There remains a lack of prognosis models for patients with chronic thromboembolic pulmonary hypertension (CTEPH). This study aims to develop a nomogram predicting 3-, 5-, and 7-year survival in patients with CTEPH and verify the prognostic model. Patients with CTEPH diagnosed in Fuwai Hospital were enrolled consecutively between May 2013 and May 2019. Among them, 70% were randomly split into a training set and the other 30% as a validation set for external validation. Cox proportional hazards model was used to identify the potential survival-related factors which were candidate variables for the establishment of nomogram and the final model was internally validated by the bootstrap method. A total of 350 patients were included in the final analysis and the median follow-up period of the whole cohort was 51.2 months. Multivariate analysis of Cox proportional hazards regression showed body mass index, mean right atrial pressure, N-terminal pro-brain natriuretic peptide (per 500 ng/ml increase in concentration), presence of anemia, and main treatment choice were the independent risk factors of mortality. The nomogram demonstrated good discrimination with the corrected C-index of 0.82 in the training set, and the C-index of 0.80 (95% CI: 0.70 to 0.91) in the external validation set. The calibration plots also showed a good agreement between predicted and actual survival in both training and validation sets. In conclusion, we developed an easy-to-use nomogram with good apparent performance using 5 readily available variables, which may help physicians to identify CTEPH patients at high risk for poor prognosis and implement medical interventions.
Chronic thromboembolic pulmonary hypertension (CTEPH) is a devastating consequence of pulmonary embolism. Within the first 2 years after symptomatic pulmonary embolism, the cumulative incidence of CTEPH is about 0.1% to 11.8%. Pulmonary endarterectomy (PEA), balloon pulmonary angioplasty (BPA), pulmonary arterial hypertension (PAH)-targeted drug and lung transplantation for selected patients are the available treatments for CTEPH. With the continuous development of medical techniques over the past decades, the prognosis of CTEPH patients has been greatly improved. , However, there remains a lack of prognosis models of CTEPH patients for clinical use. Based on the clinical data and the experience of intervention, this study sought to develop and validate an easy-to-use nomogram to help physicians identify CTEPH patients at high risk for poor prognosis and implement medical interventions.
Methods
From May 2013 to May 2019, the consecutively hospitalized patients firstly diagnosed with CTEPH at the thrombosis center of Fuwai Hospital were consecutively enrolled. The follow-up of all patients was performed annually by phone, hospitalization, or outpatient visit. Information about all-cause mortality was collected, and the cut-off date was February 22, 2021. For development and validation of the prediction model, 70% of the whole cohort was used as a training set and the other 30% as a validation set in random.
The diagnosis criterion of CTEPH is in accord with the international guidelines. , Briefly described, the diagnosis of CTEPH should simultaneously meet 3 criteria as follows: (1) the diagnosis of CTEPH is made after at least 3 months of adequate anticoagulation; (2) mean pulmonary arterial pressure ≥25 mm Hg, with pulmonary arterial wedge pressure ≤15 mm Hg in the right-sided cardiac catheterization; (3) specific diagnostic signs for CTEPH on imaging tests (including perfusion lung scan, multidetector computed tomography angiography or pulmonary angiography).
The final treatment decision (targeted drug alone, PEA or BPA) of each patient was made by an experienced multidisciplinary team and considered the choice of patients or their authorized agents. The operability of the CTEPH patients was assessed by surgeons with the expertise of PEA according to guidelines suggested, , , and PEA was performed following the standard procedure established by the University of California, San Diego group. , BPA was considered for patients who were technically inoperable or who carried an unfavorable risk to benefit ratio for PEA, and the procedure of BPA was similar to that in Japan. For patients who were not applicable to perform PEA and BPA, or who refused these treatments, targeted drugs were administrated, these drugs included phosphodiesterase type 5 inhibitors, endothelin receptor antagonists, prostacyclin analog, and guanylate stimulator riociguat.
The clinically significant data on the baseline, including demographics, laboratory tests, hemodynamic parameters, diagnoses, combination drugs, and the main choice of treatments, were obtained by reviewing the electronic medical records. The protocol was approved by the Institutional Ethics Review Board of Fuwai Hospital and all patients gave written informed consents according to the principles of the Declaration of Helsinki.
Continuous variables were expressed as mean ± SD or median (interquartile range [IQR]). The categorical variables were described as frequencies and proportions. The t test was performed for continuous variables normally distributed. Otherwise, the Mann-Whitney U test was performed. The differences of categorical variables were tested by the chi-square test or Fisher’s exact test.
Cox proportional hazards model was used to identify the potential survival-related factors, the hazard ratio (HR) and 95% confidence interval (CI) were estimated to assess the influence of different variables on survival. Spearman correlation analysis was conducted to explore the potential collinearity in the variables which stand for demographic characteristics and the variables with a p value ≤0.10 in the univariate analysis. The correlation coefficient ≥0.7 between the 2 variables was regarded as collinearity, and only 1 of them was reserved for further analysis. The selected variables without collinearity (including demographic variables and variables with p value ≤0.10 in the univariate analysis) were screened in multivariate analysis by the backward selection process.
