The staggering cardiovascular risk of kidney failure and the disappointing results of very recent and older trials are a sounding board that nephrologists should multiply efforts at identifying modifiable risk factors, to improve the dim health perspectives of dialysis patients. Moving PH from the limbo category (WHO V) where it stands now, to categories of known aetiology, may perhaps be a significant step towards this tantalizing goal.
Keywords: epidemiology, hemodialysis, pulmonary hypertension, risk factors, survival
Abstract
Background
The prevalence, determinants and prognosis of pulmonary hypertension among long-term hemodialysis patients in the USA are poorly understood.
Methods
A cross-sectional survey of prevalence and determinants of pulmonary hypertension was performed, followed by longitudinal follow-up for all-cause mortality. Pulmonary hypertension was defined as an estimated systolic pulmonary artery pressure of >35 mmHg using echocardiograms performed within an hour after the end of dialysis.
Results
Prevalent in 110/288 patients (38%), the independent determinants of pulmonary hypertension were the following: left atrial diameter (odds ratio 10.1 per cm/m2, P < 0.0001), urea reduction ratio (odds ratio 0.94 per %, P < 0.01) and vitamin D receptor activator use (odds ratio 0.41 for users, P < 0.01). Over a median follow-up of 2.15 years, 97 (34%) patients died yielding a crude mortality rate (CMR) of 114.2 per 1000 patient-years. Of these, 58 deaths occurred among 110 patients with pulmonary hypertension (53%, CMR 168.9/1000 patient-years) and 39 among 178 without pulmonary hypertension (22%, CMR 52.5/1000 patient-years) [unadjusted hazard ratio (HR) for death 2.12 (95% confidence interval 1.41–3.19), P < 0.001]. After multivariate adjustment, pulmonary hypertension remained an independent predictor for all-cause mortality [HR 2.17 (95% confidence interval 1.31–3.61), P < 0.01].
Conclusions
Among hemodialysis patients, pulmonary hypertension is common and is strongly associated with an enlarged left atrium and poor long-term survival. Reducing left atrial size such as through volume control may be an attractive target to improve pulmonary hypertension. Improving pulmonary hypertension in this group of patients may improve the dismal outcomes.
Introduction
Among hemodialysis patients, the epidemiology of pulmonary hypertension has been described in relatively small studies from outside the USA [1–9]. In these studies, a quarter to half of the patients studied had pulmonary hypertension. Only one study has reported the prevalence and outcomes of pulmonary hypertension from the USA [10]. Among these patients, half had evidence of pulmonary hypertension with a fifth having more severe disease. Pulmonary hypertension was associated with reduced survival. Even in this 90 patient study, the follow-up period was relatively short, extending to just 1 year [10].
Pulmonary hypertension is not only common, but its association with poor survival extends to a period beyond dialysis. In a study limited to renal transplant recipients, compared to patients without, severe pulmonary hypertension defined as right ventricular systolic pressure of >50 mmHg conferred a 3.75-fold hazard of death [11]. Thus, prevention of pulmonary hypertension appears to be an attractive target to improve survival in end-stage renal disease (ESRD) patients who are not just treated with hemodialysis but who will also subsequently receive a transplanted kidney.
In this study, the prevalence, determinants and outcomes of pulmonary hypertension among prevalent patients on hemodialysis are described. In these patients, a detailed echocardiographic evaluation has been performed to assess the relationship between pulmonary hypertension and markers of excess extracellular fluid volume.
Materials and methods
Participants
Part of this cohort has previously been described [12, 13]. Patients ≥18 years who had been on chronic hemodialysis for >3 months and were free of vascular, infectious or bleeding complications within 1 month of recruitment who were dialyzed three times a week at one of the four dialysis units in Indianapolis affiliated with Indiana University were enrolled in the study. Those who missed two hemodialysis treatments or more over 1 month, abused drugs, had chronic atrial fibrillation or body mass index of ≥40 kg/m2 were excluded. Patients who had a change in dry weight or anti-hypertensive drugs within 2 weeks were also excluded. The study was approved by the Institutional Review Board of Indiana University and Research and Development Committee of the Roudebush VA Medical Center, Indianapolis and all subjects gave written informed consent.
Measurements
Echocardiograms
Two-dimensional (2D)-guided M-mode echocardiograms were performed by research echocardiographic technicians, 30–60 min following dialysis, in the dialysis unit with a digital cardiac ultrasound machine (Cypress Acuson; Siemens Medical). The post-dialysis period was selected for echocardiography as it allows control over volume state of the patient since it is associated with the least intravascular volume. The day following dialysis would be associated with a variable change in the volume state and therefore pulmonary artery (PA) pressure and was not chosen for echocardiography.
The protocol specified the recording of at least six cycles of 2D parasternal long- and short-axis left atrial views with optimal orientation of the cursor beam used to derive additional M-mode recordings. Each patient underwent six M-mode measurements of inferior vena cava in inspiration and expiration, left atrial diameter in end systole, interventricular septal thickness in diastole (IVSTd), left ventricular (LV) internal diameter in diastole (LVIDd) and systole (LVIDs), LV posterior wall thickness in diastole (LVPWd) and systole (LVPWs) using standards of the American Society of Echocardiography [14]. All measurements were made over six cardiac cycles by a highly skilled echocardiographer.
Inferior vena cava (IVC) was imaged at the level just below the diaphragm in the hepatic segment by 2D-guided M-mode echocardiography. IVC diameter was measured just before the P-wave of the electrocardiogram during end expiration and end inspiration, while avoiding Valsalva-like maneurvres. The left atrial diameter indexed for body surface area and the IVC diameter in expiration also indexed for body surface area have previously been shown to be markers of volume [15].
