The current study is the first to assess reliability, precision, and CV while determining RC, MDC, and LSC for RVOTD, MPAD, RVOTVTI, and MPAVTI. It also compares RVOTSV and MPASV calculated by echocardiography. The CV of RVOTD, MPAD, RVOTVTI, and MPAVTI measurements was very low to low, according to the four observers. Nevertheless, the agreement between RVOTD, MPAD, RVOTVTI, and MPAVTI measurements was excellent from the perspectives of all four observers. The findings suggest that RVOTD, MPAD, RVOTVTI, and MPAVTI measurements are reliable. In terms of precision, these measurements were found to be satisfactory. The correlation strength between RVOTSV and MPASV ranged from moderate to strong, while the agreement between these two indices varied from poor to excellent among the four observers. A significant bias and broad limits of agreement were observed between RVOTSV and MPASV.

Yogeswaran et al. [14] compared the calculated RV output index based on RVOTD, RVOTVTI, and Doppler echocardiography-derived LV output with the Fick-derived LV output. They found these methods to be fairly to moderately correlated. In another study, Chandraratna et al. [24] measured PAD and PAVTI at the pulmonary valve level and demonstrated a strong correlation between the echocardiography-derived RV output and the RV output obtained from PA catheterization. Izzat et al. [25] assessed MPAD and MPAVTI, revealing a fair correlation between thermodilution-derived CO and Doppler echocardiography-derived LV output, as well as a strong correlation with Doppler echocardiography-derived RV output. Muhiudeen et al. [26] observed a modest correlation between RV ventricular output calculated by Doppler echocardiography at the MPA and CO measured using the thermodilution method. On the other hand, some researchers have reported a strong correlation between RV output calculated by Doppler echocardiography at the PA or the RVOT and CO derived from the thermodilution method [27,28,29,30].

Our study revealed a moderate correlation between RVOTSV and MPASV, similar to the correlation levels reported between Doppler echocardiography-derived CO and CO measured using the Fick or thermodilution methods. Nonetheless, a significant bias was observed between the two methods, with a wide range of agreement. Consequently, it can be concluded that these methods should not be considered interchangeable in clinical practice. It is noteworthy to cite that the exact comparison of calculated SV in RVOT and MPA needs another study with an appropriate design and is beyond the domain of this study.

The selection of the anatomic site for calculating right ventricular output in echocardiography should be supported by appropriate citations. Furthermore, it is essential to maintain consistency in the chosen method when performing follow-up measurements and making comparisons to ensure accuracy and reliability in clinical practice.

The proportional bias observed in the Bland-Altman plot could be attributed to either a genuine association between the difference in measured values and the mean of the two echocardiography methods or the discrepancy in within-subject standard deviations [31]. The former possibility might result from increased flow acceleration during blood flow through the pulmonary valve.

Despite calculating RVOTSV and MPASV in a single cardiac cycle, we measured the associated diameters and VTIs in three cardiac cycles. It is noteworthy that RC is a product of within-subject standard deviations. Interestingly, the RC values of RVOTD and MPAD were found to be equal, while the RC of MPAVTI was slightly higher than that of RVOTVTI. This finding indicates that variations in either within-subject standard deviations or measurement errors inherent to these methods may contribute to the observed proportional bias [31].

The discrepancies between the two methods can be attributed to several factors. Firstly, the cross-sectional areas of the RVOT and the MPA are not perfectly circular [32, 33]. Consequently, calculating the cross-sectional area using a single diameter can introduce errors. Additionally, each method involves two measurements (diameter and VTI), potentially compounding the operator-related errors in SV calculation. Several additional factors contribute to potential errors in the measurements. VTI is influenced by sample volume placement, alignment between the ultrasound beam and flow, and the flow pattern itself [11,12,13, 27]. Similarly, RVOTD measurements can be affected by near-field artifacts and trabeculation. Furthermore, RVOTD measurements are associated with axial resolution, while MPAD measurements relate to lateral resolution. Another critical aspect to consider is the dynamic nature of PAD during systole, which can introduce further variability in measurements [27, 34, 35]. The differences between the two methods may also arise from the repeatability, reliability, and precision of RVOTD, MPAD, RVOTVTI, and MPAVTI measurements. Previous studies examining CO calculation at the RVOT or the PA have not provided a complete analysis of these factors. Nevertheless, the RC of MPAD in a smaller-scale study is consistent with our findings [27]. In a separate study, the CV values for PAD and PAVTI were smaller than those reported in our study, although with wider 95% CIs [26]. Additionally, Gorcsan et al. [28] reported CVs for MPAD and MPAVTI. In comparison with our results, their CV for MPAD was higher, while their CV for MPAVTI was lower. Yogeswaran et al. [14] reported ICCs for RVOTD and RVOTVTI, similar to the values observed in our study. Mukherjee et al. [36] demonstrated that the CV and ICC of RVOTVTI were also comparable with our findings. Notably, Mukherjee et al. [36] calculated the MDC without using the square root of 2 in their calculations. In contrast, our study did not focus on a specific population (systemic sclerosis) and utilized measurements from three consecutive cardiac cycles, as opposed to two. It is generally accepted that measuring VTI in a minimum of three cardiac cycles is necessary to obtain reliable data.

