A Cross-sectional Study on Pulmonary Hypertension in Patients with Stage 5 Chronic Kidney Disease

Abstract

It is well known that chronic kidney disease (CKD) is associated with significant morbidity and mortality, predominantly due to cardiovascular complications. Recent literature report pulmonary hypertension (PH) as a common accompaniment of CKD and in majority of these cases, secondary causes of PH are not evident. In this study, we looked at the prevalence and possible risk factors of PH in stage 5 CKD patients with a special focus on unexplained PH. In this cross-sectional study, 100 stage 5 CKD patients [50 each on maintenance hemodialysis (HD) and conservative management] were included. After baseline investigations participants underwent two-dimensional transthoracic echocardiogram. The routine method of PH diagnosis based on modified Bernoulli’s formula was complemented with an alternate method based on pulmonary artery acceleration time (PAAT). Detailed workup for secondary causes was carried out in patients with PH. The prevalence of PH in the study population was 89% (56% mild, 35% moderate, and 9% severe). Asymptomatic left ventricular diastolic and systolic dysfunction were noted in 54% and 20%, respectively. Significant association with PH was found with the duration of CKD, systolic and diastolic Blood pressure, hemoglobin, transferrin saturation, maintenance HD, and dialysis vintage. In sharp contrast to the existing data this study showed a very high prevalence of PH though severe PH was present only in 9%. The inclusion of PAAT-based method enabled the detection of more cases of PH. Further evaluation carried out for common secondary causes did not show significant abnormalities except for a sizeable proportion with asymptomatic left ventricular dysfunction.

Introduction

The prevalence of chronic kidney disease (CKD) is on the rise worldwide attributable to greater proportion of old age population and lifestyle disorders such as diabetes and metabolic syndrome.^[1] The pathology of CKD is not confined to the kidneys and remarkable functional and structural changes occur in the cardiovascular system. Both CKD and cardiovascular diseases are intertwined and accentuates the progression of disease bidirectionally.^[2] Along with conventional cardiovascular risk factors, nontraditional or novel risk factors are also at play in CKD.^[3]

Pulmonary hypertension (PH) is being recognized off late as a common association of CKD. It is associated with 38% increased mortality and 23% increased risk of other cardiovascular events in CKD.^[4,5,6] CKD and PH share many epidemiologic factors, and hence both can co-exist in up to 40%–70% of the patients.^[7] In the remaining patients, PH is diagnosed without any apparent causes, referred to as “unexplained PH” by Yigla et al in 2000.^[8] In the latest classification by the Fifth World Symposium on PH in 2013 (Nice, France) CKD is a listed as a cause of Group 5 PH with unclear multifactorial mechanisms.^[9] The gold standard to diagnose PH is right heart catheterization but most of the studies used two-dimensional (2D) echo and showed a prevalence varying from 27%–65%.^[7]

The proposed risk factors include advanced age, longer duration and higher stages of CKD, and use of hemodialysis (HD).^[10,11] The exact pathogenesis is not known, and the most plausible mechanism is the synergistic effects of hemodynamic factors which increase cardiac output, increase pulmonary vascular resistance (PVR) and increase pulmonary capillary wedge pressure (PCWP).^[12] The regression of PH has been documented following kidney transplantation, arteriovenous (AV) fistula closure, and after each session of dialysis which provides an indirect evidence to the proposed mechanism.^[4,13] The PVR increase in CKD due to uremic endothelial dysfunction mediated by elevated asymmetric dimethyl arginine, thromboxane A2, and endothelin and reduced nitric oxide level.^[14,15] Hyperdynamic circulation result from the creation of AV fistula and anemia which increase cardiac output and may lead to the development of PH^[4,16] Pulmonary vascular calcification has been reported in experimental animals with hyperparathyroidism another putative mechanism for PH.^[17] Recurrent thromboembolism from fistula thrombectomy and microbubble embolism from dialyzer and tubings were also described in literature as a cause of PH in CKD.^[18,19] Elevated PCWP is a very common finding in CKD secondary to diastolic or systolic left ventricular dysfunction.^[20] The PEPPER study which used invasive hemodynamic monitoring reported that elevated PCWP is more common in CKD than reported from echo-based studies and which is the major etiology for 90% of the unexplained PH.^[13]

