Long-term follow-up in outpatients with mildly elevated pulmonary artery systolic pressure on echocardiography: a single-centre retrospective cohort study in Shanghai, China

Introduction

Echocardiography is the most frequently used technique and mainstay screening for pulmonary hypertension (PH). Doppler echocardiography is a non-invasive, widely available imaging technique that estimates pulmonary artery systolic pressure (PASP), most commonly by estimating the transtricuspid gradient from regurgitant jet velocity.1 However, right heart catheterisation (RHC) may still be required to confirm the diagnosis owing to the low positive predictive value of echocardiography. Previous studies have shown that echocardiographic measurements of elevated PASP were found in approximately 40% of patients,2 3 while the prevalence of PH was only 1% in the general population and 10% in those over 65 years of age.1 RHC remains an invasive procedure, which limits its utility in estimating pulmonary pressure in the general population or in conducting large-scale epidemiological studies. Hence, echocardiography is currently the most established non-invasive method for achieving these objectives.

Previous studies have primarily focused on the correlation between echocardiographic PASP measurements and RHC.4 Only a few studies have investigated the long-term survival of patients with mildly elevated PASP using echocardiography,4–7 and no studies have investigated the long-term changes of PASP on echocardiography. It is important to investigate the long-term prognosis and changes in PASP in patients with mildly elevated PASP by using echocardiography; hence, this study aimed to investigate the correlation between mildly elevated PASP on echocardiography and mortality, as well as long-term changes in PASP.

Materials and methods

Patient population and follow-up

This study included patients with mild echocardiographic PH (PASP, 40–49 mm Hg) who were examined at our clinic between January and June 2016. The baseline characteristics and follow-up information of patients were obtained from the electronic medical records system of our hospital. Between January and June 2016, 1869 patients underwent echocardiography at our outpatient clinic. Patients with malignant tumours, previous heart or other solid organ transplantation, previous or scheduled ventricular assist device implantation, severe kidney dysfunction (uraemia and patients on dialysis) and a life expectancy of less than 1 year for any medical condition were excluded. We then enrolled 1205 consecutive patients and referred them for follow-up studies. After excluding patients with incorrect phone numbers and those who refused to provide information over the phone, a total of 910 patients were registered. A flowchart of the patient selection process is shown in figure 1. Patients were grouped according to whether they had combined organic heart disease (OHD), which was defined as significant heart structural changes, including clinically significant valvular disease (moderate or greater aortic and/or mitral valve disease), cardiomyopathy, congenital heart disease (before or without surgical or interventional treatment), ischaemic cardiomyopathy (abnormal myocardial segment contraction, left ventricular ejection fraction (LVEF) <55%, left ventricular end-diastolic internal diameter (LVEDD) >56 mm), pericardial diseases and all cardiac tumours. We determined whether the patients had died and adjudicated the cause of death using telephone follow-up and review of hospitalisation records. The study was approved by the Biomedical Research Ethics Committee of Zhongshan Hospital affiliated with Fudan University. The ethics committee waived the requirement for informed consent for this retrospective analysis due to the use of anonymised and de-identified data; in addition, no interventions were done to the study population.

Figure 1
Figure 1

Patient selection flowchart. PH, pulmonary hypertension.

Clinical data collection

Data were collected retrospectively and obtained from our hospital. Age, sex and significant medical history, such as hypertension, diabetes mellitus, atrial fibrillation (AF) and coronary artery disease (CAD), were all included. The echocardiographic data (data from the most recent echocardiogram was used if multiple investigations) of patients from January 2021 to December 2022 were obtained, encompassing diagnostic information, PASP, LVEF, left atrial internal diameter, LVEDD and left ventricular end-systolic internal diameter.