The hepatic vein was identified and pulse wave Doppler used to obtain peak systolic and diastolic velocities. From the hepatic vein flow velocity, systolic filling fraction (HVSFF) was derived from peak velocities as peak systolic wave velocity divided by the sum of peak systolic and diastolic velocities [16]. Systolic filling fraction of <55% has been reported to have 86% sensitivity and 90% specificity in predicting mean right atrial pressure of >8 mmHg [16]. Right atrial pressure was calculated as 21.6 − 24 × (HVSFF/100) when HVSFF was >0 but <90. If HVSFF was not available or outside this range, then right atrial pressure was assumed to be 10 mmHg.
LV mass was calculated with a previously validated formula [17]: LV mass (g) = 0.832 × [(IVSTd + LVIDd + PWTd)3 − (LVIDd)3] + 0.60. This was indexed for body surface area to yield the LV mass index.
Midwall fractional shortening (mWFS) is a well-established marker of systolic ventricular dysfunction in patients with LV hypertrophy and is more sensitive than endocardial fractional shortening in detecting systolic dysfunction. In patients on hemodialysis, mWFS and change in mWFS is reported to be of prognostic importance [18]. mWFS was estimated by the elliptic model of LV geometry as described by de Simone et al. [19].
Stroke volume was calculated from the cross-sectional area of the aortic annulus, and the time-velocity integral of aortic annular flow was obtained by the pulsed Doppler recording as previously described [20]. Cardiac output was then calculated by multiplying stroke volume by heart rate. This procedure for echocardiographic determination of cardiac output has been validated against the thermodilution technique (r = 0.87–0.96) [20]. Cardiac output was divided by body surface area to yield the cardiac index. In our laboratory, this technique has excellent day-to-day reproducibility (r = 0.93, coefficient of variation = 5%).
Blood pressure measurements
Ambulatory blood pressure (BP) monitoring was performed either after the first or midweek hemodialysis session for 44 h. Ambulatory BP was recorded every 20 min during the day (6 AM–10 PM) and every 30 min during the night (10 PM–6 AM) using a Spacelab 90207 ambulatory BP monitor (SpaceLabs Medical Inc., Redmond, WA) in the non-access arm, as reported previously [21]. In this study, patients who had <8 h of ambulatory BP recordings were noted to have inadequate measurement and were excluded.
Dialysis unit BP recordings as measured by the dialysis unit staff before and after dialysis were collected prospectively at the time of the patient visit. These BP recordings were obtained using the sphygmomanometer equipped with hemodialysis machines without a specified technique and were averaged over 2 weeks. Thus, each patient had six pre-dialysis and six post-dialysis BP recordings to provide routine dialysis unit BP.
Data analysis
Descriptive statistics for demographic, clinical and hemodynamic variables related to the prevalence of pulmonary hypertension were calculated. Race was combined into two categories black and non-black. Dialysis vintage was categorized into three groups dialysis less than a year, dialysis 1–4 years and dialysis >4 years. The number of anti-hypertensives was capped at four, as generally few patients were on more than four medications. History of cardiovascular disease was defined as previous myocardial infarction, stroke, percutaneous coronary intervention or coronary artery bypass graft.
Odds ratios based on logistic regression for each covariate (demographic, clinical or hemodynamic) were computed. Those covariates with a P-value <0.2 were considered for the multivariate analysis. Stepwise forward selection logistic regression was performed with factors added at the 0.15 level of significance.
Nextly, we constructed Cox proportional regression models with each of the clinical, demographic and hemodynamic markers used in the cross-sectional study. An approach similar to the one used in the above analysis was used. In one instance where bivariate relationships were significant for two highly correlated variables (pre-dialysis diastolic BP and post-dialysis diastolic BP), we entered only one into factor into the multivariate model (using the factor with larger chi-squared likelihood ratio which was post-dialysis diastolic BP). Patients were censored on the date of transplantation (37 patients) or if they recovered renal function (1 patient). The proportionality assumption for the model was tested using Schoenfeld residuals; no violation was found.
All analyses were conducted using Stata 11.0 (Stata Corp, College Station, TX). The P-values reported are two-sided and taken to be significant at <0.05.
Results
Between September 2003 and March 2011, patients from four dialysis units staffed by the nephrology faculty of Indiana University, Indianapolis, were recruited. The study flow is shown in Figure 1. The epidemiology of pulmonary hypertension is described using 288 patients who had adequate echocardiograms.
Fig. 1.
Participant flow for the cohort study.
Table 1 shows the prevalence and bivariate determinants of pulmonary hypertension. The prevalence of hypertension was 38% (110/288). More severe pulmonary hypertension (PA systolic pressure of at least 45 mmHg) was seen in 47/288 (16%). Significant bivariate determinants of pulmonary hypertension were an older age and vitamin D receptor activator use.
Table 1.