Methods

This cross-sectional study was carried out in the Department of Nephrology, Government T. D. Medical College, Alappuzha, Kerala which is a tertiary care hospital in South India. The protocol was approved by the Institutional Research Committee and Ethics Committee. The sample size was estimated to be 100 assuming a PH prevalence of 50% based on available data in the literature. Consecutive cases were included in the study from January 2018 to November 2018 if they satisfied the inclusion criteria, i.e., >18 years, stage 5 CKD either on conservative management or on maintenance HD at least for three months. Exclusion criteria were designed to rule out cases with other potential causes for PH namely chronic lung parenchymal diseases, disorders affecting the chest wall, previous pulmonary tuberculosis, identified cardiac disorders including coronary artery diseases, valvular heart diseases, cardiac failure or congenital heart disease, connective tissue disorders, chronic liver disease, human immunodeficiency virus (HIV) infection, history of pulmonary thromboembolism, body mass index (BMI) >30 kg/m² and clinical features of sleep apnea.

Detailed history and physical examination followed by laboratory investigations were done. Antinuclear antibody (ANA) test, HIV serology, thyroid-stimulating hormone, and high-sensitivity C-reactive protein (hs-CRP) were done along with baseline investigations. All patients were subjected to 2D transthoracic echo with Doppler to assess PH and to look for any cardiac causes for PH. A single experienced cardiologist performed all echo studies using GE Vivid E9 cardiac ultrasound unit at a time when the patient was not in fluid overload clinically and immediately after dialysis session in maintenance HD patients to exclude the effect of fluid overload state. M-mode, two-dimensional, pulsed-wave, and color-flow Doppler study was done in the left lateral decubitus position. Images were obtained from subxiphoid long axis, para-sternal short- and long-axis, and apical four and two-chamber views.

Modified Bernoulli’s formula was used to calculate pulmonary artery pressure (PAP) from tricuspid jet velocity and estimated right atrial pressure in all patients with discernible tricuspid regurgitation (Estimated peak systolic PAP [EPSPAP] = 4 × TRVmax2 + 10 mm Hg). Alternatively, EPSPAP was calculated from the value of pulmonary artery acceleration time (PAAT) using the formula log10 (EPSPAP) = _0.004 (PAAT) + 2.1 in all cases. In subjects who have both readings, the highest value was taken to diagnose PH. PAP >30 mm Hg is taken as PH. Patients were divided into three groups based on the severity. Mild PH 30–50 mm Hg, moderate PH 51–70 mm Hg, and severe PH >70 mm Hg.

Patients with PH were subjected to detailed evaluation by a pulmonologist, including a pulmonary function test, 6 min walk test, and High-resolution computed tomography (HRCT) thorax to rule out obstructive and restrictive lung diseases. Ultrasonography Doppler was done to look for lower limb deep vein thrombosis as an etiological clue to chronic pulmonary thromboembolism.

Data analysis was carried out using IBM SPSS Statistics software version 25.0 (IBM Corp., Armonk, NY, USA). Categorical variables were expressed as percentage. Quantitative variables were expressed as mean ± standard deviation if data followed normal distribution pattern and as median ± interquartile range (IQR) if nonnormally distributed. Kolmogorov–Smirnov test was used if the sample size was more than 50 and Shapiro–Wilk test if sample size was <50 to test the normality of data. Means of two groups of continuous variable were compared with Student’s t-test and that of categorical variables were compared with Chi-square test. Nonparametric tests were employed in cases of nonnormally distributed data. For the correlation between two variables Pearson coefficient was used in normal distribution and Spearman coefficient in nonnormal variables. A P <0.05 was taken as the threshold for statistical significance.

Results

The final study population consisted of 50 CKD stage 5 patients on maintenance HD and another 50 patients on conservative management who satisfied the study criteria for inclusion (Figure 1). The most common etiology of CKD was diabetic nephropathy followed by ischemic nephropathy (Figure 2). The baseline characteristics of the two groups were compared and summarized in Table 1. Most of the characteristics at baseline were matching, but between the group difference was apparent with respect to gender distribution, duration of CKD, duration of systemic hypertension and diabetes, BMI, and use of erythropoietin stimulating agents. Laboratory parameters at the time of study enrolment were also compared between the groups (Table 2).

