Abstract
Mild pulmonary hypertension (as indicated by estimated right ventricular systolic pressure 30.0–39.9 mmHg) is associated with increased risk of all-cause mortality and a substantial component of premature mortality https://bit.ly/3ytwlEP
Introduction
Pulmonary hypertension is a chronic condition of increased blood pressure within the arteries of the lung due to multiple pathogenic causes [1]. Definitive diagnosis is currently predicated on mean pulmonary arterial pressure (mPAP) >20 mmHg measured via right heart catheterisation [2, 3]. Calculating the estimated right ventricular systolic pressure (eRVSP) by echocardiography based on measured tricuspid regurgitant velocity (TRV) represents a pragmatic/noninvasive means to identity potential cases of pulmonary hypertension prior to further investigation [4, 5]. A recent analysis of outcomes among 157 842 men and women captured by the National Echocardiography Database of Australia (NEDA) demonstrated that this more readily measurable parameter is independently correlated with mortality across the full spectrum of indicative pulmonary hypertension [6, 7]. These data also confirmed earlier reports (derived from disease-specific to larger patient cohort studies) that milder forms of pulmonary hypertension are indeed associated with a higher risk of mortality when compared with those with normal pulmonary arterial pressure [8–12].
Expert consensus statements currently recommend more definitive investigation if eRVSP is >40.0 mmHg or TRV is >2.8 m·s−1 in the absence of significant respiratory pathology [1, 5]. These thresholds (for more proactive management) are increasingly discordant with the scope and strength evidence [13], including the specific findings of the NEDA study [7], that suggest a higher than previously suspected risk of mortality associated with eRVSP <40.0 mmHg. Such findings would be less compelling (to change clinical practice) if the majority of deaths associated with milder forms of pulmonary hypertension 1) occurred in older individuals in whom life expectancy was already poor and/or 2) were linked to predominant forms of left heart disease (LHD) where mortality is already known to be elevated [14, 15]. A subset analysis of the original NEDA cohort suggested that this was probably not the case. Specifically, it demonstrated that due to a greater number of individuals affected overall combined with a significant (but still lower) component of premature mortality, milder forms of pulmonary hypertension (as indicated by eRVSP levels) are associated with a higher burden of premature life-years lost (LYL) relative to more severe cases [16].
To more definitively elucidate the association between eRVSP levels indicative of mild-to-severe forms of pulmonary hypertension with premature mortality and associated LYL, we analysed data from the now expanded NEDA cohort [14]. Specifically, we conducted a more granular analysis of the association between eRVSP levels determined by echocardiography and subsequent mortality in cases without evidence of LHD in order to determine 1) the overall pattern and risk of all-cause and disease-specific mortality associated with eRVSP levels above and below a pre-specified threshold of 30.0 mmHg (based on our previous research [7]), and 2) sex-specific patterns of premature mortality and subsequent LYL associated with different levels of eRVSP above and below this threshold.
Methods
Study design
As described previously, NEDA is a large observational cohort study that captures echocardiographic data from a network of centres across Australia [7, 14, 15, 17]. Individual data are combined using data linkage to derive long-term mortality outcomes [18]. With a diverse, multicultural population of approximately 26 million people, nearly all Australians have equitable (either free or subsidised) access to specialised management, including echocardiography [19]. At the time of this report, 23 participating centres contributed to the database and their patients are typically referred by a general practitioner and/or cardiologist to investigate or follow-up/manage pre-existing forms of cardiopulmonary disease. With standardised demographic profiling and routinely acquired indices of cardiac structure and function captured on all such cases, overall, NEDA represents a real-world cohort with minimal selection biases (other than localised patterns of clinical referral).
NEDA is registered with the Australian New Zealand Clinical Trials Registry with identifier number ACTRN12617001387314. Ethical approval was obtained from all relevant Human Research Ethics Committees and the study adheres to the Declaration of Helsinki [20].
