Prevalence and Mortality of Pulmonary Hypertension in ESRD: A Systematic Review and Meta-analysis

Abstract

IntroductionPulmonary hypertension (PH) in the setting of end-stage renal disease (ESRD) has important prognostic and therapeutic consequences. We estimated the prevalence of PH among patients with ESRD and compared mortality between ESRD patients with and without PH.Methods
Two independent reviewers searched three databases using a search strategy built around the medical subject headings of “hypertension, pulmonary” and “kidney failure, chronic.” Keywords and synonyms were also used. Study selection criteria included (1) Enrollment of patients with ESRD undergoing hemodialysis or peritoneal dialysis, (2) Assessment for the presence of PH using transthoracic echocardiography, and (3) Determination of PH prevalence or associated mortality. The primary outcomes were prevalence of PH or associated mortality. The Grading, Recommendations, Assessment, Development, and Evaluation (GRADE) approach was used to rate the quality of evidence.ResultsThe initial search identified 1046 publications, from which 41 studies were selected. The median prevalence of PH identified by echocardiographic criteria among patients with ESRD was 38% (range 8% to 70%), and was significantly increased in patients undergoing hemodialysis (HD) (median 40%, range 16–70%) as compared with peritoneal dialysis (PD) (median 19%, range 8–37%). Meta-analysis demonstrated that overall mortality was higher among ESRD patients with echocardiographic evidence of PH than ESRD patients without echocardiographic evidence of PH (RR 2.02; 95% CI 1.70–2.40).Conclusions
Echocardiographic evidence of PH is common among ESRD patients undergoing dialysis and associated with increased mortality. Identification of those patients with evidence of pulmonary hypertension on transthoracic echocardiography may warrant further evaluation and treatment.

Full text

Vol.:(0123456789)
1 3
Lung (2020) 198:535–545
https://doi.org/10.1007/s00408-020-00355-0
PULMONARY HYPERTENSION
Prevalence andMortality ofPulmonary Hypertension inESRD:
ASystematic Review andMeta‑analysis
NoahC.Schoenberg1 · RahulG.Argula2· ElizabethS.Klings3· KevinC.Wilson3,4· HarrisonW.Farber5
Received: 17 January 2020 / Accepted: 20 April 2020 / Published online: 4 May 2020
© Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
Introduction Pulmonary hypertension (PH) in the setting of end-stage renal disease (ESRD) has important prognostic and
therapeutic consequences. We estimated the prevalence of PH among patients with ESRD and compared mortality between
ESRD patients with and without PH.
Methods Two independent reviewers searched three databases using a search strategy built around the medical subject head-
ings of “hypertension, pulmonary” and “kidney failure, chronic.” Keywords and synonyms were also used. Study selection
criteria included (1) Enrollment of patients with ESRD undergoing hemodialysis or peritoneal dialysis, (2) Assessment for
the presence of PH using transthoracic echocardiography, and (3) Determination of PH prevalence or associated mortal-
ity. The primary outcomes were prevalence of PH or associated mortality. The Grading, Recommendations, Assessment,
Development, and Evaluation (GRADE) approach was used to rate the quality of evidence.
Results The initial search identified 1046 publications, from which 41 studies were selected. The median prevalence of
PH identified by echocardiographic criteria among patients with ESRD was 38% (range 8% to 70%), and was significantly
increased in patients undergoing hemodialysis (HD) (median 40%, range 16–70%) as compared with peritoneal dialysis (PD)
(median 19%, range 8–37%). Meta-analysis demonstrated that overall mortality was higher among ESRD patients with echo-
cardiographic evidence of PH than ESRD patients without echocardiographic evidence of PH (RR 2.02; 95% CI 1.70–2.40).
Conclusions Echocardiographic evidence of PH is common among ESRD patients undergoing dialysis and associated with
increased mortality. Identification of those patients with evidence of pulmonary hypertension on transthoracic echocardiog-
raphy may warrant further evaluation and treatment.
Keywords Hemodialysis· Echocardiography· Peritoneal dialysis· Chronic kidney failure· Pulmonary vascular disease
Introduction
Pulmonary hypertension (PH) is a complex syndrome
defined by an elevated mean pulmonary artery pressure on
right heart catheterization (RHC). It is classified into five
groups, including pulmonary arterial hypertension (Group
1), left-heart disease-associated PH (Group 2), lung dis-
ease- or hypoxemia-induced PH (Group 3), chronic throm-
boembolic PH (Group 4), and multifactorial PH (Group
Abstract Previously Presented: Abstract Poster Presentation,
American Thoracic Society 31 International Meeting, held in San
Diego, California, May 2018.
Electronic supplementary material The online version of this
article (https ://doi.org/10.1007/s0040 8-020-00355 -0) contains
supplementary material, which is available to authorized users.
* Noah C. Schoenberg
nschoenb@bidmc.harvard.edu
1 Division ofPulmonary, Critical Care, andSleep Medicine,
Beth Israel Deaconess Medical Center, 330 Brookline
Avenue, Boston02215, MA, USA
2 Pulmonary, Critical Care, Allergy & Sleep Medicine,
Medical University ofSouth Carolina, Charleston, SC, USA
3 The Pulmonary Center, Boston University School
ofMedicine, Boston, MA, USA
4 American Thoracic Society, NewYork, NY, USA
5 Pulmonary, Critical Care, andSleep Medicine, Tufts Medical
Center, Boston, MA, USA

536 Lung (2020) 198:535–545
1 3
5) [1]. Over the last 15years, it has been increasingly
recognized that chronic kidney disease (CKD), especially
end-stage renal disease (ESRD), is a risk factor for multi-
factorial pulmonary hypertension [2, 3]. The mechanism is
poorly understood, but is likely a combination of chronic
volume overload with pulmonary vascular remodeling,
diastolic dysfunction, elevated cardiac output due to
an arterio-venous fistula (AVF) or chronic anemia, and
chronic inflammation [3]. Furthermore, the presence of PH
in ESRD has been associated with worse clinical outcomes
for patients. This is of significant interest and importance
due to the large numbers of ESRD patients (in 2016 there
were more than 700,000 patients with ESRD in the US
alone, with prevalence increasing by approximately 20,000
per year) [4].
