95, 14.97)0.40NLR3.42 (2.54, 4.64)3.30 (2.65, 4.60)3.61 (2.51, 4.86)0.40UHR19.32 (15.44, 24.13)20.25 (16.68, 24.43)18.37 (13.64, 23.74)0.03URR0.71 (0.68, 0.79)0.72 (0.67, 0.76)0.71 (0.68, 0.75)0.97 Echocardiographic LVEDD, mm50.00 (46.00, 54.00)50.00 (47.00, 54.00)49.50 (45.00, 54.00)0.31LVEF, %67.00 (62.00, 73.00)67.00 (61.00, 71.50)69.00 (64.00, 74.25)0.009Average E/e’11.45 (8.61, 15.31)10.16 (7.73, 12.45)14.44 (9.20, 17.71)< 0.001PASP > 35mmHg, n (%)51 (28.5)26 (24.8)25 (33.8)0.19Note: Values were shown as n (%) or median (interquartile range)Abbreviations: AV fistula, arteriovenous fistula; HD, hemodialysis; Hs-CRP, high sensitivity C-reactive protein; PTH, parathyroid hormone; TG, triglyceride; HDL-c, high-density lipoprotein-cholesterol; LDL-c, low-density lipoprotein cholesterol; NT-proBNP, N-terminal pro-brain natriuretic peptide; NLR, neutrophil-to lymphocyte ratio; UHR, uric acid to HDL-c ratio; URR, urea clearance ratio; LVEF, left ventricular ejection fraction; E/e’, ratio between early mitral inflow velocity and mitral annular early diastolic velocity; PASP, pulmonary arterial systolic pressure Baseline characteristics of patients stratified according to age group Note: Values were shown as n (%) or median (interquartile range) Abbreviations: AV fistula, arteriovenous fistula; HD, hemodialysis; Hs-CRP, high sensitivity C-reactive protein; PTH, parathyroid hormone; TG, triglyceride; HDL-c, high-density lipoprotein-cholesterol; LDL-c, low-density lipoprotein cholesterol; NT-proBNP, N-terminal pro-brain natriuretic peptide; NLR, neutrophil-to lymphocyte ratio; UHR, uric acid to HDL-c ratio; URR, urea clearance ratio; LVEF, left ventricular ejection fraction; E/e’, ratio between early mitral inflow velocity and mitral annular early diastolic velocity; PASP, pulmonary arterial systolic pressure A comparison of the inflammatory, nutritional, and metabolic parameters between MHD patients with and without PH in different age groups was shown in Fig. 1. In both the non-elderly and elderly groups, patients with PH presented significantly lower Scr levels and higher serum NT-proBNP levels (P < 0.05). However, among non-elderly MHD patients, those with PH exhibited a significantly lower UHR (16.97 ± 4.84 vs. 23.33 ± 7.98, P < 0.001), a phenomenon that was not observed in the elderly group. Comparisons between hemodialysis patients with and without pulmonary hypertension (PH), stratified by age. a hemoglobin, b serum creatinine, c serum globulin, d serum prealbumin, e serum N-terminal pro-brain natriuretic peptide (NT-proBNP), and f uric acid to HDL-c ratio Conversely, in elderly MHD patients, those with PH demonstrated significantly lower hemoglobin levels (104.24 ± 16.69 g/L vs. 115.80 ± 13.51 g/L, P = 0.001) and higher serum globulin levels (31.23 ± 4.61 g/L vs. 27.92 ± 3.62 g/L, P = 0.007). Among the 102 enrolled patients who underwent bioelectrical impedance analysis, notably elevated ECW/TBW ratios and ECW/ICW ratios were observed both pre- and post-dialysis in patients with PH (Supplementary Tables 1, 2). Risk factors of PH stratified by age group In non-elderly patients undergoing hemodialysis, PH was negatively correlated with pre-albumin, triglyceride, Scr, and UHR, and was positively correlated with HDL-c and NT-proBNP (Supplementary Table 3). Univariate and multivariate logistic regression analyses were conducted to determine the demographic and laboratory risk factors associated with PH (Table 2). Univariate logistic regression analysis indicated that age, Scr, serum globulin, serum prealbumin, Ln NT-proBNP, and UHR were predictors of PH. Multivariate regression models further confirmed that two variables—Ln NT-proBNP (OR: 2.46, P < 0.001) and UHR (OR: 0.87, P = 0.01)—remained significant risk factors. ROC curves were drawn for NT-proBNP and UHR to predict the prevalence of PH (Fig. 2a). The AUC and cut-off value were as follows: 0.80 (bootstrap 95% CI: 0.69–0.89, P < 0.001) and 9742.00 pg/mL (sensitivity 65.4%, specificity 84.8%) for NT-proBNP; 0.78 (bootstrap 95% CI: 0.67–0.88, P < 0.001) and 18.47 (sensitivity 80.0%, specificity 77.6%) for UHR. The combined model incorporating Ln NT-proBNP and UHR outperformed the individual