In the Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness (ESCAPE) trial, use of a pulmonary artery catheter did not significantly affect advanced heart failure outcomes. However, the success of achieving the targeted hemodynamic goals of pulmonary capillary wedge pressure (PCWP) of 15 mm Hg and right atrial pressure (RAP) of 8 mm Hg and the association of these goals with clinical outcomes were not addressed. Furthermore, goals with 2 independent variables, PCWP and RAP, left room for uncertainties. We assessed the ability of a single hemodynamic target to achieve a threshold sum of PCWP and RAP as a predictor of all-cause mortality, death-or-transplantation (DT), or death-or-rehospitalization (DR) at 6 months in the pulmonary artery catheter–guided treatment arm of ESCAPE (n = 206). Patients with a posttreatment PCWP + RAP of <30 mm Hg had characteristics similar to those of the population who achieved the ESCAPE hemodynamic goals. This group had 8.7% mortality, 13.0% DT, and 58.7% DR at 6 months. The contrasting cohort with PCWP + RAP of ≥30 mm Hg had 45.3% mortality, 54.7% DT, and 84.9% DR at 6 months, with greater relative risk (RR) of death (RR 5.76), DT (RR 4.92), and DR (RR 1.80) and higher prevalence of jugular venous pulsation, edema, hepatomegaly, and ascites at admission and discharge. In conclusion, PCWP + RAP of 30 mm Hg posttreatment, obtained early in the index hospitalization, may represent as a simple congestion index that has prognostic value for heart failure survival and readmission rates at 6 months and as a warning signal for more aggressive intervention, thus warranting further validation.
Right-heart catheterization was introduced by Werner Forssmann in 1929. Together with André Cournand and Dickinson Richards, Forssmann was awarded the Nobel Prize in 1956 for this innovation. With the advent of the coronary care unit era, Forrester et al proposed that the pulmonary capillary wedge pressure (PCWP) and cardiac index be the 2 hemodynamic variables used to prognosticate acute heart failure (HF); this proposal is termed the Forrester classification. In 2000 to 2003, the National Heart, Lung, and Blood Institute (NHLBI) sponsored the Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness (ESCAPE) to address the value of hemodynamic-guided therapy using a pulmonary artery catheter (PAC). In the PAC arm of the study, the treatment goals were specified as a PCWP of 15 mm Hg and right atrial pressure (RAP) of 8 mm Hg; in the clinical management arm, the treatment goal was the resolution of clinical congestion (i.e., jugular venous pressure elevation, edema, and orthopnea). The investigators found no improvement in survival or reduction in rehospitalization rates at 6 months in patients randomized to PAC-guided therapy at the index admission. Since that study, routine PAC use has been discouraged, but it is still recommended to guide therapy in selected patients who have impaired perfusion, have an unclear filling pressure status, or require vasoactive medications or mechanical circulatory support. In the ESCAPE trial, the presence and magnitude of the improvement in PCWP and RAP varied. Therefore, we aimed to reassess the prognostic value of a single hemodynamic target; that is, the ability to achieve a threshold value for the sum of PCWP and RAP (PCWP + RAP) after treatment, early in the index hospitalization, in predicting all-cause mortality, the composite end point of death or cardiac transplantation (death-or-tranplantation [DT]), and the composite end point of death or rehospitalization (death-or-rehospitalization [DR]) at 6 months in the ESCAPE trial.
