[ Original Research ]

Transcription

1 [ Original Research ] Predicting Survival in Patients with Pulmonary Arterial Hypertension The REVEAL Risk Score Calculator 2. and Comparison With ESC/ERS-Based Risk Assessment Strategies Raymond L. Benza, MD; Mardi Gomberg-Maitland, MD; C. Greg Elliott, MD; Harrison W. Farber, MD; Aimee J. Foreman, MA; Adaani E. Frost, MD; Michael D. McGoon, MD; David J. Pasta, MS; Mona Selej, MD; Charles D. Burger, MD; and Robert P. Frantz, MD BACKGROUND: Pulmonary arterial hypertension is a progressive, fatal disease. Published treatment guidelines recommend treatment escalation on the basis of regular patient assessment with the goal of achieving or maintaining low-risk status. Various strategies are available to determine risk status. This analysis describes an update of the Registry to Evaluate Early and Long-Term PAH Disease Management (REVEAL) risk calculator (REVEAL 2.) and compares it with recently published European Society of Cardiology/ Respiratory Society guideline-derived risk assessment strategies. METHODS: A subpopulation from the US-based registry REVEAL that survived $ 1 year postenrollment (baseline for this cohort) was analyzed. For REVEAL 2., point values and cutpoints were reassessed, and new variables were evaluated. The Kaplan-Meier method was used to estimate survival at 12 months postbaseline; discrimination was quantified using the c-statistic. Mortality estimates and discrimination were compared between REVEAL 2. and Comparative, Prospective Registry of Newly Initiated Therapies for Pulmonary Hypertension (COMPERA) and French Pulmonary Hypertension Registry (FPHR) risk assessment strategies. For this comparison, a three-category REVEAL 2. score was computed in which patients were classified as low-, intermediate-, or high-risk. RESULTS: REVEAL 2. demonstrated similar discrimination as the original calculator in this subpopulation (c-statistic ¼.76 vs.74), provided excellent separation of risk among the risk categories, and predicted clinical worsening as well as mortality in patients who were followed $ 1 year. The REVEAL 2. three-category score had greater discrimination (c-statistic ¼.73) than COMPERA (c-statistic ¼.62) or FPHR (c-statistic ¼.64). Compared with REVEAL 2., COMPERA and FPHR both underestimated and overestimated risk. CONCLUSIONS: REVEAL 2. demonstrates greater risk discrimination than the COMPERA and FPHR risk assessment strategies in patients enrolled in REVEAL. After external validation, the REVEAL 2. calculator can assist clinicians and patients in making informed treatment decisions on the basis of individual risk profiles. TRIAL REGISTRY: ClinicalTrials.gov; No. NCT37214; URL: CHEST 219; -(-):--- KEY WORDS: ESC/ERS-derived risk assessment; pulmonary arterial hypertension; REVEAL; registry; risk score calculator chestjournal.org 1

2 Pulmonary arterial hypertension (PAH; ie, World Health Organization [WHO] group 1 pulmonary hypertension) is a rare, progressive disease characterized by elevated pulmonary vascular resistance, which leads to right ventricular failure and death. Although treatment advances have lengthened survival times, 1-3 PAH remains incurable. Regular risk assessment is recommended to guide treatment decisions and potentially improve morbidity and mortality. 4-6 Investigators have used data from PAH patient registries to derive algorithms that estimate survival in patients with PAH 7-13 and to inform treatment guidelines. 4 The first algorithm used to predict survival was derived from hemodynamic data collected by the National Institutes of Health Primary Pulmonary Hypertension Registry investigators. 8 Subsequently, the Registry to Evaluate Early and Long-Term PAH Disease Management (REVEAL) risk calculator was derived and validated on the basis of up to 12 clinically relevant variables in incident or prevalent patients in the registry. 7,14 The REVEAL risk calculator predicts survival in diverse PAH populations and provides useful serial survival assessments. 7,1,13-2 More recently, investigators have derived three additional risk assessment strategies using incident patient populations from the Swedish Pulmonary Arterial Hypertension Register (SPAHR), 21 the Comparative, Prospective Registry of Newly Initiated Therapies for Pulmonary Hypertension (COMPERA), 22 and the French Pulmonary Hypertension Registry (FPHR). 23 These risk assessment strategies are made on the basis of thresholds defined by the European Society of Cardiology and European Respiratory Society (ESC/ ERS) 4 in four to eight variables, and are calculated by various methods The original REVEAL risk calculator allowed for periodic refinements that include reassessing cutpoints and number of points assigned to existing variables and added new variables. Recognizing the potential for enhanced risk prediction, a new variable (all-cause hospitalization) and a revised variable (renal function measured by estimated glomerular filtration rate [egfr]) were identified for potential inclusion in an updated REVEAL risk calculator, REVEAL 2., on the basis of data showing a relationship between these variables and mortality in patients with PAH. 7,14,24-28 In addition, cutpoints and point values of existing variables were reassessed in an effort to optimize clinical relevance and to achieve better discrimination and delineation between risk categories. The purpose of this report is to describe the development of the updated REVEAL risk score calculator, REVEAL 2., and compare it with the original REVEAL calculator. We also compared the risk discrimination of REVEAL 2. with other contemporary risk assessment strategies to provide clinicians with information on the relative strength of each risk assessment strategy. Materials and Methods The REVEAL registry design and development of the original risk calculator have been described previously. 7,14,29 REVEAL was conducted in accordance with the amended Declaration of Helsinki. Institutional review boards at each study site approved the REVEAL protocol (e-appendix 1). Written informed consent was obtained from all patients. 29 Development of the REVEAL 2. calculator was made on the basis of the final study data (released February 4, 213) and included patients who survived $ 1 year postenrollment to ABBREVIATIONS: 6MWD = 6-min walk distance; BNP = brain natriuretic peptide; COMPERA = Comparative, Prospective Registry of Newly Initiated Therapies for Pulmonary Hypertension; DLCO = diffusing capacity of the lungs for carbon monoxide; egfr = estimated glomerular filtration rate; ESC = European Society of Cardiology; ERS = European Respiratory Society; FC = functional class; FPHR = French Pulmonary Hypertension Registry; PAH = pulmonary arterial hypertension; PVR = pulmonary vascular resistance; RAP = right atrial pressure; REVEAL = Registry to Evaluate Early and Long-Term PAH Disease Management; SPAHR = Swedish Pulmonary Arterial Hypertension Register; TIMI = thrombolysis in myocardial infarction; WHO = World Health Organization AFFILIATIONS: From the Allegheny General Hospital (Dr Benza), Pittsburgh, PA; George Washington University (Dr Gomberg- Maitland), Washington, DC; Intermountain Medical Center and the University of Utah (Dr Elliott), Salt Lake City, UT; Tufts Medical Center (Dr Farber), Boston, MA; ICON Clinical Research (Ms Foreman and Dr Pasta), San Francisco, CA; Houston Methodist Hospital Lung Center (Dr Frost), Weill-Cornell Medical College, Houston, TX; Mayo Clinic (Drs McGoon and Frantz), Rochester, MN; Actelion Pharmaceuticals US Inc. (Dr Selej), South San Francisco, CA; and the Mayo Clinic (Dr Burger), Jacksonville, FL. Preliminary data were presented at the International Society for Heart and Lung Transplantation 37th Annual Meeting, April 5-8, 217, San Diego, CA; The American Thoracic Society International Conference, May 19-24, 217, Washington DC; and at The American Thoracic Society International Conference, May 18-23, 218, San Diego, CA. FUNDING/SUPPORT: Actelion Pharmaceuticals US, Inc. (A Janssen Pharmaceutical Company of Johnson and Johnson) is the sponsor of the Registry to Evaluate Early and Long-Term PAH Disease Management Registry and provided funding and support for the analysis presented. CORRESPONDENCE TO: Raymond L. Benza, MD, Cardiovascular Institute, Allegheny General Hospital, 32 E North Ave, 16th Floor, South Tower, Pittsburgh, PA 15212; Raymond.Benza@ahn.org Copyright Ó 219 The Authors. Published by Elsevier Inc under license from the American College of Chest Physicians. This is an open access article under the CC BY-NC-ND license ( creativecommons.org/licenses/by-nc-nd/4./). DOI: 2 Original Research [ - # - CHEST ]