Descriptive characteristics of the study population and bivariate ORs for prevalence of pulmonary hypertensiona
| Characteristic | PHTN + | PHTN − | OR (95% CI) | P-value | Total |
|---|---|---|---|---|---|
| N | 110 (38%) | 178 (62%) | 288 (100%) | ||
| Age (years) | 56.5 ± 12.6 | 53.1 ± 12.7 | 1.02 (1.00–1.04) | 0.03 | 54.4 ± 12.8 |
| Male | 72 (65%) | 112 (63%) | 1.12 (0.68–1.84) | 0.7 | 184 (64%) |
| Racial category | 0.2 | ||||
| Non-black | 14 (13%) | 32 (18%) | 1.00 (ref cat) | 46 (16%) | |
| Black | 96 (87%) | 146 (82%) | 1.50 (0.76–2.96) | 0.2 | 242 (84%) |
| Dialysis access | 0.3 | ||||
| Fistula | 54 (49%) | 72 (40%) | 1.60 (0.87–2.94) | 0.1 | 126 (44%) |
| Graft | 31 (28%) | 40 (22%) | 1.65 (0.83–3.27) | 0.1 | 71 (25%) |
| Catheter | 23 (21%) | 49 (28%) | 1.00 (ref cat) | 72 (25%) | |
| History of smoking | 0.4 | ||||
| Current | 40 (36%) | 59 (33%) | 1.33 (0.74–2.40) | 0.3 | 99 (34%) |
| Past | 37 (34%) | 49 (28%) | 1.49 (0.81–2.72) | 0.2 | 86 (30%) |
| Never | 32 (29%) | 63 (35%) | 1.00 (ref cat) | 95 (33%) | |
| History of cardiovascular disease | 44 (40%) | 54 (30%) | 1.47 (0.89–2.42) | 0.1 | 98 (34%) |
| History of diabetes mellitus | 43 (39%) | 87 (49%) | 0.64 (0.40–1.04) | 0.07 | 130 (45%) |
| Years on dialysis | 0.1 | ||||
| <1 | 25 (23%) | 57 (32%) | 1.00 (ref cat) | 82 (28%) | |
| 1–4 | 47 (43%) | 74 (42%) | 1.45 (0.80–2.63) | 0.2 | 121 (42%) |
| 4+ | 38 (35%) | 44 (25%) | 1.97 (1.04–3.73) | 0.04 | 82 (28%) |
| Pre-HD weight (kg) | 83.1 ± 18.3 | 83.5 ± 19.0 | 1.00 (0.99–1.01) | 0.9 | 83.3 ± 18.7 |
| Post-HD weight (kg) | 79.5 ± 17.5 | 80.3 ± 18.6 | 1.00 (0.98–1.01) | 0.7 | 80.0 ± 18.2 |
| Body mass index (kg/m2) | 28.2 ± 5.8 | 28.4 ± 6.0 | 0.99 (0.95–1.04) | 0.7 | 28.3 ± 5.9 |
| Etiology of ESRD | 0.4 | ||||
| Diabetes mellitus | 36 (33%) | 63 (35%) | 1.00 (ref cat) | 99 (34%) | |
| Hypertensive nephrosclerosis | 58 (53%) | 76 (43%) | 1.34 (0.78–2.28) | 0.3 | 134 (47%) |
| Glomerulonephritis | 7 (6%) | 8 (4%) | 1.53 (0.51–4.57) | 0.4 | 15 (5%) |
| Adult autosomal polycystic kidney disease | 1 (1%) | 2 (1%) | 0.88 (0.08–9.99) | 0.9 | 3 (1%) |
| Other | 8 (7%) | 24 (13%) | 0.58 (0.24–1.43) | 0.2 | 32 (11%) |
| Calcium (mg/dL) | 9.0 ± 0.8 | 9.0 ± 0.8 | 1.01 (0.74–1.37) | 1 | 9.0 ± 0.8 |
| Phosphorus (mg/dL) | 5.3 ± 1.6 | 5.5 ± 1.8 | 0.94 (0.82–1.08) | 0.4 | 5.5 ± 1.8 |
| PTH intact (ng/L) | 416.9 ± 379.2 | 478.8 ± 573.3 | 1.00 (0.77–1.31)b | 1 | 453.6 ± 503.5 |
| Urea reduction ratio (%) | 73.1 ± 7.1 | 74.9 ± 7.8 | 0.97 (0.94–1.00) | 0.07 | 74.2 ± 7.6 |
| Serum albumin (g/dL) | 3.7 ± 0.4 | 3.7 ± 0.5 | 1.33 (0.77–2.32) | 0.3 | 3.7 ± 0.4 |
| Hemoglobin (g/dL) | 11.9 ± 1.5 | 12.2 ± 1.5 | 0.90 (0.76–1.06) | 0.2 | 12.1 ± 1.5 |
| On anti-hypertensive medications | 91 (83%) | 142 (80%) | 1.10 (0.58–2.09) | 0.8 | 233 (81%) |
| Number of anti-hypertensives | 0.6 | ||||
| 0 anti-hypertensives | 18 (16%) | 31 (17%) | 1.00 (ref cat) | 49 (17%) | |
| 1 anti-hypertensives | 30 (27%) | 35 (20%) | 1.48 (0.69–3.15) | 0.3 | 65 (23%) |
| 2 anti-hypertensives | 28 (25%) | 43 (24%) | 1.12 (0.53–2.38) | 0.8 | 71 (25%) |
| 3 anti-hypertensives | 12 (11%) | 28 (16%) | 0.74 (0.30–1.80) | 40 (14%) | |
| ≥4 anti-hypertensives | 21 (19%) | 36 (20%) | 1.00 (0.46–2.22) | 1 | 57 (20%) |
| Aspirin use | 47 (43%) | 66 (37%) | 1.23 (0.75–2.00) | 0.4 | 113 (39%) |
| Statin use | 37 (34%) | 74 (42%) | 0.69 (0.42–1.13) | 0.1 | 111 (39%) |
| Vitamin D receptor activator use | 24 (22%) | 77 (43%) | 0.35 (0.20–0.61) | <0.001 | 101 (35%) |
| Epoetin use | 61 (55%) | 104 (58%) | 0.84 (0.52–1.37) | 0.5 | 165 (57%) |
aHD, hemodialysis; PTH, parathyroid hormone; OR, odds ratio.
bOR computed using log-transformed values.
Table 2 shows the bivariate hemodynamic (BP and echocardiographic) determinants of pulmonary hypertension. Mean PA systolic pressure was 33.9 mmHg (44.7 mmHg among those with and 27.2 mmHg among those without pulmonary hypertension). The right atrial pressure was not significantly different between groups. Significant bivariate determinants of pulmonary hypertension were the following: higher ambulatory systolic BP, increased inferior vena cava diameter in expiration, increased left atrial diameter, higher mWFS and increased cardiac index.
Table 2.