Study cohort
As shown in figure 1, profiling data (as of January 2020) were used to identify all adult men and women aged >18 years who had at least one echocardiogram (data from the most recent echocardiogram was used if multiple investigations) captured by NEDA. As in previous reports [7, 14, 15, 17], only those individuals with both the primary variable(s) of interest (thereby reflecting real-world practice and negating the need to impute data) and with data linkage to mortality outcomes were considered for inclusion. With a specific focus on pulmonary hypertension, subjects with a calculable eRVSP were potentially eligible. Moreover, given the distinctive features and confounding of outcomes of those presenting with pulmonary hypertension due to LHD [21], applying the same criteria used in previous NEDA analyses, we specifically focused on those individuals without evidence of LHD [7]. Specifically, subjects were excluded if they had 1) left ventricular ejection fraction <55% [22], 2) signs of increased left ventricular filling pressure (manifesting in a ratio of mitral inflow E-wave peak velocity to peak early relaxation tissue Doppler velocity (E/e′) >12) [23], 3) left atrial volume index (LAVi) >34 mL·m−2 [22] and/or 4) moderate to severe mitral or aortic valve disease [24]. On this basis, a total of 154 956 eligible subjects (age 61.4±18.1 years) with a documented peak TRV to derive a valid eRVSP level were identified. Consistent with sex-based differences in the pattern of pulmonary hypertension [1], the proportion of men (48.8% versus 51.2%) and women (40% versus 60%) with and without a valid eRVSP level was markedly different (p<0.001). Alternatively, in both men (58.1±16.5 versus 64.2±16.5 years) and women (57.3±18.2 versus 64.6±17.9 years), those with a valid eRVSP were older compared with those without this parameter (p<0.001 for both comparisons).
Study data
All echocardiographic measurement and report data, including basic demographic profiling of subjects collected by participating centres during the period 1 January 2000 to 21 May 2019, were transferred into a central NEDA database. All data were then cleaned and transformed into standard NEDA format to generate uniform echocardiographic profiling data and to remove duplicate and/or impossible measurements/investigations. All subjects contributing to NEDA receive a unique identifier linked to their echocardiograms and their anonymity is protected by stringent security protocols [6].
Consistent with our previous reports [7], a consistent method was used to derive eRVSP by using the Bernoulli equation (eRVSP=4×(TRV)2+5 mmHg). A right atrial pressure of 5 mmHg approximates the average value recorded overall and removes any variation between laboratories. All eligible subjects with a calculated eRVSP derived from their last recorded echocardiographic examinations were included. The following thresholds of eRVSP indicative of increasing levels of pulmonary hypertension (mild-to-severe) were applied to create four main groups for initial comparisons: 1) normal/no pulmonary hypertension (eRVSP <30.0 mmHg), 2) mildly elevated (eRVSP 30.0–39.9 mmHg), 3) moderately elevated (eRVSP 40.0–49.9 mmHg) and 4) severely elevated (eRVSP ≥50.0 mmHg).
To derive all survival data, data linkage was performed via the (well-validated) National Death Index of Australia [18]. Specifically, reliable data on the survival status of subjects up to the study census (21 May 2019) were generated. Subsequently, with very low emigration rates, there was minimal loss to follow-up. If a subject had died, the listed causes of death were categorised according to International Statistical Classification of Diseases, 10th Revision (ICD-10) coding. Based on the primary cause of death, all ICD-10 Australian Modification [25] chapter codes in the range of C00–C97, I00–I99 and J00–J99 were categorised as a cancer-, cardiovascular- and respiratory-related death, respectively.
