Prevalence, Predictors, and Outcomes of Pulmonary Hypertension in CKD

Abstract

Pulmonary hypertension (PH) is associated with poor outcomes in the dialysis and general populations, but its effect in CKD is unclear. We evaluated the prevalence and predictors of PH measures and their associations with long–term clinical outcomes in patients with nondialysis-dependent CKD. Chronic Renal Insufficiency Cohort (CRIC) Study participants who had Doppler echocardiography performed were considered for inclusion. PH was defined as the presence of estimated pulmonary artery systolic pressure (PASP) >35 mmHg and/or tricuspid regurgitant velocity (TRV) >2.5 m/s. Associations between PH, PASP, and TRV and cardiovascular events, renal events, and all-cause mortality were examined using Cox proportional hazards models. Of 2959 eligible participants, 21% (n=625) had PH, with higher rates among those with lower levels of kidney function. In the multivariate model, older age, anemia, lower left ventricular ejection fraction, and presence of left ventricular hypertrophy were associated with greater odds of having PH. After adjusting for relevant confounding variables, PH was independently associated with higher risk for death (hazard ratio, 1.38; 95% confidence interval, 1.10 to 1.72) and cardiovascular events (hazard ratio, 1.23; 95% confidence interval, 1.00 to 1.52) but not renal events. Similarly, TRV and PASP were associated with death and cardiovascular events but not renal events. In this study of patients with CKD and preserved left ventricular systolic function, we report a high prevalence of PH. PH and higher TRV and PASP (echocardiographic measures of PH) are associated with adverse outcomes in CKD. Future studies may explain the mechanisms that underlie these findings.

CKD is present in over 25 million Americans, and it is associated with an increased risk of ESRD, cardiovascular disease, and mortality.^1,2 Factors contributing to such high risk of adverse outcomes in CKD remain poorly defined. Interventions to lower cardiovascular risk in CKD by targeting the traditional cardiovascular risk factors have been only modestly successful at best.^3–6 Therefore, there is an urgent need to evaluate novel risk factors for cardiovascular disease and mortality in CKD that may eventually serve as therapeutic targets.

There has been increasing interest recently in studying the association between pulmonary hypertension (PH) and CKD.^7–10 PH, defined by elevated pulmonary arterial pressure, is rare in the general population but seen in 30%–40% of patients with ESRD (on the basis of echocardiographic studies).^11,12 After it is established, PH is often progressive and associated with high morbidity and mortality.¹³ Risk factors for PH, such as left ventricular hypertrophy (LVH) and diastolic dysfunction, are common in patients with CKD and may predispose them to PH.¹⁴ However, there are very limited data available evaluating the prevalence of PH in nondialysis-dependent CKD, factors that contribute to PH in this setting, and more importantly, whether the presence of PH is associated with a greater risk of adverse outcomes in this population.^7,8,15–18

Therefore, we examined the prevalence, predictors, and associations between PH parameters and clinical outcomes in participants enrolled in the Chronic Renal Insufficiency Cohort (CRIC) Study. We hypothesized that there is a high prevalence of PH in those with nondialysis-dependent CKD and that the presence of PH (measured using estimated pulmonary artery systolic pressure [PASP] and tricuspid regurgitant velocity [TRV]) is associated with an increased risk for cardiovascular and renal events as well as death.

Results

Prevalence and Patient Characteristics

PH was present in 21.1% (625 of 2959) of the study population. Prevalence of PH was 5.88%, 10.9%, 20.97%, 21.85%, 26.5%, and 32.8% in those with stages 1, 2, 3a, 3b, 4, and 5, respectively. Participants with PH were older, were more likely to be of non–Hispanic black origin, and had lower eGFR. In addition, participants with PH were more likely to have underlying comorbid conditions and use cardiovascular medications, including statins, β-blockers, and diuretics (Table 1). Participants with PH had lower left ventricular ejection fraction (LVEF), higher left ventricular mass index, and higher prevalence of LVH than those without PH (Table 1). There were also significant differences between participants who did not have PH measures performed compared with those with PH measures who are included in this analysis (Supplemental Table 1).