3 allow sufficient capture of all-cause hospitalization data in the previous 6 months. Baseline for this analysis was therefore defined as 1 year postenrollment in REVEAL. Changes to the original risk calculator were evaluated one variable at a time using a multivariable Cox proportional hazards model with allcause mortality as the outcome. Hospitalization and egfr were evaluated a priori for the updated REVEAL 2. model. We were not able to evaluate additional hemodynamically derived variables because of the absence of heart rate collection during the right heart catheterization. Risk calculator adjustments were made if a variable parameter estimate either reached the threshold for an additional point or no longer reached the threshold for its current point value. If an indicator variable was no longer significantly associated with mortality, we examined whether an alternative cutpoint was predictive. Potential cutpoints were explored on the basis of both clinical input and patterns in the data. The potential new variable of all-cause hospitalization within the previous 6 months was tested to determine if it was associated with risk. At each step, variables were reassessed to determine whether a change necessitated an adjustment to another variable. The most recent recorded value for each variable was used in the development of the REVEAL 2. model. Values at enrollment were used for patients with no assessments during the analysis period. The Kaplan-Meier method was used to evaluate 12-month survival for risk score groups (lowest to highest risk: # 6, 7, 8, 9, 1, 11, 12, and 13þ), starting from 1 year postenrollment (baseline). The c-statistic, a global measure of model discrimination that evaluates concordance between model predictions of survival time and actual survival time (.5 ¼ random concordance, 1 ¼ perfect concordance), 3 was calculated for REVEAL and REVEAL 2.. A sensitivity analysis was conducted to determine whether REVEAL 2. predicted clinical worsening, which is defined as subsequent all-cause hospitalization or initiation of parenteral prostacyclin therapy for any reason. 24 In contrast to the definition from Frost et al, 24 New York Heart Association/WHO functional class (FC) and 6-min walk distance (6MWD) were not included in the definition of clinical worsening in this sensitivity analysis because they are variables that are already part of the REVEAL risk calculator. Additional sensitivity analyses were conducted to compare REVEAL 2. risk discrimination in patients with known egfr to those for whom egfr was unknown. Sensitivity analyses were also conducted to assess REVEAL 2. risk discrimination in patients newly receiving a diagnosis in the REVEAL 2. cohort (ie, patients receiving a diagnosis of PAH # 3 months before enrollment). To examine whether REVEAL 2. can be used at time of diagnosis when hospitalization post-pah diagnosis does not apply, we also applied REVEAL 2. in patients newly receiving a diagnosis using study enrollment as baseline and assuming no hospitalization (score was set to ). The COMPERA 22 risk assessment strategy categorizes patients as low, intermediate, or high risk by assigning a grade (low ¼ 1, intermediate ¼ 2, high ¼ 3) to specific variables according to thresholds prescribed by the 215 ESC/ERS guidelines. 4 These variables are WHO FC, 6MWD, brain natriuretic peptide (BNP) or N-terminal fragment of pro-bnp, right atrial pressure (RAP), cardiac index, and mixed venous oxygen saturation. The overall risk category is determined by computing the mean of the risk grades from available variables for each patient and rounding to the nearest integer. The FPHR strategy 23 assesses WHO FC, 6MWD, RAP, and cardiac index for each patient and determines whether the values for these variables fall into the low-risk category according to thresholds prescribed by the 215 ESC/ERS guidelines. 4 Risk is defined by how many low-risk values are assigned to a patient. 23 In our analysis, we calculated the number of low-risk variables according to the same thresholds and considered the presence of three or four low-risk variables to be low risk, the presence of one or two low-risk variables to be intermediate risk, and no low-risk variables to be high risk. To compare REVEAL 2. with these strategies, a three-category REVEAL 2. risk score was developed by applying the mortality risk thresholds (ie, low risk < 5% 12-month mortality; intermediate risk 5% to 1% 12-month mortality; high risk > 1% 12-month mortality) from the 215 ESC/ ERS guidelines. 4 Specifically, REVEAL 2. low risk was defined as a score of # 6 (12-month mortality risk # 2.6%), intermediate risk was defined as a score of 7 or 8 (12-month mortality risk of 6.2% and 7.%, respectively), and high risk was defined as a score of $ 9 (12-month mortality risk of $ 1.7%). Kaplan-Meier survival curves were calculated for each risk group predicted by each of the three strategies. Receiver operating characteristic curves at 12 months, which show true positives (sensitivity) against false positives (1-specificity), were graphed for each strategy. The c-statistic was used to compare discrimination among the strategies. All analyses were conducted using SAS v9.4 (SAS Institute, Cary, NC). Results Of the 3,515 patients enrolled in REVEAL, 2,529 were in the analysis cohort (Fig 1); 73.7% had previously received a diagnosis (ie, > 3 months before enrollment) and 26.3% had newly received a diagnosis (ie, # 3 months before enrollment). The majority of the patients were women (8%), were categorized as New York Heart Association/WHO FC II (41.3%) or FC III (45.9%), and had a mean age of 53.6 years (Table 1). Disease characteristics are shown in Table 2. REVEAL Risk Score Calculator Update The modifications to the original REVEAL risk calculator and the resulting REVEAL 2. calculator are shown in Figure 2. The new variable of all-cause hospitalization within the previous 6 months and the revised renal insufficiency variable (defined as egfr with an explicit cutpoint [< 6 ml/min/1.73 m 2 ]or designated as renal insufficiency when egfr was unavailable) added significantly (P <.5) to the predictive capability of the REVEAL 2. calculator. Revising the cutpoints for BNP/N-terminal fragment of pro-bnp level, heart rate, diffusing capacity of the lungs for carbon monoxide (DLCO), 6MWD, and pulmonary vascular resistance (PVR) also improved the predictive power of the calculator. The mean (SD) REVEAL 2. risk score was 7.1 (2.8). Twelve-month survival rates from 1 year postenrollment were greater for patients with lower risk scores and poorer for those with higher risk scores (P <.1), with excellent separation between low- (# 6), intermediate- (7-8), and high-risk ($ 9) groups (Fig 3A). A similar result was observed with chestjournal.org 3