To date, the majority of studies investigating PH in iso-
lated ESRD populations have been limited in scope, and
have primarily employed transthoracic echocardiography
(TTE) to assess for PH. Measurement of the tricuspid
regurgitant jet velocity (TRV) can allow for estimation
of the systolic pulmonary artery pressure (sPAP), and an
elevated TRV is often used as a surrogate for the presence
of pulmonary hypertension. Therefore, we conducted a
quantitative systemic review of the evidence to address
two questions: “How prevalent is echocardiographic PH
among patients with ESRD who are undergoing hemodi-
alysis (HD) or peritoneal dialysis (PD)?” and “Are ESRD
patients with echocardiographic PH at higher risk for mor-
tality than ESRD patients who do not have PH?” Our goal
was to estimate the prevalence and mortality risk from
the entire body of evidence, as estimates from individual
studies have been highly variable [3].
Materials andMethods
Literature Search
A search strategy was constructed using the medical sub-
ject headings (MeSH) of “hypertension, pulmonary,” and
“kidney failure, chronic,” as well as keywords (“pulmo-
nary hypertension,” “pulmonary vascular disease,” “end-
stage renal disease,” “chronic kidney disease,” “dialysis”)
and synonyms (Tables1–3 in online supplement). Three
databases (Medline via PubMed, Cochrane Library, and
Cumulative Index to Nursing and Allied Health Litera-
tures [CINAHL]) were searched independently by two
investigators in September 2016, and October 2016, and
updated in November 2017 and November 2018. Searches
were restricted to English language publications involving
human subjects, but not restricted by date or publication
type.
Study Selection
Pre-specified study selection criteria included (1) Enroll-
ment of ESRD patients actively treated with HD or PD;
(2) Assessment for PH by estimating the sPAP via tran-
sthoracic echocardiography, and (3) Determination of the
prevalence of PH or associated mortality. Studies were
included irrespective of HD access type (i.e., fistula, graft,
or indwelling catheter). Those utilizing a pre-transplant
cohort were excluded from the primary analysis due to
concerns of selection bias.
Two investigators (NS and RA) independently screened
the search results after the searches were merged and dupli-
cates removed. Most publications were excluded by title
and abstract. Those that could not be excluded by title and
abstract underwent full text review and then were included
or excluded. A third investigator (KW) was assigned to
adjudicate differences of opinion (not required).
Evidence Synthesis
Data were extracted from the selected studies into a Micro-
soft Excel spreadsheet. Data extracted included the follow-
ing: location; inclusion and exclusion criteria; mean age
and sex of all patients, and patients with and without PH;
diagnostic test parameters used to identify PH; number of
ESRD patients with and without PH (total, on HD, and
on PD); follow-up duration; deaths among ESRD patients
with and without PH; and number of patients with and
without diastolic dysfunction (where available).
Prevalence was pooled by generic inverse variance
using a random effects model and reported as a propor-
tion. Pre-specified subgroup analyses included prevalence
among patients receiving HD and PD. Mortality was
pooled by the Mantel–Haenszel method using a random
effects model and reported as a relative risk (RR). All anal-
yses were performed using the Cochrane Review Manager
(i.e., RevMan), version 5.3 [5]. Regardless of the approach
used to pool individual studies, the accompanying 95%
confidence interval (CI) was determined.
Statistical heterogeneity of the pooled results was meas-
ured using the I2 test, considering an I2 value of ≥ 50%
to indicate significant heterogeneity. When significant
heterogeneity was encountered, sensitivity analyses were
performed to assess potential sources of heterogeneity.
If the source of heterogeneity could not be determined,
the median and range were used to inform conclusions
although the pooled estimates were also reported.
The Grading, Recommendations, Assessment, Devel-
opment, and Evaluation (GRADE) approach was used to
assess certainty in the estimated effects (i.e., the quality of
evidence) for each outcome of interest [6]. The certainty
in the estimated effects was categorized into one of four

537Lung (2020) 198:535–545
1 3
levels: high, moderate, low, or very low. This categori-
zation was based upon multiple criteria including study
design, risk of bias, directness, consistency, precision,
magnitude of effect, dose–response gradient, and residual
confounding.
Manuscript Preparation
The manuscript was written to adhere with the Preferred
Reporting Items for Systematic Reviews and Meta-Analyses
(PRISMA) statement [7]. The initial draft of the manuscript
was written by the first author and then circulated to the
other authors; multiple cycles of review and revision ensued.
All authors approved the manuscript for submission.
Results
Literature Search
The search identified 1046 studies (413 studies in Med-
line, 599 studies in the Cochrane Library, and 34 studies in
CINAHL). Most studies were excluded based upon title and
abstract alone. The full texts of 67 studies were reviewed,
from which 41 studies were selected (Fig.1) [2, 8–47]. Eight
studies reported both prevalence and mortality data, 32 stud-
ies reported prevalence data only, and one study reported
mortality data only [26].
Most studies originated from either the Middle East and
North Africa (25 studies with 2183 patients) or East Asia (10
studies with 2215 patients). The remaining studies (6 stud-
ies with 839 patients) originated from the North America,
Brazil, and Western Europe. The median sample size was 90
patients, with a range from 32 to 618 patients.