prognostic factors in predicting PH (AUC: 0.86, bootstrap 95% CI 0.78–0.93, P < 0.001). Table 2Logistic regression analysis of the risk factors associated with pulmonary hypertension in non-elderly patients undergoing maintenance hemodialysis (n = 105, 26 cases with pulmonary hypertension)VariablesUnivariateMultivariate aOR (95% CI)P valueOR (95% CI)P valueAge, years1.07 (1.00, 1.14)0.05//Serum creatinine, mg/dL1.00 (1.00, 1.00)0.03//Globulin, g/L1.13 (1.01, 1.26)0.03//Pre-albumin, mg/dL0.87 (0.80, 0.95)0.003//Ln NT-proBNP2.65 (1.70, 4.13)< 0.0012.46 (1.52, 3.98)< 0.001UHR0.83 (0.74, 0.92)0.0010.87 (0.77, 0.97)0.01URR, %0.94 (0.87, 1.02)0.13//Note: a The multivariate model was restricted to clinically essential predictors with an events per variable ratio of 13. Negelkerke R Square: 0.43; P value of Hosmer-Lemeshow test: 0.79Abbreviations: OR, odds ratio; CI, confidence interval; NT-proBNP, N-terminal pro-brain natriuretic peptide; UHR, uric acid to HDL-c ratio; URR, urea clearance ratio Logistic regression analysis of the risk factors associated with pulmonary hypertension in non-elderly patients undergoing maintenance hemodialysis (n = 105, 26 cases with pulmonary hypertension) Note: a The multivariate model was restricted to clinically essential predictors with an events per variable ratio of 13. Negelkerke R Square: 0.43; P value of Hosmer-Lemeshow test: 0.79 Abbreviations: OR, odds ratio; CI, confidence interval; NT-proBNP, N-terminal pro-brain natriuretic peptide; UHR, uric acid to HDL-c ratio; URR, urea clearance ratio Receiver operating characteristic curves of risk factors predicting the prevalence of pulmonary hypertension. a non-elderly hemodialysis patients, b elderly hemodialysis patients. Abbreviations: AUC, area under the curve; NT-proBNP, N-terminal pro-brain natriuretic peptide; UHR, uric acid to high-density lipoprotein ratio; DM, diabetes mellitus In elderly patients undergoing hemodialysis, PH was negatively correlated with hemoglobin, triglyceride, and Scr, and was positively correlated with serum globulin and NT-proBNP (Supplementary Table 4). Univariate logistic regression analysis identified the following risk factors for PH: diabetes mellitus (DM), hemoglobin, Scr, serum globulin, serum prealbumin, Ln NT-proBNP, and NLR. Further multivariate analysis revealed that two variables, diabetes mellitus (OR: 6.71, P = 0.01), and Ln NT-proBNP (OR: 4.37, P < 0.001), were strong predictors of PH. (Table 3) As depicted in Fig. 2b, the AUC and cut-off value for predicting the prevalence of PH in elderly patients were as follows: 0.87 (bootstrap 95% CI: 0.77–0.94, P < 0.001) and 6650.2 pg/mL (sensitivity 88.0%, specificity 79.6%) for NT-proBNP. The combined model, which integrated Ln NT-proBNP and DM demonstrated superior performance compared to individual prognostic factors in predicting PH (AUC: 0.90, bootstrap 95% CI 0.82–0.97, P < 0.001). Table 3Logistic regression analysis of the risk factors correlated with pulmonary hypertension in elderly patients receiving maintenance hemodialysis. (n = 74, 25 cases with pulmonary hypertension)VariablesUnivariateMultivariate aOR (95%CI)P valueOR (95%CI)P valueDiabetes, yes3.06 (1.07,8.77)0.046.71 (1.53, 29.35)0.01Hemoglobin, g/L0.95 (0.91, 0.98)0.005//Serum creatinine, mg/dL1.00 (0.99, 1.00)0.001//Globulin, g/L1.23 (1.07, 1.41)0.003//Pre-albumin, mg/dL0.91 (0.84, 0.99)0.03//Ln NT-proBNP3.75 (2.03, 6.92)< 0.0014.37(2.21, 8.63)< 0.001NLR1.41 (1.05, 1.90)0.02//URR, %0.99 (0.91, 1.08)0.77//Note: a The multivariate model was restricted to clinically essential predictors with an events per variable ratio of 12.5. Negelkerke R Square: 0.58; P value of Hosmer-Lemeshow test: 0.59Abbreviations: OR, odds ratio; CI, confidence interval; NT-proBNP, N-terminal pro-brain natriuretic peptide; NLR, neutrophil-to lymphocyte ratio; URR, urea clearance ratio Logistic regression analysis of the risk factors correlated with pulmonary hypertension in elderly patients receiving maintenance hemodialysis. (n = 74, 25 cases with pulmonary hypertension) Note: a The multivariate model was restricted to clinically essential predictors with an events per variable ratio of 12.5. Negelkerke R Square: 0.58; P value