4 All patients enrolled in REVEAL N = 3,515 Meets traditional hemodynamic criteria (PCWP 15) n = 3,211 PCWP >15 or met entry criteria based only on hemodynamic data during exercise n = 34 Adult at diagnosis n = 2,996 Pediatric patient at diagnosis n = 215 At least 12 mo of follow-up n = 2,64 Without 12 mo of follow-up from enrollment n = 356 Not on a blinded clinical trial at enrollment n = 2,55 Patient was on blinded clinical trial at enrollment n = 9 Patient was not transplanted within 1 y after enrollment n = 2,529 Patient was transplanted in first year after enrollment n = 21 print & web 4C=FPO Analysis cohort n = 2,529 Figure 1 Strengthening the Reporting of Observational Studies in Epidemiology diagram showing analysis population. Legend Included Excluded 6-month survival rates (Fig 3B). The c-statistic for the REVEAL 2. risk score indicated similar discrimination in predicting mortality as the original REVEAL risk score (.76 [95% CI, ] vs.74 [95% CI, ]) in this cohort. The original REVEAL risk calculator had a c-statistic of.72 (95% CI, ) when applied to the original validation cohort. 14 A sensitivity analysis compared risk discrimination with REVEAL 2. in patients with egfr values and those for whom egfr was unknown. Discrimination was improved in the 1, patients with egfr (c-statistic ¼.77 [95% CI,.75-.8]), but still high in patients without an egfr value (c-statistic ¼.74 [95% CI, ]). An additional sensitivity analysis showed that REVEAL 2. discriminated risk in patients recently receiving a diagnosis (n ¼ 666; c-statistic ¼.76 [95% CI, ]) as it did in the overall analysis population (N ¼ 2529, c-statistic ¼.76). REVEAL 2. also discriminated risk well in patients recently receiving a diagnosis for which study enrollment was considered baseline (n ¼ 8; c-statistic ¼.7 [95% CI, ]). The hospitalization variable was not considered in this sensitivity analysis. The REVEAL 2. score also predicted clinical worsening (hazard ratio, 1.23; 95% CI, ) (Fig 4A) with excellent separation among the lowest risk categories at 12 months (ie, -4 vs 5-6) (Fig 4B). Comparison of Risk Assessment Strategies To compare risk assessment strategies, patients in each of the REVEAL 2. three-category score risk groups were stratified according to low-, intermediate-, and high-risk scores per both COMPERA and FPHR strategies. In the REVEAL 2. low-risk group (< 5% 12-month mortality), both COMPERA and FPHR overestimated risk in 51% (549 patients) and 6% (644 patients) of the patient population, respectively. In the REVEAL 2. intermediate-risk group (5%-1% 4 Original Research [ - # - CHEST ]

5 TABLE 1 ] Patient Demographics and Clinical Characteristics at 1 y Postenrollment Characteristic Patients at 1 y Postenrollment (n ¼ 2,529) a Full REVEAL Cohort at Study Enrollment (n ¼ 3,515) Diagnosed with PAH Previously diagnosed b 1,863 (73.7) 2,555 (72.7) Newly diagnosed c 666 (26.3) 96 (27.3) Age, y, mean (SD) 53.6 (14.3) 51.1 (16.8) Sex Men 55 (2.) 788 (22.4) Women 2,24 (8.) 2,727 (77.6) Race White 1,89 (71.5) 2,544 (72.4) Black 33 (13.) 451 (12.8) Hispanic 228 (9.) 34 (8.6) Asian or Pacific Islander 85 (3.4) 17 (3.) Native American or Native Alaskan 16 (.6) 2 (.6) Other 22 (.9) 33 (.9) Unknown 39 (1.5) 56 (1.6) WHO group I PH etiology Idiopathic 1,171 (46.3) 1,46 (47.) Heritable 74 (2.9) 88 (2.8) Other 18 (.7) 14 (.5) PAH associated with Connective tissue disease 649 (25.7) 759 (24.4) Congenital heart disease 244 (9.6) 344 (11.1) Portal hypertension 139 (5.5) 179 (5.8) HIV 48 (1.9) 54 (1.7) Other 186 (7.4). Modified NYHA/WHO FC d I 23 (8.4) 232 (8.1) II 1,3 (41.3) 1,55 (36.9) III 1,116 (45.9) 1,46 (49.1) IV 18 (4.4) 169 (5.9) 6MWD, mean (SD), m n ¼ 2, (124.5) n ¼ 2, (124.5) 25th, 5th, 75th percentile 296, 385, , 376, 451 Data are presented as No. (%) unless otherwise indicated. 6MWD ¼ 6-min walk distance; FC ¼ functional class; NYHA ¼ New York Heart Association; PAH ¼ pulmonary arterial hypertension; PH ¼ pulmonary hypertension; REVEAL ¼ Registry to Evaluate Early and Long-Term PAH Disease Management; WHO ¼ World Health Organization. a Data at 1 y postenrollment, which is considered baseline for this analysis. b Previously diagnosed is defined as diagnosed by right heart catheterization > 3 mo before study enrollment. c Newly diagnosed is defined as diagnosed by right heart catheterization # 3 mo before enrollment. d For the category of modified NYHA/WHO FC, data were missing for 99 patients. 12-month mortality), COMPERA and FPHR underestimated risk in 22% (152 patients) and 15% (12 patients) of patients, and overestimated risk in 4% (31 patients) and 19% (131 patients) of the patients, respectively. Finally, in the highest risk group (> 1% 12-month mortality), COMPERA and FPHR underestimated risk in 8% (611 patients) and 58% (44 patients) of the patient population, respectively (Table 3). The Kaplan-Meier survival estimates (Fig 5) show that the REVEAL 2. three-category score provided better characterization of the actual risk of mortality at 12 months than either COMPERA or FPHR. For chestjournal.org 5