All 41 studies enrolled ESRD patients requiring dialysis.
36 studies included patients undergoing HD and 11 studies
included patients undergoing PD. A total of 5237 patients
were included in the final analyses, among whom 3850 were
undergoing HD and 1387 were undergoing PD. The sPAP
used to diagnose PH varied among studies; the most com-
mon threshold was an sPAP > 35mmHg (used in 30 studies)
but ranged from > 30mmHg to > 45mmHg (Table1).
Prevalence
Forty studies based on retrospective chart reviews or pro-
spective case series reported the proportion of ESRD
patients with TTE evidence of PH. None of the studies
Fig. 1 Flow of information
diagram

538 Lung (2020) 198:535–545
1 3
Table 1 Selected studies
Author Year Design Number Dialysis type (num-
ber)
Trans-
plant
cohort?
Definition of PH Outcome
Abdelwhab etal. 2008 Case series 45 HD (45) No sPAP > 35mmHg Prevalence
Abedini etal. 2013 Case series 120 HD (60) PD (60) No mPAP > 25mmHg,
(~ sPAP > 37mmHg)
Prevalence
Acarturk etal. 2008 Case series 32 HD (32) No mPAP > 25mmHg,
(~ sPAP > 37mmHg)
Prevalence
Agarwal etal. 2012 Hybrid: observa-
tional study and
case series
288 HD (288) No sPAP > 35mmHg Prevalence, Mortality
Alhamad etal. 2014 Case series 72 HD (55) PD (17) No sPAP > 40mmHg Prevalence
Amin etal. 2003 Case series 51 HD (51) No sPAP > 35mmHg Prevalence
Bozbas etal. 2009 Case series 500 HD (432) PD (68) Yes sPAP > 30mmHg Prevalence
Casas-Aparicio etal. 2010 Case series 34 HD (27) PD (7) Yes sPAP > 40mmHg Prevalence
Dagli etal. 2009 Case series 116 HD (116) No sPAP > 30mmHg Prevalence
Etemadi etal. 2012 Case series 66 HD (34), PD (32) No sPAP > 35mmHg Prevalence
Fabbian etal. 2010 Case series 56 HD (29) PD (27) No sPAP > 35mmHg Prevalence
Fadaii etal. 2013 Case series 102 HD (102) No sPAP > 35mmHg Prevalence
Faqih etal. 2016 Case series 111 HD (111) No sPAP > 35mmHg Prevalence
Genctoy etal. 2015 Case series 179 HD (179) No sPAP > 35mmHg Prevalence
Hassanin etal. 2016 Case series 100 HD (100) No Not specified Prevalence
Hayati etal. 2017 Case series 69 HD (69) No mPAP > 25mmHg,
(~ sPAP > 37mmHg)
Prevalence
He etal. 2015 Hybrid: observa-
tional study and
case series
136 HD (136) No sPAP > 35mmHg Prevalence, Mortality
Issa etal. 2008 Case series 215 Not specified Yes sPAP > 35mmHg Prevalence
Kim etal. 2015 Case series 172 HD (84) PD (88) No sPAP > 37mmHg Prevalence
Kiykim etal. 2010 Case series 74 HD (74) No sPAP > 30mmHg Prevalence
Li etal. 2014 Case series 485 HD (485) No sPAP > 35mmHg Prevalence
Li etal. 2013 Observational study 278 Not specified No sPAP > 35mmHg Mortality
Mahdavi-Mazdeh
etal.
2008 Case series 62 HD (62) No sPAP > 35mmHg Prevalence
Mukhtar etal. 2014 Case series 80 HD (80) No sPAP > 30mmHg Prevalence
Nakhoul etal. 2005 Hybrid: observa-
tional study and
case series
42 HD (42) No sPAP > 35mmHg Prevalence, Mortality
Omrani etal. 2016 Case series 150 HD (150) No Not specified Prevalence
Oygar etal. 2012 Case series 105 HD (77) PD (28) No sPAP > 35mmHg Prevalence
Ramasubbu etal. 2010 Hybrid: observa-
tional study and
case series
90 HD (90) No sPAP > 35mmHg Prevalence, Mortality
Reddy etal. 2013 Case series 124 Not specified Ye s sPAP > 35mmHg Prevalence
Reque etal. 2016 Hybrid: observa-
tional study and
case series
211 HD (211) No sPAP > 35mmHg Prevalence, Mortality
Shen etal. 2015 Case series 60 HD (60) No sPAP > 35mmHg Prevalence
Stallworthy etal. 2013 Case series 739 Not specified Ye s sPAP > 30mmHg Prevalence
Tarrass etal. 2006 Case series 86 HD (86) No sPAP > 35mmHg Prevalence
Unal etal. 2013 Case series 70 HD (50) PD (20) No sPAP > 35mmHg Prevalence
Unal etal. 2009 Case series 135 PD (135) No sPAP > 35mmHg Prevalence
Xu etal. 2015 Case series 618 PD (618) No sPAP > 35mmHg Prevalence

539Lung (2020) 198:535–545
1 3
included a control group intended to determine the preva-
lence of PH among patients without ESRD.
The median prevalence of PH was 38% (range 8% to
70%) among patients undergoing any type of dialysis,
40% (range 16–70%) among patients undergoing HD, and
19% (range 8–37%) among patients undergoing PD. Using
meta-analysis, the pooled prevalence estimates of PH
were similar to the median prevalence estimates but had
high statistical heterogeneity. Specifically, the prevalence
among patients undergoing any type of dialysis was 37%
(95% CI 3–42%, I2 93%), among patients receiving HD
was 42% (95% CI 37–47%, I2 90%), and among patients
receiving PD was 21% (95% CI 15–27%, I2 85%) (Fig.2).
