Association of red cell distribution width with pulmonary arterial hypertension in patients with mixed connective tissue disease

Background Pulmonary arterial hypertension (PAH) is a severe complication of mixed connective tissue disease (MCTD) and contributes to increased morbidity and mortality. Still, the demographic characteristics and risk factors of PAH in MCTD remain poorly understood. This study explored risk factors for PAH development in MCTD. Methods Data from patients with MCTD and PAH hospitalized from May 2009 to December 2022 in a single center were collected and compared with patients with MCTD without PAH. The variables were analyzed by logistic regression to identify the factors associated with PAH in patients with MCTD. The receiver-operating characteristic (ROC) curve was used to assess the diagnostic value of the identified factors. Results Finally, 119 patients with MCTD were included; 46 had PAH. The mean age at PAH onset and diagnosis was 38.9 ± 13.4 and 39.9 ± 13.7 years, respectively. The median pulmonary arterial systolic pressure (PASP) was 67.0 mmHg. The median brain natriuretic peptide (BNP) level was 180.0 pg/ml at PAH diagnosis. Red cell distribution width (RDW) (OR: 2.128; 95% confidence interval: 1.497–3.026; P < 0.001) was associated with PAH in patients with MCTD. There was a positive correlation between RDW and PASP (r = 0.716, P < 0.001). At a cutoff of 15.2%, RDW had the best sensitivity (80.4%) and specificity (82.2%) for PAH. Conclusion RDW may serve as a sensitive index to predict PAH in patients with MCTD. Supplementary Information The online version contains supplementary material available at 10.1186/s12890-023-02597-z.


Background
Mixed connective tissue disease (MCTD) is one of the most complicated autoimmune diseases and involves various overlapping symptoms [1].It is characterized by diverse clinical manifestations of systemic sclerosis (SSc), systemic lupus erythematosus (SLE), polymyositis or dermatomyositis (PM/DM), and rheumatoid arthritis (RA), which can also affect multiple organs, including the pulmonary system [2].Pulmonary involvements, especially pulmonary arterial hypertension (PAH), are devastating complications among patients with MCTD, resulting in increased morbidity and mortality [3][4][5].The prevalence of PAH is 2-24% in patients with MCTD, depending on the measurement methods used [3].So far, little knowledge is available about the clinical characteristics of PAH patients associated with the MCTD population.The present study aimed to explore the factors associated with PAH in patients with MCTD and determine the diagnostic value of the identified factors.

Patients and diagnostic criteria
Patients with MCTD hospitalized in the rheumatology and immunology department of the affiliated Drum Tower Hospital of Nanjing University Medical School between May 2009 and December 2022 were included in this study.The exclusion criteria were (a) < 18 years of age, (b) complicated with other connective tissue diseases (CTD) such as SSc, SLE, or primary Sjögren's syndrome (pSS), (c) PAH caused by cardiac structural lesions, such as left heart disease or valvular heart disease, (d) with severe pulmonary diseases resulted from lung cancer or chronic obstructive pulmonary disease, or (e) PAH due to liver cirrhosis, pulmonary venous occlusive disease, or chronic thromboembolism.All patients met the established MCTD criteria [6].
PAH was defined according to the 2022 ESC/ERS guidelines, using tricuspid regurgitation velocity (TRV) > 2.8 m/s [7].The initial diagnosis or screening of PAH was established based on transthoracic echocardiography (TTE)-estimated resting pulmonary arterial systolic pressure (PASP) ≥ 36 mmHg [7], while those with PASP < 36 mmHg and without any other echocardiographic indicators suggesting PAH were included in the MCTD-non-PAH group.Written informed consent was obtained from each subject following a detailed explanation of the objectives and protocol of the study.The study was conducted in accordance with the Declaration of Helsinki and was approved by the ethics committee of Nanjing Drum Tower Hospital (No.2020-093-01).