6 TABLE 2 ] Patient Assessments at 1 y Postenrollment Parameter Patients at 1 y Postenrollment a Patients With Available Data at (N ¼ 2,529) a 1 y Postenrollment a RAP, mm Hg 8.7 (5.3) 2,38 mpap, mm Hg 48.5 (14.2) 2,525 PVR, Wood units 9.6 (6.5) 2,411 PAWP, mm Hg 1.1 (4.3) 2,468 Serum BNP level, pg/ml 321 (834.5) 1,598 N-terminal probnp level, pg/ml 1,729.6 (6,216) 318 DLCO, % predicted 58.7 (22.9) 1,625 Systolic blood pressure, mm Hg (16.8) 2,521 Heart rate, beats/min 82. (14.6) 2,523 egfr, ml/min/1.73 m (26.4) 1,77 Months from 1 y postenrollment to death/eos 4.9 (2.2) 2,529 Pericardial effusion, No. (%). 2,216 No 1,651 (74.5) Yes 565 (25.5) Renal insufficiency, No. (%) b 2,529 No 2,46 (95.1) Yes 123 (4.9) Data are presented as mean (SD) unless otherwise indicated. BNP ¼ brain natriuretic peptide; DLCO ¼ diffusing capacity of the lungs for carbon monoxide; egfr ¼ estimated glomerular filtration rate; EOS ¼ end of study; mpap ¼ mean pulmonary artery pressure; PAWP ¼ pulmonary arterial wedge pressure; PVR ¼ pulmonary vascular resistance; RAP ¼ right arterial pressure. a Data at 1 y postenrollment, which is considered baseline for this analysis. b Renal insufficiency was determined by qualitative investigator assessment. example, the REVEAL 2. intermediate-risk category demonstrated a 6.5% 12-month mortality, but both intermediate-risk categories for COMPERA and FPHR estimated mortality consistent with that seen in highrisk patients (12-month mortality > 1%; COMPERA, 11.3%; and FPHR, 1.3%, respectively) as defined by the 215 ESC/ERS guidelines. 4 The FPHR classification showed the least discrimination in 12-month mortality between intermediate- (1.3%) and high-risk (17.%) categories (Fig 5). A wider separation was seen with the REVEAL 2. three-category score (6.5% vs 25.8%) and with COMPERA (11.3% vs 29.1%) (Fig 5). Discrimination, as measured by the c-statistic, was greater for both the full REVEAL 2. score (c-statistic ¼.76) and the REVEAL 2. three-category score (.73) than for either the FPHR (.64) or COMPERA (.62) risk assessment strategies (Fig 6). Because the FPHR population was restricted to patients with idiopathic, heritable, and drug- or toxin-induced PAH, we assessed risk discrimination in patients in those PAH subgroups in the REVEAL dataset. REVEAL 2. discriminated risk equally well in patients in those PAH subgroups compared with those not in those subgroups (c-statistic ¼.76 [95% CI, ] in 1,413 patients in those PAH subgroups vs c-statistic ¼.76 [95% CI, ] in 1,116 patients not in those PAH subgroups) and better than FPHR. FPHR showed similar risk discrimination in both populations (cstatistic ¼.65 [95% CI, ] in 1,413 patients in those PAH subgroups vs c-statistic ¼.63 [95% CI,.6-.66] in 1,116 patients not in those PAH subgroups). When the population was restricted to patients recently receiving a diagnosis, REVEAL 2. showed similarly better risk discrimination regardless of PAH subgroup (c-statistic ¼.77 [95% CI, ] in 379 patients with idiopathic, heritable, and drug- or toxin-induced PAH vs c-statistic ¼.73 [95% CI, ] in 287 patients with PAH of other causes) compared with FPHR (c-statistic ¼.62 [95% CI, ] in 379 patients with idiopathic, heritable, and drug- or toxin-induced PAH vs c-statistic ¼.58 [95% CI, ] in 287 patients with PAH of other causes), although the sample size was smaller. Discussion Regular and accurate risk assessment affords clinicians a consistent method to follow their patients over time, informs treatment goals and decision-making, and 6 Original Research [ - # - CHEST ]

7 A PAH Risk Score B WHO Group 1 Subgroup Demographics (no change) CTD-PAH PoPH CTD-PAH PoPH +3 Males age >6 y Renal insufficiency Heritable Heritable Existing variables with unchanged risk points/cut-points PAH associated with connective tissue disease Heritable PAH Renal insufficiency Males age >6 y Systolic blood pressure <11 mm Hg Pericardial effusion Mean right atrial pressure 2 mm Hg within 1 y Existing variables with revised risk points/cut-points BNP/NT-proBNP Heart rate 6MWD Pulmonary vascular resistance PAH associated with portopulmonary hypertension Percent predicted DLCO NYHA/WHO functional class Comorbidities egfr <6 ml/min/1.73 m 2 or renal insufficiency (if egfr is unavailable) New/revised variables Hospitalizations within the last 6 mo egfr <6 ml/min/1.73 m 2 or renal insufficiency if missing egfr NYHA/WHO Functional Class FC I FC III 2 FC I FC III 1 FC IV FC IV C Updated PAH Risk Score All-Cause Hospitalizations 6 Mo New variable All-cause hospitalizations within 6 mo WHO Group I Subgroup CTD-PAH PoPH +3 Heritable Vital Signs SBP <11 mm Hg SBP <11 mm Hg HR >92 BPM HR >96 BPM Demographics Comorbidities Males age >6 y egfr <6 ml/min/1.73 m 2 or renal inefficiency (if egfr is unavailable) 6MWD 44 m 1 44 m 2 32 to <44 m 1 <165 m <165 m NYHA/WHO Functional Class Vital Signs I III 1 SBP <11 mm Hg IV HR >96 BPM BNP Echocardiogram (no change) <5 pg/ml >18 pg/ml 1 <5 pg/ml or NTproBNP 2 to <8 8 pg/ml or NT- <3 pg/ml pg/ml probnp 11 pg/ml 2 Pericardial effusion All-cause Hospitalizations 6 mo 6-Minute Walk Test BNP All-cause hospitalizations within 6 mo 44 m 32 to <44 m <165 m 2 1 <5 pg/ml or 8 pg/ml or 2 to NT-proBNP NT-proBNP <8 pg/ml <3 pg/ml 1,1 pg/ml 2 Pulmonary Function Test % predicted DLCO 8% 1 % predicted DLCO <4% % predicted DLCO 32% Echocardiogram Pulmonary Function Test Pericardial effusion % predicted DLCO <4% print & web 4C=FPO Right Heart Catheterization Indicates change in cutoff mrap 2 mm Hg within 1 y mrap 2 mm Hg within 1 y Indicates change in risk point PVR >32 Wood units PVR <5 Wood units 1 Right Heart Catheterization mrap >2 mm Hg within 1 y PVR <5 Wood units 1 SUM OF ABOVE + = RISK SCORE Figure 2 A, Comparison of the original REVEAL risk calculator and REVEAL 2.; B, overview of modifications in REVEAL 2.; C, and REVEAL 2. calculator. 6MWD ¼ 6-min walk distance; BNP ¼ brain natriuretic peptide; BPM ¼ beats per min; CTD-PAH ¼ PAH associated with connective tissue disease; DLCO ¼ diffusing capacity of the lungs for carbon monoxide; egfr ¼ estimated glomerular filtration rate; FC ¼ functional category; HR ¼ heart rate; mrap ¼ mean right atrial pressure; NT-proBNP ¼ N-terminal fragment of pro-brain natriuretic peptide; NYHA ¼ New York Heart Association; PAH ¼ pulmonary arterial hypertension; PoPH ¼ pulmonary arterial hypertension associated with portopulmonary hypertension; PVR ¼ pulmonary vascular resistance; REVEAL ¼ Registry to Evaluate Early and Long-Term Pulmonary Arterial Hypertension Disease Management; SBP ¼ systolic BP; WHO ¼ World Health Organization. 6 chestjournal.org 7

