Introduction

Pulmonary hypertension (PH) is currently defined by a mean pulmonary arterial pressure (mPAP) > 20 mmHg at rest. Pulmonary arterial hypertension (PAH) is a subtype of PH (WHO Group 1 PH) characterized by pulmonary arterial remodeling [1], which may present with underlying diseases [2]. PAH is not only a dreaded complication of connective tissue diseases (CTDs), but also a major cause of morbidity and mortality in CTDs [3]. It is reported that around 30% of scleroderma-related deaths are due to PAH [4]. In a France national multicenter PAH registry study, CTD-PAH counted for 15.3% in the PAH cohort [5]. PAH-CTD is the second most prevalent type of PAH in Western countries, and systemic sclerosis (SSc) is the leading cause of CTD-PAH in Western countries [6], which accounting for almost 75% of CTD-PAH. Different from Western countries, CTD-PAH is the most prevalent type of PAH in China, and the most common cause of CTD-PAH in Chinese people is systemic lupus erythematosus (SLE) [7], while the epidemiological data of CTD-PAH in Chinese people is still limited up to now.

Since the mid-1990s, medications have become available to target three key pathophysiologic derangements in PAH-the prostacyclin, endothelin, and nitric oxide pathways [8], which improved quality of life, exercise capacity, and survival of these PAH patients [9]. CTD-PAH patients constitute at least 20% of patients included in all major trials of PH-specific targeted therapy [3]. However, despite the growing numbers of available PAH-targeted medications, many patients with PAH continue to deteriorate and the disease ultimately remains fatal; the outcome of patients with CTD-PAH remains poor [10]. Furthermore, the unadjusted risk of death for SSc-associated PAH compared with Idiopathic pulmonary arterial hypertension (IPAH) is 2.9, and the predictors of outcome are broadly similar to those for IPAH [10, 11]. For SSc-associated PAH, overall survival was 90%, 78%, and 56% at 1, 2, and 3 years from PAH diagnosis, respectively [11]. Therefore, it is important to recognize prognostic predictors for CTD-PAH patients, which means a lot to improve patients’ prognosis and long-term survival.

The objectives of this retrospective study are to analyze clinical characteristics and recognize survival and prognostic factors of CTD-PAH patients in a tertiary teaching hospital of Northwest of China.

Materials and methods

Study design

This is a retrospective cohort study of consecutive patients with documented CTD-PAH diagnosed at Department of Rheumatology at a tertiary teaching hospital in Xi’an, center of Northwest of China, from Jan 2014 to Dec 2019. All patient’s demographic, clinical characteristics, and clinical measures were retrieved from the hospital electronic medical records (EMR).

The study received favorable opinion from the Ethics Committee of First affiliated hospital of Xi’an Jiaotong University. The study protocol and data collection instruments were submitted and approved by the Data Protection Commission of Xi’an Jiaotong University. All patients provided their written informed consent prior to inclusion in the study.

Patient population

Patients of CTD were eligible to include in the study if they were ≥ 18 years old and discharge diagnosis including pulmonary hypertension. Patients were excluded as follows: (1) systolic pulmonary artery pressure (sPAP) estimated by first documented echocardiography when PAH diagnosis was absent; (2) HIV infection, congenital heart disease, acute heart failure and portal hypertension associated pulmonary arterial hypertension, pulmonary veno-occlusive disease, and drugs and toxins induced pulmonary arterial hypertension; (3) left heart disease; (4) chronic obstructive pulmonary disease, interstitial lung disease and sleep apnea; (5) when there is a high clinical suspicion for chronic pulmonary embolism, computed tomography pulmonary angiography (CTPA) or ventilation/perfusion (V/Q) imaging has been undertaken, and chronic pulmonary embolism patients in this study was excluded.