Figure 1: Schematic representation of the study procedure.MHD: Maintenance hemodialysis, COPD: Chronic obstructive pulmonary disease, PTB: Pulmonary tuberculosis, SLE: Systemic lupus erythematosus.

Figure 2: Etiology of chronic kidney disease.CGN: Chronic glomerulonephritis, ADPKD: Autosomal dominant polycystic kidney disease, CTID: Chronic tubulointerstitial kidney disease.

Table 1: Baseline characteristics.

Table 2: Baseline laboratory parameters.

In the maintenance HD group, 60% had a maintenance HD vintage between one and five years, and 22% were on maintenance HD for more than five years. Patients having either fistula or tunneled dialysis catheter as vascular access were compared to study the association between AV fistula and risk of PH. The proportion of cases with different types of vascular access is depicted in Figure 3. None of the patients in the conservative group had a functional AV fistula. Patients on twice-perweek and thrice-per-week maintenance HD were distributed with nearly equal proportions.

Figure 3: Vascular access in maintenance hemodialysis patients.AVF: Arteriovenous fistula.

After detailed echocardiographic evaluation 89% of the cases were found to have PH with 30 mm Hg as the cut off PAP. Based on severity of PH, 9% had severe PH, 35% moderate and rest of the cases had mild PH (Figure 4). The proportion of PH was 84% in the conservative group whereas it was 94% in the maintenance HD group. The prevalence of PH based on TR jet velocity (Bernoulli’s formula) was 65% and 24 additional cases were picked up when it was complemented with PAAT based method. The two methods of PAH measurement showed statistically significant correlation (Pearson coefficient r = 0. 643, P <0.01). This relationship was further established by a significant agreement in the Cohen’s Kappa test (k = 0.701, P <0.01) and a negative McNemar’s test (P = 1.00).

Figure 4: Distribution of number of cases with pulmonary hypertension in conservative and maintenance hemodialysis group.MHD: Maintenance hemodialysis, PH: Pulmonary hypertension.

Figure 5: Correlation of pulmonary arterial pressure from TR jet velocity and pulmonary artery acceleration time.PAAT: Pulmonary artery acceleration time, PAP: Pulmonary artery pressure, TR: Tricuspid regurgitation

Other important findings in echo are left ventricular hypertrophy in 43%. The left ventricular systolic dysfunction (LVSD) was diagnosed in 22% in the conservative group and 18% in the maintenance HD group. However, no patient in the study had severe LVSD characterized by an ejection fraction (EF) of <30% and median EF was 63.5% with an IQR of 19. The left ventricular diastolic dysfunction was more common with an overall prevalence of 54% in the study group, 50% and 42% in the maintenance HD and conservative group, respectively.

In the pulmonology evaluation, the HRCT thorax did not show evidence of interstitial lung disease or other significant lung parenchymal changes to account for PH. Exercise desaturation was tested by 6 min walk test which was also normal in all the subjects. Pulmonary function test showed a restrictive pattern in two patients and obstructive pattern in one patient.

Lower limb venous Doppler study did not reveal evidence of deep vein thrombosis in any case. HIV serology, ANA test, and thyroid function test were also normal in the study subjects.

Many clinical and laboratory variables were compared in patients with and without PH. Blood pressure (BP), anemia, duration of CKD, being treated by maintenance HD were all found to be statistically significant (Table 3). Various other parameters were analyzed but were found to be nonsignificant (Table 4).

Table 3: Risk factors for pulmonary hypertension mean value.

Table 4: Variables without statistical significance.

Discussions

Patients in this study were chosen after excluding other common risk factors for PH. This step could help us in eliminating the confounding factors for PH to a great extent which is very common in CKD population. Two equal-sized populations were chosen from patients on maintenance HD and on conservative management to analyze the effect of HD on PH. The two groups were comparable at baseline with respect to other clinical and laboratory parameters.

Laboratory parameters pertaining to various aspects of CKD were estimated and analyzed for any association with unexplained PH in CKD. They include:

Anemia and iron deficiency (hemoglobin, transferrin saturation),

1. Inflammation (ESR, hs-CRP, ferritin)
2. Nutrition (serum albumin, blood urea, and creatinine) and
3. CKD-mineral bone disease (calcium, phosphorus, alkaline phosphatase, intact parathyroid hormone.