8 A Survival (%) Low risk Intermediate risk High risk No. at risk Score = -6 Score = 7 Score = 8 Score = 9 Score = 1 Score = 11 Score = 12 Score = 13+ 1, Time from 1 y postenrollment (mo) 1, , Score = -6 Score = 7 Score = 8 Score = 9 Score = 1 Score = 11 Score = 12 Score = 13+ print & web 4C=FPO B Time from 1 y postenrollment (mo) No. at risk Score = -6 1,73 1, Score = Score = Score = Score = Score = Score = Score = Survival (%) Score = -6 Score = 7 Score = 8 Score = 9 Score = 1 Score = 11 Score = 12 Score = 13+ Figure 3 Kaplan-Meier survival estimates by Registry to Evaluate Early and Long-Term PAH Disease Management 2. risk scores calculated at 1 y postenrollment (baseline for this analysis). A, Twelve-month survival estimate and (B) 6-month survival estimate assists in the timely referral of patients for transplantation. 5,6 Current ESC/ERS guidelines recommend regular risk assessment of patients with PAH, with an overall treatment goal of achieving a lowrisk status. 4 The REVEAL 1. risk calculator is a validated tool with good discrimination and utility in serial risk assessments 15 and the ability to predict long-term outcomes. The predictive strength of REVEAL 1. is also maintained when patient data are missing for some 8 Original Research [ - # - CHEST ]

9 A Freedom from event (%) No. at risk Score = -6 Score = 7 Score = 8 Score = 9 Score = 1 Score = 11 Score = 12 Score = , Time from 1 y postenrollment (mo) Score = -6 Score = 7 Score = 8 Score = 9 Score = 1 Score = 11 Score = 12 Score = 13+ print & web 4C=FPO B Freedom from event (%) No. at risk Score = -4 Score = 5-6 Score = 7 Score = Time from 1 y postenrollment (mo) Score = -4 Score = 5-6 Score = 7 Score = 8 Figure 4 Kaplan-Meier estimates of clinical worsening by Registry to Evaluate Early and Long-Term PAH Disease Management 2. risk scores calculated at 1 y postenrollment (baseline for this analysis). A, Clinical worsening estimates by all risk scores and (B) clinical worsening among patients with risk scores of to 8. Clinical worsening was defined as all-cause hospitalization or the introduction of a parenteral prostacyclin analog for any reason variables included in the calculator. 7,14 It is broadly applicable to the general PAH population because it predicts risk in patients recently receiving a diagnosis and those previously receiving a diagnosis with various PAH etiologies. This feature distinguishes both REVEAL 1. and 2. from the COMPERA and the FPHR risk assessment strategies because the COMPERA and the FPHR populations were newly diagnosed 22,23 and FPHR was restricted to patients with idiopathic, heritable, or drug-induced PAH. 23 Periodic refinements of risk assessment tools are essential to incorporate new data on predictors of disease progression and mortality and thereby maintain their clinical utility. For example, refinements have been made in the thrombolysis in myocardial infarction chestjournal.org 9

10 TABLE 3 ] Mortality at 12 mo Postbaseline With Each REVEAL Three-Category Score Risk Group Subdivided by the Risk Categorization of the COMPERA and FPHR Strategies No. a 12-mo Mortality, % No. a 12-mo Mortality, % REVEAL 2. Three-Category Score Risk Group Comparator Risk Group COMPERA Risk Group FPHR Risk Group Low (score # 6) Low Intermediate b 544 b 1.7 b 573 b 1.9 b High b 5 b. b 71 b. b Intermediate (score 7-8) Low c 152 c 6. c 12 c 6.9 c Intermediate High b 31 b 6.% b 131 b 7.1 b High (score $ 9) Low c 43 c 11.6 c 26 c 26.9 c Intermediate c 568 c 24.6 c 414 c 26.6 c High COMPERA ¼ Comparative, Prospective Registry of Newly Initiated Therapies for Pulmonary Hypertension; FPHR ¼ French Pulmonary Hypertension Registry; REVEAL ¼ Registry to Evaluate Early and Long-Term PAH Disease Management. a At 1-y postenrollment into REVEAL. b Overestimation of risk. c Underestimation of risk. (TIMI) risk score for patients with ST-elevation myocardial infarction 31 and those refinements that incorporated additional variables have increased its risk assessment capability over time. 32,33 Similarly, REVEAL 2. incorporates a new variable of all-cause hospitalization and refines the definition of renal insufficiency using egfr, an objective, laboratoryconfirmed measure of renal function universally available to most clinicians and patients that improved risk assessment capability. In addition, REVEAL 2. expands the cutpoints for 6MWD, BNP/N-terminal fragment of pro-bnp, and changes the cutpoints for DLCO and PVR in an effort to optimize clinical relevance. REVEAL 2. offers similar discrimination as the original calculator, provides excellent separation of risk among the risk strata, and predicts clinical worsening as well as mortality in patients with at least 1 year of follow-up. Data show that clinical worsening predicts mortality, 24,25,34 demonstrating the ability of REVEAL 2. to predict clinical worsening enhances its utility, and may prompt more intensive intervention earlier in the disease course. Although the definition of clinical worsening differs among these studies 24,25,34 and from the definition used in our analysis, each definition includes accepted indicators of clinical worsening. Our data also suggest that REVEAL 2. discriminates risk in the REVEAL cohort better than other contemporary risk assessment strategies. 22,23 Three European PAH registries (SPAHR, 21 COMPERA, 22 and FPHR 23 ) have published strategies to estimate mortality risk in patients newly receiving a diagnosis on the basis of a subset of variables and the associated thresholds defined in the ESC/ERS guidelines. 4 The current analysis showed that the REVEAL 2. three-category score discriminates risk better than the COMPERA and FPHR risk assessment strategies in a mixed population of patients newly receiving a diagnosis and those with a previous diagnosis, although all patients in the analysis cohort had been followed for at least 1 year. Demonstrating utility in a mixed cohort of patients is important, particularly a cohort in which patients previously receiving a diagnosis predominate, because this population constitutes the majority of patients seen in a routine clinical practice. REVEAL 2. also discriminated risk better than FPHR in a population of patients recently receiving a diagnosis of idiopathic, heritable, and drug- or toxin-induced PAH, the same type of population on which the FPHR strategy was developed. REVEAL 2. therefore presents a more pragmatic platform for risk assessment in the typical clinical environment. We did not perform a comparison with the SPAHR strategy because it is similar to the COMPERA strategy. Within this REVEAL population, COMPERA and FPHR overestimated and underestimated mortality risk when compared with REVEAL 2.. Notably, COMPERA underestimated risk in 8% of high-risk patients and FPHR underestimated risk in 58% of high-risk patients; such miscalculations of risk could have serious implications for patient care. A larger population of patients recently receiving a diagnosis in COMPERA and FPHR does not likely 1 Original Research [ - # - CHEST ]