Table 1.

Clinical characteristics of CRIC Study participants with and without PH (PASP>35 mmHg and/or TRV>2.5 m/s)

Variables	With PH (n=625; Mean [SD] or %)	Without PH (n=2334; Mean [SD] or %)	P Value
Demographics
Age (yr)	63.2 (9.1)	58.6 (10.8)	<0.001
Sex (men/women; %)	52.2/47.8	54.3/45.7	0.34
Race (%)
Non-Hispanic white	38.9	47.9	<0.001
Non-Hispanic black	52.3	43.0
Hispanic	5.1	4.7
Other	3.7	4.5
Comorbid conditions
Cardiovascular disease (%)	46.1	34.5	<0.001
Stroke (%)	13.4	10.3	0.02
Congestive heart failure (%)	16.0	9.6	<0.001
Peripheral vascular disease (%)	8.8	7.5	0.26
COPD (%)	7.9	4.3	<0.001
Systolic BP (mmHg)	130.9 (22.3)	125.2 (21.0)	<0.001
Diastolic BP (mmHg)	67.6 (12.5)	70.3 (12.7)	<0.001
Diabetes (%)	53.1	47.0	<0.01
Body mass index (kg/m2)	32.0 (7.3)	32.0 (7.8)	0.97
Smoking (%)	10.7	13.1	0.10
Laboratory parameters
eGFR (ml/min per 1.73 m2)	38.8 (15.2)	43.8 (17.4)	<0.001
BUN	33.8 (17.5)	30.0 (15.3)	<0.001
24-h Proteinuria (g)	0.2 (0.1–0.9)	0.2 (0.1–0.8)	0.10
Serum phosphorous (mg/dl)	3.7 (0.6)	3.7 (0.7)	0.06
Serum bicarbonate (mg/dl)	24.4 (3.4)	24.4 (3.2)	0.68
Serum uric acid (mg/dl)	7.6 (1.9)	7.3 (1.9)	<0.001
Serum LDL cholesterol (mg/dl)	93.6 (32.4)	100.8 (34.8)	<0.001
Serum HDL cholesterol (mg/dl)	48.7 (15.6)	49.0 (15.8)	0.75
Hs C–reactive protein	5.9 (10.9)	5.2 (8.9)	0.08
BNP	152.3 (337.0)	71.0 (193.6)	<0.001
Medication use (%)
Angiotensin–converting enzyme inhibitors/angiotensin receptor use	70.7	69.0	0.40
Statins use	67.5	57.6	<0.001
β-Blocker use	64.5	47.5	<0.001
Diuretic use	67.8	57.5	<0.001
Echo findings
Ejection fraction (%)	53.1 (10.6)	54.6 (8.0)	<0.001
PASP (mmHg)	34.9 (8.9)	22.6 (4.1)	<0.001
TRV (m/s)	2.81 (0.29)	2.25 (0.15)	<0.001
Right atrial pressure (mmHg)	2.8 (3.4)	2.0 (3.0)	<0.001
LVH (%)	64.8	47.7	<0.001
Left ventricular mass index (g/m2.7)	55.9 (15.1)	50.2 (13.4)	<0.001

COPD, chronic obstructive pulmonary disease; Hs, highly sensitive; BNP, brain natriuretic peptide.

Cross-Sectional Association of Clinical Characteristics with PH

In the cross–sectional multivariable adjusted analyses, older age, anemia (hemoglobin <10 g/dl), lower LVEF, and presence of LVH were independently associated with higher odds of PH (Table 2). Supplemental Table 2 shows the results of a multivariable adjusted analysis, in which diastolic relaxation was used as a marker of diastolic function (instead of LVH).

Table 2.