The difference in pooled prevalence among those receiv-
ing HD versus PD was significant (Chi-squared 27.53, df
1, p < 0.00001). Sensitivity analyses were performed to
identify the cause of the heterogeneity; potential causes
that were explored included geography, patient age, study
design, dialysis modality, timing of dialysis relative to
echocardiography, and the sPAP threshold employed. The
sensitivity analyses failed to identify the cause of hetero-
geneity. However, visual inspection of the Forest Plots
suggested that the heterogeneity may be attributable to
there being three categories of studies, those reporting
low, moderate, and high prevalence of PH. When the three
categories of studies were pooled separately, the hetero-
geneity nearly disappeared. No study characteristics were
identified that explained the three categories of results.
Notably, the moderate prevalence group estimates reca-
pitulated those of the combined group.
Studies from the Middle East and North Africa had a
pooled prevalence among patients undergoing any type of
dialysis of 38% (95% CI 30–45%), among patients receiving
HD of 42% (95% CI 35–50%), and among patients receiving
PD of 15% (95% CI 9–21%). Studies from East Asia had a
pooled prevalence among patients undergoing any type of
dialysis of 35% (95% CI 27–44%), among patients receiving
HD of 44% (95% CI 38–51%), and among patients receiving
PD of 24% (95% CI 14–34%). The remaining studies had a
pooled prevalence among patients undergoing any type of
dialysis of 36% (95% CI 27–45%), among patients receiving
HD of 38% (95% CI 29–48%), and among patients receiving
PD of 19% (1 study).
Studies defining PH as an estimated PASP > 35mmHg
(i.e., studies that defined PH as an estimated PASP > 25
or > 30 mmHg were excluded) had a pooled prevalence
among patients undergoing any type of dialysis of 35% (95%
CI 31–39%), among patients receiving HD of 40% (95% CI
35–44%), and among patients receiving PD of 21% (95%
CI 15–27%).
Five additional studies were identified that included renal
pre-transplant cohorts [48–52]. These studies were excluded
from the primary analyses due to concerns of selection bias
arising from the transplant evaluation. As a secondary anal-
ysis, we pooled these studies by meta-analysis and found
the prevalence of PH by TTE criteria to be 25% (95% CI
17–34%) in the pre-transplant population.
Mortality
Nine of the 41 studies contained adequate data to compare
mortality among those with and without TTE evidence
of PH in ESRD. Follow-up in these studies ranged from
12months to > 5years. In ESRD patients with TTE evidence
Table 1 (continued)
Author Year Design Number Dialysis type (num-
ber)
Trans-
plant
cohort?
Definition of PH Outcome
Yigla etal. 2009 Hybrid: observa-
tional study and
case series
127 HD (127) No sPAP > 45mmHg Prevalence, Mortality
Yigla etal. 2004 Case series 49 HD (49) No sPAP > 35mmHg Prevalence
Yigla etal. 2003 Hybrid: observa-
tional study and
case series
63 HD (58) PD (5) No sPAP > 35mmHg Prevalence, Mortality
Yilmaz etal. 2016 Case series 77 HD (77) No sPAP > 35mmHg Prevalence
Yoo etal. 2012 Hybrid: observa-
tional study and
case series
75 HD (75) No sPAP > 35mmHg Prevalence, Mortality
Yoo etal. 2017 Case series 119 HD (119) No sPAP > 35mmHg Prevalence
Yu etal. 2009 Case series 39 HD (39) No sPAP > 35mmHg Prevalence
Zeng etal. 2016 Case series 180 PD (180) No sPAP > 35mmHg Prevalence
Zhang etal. 2016 Case series 177 PD (177) No sPAP > 35mmHg Prevalence
Zhao etal. 2014 Case series 70 HD (70) No sPAP > 35mmHg Prevalence

540 Lung (2020) 198:535–545
1 3
Fig. 2 Prevalence of PH in ESRD

541Lung (2020) 198:535–545
1 3
of PH, there were 206 deaths from a total of 487 patients
(42.3% mortality), while in ESRD patients without TTE
evidence of PH, there were 172 deaths from a total of 820
patients (20.9% mortality). Thus, TTE evidence of PH was
associated with an increased risk of death from all causes
(RR 2.02; 95% CI 1.70–2.40) (Fig.3).
Quality ofEvidence
The evidence provided very low certainty in the prevalence
and mortality estimates. The evidence pertaining to preva-
lence was uncontrolled with inconsistency and contained
risk of selection bias (some patients were chosen for evalu-
ation by clinician discretion, rather than consecutively or
randomly). The evidence pertaining to mortality was also
observational with a risk of selection bias (Table2).
Discussion
Summary
A systematic review was performed to clarify the preva-
lence and clinical importance of TTE signs of PH in ESRD
patients undergoing dialysis. We identified 40 relevant stud-
ies that collectively yielded a median prevalence of 38%
(range 8% to 70%) among patients undergoing dialysis,
which is markedly higher than the prevalence of echocar-
diographic PH in the general population, estimated at ~ 2.5%
[53]. Patients undergoing HD were twice as likely to have
TTE evidence of PH (median 40%; range 16–70%) as
patients undergoing PD (median 19%; range 8–37%). Nine
studies demonstrated that TTE evidence of PH was associ-
ated with twice the risk of death (RR 2.02; 95% CI 1.70,
2.40).