Study outcomes
Study outcomes were derived from a median (interquartile range (IQR)) follow-up of 5.7 (3.2–8.9) years. During this timeframe, we examined all-cause and disease-specific deaths (including respiratory and cardiovascular illnesses) occurring at the fixed time-points of 1 and 5 years, and at any time during follow-up, according to the four pre-specified eRVSP groups. We then conducted more granular analyses of the association between eRVSP levels (5 mmHg increments) from <30.0 to ≥60.0 mmHg (highest increment measured) with all-cause and cardiovascular-related mortality. Applying sex-specific life expectancy for the Australian population in 2020, premature mortality was defined as any death occurring below the age of 80.7 years in men and 84.9 years in women. If prematurely mortality did occur, the number of subsequent LYL was calculated by subtracting these age-specific thresholds with actual age (in years) at death.
Statistical analyses
NEDA analyses and reports conform to the relevant STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines [26]. All variables used in study analyses are without data imputation. Standard methods for describing and comparing continuous and grouped data, including mean with standard deviation and median (IQR) for normally and non-Gaussian distributed continuous variables, and proportions for categorical data according to baseline profiling were applied. Time zero for follow-up was set at the last recorded echocardiogram. Age and sex-adjusted odds ratios plus 95% confidence intervals for all-cause and disease-specific mortality at 1 and 5 years (150 062 and 99 372 cases, respectively, with complete follow-up at these timeframes) according to the four pre-specified eRVSP groups indicative of no (reference group) versus increasing levels of pulmonary hypertension were derived from multiple logistic regression (entry model). The Kaplan–Meier method followed by Cox proportional hazard models (entry method) was used to derive adjusted hazard ratios and 95% confidence intervals for the risk of all-cause and cardiovascular-related mortality during the entire period of follow-up when also adjusting for age and sex according to 1) the four pre-specified eRVSP groups and 2) each 5 mmHg increment in eRVSP above 30.0 mmHg (reference group). In a more granular, sensitivity analyses of mortality above and below this threshold (using the same methods), the reference eRVSP group was 30.0–31.9 mmHg. Multiple logistic regression (entry models) was also used to calculate the age-adjusted risk of premature mortality for men and women separately, according to 5 mmHg increments in eRVSP above 30.0 mmHg (30.0–34.9 mmHg; reference group). We used the comparative risk assessment method to then calculate the population attributable risk (PAR) and associated PAR% for each discrete eRVSP group [27]. All statistical analyses were performed using SPSS version 26.0 (IBM, Armonk, NY, USA). Statistical significance was accepted at a two-sided α=0.05.
Results
Study cohort
Overall, the study cohort comprised 70 826 men (45.7%) and 84 130 women (54.3%) with a similar age profile (61.3±17.7 and 61.4±18.4 years, respectively). Just over half (85 173/154 956 cases (55%)) had normal eRVSP <30.0 mmHg indicative of no pulmonary hypertension. Alternatively, 49 276 (31.8%), 13 060 (8.4%) and 7447 (4.8%) had mildly, moderately and severely elevated eRVSP levels, respectively.
Table 1 summarises the demographic and echocardiographic characteristics of the cohort on a sex-specific basis and according to the four pre-specified eRVSP levels. Overall, mean age rose steadily with increasing eRVSP (range 56–71 years). A dilated right ventricle and impaired right ventricular function were documented in 10 618 (9.6%) and 1972 (1.8%) cases, respectively. The prevalence of impaired right ventricular function was associated with increasing eRVSP (0.7%, 1.4%, 4.6% and 12.3% in those with eRVSP <30.0, 30.0–39.9, 40.0–49.9 and ≥50.0 mmHg, respectively; p<0.001). Compared with those with eRVSP <30.0 mmHg, those with eRVSP indicative of mild pulmonary hypertension (30.0–39.9 mmHg) had a higher prevalence of right ventricular dilation (8.7% versus 4.6%; p<0.001). Minor increases in left and right atrial volumes along with markers of left ventricular filling pressure were also noted with increasing eRVSP.