Cross-sectional association between clinical and demographic factors and prevalent PH in CRIC Study participants

Variable	PH Odds Ratio (95% Confidence Interval)a
Age (per 5-yr increase)	1.25 (1.18 to 1.32)
Men	0.99 (0.80 to 1.23)
Race
Non-Hispanic black versus non-Hispanic white	1.24 (0.99 to 1.55)
Hispanic versus non-Hispanic white	1.18 (0.73 to 1.90)
Other versus non-Hispanic white	0.98 (0.58 to 1.68)
eGFR (per SD decrease)	1.11 (0.97 to 1.26)
Proteinuria (per SD increase)	1.06 (0.96 to 1.18)
Diabetes	0.92 (0.74 to 1.14)
Hypertension	1.24 (0.83 to 1.85)
Hyperlipidemia	0.99 (0.71 to 1.39)
Chronic obstructive pulmonary disease	1.43 (0.93 to 2.19)
Smoking	0.75 (0.54 to 1.05)
Body mass index (kg/m2)
<18.5 versus 18.5–24.9	1.46 (0.45 to 4.68)
25–29.9 versus 18.5–24.9	1.08 (0.78 to 1.50)
>30 versus 18.5–24.9	0.85 (0.61 to 1.17)
LVEF (each SD decrease)	1.12 (1.01 to 1.23)
LVH	1.55 (1.22 to 1.95)
Hemoglobin (g/dl)
<10 versus 10–12	1.60 (1.05 to 2.45)
12 versus 10–12	0.76 (0.60 to 0.97)

COPD, chronic obstructive pulmonary disease.

^aModel adjusted for age, sex, race, eGFR, proteinuria, diabetes, hypertension, smoking, COPD, body mass index, LVEF, LVH, and hemoglobin levels.

Associations between PH Measures and Clinical Outcomes

PH and Outcomes

After a median follow-up of 4.06 years, 497 participants died, 575 cardiovascular events occurred, and 705 renal events occurred. Participants with PH had higher rates of renal events, cardiovascular events, and all-cause mortality (log-rank P<0.001) (Figure 1). After adjusting for relevant confounding variables, the presence of PH was independently associated with a 38% and a 23% increased risk for all-cause mortality and cardiovascular events, respectively, but was not associated with an increased risk for renal events (Table 3).

Figure 1.

Clinical outcomes in those with and without Pulmonary Hypertension. Kaplan-Meier curve for mortality, cardiovascular (CV) events, and renal events for those with and without PH (defined as PASP>35 mmHg and/or TRV>2.5 m/s).

Table 3.

Associations between PH (defined as PASP>35 mmHg and/or TRV>2.5 m/s), TRV, PASP, and outcomes