11 Survival (%) No. at risk Low REVEAL 2. Low COMPERA Low FPHR Intermediate REVEAL 2. Intermediate COMPERA Intermediate FPHR High REVEAL 2. High COMPERA High FPHR 1, ,621 1, Time from 1 y postenrollment (mo) 1, ,56 1, , ,389 1, mo mortality rates for REVEAL 2., COMPERA, and FPHR risk scores Risk score 12-mo mortality estimate (95% CI) Low REVEAL % (1.1%-2.7%) Low COMPERA 3.5% (2.2%-4.9%) Low FPHR 4.2% (2.5%-5.9%) Intermediate REVEAL % (4.7%-8.4%) Intermediate COMPERA 11.3% (9.7%-12.8%) Intermediate FPHR 1.3% (8.8%-11.9%) High REVEAL % (22.7%-28.9%) High COMPERA 29.1% (22.5%-35.6%) High FPHR 17.% (13.8%-2.3%) print & web 4C=FPO Low REVEAL 2. Low COMPERA Low FPHR Intermediate REVEAL 2. Intermediate COMPERA Intermediate FPHR High REVEAL 2. High COMPERA High FPHR Figure 5 Kaplan-Meier survival curves by REVEAL 2. (three-category score), COMPERA, and FPHR risk categories from 1 y postenrollment. COMPERA ¼ Comparative, Prospective Registry of Newly Initiated Therapies for Pulmonary Hypertension; FPHR ¼ French Pulmonary Hypertension Registry. See Figure 1 legend for expansion of abbreviations. account for the difference in risk discrimination between these strategies and REVEAL 2.. Although the patient populations in COMPERA and FPHR are classified as newly diagnosed, albeit with varying definitions of this term, patients were assessed at a defined baseline and 1 to 2 years later. Patients recently receiving a diagnosis in REVEAL were similarly assessed 2 years postenrollment in the main analysis and REVEAL 2. had equally good discrimination in this population as in the overall cohort. In addition, our sensitivity analyses of patients recently receiving a diagnosis using study enrollment as baseline showed that REVEAL 2. has good discrimination in this cohort, although less than what is attained when the hospitalization variable is included. There may be several reasons why REVEAL 2. discriminated risk more accurately than the COMPERA and FPHR strategies. In REVEAL 2., unlike the other strategies, 22,23 a patient s risk score is calculated using modifiable and nonmodifiable variables. Nonmodifiable factors such as sex, race, etiology, and genomic fingerprints are important because they determine a patient s immutable and inherent risk and are at the heart of personalized medicine (ie, it is well known that the etiology of group 1 PH is associated with differential survival). Patients with connective tissue diseaseassociated PAH have markedly lower 1-year survival rates than patients with other etiologies. 35,36 Similarly, compared with patients with idiopathic PAH, those with chestjournal.org 11

12 print & web 4C=FPO Sensitivity Low-risk High-risk Specificity Risk Score c-statistic REVEAL REVEAL 2. 3-category risk score.73 COMPERA.62 FPHR.64 Figure 6 ROC curves and c-statistics at 12 months postbaseline for REVEAL 2., REVEAL 2. three-category risk score, COMPERA. and FPHR scores. ROC ¼ receiver operating characteristic. See Figure 1 and 5 legends for expansion of other abbreviations. heritable PAH or portopulmonary hypertension are known to have a worse prognosis despite similar or better risk profiles made on the basis of modifiable risk factors. 37 Nonmodifiable risk factors are included in most contemporary risk algorithms used in clinical practice to estimate cumulative risk in other disease states, such as heart failure and thrombolysis in myocardial infarction; therefore, its incorporation in stratifying PAH is sensible and reasonable. 31,38 In addition, REVEAL 2. is built on the basis of more patient variables than the COMPERA and FPHR strategies. Incorporation of more patient data provides a more complete clinical assessment that likely results in more accurate risk discrimination. Finally, REVEAL/ REVEAL 2. use weighting of variables by assigning an integer score to a risk factor in proportion to its contribution to the overall risk rating. Weighting improves a tool s calibration, which is the agreement between observed and predicted risk. The risk strategies derived from the ESC/ERS guidelines 22,23 do not use weighting, which limits their calibration and ability to accurately segregate risk, particularly in the intermediate-risk range. The use of risk assessment strategies with inadequate discrimination may misinform risk assessment and clinical management; for example, consider the case of a 52-year-old white woman recently receiving a diagnosis of PAH associated with scleroderma. 39 Her risk assessment is made on the basis of the following baseline data points: connective tissue disease-associated PAH, creatinine of 1.4 mg/dl (egfr by chronic kidney disease epidemiology collaboration ¼ 43.1 ml/min/1.73 m 2 ), WHO FC III, 6MWD of 3 m, BNP of 6 pg/ml, DLCO of 45%, presence of pericardial effusion on echocardiogram, RAP of 15 mm Hg, cardiac index of 2.2 L/min/m 2, and PVR of 9.5 Wood units. Her REVEAL 2. risk score is 11, corresponding to a high-risk classification with a predicted 12-month survival of 71%. This high-risk patient should be started on upfront combination therapy, including a parenteral prostacyclin, and be referred for lung transplantation evaluation. 4 Comparatively, using the COMPERA strategy, this patient would be classified as intermediaterisk with a score of 2.2 and a 12-month survival of 9.1%. 22 Applying the FPHR strategy, 23 this patient would be characterized as having no low-risk criteria, which corresponds to a 12-month survival of approximately 94%. Neglecting to account for PAH etiology and renal dysfunction in this patient resulted in an underestimation of risk by COMPERA and FPHR that could have important treatment implications, including failure to use parenteral prostacyclin therapy. This scenario illustrates how a tool that provides a high level of discrimination between risk categories, particularly among patients in the intermediate-risk range and during the first 12 months of follow-up, can guide clinical decisions at the individual patient level. The current analysis has several limitations. Patients in REVEAL were treated at centers with PAH expertise in the United States, which may limit the generalizability of the findings to all PAH patients. Another limitation is that all-cause hospitalization (within the previous 6 months) was added as a variable to REVEAL 2. instead of PAH-related hospitalization. Although PAH-related hospitalization may be more relevant to this population, only all-cause hospitalization data were collected for the entire cohort. Nonetheless, there is a high incidence of all-cause hospitalization in PAH patients with many admissions precipitated by PAH-related comorbidities 25 ; this observation suggests that the all-cause hospitalization variable provides a reasonable approximation of PAH-related hospitalization. Another limitation is that 12-month follow-up was made on the basis of patients who survived for at least 1 year from enrollment to account for all-cause hospitalization data in the previous 6 months; therefore, there may be concern that the results are subject to survival bias. Risk prognostication 12 Original Research [ - # - CHEST ]