Data collection and instruments

Clinical and laboratory data during hospitalization at first-time PAH diagnosis were retrieved from the database of the dedicated EMR software of the hospital. The primary outcome of interest was all-cause mortality, which was ascertained from our electronic medical record and telephone follow-up calls to patients or bereaved family members.

sPAP estimation by echocardiography

Pulmonary artery systolic pressure (sPAP) was estimated using a modified Bernoulli equation: sPAP = 4 × (tricuspid systolic jet)2 + estimated right atrial pressure (RAP) [12]. Pulmonary hypertension was defined as a systolic pulmonary artery pressure (sPAP) > 35 mmHg estimated using echocardiograms [13]. Various sPAP cutoffs have been applied in previous studies, but a Doppler-derived sPAP of 35 mmHg is the most widely accepted threshold [14, 15]. PH severity was categorized according to the sPAP as follows: mild (35–50 mmHg), moderate (50–70 mmHg), and severe (> 70 mmHg) [16].

Statistical analysis

Normally distributed continuous variables were presented as mean (SD) and compared using t-tests. Nonnormally distributed continuous variables were presented as median (IQR) and compared using rank-sum tests. Categorical variables were presented as number (percentage) and were compared using the chi-square test. Survival functions were plotted using the Kaplan–Meier method and comparison of survival functions was performed by the log-rank test. The predictive values of biomarkers were assessed through time-dependent receiver operating characteristic analyses (ROC). Univariable Cox regression analysis and multivariable Cox analysis were applied to determine the survival-related factors. The univariable Cox regression models included Hb, HCT, pH, PaCO2, PaO2, Ln (NT-pro BNP), IgG, Fib, and sPAP. Factors with statistical significance in univariate analysis were included in a multivariate analysis. For data with skewed distribution (i.e., NT-pro BNP), logarithmic transformation with natural constant as the base was done in the Cox regression analysis and ROC analysis. A 2-tailed P < 0.05 was considered statistically significant in all analyses. R software, version 4.2.2 (http://www.R-project.org/), was used for all statistical analyses. Univariate log-rank test and cox regression analysis were performed in R using the survival package and the survminer package.

Results

  1. 1.

    Prevalence of PAH in CTD patients

From Jan 2014 to Dec 2019, there were 144 patients (2.80%) diagnosed as PAH among a total of 5142 CTD patients; we also collected 24 non-PAH-CTD patients with sPAP value estimated by echocardiography as control. The average follow-up period of all patients was 3.33 years. Of the 144 CTD-PAH patients (Table 1), there were 89 SLE patients (60.0%), 19 SSc patients (14.7%), 11 mixed connective tissue disease (MCTD) patients (7.6%), 7 rheumatoid arthritis (RA) patients (4.1%), 6 Sjögren’s syndrome (SS) patients (3.5%), 5 dermatomyositis (DM) patients (3.5%), 4 vasculitis patients (4.7%, 3 cases of microscopic polyangiitis and one case of Takayasu arteritis), and 3 primary biliary cholangitis (PBC) patients (1.8%). In this study, the prevalence of PAH was 7.97% in SLE, 15.32% in SSc, 14.10% in MCTD, 0.55% in RA, 2.74% in SS, 2.56% in DM, 2.63% in vasculitis, and 11.54% in PBC.

Table 1 Prevalence of PAH among CTDs
Full size table

The duration from CTD onset to PAH diagnosis was 75.4 ± 81.4 months (Table 1). The duration for SLE, SSc, MCTD, DM, SS, RA, vasculitis, and PBC was 39.0 (10.0, 101.0) months, 92.0 (6.5, 153.5) months, 27.0 (13.5, 98.0) months, 6.0 (0, 12.0) months, 127.0 (78.3, 139.0) months, 126.7 ± 100.5 months, 60.8 ± 88.5 months, and 65.0 ± 59.2 months respectively.

  1. 2.