Methods
Our analysis was conducted using the public-use copy of the ESCAPE study from the NHLBI Biologic Specimen and Data Repository Information Coordinating Center (Bethesda, Maryland). By trial design, patients enrolled in the ESCAPE trial had a left ventricular ejection fraction of ≤30% and systolic blood pressure (SBP) of ≤125 mm Hg; the trial excluded patients with a baseline creatinine (Cr_B) level of >3.5 mg/dL. In our study, as in ESCAPE, the admission variables were labeled with “_B” and the posttreatment variables were labeled with “_L.” The study group comprised 206 patients who received a PAC during the index admission: 166 patients had RAP_L readings, 159 had PCWP_L readings, 180 had at least an RAP_L or PCWP_L reading, and 145 patients had simultaneous RAP_L and PCWP_L readings. We developed a single-pressure/congestion index that incorporated the ESCAPE hemodynamic goals for PCWP and RAP (PCWP + RAP). PCWP + CVP <30 mm Hg corresponds to the ESCAPE goals of PCWP of 15 mm Hg or RAP of 8 mm Hg ( Supplementary Tables 1 and 2 ). The analysis groups were accordingly termed PCWP + CVP <30 mm Hg and PCWP + CVP ≥30 mm Hg (after treatment). The following variables were analyzed: gender, age, implantable cardioverter-defibrillator status, ejection fraction at baseline (EF_B), systolic pressure at baseline (SB_B), blood urea nitrogen at baseline (BUN_B), serum Cr_B, cardiac index at baseline (CI_B), inotrope/vasodilator use in hospital stay, and change in weight with treatment (ΔWt). Additionally, 1.9 and 0.30 L/min/m 2 and 7.5 mm Hg were determined as the respective median levels for CI_B, ΔCI (L-B), and ΔPCWP + RAP (B-L) of the PAC population, where Δ denotes the difference between discharge (_L) and admission (_B) value (L-B), or vice versa (B-L), and these discriminating variables were also analyzed. We used the Physical Diagnosis data from ESCAPE, including jugular venous pulsation (JVP), peripheral edema (Edema), hepatomegaly status (Hepatomegaly), and ascites status (Ascites) for discrete data analysis.
Unpaired t testing was used to assess intergroup differences for continuous variables. Fisher’s exact test was used to assess categorical variables and to analyze the Physical Diagnosis discrete data using a 2 × 2 contingency table ( http://graphpad.com/quickcalcs/contingency1.cfm ). JVP, edema, and hepatomegaly were dichotomized as moderate (<13 cm, <2+, <2 fingerbreadths, respectively) or severe (≥13 cm, ≥2+, ≥2 fingerbreadths, respectively); the grade of edema is descriptive and unitless, as was prescribed by the ESCAPE study. Ascites was dichotomized as normal (none) or abnormal (present). Receiver-operating characteristic (ROC) analysis was used for general sensitivity and specificity analysis. Kaplan-Meier plots were presented for comparing survival curves, and the intergroup differences were assessed with the log-rank test. A multivariate Cox regression model was used with adjustments of the following covariates: BUN_B (≥43 vs <43 mg/dl), gender, SBP (<115 vs ≥115 mm Hg), age (<60 vs ≥60 years), cardiac index (<2.2 vs ≥2.2 L/min/m 2 ), and serum Cr_B (≥2.75 vs <2.75 mg/dl). Statview, version 5.01 (SAS Institute, Inc., Cary, North Carolina) on a Windows 7 platform was used for statistical and survival analyses. The R program, version 3.2.2 ( http://www.r-project.org/ ), with the epiR package was used to determine the odds ratio (OR) with the 2 × 2 table and to perform the Breslow-Day test of homogeneity using a 2 × 2 × k contingency table.
Results
Table 1 lists the individual RAP goal (RAP_L ≤ 8 mm Hg) or PCWP goal (PCWP_L ≤ 15 mm Hg), which had low success rates in reaching the target and also showed discordant success rates. Combined ESCAPE goals, interpreted as either target-1 (RAP_L ≤ 8 mm Hg and PCWP_L ≤ 15 mm Hg) or target-2 (RAP_L ≤ 8 mm Hg or PCWP_L ≤ 15 mm Hg), where target-1 requires both goal measurements be met but target-2 only requires one of the goal measurements be met, were inferior to the unitary target PCWP + CVP at 30 mm Hg, with the unitary target showing more optimal sensitivity and specificity. Neither median CI_B or median ΔCI_B from treatment nor median decrease in PCWP + RAP was a satisfactory index for optimizing sensitivity/specificity.