Age and sex-specific risk of mortality
Table 2 summarises the overall pattern of all-cause mortality according to the four pre-specified eRVSP groups. As expected, both absolute and age and sex-adjusted risk of mortality steadily increased with higher eRVSP levels. This was evidenced by the large differential in actual 1- and 5-year mortality (3.9% and 16.7%, respectively) in those with eRVSP <30.0 mmHg compared with those with eRVSP ≥50.0 mmHg (32.5% and 74.5%, respectively). Figure 2 shows the age and sex-adjusted survival curves for all-cause mortality over the longer term according to eRVSP levels. When examined on a more granular basis (5 mmHg increments), those with eRVSP 35.0–39.9 mmHg were almost twice as likely to die from all causes (HR 1.90, 95% CI 1.84–1.96) and cardiovascular disease (HR 1.85, 95% CI 1.74–1.97) when compared with those with eRVSP <30.0 mmHg (p<0.001 for both comparisons) (figure 3). This associated risk of mortality rose markedly among those with eRVSP ≥50.0 mmHg (HR 4.79, 95% CI 4.57–5.02 and HR 5.63, 95% CI 5.20–6.11 for all-cause and cardiovascular-related mortality, respectively) during long-term follow-up. Additional granular assessments of the age and sex-adjusted risk for all-cause mortality in those cases with eRVSP 10 mmHg above and below the selected threshold of 30.0 mmHg reconfirmed that this level was a natural, if not conservative, reference point for survival analyses (supplementary figure S1).
Sex-specific pattern of mortality
Figure 4 shows the pattern (overall and cause-specific) of increasing mortality associated with each 5 mmHg increase in eRVSP above 30.0 mmHg on a sex-specific basis. Overall, the proportional contribution of malignancy-related deaths declined from 22.6–27.6% of deaths in men and women with eRVSP <30.0 mmHg to around half (10.5–12.7% of deaths) in those with the highest eRVSP levels. Alternatively, for both men and women, the absolute frequency and proportional contributions of respiratory-related deaths (from ∼9.0% to 16.9–23.6%) and cardiovascular-related deaths (from 25.2–28.2% to 33.5–40.1%) rose markedly with increasing eRVSP levels.
Premature mortality
Overall, 54% of men and 55% of women died prematurely. Figure 5 shows the age-adjusted risk for premature mortality among those cases with eRVSP ≥30.0 mmHg on a sex-specific basis. As expected, each 5 mmHg increment in eRVSP was associated with increasingly more premature mortality as a proportion of all deaths. Accordingly, premature mortality occurred in 46.7–79.2% of all deaths among those cases with eRVSP 30.0–34.9 mmHg (reference group) versus those with the highest eRVSP levels (≥60.0 mmHg). Within the entire cohort, 34% of premature deaths were cancer-related (mean age at death 70.9 years) and 22% of premature deaths were cardiovascular-related (74.0 years). However, the distribution of cause-specific contributions to premature mortality changed with rising eRVSP levels. Among cases with eRVSP ≥60.0 mmHg, premature mortality was predominantly attributable to cardiovascular (34% of deaths with a mean age at death of 70.2 years) and respiratory illnesses (25%, 71.5 years). Overall, for every 1000 cases at risk, the rate of premature mortality increased by three (0.5%), 32 (6.2%) and 53 (9.8%) cases for those with eRVSP 30.0–39.9, 40.0–49.9 and ≥50.0 mmHg, respectively, compared with those with eRVSP <30.0 mmHg.
Life-years lost
Figure 6 shows the relationship between increasing eRVSP levels and LYL due to premature mortality among the 11 607 men and 13 588 women with eRVSP ≥30.0 mmHg. Overall, a total of 158 587 LYL were accumulated by these cases, comprising 70 019 LYL among men and 88 569 LYL among women. As expected, the average LYL due to premature mortality positively correlated with increasing eRVSP levels, rising from a mean of 5.4 to 11.4 LYL and 5.1 to 10.4 LYL among men and women, respectively, associated with eRVSP 30.0–34.9 to ≥60.0 mmHg. However, due to a much higher number of affected cases, those with eRVSP 30.0–39.9 mmHg accounted for 58% (40 606/70 019) of total LYL among men and 53% (47 333/88 568) of total LYL occurring within the broader group of cases with eRVSP ≥30.0 mmHg indicative of mild-to-severe pulmonary hypertension.