    Clinical characteristics of CTD-PAH patients

Baseline clinical characteristics for these patients divided into CTD-PAH and non-PAH-CTD groups are summarized in Table 2. Demographic analysis showed that the median age of CTD-PAH patients was 46.0 years old and had high female predominance (94.4%); the median sPAP value of CTD-PAH patients was 52.5 (44.0,71.0) mmHg. Compared with non-PAH-CTD groups, CTD-PAH patients had worse cardiac function (P = 0.019), higher NT-pro BNP (P = 0.022), and γ-globulin (P = 0.023) level and lower PaCO2 level (P = 0.045). Although without statistically significant difference, the CTD-PAH group had higher D-dimer and IgG, lower serum vitamin D levels.

Table 2 Comparison of baseline clinical characteristics between non CTD-PAH and CTD-PAH patients
Full size table

The disease severity was classified as “mild” in 61 (42.36%) patients and “moderate-severe” in 83 (57.64%) patients using sPAP cutoff of 50 mmHg (Table 3). The average follow-up period of CTD-PAH patients was 3.19 years. Both groups had similarly high female predominance (96.7% vs 92.8%). Compared with mild group, moderate-severe group had worse cardiac function (P < 0.001), increased Hb (P = 0.014), HCT (P = 0.005), and NP-pro BNP (P = 0.011) level, decreased PaO2 (P = 0.023). Although without statistically significant difference, the moderate-severe group had decreased PaCO2 and dsDNA level. The most common cause of CTD-PAH in both groups is SLE (55.7% vs 66.3%).

Table 3 Comparison of baseline clinical characteristics between mild and moderate-severe CTD-PAH patients
Full size table

The targeted drugs used in this study were PDE-5Is (Sildenafil), ERAs (Ambrisentan, Bosentan), and PGI2 analogues (Beraprost, Treprostinil), the overall targeted drug usage rate was 55.6% (80/144), and the most used drug was Sildenafil. Only 27.5% patients were given combination, compared with mild patients; moderate-severe patients tended to use combination. In addition, only one patient used oral triple combination therapy.

  1. 3.

    Evaluation of outcomes

Kaplan–Meier curves were generated for analysis of final follow-up sPAP > 35 mmHg and death for the non-PAH-CTD, mild CTD-PAH, and moderate-severe CTD-PAH groups; patients lost to follow up were excluded from analysis. As showed in Fig. 1A, in terms of final follow-up sPAP > 35 mmHg, no significant difference was found among the 3 groups. While significant difference for survival was found among the 3 groups (Fig. 1B), it indicates that baseline sPAP value affects patients’ survival.

  1. 4.

    Determination of prognostic factors

Fig. 1
figure 1

Kaplan–Meier curves in subgroups. A Final follow-up sPAP value ≤ 35 mmHg, B survival. Log-rank test was conducted for comparison among the 3 groups (sPAP ≤ 35 mmHg, sPAP 35–50 mmHg, sPAP ≥ 50 mmHg)

Full size image

The results of the univariate and multivariate analyses using Cox regression model of the data for the CTD-PAH group are summarized in Table 4. Patients lost to follow up were excluded from Cox regression analysis. The variables with P value < 0.1 between groups in Table 1 were selected to be included in the Cox regression model. When as continuous variables, Hb, HCT, pH, Ln (NT-pro BNP), and sPAP were identified as factors significantly associated with survival. Factors with statistical significance in univariate analysis were included in a multivariate analysis. Due to the strong correlation between Hb and HCT, in order to prevent collinearity, only Hb, pH, Ln (NT-pro BNP), and sPAP were selected to be included in the multivariate model. Results showed significant association between pH and risk of death.

Table 4 Prognostic factors for survival in PAH-CTD patients
Full size table

For the 4 prognostic factors identified as impacting patients’ survival by univariate analysis as continuous variables (i.e., Hb, pH, Ln (NT-pro BNP) and sPAP), the areas under the ROC curves (AUCs) were calculated to be 0.655, 0.667, 0.678, and 0.669, respectively (Fig. 2A–D). Based on the ROC curves, the cutoff values were determined to be 109.0 g/L, 7.457, 7.267, and 62.5 mmHg, respectively. When using cutoff values of Hb, pH, Ln (NT-pro BNP) and sPAP, Hb, pH, and Ln (NT-pro BNP) were identified as factors significantly associated with survival by univariate analysis (Table 5). In the multivariate model, Hb and pH showed significant association with risk of death. Kaplan–Meier curves according to the cutoff values of Hb and pH were shown in Fig. 3, which proved further that Hb > 109.0 g/L (Fig. 3A) and pH > 7.457 (Fig. 3B) were independent unfavorable predictors of survival.