Outcome	Unadjusted		Model Aa		Model Bb		Model Cc		Model Dd
	HR (95% CI)	P Value	HR (95% CI)	P Value	HR (95% CI)	P Value	HR (95% CI)	P Value	HR (95% CI)	P Value
PH versus no PH
All-cause mortality	2.13 (1.77 to 2.57)	<0.001	1.68 (1.38 to 2.05)	<0.001	1.48 (1.21 to 1.81)	<0.001	1.40 (1.12 to 1.74)	0.003	1.38 (1.10 to 1.72)e	0.004e
Cardiovascular events	1.71 (1.42 to 2.05)	<0.001	1.45 (1.20 to 1.75)	<0.001	1.34 (1.11 to 1.62)	0.002	1.31 (1.07 to 1.61)	0.01	1.23 (1.00 to 1.52)e	0.04e
Renal events	1.38 (1.15 to 1.65)	<0.001	1.31 (1.09 to 1.58)	0.003	1.13 (0.93 to 1.36)	0.21	1.12 (0.91 to 1.37)	0.28	1.16 (0.95, 1.43)	0.14
CHF	2.08 (1.69 to 2.56)	<0.001	1.77 (1.42 to 2.19)	<0.001	1.59 (1.28 to 1.97)	<0.001	1.47 (1.16 to 1.86)	0.001	1.38 (1.09 to 1.77)e	<0.01e
MI	1.33 (0.96 to 1.82)	0.08	1.12 (0.80 to 1.55)	0.51	1.02 (0.73 to 1.42)	0.91	1.07 (0.75 to 1.52)	0.71	1.04 (0.73 to 1.49)	0.80
CVA	1.33 (0.83 to 2.13)	0.23	1.16 (0.71 to 1.89)	0.54	1.07 (0.65 to 1.76)	0.78	1.04 (0.61 to 1.75)	0.89	0.99 (0.58 to 1.70)	0.98
TRV (per 1-SD increase)e
All-cause mortality	1.58 (1.44 to 1.73)	<0.001	1.45 (1.31 to 1.60)	<0.001	1.28 (1.16 to 1.42)	<0.001	1.19 (1.06 to 1.33)	0.003	1.18 (1.05 to 1.32)e	<0.01e
Cardiovascular events	1.69 (1.53 to 1.86)	<0.001	1.61 (1.45 to 1.78)	<0.001	1.41 (1.27 to 1.56)	<0.001	1.31 (1.17 to 1.47)	<0.001	1.26 (1.12 to 1.42)e	<0.001e
Renal events	1.36 (1.24 to 1.50)	<0.001	1.32 (1.19 to 1.46)	<0.001	1.16 (1.03 to 1.29)	0.01	1.14 (1.00 to 1.29)	0.04	1.12 (0.99 to 1.27)	0.06
CHF	1.85 (1.67 to 2.06)	<0.001	1.78 (1.59 to 1.99)	<0.001	1.51 (1.34 to 1.69)	<0.001	1.36 (1.19 to 1.55)	<0.001	1.30 (1.14 to 1.49)e	<0.001e
MI	1.31 (1.11 to 1.54)	0.001	1.22 (1.02 to 1.46)	0.03	1.09 (0.91 to 1.30)	0.36	0.99 (0.82 to 1.20)	0.95	0.97 (0.81 to 1.18)	0.79
CVA	1.46 (1.15 to 1.86)	0.002	1.34 (1.04 to 1.72)	0.03	1.20 (0.94 to 1.55)	0.14	1.30 (0.97 to 1.74)	0.07	1.29 (0.95 to 1.74)	0.10
PASP (per 1-SD increase)e
All-cause mortality	1.50 (1.37 to 1.64)	<0.001	1.41 (1.28 to 1.55)	<0.001	1.25 (1.12 to 1.38)	<0.001	1.15 (1.02 to 1.29)	0.02	1.15 (1.02 to 1.29)e	0.02e
Cardiovascular events	1.69 (1.54 to 1.86)	<0.001	1.64 (1.48 to 1.81)	<0.001	1.42 (1.28 to 1.58)	<0.001	1.34 (1.19 to 1.51)	<0.001	1.28 (1.13 to 1.45)e	<0.001e
Renal events	1.27 (1.15 to 1.40)	<0.001	1.24 (1.11 to 1.38)	<0.001	1.08 (0.95 to 1.22)	0.24	1.07 (0.93 to 1.23)	0.34	1.10 (0.95 to 1.27)	0.18
CHF	1.85 (1.68 to 2.04)	<0.001	1.81 (1.62 to 2.01)	<0.001	1.55 (1.38 to 1.75)	<0.001	1.40 (1.22 to 1.61)	<0.001	1.35 (1.17 to 1.56)e	<0.001e
MI	1.32 (1.12 to 1.56)	0.001	1.24 (1.03 to 1.48)	0.01	1.1 (0.92 to 1.34)	0.26	1.03 (0.84 to 1.26)	0.78	1.00 (0.82 to 1.22)	0.98
CVA	1.46 (1.18 to 1.82)	<0.001	1.36 (1.08 to 1.71)	<0.01	1.20 (0.94 to 1.52)	0.13	1.30 (0.98 to 1.71)	0.06	1.25 (0.94 to 1.68)	0.12

Each SD increase in TRV is 0.35 m/s, and each SD increase in PAP is 8.8 mmHg. CHF, congestive heart failure; MI, myocardial infarction; CVA, cerebrovascular accidents; COPD, chronic obstructive pulmonary disease; RAS, renin-angiotension system; BNP, brain natriuretic peptide; HR, hazard ratio; 95% CI, 95% confidence interval.

^aModel A: adjusted for age, sex, race, and clinical center.

