Association of pulmonary transit time by cardiac magnetic resonance with heart failure hospitalization in a large prospective cohort with diverse cardiac conditions

There is an extensive body of work examining the central circulation time spanning nearly a century using a predominantly invasive approach in the classic literature [1‐7] and non-invasive modalities in the modern era [8‐17]. Most investigations have focused on the transit time between the right and left circulation, commonly referred as pulmonary transit time (PTT) where a longer duration has been associated with reduced cardiac function or stroke volume, as well as pulmonary hypertension and heart failure [8‐17]. To date, the prognostic value of PTT has begun to emerge in selected cardiac conditions [18‐22] in determining adverse clinical outcomes. As a biomarker, PTT is relatively easy to acquire. However, longer PTT duration is not specific to a particular cardiac disease as many conditions associated with altered cardiopulmonary hemodynamics can influence the duration of PTT including ventricular preload, blood volumes, global ventricular function, valvular function, and pulmonary diseases that affect the pulmonary circulation [1‐19]. Since all these factors may also contribute to the development of heart failure, we examined the long-term risk of heart failure hospitalization associated with longer PTT duration in a large prospective patient cohort with broad spectrum of cardiac conditions.

There were 506 consecutive subjects included in the analysis. The mean age was 63 ± 14 years and 373 (73%) were male. The mean LVEF was 49 ± 12% and there were 209 (41%) subjects with LVEF < 50%. Among diverse cardiac conditions identified by CMR the top three findings were dilated cardiomyopathy, myocardial infarction, and valvular heart disease (Table 1). After a mean follow up duration of 4.5 ± 3.0 years after their CMR exam, 70 (14%) subjects developed heart failure requiring hospitalization and among them 6 died.

We found that RR-interval normalized PTT ≥ 8.2 was the optimal cut point for predicting hospitalized heart failure. Among those with longer PTT (≥ 8.2), left and right ventricular chamber sizes as well as left atrial size were larger, stroke volume smaller and EF lower. The prevalence of infarct scar by LGE was higher. In contrast, the prevalence of non-infarct LGE was slightly lower although it was common in both groups (Table 2).

As shown in the Kaplan Meier curves, longer PTT (≥ 8.2) was associated with reduced survival during follow up compared to those with PTT < 8.2 (Fig. 2A, log-rank p < 0.001). The difference was observed consistently in subjects with preserved and reduced LVEF (Fig. 2B and C, log-rank p = 0.0242 and p = 0.0011, respectively). In the Cox proportional hazard analysis longer PTT (≥ 8.2) was associated with a hazard ratio [95% confidence interval (CI)] of 3.90 (2.32, 6.56) when compared with a lower PTT. In the fully adjusted model including covariates of age, gender, hypertension, diabetes, smoking history, hyperlipidemia, coronary artery disease and LVEF, longer PTT remained to be an independent risk factor for heart failure with a hazard ratio of 3.69 (2.02, 6.73). Longer PTT was associated with increased hazard for both preserved and reduced LVEF yet the hazard was greater among those with preserved LVEF [HR 3.59 (1.09, 11.79)] than those with reduced LVEF [HR 2.55 (1.42, 4.56)] in the stratified analysis. The difference remained in the fully adjusted model, HR 4.18 (1.02, 17.14) vs 2.91 (1.46, 5.80), respectively.

When assessing PTT as a continuous variable in the Cox proportional hazard analysis, each 1 unit or 1 heartbeat increase in PTT was associated with a hazard ratio of 1.26 (95% CI 1.18, 1.33) in the fully adjusted model (Table 3). In the LVEF-stratified analysis, the hazard was greater among those with preserved LVEF where each unit increase of PTT was associated with 70% increase in the hazard of heart failure admission compared to 18% increase among those with reduced LVEF (Table 3) in the fully adjusted model. After additionally adjusting for LV global longitudinal strain and NT-proBNP, PTT remained an independent predictor of hospitalized heart failure (Table 3). PTT calculated using the full width half maximum method was not an independent risk factor for heart failure hospitalization (data not shown).

We performed post hoc analyses to assess the incremental benefit of adding PTT to models containing known risk factors or predictors of heart failure including the estimates of cardiac function. Compared to reduced models containing baseline risk factors described above along with 1 additional measurement such as cardiac chamber size, EF, stroke volume, myocardial mass, LV global longitudinal strain, and NT-proBNP, adding PTT to these models significantly improved model fit and predictability (LRT p-values all < 0.05 and NRI demonstrating similar results). This was also true for a model with variables selected through the use of principal component analysis, whereby after adjusting for age, gender, LV end diastolic volume index, RV end systolic volume index, LA volume index, LV global longitudinal strain, and presence of LGE, PTT remained a significant risk factor for hospitalized heart failure with HR 1.15 (1.05, 1.25) per 1 unit increase and HR 1.92 (1.02, 3.61) for PTT ≥ 8.2 vs. PTT < 8.2.