Fig. 2
figure 2

ROC curves based on survival. A Hemoglobin, B pH, C Ln (NT-pro BNP), D baseline pulmonary artery systolic pressure

Full size image
Table 5 Prognostic factors for survival in PAH-CTD patients (based on cutoff values)
Full size table
Fig. 3
figure 3

Kaplan–Meier curves divided by cutoff values of Hb and pH in CTD-PAH. A log-rank test was conducted in CTD-PAH for comparisons among the groups divided by each cutoff value. A Hemoglobin, B pH

Full size image

Discussion

This study describes the characteristics of CTD-PAH patients and provides insight into their PAH-targeted treatment status and survival. Our data reveal the protopathy distribution of CTDs, PAH prevalence in CTDs, PAH-target drugs usage, survival, and prognostic factors for death of CTD-PAH. Knowing these features may prove helpful for clinicians evaluating CTD patients who present with PAH timely, risk assessment, and predicting prognosis in patients with CTD-PAH.

We showed that SLE is the leading cause of CTD-PAH, which is consist with previous data. The second cause of CTD-PAH is SSc. Previous cohort studies in Asia reported that PAH is most commonly associated SLE, followed by SS and MCTD [17,18,19]. In Europe, a study showed that of CTD-PAH patients, 75% had SSc, 7% had SLE, 9% had MCTD, and 12% had another CTD [20]. In the USA, study showed that of patients with CTD-PAH, 67.7% had SSc, 18.7% had SLE, and 8.8% had MCTD [21]. The relationship between RA and PAH is not established [22], this study showed that 7 RA patients (7/144, 4.1%) were presented with PAH, and the prevalence of PAH in RA was only 0.55%. Few data describe when PAH occurs in CTDs, our results showed that PAH can occur early after DM onset. In addition, our study also showed that PAH could occur in ANCA-associated vasculitis and Takayasu’s arteritis. PAH is a rare occurrence in vasculitis, previous studies also showed that PAH could occur in vasculitis [23,24,25], and CTDs-PAH studies also included vasculitis patients [19, 26]. The exact pathophysiology of how PAH develops in vasculitis is yet unknown which needs to be investigated further, as has been hypothesized with other CTDs, endothelial dysfunction and remodeling of pulmonary arteries, and immune dysregulation can all play a role [27]. And researcher also showed that pulmonary artery vasculitis, pulmonary artery sclerosis/stenoses and elevated left ventricular filling pressure during Takayasu’s arteritis may contribute to PAH development [28]. Clinicians should be aware that PAH can occur at early stage of CTDs, and patients with RA and vasculitis should also be alert to the occurrence of PAH. Regular echocardiography monitor is recommended strongly in the CTD patients management to identify CTD-PAH as early as possible.

PAH has generally been thought to affect predominantly younger individuals, mostly females [29], female sex is a risk factor for PAH, yet females have better survival than males [30]. In industrialized countries, PAH affects mainly elderly people, whereas mostly young people are diagnosed in the developing world including China. In our study, the median age of CTD-PAH was 46 years old with high female predominance, but older age and female was not found to be an independent risk factor for disease severity of PAH in our study.

Worse World Health Organization functional class (WHO-FC) and NT-pro BNP were found in this study in CTD-PH patients, especially for the moderate-severe CTD-PH patients. The WHO-FC is one of the strongest predictors of survival both at diagnosis and follow-up of PH patients; worsening WHO-FC is one of the most alarming indicators of disease progression. In addition, although BNP and NT-proBNP are not specific for PH, they remain the only biomarkers routinely and frequently used for PH patients, which is correlating with myocardial stress and providing prognostic information [1]. Both WHO-FC and NT-pro BNP are one of the three important indicators of the simplified non-invasive risk-assessment tool for prognostic evaluation, risk assessment, and treatment evaluation of PH [31], which need to be assessed every 3–6 months.