^bModel B: adjusted for model A plus smoking, diabetes, hypertension, previous cardiovascular disease, COPD, chronic obstructive pulmonary disease, RAS: Renin-angiotension system; BNP - brain natriuretic peptide, body mass index, albumin, hemoglobin, LDL cholesterol, use of RAS blockers, diuretics, statins, β-blockers, and BNP.

^cModel C: adjusted for model B plus LVEF and LVH.

^dModel D: adjusted for model C plus eGFR and urine albumin to creatinine ratio.

^eEach SD increase in TRV 0.35 m/sec and SD increase in PAP 8.8 mm Hg.

TRV and Outcomes

Participants with TRV>2.5 m/s had higher rates of renal events, cardiovascular events, and all-cause mortality than those with TRV<2.5 m/s (log-rank P<0.001) (Supplemental Figure 1). We did not find evidence against linearity of TRV as a predictor of the outcomes and therefore, treated it as a single continuous variable in the primary analyses. After adjusting for confounding variables, each SD increase in TRV was associated with an 18% increased risk for death and a 26% increased risk for composite cardiovascular events (Table 3). Among the individual cardiovascular events, each SD increase in TRV was associated with higher risk for congestive heart failure (CHF) but not myocardial infarction (MI) or cerebrovascular accidents (CVA) (Table 3). There was a graded association between TRV and composite cardiovascular events at all levels of eGFR; participants with TRV>3.4 m/s had almost a 3-fold higher rate of composite cardiovascular outcomes compared with those with TRV<2.5 m/s (Figure 2A, Supplemental Table 3). Although there was a trend toward significance, there was no significant association between TRV and renal events, except for in those with TRV>3.0 m/s (Table 3, Supplemental Figure 2).

Figure 2.

Cardiovascular (CV) event rates. (A) TRV and CV event rates among various stages of CKD. (B) PASP and CV event rates among various stages of CKD.

PASP and Outcomes

Participants with higher PASP had higher rates of composite cardiovascular outcomes (Figure 2B). The Kaplan-Meier analysis showed significant differences in all-cause mortality and cardiovascular and renal events for patients with CKD and PASP>35 versus <35 mmHg (Supplemental Figure 3). After adjusting for confounding variables, each SD increase in PASP was associated with a 15% and a 28% increased risk for death and cardiovascular events, respectively, but was not associated with renal events. Each SD increase in PASP was associated with increased risk of CHF but not MI or CVA. Mortality and cardiovascular risk were significantly higher for those with PASP>55 mmHg (Supplemental Table 4).

Sensitivity Analyses

Associations between PH, TRV, and PASP and congestive heart failure were consistent in sensitivity analyses, which excluded participants who had a history of cardiovascular disease at the year 1 CRIC Study visit (Supplemental Table 5). Although the effect estimate suggested higher risk for composite cardiovascular events with PH and each SD increase in TRV and PASP, it was not statistically significant (Supplemental Table 5). In another sensitivity analysis, presence of PH and each SD increase in TRV were associated with the ESRD/death composite (Supplemental Table 6). Results were also consistent by subgroups of age, ethnicity, and eGFR at baseline (interaction P values ≥0.05 for all subgroups), except that the associations between PH and mortality differed between men and women (hazard ratio for men, 1.05; 95% confidence interval, 0.76 to 1.45 versus hazard ratio for women, 1.92; 95% confidence interval, 1.32 to 2.79) (Supplemental Figures 4 and 5).

Discussion

In this large, prospective study of patients with CKD and preserved left ventricular systolic function, we report a high prevalence of PH and note that the prevalence of PH increases with lower levels of kidney function. We show, for the first time, that the presence of PH is an independent predictor of mortality and cardiovascular events in patients with CKD. We also show that TRV and PASP (echocardiographic measures of PH) are also associated with mortality and cardiovascular events (primarily heart failure).

