Connective Tissue Disease in Pulmonary Arterial Hypertension
Connective Tissue Disease in Pulmonary Arterial Hypertension
Our search of the database identified 115 patients, 65 (56%) with IPAH and 50 (44%) with CTD-PAH during the study period. Patient demographics for the entire cohort and the subgroups with IPAH and CTD-PAH are summarized in Table 1. Most of the patients were middle-aged women with a body mass index <30 kg/m. There were no significant demographic differences between the IPAH and CTD-PAH subgroups.
Standard measures of PAH severity are shown in Table 2. Most patients (64%) were in WHO functional class III or IV and had increased values for BNP (reference range 0–112 pg/mL), right atrial pressure (reference range 0–5 mm Hg), MPAP (reference range <25 mm Hg), and pulmonary vascular resistance (reference range <3 WU), with a correspondingly low 6MWD. The IPAH and CTD-PAH subgroups had similar severity of disease on the basis of WHO functional class, 6MWD, and MPAP. The CTD-PAH group, however, had significantly higher BNP levels (mean standard deviation [SD], 635 [715] vs 325 [489] pg/mL; P = 0.02) and significantly lower pulmonary vascular resistance (mean [SD], 6 [3] vs 9 [5] WU; P = 0.04) than the IPAH group.
The median quantitative ANA level also was significantly higher in the CTD-PAH group than in the IPAH group (7 U vs 0 U; P < 0.001; Fig. 1). The AUC was excellent, at 0.91 (Fig. 2). The ROC for a positive ANA indicated that a quantitative cutoff value of 5 U provided for optimal specificity (94%) and sensitivity (70%). Compared with patients with a quantitative ANA <5 U, patients with higher ANA levels were four times more likely to have CTD-PAH (relative risk 4.41). As such, the positive likelihood ratio was 12 (P < 0.001) and the negative likelihood ratio was 0.32. Overall, an ANA of at least 5 U had a positive predictive value of 91% and a negative predictive value of 79%.
(Enlarge Image)
Figure 1.
Box-and-whisker plots of anti-nuclear antibody (ANA) results. Median ANA level was 7 U in the connective tissue disease–related pulmonary arterial hypertension group (CTDPAH) and 0 U in the idiopathic pulmonary arterial hypertension group (IPAH).
(Enlarge Image)
Figure 2.
Receiver operating characteristic curve for positive antinuclear antibody to identify connective tissue disease–related pulmonary arterial hypertension. Area under the curve was 0.91.
Using a model containing BNP and ANA as covariates, BNP was not an independently significant predictor (P = 0.72), although ANA remained significant (P < 0.001; data not shown). In fact, the addition of BNP in the covariate model produced less-favorable correlations than the model by ANA alone, based on a decrease in the r value, to 0.400 from 0.423. BNP also was examined as an independent predictor of CTD-PAH; however, the AUC was poor (0.66), with decreased values for both sensitivity and specificity, as compared with ANA as a predictor (data not shown).
Results
Our search of the database identified 115 patients, 65 (56%) with IPAH and 50 (44%) with CTD-PAH during the study period. Patient demographics for the entire cohort and the subgroups with IPAH and CTD-PAH are summarized in Table 1. Most of the patients were middle-aged women with a body mass index <30 kg/m. There were no significant demographic differences between the IPAH and CTD-PAH subgroups.
Standard measures of PAH severity are shown in Table 2. Most patients (64%) were in WHO functional class III or IV and had increased values for BNP (reference range 0–112 pg/mL), right atrial pressure (reference range 0–5 mm Hg), MPAP (reference range <25 mm Hg), and pulmonary vascular resistance (reference range <3 WU), with a correspondingly low 6MWD. The IPAH and CTD-PAH subgroups had similar severity of disease on the basis of WHO functional class, 6MWD, and MPAP. The CTD-PAH group, however, had significantly higher BNP levels (mean standard deviation [SD], 635 [715] vs 325 [489] pg/mL; P = 0.02) and significantly lower pulmonary vascular resistance (mean [SD], 6 [3] vs 9 [5] WU; P = 0.04) than the IPAH group.
The median quantitative ANA level also was significantly higher in the CTD-PAH group than in the IPAH group (7 U vs 0 U; P < 0.001; Fig. 1). The AUC was excellent, at 0.91 (Fig. 2). The ROC for a positive ANA indicated that a quantitative cutoff value of 5 U provided for optimal specificity (94%) and sensitivity (70%). Compared with patients with a quantitative ANA <5 U, patients with higher ANA levels were four times more likely to have CTD-PAH (relative risk 4.41). As such, the positive likelihood ratio was 12 (P < 0.001) and the negative likelihood ratio was 0.32. Overall, an ANA of at least 5 U had a positive predictive value of 91% and a negative predictive value of 79%.
(Enlarge Image)
Figure 1.
Box-and-whisker plots of anti-nuclear antibody (ANA) results. Median ANA level was 7 U in the connective tissue disease–related pulmonary arterial hypertension group (CTDPAH) and 0 U in the idiopathic pulmonary arterial hypertension group (IPAH).
(Enlarge Image)
Figure 2.
Receiver operating characteristic curve for positive antinuclear antibody to identify connective tissue disease–related pulmonary arterial hypertension. Area under the curve was 0.91.
Using a model containing BNP and ANA as covariates, BNP was not an independently significant predictor (P = 0.72), although ANA remained significant (P < 0.001; data not shown). In fact, the addition of BNP in the covariate model produced less-favorable correlations than the model by ANA alone, based on a decrease in the r value, to 0.400 from 0.423. BNP also was examined as an independent predictor of CTD-PAH; however, the AUC was poor (0.66), with decreased values for both sensitivity and specificity, as compared with ANA as a predictor (data not shown).