of Hosmer-Lemeshow test: 0.59 Abbreviations: OR, odds ratio; CI, confidence interval; NT-proBNP, N-terminal pro-brain natriuretic peptide; NLR, neutrophil-to lymphocyte ratio; URR, urea clearance ratio All-cause mortality, all-cause hospitalizations, and CV hospitalization associated with PH During the follow-up period, a total of 35 all-cause deaths (9 cases among non-elderly and 26 cases among the elderly), 105 all-cause hospitalizations (56 cases among non-elderly and 49 cases among the elderly), and 26 cardiovascular hospitalizations (11 cases among non-elderly and 15 cases among the elderly) were recorded in the study population. All-cause mortality was more prevalent among the elderly than among the non-elderly (35.1% vs. 8.6%, P < 0.001). In contrast, a marginally significant difference was noted in the prevalence of all-cause hospitalization (66.2% vs. 53.3%, P = 0.08) and cardiovascular hospitalization (20.3% vs. 10.5%, P = 0.07) between the elderly and non-elderly groups. Kaplan-Meier analysis indicated that the cumulative incidences of all-cause mortality and cardiovascular hospitalization were significantly higher in both non-elderly and elderly patients with PH than in those without PH, while the cumulative incidence of all-cause hospitalization was only significantly higher in non-elderly patients (Figs. 3 and 4). Survival curves of three endpoints Kaplan-Meier survival curves for non-elderly patients undergoing maintenance hemodialysis with and without pulmonary hypertension. a Cumulative survival curves for all-cause mortality. b Cumulative survival curves for all-cause hospitalization. c Cumulative survival curves for cardiovascular hospitalization. The Log-rank P value indicated a comparison between patients with and without pulmonary hypertension. Abbreviations: PH, pulmonary hypertension; HR, hazard ratio Survival curves of three endpoints Kaplan-Meier survival curves for elderly patients undergoing maintenance hemodialysis with and without pulmonary hypertension. a Cumulative survival curves for all-cause mortality. b Cumulative survival curves for all-cause hospitalization. c Cumulative survival curves for cardiovascular hospitalization. The Log-rank P value indicated a comparison between patients with and without pulmonary hypertension. Abbreviations: PH, pulmonary hypertension; HR, hazard ratio In both unadjusted and adjusted Cox regression models, pulmonary hypertension (PH) was correlated with an elevated risk of all-cause mortality and cardiovascular hospitalization among non-elderly and elderly MHD patients. In the non-elderly group, PH also exhibits a strong association with an elevated risk of all-cause hospitalization (HR: 2.24, 95% CI 1.26–3.98, P = 0.006), a phenomenon not detected in the elderly patients. (Table 4) Table 4Association of pulmonary hypertension with all-cause mortality, all-cause hospitalization, and cardiovascular hospitalization among non-elderly and elderly hemodialysis patientsNumber of eventsn (%)UnadjustedHR (95%CI)P valueAdjusted aHR (95%CI)P value All-cause Mortality Age < 60 years With PH7 (26.9)8.66 (1.74, 43.08)0.0089.55 (1.65, 55.36)0.01No PH2 (2.50)1.00 (Referent)1.00 (Referent) Age ≥ 60 years With PH16 (64.0)3.72 (1.68, 8.27)0.0013.33 (1.44,7.67)0.005No PH10 (20.4)1.00 (Referent)1.00 (Referent) All-cause hospitalization Age < 60 years With PH20 (76.9)2.30 (1.33, 4.00)0.0032.24 (1.26, 3.98)0.006No PH36 (45.6)1.00 (Referent)1.00 (Referent) Age ≥ 60 years With PH17 (68.0)1.30 (0.72, 2.35)0.381.26 (0.69, 2.29)0.46No PH32 (65.3)1.00 (Referent)1.00 (Referent) CV Hospitalization Age < 60 years With PH6 (23.1)3.72 (1.68, 8.27)0.0013.42 (1.01, 11.58)0.048No PH5 (6.30)1.00 (Referent)1.00 (Referent) Age ≥ 60 years With PH11 (44.0)7.23 (2.30, 22.82)0.0016.80 (2.06, 22.45)0.002No PH4 (8.2)1.00 (Referent)1.00 (Referent)Note: a The values shown were adjusted for age, sex, and diabetes mellitusAbbreviations: HR, hazard ratio; CI, confidence interval; PH, pulmonary hypertension; CV, cardiovascular Association of pulmonary hypertension with all-cause mortality, all-cause hospitalization, and cardiovascular hospitalization among non-elderly and elderly hemodialysis patients Note: a The values