Data collection
Demographic data, including sex, age at MCTD onset, and disease duration, height, weight, body mass index (BMI), body surface area (BSA), and disease activity were collected.In addition, age at PAH onset and diagnosis was also included for patients with PAH, in which the former was defined as the time when the initial symptoms appeared, while the latter was when TTE examination showed abnormalities.The nonspecific symptoms were recorded, such as Raynaud's phenomenon, sclerodactyly, serositis, limb edema, fatigue, arthralgia, and chest tightness.The World Health Organization functional classification (WHO-FC) was independently evaluated by two authors (YJ and BJ) for each patient with PAH.
TTE was performed using a Philips IE Elite machine.The included patients were all examined by TTE to evaluate the likelihood of PAH.The peak PASP was calculated using the simplified Bernoulli equation (PASP = 4 tricuspid regurgitation 2 + right atrial pressure) [8].The other parameters based on TTE were included.Autoantibodies were routinely tested.Other laboratory data, including complete blood count, uric acid (UA), blood lipid analysis, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), thyroid function test, complement, immunoglobulin, blood coagulation function test, and brain natriuretic peptide (BNP) levels were also documented.The hemoglobin and mean corpuscular volume (MCV) were used to evaluate the iron deficiency.The Systemic Lupus Erythematosus Disease Activity Index 2000 (SLE-DAI-2 K) was used to evaluate disease activity [9].

Statistical analysis
Analyses were done using SPSS 19.0 (IBM, Armonk, NY, USA) and MedCalc 19.1 (MedCalc, Ostend, Belgium).The variables were presented as means ± standard deviations ± or medians with interquartile ranges.Student's t-test or Mann-Whitney U-test was used for comparisons.The categorical variables were presented as n (%) and analyzed using the chi-square test.The correlations between two continuous variables were determined using Spearman's correlation coefficient.Factors associated with PAH in the univariable logistic regression analyses were evaluated using multivariable logistic regression analysis.The receiver operating characteristic (ROC) curve with the area under the curves (AUC) was plotted to investigate the diagnostic value of the factor.The optimal cutoff was assessed using Youden's index in ROC analysis.Survival was analyzed using the Kaplan-Meier method and the log-rank test.P < 0.05 was considered statistically significant.As for the symptoms (Table 1), serositis (63.0% vs. 27.4%,P < 0.001) and chest tightness (76.1% vs. 32.9%,P < 0.001) in the MCTD-PAH group were significantly more common than in the MCTD-non-PAH group.Raynaud's phenomenon, sclerodactyly, limb edema, fatigue, disease activity and arthralgia showed no differences between the two groups.
As shown in Table 2, many patients with PAH had severe cardiac involvement at PAH diagnosis, with a median PASP of 67.0 (interquartile range 48.5-90.0)mmHg and a median level of BNP of 180.0 (interquartile range 63.8-478.5)pg/mL.The BNP levels were higher in the MCTD-PAH group compared with the MCTD-non-PAH group (median, 180.0 vs. 57.3pg/mL, P < 0.001).The values of PASP were not routinely tested in the MCTDnon-PAH group, and the data could not be compared.When PAH was diagnosed, nearly half (45.7%) of patients were in WHO Fc III/IV.Meanwhile, only one (1.4%)patient in the MCTD-non-PAH group reached WHO Fc III/IV.

Factors associated with PAH
Univariable and multivariable binary logistic regression analyses were performed to evaluate the clinical, laboratory, and echocardiographic factors independently associated with PAH.As shown in Table 3, only RDW (OR: 2.128; 95% confidence interval: 1.497-3.026;P < 0.001) was significant.Consistently, there was a positive correlation between RDW and PASP (r = 0.716, P < 0.001) by Spearman correlations analysis (Fig. 1).
Next, we analyzed the ROC curve of the diagnostic accuracy for RDW (Fig. 2).Using 15.2% as a cutoff level, the RDW value had 80.4% sensitivity and 82.2% specificity for PAH diagnosis.The AUC estimate for RDW was 0.9 (P < 0.001), with a Youden index of 0.6.

Survival
Nineteen out of 119 patients (16.0%) died during a mean follow-up period of 80 ± 38 months.Eight patients were  lost to follow-up.One patient committed suicide, and one died in a car accident.The survival rate was higher in the MCTD-non-PAH group compared with the MCTD-PAH group (log-rank P < 0.001) (Supplementary Figure S1).Elevated RDW values were associated with an increased risk of all-cause death.Fourteen out of 46 patients (30.4%) died during a mean follow-up period of 68 ± 40 months.Two patients were lost to follow-up.One patient committed suicide.According to the Kaplan-Meier survival analysis, a significant difference was found between the two groups(patients with RDW > 15.2% and patients with RDW ≤ 15.2%) in terms of survival rate (logrank P = 0.049) (Supplementary Figure S2).