The final candidates for our formulation of nomogram were those variables that were independent risk factors associated with mortality showed by the multivariate analysis. A nomogram for predicting the long-term survival of CTEPH patients was built through the “RMS” package in R software (Version 4.0.0, Foundation for Statistical Computing, Vienna, Austria). We examined the calibration and discrimination of the nomogram by calibration plots and C-index, respectively. The final model was internally validated using bootstrap with 1,000 resamples to calculate the corrected C-index in the training set and the external validation was performed in the validation set.
Sensitivity analyses were conducted to examine the robustness of our results by including potential confounding variables in the Cox proportional hazards model, and the net reclassification index (NRI) was calculated to assess whether the model was promoted after including these variables.
All the p values were two-tailed, and a p value <0.05 was believed to be statistically significant. Bonferroni correction was used to adjust p values in multiple comparisons. Statistical analyses were performed using IBM SPSS Statistics for Windows, Version 25.0. (Armonk, New York. IBM Corp.) and R software.
Results
From May 2013 to May 2019, this study included 364 patients diagnosed with CTEPH at the thrombosis center of Fuwai Hospital, 14 patients were lost and 350 patients (96.2%) completed follow-up. From the date of diagnosis to the date of death or last observation, the median follow-up time of all included patients was 51.2 (range: 0.3 to 93.9) months. The baseline characteristics of all patients are listed in Table 1 . The mean age of the overall cohort was 52.6 ± 14.8, 160 (45.7%) of them were female. In the 350 CTEPH patients, 121 (34.6%) underwent BPA, 123 (35.1%) received PEA, and 106 (30.3%) were treated with a targeted drug alone.
| Variables | Total (n=350) | Survival (n=305) | Non-survival (n=45) | p value |
|---|---|---|---|---|
| Age (years) | 52.6±14.8 | 52.0±14.6 | 56.4±15.3 | 0.075 |
| Women | 160 (45.7%) | 138 (45.2%) | 22 (48.9%) | 0.647 |
| BMI (kg/m 2) | 23.90±3.52 | 24.10±3.49 | 22.54±3.44 | 0.005 |
| Smoker | 103 (29.4%) | 90 (29.5%) | 13 (28.9%) | 0.932 |
| Alcohol drink | 60 (17.1%) | 53 (17.4%) | 7 (15.6%) | 0.762 |
| History of VTE | 270 (77.1%) | 238 (78.0%) | 32 (71.1%) | 0.302 |
| Time from first symptom to diagnosis (months) | 24.33 [8.47, 57.80] | 23.33 [7.63, 49.73] | 36.53 [24.00, 60.90] | 0.026 |
| WHO FC III/IV | 213 (60.9%) | 178 (58.4%) | 35 (77.8%) | 0.013 |
| 6MWD (m) | 390.00 [342.00, 442.00] | 402.00 [342.00, 445.00] | 342.00 [253.00, 402.00] | <0.001 |
| Systolic PAP (mm Hg) | 83.69±22.34 | 83.32±22.61 | 86.20±20.48 | 0.420 |
| Diastolic PAP (mm Hg) | 31.00 [25.00, 37.00] | 31.00 [25.00, 37.00] | 33.00 [26.00, 40.00] | 0.204 |
| Mean PAP (mm Hg) | 49.87±12.42 | 49.50±12.50 | 52.36±11.70 | 0.151 |
| Mean RAP (mm Hg) | 7.00 [5.00, 10.00] | 7.00 [5.00, 10.00] | 10.00 [7.00, 14.00] | <0.001 |
| PAWP (mm Hg) | 10.00 [8.00, 12.00] | 10.00 [8.00, 12.00] | 10.00 [9.00, 12.00] | 0.208 |
| PVR (Wood units) | 8.73 [6.45, 11.84] | 8.67 [6.45, 11.50] | 9.94 [6.64, 13.14] | 0.055 |
| CO (L/min) | 4.43 [3.77, 5.23] | 4.47 [3.83, 5.27] | 4.40 [3.37, 4.90] | 0.070 |
| CI (L/(min.m ) | 2.47 [2.10, 2.82] | 2.47 [2.12, 2.82] | 2.40 [1.94, 2.80] | 0.262 |