Hemodynamic characteristics of the study population and bivariate ORs for prevalence of pulmonary hypertensiona
| Characteristic | PHTN + | PHTN − | OR (95% CI) | P-value | Total |
|---|---|---|---|---|---|
| N | 110 (38%) | 178 (62%) | 288 (100%) | ||
| PA systolic pressure (mmHg) | 44.7 ± 7.9 | 27.2 ± 4.9 | N/A | 33.9 ± 10.6 | |
| Right atrial pressure (mmHg) | 8.1 ± 2.0 | 8.3 ± 1.8 | 0.95 (0.84–1.08) | 0.4 | 8.2 ± 1.9 |
| Ambulatory systolic BP (mmHg) | 138.9 ± 24.1 | 133.7 ± 18.5 | 1.01 (1.00–1.02) | 0.05 | 135.7 ± 21.0 |
| Ambulatory diastolic BP (mmHg) | 78.4 ± 16.5 | 77.8 ± 12.0 | 1.00 (0.99–1.02) | 0.7 | 78.0 ± 13.9 |
| Ambulatory heart rate (b.p.m.) | 80.0 ± 11.4 | 79.9 ± 12.2 | 1.00 (0.98–1.02) | 0.9 | 79.9 ± 11.9 |
| Pre-HD systolic BP (mmHg) | 149.5 ± 24.2 | 150.7 ± 20.0 | 1.00 (0.99–1.01) | 0.7 | 150.2 ± 21.6 |
| Pre-HD diastolic BP (mmHg) | 81.6 ± 16.2 | 82.3 ± 13.3 | 1.00 (0.98–1.01) | 0.7 | 82.0 ± 14.5 |
| Pre-HD heart rate (b.p.m.) | 79.0 ± 12.4 | 78.8 ± 10.8 | 1.00 (0.98–1.02) | 0.9 | 78.9 ± 11.4 |
| Post-HD systolic BP (mmHg) | 137.9 ± 24.2 | 135.9 ± 19.7 | 1.00 (0.99–1.02) | 0.5 | 136.6 ± 21.5 |
| Post-HD diastolic BP (mmHg) | 74.5 ± 14.6 | 75.4 ± 11.4 | 0.99 (0.97–1.01) | 0.6 | 75.1 ± 12.7 |
| Post-HD heart rate (b.p.m.) | 77.7 ± 10.3 | 78.9 ± 11.2 | 0.99 (0.97–1.01) | 0.4 | 78.4 ± 10.9 |
| Inferior vena cava at expiration (mm/m2) | 8.8 ± 3.3 | 7.5 ± 3.0 | 1.13 (1.05–1.23) | <0.01 | 8.0 ± 3.2 |
| Left atrial diameter (cm/m2) | 2.4 ± 0.4 | 2.0 ± 0.4 | 7.20 (3.74–13.84) | <0.0001 | 2.2 ± 0.4 |
| LV mass index (g/m2) | 147.3 ± 44.1 | 144.1 ± 48.1 | 1.00 (1.00–1.01) | 0.6 | 145.3 ± 46.5 |
| mWFS (observed) (%) | 13.7 ± 2.9 | 12.6 ± 3.3 | 1.12 (1.03–1.21) | <0.01 | 13.0 ± 3.2 |
| Cardiac index (L/min/m2) | 2.9 ± 1.0 | 2.5 ± 0.8 | 1.63 (1.21–2.22) | <0.01 | 2.7 ± 0.9 |
aHD, hemodialysis; OR, odds ratio.
The multivariate determinants of prevalence of pulmonary hypertension are shown in Table 3. Only three variables, left atrial diameter, urea reduction ratio and vitamin D receptor activator use, were significant determinants of pulmonary hypertension. The strongest association was seen with increased left atrial diameter.
Table 3.
Multivariate ORs for prevalence of pulmonary hypertension
| Determinants | OR (95% CI) | P-value |
|---|---|---|
| Left atrial diameter (cm/m2) | 10.10 (4.29–23.74) | <0.0001 |
| Urea reduction ratio (%) | 0.94 (0.90–0.98) | <0.01 |
| Vitamin D receptor activator use (yes/no) | 0.41 (0.21–0.81) | <0.01 |
| LR chi-square | 54.59 | |
| Pseudo r2 | 0.18 |
aOR, odds ratio; CI, confidence interval.
Over a median follow-up of 2.15 years, 97 (34%) died yielding a crude mortality rate (CMR) of 114.2 per 1000 patient-years. Of these, 58 deaths occurred among 110 patients with pulmonary hypertension (53%, CMR 168.9/1000 patient-years) and 39 among 178 without pulmonary hypertension (22%, CMR 52.5/1000 patient-years). The cumulative hazard estimates are shown in Figure 2.
Fig. 2.
Cumulative hazard estimates for all-cause mortality for participants with (dotted line) and without (solid line) pulmonary hypertension. The table at the bottom reflects the number at risk at each time point.
Table 4 shows the bivariate hazard ratios (HRs) for all-cause mortality. Significant bivariate predictors for all-cause mortality were the following: pulmonary hypertension, age, dialysis access, history of cardiovascular disease, history of diabetes mellitus and vitamin D receptor activator use. Hemodynamic variables associated with all-cause mortality (Table 5) were the following: low pre-dialysis diastolic BP, low post-dialysis diastolic BP, increased inferior vena cava diameter in expiration, increased left atrial diameter and increased cardiac index.
Table 4.