The prevalence of PH in our study population was 89%. None of the previously published studies had shown such a high prevalence in maintenance HD patients or in conservatively managed patients. In studies conducted in stage 5 CKD patients based on echocardiogram for diagnosis, the reported prevalence varied from 16%–64.4%. Few representative studies conducted in stage 5 CKD are summarized in Table 5.

Table 5: Studies reporting prevalence of pulmonary hypertension in chronic kidney diseases.

We consider the following reasons as the possible cause of the reported high prevalence in our study.

1. Use of PAAT-based measurement of PAP
It is very evident that the addition of this method has picked up 24 more cases than the conventional method and altogether improved the sensitivity of the echo-based PH diagnosis. Continuous-wave Doppler to measure the maximum velocity of tricuspid regurgitation (TRVmax) is widely used for PAP pressure estimation as it provides a direct estimate of the right ventricular systolic pressure (RVSP) which correlates with invasive hemodynamic measurements. When the TR jet is absent or trivial, one of the alternate methods proposed is the measurement of PAAT. This method is attractive as it does not depend on any anatomic defect or valve regurgitation and so making it possible to be measured in all cases. Yared et al showed in their paper that PAAT inversely correlates with TRVmax and EPSPAP. An equation to calculate EPSPAP was devised from linear regression analysis and the data were validated in a population of 498 patients (log10[EPSPAP] = _0.004 [PAAT] + 2.1).^[27]
There are other studies both from India and abroad supporting the relationship between PAAT and PAP.^[28] Inclusion of this method is definitely the main reason for detecting more cases of PH in our study.

2. Meticulously performed echocardiograms dedicated to PAP measurement
Right heart catheterization is the gold standard for the diagnosis of PH and accurately determine the severity of the hemodynamic derangements. PH is confirmed when the mean PAP is ≥25 mm Hg at rest.^[29] Most of the studies relied upon echo for the diagnosis of PH and PAP is estimated from the maximum TR jet velocity (TRVmax). Errors can occur in patient with only trivial TR as both underestimation and overestimation have been reported.^[30] In addition, in a casually performed echo PH may be overlooked especially when it is mild.

3. High prevalence of PH in CKD in the population studied Our population may be at higher risk of vascular diseases associated with CKD and hence may have a higher prevalence of PH. However, at present, this remains a hypothesis as we do not have data about baseline PH prevalence of our population, and definitely warrants further investigations.
4. The use of a lower threshold PAP value for PH diagnosis Different studies employed varying cut off PAP to diagnose PH ranging from 30–45 mm Hg although the suggested cutoff is 25 mm Hg.^[31] We used a cutoff of 30 based on our hospital protocol and might have led to the diagnosis of few additional cases of mild PH. Still the prevalence would drop only to 85% even if we choose 35 mmHg as the cutoff and which is still higher than the prevalence reported in other studies.
Few studies done with right heart catheterization as a diagnostic tool had reported a higher prevalence of PH.^[13]

A good number of patients with PH have diastolic and systolic dysfunction and this indicates that undiagnosed left ventricular dysfunction is the likely major cause of unexplained PH.

Most of the risk factors cited in the literature were not found to be significantly different in the PH group except for anemia, dialysis as the treatment modality, duration of dialysis, and systemic hypertension. We hope studies with dedicated design would examine this in future.

The important point we infer from this study is that PH is a near-universal phenomenon accompanying CKD, although severe and symptomatic forms are rare. The use of carefully performed echocardiogram in the diagnosis cannot be overemphasized as the most important noninvasive tool, as PAAT-based measurement can diagnose PH with exceptional sensitivity. The study also hints at undiagnosed LV dysfunction as a potential cause for the high prevalence of unexplained PH in CKD.

The limitations of the study are mainly due to its cross-sectional design which precludes us from drawing any definite conclusions about risk factors. The prognosis of patients with PH, including their cardiovascular mortality, survival on dialysis, effect of renal transplantation, and role of pharmacological agents for PH are unresolved queries which could not be addressed by our study. We could not perform CT pulmonary angiogram to detect chronic thromboembolism due to ethical and financial reasons, and hence this important cause was not ruled out with certainty. Finally, patients on peritoneal dialysis one of the prevalent mode of treatment in CKD was not included in this study. Studies with the capability to answer these important queries and fill the knowledge gap are the need of the hour.

Conflict of interest:

None declared.