Background:
Pulmonary hypertension (PH) mostly causes distressing symptoms, a reduced quality of life and an increased mortality rate. PH can also cause characteristic electrocardiographic changes. Although the electrocardiogram (ECG) is a rapid, simple, non-invasive and ubiquitously available test and studies suggest that the echocardiogram only diagnoses PH with a sensitivity and specificity of 83-85% and 72-74%, the ECG does not play a major role in the current ESC/ERS guidelines. The aim of the present study was to explore the diagnostic accuracy of ECG parameters for the diagnosis of PH.
Methods:
The present study was conducted as a bicentric, retrospective study in two German high volume referral centres for pulmonary arterial hypertension (PAH) and chronic thromboembolic pulmonary hypertension (CTEPH). Between 07/2012 and 11/2023, each 100 patients without PH (control group), with PAH (PAH group) and with CTEPH (CTEPH group), all confirmed by right heart catheterisation, were retrospectively enrolled.
Results:
A total of 300 patients were included. The control group [median pulmonary artery pressure (mPAP): 16 mm Hg, median pulmonary vascular resistance (PVR): 128 dyn∗sec∗cm−5], PAH (median mPAP: 47 mm Hg, median PVR: 716 dyn∗sec∗cm−5) and CTEPH (median mPAP: 40 mm Hg, median PVR: 536 dyn∗sec∗cm−5) groups consisted of 55%, 57% and 52% women (median age: 60.8, 65.7 and 63.2 years). Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of electrocardiographic parameters for the diagnosis of PH varied from 3-98%, 3-100%, 43-100% and 36-70% with mean values of 39%, 87%, 83% and 59%. ‘’QRS axis associated with right heart strain” showed the highest Youden index (0.53) and was able to diagnose PH with a sensitivity, specificity, PPV and NPV of 60%, 93%, 90% and 70%. In severe PH, sensitivity and PPV was higher than in non-severe PH. The currently applied cut-off’s of all analysed ECG parameters were found to be not optimal for PH diagnosis and were adjusted. After optimisation, sensitivity (range: from 3-98% to 29-81%, mean: from 39% to 66%) and Youden index (range: from -0.18-0.53 to -0.22-0.68, mean: from 0.26 to 0.40) increased significantly but specificity (range: from 3-100% to 37-92%, mean: from 87% to 74%) slightly decreased. After optimisation, the parameter “(RI + SIII) - (SI + RIII)” showed the highest Youden index (0.68). A new cut-off (< -0.05 mV) was able to diagnose PH with a sensitivity of 76% and specificity of 92% (AUC: 0.86, accuracy: 0.78).
Conclusions:
The currently applied cut-off’s seem more suitable for excluding PH, while some ECG parameters with a more optimal cut-off may have the same diagnostic accuracy as the echocardiogram for the diagnosis of PH.
Table 1 Sensitivity and specificity of selected electrocardiographic parameters for the diagnosis of pulmonary hypertension
Sensitivity Specificity PPV NPV
QRS axis associated with right heart strain*, % 60 93 90 70
P dextroatriale or P biatriale, % 42 99 98 63
Right or biventricular hypertrophy (SLI), % 45 92 85 63
qR pattern in V1, % 14 98 88 53
Right bundle branch block, % 35 89 76 58
R/S in V1 > 1.0, % 41 99 98 63
R V1, V2 + S I, aVL - S V1 > 0.6 mV, % 52 97 95 67
R V1 + S V5, V6 > 1.05 mV, % 32 99 97 59
Time to R peak in V1 (QRS < 120 ms) > 35 ms, % 61 85 80 69
*: QRS axis > 90°, SIQIII type or SISIISIII type, NPV: negative predictive value, PPV: positive predictive value, SLI: Sokolow-Lyon index