Descriptive characteristics of the study population and bivariate HRs for deatha
| Characteristic | Dead | Alive | HR (95% CI) | P-value | Total |
|---|---|---|---|---|---|
| N | 97 (34%) | 191 (66%) | 288 (100%) | ||
| Pulmonary hypertension | <0.001 | ||||
| Yes | 58 (60%) | 52 (27%) | 2.12 (1.41–3.19) | 110 (38%) | |
| No | 39 (40%) | 139 (73%) | 1.00 (ref cat) | 178 (62%) | |
| Age (years) | 60.0 ± 13.2 | 51.5 ± 11.6 | 1.04 (1.03–1.06) | <0.0001 | 54.4 ± 12.8 |
| Gender | 0.7 | ||||
| Female | 38 (39%) | 66 (35%) | 1.00 (ref cat) | 104 (36%) | |
| Male | 59 (61%) | 125 (65%) | 0.93 (0.62–1.40) | 184 (64%) | |
| Racial category | 0.07 | ||||
| Non-black | 13 (13%) | 33 (17%) | 1.00 (ref cat) | 46 (16%) | |
| Black | 84 (87%) | 158 (83%) | 0.57 (0.32–1.04) | 242 (84%) | |
| Dialysis access | 0.03 | ||||
| Fistula | 33 (34%) | 93 (49%) | 0.62 (0.37–1.03) | 0.07 | 126 (44%) |
| Graft | 35 (36%) | 36 (19%) | 1.16 (0.69–1.92) | 0.6 | 71 (25%) |
| Catheter | 26 (27%) | 46 (24%) | 1.00 (ref cat) | 72 (25%) | |
| History of smoking | 1 | ||||
| Current | 36 (37%) | 63 (33%) | 1.01 (0.61–1.66) | 1 | 99 (34%) |
| Past | 33 (34%) | 53 (28%) | 1.05 (0.63–1.75) | 0.8 | 86 (30%) |
| Never | 27 (28%) | 68 (36%) | 1.00 (ref cat) | 95 (33%) | |
| History of cardiovascular disease | <0.01 | ||||
| Yes | 45 (46%) | 53 (28%) | 1.77 (1.19–2.65) | 98 (34%) | |
| No | 52 (54%) | 130 (68%) | 1.00 (ref cat) | 182 (63%) | |
| History of diabetes mellitus | 0.04 | ||||
| Yes | 50 (52%) | 80 (42%) | 1.54 (1.03–2.30) | 130 (45%) | |
| No | 47 (48%) | 107 (56%) | 1.00 (ref cat) | 154 (53%) | |
| Years on dialysis | 0.2 | ||||
| <1 | 19 (20%) | 63 (33%) | 1.00 (ref cat) | 82 (28%) | |
| 1–4 | 41 (42%) | 80 (42%) | 1.24 (0.72–2.14) | 0.4 | 121 (42%) |
| 4+ | 37 (38%) | 45 (24%) | 1.67 (0.96–2.91) | 0.07 | 82 (28%) |
| Pre-HD weight (kg) | 80.5 ± 18.5 | 84.9 ± 18.7 | 0.99 (0.98–1.00) | 0.08 | 83.3 ± 18.7 |
| Post-HD weight (kg) | 77.8 ± 18.2 | 81.1 ± 18.1 | 0.99 (0.98–1.00) | 0.1 | 80.0 ± 18.2 |
| Body mass index (kg/m2) | 28.1 ± 6.1 | 28.5 ± 5.8 | 0.98 (0.95–1.02) | 0.4 | 28.3 ± 5.9 |
| Etiology of ESRD | 0.5 | ||||
| Diabetes mellitus | 38 (39%) | 61 (32%) | 1.00 (ref cat) | 99 (34%) | |
| Hypertensive nephrosclerosis | 44 (45%) | 90 (47%) | 0.75 (0.48–1.16) | 0.2 | 134 (47%) |
| Glomerulonephritis | 4 (4%) | 11 (6%) | 0.60 (0.21–1.68) | 0.3 | 15 (5%) |
| Adult autosomal polycystic kidney disease | 1 (1%) | 2 (1%) | 1.19 (0.16–8.66) | 0.9 | 3 (1%) |
| Other | 10 (10%) | 22 (12%) | 1.16 (0.58–2.33) | 0.7 | 32 (11%) |
| Calcium (mg/dL) | 9.1 ± 0.7 | 9.0 ± 0.9 | 1.20 (0.95–1.51) | 0.1 | 9.0 ± 0.8 |
| Phosphorus (mg/dL) | 5.2 ± 1.7 | 5.6 ± 1.8 | 0.96 (0.85–1.08) | 0.5 | 5.5 ± 1.8 |
| PTH intact (ng/L) | 438.7 ± 465.3 | 460.8 ± 522.0 | 1.01 (0.79–1.28)b | 0.9 | 453.6 ± 503.5 |
| Urea reduction ratio (%) | 74.4 ± 7.3 | 74.0 ± 7.7 | 1.02 (0.99–1.05) | 0.1 | 74.2 ± 7.6 |
| Serum albumin (g/dL) | 3.7 ± 0.4 | 3.7 ± 0.5 | 0.65 (0.42–1.02) | 0.06 | 3.7 ± 0.4 |
| Hemoglobin (g/dL) | 12.2 ± 1.5 | 12.0 ± 1.5 | 0.92 (0.80–1.06) | 0.2 | 12.1 ± 1.5 |
| On anti-hypertensive medications | 0.9 | ||||
| Yes | 76 (78%) | 157 (82%) | 0.98 (0.60–1.61) | 233 (81%) | |
| No | 20 (21%) | 29 (15%) | 1.00 (ref cat) | 49 (17%) | |
| Number of anti-hypertensives | 0.3 | ||||
| 0 anti-hypertensives | 20 (21%) | 29 (15%) | 1.00 (ref cat) | 49 (17%) | |
| 1 anti-hypertensives | 27 (28%) | 38 (20%) | 0.96 (0.54–1.72) | 0.9 | 65 (23%) |
| 2 anti-hypertensives | 27 (28%) | 44 (23%) | 1.37 (0.77–2.46) | 0.3 | 71 (25%) |
| 3 anti-hypertensives | 7 (7%) | 33 (17%) | 0.66 (0.28–1.57) | 0.4 | 40 (14%) |
| ≥4 anti-hypertensives | 15 (15%) | 42 (22%) | 0.79 (0.40–1.54) | 0.5 | 57 (20%) |
| Aspirin use | 0.7 | ||||
| Yes | 41 (42%) | 72 (38%) | 1.08 (0.72–1.61) | 113 (39%) | |
| No | 55 (57%) | 114 (60%) | 1.00 (ref cat) | 169 (59%) | |
| Statin use | 0.4 | ||||
| Yes | 41 (42%) | 70 (37%) | 1.20 (0.80–1.80) | 111 (39%) | |
| No | 55 (57%) | 116 (61%) | 1.00 (ref cat) | 171 (59%) | |
| Vitamin D receptor activator use | <0.01 | ||||
| Yes | 15 (15%) | 86 (45%) | 0.48 (0.28–0.84) | 101 (35%) | |
| No | 81 (84%) | 100 (52%) | 1.00 (ref cat) | 181 (63%) | |
| Epoetin use | 0.7 | ||||
| Yes | 54 (56%) | 111 (58%) | 1.10 (0.73–1.65) | 165 (57%) | |
| No | 42 (43%) | 75 (39%) | 1.00 (ref cat) | 117 (41%) |
aHD, hemodialysis; PTH, parathyroid hormone; CI, confidence interval.
bHR computed using log-transformed values.