Our findings that the prevalence of PH is increased among
patients with ESRD compared to the general population and
that mortality is increased among ESRD patients with PH
compared to ESRD patients without PH is consistent with
a recently published systematic review by Tang etal. [54]
Our study differs, however, in that we included more studies,
focused on patients with more severe kidney disease (i.e.,
ESRD rather than CKD, which may explain why our esti-
mates are slightly higher), and also differentiated between
patients undergoing HD and PD [54]. Additionally, we per-
formed a secondary analysis of patients in pre-transplant
cohorts.
Implications
The lower prevalence of PH among those patients under-
going PD compared with HD raises important questions,
including whether patients undergoing PD tend to be
younger or “healthier” with fewer comorbidities, or whether
there is an intrinsic risk associated with HD, such as the
presence of an AV fistula/graft, the chronic intermittent vol-
ume overload, or the process of hemodialysis itself. Notably,
no clear survival advantage has been shown in the general
ESRD population between PD and HD. This would suggest
either that an alternative mechanism exists during PD to
balance the higher rate of mortality of PH in the HD popu-
lation, or that our results are artifactual in the context of
limited data. Further research is warranted to explore this
intriguing question.
The combination of a high prevalence of PH by TTE
criteria and increased mortality associated with echocar-
diographic findings of PH suggests that screening ESRD
patients for PH by TTE may identify a large subpopula-
tion of patients who are at risk for increased mortality.
An echocardiographic diagnosis of “PH” may be attrib-
utable to a number of contributing factors, such as com-
mon comorbidities (hypertension, diabetes, and diastolic
dysfunction), chronic under-dialysis with occult volume
overload, high flow through the arterio-venous fistula,
chronic anemia, and systemic inflammation due to renal
disease [3]. While the vast majority of ESRD patients will
Fig. 3 Mortality associated with PH in ESRD

542 Lung (2020) 198:535–545
1 3
not have Group 1 PAH (and therefore will not benefit nor
should be considered for treatment with pulmonary vaso-
dilators), screening by TTE may prompt invasive hemody-
namic assessment, which in turn can identify other, poten-
tially treatable factors contributing to the excess mortality,
such as chronic volume overload (under-dialysis), or high-
output heart failure due to AVF size.
Several studies using RHC to better characterize hemo-
dynamics in patients with underlying renal disease support
this concept. In a 2017 study by Nishimura etal., 19 of 85
patients screened with echocardiographic PH were felt to
warrant RHC, of which 15 had pre-capillary disease and
3 had post-capillary PH (one had a mPAP < 25 mmHg)
[55]. Two additional studies retrospectively analyzed
separate RHC databases, examining patients with CKD
overall. Both studies identified a mixture of pre- and post-
capillary pulmonary hypertension, with post-capillary
disease (or combined pre- and post-capillary disease) as
the predominant phenotype in both cohorts; increased
mortality was associated with these hemodynamic find-
ings [56, 57]. Neither study analyzed ESRD patients sepa-
rately (both grouped them with CKD stage V patients);
however, more advanced CKD/ESRD was generally asso-
ciated with higher rates of PH and more predominantly
post-capillary PH. These observations suggest that the
finding of increased mortality associated with echocar-
diographic PH in an ESRD population is valid [56, 57].
Thus, screening with TTE to identify factors contributing
to the elevated PASP and then to optimize them (increased
volume removal, AV fistula modification, transition to PD,
or renal transplantation referral) could potentially improve
outcomes, although such data are conflicting, and further
study is warranted [3, 49–51].
It has been suggested that patients who have evidence of
PH may have their symptoms improved by renal transplanta-
tion (although it is not clear if this translates to a mortality
benefit) [49–51]. Our analysis of five pre-transplant cohorts
identified a lower rate of PH among patients referred for
transplantation than among those undergoing hemodialysis,
possibly due to some form of selection bias. Given the mor-
tality in ESRD patients with PH, if renal transplantation is
truly an effective intervention, it should be considered more
frequently and earlier in those patients who would otherwise
be considered transplant candidates.
Limitations
Our study has several important limitations. First, our meta-
analyses of prevalence had a high degree of heterogene-
ity. Despite numerous sensitivity analyses and confirming
that our results congregated into three categories, we were
unable to identify the source of heterogeneity across stud-
ies. Therefore, we used the median prevalence estimates as
Table 2 Evidence profile
a Selection bias—testing at clinician’s discretion; did not select patients for testing consecutively or randomly
b Inconsistency—I2 value > 90%
Quality assessment Groups Effect Quality Importance
No of studies Design Risk of bias Inconsistency Indirectness Imprecision Other ESRD w/ PH ESRD w/o PH
Prevalence
40 Case series SeriousaSeriousbNone None None 1835/4959 – 37% (95% CI 33–42%) VERY LOW CRITICAL
Mortality
9 Observational studies SeriousaNone None None None 206/487 (42.3%) 172/820 (20.9%) RR 2.02 (95% CI
1.70–2.40)
VERY LOW CRITICAL

543Lung (2020) 198:535–545
1 3
the outcome to inform our conclusions. It is noteworthy,
however, that the median prevalence estimates are nearly
identical to the pooled prevalence estimates in all cases,
thus increasing our confidence in the prevalence estimates,
despite the heterogeneity.
Second, none of the studies determined the prevalence
of PH in patients without ESRD; thus, we were unable to
directly compare the prevalence of PH among patients with
and without ESRD. Third, the studies in the mortality analy-
sis had a highly variable length of follow-up, which may
have contributed to the mild heterogeneity, without necessar-
ily affecting the differential mortality. Fourth, the cause-of-
death was not universally available for those studies report-
ing mortality; thus, it is unknown if the increased mortality
in this population is directly attributable to their PH. Finally,
the studies in our analysis identified PH by TTE rather than
by the gold-standard RHC. Although TTE is commonly used
as a surrogate for PH, it is far from adequate; the absence
of a measurable TRV does not preclude hemodynamically
significant PH [58]. Moreover, there are insufficient studies
using comparable RHC data in an ESRD patient popula-
tion to allow for adequate comparison between echocardio-
graphic PH and RHC.