Immunity/inflammation is an important mechanism in the pathogenesis of PAH [32]; anti-inflammatory therapy has been evaluated for years. The most significant γ-globulins are immunoglobulins (IgA, IgG, and IgM antibodies), although some immunoglobulins are not gamma globulins, and some gamma globulins are not immunoglobulins. Hypergammaglobulinemia is often found in patients with autoimmune diseases, and its level may correlate with disease activity, such as lymphoid cell infiltration, treatment responsiveness, and also PAH [33, 34]. It is known that SLE and MCTD often display hypergammaglobulinemia or a high IgG titer; auto-antibodies participate in the CTD-PAH development [35]. In this study, γ-globulin was significantly increased in CTD-PH patients, and serum IgG level was higher in the moderate-severe group than the mild group although without statistical difference. Nowadays, elucidating the process of immunoglobulin production is in great need to provide new treatment strategies for PH.

D-Dimer level was higher and Vitamin D3 (VitD3) was lower in CTD-PAH patients compared with CTD patients without PAH, although without statistical differences. We have found previously that SLE-PAH had high D-dimer and FDP levels, and D-dimer was an independent predictor for SLE-PAH, which mediated effect of low low-density lipoprotein on SLE-PAH [36]. While the latest guideline of the long-term risk/benefit ratio of oral anticoagulation is unfavorable because of an increased risk of bleeding, vitamin K antagonists are recommended in PAH-CTD with a thrombophilic predisposition [37]. For VitD3, studies have proved that VitD3 deficiency is frequently seen and indicates worse prognosis in PH patient. VitD3 supplementation could improve PH survival in rat and endothelial function, while there is still controversy about that whether VitD3 could ameliorate pulmonary vascular remodeling and right ventricle hypertrophy [38,39,40]. The effect of VitD3 on pulmonary hemodynamics, pulmonary vascular remodeling, and right ventricle hypertrophy in PH and its potential mechanisms need to be clarified further.

Except WHO-FC and NT-pro BNP, increased Hb and decreased PaO2 were identified to be associated with PAH severity in CTD patients in the present study; PaCO2 was decreased significantly in CTD-PAH patients without association with PAH severity. In addition, we found that the factors associated with a higher risk for death in CTD-PAH patients (in the multivariate model) were Hb and pH. Our results indicated that hemoglobin in blood and arterial blood gas analysis was important for risk and prognosis assessment of CTD-PAH patients. Both iron deficiency and anemia have been linked to the clinical course of PH [41], we also found that Hb level was decreased in CTD-PAH patients compared CTD patients without PAH, and current treatment guidelines suggest regular iron status assessment and the implementation of iron supplementation strategies in these patients [1]. While it is reported that as the Hb level decreases, blood viscosity decreases, cardiac output increases, filling pressures tend to decrease, and systemic and pulmonary vascular resistances decrease substantially [42, 43]. Previous study proved that PH patients who died had significantly lower hemoglobin levels than those survived, and anemic PH patients were 3.3 times more likely to die than non-anemic PH patients [44]; it is also reported that higher Hb levels associated with higher risk for cardiovascular disease-related mortality [45]. While the optimal hemoglobin level for prognosis of PH patients is still unknown, in this study, we showed that Hb > 109.0 g/L was an independent unfavorable predictors of survival. Hypoxia induces erythropoietin and secondary increase in Hb [46]; it is reported that erythropoietin attenuates pulmonary vascular remodeling and induces a reduction in pulmonary arterial pressure [47, 48], while higher level of Hb may inhibit the increase of erythropoietin level. In the meanwhile, chronic hypoxia is a well-known cause of PH; anemia can lead to a hypoxic state which leading to development of PH. Therefore, the complex interaction of hemoglobin level and PH need to be investigated further, aiming to give guidance on how to select PH patients who might benefit from iron supplementation, and help to set treatment goal.