Discussion
In our study of 154 956 individuals referred for routine echocardiography, we confirmed that milder forms of pulmonary hypertension (based on indicative eRVSP levels and in the absence of significant LHD) are associated with an increased risk of morality. We then confirmed, for the first time, that this phenomenon is associated with a significant component of premature mortality and associated LYL in both sexes. Specifically, above a clear inflection point indicative of no versus mild pulmonary hypertension, eRVSP 30.0–34.9 mmHg was associated with a 38% increase in the age and sex-adjusted risk of all-cause mortality over the longer term compared with a normal eRVSP. This specific finding (when applying eRVSP 30.0 mmHg as our reference point for all comparisons) is consistent with previous analyses of an earlier iteration of the NEDA cohort [7]. These findings suggest that the current echocardiographic thresholds for defining pulmonary hypertension (eRVSP 40 mmHg which approximates to mPAP 25 mmHg) do not yet fully capture clinical risk related to those presenting with mildly elevated eRVSP. Of relevance, >50% of deaths were premature among those with eRVSP ≥30.0 mmHg and this generated a significant component of LYL. This was particularly true for those cases with eRVSP levels indicative of mild pulmonary hypertension (30.0–39.9 mmHg) who contributed to more than half the total number of LYL associated with eRVSP ≥30.0 mmHg. Collectively, noting the exclusion of cases with LHD, our findings suggest links between premature mortality and pulmonary hypertension not only with advanced disease states associated with impairment of cardiac (right ventricular) haemodynamics, but also with earlier, subclinical stages within the natural history and progression of pulmonary hypertension in affected individuals.
Our findings are consistent with a large, well-characterised patient cohort from Huston et al. [12], who demonstrated that the increased risk of clinical events among patients with mild pulmonary hypertension is not driven solely by an increased burden of comorbidities. Rather it represents a pathological response of the right ventricle to increasing pulmonary pressure. Unfortunately, since we do not have complete clinical data, it is unclear if patients in this study died specifically from mildly elevated pulmonary pressure, due to other concomitant conditions and/or subsequent development of LHD (after baseline exclusion of significant LHD, higher eRVSP was associated with small increases in LAVi and E/e′ ratio). Alternatively, we were able to specifically analyse eRVSP as a continuous variable to determine at what haemodynamic pressure the risk of mortality increases. Subsequently, we have identified substantial risk at eRVSP levels that would traditionally be considered as normal or of no clinical concern. Our finding of increased adjusted mortality risk starting at eRVSP 30.0 mmHg is consistent with our previous NEDA report [7]. These findings are also consistent with equivalent studies using echocardiographic estimates of pulmonary pressure in populations at high risk of pulmonary hypertension [12, 28]. Although the gold standard to accurately measure right ventricular haemodynamics is by right heart catheterisation [2], this procedure is invasive and its potential complications make it unsuitable for screening or first-line evaluation of pulmonary hypertension. Accordingly, the role of echocardiography in evaluating such patients with pulmonary hypertension is well established. Our data reaffirm the value of echocardiography to inform the evidence-based clinical management of pulmonary hypertension [29].