shown were adjusted for age, sex, and diabetes mellitus Abbreviations: HR, hazard ratio; CI, confidence interval; PH, pulmonary hypertension; CV, cardiovascular In elderly MHD patients, the presence of PH was strongly correlated with an increased risk of cardiovascular hospitalization. Nevertheless, there were 15 patients (20.3%) died without CV hospitalization. The Cox regression analysis treating death as a censoring event may have led to an overestimation of the incidence of CV hospitalization, particularly in the elderly patients. Thus, we conducted the Fine-Gray competing risk regression analysis in these patients. The Fine-Gray model identified PH as a strong independent risk factor of CV hospitalization after accounting for all-cause mortality (SHR: 5.69, 95%CI 1.72–18.88, P = 0.004). (Table 5, Supplementary Fig. 1) Table 5Association between pulmonary hypertension and cardiovascular hospitalization among elderly hemodialysis patients: a Fine-Gray regression analysisVariableUnadjustedSHR (95%CI)P valueAdjusted aSHR (95%CI)P valueWith PH6.70 (2.22, 20.20)< 0.0015.69 (1.72, 18.88)0.004No PH1.00 (Referent)1.00 (Referent)Note: The competing risk event was all-cause mortalitya The values shown were adjusted for age, sex, and diabetes mellitusAbbreviations: SHR, subdistribution hazard ratio; CI, confidence interval; PH, pulmonary hypertension Association between pulmonary hypertension and cardiovascular hospitalization among elderly hemodialysis patients: a Fine-Gray regression analysis Note: The competing risk event was all-cause mortality a The values shown were adjusted for age, sex, and diabetes mellitus Abbreviations: SHR, subdistribution hazard ratio; CI, confidence interval; PH, pulmonary hypertension Discussion In the current study, we reported, for the first time, the distinct risk profiles of PH between non-elderly and elderly patients with chronic renal failure who are undergoing maintenance hemodialysis. We discovered that serum NT-proBNP was the predominant risk factor for PH in patients undergoing hemodialysis. Additionally, a lower UHR was strongly associated with the prevalence of PH in non-elderly patients, whereas diabetes mellitus was identified as a specific risk factor for PH in elderly patients. We also demonstrated that PH was strongly associated with an elevated risk of all-cause mortality, all-cause hospitalization in non-elderly patients, and an increased risk of cardiovascular hospitalization in elderly patients. Serum NT-proBNP, a crucial biomarker for evaluating volume status and left ventricular diastolic dysfunction, is recommended in the current guidelines for the diagnosis and severity assessment of pulmonary arterial hypertension [1]. Deterioration of renal function leads to gradual sodium retention, resulting in the release of compensatory natriuretic peptides in response to the increased stress and distension of left ventricular wall. NT-proBNP has been associated with the extensive pathophysiological processes in hemodialysis patients, including fluid overload, malnutrition, and inflammation [19]. Previous studies have confirmed that patients with ESRD and PH have significantly elevated serum NT-proBNP levels [6, 20–22]. In this study, the most crucial determinant of PH in MHD patients was serum NT-proBNP, with a substantial AUC of 0.80 in non-elderly patients and 0.87 in elderly patients, supporting the predominant role of volume overload in patients undergoing hemodialysis complicated with PH. Volume overload is a common complication and an independent risk factor for mortality in patients with ESRD. When assessing echocardiographic parameters in patients undergoing maintenance hemodialysis, we found that increased levels of LVEDD and average E/e’, which indicate volume overload and cardiac diastolic function, were strong predictors of PH (Supplementary Table 5). These findings were consistent with previous reports [16, 23]. Cardiac diastolic dysfunction could contribute to the development of PH by inducing an elevated left atrial pressure. Chronic and progressive volume overload with diastolic dysfunction can elevate the pulmonary capillary wedge pressure, resulting in PH. Although the pathophysiological