Discussion
CTD-PAH constitutes the second most common subtype of PAH [10], most commonly associated with SSc, SLE, and MCTD [11].While SSc and SLE are the leading cause of CTD-PAH [12], little is known about MCTD-PAH, despite reports showing that MCTD-PAH was more frequent than SSc and SLE.In this study, PAH diagnosis according to TTE was 1.0 ± 1.6 years' delayed from PAH onset, accompanied by a significantly higher percentage The first line of diagnostic testing for patients with suspected PAH involves undergoing TTE.Previous reports showed that noninvasive echocardiography had high specificity and sensitivity in detecting PAH when above 36 mmHg [14].The median PASP by echocardiography of the MCTD-PAH patients was 67.0 mmHg, greatly higher than the diagnostic critical value.
On the other hand, significantly higher pericardial effusion values and lower indexes of LVDd, PA, LAD, and LVEF were also observed in MCTD-PAH patients.Thus, it is likely that the pathophysiology of PAH involves remodeling not only the right heart but also the left heart.In addition, in the present study, 17 of 46 patients (37.0%) had pericardial effusion by echocardiography.
Fig. 2 The ROC curve of diagnostic accuracy with RDW for predicting the presence of PAH in MCTD patients.An RDW value > 15.2% was 80.4% sensitive and 82.2% specific for diagnosing PAH, with an AUC of 0.9 (95% confidence interval: 0.835-0.949;P < 0.001).Youden index was 0.6 The literature suggests that pericardial effusion relates to RV failure and immunologically mediated inflammatory conditions [15].Pericardial effusion is an increasingly recognized risk factor of PAH related to CTD [16].
Several biomarkers have been recommended to assess the presence and prognosis of PAH [17].Nagaya et al. reported that PAH patients with BNP > 150 pg/mL at baseline and > 180 pg/mL after therapy had a poor prognosis [18].The median plasma levels of BNP in the MCTD-PAH group was 180.0 pg/mL, suggesting these patients were in poor condition.UA [19], thyroid dysfunction [20], and the presence of specific autoantibodies [21] were also proposed to be related to PAH.However, the risk of both UA and thyroid dysfunction has not been confirmed in this study, and autoantibody profiles were not significantly different between the two groups.
RDW is an index of blood routine [22], which has been suggested as a biomarker of PAH in patients with other CTD [23,24].To our knowledge, this is the first study to identify the importance of RDW in PAH related to MCTD.The sensitivity and specificity of RDW value > 15.2% were 80.4% and 82.2%, respectively, for the prediction of MCTD-PAH.Given its positive correlation with the PASP, RDW may significantly affect the progression of PAH associated with MCTD.Recently, raised RDW values are related to the mortality of PAH and proposed as a prognostic biomarker of poor prognosis in PAH with various etiologies [23,25], which was even superior to pro-BNP [26].Xi et al. [27] suggested that RDW independently predicted the responsiveness of acute pulmonary vasodilator testing in idiopathic PAH patients.
The exact pathologic mechanism linking RDW to PAH is still unclear but may be related to vascular injury [28].Routinely, the most common cause of an increased RDW is anemia related to ineffective erythropoiesis [29].In the present study, all patients had normal hemoglobin levels with a normal range of MCV in the whole group.There was no iron deficiency; as such, the influence of iron was excluded.It was hypothesized that RDW was involved in inflammation [30].The elevation of RDW level could reflect the circulating levels of proinflammatory cytokines, such as tumor necrosis factor (TNF)-α, interleukin (IL)-1, and IL-6, thus leading to the occurrence of PAH [23,31,32].Further studies focusing on mechanisms correlated with RDW and MCTD-PAH are necessary.
FT3 was significantly decreased in MCTD-PAH patients, while TSH was significantly higher in the MCTD-PAH group.The literature suggests that thyroid dysfunction is associated with PAH [33].Indeed, thyroid hormone impacts cardiac contractility, cardiac output, and pulmonary and systemic vascular resistance [20,34], and up to 20% of patients with PAH also have thyroid disease [35].The present study report reinforces the important observation that TSH is associated with PAH, and a future study of thyroid dysfunction as a potential remediable contributor to mortality in PAH is warranted.

Limitation of the study
A major limitation of this study is not routinely applying the gold standard test for PAH-right-sided heart catheterization (RHC).As an invasive and expensive operation, RHC was not common consent by most patients.Some patients in the present study were too serious to lie flat.The retrospective nature of the study limited the data to those available in the patient charts.

Conclusion
MCTD patients with elevated RDW had a high risk for PAH.RDW can be used as a simple index for monitoring PAH in clinical practice.

Fig. 1
Fig. 1 Association between RDW and PASP.Spearman correlations analysis showed a significant positive correlation between RDW and the PASP (r = 0.716, P < 0.001) in the study population of MCTD-PAH (n = 46)

Table 3
Multivariable binary logistic regression analysis of statistically significant candidates