Echocardiography is commonly used for the initial evaluation of PH in clinical practice and in epidemiologic studies. Using Doppler–based echocardiographic measures, we report that the prevalence of PH is 21% in a large cohort of patients with nondialysis-dependent CKD. Comparison of prevalence of PH across study populations is limited by the varying definitions of PH used in different studies. Our findings are consistent with previous study results in patients with sickle cell that reported a prevalence of 25%–30% PH using a similar definition.¹⁹ However, it is important to point out that the sensitivity and specificity of echocardiography to diagnose PH are modest (83% and 72%, respectively).^20,21 This suggests that confirmation by right heart catheterization is needed for additional clinical evaluation. However, as discussed below, the predictive value of the echocardiography-based measures shown in our study is an important advance and highlights the need for additional study in this area.

In the CRIC Study, the likelihood of prevalent PH increased with older age, presence of anemia, and lower levels of LVEF and LVH. In the general population, PASP assessed using Doppler echocardiography increased with age.²² LVH and systolic dysfunction are important factors contributing to the development of PH, particularly with the underlying volume overload observed in this population. It is important to note that most CRIC Study participants had preserved LVEF, and patients with preexisting New York Heart Association (NYHA) classes III and IV heart failure were excluded. However, LVEF was still associated with PH in our study.²³ However, diastolic dysfunction (measured as LVH) or possibly, higher rates of precapillary hypertension (pulmonary arterial hypertension) with underlying vascular calcification noted in CKD may be contributing factors. Furthermore, CKD is associated with endothelial dysfunction, arterial stiffness, and elevated parathyroid hormone, factors implicated in the development of PH.^24–27 Our ability to identify the etiology of PH was limited in this large epidemiologic study. Hence, future mechanistic studies, possibly including right heart catheterization, may offer additional insights.

PH is associated with a high morbidity and mortality burden in the general population, those with sickle cell disease, chronic lung diseases, or ambulatory heart failure, and those on dialysis.^{19,22,28–30} Consistent with these studies, we now show that the presence of PH and higher TRV and PASP values are independent risk factors for mortality and cardiovascular events in patients with CKD. Interestingly, the risk for cardiovascular events and mortality observed in our study (in those without heart failure) is similar to what was noted in other cohorts with heart failure, highlighting the importance of PH in the CKD population. Furthermore, the stage of CKD did not modify the associations between PH measures and cardiovascular outcomes, suggesting that the risk for adverse outcomes begins even with mild to moderate CKD. We noted that the associations between PH and mortality are different for men and women, and additional studies are needed to explain these sex differences. Because most patients with early stages of CKD are cared for by primary care physicians, these results emphasize the need for them to understand the implications of PH in kidney disease.

Another important finding in our study was the association between PH, TRV, PASP, and the risk for heart failure. With left ventricular dysfunction, an increased left ventricular filling pressure leads to elevated left atrial pressure, which in turn, raises pulmonary venous pressure. This hemodynamic alteration that decreases filling pressure leads to pulmonary vascular remodeling in the long term, with resultant worsening of underlying PH and pulmonary edema. Our findings of increased hazard for CHF (but not MI or CVA in the cardiovascular event composite) with higher TRV are consistent with this sequence leading to heart failure in patients with CKD and PH. Furthermore, PH and TRV were also associated with incident cardiovascular events (in those without prior cardiovascular disease), strengthening these findings. In contrast to other cohorts, a higher proportion of CRIC Study participants had greater left ventricular mass index and LVH but with preserved LVEF. Future studies should clarify if these CHF events were related to right–sided heart failure caused by PH or congestive heart failure caused by left ventricular systolic or diastolic dysfunction.