Table 5.
Hemodynamic characteristics of the study population and bivariate HRs for deatha
| Characteristic | Dead | Alive | HR (95% CI) | P-value | Total |
|---|---|---|---|---|---|
| N | 97 (34%) | 191 (66%) | 288 (100%) | ||
| Ambulatory systolic BP (mmHg) | 135.5 ± 22.0 | 135.8 ± 20.5 | 1.01 (1.00–1.02) | 0.07 | 135.7 ± 21.0 |
| Ambulatory diastolic BP (mmHg) | 74.5 ± 14.0 | 79.9 ± 13.5 | 0.99 (0.98–1.01) | 0.3 | 78.0 ± 13.9 |
| Ambulatory heart rate (b.p.m.) | 78.5 ± 11.6 | 80.7 ± 11.9 | 0.99 (0.97–1.01) | 0.2 | 79.9 ± 11.9 |
| Pre-HD systolic BP (mmHg) | 147.3 ± 22.4 | 151.8 ± 21.1 | 1.00 (0.99–1.01) | 0.9 | 150.2 ± 21.6 |
| Pre-HD diastolic BP (mmHg) | 77.4 ± 13.5 | 84.5 ± 14.4 | 0.98 (0.96–0.99) | <0.01 | 82.0 ± 14.5 |
| Pre-HD heart rate (b.p.m.) | 78.6 ± 11.2 | 79.1 ± 11.5 | 1.00 (0.98–1.02) | 0.9 | 78.9 ± 11.4 |
| Post-HD systolic BP (mmHg) | 135.3 ± 21.6 | 137.3 ± 21.4 | 1.00 (0.99–1.01) | 0.6 | 136.6 ± 21.5 |
| Post-HD diastolic BP (mmHg) | 71.2 ± 11.4 | 77.2 ± 12.9 | 0.97 (0.95–0.99) | <0.01 | 75.1 ± 12.7 |
| Post-HD heart rate (b.p.m.) | 76.6 ± 10.9 | 79.4 ± 10.8 | 0.99 (0.97–1.01) | 0.2 | 78.4 ± 10.9 |
| Inferior vena cava at expiration (mm/m2) | 8.9 ± 3.7 | 7.6 ± 2.8 | 1.07 (1.02–1.13) | <0.01 | 8.0 ± 3.2 |
| Left atrial diameter (cm/m2) | 2.3 ± 0.5 | 2.1 ± 0.4 | 2.22 (1.39–3.55) | <0.001 | 2.2 ± 0.4 |
| LV mass index (g/m2) | 137.9 ± 42.0 | 149.1 ± 48.3 | 1.00 (1.00–1.00) | 1 | 145.3 ± 46.5 |
| mWFS (observed) (%) | 13.3 ± 2.9 | 12.9 ± 3.3 | 0.97 (0.91–1.03) | 0.3 | 13.0 ± 3.2 |
| Cardiac index (L/min/m2) | 3.0 ± 1.0 | 2.5 ± 0.8 | 1.29 (1.05–1.58) | 0.02 | 2.7 ± 0.9 |
aHD, hemodialysis; CI, confidence interval; mWFS, midwall fractional shortening.
Table 6 shows the multivariate HR for all-cause mortality. Pulmonary hypertension was associated with increased risk for all-cause mortality. Besides this, other risk factors included the following: age, race, access type, serum albumin and history of cardiovascular disease. Forcing IVC diameter in expiration or left atrial diameter reduced the HRs slightly (HRs for death with pulmonary hypertension were 1.99 and 2.02, respectively) but did not remove the statistical significance of pulmonary hypertension.
Table 6.
Multivariate HR for death
| Determinants | HR (95% CI) | P-value |
|---|---|---|
| Pulmonary hypertension | 2.17 (1.31–3.61) | <0.01 |
| Age (years) | 1.03 (1.01–1.05) | 0.02 |
| Black | 0.51 (0.24–1.10) | 0.09 |
| Access type | 0.06 | |
| Catheter | 1.00 (ref cat) | |
| Fistula | 0.54 (0.29–1.01) | 0.05 |
| Graft | 1.05 (0.57–1.94) | 0.9 |
| Serum albumin (g/dL) | 0.65 (0.35–1.18) | 0.2 |
| History of cardiovascular disease | 1.67 (1.02–2.75) | 0.04 |
| LR chi-square | 44.10 |
aCI, confidence interval; LR, Likelihood ratio.
Discussion
The major findings of our study are the following: (i) the prevalence of pulmonary hypertension among ESRD patients on hemodialysis is high (38%); (ii) the independent determinants of pulmonary hypertension are enlarged left atrial diameter, low urea reduction ratio and no vitamin D receptor activator use and (iii) pulmonary hypertension is associated with increased hazards for all-cause mortality >2-fold independent of other risk factors.