In this prospective study, we demonstrated that longer PTT duration is an independent predictor of hospitalized heart failure in a large clinical cohort with a broad spectrum of cardiac conditions. Among those with longer PTT, there is significantly increased hazard of heart failure hospitalization with the greater hazard among those with LVEF ≥ 50%. We also demonstrated the incremental benefit of using PTT in models predicting hospitalized heart failure after considering not only traditional risk factors such as cardiac volume and function but also LV global longitudinal strain and NT-proBNP.

There is extensive literature examining the relationship of PTT with cardiomegaly, systolic dysfunction, pulmonary hypertension and heart failure due to reduced LVEF [8‐17]. PTT is also correlated with B-type natriuretic peptide [29]. Recently, longer PTT duration has been observed in heart failure patients with preserved LVEF [24] as it is closely associated with altered hemodynamics such as elevated pulmonary artery wedge pressure [17, 24]. As described previously, PTT consists of 3 major components including the transit times through the right heart, the pulmonary vasculature and the left heart where the transit time through the pulmonary vasculature accounts for the largest proportion of PTT [24]. Hence, a longer PTT duration represents the collective abnormality of blood flow transit in the cardio-pulmonary circulation including the cardiac chambers and the pulmonary vasculature. It has been shown that cardiovascular risk stratification can be improved significantly when pulmonary vascular response or simply pulmonary function is incorporated into the cardiac function evaluation [30]. For example, among patients with preserved LVEF, those with increased pulmonary wedge pressure during exercise are associated with significantly increased long-term mortality risk [30]. In a large epidemiological study, the subjects with both reduced LVEF and impaired pulmonary function were associated with a far greater risk of heart failure than those with reduced LVEF or impaired pulmonary function alone [31]. The LV diastolic dysfunction algorithm recommended by the American Echocardiography Society includes pulmonary pressure as one of the four essential evaluation criteria that has effectively differentiated the risk of clinical outcome above and beyond LVEF [32‐34].

There have been a number of small studies that reported outcome risk associated with longer PTT duration among patients with pulmonary hypertension [19] and those with congenital heart disease [18]. A recent publication examined a large cohort referred for chest pain evaluation using stress CMR and reported adverse clinical risk associated with longer PTT duration [20]. Similar observations were made recently among subjects with advanced heart failure [21] and with acute myocardial infarction [22]. In the present study we demonstrated in a large prospective clinical cohort with diverse common cardiac conditions that a longer PTT duration is associated with increased long-term risk of heart failure admission. Our work is an important validation of clinical relevance of assessing PTT in a general cardiac patient population thereby establishing generalizability for use with clinical cohorts. In addition, it highlights the importance of cardiopulmonary evaluation in the assessment of heart failure risk. PTT may have the potential to be complementary to routine CMR examination of cardiac structure, function and tissue characterization for heart failure risk assessment. While heart failure risk was increased independent of LVEF, the hazard ratio appears to be greater among those with preserved LVEF. We speculate that LVEF may act as an effect modifier on the relationship between PTT and heart failure risk [35]. When LVEF is reduced, LVEF presumably becomes the primary or sufficient cause of the HF hospitalizations outweighing the effect of PTT. When LVEF is preserved, PTT becomes the primary or sufficient cause of the HF hospitalizations. Our observation may have important clinical implication in the risk stratification of heart failure for patients with preserved LVEF as heart failure risk prediction for preserved EF is more complex than for those with reduced EF. In recent years, a number of powerful biomarkers have emerged in predicting heart failure risk including NT-proBNP [36, 37] and LV global longitudinal strain [38, 39]. To understand the clinical relevance of PTT, we tested the incremental value of PTT in the context of those well-established biomarkers and found PTT remains to be an important predictor of hospitalized heart failure.

Technically, PTT can be easily incorporated into routine clinical CMR examination, as the sequence of first pass perfusion is commercially available and the contrast needed is trivial. In our study we only used gadolinium contrast at 0.01 mmol/kg, a small fraction (5–6%) of the dose for late gadolinium enhancement imaging, which provides the signal intensity or image quality adequate for PTT evaluation. In addition, the time-intensity curve to calculate PTT is easily accessible using commercial software and does not require a proprietary program. Therefore, to incorporate PTT evaluation to routine clinical CMR study is feasible. We used sagittal views which are copied from the standard sagittal localizer available for all clinical CMR scans to alleviate the need for acquisition planning.

To conclude, longer PTT duration can predict long-term risk of heart failure admission independent of LVEF in patients with broad spectrum of cardiac conditions. We also demonstrated the incremental benefit of using PTT in models predicting hospitalized heart failure after considering not only traditional risk factors such as cardiac volume and function but also LV global longitudinal strain and NT-proBNP. Our findings underscore the value of cardiopulmonary circulation evaluation in the heart failure risk assessment.