Echocardiography

Two-dimensional transthoracic echocardiography (Philips Sonos 5500 or IE33; GE Vingmed Ultrasound AS; Horten, Norway) was also performed. Standardised echocardiographic examinations included apical four-chamber, parasternal right ventricular inflow, parasternal short-axis and subcostal views. PASP was derived from right ventricular (RV) systolic pressure estimates using the tricuspid regurgitation velocity (V) and the Bernoulli equation as follows: 4V² + right atrial pressure (5 mm Hg if complete, 10 mm Hg if partial and 15 mm Hg if there was no collapse on inspiration).8 PH was defined as PASP ≥40 mm Hg, wherein PASP ≥40 mm Hg but <50 mm Hg was considered as mild PH; PASP ≥50 mm Hg but <70 mm Hg was considered as moderate PH; and PASP ≥70 mm Hg was considered as severe PH. The RV systolic function was assessed using the tricuspid annular plane systolic excursion (TAPSE), wherein TAPSE <17 mm indicated RV systolic dysfunction. Right atrial enlargement was defined as a left-right diameter >44 mm or a superior-inferior diameter >53 mm. Pulmonary artery widening was defined as an inner diameter of >22 mm. All echocardiographic measurements were performed in accordance with American Society of Echocardiography recommendations and interpreted by trained echocardiographers.

Statistical analysis

Stata 17.0 was used for the statistical analyses, and GraphPad Prism10 (GraphPad Software, San Diego, CA) was used for mapping. Continuous variables were assessed for normality and presented as mean±SD if normally distributed. The Shapiro–Wilk test was used to assess normality. Categorical variables were presented as frequencies (percentages) and were compared using the chi-square test. Survival curves were constructed using the Kaplan–Meier method and compared using the log-rank test. The significance level for intergroup comparisons was corrected using the Bonferroni method. The predictors of death in patients with mild echocardiographic PH were analysed using univariate and multivariate Cox regression analyses. Paired t-tests were used to calculate changes in the PASP values at baseline and follow-up for different patient groups. Statistically, significance was set at a two-sided p<0.05.

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting or dissemination plans of this research.

Results

Baseline data

Of the 910 patients with mild PH, 65.5% (596 of 910) were categorised as patients without OHD and 34.5% (314 of 910) were categorised as patients with OHD. Table 1 lists the baseline patient characteristics. The patients without OHD group had a significantly higher mean (SD) age (65.3±14.1 vs . 61.7±14.2 years; p<0.001) and LVEF (66.5±4.5% vs 60.8±10.85%; p<0.001). Right atrial enlargement (10.6% (65 of 596) vs 22.6% (71 of 314)), left ventricular dysfunction, widening of the pulmonary artery (14.3% (45 of 314) vs 6.4% (38 of 596)), and moderate or severe tricuspid regurgitation (TR) (19.7% (49 of 314) vs 4.4% (21 of 596)) were more likely to develop in the patients with OHD group compared with the patients without OHD group. The AF, CAD and left ventricular systolic and diastolic performances differed significantly between the groups (all p<0.05). No differences were found in hypertension, diabetes, cardiac surgery or PASP (42.9±2.5 mm Hg vs 42.7±2.3 mm Hg; p=0.43) between the different patient groups.

Table 1

Baseline characteristics among the different groups

Survival rates among the different groups

Among the 910 patients who followed-up, 62 (6.8%) died within 5 years. The patients with OHD group showed a significantly higher 5-year mortality rate than the patients without OHD group (10.8% (34 of 314) vs 4.7% (28 of 596); p<0.001) in the Kaplan–Meier analysis (figure 2). There were four cardiac deaths in the patients without OHD group (4 of 28) compared with the 19 cardiac deaths in the patients with OHD group (19 of 34).

Figure 2
Figure 2

Survival rates of patients with mild echocardiographic pulmonary hypertension among the different patient groups. OHD, organic heart disease.

Predictors of all-cause mortality

We analysed different groups using univariate (online supplemental table 1) and multivariate COX regression analyses (table 2) and compared the different factors affecting survival outcomes between the different groups. The multivariable regression analysis of the total patient group did not show PASP value as a risk factor for all-cause mortality (HR: 0.97; 95% CI: 0.87–1.08; p=0.581). For patients without OHD, age (HR: 1.13; 95% CI: 1.08 to 1.18; p<0.001), widening of the inferior vena cava (HR: 22.62; 95% CI: 2.89 to 176.78 p=0.003) and widening of the pulmonary artery (HR: 3.56; 95% CI: 1.30 to 9.74; p=0.013) were significant independent predictors of all-cause mortality. For patients with OHD, the multivariate Cox regression analysis showed that age (HR: 1.08; 95% CI: 1.04 to 1.11; p<0.001), PASP value (HR: 1.02; 95% CI: 1.01 to 1.03, p=0.038), LVEF (HR: 0.96; 95% CI: 0.93 to 0.99, p=0.008) and RV systolic dysfunction (HR: 17.18; 95% CI: 3.73 to 79.03; p<0.001) were the independent predictors of all-cause mortality.