mechanisms remain incompletely elucidated, the primary disturbance is associated with the backward transmission of elevated left-sided filling pressures. Chronic elevation of left-sided pressure can prompt reactive vasoconstriction and remodeling of the pulmonary vasculature. For patients with ESRD undergoing hemodialysis, interventions aimed at addressing volume overload should be prompt and aggressive [24]. Volume management in patients undergoing maintenance dialysis, including the establishment and maintenance of dry weights, is challenging and largely depends on physician judgment and shared decision-making [9]. Clinicians should continuously reevaluate the target weights for hemodialysis, and novel volume assessment methods, such as bioimpedance, may be useful in guiding volume assessment in clinically challenging cases. To the best of our knowledge, this is the first study to establish a significant association between the lower UHR and pulmonary hypertension in patients undergoing hemodialysis. UA and HDL-c are two common metabolism-related indicators closely associated with cardiovascular diseases. The UHR has emerged as a novel biomarker for cardiometabolic risk, reflecting the complex interplay between oxidative stress, endothelial dysfunction, systemic inflammation, and lipid metabolism [25]. Studies on the general population have indicated that an elevated UHR is associated with a higher risk of cardiovascular events and all-cause mortality [26–28]. Contrary to the findings in the general population, studies on hemodialysis patients have shown that elevated serum UA appears to be strongly associated with a lower risk of cardiovascular mortality and a U-shaped association exists between HDL-c levels and mortality [29, 30], similar to a reverse epidemiology phenomenon (paradoxical relationships between nutritional markers such as increased serum creatinine and high serum cholesterol levels that are protective in dialysis patients) [31]. In patients undergoing hemodialysis, serum creatinine, which reflects muscle mass, meat ingestion, and/or the degree of dialysis efficiency, was significantly reduced in both non-elderly and elderly patients with PH in this study. This finding was contrary to the observations in non-dialysis CKD patients [21, 32]. The mechanism underlying the paradoxical association between serum UA, HDL-c, and cardiovascular risk among hemodialysis patients has not been clearly elucidated. Based on this study, several potential mechanisms can be proposed. First, both UA and HDL-c can be regarded as indicators of the nutritional status of hemodialysis patients. It has been reported that HD patients with hyperuricemia exhibited better nutritional status than patients with normal UA levels, characterized by higher levels of creatinine, protein intake, along with a lower prevalence of cachexia [33]. A growing body of evidence has revealed a complex and even paradoxical association related to dyslipidemia in patients with ESRD on MHD, which is marked by abnormal TG production and metabolism, as well as HDL-c deficiency and dysfunction [30, 34, 35]. It is possible that the lower UHR ratio detected in hemodialysis patients with PH is not a cause but a consequence of underlying conditions, such as malnutrition and protein-energy wasting. In non-elderly patients with comorbid PH, significantly lower levels of serum UA, TG/HDL-c, Scr, and serum prealbumin, were detected, suggesting malnutrition. Malnutrition plays a crucial role in PH and has been demonstrated to be a predictor of poor clinical outcomes in PH [36, 37]. Second, both serum UA and HDL-c levels serve as indicators of oxidative stress. Serum UA functions as a powerful scavenger of oxygen-free radicals and plays a crucial role as an antioxidant. Angeles et al. reported that hemodialysis patients with hyperuricemia exhibited higher antioxidant capacity and less oxidant damage [33]. Meanwhile, a growing body of evidence indicates that in the context of uremia, HDL can transform from an antioxidant and anti-inflammatory lipoprotein into a pro-oxidant and proinflammatory particles [38–40]. Taken together, it can be hypothesized that MHD patients with PH are more