| Criteria of the Pulmonary Artery Catheter (PAC) Group | Reaching target | Mortality in failed-target group | OR ∗ | (95% C.I.) | Sensitivity (predicting death when failed) | Specificity (excluding death when successful) |
|---|---|---|---|---|---|---|
| RAP_L≤8 mmHg † | 86/166= 53% | 27/80=36% | 4.32 | (1.80, 11.32) | 0.75 | 0.59 |
| PCWP_L≤15 mmHg ‡ | 67/159= 42% | 29/92=32% | 3.02 | (1.41, 6.47) | 0.81 | 0.49 |
| ESCAPE target-1 § | 49/145= 34% | 28/96=29% | 4.59 | (1.46, 19.22) | 0.88 | 0.40 |
| ESCAPE, target-2 ¶ | 104/162= 64% | 23/58=40% | 4.98 | (2.12, 12.27) | 0.66 | 0.72 |
| PCWP+RAP_L<30 mmHg ‖ | 92/145= 63% | 24/53=45% | 8.54 | (3.27, 24.57) | 0.75 | 0.74 |
| CI_B>1.9 ∗∗ | 74/157=47% | 23/83=28% | 1.80 | (0.83, 3.88) | 0.64 | 0.50 |
| ΔCI (L-B) >0.30 †† | 78/157=50% | 18/79=23% | 0.98 | (0.47, 2.07) | 0.50 | 0.50 |
| ΔPCWP+RAP (B-L) >7.5 ‡‡ | 71/142=50% | 20/71=28% | 1.93 | (0.86, 4.32) | 0.63 | 0.54 |
∗ OR = odds ratio for odds of death failed at target to odds of death achieving target; epiR package of R program (Methods).
† RAP_L ≤ 8 mm Hg; 37 patients (43.0%) of those had PCWP_L > 15 mm Hg (n = 86-49 = 37).
‡ PCWP_L ≤ 15 mm Hg; 18 patients (26.9%) of those had RAP_L > 8 mm Hg (n = 67-49 = 18).
§ ESCAPE target-1: PCWP_L ≤ 15 mm Hg AND RAP_L ≤ 8 mm Hg.
¶ ESCAPE target-2: PCWP_L ≤ 15 mm Hg OR RAP_L ≤ 8 mm Hg.
‖ PCWP + RAP_L < 30 mm Hg: the sum of PCWP and RAP at discharge less than 30 mm Hg.
∗∗ CI_B > 1.9 L/min/m 2 : Baseline cardiac index greater than 1.9 L/min/m 2 , where 1.9 L/min/m 2 was the median value for the PAC Group CI_B.
†† ΔCI (L-B) >0.30 L/min/m 2 : the difference of CI calculated at post-treatment (_L) from that at baseline (_B) greater than 0.30 L/min/m 2 , where 0.30 L/min/m 2 was the median value for the PAC Group ΔCI (L-B).
‡‡ ΔPCWP + RAP (B-L) >7.5 mm Hg: the difference of PCWP + RAP calculated at baseline (_B) from that at post-treatment (_L) greater than 7.5 mm Hg, where 7.5 mm Hg was the median value for the PAC Group Δ PCWP + RAP (B-L).
On ROC analysis (data not shown), PCWP + RAP_L had a sensitivity of 0.75, a specificity of 0.74, and an area under the curve of 0.80 as a discriminator for death. All other admitting indexes (PCWP + RAP_B, RAP_B, PCWP_B, SBP_B, CI_B, age), posttreatment indexes (CI_L), and treatment difference indexes (ΔPCWP + RAP, ΔRAP, ΔPCWP; ΔWt) resulted in poor or failed tests as mortality discriminators (negative results not shown). The individual ESCAPE targets, RAP_L at 8 mm Hg or PCWP_L at 15 mm Hg, were fair tests (area under the curve 0.73) but had low specificities of 0.59 and 0.49, respectively.
The PCWP + RAP <30 mm Hg group and the patients who achieved the ESCAPE goals had similar distribution of baseline characteristics ( Table 2 ); the PCWP + RAP ≥30 mm Hg group and the patients who failed to achieve the ESCAPE goals also had similar distribution of baseline characteristics. Patients with PCWP + RAP ≥30 mm Hg posttreatment had a higher PCWP and higher RAP mean pressures on admission. Compared with patients with PCWP + RAP <30 mm Hg posttreatment, they also had a higher prevalence of JVP ≥13 cm, edema ≥2+, hepatomegaly ≥2 fingerbreadths, and ascites at both admission and discharge. No difference was found in inotrope/vasodilator use between groups defined by the PCWP + RAP goal versus those defined by the ESCAPE goals.