In conclusion, our study found that TTE evidence of PH
is common across a wide variety of geographic and institu-
tional settings, occurring, on average, in more than 1 out of
every 3 ESRD patients on dialysis. Although the potential
causes for this finding are myriad (and deserving of further
research), there exists significant evidence that despite the
inherent flaws in the echocardiographic assessment of PH, it
nonetheless can identify a population of patients at increased
risk for poor outcomes. Although it is not currently known
which ESRD patients should be considered for screening
for PH, it is clear that there are phenotypes of PH in this
population that can be impacted by evaluation and directed
treatment. Given the large number of extant ESRD patients,
this is an area that necessitates further study.
Author Contributions NS: Conceptualization, Data Collection, Data
Analysis, Manuscript Writing, Guarantor. RA: Data Collection, Manu-
script Editing. EK: Manuscript Editing. KW: Conceptualization, Data
Analysis, Manuscript Editing. HF: Conceptualization, Manuscript
Editing.
Funding None.
Compliance with Ethical Standards
Conflict of interest The authors declare that there is no conflict of in-
terest.
Ethical Approval IRB approval was not required for this systematic
review as human subjects were not directly involved.
References
1. Simonneau G, Montani D, Celermajer DS etal. (2019) Haemo-
dynamic definitions and updated clinical classification of pul-
monary hypertension. Eur Resp J. 53:1801913. https ://doi.
org/10.1183/13993 003.01913 -2018]10.1016/j.jacc.2013.10.029
2. Yigla M, Nakhoul F, Sabag A etal. (2003) Pulmonary hyper-
tension in patients with end-stage renal disease. Chest
123(5):1577–1582
3. Sise ME, Courtwright AM, Channick RN (2013) Pulmonary
hypertension in patients with chronic and end-stage kidney
disease. Kidney Int 84(4):682–692. https ://doi.org/10.1038/
ki.2013.186
4. United States Renal Data System. 2018 USRDS annual data
report: executive summary. https ://www.usrds .org/2018/view/
v1_00.aspx.
5. The Cochrane Collaboration (2014) Review manager (RevMan)
[computer program]
6. Balshem H, Helfand M, Schunemann HJ etal. (2011) GRADE
guidelines: 3, rating the quality of evidence. J Clin Epidemiol
64(4):401–406. https ://doi.org/10.1016/j.jclin epi.2010.07.015
7. Stewart LA, Clarke M, Rovers M etal. (2015) Preferred reporting
items for systematic review and meta-analyses of individual par-
ticipant data: the PRISMA-IPD statement. JAMA 313(16):1657–
1665. https ://doi.org/10.1001/jama.2015.3656
8. Abdelwhab S, Elshinnawy S (2008) Pulmonary hypertension in
chronic renal failure patients. Am J Nephrol 28(6):990–997. https
://doi.org/10.1159/00014 6076
9. Abedini M, Sadeghi M, Naini AE, Atapour A, Golshahi J (2013)
Pulmonary hypertension among patients on dialysis and kid-
ney transplant recipients. Ren Fail 35(4):560–565. https ://doi.
org/10.3109/08860 22X.2013.76656 7
10. Acarturk G, Albayrak R, Melek M etal. (2008) The relation-
ship between arteriovenous fistula blood flow rate and pulmo-
nary artery pressure in hemodialysis patients. Int Urol Nephrol
40(2):509–513. https ://doi.org/10.1007/s1125 5-007-9269-8
11. Agarwal R (2012) Prevalence, determinants and prognosis of pul-
monary hypertension among hemodialysis patients. Nephrol Dial
Transplant 27(10):3908–3914. https ://doi.org/10.1093/ndt/gfr66 1
12. Alhamad EH, Al-Ghonaim M, Alfaleh HF, Cal JP, Said N (2014)
Pulmonary hypertension in end-stage renal disease and post renal
transplantation patients. J Thorac Dis 6(6):606–616. https ://doi.
org/10.3978/j.issn.2072-1439.2014.04.29
13. Amin M, Fawzy A, Hamid MA, Elhendy A (2003) Pulmonary
hypertension in patients with chronic renal failure: role of para-
thyroid hormone and pulmonary artery calcifications. Chest
124(6):2093–2097
14. Dagli CE, Sayarlioglu H, Dogan E etal. (2009) Prevalence of and
factors affecting pulmonary hypertension in hemodialysis patients.
Respiration 78(4):411–415. https ://doi.org/10.1159/00024 7334
15. Etemadi J, Zolfaghari H, Firoozi R etal. (2012) Unexplained
pulmonary hypertension in peritoneal dialysis and hemodialysis
patients. Rev Port Pneumol 18(1):10–14. https ://doi.org/10.1016/j.
rppne u.2011.07.002
16. Fabbian F, Cantelli S, Molino C, Pala M, Longhini C, Portaluppi
F (2010) Pulmonary hypertension in dialysis patients: A cross-
sectional italian study. Int J Nephrol 2011:283475. https ://doi.