To the best of our knowledge, there is a paucity of data describing echocardiographic pulmonary pressure estimates and examining the link between mildly elevated eRVSP and premature mortality at both the population and clinical cohort level. However, our findings, derived from a large unselected clinical cohort, suggest that even modest increases in eRVSP are associated with a significant rise in premature deaths and considerable potential for LYL without active intervention. Our data are consistent with similar studies in systemic hypertension [30], in which minor elevations in systemic blood pressure have profound implications on LYL when applied across an entire cohort. Despite the inherent selection bias of being investigated with echocardiography, our findings suggest a significant group of individuals within the general population who are adversely affected by milder forms of pulmonary hypertension and remain undiagnosed and treated. Although consistent with the current therapeutic focus on patients with severe forms of pulmonary hypertension, in whom the mean LYL was highest, we found that individuals with eRVSP 30.0–39.9 mmHg (representing the highest proportion of cases) accumulated more than half of the total LYL within the overall cohort. Moreover, >50% of deaths were premature and, for many individual cases, were associated with a significant component of LYL. On this basis, if targeted treatments can slow disease progression towards right heart failure and the subsequent clinical sequalae, early more aggressive management of mild-to-moderately elevated pulmonary pressure [31] could potentially yield enormous health benefits and substantial reductions in premature mortality within a variety of high-risk clinical populations [16].
Limitations
We acknowledge that our study reports outcomes in a cohort of subjects being investigated for possible/pre-existing cardiopulmonary conditions referred for echocardiography, which may not be generalisable for the wider population. By virtue of our de-identified NEDA electronic record interface, we were unable to directly review echocardiographic images related to pressure estimates or other cardiac functional parameters and this is a methodological drawback. As highlighted by a recent analysis of tricuspid regurgitant gradient in predicting pulmonary hypertension in clinical practice, there is a critical need to consider all echocardiographic and clinical factors in evaluating the probability of underlying pulmonary hypertension [32]. As such, we relied on the accuracy of data input by physicians into echocardiographic reports and the accuracy of ICD-10 coding of cause of death. While NEDA can capture detailed echocardiographic data with reliable individual linkage to long-term mortality, at the time of preparing this article, it has yet to capture some important clinical details pivotal to health outcomes. These include an individual's clinical comorbidities, pattern of hospital episodes, pharmacological treatment and surgical management. NEDA also lacks potentially important socioeconomic variables such as income and occupation (although access to the healthcare system is subsidised for lower socioeconomic groups). While we have excluded subjects with echocardiographic evidence of LHD, we were unable to completely exclude minor valvular disease that might develop further. Given that the absence of a TRV does not exclude pulmonary hypertension [33], our estimates of the prevalence and prognostic impact of pulmonary hypertension should be interpreted as the minimum indicative prevalence from a clinical cohort perspective. While we were able to confirm that those without a calculable eRVSP represent a lower risk group overall (supplementary figure S2), our findings reinforce the need for routine documentation of the TRV and eRVSP. Moreover, we relied on the most recently recorded eRVSP for our outcome analyses. Using data from the 37.1% of men and 32.4% of women with multiple echocardiograms, we plan a future analysis of the prognostic importance of the rate of change in eRVSP over the longer term. As shown in supplementary figure S3, in a sensitivity analysis of those cases with only one recorded echocardiogram, we found the same pattern of mortality according to eRVSP levels. With limited clinical data available and the absence of pulmonary vascular resistance information, we were unable to identify the specific causes of elevated eRVSP and the distinct type of pulmonary hypertension (including pulmonary arterial hypertension) present. Nevertheless, consistent with an overall increased risk of mortality among those presenting with eRVSP ≥30 mmHg, it has been recently shown that patients presenting with mild pulmonary arterial hypertension associated with relatively low pulmonary vascular resistance still have poor outcomes that may be amenable to treatment [34]. Finally, we chose 5 mmHg as the most representative right atrial pressure across the NEDA cohort to avoid variation across readers and laboratories. This is unlikely to have resulted in underestimation of our identified eRVSP risk threshold around 30 mmHg, since the most frequently allocated American Society of Echocardiography guideline-directed right atrial pressure estimation is lower than our estimate at 3 mmHg [22].