likely to have a lower UHR due to the complex pathophysiological conditions in hemodialysis patients, such as malnutrition (lower uric acid) and the altered antioxidant effect of HDL, rather than a causal effect. Another important finding of this study was the strong association between diabetes mellitus and PH in elderly patients undergoing hemodialysis. Increasing evidence indicates that impaired glucose homeostasis and DM contribute significantly to the pathogenesis and prognosis of PH and the development of right ventricular failure [41–44]. The proposed mechanisms by which DM contributes to PH include chronic inflammation, endothelial cell dysfunction, disrupted lipid homeostasis, and impaired mitochondrial function [45]. Bagheri et al. [43] conducted an unbiased phenome-wide association study involving 14,861 participants and identified diabetes as the most statistically significant clinical code associated with polygenic risk for increased pulmonary pressure. Moreover, their findings showed that genetic risk for diabetes is the only significant independent causative driver of genetic risk for increased pulmonary pressure and decreased right ventricle load stress. These findings emphasize the necessity of heightened attention to the risk assessment and management of PH in elderly hemodialysis patients with comorbid diabetes mellitus. Pulmonary hypertension is a prevalent and devastating complication of chronic kidney disease, particularly in patients undergoing maintenance hemodialysis. The existence of PH significantly increased the risk of all-cause mortality and cardiovascular hospitalization in patients undergoing hemodialysis. Moreover, it substantially augmented the risk of all-cause hospitalization in the non-elderly. Volume overload and left heart disease are the major focuses of pulmonary hypertension management in CKD-associated pulmonary hypertension [24]. In addition to the routine changes in the pulmonary and cardiovascular systems associated with age, chronic diseases and comorbidities complicate the diagnostic and therapeutic approaches for PH. Advances in pharmacological therapies have substantially improved the morbidity and mortality rates of patients with PH. Nevertheless, older patients with PH are less likely to receive combined drug regimens, demonstrate poorer treatment responses, and experience inferior survival outcomes than younger patients. Therefore, early identification of PH risk in patients undergoing hemodialysis and early intervention are of great significance. Our study has some limitations. First, in clinical practice, we estimated PASP via echocardiography rather than directly measuring it through right heart catheterization. Echocardiography was randomly conducted by unbiased, proficient technicians in the hospital. The inter-observer and intra-observer variability of echocardiographic measurements could not be excluded. The gold standard for diagnosing PH is right heart catheterization, which can further categorize PH into pre-capillary PH, isolated post-capillary PH, combined post- and pre-capillary PH, and exercise PH [1]. However, it is an invasive procedure with a high risk, and is not recommended for clinical monitoring. Consequently, we lacked data to characterize the different types of PH. Second, we lacked information regarding the location and blood flow of vascular access. Arteriovenous fistulas were utilized for dialysis access in 95.5% of the patients in this study. Previous studies have reported positive associations between access blood flow and the severity of PH, attributable to their capacity to increase cardiac output and reduce compliance of pulmonary vasculature [9, 11]. However, the impact of vascular access on PH could not be evaluated precisely in this study. Third, this was a single-center retrospective study with a limited number of outcome events in the subgroup analysis. This may have reduced the precision of some estimates, as reflected in the wider confidence intervals (such as diabetes mellitus in the elderly subgroup and all-cause mortality in the non-elderly subgroup). These intervals reflect limited precision rather than a strong effect magnitude. Therapeutic