org/10.4061/2011/28347 5
17. Fadaii A, Koohi-Kamali H, Bagheri B, Hamidimanii F,
Taherkhanchi B (2013) Prevalence of pulmonary hyperten-
sion in patients undergoing hemodialysis. Iran J Kidney Dis
7(1):60–63
18. Faqih SA, Noto-Kadou-Kaza B, Abouamrane LM etal. (2016)
Pulmonary hypertension: prevalence and risk factors. Int J Cardiol
Heart Vasc 11:87–89. https ://doi.org/10.1016/j.ijcha .2016.05.012

544 Lung (2020) 198:535–545
1 3
19. Genctoy G, Arikan S, Eldem O (2015) Pulmonary hypertension
associates with malnutrition and body composition hemodialysis
patients. Ren Fail 37(2):273–279. https ://doi.org/10.3109/08860
22X.2014.98670 5
20. Hassanin N, Alkemary A (2016) Evaluation of pulmonary artery
pressure and resistance by pulsed doppler echocardiography
in patients with end-stage renal disease on dialysis therapy. J
Saudi Heart Assoc 28(2):101–112. https ://doi.org/10.1016/j.
jsha.2015.09.002
21. Hayati F, Beladi Mousavi SS, Mousavi Movahed SM, Mofrad
BM (2016) Pulmonary hypertension among patients undergo-
ing hemodialysis. J Renal Inj Prev 6(2):122–126. https ://doi.
org/10.15171 /jrip.2017.24
22. He Y, Wang Y, Luo X, Ke J, Du Y, Li M (2015) Risk factors for
pulmonary hypertension in maintenance hemodialysis patients: a
cross-sectional study. Int Urol Nephrol 47(11):1889–1897. https
://doi.org/10.1007/s1125 5-015-1119-5
23. Kim SC, Chang HJ, Kim MG, Jo SK, Cho WY, Kim HK (2015)
Relationship between pulmonary hypertension, peripheral vas-
cular calcification, and major cardiovascular events in dialy-
sis patients. Kidney Res Clin Pract 34(1):28–34. https ://doi.
org/10.1016/j.krcp.2015.01.003
24. Kiykim AA, Horoz M, Ozcan T, Yildiz I, Sari S, Genctoy G
(2010) Pulmonary hypertension in hemodialysis patients with-
out arteriovenous fistula: The effect of dialyzer composition.
Ren Fail 32(10):1148–1152. https ://doi.org/10.3109/08860
22X.2010.51685 4
25. Li Z, Liang X, Liu S etal. (2014) Pulmonary hypertension: Epi-
demiology in different CKD stages and its association with car-
diovascular morbidity. PLoS ONE 9(12):e114392. https ://doi.
org/10.1371/journ al.pone.01143 92
26. Li Z, Liu S, Liang X etal. (2014) Pulmonary hypertension as an
independent predictor of cardiovascular mortality and events in
hemodialysis patients. Int Urol Nephrol 46(1):141–149. https ://
doi.org/10.1007/s1125 5-013-0486-z
27. Mahdavi-Mazdeh M, Alijavad-Mousavi S, Yahyazadeh H, Azadi
M, Yoosefnejad H, Ataiipoor Y (2008) Pulmonary hypertension in
hemodialysis patients. Saudi J Kidney Dis Transpl 19(2):189–193
28. Mukhtar KN, Mohkumuddin S, Mahmood SN (2014) Frequency
of pulmonary hypertension in hemodialysis patients. Pak J Med
Sci 30(6):1319–1322. https ://doi.org/10.12669 /pjms.306.5525
29. Nakhoul F, Yigla M, Gilman R, Reisner SA, Abassi Z (2005)
The pathogenesis of pulmonary hypertension in haemodialy-
sis patients via arterio-venous access. Nephrol Dial Transplant
20(8):1686–1692
30. Omrani H, Golshani S, Sharifi V, Almasi A, Sadeghi M (2016)
The relationship between hemodialysis and the echocardiographic
findings in patients with chronic kidney disease. Med Arch
70(5):328–331. https ://doi.org/10.5455/medar h.2016.70.328-331
31. Oygar DD, Zekican G (2012) Pulmonary hypertension in dialysis
patients. Ren Fail 34(7):840–844. https ://doi.org/10.3109/08860
22X.2012.69071 5
32. Ramasubbu K, Deswal A, Herdejurgen C, Aguilar D, Frost AE
(2010) A prospective echocardiographic evaluation of pulmonary
hypertension in chronic hemodialysis patients in the united states:
Prevalence and clinical significance. Int J Gen Med 3:279–286.
https ://doi.org/10.2147/IJGM.S1294 6
33. Reque J, Quiroga B, Ruiz C etal. (2016) Pulmonary hypertension
is an independent predictor of cardiovascular events and mortality
in haemodialysis patients. Nephrology (Carlton) 21(4):321–326.
https ://doi.org/10.1111/nep.12595
34. Shen S, Sun Q (2015) Analysis of clinically relevant factors for
pulmonary hypertension in maintenance hemodialysis patients.
Med Sci Monit 21:4050–4056
35. Tarrass F, Benjelloun M, Medkouri G, Hachim K, Benghanem
MG, Ramdani B (2006) Doppler echocardiograph evaluation of
pulmonary hypertension in patients undergoing hemodialysis.
Hemodial Int 10(4):356–359
36. Unal A, Duran M, Tasdemir K etal. (2013) Does arterio-venous
fistula creation affects development of pulmonary hypertension
in hemodialysis patients? Ren Fail 35(3):344–351. https ://doi.
org/10.3109/08860 22X.2012.76040 7
37. Unal A, Sipahioglu M, Oguz F etal. (2009) Pulmonary hyperten-
sion in peritoneal dialysis patients: Prevalence and risk factors.