Supplemental material

Table 2

Multivariate Cox regression models for mortality of all causes

PASP changes after 5–6 years of follow-up

A fraction of the patients (46% (419 of 910)) continued to undergo echocardiography at our hospital. Of the 419 patients, 281 (67%) were patients without OHD and 138 (33%) were patients with OHD (table 3). Patients without OHD had an average age of 65.1±11.7, while patients with OHD were slightly younger at 63.2±12.5 years. No differences were found in sex, diabetes, AF or baseline PASP (42.8±2.4 mm Hg vs 42.8±2.5 mm Hg; p>0.05) between the different patient groups. Pulmonary artery widening was more likely to occur in the patients with OHD group than in the patients without OHD group. After 5–6 years, 36.6% of the patients without OHD group had a mild PASP elevation, whereas, 10.0% had a moderate or severe elevation. In the patients with OHD group, 55.8% of the patients had mild, 13.3% had moderate or severe, and 2.2% had severe elevation. A slight reduction in PASP was shown in the patients without OHD group (42.8±2.4 mm Hg vs 39.3±8.2 mm Hg; p=0.000), but no change was seen in the patients with OHD group (42.8±2.5 mm Hg vs 42.4±8.8 mm Hg, p=0.339). In addition, 39 (28.3%) patients in the OHD group underwent cardiac surgery during follow-up. Medication use was recorded during the follow-up period. Among the 281 patients without OHD group, 161 (57.2%) were treated with beta-blockers, 115 (40.9%) with diuretics, 111 (39.5%) with calcium channel blockers, 90 (32.1%) with angiotensin receptor blockers, 70 (24.9%) with oral anticoagulants and 4 (1.4%) with targeted PH therapy (online supplemental table 2).

Table 3

Baseline characteristics and 5–6 years of follow-up among the different subgroups

We further analysed changes in PASP in the patients without OHD by categorising them into three age groups: ≤50 years (young), 50–65 years (middle-aged) and >65 years (older adults) (online supplemental table 3). The results indicated no significant differences in baseline PASP across age groups (42.5±2.7 mm Hg for ≤50 years, 42.5±2.4 mm Hg for 50–65 years and 43.2±2.4 mm Hg for >65 years; p>0.05). Significant differences were observed in sex, hypertension prevalence and history of cardiac-related procedure between the groups (p<0.05). Notably, younger patients had a significantly higher history of cardiac-related procedures than the other age groups. After 5–6 years of follow-up, the PASP values remained statistically similar across age groups (38.0±10.2 mm Hg for ≤50 years, 39.1±8.3 mm Hg for 50–65 years and 40.1±8.0 mm Hg for>65 years; p>0.05).

Discussion

Prior studies have shown a connection between slightly elevated PASP and negative outcomes, such as RV dysfunction2 and premature death.5 However, more research is needed to identify which patients with mildly elevated PASP levels are at risk of these outcomes. This real-world study categorised patients based on the presence or absence of OHD. We found that PASP values on echocardiography were not associated with all-cause mortality in patients without OHD despite having mildly elevated PASP levels. Furthermore, after 5–6 years of follow-up, PASP on echocardiography was not further elevated in patients without OHD.