Prevalence of pulmonary hypertension
The prevalence of pulmonary hypertension has ranged from 27 [6] to 56% [5]; our study found the prevalence in the middle of the range of 38%. Prevalence estimates are variable due to several reasons. Firstly, the timing of performance of echocardiograms may influence prevalence estimates. The interdialytic increase in volume may increase PA pressures. For example, the lower estimates from Egypt of 29% may be because measurements were performed within 4 h of the end of dialysis [1], whereas higher estimates of 58.6% from Italy may be related to performance of echocardiograms the day following dialysis [7]. Secondly, varying definitions of pulmonary hypertension may account for varying prevalence estimates. For example, in a study from the USA, using a PA systolic pressure of ≥35 mmHg yielded a prevalence of 47% pulmonary hypertension [10]. When pulmonary hypertension was defined as a PA systolic pressure of ≥45 mmHg, prevalence was only 20% [10]. This is similar to the 16% prevalence estimate seen in our study, when pulmonary hypertension was defined as PA systolic pressure of ≥45 mmHg. Thirdly, variable degrees of chronic volume overload may also confound the prevalence. Since there are no agreed upon markers for chronic volume overload, it is difficult to account for this variable among studies.
Determinants of pulmonary hypertension
Independent determinants of pulmonary hypertension vary according to the study and have been noted to be the following: increased age [10], female gender [1], lower body mass index [10], higher cardiac output [2, 3], lower hemoglobin [2], lower nitric oxide metabolites [3], higher dialysis vintage [7], lower diastolic BP [7] and LV diastolic dysfunction [8]. In our study, we found left atrial diameter to be strongly associated with pulmonary hypertension. This may reflect either chronic volume overload or poor myocardial performance. Poor systolic myocardial performance is unlikely to be a reason because those with pulmonary hypertension had a greater mWFS and a better cardiac index indicating better systolic myocardial function. LV mass index was similar between the two groups therefore a strong case cannot be made for greater diastolic dysfunction among those with pulmonary hypertension. Since inferior vena cava diameter in expiration was elevated in those with pulmonary hypertension, chronic volume overload may be more likely to be the cause of association between elevated left atrial diameter and pulmonary hypertension. Under-dialysis, as indicated by low urea reduction ratio, may aggravate this situation. An increase in circulating inhibitors of nitric oxide synthase may accumulate and low availability of nitric oxide may aggravate pulmonary hypertension [3]. The association between vitamin D receptor activator use and lower odds for pulmonary hypertension is also biologically plausible. Vitamin D receptor activators can improve diastolic function and are therefore likely to reduce afterload for the right ventricle [22]. Use of these drugs therefore may be causally associated with less pulmonary hypertension (although we did not find differences in LV mass index between groups). As noted by several other investigators, we did not find calcium, phosphorus or parathyroid hormone to be associated with pulmonary hypertension [1, 3, 6–9].
Prognosis of pulmonary hypertension
There are only two studies that have examined the association of pulmonary hypertension with overall survival in studies with >75 patients. Yigla et al. [4] followed 127 hemodialysis patients among which pulmonary hypertension defined as a PA systolic pressure of 45 mmHg was seen in 37 (29%); of these, 17 had pulmonary hypertension before starting dialysis, whereas 20 had incident pulmonary hypertension. Both incident and prevalent pulmonary hypertension was associated with increased all-cause mortality (HRs 2.4 and 3.6, respectively). Ramasubbu et al. [10] followed 90 hemodialysis patients prospectively after performing echocardiogram just before a dialysis session. Pulmonary hypertension defined as tricuspid regurgitation jet of ≥2.5 m/s (that correlates with PA systolic pressure of ≥35 mmHg) was seen in 47%. At 12 months, mortality was 26% in those with pulmonary hypertension and 6% in those without pulmonary hypertension (log-rank test significant). All-cause hospitalizations was not different between groups.
The results of this larger study support the results of the above investigators. Prevalent pulmonary hypertension was associated with ∼2-fold hazards for all-cause mortality. These results are particularly important because pulmonary hypertension also portends a poor survival among renal transplant recipients [11]. For example, recipients of renal transplants had nearly 4-fold higher hazards for death when severe pulmonary hypertension was present [11]. Adjustment for IVC diameter or left atrial diameter did not remove the statistical significance of pulmonary hypertension. This suggests that once pulmonary hypertension is established, markers of volume excess do not influence mortal outcomes.
Limitations
Although this study is the largest study on pulmonary hypertension reported to date, it has a number of limitations. The diagnosis of pulmonary hypertension was only based on indirect echocardiographic estimates of PA systolic pressure. No measurements through right heart catheterizations were made to directly quantify PA pressure to confirm the echocardiographic findings. Estimations of right atrial pressure were made from hepatic vein Doppler studies; these may be erroneous. The etiology of pulmonary hypertension was not accurately determined. Thus, whether pulmonary hypertension reflects diastolic dysfunction, volume overload, sleep apnea, other etiologies or a combination of them is uncertain. Being a non-interventional trial, cause and effect relationship between pulmonary hypertension and mortality cannot be established.
Conclusions
We speculate that pulmonary hypertension among hemodialysis patients may be in response to chronic volume overload. Increase in cardiac index, elevated inferior vena cava diameter and larger left atrial diameter all suggest this possibility. If this is true, then it would be desirable to treat volume excess early among hemodialysis patients because once pulmonary hypertension is established, the long-term survival is poor. We have previously shown that left atrial size and inferior vena cava diameter improve with probing dry weight [15]. Whether reduction in excess volume with probing dry weight can reduce incident pulmonary hypertension should be explored in future studies.
Acknowledgements
Funding. National Institutes of Health (2R01-DK062030-07).
Conflict of interest statement. None declared.
(See related article by Zoccali. Pulmonary hypertension in dialysis patients: a prevalent, risky but still uncharacterized disorder. Nephrol Dial Transplant 2012; 27: 3675–3677.)