In studies focusing on patients with pulmonary thromboembolism [37, 38], the mean RVOTVTI differences between various subgroups exceeded the MDC that we reported for RVOTVTI. It is important to note that MDC does not signify the minimum clinically important difference but rather serves as an estimate of the difference beyond measurement errors. Interestingly, despite evaluating different patients, the four observers in our study demonstrated similar RCs and MDCs for RVOTD, MPAD, RVOTVTI, and MPAVTI. Cheong et al. [39] found that a 15% increase in RVOTVTI during passive leg-raising accurately predicted a patient’s response to a fluid bolus. Notably, the LSC for RVOTVTI in their study was higher than that observed in our investigation. Further, despite assessing different patients, the four observers in our study reported similar precision levels and LSCs for RVOTD, MPAD, RVOTVTI, and MPAVTI, which remained within a narrow range.

The current literature is deficient in scientific evidence regarding the validation of noninvasive assessment of RVCO [40]. To address this knowledge gap, our study focused on comparing SV derived from two distinct sites. We aimed to contribute essential data on the repeatability, reliability, and precision of these measurements. By providing these valuable insights, our findings can serve as a foundation for future investigations. Researchers can utilize the RC, MDC, and LSC established in our study to interpret their results and accurately calculate sample sizes for forthcoming investigations.

From a clinical perspective, the indices of repeatability, reliability, and precision can be used in interpreting the time serial measurements. The time serial measurements are used for monitoring the progression of the diseases, treatment of disease, and the evaluation of physiological changes such as the effect of age and exercise. In addition, they provide a primary baseline for assessing quality controls in the echocardiography laboratory. For example, in the case of RVOT VTI and RVOTD in a patient with arrhythmogenic RV cardiomyopathy where echocardiography is done serially, the MDC can determine that the difference among measurements is due to measurement error or indicating an actual change. Identifying this cut point (for example, according to our study, approximately 3 cm and 0.3 cm for RVOT VTI and RVOTD, respectively) helps a physician know whether the disease has progressed, ameliorated, or stopped. This is true regarding critically ill patients in intensive care units where hemodynamic challenges and medical treatments are done. The changes more than MDC indicate that with 95% confidence, the change is true and is not due to measurement errors. The result of our study indicates that in multiple evaluations (for example patients with congenital disease) by echocardiography the calculation of RVSV in the distal part of RVOT cannot be replaced by the calculation of RVSV in MPA. In other words, if the first site of measurements is distal of RVOT, the second measurement should be done in this place. In addition, if the difference between the first and second measurements of RVOT VTI and RVOTD was more than the aforementioned value for MDC, it should be considered true. Also, our results can be applied in busy echocardiography laboratories (where usually only one measurement is done), because they provide a frame for interpreting the measured values.

The rate of poor-quality echocardiograms in our study was 13%. The previous studies reported a rate of poor echocardiograms of 10–15%, so the proportion of poor echocardiograms was in the acceptable reported range in our study [41, 42].

Firstly, although our study included a relatively large sample size, the selection of observers was not random, which could potentially introduce bias in the results. Secondly, the single-center design of the study limited the opportunity for comparing our findings with a gold-standard or invasive method. Thirdly, since LVSV was not calculated, we could not compare it with RVSV. Fourthly, since the RVOT and the MPA are not perfectly circular, employing 3D echocardiography or cardiac magnetic resonance imaging could provide more accurate cross-sectional area measurements. Fifthly, our study did not calculate interobserver variability for RVOTSV, MPASV, and related measurements, which might influence the overall reliability of the findings. Sixthly, while we provided indices of repeatability, reliability, and precision for basic measurements used to calculate RVSV at two sites, we were unable to compare calculated RVSV with cardiac magnetic resonance imaging or the thermodilution method, which could have served as valuable references. Seventhly, the lack of interobserver variability evaluation is another limitation in our study, as it could impact the generalizability of the results to other operators and settings. Eighthly, since observers were aware of their measurements during the study, the potential for bias should be considered when interpreting the results, as this awareness could have influenced the measurements and introduced observer-related errors. Finally, although the observers’ experiences were not largely different, they may affect our results.