Tan et al
6 evaluated long-term survival in patients with suspected heart failure and an incidental finding of PH on echocardiography. The results revealed no significant difference in mortality between the mild and no PH groups, and mild PH was not an independent predictor of mortality. Our study provides supplementary insight into these findings. In our study, a multivariable regression analysis of the total patient group did not show mild PH as a risk factor for all-cause mortality (HR: 0.97; 95% CI: 0.87 to 1.08; p=0.581). However, stratifying individuals based on the presence or absence of OHD showed that mildly elevated PASP value was an independent predictor of all-cause mortality in patients with OHD (HR: 1.02; 95% CI: 1.01 to 1.03, p=0.038), which is consistent with those of previous studies. Even a mild elevation in PASP could be an independent predictor of all-cause mortality in patients with valvular disease or in those who have undergone surgery for valve disease.9–11 Pressure elevation is unlikely to be the cause for impaired survival, but it reflects an underlying pathology that leads to both increased pulmonary pressure and poor prognosis.12 The finding of increased PASP on echocardiography should not be trivialised, and further clinical evaluations or surgical interventions are justified. After a 5–6-year follow-up, PASP on echocardiography was not further elevated in patients with OHD (42.8±2.5 mm Hg vs 42.4±8.8 mm Hg; p=0.339). This may be because many patients with OHD accept active treatments, such as surgery for valvular heart disease and interventions for congenital heart disease. Our subgroup analysis included only surviving patients, despite the fact that this population may have had a lower PASP than those who died.

For patients without OHD, the PASP on echocardiography was slightly reduced after 5–6 years of follow-up (42.8±2.4 mm Hg to 39.3±8.2 mm Hg; p<0.001). This work is the first evidence supporting this finding. Patients with mildly elevated PASP are not infrequently observed clinically, accounting for approximately 30%–40% of patients with echocardiography,4 5 which was observed in the present study. Previous studies have found that elevated PASP is associated with increasing age, female sex, systemic pulse pressure, diabetes, obesity, chronic lung disease and age.4–6 However, elevated PASP does not necessarily represent obvious heart disease or progressive pulmonary hypertension. Notably, this was a retrospective study based on the database from our institution, and half of the patients did not have an available 5–6-year follow-up echocardiographic data; hence, a bias may have been present. The correlation between PASP measured by echocardiography and RHC-derived data was not significant, especially in patients with mild echocardiographic PH.7 13 Therefore, when assessing the probability of PH, the measurement of TR velocity should be used in conjunction with other echocardiographic markers of PH, including the ventricles, pulmonary artery, inferior vena cava and right atrium.1 Finally, the medications administered to the patients after referral to our hospital may have been another reason for the findings. Half of the patients were taking beta-blockers, diuretics and calcium channel blockers, which may reduce PASP by reducing cardiac output or volume.

In patients without OHD but with mildly elevated PASP levels, our study found that PASP values on echocardiography were not associated with all-cause mortality. Only 10.0% (28 of 281) of the patients had PASP ≥50 mm Hg and 0.71% (2 of 281) of the patients had PASP ≥70 mm Hg after 5–6 years of follow-up. These results indicate that mildly elevated PASP in patients without OHD is not a poor predictor of prognosis, and we may not need to overly worry about these patients nor do we need to take proactive interventions, such as RHC and frequent follow-up. The PASP remains a single parameter that should always be interpreted in the context of all haemodynamic variables and the clinical picture. All relevant clinical circumstances, including age, comorbidities and various haemodynamic and clinical parameters, should be considered before making diagnostic and treatment choices.14 However, we found that widening of the pulmonary artery was associated with all-cause mortality and elevation of PASP on echocardiography at 5–6 years of follow-up, which is also consistent with previous research.1 2 15 This indicates that for patients without OHD but with a mildly elevated PASP, more attention should be paid to widening of the pulmonary artery.

Limitations

This study had several limitations. First, our study relied solely on telephone follow-ups, which has the potential for underestimating mortality rates. Second, more than 20% of the patients were not enrolled because they had no valid phone number or refused follow-up. However, the included patients did not differ significantly in age, baseline PASP or LVEF compared with the excluded patients, but they were more likely to have AF, history of pacemaker implantation, previous cardiac surgery or organic heart disease, as shown in online supplemental table 4. Moreover, this was not a prospective cohort study, and there is a potential for selection bias in echocardiographic follow-up, which may have been influenced by a number of factors, including patient clinical status. Only 46% (419 of 910) of the patients had a greater than 5 years of follow-up echocardiography. This could have resulted in bias in the analysis. Finally, our study population was predominantly the elderly, which restricts the generalisability of our findings to younger populations.

This post was originally published on https://bmjopen.bmj.com