Hypoxic stress is a key driving force in the vascular remodeling observed in PH, which could induce hyperventilation/alkalosis compensatorily; respiratory alkalosis is observed commonly in PH. In this study, abnormal arterial blood gas analysis was found in CTD-PAH patients, PaO2 was negatively associated with PAH severity as expected, and interestingly, pH > 7.457 was an independent unfavorable predictor of survival. It has been proved that acute hypoxic pulmonary vasoconstriction is attenuated by respiratory alkalosis, but respiratory alkalosis failed to attenuate vascular reactivity to subsequent pressor stimuli [49]. Although hypoxia-induced vasoconstriction leads to compensatory respiratory alkalosis that has vasodilating effect, it also can worsen intrapulmonary shunt and systemic oxygenation [50]. Therefore, it is important to correct hypoxemia for PH patients to reverse alkalosis.

Since 2007, PAH-targeted medications have been available in China, but because of low disease awareness, difficult access to experienced diagnostic facilities, and financial constraints, quite a few PAH patients have not received sufficient targeted drug use, which affects patients’ prognosis further and needs to be improved. In this study, PAH-targeted treatment rate was 55.6%, which needs to be improved. One study based on Chinese population had showed that over 50% of 174 participants did not use any kind of PAH-specific therapies because of the financial burden [51]. Most patients of this study chose monotherapy, PDE5I was used most often in both monotherapy and combination therapy because of its relatively low price, and only one patient with combination therapy did not use PDE5I. Combining drugs with different mechanisms of action, in order to optimize clinical benefit while minimizing side effects, a study in the USA states that only 5% of patients began initial combination therapy for PAH [52]. Treatment strategy for PAH has thereby changed significantly over the past decade; clinicians should know that combination therapy is becoming progressively the gold standard of care in patients with PAH and is becoming widely used in clinical practice [53]. The benefit of combination over monotherapy was demonstrated in patients with CTD-PAH, particularly in those with typical PAH hemodynamics characteristics at baseline [54], while randomized controlled trials (RCTs) showed that the magnitude of the targeted drugs response in the PAH-CTD was lower than in the IPAH patients [55]. The limited data precludes consensus on which combined treatment or strategy should be preferred; drugs targeting the ET and NO pathways are most commonly used as first-line combination therapy, but the association of ERA and PDE5I, both available in oral form, offers an interesting option [56]. In our study, the majority of patients using combination treatment is treated with an ERA combined with a PDE-5I. Also, no long-term evaluation of combination therapy is available; it is reported that combination therapy had no proven effect on mortality and had a much higher incidence of withdrawal due to adverse effects than monotherapy [57]. With this in mind, treating patients with more than one of these drugs may result in additive side effects; it is reported that edema and headache occurred more frequently in patients taking a combination of ambrisentan and tadalafil compared with patients treated with either drug as monotherapy in the AMBITION study [58].

The results of the present study should be interpreted in the context of several limitations. First, due to the invasive nature and cost of RHC, echocardiography was used to diagnose PAH in patients in this study. Estimates of sPAP by echocardiography are based on the peak TRV. This technique can lead to under- or over-diagnosis of PAH rates as TR may not be present in mild PAH cases without signs of right heart failure and also could be found in elderly or obese patients without PAH. Second, the present study concerns the small sample size; the results would be not vigorous. Third, as a non-interventional trial, a cause-and-effect relationship between PAH and mortality cannot be established. Fourth, although multiple Cox regression analysis was done to control for the effect of potential confounders, there might be other potential confounding variables that were not considered in this study.

In conclusion, PAH is not rare in CTDs patients; PAH affects CTD patients’ prognosis significantly. Higher Hb and pH were associated with an increased risk of death. Future studies are needed to explore the mechanisms underlying these associations to identify some potential therapeutic interventions for this high-risk population.