The principal strength of this study is the availability of comprehensive clinical data and echocardiographic assessment of cardiac function using a standard protocol in a large, well characterized, contemporary cohort of patients with CKD without preexisting NYHA class III or IV heart failure. In addition, long-term follow-up, availability of adjudicated cardiovascular and renal events, and adequate adjustment for confounding variables in the analysis enhance the quality of the findings. However, there are important limitations in our study. The lack of a standardized definition of PH on the basis of echocardiographic data is a limitation in this field. To address this, we also considered PASP and TRV as continuous measures in addition to the categorical definition of PH. Although some guidelines suggest the use of TRV>2.7 m/s as possible PH, we used a lower cutoff, which was used by other studies.^19,31 However, our categorical analysis confirms higher risk for cardiovascular events and death in those with TRV>2.9 m/s. It is well known that Doppler echocardiograms tend to overestimate PH measures, but right heart catheterization is not feasible in such large cohorts. Given the nature and details collected in the parent study, we did not have information about vasculitis, chronic lung disease, details of AKI and duration of CKD, obstructive sleep apnea and continuous positive airway pressure use, and evaluation for other etiologies of PH, precluding us from adjusting for and classifying our patients into the diagnostic categories recommended by the World Health Organization.³² Because of the nature of the cohort (followed in health care systems), additional studies should confirm these findings. Although pulse–wave velocity measures (to assess arterial stiffness) were performed in the CRIC Study, the differences in timing of these studies limited our availability to evaluate the association between arterial stiffness and PH.

In summary, the prevalence of PH (defined as PASP>35 mmHg and/or TRV>2.5 m/s) was high in CRIC Study participants on the basis of echocardiographic assessment. PH, TRV, and PASP are independent predictors of cardiovascular events and mortality in this CKD cohort without preexisting heart failure. Importantly, additional studies are warranted to examine the mechanisms, perhaps including right heart catheterization, which might explain these associations, and explore the potential therapeutic implications.

Concise Methods

Study Design and Population

The CRIC Study enrolled 3939 individuals ages 21–74 years old with eGFRs of 20–70 ml/min per 1.73 m² from June of 2003 to December of 2008 at seven clinical centers across the United States (Ann Arbor, MI; Baltimore, MD; Chicago, IL; Cleveland, OH; New Orleans, LA; Philadelphia, PA; and Oakland, CA). Study design and baseline participant characteristics were previously published.^23,33 Major exclusion criteria included prior dialysis lasting >1 month, NYHA class III/IV heart failure, polycystic kidney disease or other primary renal diseases requiring active immunosuppression, HIV, and pregnancy. Participants underwent annual study visits and telephone follow-up two times per year. Study participants provided written informed consent and authorized release of medical records for study purposes, and they were followed annually under protocols approved by institutional review boards at each of the CRIC Study clinical centers. Patients who underwent Doppler echocardiography study at the year 1 visit of the study with available data relating to PH (TRV and PASP) were included in this analysis (Supplemental Figure 6).

Doppler Echocardiogram

Doppler echocardiography was performed 1 year after enrollment according to the American Society of Echocardiography guidelines to assess heart structure and function.¹⁴ Data were sent to a core echocardiography laboratory for measurement and analysis (University of Pennsylvania). The peak difference between right ventricular and right atrial pressure was assessed from the continuous-wave signal of tricuspid regurgitation. Right atrial pressure was estimated by visual inspection of the inferior vena caval diameter through several respiratory cycles. This was added to the right atrial/right ventricular pressure gradient obtained from the tricuspid regurgitation velocity to yield the PASP. Left ventricular mass and LVH were assessed as described previously.¹⁴ Left ventricular diastolic function was categorized into normal or mildly, moderately, or severely abnormal (corresponding to normal and grades 1–3 diastolic dysfunction, respectively).³⁴ Multiple reproducibility, inter- and intrareader reliability, and reader drift analyses were performed throughout the course of this study on a 2% random sample of the entire cohort each year. The intraclass correlation coefficients (and κ-statistics) for the echocardiographic measures are LVH=0.759 (κ=0.61), TRV=0.97 (κ not available), and LVEF=0.854 (κ not applicable).

PH Definition

PH was defined as the presence of PASP>35 mmHg and/or TRV>2.5 m/s in the echocardiogram obtained during the year 1 CRIC Study visit.¹⁹ We also examined TRV and PASP, key PH measures, as a continuous measure (each SD higher). Tricuspid regurgitant Doppler was graded for quality as not done, not analyzable, good, fair, and poor. We excluded the not done and not analyzable groups for these analyses (Supplemental Figure 6).

