Acceleration time from the Doppler registration in the RVOT close to the pulmonary valve was found to correlate with both the echocardiographic estimation of SPAP and the invasively measured SPAP and MPAP, supported by previous findings . Importantly, the present study evaluated the relationship to invasively measured SPAP, whereas some previous studies have focused on MPAP [11, 12]. Moreover, the present study shows that this correlation was equally strong when evaluated in the RHC subgroup with high prevalence of PH. It has previously been shown that AT should be possible to measure in 99% of patients out of which 25% has no measurable TR and thus provide a way of estimating the pulmonary pressure non-invasively . Our data on sensitivity and specificity with different cut-off levels of AT could be used for different clinical situations. An important clinical issue is the screening of high risk populations (SSc and HIV patients for example) with asymptomatic or mildly symptomatic PAH. Previous echocardiographic screening studies using the TR velocity (TTVG) claimed that a diagnosis of PH would be unlikely when TTVG was less than 31 mmHg (TR velocity < 2.8 m/s) and no other echocardiographic signs suggesting PH were present [2, 10, 13, 14]. When a higher cut-off level of TTVG was applied, fewer false positive PH patients were diagnosed on subsequent RHC [1, 15] but importantly the number of false negatives was not assessed.
The clear correlation between AT and SPAP (and TTVG) in both the whole population and the RHC subgroup in the present study has important implications for patients without a measurable TR. An AT of less than 100 ms would detect most patients with PH (TTVG >30 mmHg, corresponding SPAP of >38 mmg or MPAP > 25 mmHg) with an acceptable specificity as well. Furthermore, almost all patients with severe PH (SPAP >58 mmHg) could be identified using this cut-off for AT but at the price of lower specificity. However, when screening a high-risk population for PAH, the most important quality would be to have a high sensitivity for selecting patients for further evaluation with RHC. In other clinical scenarios, another cut-off level might be more appropriate. Although some early reports suggested that AT would have a good correlation to MPAP [11, 12], the use of AT in clinical practise has remained limited. The study by Yared et al. emphasized the availability of AT which together with our data supports the use of AT in clinical and research practise as a promising way of estimating SPAP.
Echocardiographic estimation of SPAP has shown a good correlation to RHC in previous studies [16, 17] as well as in our study. However, we suggest that adding RAP estimation to the TTVG did not improve the correlation and used a general 8 mmHg added to the TTVG in agreement with some previous studies when assessing SPAP [6, 18, 19]. RAP did not correlate sufficiently well with the RHC measured RAP, when using a similar estimation from ICV registrations in our population as shown earlier [9, 10]. However, as shown in our study, both ways of estimating SPAP by echocardiography showed a good correlation with the invasively measured SPAP.
Our estimation of PVR using TTVG and TVI RVOT had a better correlation than PVR estimated by the formula of Abbas. Estimating PVR by echocardiography provides additional noninvasive hemodyanamic information. Invasive measurements are essential for exact values for example pretransplantation. Being able to estimate increased PVR noninvasively could have additional value when following patients and changes of treatments. It could be more accurate to use a pressure estimate rather than velocity estimate as in the formula of Abbas when calculating resistance. Different methods could be applied to make the PVR and in our study design it seems more useful to apply the TTVG and TVI RVOT ratio. Regardless of how the PVR estimation is performed the four patients with PVR > 10 WU in the RHC-population were outliers. One reason for this might be that these four patients had poor CO on RHC (CI < 2.0). To detect this noninvasively, the tracing of the pulmonary flow might not be sensitive enough. This agrees with the findings reported by Rajagopalan  who showed that PVR estimation using the formula of Abbas  was not accurate for PVR levels above 8 WU. Our study also suggests that the PVR estimation should be avoided in patients with PVR > 10 WU or very poor CO.
Despite the clear and interesting findings the retrospective design introduces some limitations. The accuracy when evaluating AT could likely be improved if it was a standard procedure in clinical practise and performed routinely in most patients. Furthermore, even if there was a short time between RHC and echocardiography, the examinations were not simultaneous which might influence data. The number of patients without measurable AT is not available and could reflect some bias. However, it has been shown that the AT measurement is possible to measure in 99% of patients  when decided to be routinely performed. The finding that estimated RAP was not significantly correlated to invasive RAP measurement could be due to that they are not simultaneously measured. However, RAP estimated using the ICV did not change the correlation of SPAP to invasive SPAP. Furthermore, our main correlation in the total group involves the TTVG rather than SPAP. Although a prospective study design could improve the quality of measurements, our study with data from regular clinical practise allowed evaluating correlations of clinical importance. On the other hand, when evaluating methods, echocardiography performed in the catheterization laboratory should be done previous to RHC and in a blinded way.