Perit Dial Int 29(2):191–198
38. Xu Q, Xiong L, Fan L etal. (2015) Association of pulmonary
hypertension with mortality in incident peritoneal dialysis
patients. Perit Dial Int 35(5):537–544. https ://doi.org/10.3747/
pdi.2013.00332
39. Yigla M, Fruchter O, Aharonson D etal. (2009) Pulmonary hyper-
tension is an independent predictor of mortality in hemodialy-
sis patients. Kidney Int 75(9):969–975. https ://doi.org/10.1038/
ki.2009.10
40. Yigla M, Keidar Z, Safadi I, Tov N, Reisner SA, Nakhoul F (2004)
Pulmonary calcification in hemodialysis patients: Correlation with
pulmonary artery pressure values. Kidney Int 66(2):806–810.
https ://doi.org/10.1111/j.1523-1755.2004.00807 .x
41. Yilmaz S, Yildirim Y, Taylan M etal. (2016) The relationship
of fluid overload as assessed by bioelectrical impedance analysis
with pulmonary arterial hypertension in hemodialysis patients.
Med Sci Monit 22:488–494
42. Yoo HH, Martin LC, Kochi AC etal. (2012) Could albumin level
explain the higher mortality in hemodialysis patients with pul-
monary hypertension? BMC Nephrol 13:80–2369. https ://doi.
org/10.1186/1471-2369-13-80
43. Yoo HHB, Dos Reis R, Telini WM etal. (2017) Association of
pulmonary hypertension with inflammation and fluid overload in
hemodialysis patients. Iran J Kidney Dis 11(4):303–308
44. Yu TM, Chen YH, Hsu JY etal. (2009) Systemic inflammation is
associated with pulmonary hypertension in patients undergoing
haemodialysis. Nephrol Dial Transplant 24(6):1946–1951. https
://doi.org/10.1093/ndt/gfn75 1
45. Zeng Y, Yang DD, Feng S etal. (2016) Risk factors for pulmonary
hypertension in patients receiving maintenance peritoneal dialysis.
Braz J Med Biol Res. https ://doi.org/10.1590/1414-431X2 01547
33
46. Zhang L, Zhao S, Ma J etal. (2016) Prevalence and risk fac-
tors for pulmonary arterial hypertension in end-stage renal
disease patients undergoing continuous ambulatory peritoneal
dialysis. Ren Fail 38(5):815–821. https ://doi.org/10.3109/08860
22X.2015.11036 37
47. Zhao LJ, Huang SM, Liang T, Tang H (2014) Pulmonary hyper-
tension and right ventricular dysfunction in hemodialysis patients.
Eur Rev Med Pharmacol Sci 18(21):3267–3273
48. Bozbas SS, Akcay S, Altin C etal. (2009) Pulmonary hyperten-
sion in patients with end-stage renal disease undergoing renal
transplantation. Transplant Proc 41(7):2753–2756. https ://doi.
org/10.1016/j.trans proce ed.2009.07.049
49. Casas-Aparicio G, Castillo-Martinez L, Orea-Tejeda A, Abasta-
Jimenez M, Keirns-Davies C, Rebollar-Gonzalez V (2010) The
effect of successful kidney transplantation on ventricular dysfunc-
tion and pulmonary hypertension. Transplant Proc 42(9):3524–
3528. https ://doi.org/10.1016/j.trans proce ed.2010.06.026
50. Issa N, Krowka MJ, Griffin MD, Hickson LJ, Stegall MD, Cosio
FG (2008) Pulmonary hypertension is associated with reduced
patient survival after kidney transplantation. Transplantation
86(10):1384–1388. https ://doi.org/10.1097/TP.0b013 e3181 88d64
0
51. Reddy YN, Lunawat D, Abraham G etal. (2013) Progressive
pulmonary hypertension: Another criterion for expeditious renal
transplantation. Saudi J Kidney Dis Transpl 24(5):925–929

545Lung (2020) 198:535–545
1 3
52. Stallworthy EJ, Pilmore HL, Webster MW etal. (2013) Do echo-
cardiographic parameters predict mortality in patients with end-
stage renal disease? Transplantation 95(10):1225–1232. https ://
doi.org/10.1097/TP.0b013 e3182 8dbbb e
53. Moreira EM, Gall H, Leening MJ etal. (2015) Prevalence of pul-
monary hypertension in the general population: The rotterdam
study. PLoS ONE 10(6):e0130072. https ://doi.org/10.1371/journ
al.pone.01300 72
54. Tang M, Batty JA, Lin C, Fan X, Chan KE, Kalim S (2018) Pul-
monary hypertension, mortality, and cardiovascular disease in
CKD and ESRD patients: a systematic review and meta-analysis.
Am J Kidney Dis 72(1):75–83
55. Nishimura M, Tokoro T, Yamazaki S, Hashimoto T, Kobayashi
H, Ono T (2017) Idiopathic pre-capillary pulmonary hypertension
in patients with end-stage kidney disease: Effect of endothelin
receptor antagonists. Clin Exp Nephrol 21(6):1088–1096. https
://doi.org/10.1007/s1015 7-016-1344-y
56. O’Leary JM, Assad TR, Xu M, Birdwell KA, Farber-Eger E, Wells
QS, Hemnes AR, Brittain EL (2017) Pulmonary hypertension in
patients with chronic kidney disease: invasive hemodynamic etiol-
ogy and outcomes. Pulm Circ 7(3):674–683
57. Edmonston DL, Parikh KS, Rajagopal S, Shaw LK, Abraham D,
Grabner A, Sparks MA, Wolf M (2019) Pulmonary hypertension
subtypes and mortality in CKD. Am J Kidney Dis. https ://doi.
org/10.1053/j.ajkd.2019.08.027
58. O’Leary JM, Assad TR, Xu M, Farber-Eger E, Wells QS, Hemnes
AR, Brittain EL (2018) Lack of a tricuspid regurgitation doppler
signal and pulmonary hypertension by invasive measurement. J
Am Heart Assoc. 7(13):e009362
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.