interventions were not evaluated in this study. Despite adjusting for several relevant covariates, residual confounding cannot be excluded from this study. Further prospective controlled studies with large cohorts are required to validate these findings. Finally, the findings of this study were within the framework of CKD-associated pulmonary hypertension, in which volume overload and left-heart disease dominate the current paradigm and cannot be generalized to pulmonary arterial hypertension. In summary, this study explicitly investigated the differences in the prevalence, risk factors, and prognostic impact of pulmonary hypertension between non-elderly and elderly patients undergoing maintenance hemodialysis. Volume overload is the predominant risk factor for the development of PH in patients undergoing hemodialysis. Malnutrition and oxidative stress are implicated in the increased risk of PH in non-elderly hemodialysis patients, whereas comorbid DM significantly contributes to the elevated risk of PH in elderly patients. This study provides valuable clinical reference data for the early identification and intervention of PH in patients undergoing MHD based on age stratification. An age-adapted PH screening strategy, such as enhanced assessment of the nutritional status for non-elderly patients and heightened surveillance in elderly patients with diabetes mellitus, leads to more appropriate attention and allocates resources to those who are at a higher risk of adverse outcomes associated with PH. Further studies are necessary to clarify the mechanisms underlying the observed associations and explore the potential utility of novel therapeutic strategies for PH in patients undergoing maintenance hemodialysis. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements We thank all nephrologists and nurses at The First Affiliated Hospital of Sun Yat-sen University for their excellent management of patients undergoing hemodialysis. Abbreviations Pulmonary Hypertension Maintenance Hemodialysis N-terminal Pro-brain Natriuretic Peptide Uric Acid-to-High-density Lipoprotein Ratio Confidence Interval Chronic Kidney Disease End-stage Renal Disease High-sensitivity C-reactive Protein White Blood Cell Hemoglobin Platelet count Intact Parathyroid Hormone Serum Creatinine Uric Acid Total Cholesterol Triglyceride High-density Lipoprotein-cholesterol Low-density Lipoprotein Cholesterol Neutrophil-to-Lymphocyte Ratio Left Ventricular End-diastolic Diameter Interventricular Septum Thickness Left Ventricular Posterior Wall Diameter Tricuspid Annular Plane Systolic Excursion Left Ventricular Ejection Fraction Early Mitral Inflow Velocity-to-Mitral Annular Early Diastolic Velocity Ratio Pulmonary Arterial Systolic Pressure Bioelectrical Impedance Analysis Extracellular Water Intracellular Water Total Boday Water Fat-free Mass Cardiovascular Standard Deviation Interquartile Range Receiver Operating Characteristic Area Under The Curve Hazard Ratio Subdistribution Hazard Ratio Diabetes Mellitus Author contributions Z.A., Y.X. and J.W. contributed to the study design and prepared the study protocol. Q.J., S.W., Y.Z. and R.W. were involved in patient recruitment and data collection. Z.A. and Q.J. performed statistical analysis. D.Z., Y.X, and J.W. provided advice on the analysis and made substantial improvements to the paper. Z.A. wrote the main manuscript. All authors reviewed and approved the final manuscript. Funding None of the authors received any funding for this study. Data availability The data that support the findings of this study are available from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate This study was conducted as per the Declaration of Helsinki and approved by the Clinical Research and Laboratory Animal Ethics Committee of The First Affiliated Hospital of Sun Yat-sen University (approval number: [2024]847). All patients enrolled in the study provided informed consent. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Zhen Ai and Qingling Jia contributed equally to this work. Contributor Information Jianhua Wu, Email: wujianhuasara@163.com. Yuanwen Xu, Email: xuyw@mail.sysu.edu.cn. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.