| S V O 2 (%) | 62.55 [58.50, 67.30] | 63.00 [59.00, 67.70] | 58.80 [54.90, 65.53] | <0.001 |
| Laboratory tests | ||||
| HGB (g/L) | 150.00 [135.00, 163.00] | 152.00 [137.00, 164.00] | 142.00 [115.00, 152.00] | <0.001 |
| WBC (× 10 9 ) | 6.10 [4.94, 7.38] | 6.17 [4.92, 7.41] | 5.92 [5.14, 7.06] | 0.538 |
| PLT (× 10 9 ) | 187.50 [152.00, 240.00] | 189.00 [152.00, 241.00] | 183.00 [147.00, 231.00] | 0.379 |
| NT-proBNP (ng/ml) | 1,249.0 [361.00, 2,858.0] | 1,086.0 [325.00, 2,782.0] | 2,026.0 [858.80, 4,150.3] | 0.004 |
| d -dimer (μg/ml) | 0.50 [0.23, 1.01] | 0.48 [0.23, 0.93] | 0.82 [0.23, 1.20] | 0.129 |
| Total protein (g/L) | 65.05 [59.80, 69.30] | 65.20 [59.90, 69.30] | 63.10 [59.00, 68.60] | 0.193 |
| Albumin (g/L) | 39.75 [36.70, 42.80] | 40.10 [37.00, 43.20] | 37.00 [34.40, 40.20] | <0.001 |
| A/G | 1.61 [1.42, 1.84] | 1.61 [1.45, 1.84] | 1.48 [1.30, 1.71] | 0.013 |
| ALT (IU/L) | 23.00 [16.00, 34.00] | 24.00 [16.00, 35.00] | 19.00 [12.00, 27.00] | 0.010 |
| AST (IU/L) | 25.00 [20.00, 31.00] | 25.00 [20.00, 31.00] | 25.00 [19.00, 31.00] | 0.548 |
| ALP (IU/L) | 68.00 [55.00, 88.00] | 67.00 [55.00, 84.00] | 83.00 [57.00, 115.00] | 0.010 |
| Total bilirubin (μmol/L) | 19.46 [14.10, 30.61] | 19.10 [13.80, 29.84] | 23.10 [16.60, 40.00] | 0.009 |
| Direct bilirubin (μmol/L) | 4.44 [3.00, 7.70] | 4.31 [2.90, 7.31] | 5.70 [3.41, 11.10] | 0.012 |
| Creatinine (μmol/L) | 79.15 [69.00, 91.90] | 79.71 [69.38, 91.90] | 77.73 [64.70, 91.00] | 0.527 |
| BUN (mmol/L) | 5.90 [5.06, 7.14] | 5.81 [5.02, 7.00] | 6.41 [5.50, 8.10] | 0.010 |
| Uric acid (μmol/L) | 417.00 [341.00, 509.66] | 416.00 [347.87, 507.00] | 427.95 [314.16, 546.20] | 0.890 |
| S A O 2 (%) | 90.35 [87.80, 93.20] | 90.50 [88.00, 93.40] | 89.70 [86.50, 91.80] | 0.062 |
| Comorbidities | ||||
| Thrombophilia | 31 (8.9%) | 27 (8.9%) | 4 (8.9%) | >0.999 |
| APS | 26 (7.4%) | 24 (7.9%) | 2 (4.4%) | 0.553 |
| Coronary disease | 23 (6.6%) | 19 (6.2%) | 4 (8.9%) | 0.516 |
| Diabetes mellitus | 12 (3.4%) | 10 (3.3%) | 2 (4.4%) | 0.658 |
| Hypertension | 84 (24.0%) | 75 (24.6%) | 9 (20.0%) | 0.501 |
| Gastrointestinal disease | 21 (6.0%) | 19 (6.2%) | 2 (4.4%) | >0.999 |
| Chronic kidney disease | 8 (2.3%) | 5 (1.6%) | 3 (6.7%) | 0.070 |
| Cancer | 4 (1.1%) | 3 (1.0%) | 1 (2.2%) | 0.425 |
| Hyperhomocysteinemia | 119 (34.0%) | 101 (33.1%) | 18 (40.0%) | 0.363 |
| Anemia | 28 (8.0%) | 18 (5.9%) | 10 (22.2%) | 0.001 |
| Main treatment modality | <0.001 | |||
| PEA | 123 (35.1%) | 117 (95.1%) | 6 (4.9%) | |
| BPA | 121 (34.6%) | 115 (95.0%) | 6 (5.0%) | |
| Targeted drug | 106 (30.3%) | 73 (68.9%) | 33 (31.1%) | |
| Targeted drug at diagnosis (No drugs/monotherapy/Combination therapy) | 23/119/208 | 23/97/185 | 0/22/23 | 0.465 |
| Anticoagulants † | 0.026 | |||
| Warfarin | 141 (40.3%) | 130 (42.6%) | 11 (24.4%) | |
| NOAC | 208 (59.4%) | 175 (57.4%) | 33 (73.3%) | |
| Combination medicine | ||||
| Antiplatelet | 22 (6.3%) | 21 (6.9%) | 1 (2.2%) | 0.333 |
| NASID | 7 (2.0%) | 7 (2.3%) | 0 (0.0%) | 0.602 |
| Steroidhormones | 20 (5.7%) | 20 (6.6%) | 0 (0.0%) | 0.090 |
| CCB | 3 (0.9%) | 3 (1.0%) | 0 (0.0%) | >0.999 |
| Diuretic | 306 (87.4%) | 266 (87.2%) | 40 (88.9%) | 0.752 |
| Digoxin | 196 (56.0%) | 173 (56.7%) | 23 (51.1%) | 0.479 |
| IVC filter | 52 (14.9%) | 48 (15.7%) | 4 (8.9%) | 0.228 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