References
- 1.Amin M, Fawzy A, Hamid MA, et al. Pulmonary hypertension in patients with chronic renal failure: role of parathyroid hormone and pulmonary artery calcifications. Chest. 2003;124:2093–2097. doi: 10.1378/chest.124.6.2093. [DOI] [PubMed] [Google Scholar]
- 2.Yigla M, Nakhoul F, Sabag A, et al. Pulmonary hypertension in patients with end-stage renal disease. Chest. 2003;123:1577–1582. doi: 10.1378/chest.123.5.1577. [DOI] [PubMed] [Google Scholar]
- 3.Nakhoul F, Yigla M, Gilman R, et al. The pathogenesis of pulmonary hypertension in haemodialysis patients via arterio-venous access. Nephrol Dial Transplant. 2005;20:1686–1692. doi: 10.1093/ndt/gfh840. [DOI] [PubMed] [Google Scholar]
- 4.Yigla M, Fruchter O, Aharonson D, et al. Pulmonary hypertension is an independent predictor of mortality in hemodialysis patients. Kidney Int. 2009;75:969–975. doi: 10.1038/ki.2009.10. [DOI] [PubMed] [Google Scholar]
- 5.Havlucu Y, Kursat S, Ekmekci C, et al. Pulmonary hypertension in patients with chronic renal failure. Respiration. 2007;74:503–510. doi: 10.1159/000102953. [DOI] [PubMed] [Google Scholar]
- 6.Tarrass F, Benjelloun M, Medkouri G, et al. Doppler echocardiograph evaluation of pulmonary hypertension in patients undergoing hemodialysis. Hemodial Int. 2006;10:356–359. doi: 10.1111/j.1542-4758.2006.00129.x. [DOI] [PubMed] [Google Scholar]
- 7.Fabbian F, Cantelli S, Molino C, et al. Pulmonary hypertension in dialysis patients: a cross-sectional Italian study. Int J Nephrol. 2011;2011:283475. doi: 10.4061/2011/283475. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Abdelwhab S, Elshinnawy S., et al. Pulmonary hypertension in chronic renal failure patients. Am J Nephrol. 2008;28:990–997. doi: 10.1159/000146076. [DOI] [PubMed] [Google Scholar]
- 9.Mazdeh MM, Mousavi SA, Yahyazadeh H, et al. Pulmonary hypertension in hemodialysis patients. Saudi J Kidney Dis Transpl. 2008;19:189–193. [PubMed] [Google Scholar]
- 10.Ramasubbu K, Deswal A, Herdejurgen C, et al. A prospective echocardiographic evaluation of pulmonary hypertension in chronic hemodialysis patients in the United States: prevalence and clinical significance. Int J Gen Med. 2010;3:279–286. doi: 10.2147/IJGM.S12946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Issa N, Krowka MJ, Griffin MD, et al. Pulmonary hypertension is associated with reduced patient survival after kidney transplantation. Transplantation. 2008;86:1384–1388. doi: 10.1097/TP.0b013e318188d640. [DOI] [PubMed] [Google Scholar]
- 12.Agarwal R, Brim NJ, Mahenthiran J, et al. Out-of-hemodialysis-unit blood pressure is a superior determinant of left ventricular hypertrophy. Hypertension. 2006;47:62–68. doi: 10.1161/01.HYP.0000196279.29758.f4. [DOI] [PubMed] [Google Scholar]
- 13.Agarwal R, Andersen MJ, Bishu K, et al. Home blood pressure monitoring improves the diagnosis of hypertension in hemodialysis patients. Kidney Int. 2006;69:900–906. doi: 10.1038/sj.ki.5000145. [DOI] [PubMed] [Google Scholar]
- 14.Sahn DJ, DeMaria A, Kisslo J, et al. Recommendations regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation. 1978;58:1072–1083. doi: 10.1161/01.cir.58.6.1072. [DOI] [PubMed] [Google Scholar]
- 15.Agarwal R, Bouldin JM, Light RP, et al. Inferior vena cava diameter and left atrial diameter measure volume but not dry weight. Clin J Am Soc Nephrol. 2011;6:1066–1072. doi: 10.2215/CJN.09321010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Nagueh SF, Kopelen HA, Zoghbi WA, et al. Relation of mean right atrial pressure to echocardiographic and Doppler parameters of right atrial and right ventricular function. Circulation. 1996;93:1160–1169. doi: 10.1161/01.cir.93.6.1160. [DOI] [PubMed] [Google Scholar]
- 17.Devereux RB, Alonso DR, Lutas EM, et al. Echocardiographic assessment of left ventricular hypertrophy: comparison to necropsy findings. Am J Cardiol. 1986;57:450–458. doi: 10.1016/0002-9149(86)90771-x. [DOI] [PubMed] [Google Scholar]
- 18.Zoccali C, Benedetto FA, Tripepi G, et al. Left ventricular systolic function monitoring in asymptomatic dialysis patients: a prospective cohort study. J Am Soc Nephrol. 2006;17:1460–1465. doi: 10.1681/ASN.2005111240. [DOI] [PubMed] [Google Scholar]
- 19.de Simone G, Devereux RB, Roman MJ, et al. Assessment of left ventricular function by the midwall fractional shortening/end-systolic stress relation in human hypertension. J Am Coll Cardiol. 1994;23:1444–1451. doi: 10.1016/0735-1097(94)90390-5. [DOI] [PubMed] [Google Scholar]
- 20.Lewis JF, Kuo LC, Nelson JG, et al. Pulsed Doppler echocardiographic determination of stroke volume and cardiac output: clinical validation of two new methods using the apical window. Circulation. 1984;70:425–431. doi: 10.1161/01.cir.70.3.425. [DOI] [PubMed] [Google Scholar]
- 21.Agarwal R., et al. Supervised atenolol therapy in the management of hemodialysis hypertension. Kidney Int. 1999;55:1528–1535. doi: 10.1046/j.1523-1755.1999.00359.x. [DOI] [PubMed] [Google Scholar]
- 22.Bodyak N, Ayus JC, Achinger S, et al. Activated vitamin D attenuates left ventricular abnormalities induced by dietary sodium in Dahl salt-sensitive animals. Proc Natl Acad Sci U S A. 2007;104:16810–16815. doi: 10.1073/pnas.0611202104. [DOI] [PMC free article] [PubMed] [Google Scholar]