13 is typically not subject to survivor bias because risk is assessed only during the time the patient has participated in the registry. In addition, the sensitivity analyses assessing risk in patients recently receiving a diagnosis showed similar risk discrimination with REVEAL 2. as in the full analysis cohort. Comparisons between the risk assessment strategies were also conducted in one of the derivative cohorts (REVEAL cohort); thus, these findings need to be validated in a nonderivative cohort. No comparison was made with the SPAHR risk assessment strategy 21 because of the extensive overlap between the SPAHR and COMPERA strategies. 21,22 Finally, prospective, external validation of all risk assessment strategies is required to support the integration of these tools into real-world clinical practice. Quantitative risk assessment tools such as the REVEAL 2. risk calculator can provide a valuable adjunct to standard clinical assessments because they provide a quantitative measure that can be easily tracked over time. REVEAL 2. more accurately stratifies individual patient risk because, in part, of inclusion of multiple, weighted modifiable and nonmodifiable variables. It also offers the advantage of predicting clinical worsening, which may enhance the capacity of the calculator to identify high-risk patients earlier in the disease course and support more informed treatment decisions. For clinicians already routinely using risk assessment algorithms, the REVEAL 2. calculator represents a useful addition to previously published strategies. In the future, with efforts to prospectively validate these tools, their use in routine clinical practice and in clinical trials will increase. Conclusions The current analysis describes an updated version of the REVEAL risk score calculator, which includes an additional variable, expands and refines cutpoints for existing variables, and predicts both clinical worsening and survival in patients who survived at least 1 year from enrollment. REVEAL 2. has similar discrimination as the original calculator and has better discrimination than the COMPERA and FPHR strategies when applied to a mixed cohort of patients recently or previously receiving a diagnosis of PAH. It is important that the enhanced discrimination allows for more informed treatment decisions that can be sacrificed when risk stratification is oversimplified. Following external validation, REVEAL 2., when used as part of a regular assessment strategy, may assist clinicians and patients to tailor treatment decisions aimed toward the overall goal of achieving a low-risk status. 4-6 In post hoc analyses, interventions in clinical trials have been shown to correlate with improvements in the REVEAL risk score. 17 Validation of risk assessment strategies as a surrogate end point should be explored in future clinical trials. 6 Acknowledgments Author contributions: R. L. B. takes responsibility for the content of the manuscript, including the data and analysis. R. L. B., M. G.-M., C. G. E., H. W. F., A. J. F., A. E. F., M. D. M., D. J. P., M. S., C. D. B., and R. P. F. were involved in study concept and design, preparation of the draft manuscript, and critical revision and approval of the final manuscript for interpretation of the data and important intellectual input. Financial/nonfinancial disclosures: The authors have reported to CHEST the following: R. L. B. has received honoraria from Actelion Pharmaceuticals US, Inc., Gilead Sciences, Inc., and United Therapeutics Corporation; his institution received or is pending receipt of grants from Actelion, the American Heart Association, Bayer Corporation, the National Institutes of Health/ National Heart, Lung, and Blood Institute (NIH/NHLBI), and United Therapeutics Corporation. M.G.-M. is a consultant to and a steering committee member/dsmb for Arena, Acceleron, Actelion Pharmaceuticals US, Inc., Bayer Corporation, Janssen, Liquidia, Medtronic, Inc., Merck, Reata, St. Jude s, and United Therapeutics Corporation; her institution during development of the manuscript (Inova Heart and Vascular Institute) receives grant support from AADi, Actelion, and United Therapeutics for her to conduct clinical trials. C. G. E. is a consultant for Bellerophon Therapeutics, Inc., Actelion Pharmaceuticals US, Inc., and Bayer Corporation; he has received grant/research support from Actelion Pharmaceuticals US, Inc., Gilead Sciences, Inc., United Therapeutics Corporation, NIH/NHLBI, and Intermountain Research and Medical Foundation. H. W. F. has received grant/ research support from Actelion Pharmaceuticals US, Inc., Gilead Sciences, Inc., and United Therapeutics Corporation; is a member of the Speakers Bureau for Gilead Sciences, Inc., Bayer Corporation, and Actelion Pharmaceuticals US, Inc.; is an advisory board member for Bellerophon Therapeutics, Inc., Actelion Pharmaceuticals US, Inc., Bayer Corporation, Gilead Sciences, Inc., and United Therapeutics Corporation. A. J. F. and D. J. P. are former employees and former stockholders in ICON Clinical Research, which was paid by Actelion Pharmaceuticals US, Inc., to provide analytic services. A. E. F. is a consultant for Actelion Pharmaceuticals US, Inc., and Gilead Sciences, Inc., and received grant/research support for Bayer Corporation, Gilead Sciences, Inc., Actelion Pharmaceuticals US, Inc., Reata Pharmaceuticals, Complexa, Inc., and United Therapeutic Corporation. M. D. M. is on a data monitoring committee for Pfizer and Acceleron and on an advisory committee for Lung Biotechnology. M. S. is an employee of and stockholder in Actelion Pharmaceuticals US, Inc. (a Janssen Pharmaceutical Company of Johnson and Johnson). C. D. B. served as an advisory board member for Actelion Pharmaceuticals US, Inc., and Gilead Sciences, Inc., and received grant/research support for Gilead Sciences, Inc., Actelion Pharmaceuticals US, Inc., and United Therapeutics Corporation. R. P. F. is a consultant and steering committee member for Actelion Pharmaceuticals US, Inc., United Therapeutics Corporation, and St. Jude chestjournal.org 13