Outcomes

The outcome measures for this study included (1) cardiovascular events (heart failure, myocardial infarction, stroke, and peripheral vascular disease), (2) renal outcomes (50% decrease in eGFR or ESRD), and (3) all-cause mortality. These events were adjudicated from study entry to March of 2012. Participants’ follow-up was censored at the end of the follow-up time period, loss to follow-up, or death, whichever occurred first. Two independent reviewers evaluated the clinical data on the basis of predefined criteria to adjudicate each outcome. Detailed outcome ascertainment details are reported in Supplemental Material.

Covariates

Demographics and clinical information were obtained at the baseline and follow-up study visits by questionnaires, interviews, and physical examination. Data obtained at the year 1 visit for the CRIC Study were used for this analysis to coincide with the echocardiogram. History of any cardiovascular disease included prior myocardial infarction, revascularization, heart failure, stroke, or peripheral arterial disease. Current smoking was defined as self-report of current use of cigarettes and at least 100 cigarettes smoked. At each study visit, participants were queried about any medication usage in the prior 30 days. All antihypertensive medications were categorized into drug classes, and the total number of antihypertensive drug classes was calculated. Serum creatinine and 24-hour urine total protein were measured at the CRIC Central Laboratory using standard methods. Diabetes mellitus was defined as a fasting glucose >126 mg/dl, a nonfasting glucose >200 mg/dl, or use of insulin or other antidiabetic medication. Except for high-sensitivity C–reactive protein, serum phosphorus, and uric acid (for which baseline visit data were used), all other laboratory data obtained at the year 1 visit were used in this analysis.

Statistical Analyses

We calculated the overall prevalence rates of PH on the basis of age, sex, race, and stage of kidney disease. Descriptive statistics for baseline variables were compared among participants with and without PH using chi-squared and ANOVA models for categorical and continuous variables, respectively. A two–sided type I error probability of 0.05 was used throughout the analyses. Multivariate regression analyses were conducted to examine the predictors of PH. This model was adjusted for potential confounding variables, such as age, sex, race, eGFR, proteinuria, diabetes, hypertension, smoking, COPD, body mass index, LVEF, LVH, and hemoglobin levels. We also conducted a multivariate regression analysis, in which diastolic relaxation was used instead of LVH as a marker of diastolic relaxation.

To evaluate whether PH was associated with the individual primary outcome measures among this CKD population, Kaplan-Meier plots and log-rank tests were used. The associations of PH with individual outcomes of interest were examined using the Cox proportional hazards model while adjusting for age; sex; race; clinical center; diabetes; hypertension; smoking; body mass index; previous cardiovascular disease; chronic obstructive pulmonary disease; angiotensin–converting enzyme inhibitors/angiotensin receptor blockers use; use of diuretics, statins, and β-blockers; serum albumin; hemoglobin; LDL cholesterol; eGFR; urine albumin to creatinine ratio; LVEF; and LVH. We performed sequential modeling by first adjusting for demographics and then including comorbid conditions followed by adjustment for cardiac and kidney function as described above. A separate analysis was conducted by considering TRV as a continuous measure (each SD increase). Prespecified interactions between PH and age, sex, race, hypertension, and eGFR were tested, because these variables might modify the associations of PH with outcomes in CKD.

Sensitivity Analyses

In addition to examining TRV in a continuous analysis, we also compared the different categories of TRV and its relationship with outcomes (TRV=2.6–2.9, 3.0–3.4, and >3.4 versus <2.5 m/s). Similarly, we examined the associations between PASP and outcomes as continuous (each SD increase) and categorical variables (PASP=36–45, 46–55, and >55 versus <35 mmHg). We also conducted separate analysis to examine the associations of PH, PASP, and TRV with incident cardiovascular events (among those without previous cardiovascular disease at the year 1 CRIC Study visit) and the ESRD/death composite in the CRIC Study.

Disclosures

G.S. received research grants (principal investigator for multicenter studies) from Actelion and Gilead; was part of the Lectures/Speaker Bureau for Gilead, United Therapeutics, and Bayer; and also, reports travel arrangements to attend investigator/research meetings by Actelion. All other authors report no competing financial interests relevant to this manuscript.