| Characteristics | PCWP+RAP groups (N=145) | P value | ESCAPE goals (N=162) | P value | ||
|---|---|---|---|---|---|---|
| PCWP+RAP_L<30 mmHg (N=92) | PCWP+RAP_L≥30 mmHg (N=53) | PCWP_L ≤15 or RAP_L ≤8 mmHg (N=104) | PCWP_L >15 and RAP_L >8 mmHg (N=58) | |||
| Baseline | ||||||
| Male | 70 (76%) | 39 (74%) | 0.84 ∗ | 74 (71%) | 45 (82%) | 0.18 ∗ |
| Age, mean (SD) (years) | 56.7 (12.9) | 57.6 (12.8) | 0.70 | 56.5 (12.9) | 57.7 (12.4) | 0.56 |
| ICD status | 34 (37%) | 14 (26%) | 0.21 ∗ | 36 (35%) | 16 (28%) | 0.13 ∗ |
| Ischemic etiology | 47 (51%) | 30 (57%) | 0.60 ∗ | 57 (55%) | 30 (52%) | 0.74 ∗ |
| EF_B, mean (SD) (%) | 19.6 (6.8) | 19.2 (7.1) | 0.69 | 20.1 (6.8) | 19.1 (6.8) | 0.37 |
| SBP_B, mean (SD) (mmHg) | 108.4 (15.4) | 102.6 (17.5) | 0.04 | 107.3 (17.2) | 103.8 (17.2) | 0.21 |
| BUN_B, mean (SD) (mg/dL) | 30.9 (20.2) | 39.5 (20.0) | 0.0152 | 31.7 (20.8) | 39.7 (22.0) | 0.0238 |
| CI_B, mean (SD) (L/min/m 2 ) | 1.98 (0.60) | 1.96 (0.51) | 0.78 | 1.96 (0.58) | 1.97 (0.52) | 0.94 |
| Cr_B, mean (SD) (mg/dL) | 1.45(0.58) | 1.62(0.63) | 0.10 | 1.52(0.64) | 1.57(0.61) | 0.60 |
| PCWP_B, mean (SD) (mmHg) | 22.2 (9.0) | 28.7 (9.0) | <0.0001 | 22.1(8.2) | 29.0(9.7) | <0.0001 |
| RAP_B, mean (SD) (mmHg) | 11.3 (6.9) | 16.1(6.7) | <0.0001 | 11.0 (6.9) | 16.0(6.4) | <0.0001 |
| JVP_B (≥13cm) | 44 (49%) | 38(72%) | 0.0043 ∗ | 46 (45%) | 42(75%) | 0.0004 ∗ |
| Edema_B (≥2+) † | 30 (33%) | 29 (55%) | 0.0136 ∗ | 33 (32%) | 30(52%) | 0.0182 ∗ |
| Hepatomegaly_B (present) | 44 (48%) | 38 (72%) | 0.0086 ∗ | 52 (50%) | 39(68%) | 0.0306 ∗ |
| Ascites_B (present) | 24 (26%) | 33 (62%) | 0.0001 ∗ | 24 (23%) | 36(62%) | 0.0001 ∗ |
| Treatment | ||||||
| Inotrope/vasodilator use | 58 (63%) | 42 (79%) | 0.062 | 68 (65%) | 45 (78%) | 0.11 |
| Discharge | ||||||
| JVP_L (≥13cm) | 1 (1.2%) | 8 (18%) | 0.0008 ∗ | 3 (3.1%) | 7 (15%) | 0.0162 ∗ |
| Edema_L (≥2+) † | 0(0%) | 5(10%) | 0.0055 ∗ | 1(1.0%) | 5(9.1%) | 0.0229 ∗ |
| Hepatomegaly_L (present) | 9(11%) | 15(31%) | 0.0090 ∗ | 9(9.6%) | 16(30%) | 0.0025 ∗ |
| Ascites_L (present) | 2 (2.3%) | 13 (27%) | 0.0001 ∗ | 2 (2.1%) | 13(25%) | 0.0001 ∗ |
| ΔWt, mean (SD) (kg) | -4.05(4.68) | -3.77(4.66) | 0.75 | -4.02(4.69) | -3.89(4.60) | 0.88 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