14 Medical, Inc.; is a consultant for Arena and Bayer Pharmaceuticals. Third-party medical editorial assistance was provided by Twist Medical and by Donna Simcoe of Simcoe Consultants, Inc., which were supported by Actelion Pharmaceuticals US, Inc. Role of the sponsor: The sponsor had no role in the design of the study, or the collection and analysis of the data One of the authors (Dr Selej) is an employee of the sponsor and participated to a similar degree as the other authors in the preparation of the manuscript. Additional information: The e-appendix can be found in the Supplemental Materials section of the online article. References 1. Benza RL, Miller DP, Barst RJ, Badesch DB, Frost AE, McGoon MD. An evaluation of long-term survival from time of diagnosis in pulmonary arterial hypertension from the REVEAL Registry. Chest. 212;142(2): Farber HW, Miller DP, Poms AD, et al. Five-year outcomes of patients enrolled in the REVEAL Registry. Chest. 215;148(4): Lau EMT, Giannoulatou E, Celermajer DS, Humbert M. Epidemiology and treatment of pulmonary arterial hypertension. Nat Rev Cardiol. 217;14(1): Galiè N, Humbert M, Vachiery JL, et al. 215 ESC/ERS guidelines for the diagnosis and treatment of pulmonary hypertension: the Joint Task Force for the Diagnosis and Treatment of Pulmonary Hypertension of the European Society of Cardiology (ESC) and the European Respiratory Society (ERS): Endorsed by: Association for European Paediatric and Congenital Cardiology (AEPC), International Society for Heart and Lung Transplantation (ISHLT). Eur Heart J. 216;37(1): Farber HW, Benza RL. Risk assessment tools in pulmonary arterial hypertension. Prognosis for prospective trials? Am J Respir Crit Care Med. 218;197(7): Weatherald J, Boucly A, Sahay S, Humbert M, Sitbon O. The low-risk profile in pulmonary arterial hypertension. Time for a paradigm shift to goal-oriented clinical trial endpoints? Am J Respir Crit Care Med. 218;197(7): Benza RL, Miller DP, Gomberg- Maitland M, et al. Predicting survival in pulmonary arterial hypertension: insights from the Registry to Evaluate Early and Long-Term Pulmonary Arterial Hypertension Disease Management (REVEAL). Circulation. 21;122(2): D Alonzo GE, Barst RJ, Ayres SM, et al. Survival in patients with primary pulmonary hypertension. Results from a national prospective registry. Ann Intern Med. 1991;115(5): Lee WT, Ling Y, Sheares KK, Pepke- Zaba J, Peacock AJ, Johnson MK. Predicting survival in pulmonary arterial hypertension in the UK. Eur Respir J. 212;4(3): Ling Y, Johnson MK, Kiely DG, et al. Changing demographics, epidemiology, and survival of incident pulmonary arterial hypertension: results from the pulmonary hypertension registry of the United Kingdom and Ireland. Am J Respir Crit Care Med. 212;186(8): Thenappan T, Glassner C, Gomberg- Maitland M. Validation of the pulmonary hypertension connection equation for survival prediction in pulmonary arterial hypertension. Chest. 212;141: Humbert M, Sitbon O, Yaïci A, et al. Survival in incident and prevalent cohorts of patients with pulmonary arterial hypertension. Eur Respir J. 21;36(3): Sitbon O, Benza RL, Badesch DB, et al. Validation of two predictive models for survival in pulmonary arterial hypertension. Eur Respir J. 215;46(1): Benza RL, Gomberg-Maitland M, Miller DP, et al. The REVEAL Registry risk score calculator in patients newly diagnosed with pulmonary arterial hypertension. Chest. 212;141(2): Benza RL, Miller DP, Foreman AJ, et al. Prognostic implications of serial risk score assessments in patients with pulmonary arterial hypertension: a Registry to Evaluate Early and Long-Term Pulmonary Arterial Hypertension Disease Management (REVEAL) analysis. J Heart Lung Transplant. 215;34(3): Benza RL, Lohmueller LC, Kraisangka J, Kanwar M. Risk assessment in pulmonary arterial hypertension patients: the long and short of it. Adv Pulm Hypertens. 218;16(3): Benza RL, Farber HW, Frost A, et al. REVEAL risk scores applied to riociguattreated patients in PATENT-2: impact of changes in risk score on survival. J Heart Lung Transplant. 218;37(4): Benza RL, Farber HW, Frost A, et al. REVEAL risk score in patients with chronic thromboembolic pulmonary hypertension receiving riociguat. J Heart Lung Transplant. 218;37(7): Kane GC, Maradit-Kremers H, Slusser JP, Scott CG, Frantz RP, McGoon MD. Integration of clinical and hemodynamic parameters in the prediction of long-term survival in patients with pulmonary arterial hypertension. Chest. 211;139(6): Escribano-Subias P, Blanco I, López- Meseguer M, et al. REHAP investigators. Survival in pulmonary hypertension in Spain: insights from the Spanish registry. Eur Respir J. 212;4(3): Kylhammar D, Kjellström B, Hjalmarsson C, et al. A comprehensive risk stratification at early follow-up determines prognosis in pulmonary arterial hypertension. Eur Heart J. 218;39(47): Hoeper MM, Kramer T, Pan Z, et al. Mortality in pulmonary arterial hypertension: prediction by the 215 European pulmonary hypertension guidelines risk stratification model. Eur Respir J. 217;5(2): Boucly A, Weatherald J, Savale L, et al. Risk assessment, prognosis and guideline implementation in pulmonary arterial hypertension. EurRespirJ. 217;5(2): Frost AE, Badesch DB, Miller DP, Benza RL, Meltzer LA, McGoon MD. Evaluation of the predictive value of a clinical worsening definition using 2-year outcomes in patients with pulmonary arterial hypertension: a REVEAL Registry analysis. Chest. 213;144(5): Burger CD, Long PK, Shah MR, et al. Characterization of first-time hospitalizations in patients with newly diagnosed pulmonary arterial hypertension in the REVEAL registry. Chest. 214;146(5): Shah SJ, Thenappan T, Rich S, Tian L, Archer SL, Gomberg-Maitland M. Association of serum creatinine with abnormal hemodynamics and mortality in pulmonary arterial hypertension. Circulation. 28;117(19): Haddad F, Peterson T, Fuh E, et al. Characteristics and outcome after hospitalization for acute right heart failure in patients with pulmonary arterial hypertension. Circ Heart Fail. 211;4(6): Chakinala MM, Coyne DW, Benza RL, et al. Impact of declining renal function on outcomes in pulmonary arterial hypertension: a REVEAL registry analysis. J Heart Lung Transplant. 218;37(6): McGoon MD, Krichman A, Farber HW, et al. Design of the REVEAL registry for US patients with pulmonary arterial hypertension. Mayo Clin Proc. 28;83(8): Pencina MJ, D Agostino RB Sr. Evaluating discrimination of risk prediction models: the C statistic. JAMA. 215;314(1): Morrow DA, Antman EM, Charlesworth A, et al. TIMI risk score for ST-elevation myocardial infarction: a convenient, bedside, clinical score for risk assessment at presentation. Circulation. 2;12(17): Grabowski M, Filipiak KJ, Malek LA, et al. Admission B-type natriuretic peptide assessment improves early risk stratification by Killip classes and TIMI risk score in patients with acute ST elevation myocardial infarction treated with primary angioplasty. Int J Cardiol. 27;115(3): Giraldez RR, Sabatine MS, Morrow DA, et al. Baseline hemoglobin concentration and creatinine clearance composite laboratory index improves risk stratification in ST-elevation myocardial infarction. Am Heart J. 29;157(3): Original Research [ - # - CHEST ]