Conclusions
This large real-world echocardiographic database study points to a high mortality burden and consequential premature deaths in individuals routinely presenting with mildly elevated eRVSP. Our findings support the contention that even subclinical pulmonary hypertension has an extensive clinical impact. Specifically, we propose increased clinical risks starting at eRVSP levels around 30 mmHg and recommend early monitoring from treating clinicians with efforts to modify risk factors and improve outcome weighted against the likely increased economic burden of additional screening and increased referrals of advanced pulmonary hypertension. Furthermore, more granular work is warranted to determine if early aggressive management of risk factors in individuals with mildly elevated eRVSP can significantly increase survival and reduce a high burden of premature mortality and associated LYL.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary figure S1. Adjusted age and sex-adjusted risk of all-cause mortality above and below an eRVSP of 30.0 to 31.9 mmHg. Age and sex-adjusted risk (hazard ratios shown in top left corner; p<0.001 for both) of all-cause mortality 10 mmHg above and below the reference group (REF/green shade) eRVSP 30.0 to 31.9 mmHg (n=129 581) according to 2-mmHg eRVSP groups. Actual 1- and 5-year mortality shown on the x-axis for each group. ERJ-00832-2021.Figure_S1
Supplementary figure S2. Age and sex-adjusted all-cause mortality per eRVSP group (assuming a non-recorded eRVSP equals <30.0 mmHg). The main graph shows age and sex-adjusted (hazard ratios shown in top right corner; p<0.001 for both) plots for all-cause mortality during long-term follow-up shown for an expanded reference group of 273 425 cases (this includes those cases with no eRVSP calculable but are assumed to have an eRVSP <30.0 mmHg) and no evidence of LHD versus the three other existing/pre-specified eRVSP groups. Overall numbers at risk (in 3-year intervals) and mortality rates during these specific intervals are shown below and above the x-axis. Compared to the new reference group, the HR for all-cause mortality (p<0.001 for all) associated with an eRVSP of 30.0–39.9, 40.0821140.9 and ≥50.0 mmHg was 1.38 (95% CI 1.33–1.422), 1.46 (95% CI 1.39–1.54) and 2.33 (95% CI 2.23–2.45), respectively. ERJ-00832-2021.Figure_S2
Supplementary figure S3. Adjusted age and sex-adjusted risk of all-cause mortality above and below an eRVSP of 30.0 to 31.9 mmHg (cohort with one echocardiogram only). Age and sex-adjusted risk (hazard ratios shown in top left corner; p<0.001 for both) of all-cause mortality 10 mmHg above and below the reference group (REF/green shade) eRVSP 30.0 to 31.9 mmHg (n=129 581) according to 2-mmHg eRVSP groups are shown for the 98 074 cases with only one echocardiogram. Actual 1- and 5-year mortality shown the x-axis for each group. ERJ-00832-2021.Figure_S3
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-00832-2021.Shareable
Footnotes
This article has supplementary material available from erj.ersjournals.com
This article has an editorial commentary: https://doi.org/10.1183/13993003.02064-2021
Author contributions: S. Stewart, G.A. Strange and D. Playford conceived and designed the study. G.A. Strange and D. Playford acquired data as part of the overall NEDA study. Y-K. Chan and S. Stewart analysed the study and drafted the first version of the manuscript. All authors critically revised the manuscript and approved the final version for publication.
Conflict of interest: S. Stewart reports a Senior Principal Research Fellowship from the NHMRC Australia, consultancy fees from NEDA, and honoraria for presentations from Novartis Pharmaceuticals, outside the submitted work.
Conflict of interest: Y-K. Chan has nothing to disclose.
Conflict of interest: D. Playford reports an investigator-initiated grant from Johnson & Johnson, during the submitted work.
Conflict of interest: G.A. Strange reports an investigator-initiated grant from Johnson & Johnson, during the submitted work.
Support statement: This study was supported by an investigator-initiated grant from Johnson & Johnson. No commercial entity had any insight into the conception, analysis or writing of this article. S. Stewart is supported by the NHMRC Australia (GNT1135894). Funding information for this article has been deposited with the Crossref Funder Registry.
- Received March 21, 2021.
- Accepted May 16, 2021.
- Copyright ©The authors 2022. For reproduction rights and permissions contact permissions{at}ersnet.org