H2 Receptor Antagonist Use and Mortality in Pulmonary Hypertension: Insight from the VA-CART Program

To the Editor:

Pulmonary hypertension (PH) is common and predicts or mediates poor outcomes in many medical disorders (1). Independent of the underlying etiology, persons affected by PH have increased right-ventricular load and often develop right-heart failure. Pulmonary vasodilators are beneficial in some forms of PH (e.g., pulmonary arterial hypertension), but they can be harmful or have no impact in other forms (2).

Histamine H2 receptor antagonism may be relevant in the myocardial stress response and beneficial in right-heart dysfunction and right-heart failure (3–6). To explore this possibility, we examined relationships between H2 receptor antagonist (H2RA) use and mortality in a cohort of veterans with PH confirmed at right-heart catheterization (RHC).

Methods

We included veterans from the Veterans Affairs (VA) Clinical Assessment, Reporting, and Tracking (VA-CART) program who received RHC at a VA center between 2008 and 2014, had a mean pulmonary artery pressure (mPAP) of ≥25 mm Hg, and had a pulmonary artery wedge pressure (PAWP) recorded. The Colorado Multiple Institutional Review Board approved this study (#14-1649).

Exposure

Participants were considered to have used an H2RA if an outpatient prescription was filled within 90 days of the RHC. Participants were excluded if they died within 90 days or had a hospitalization lasting longer than 60 days after catheterization. This provided at least 30 days to detect outpatient medication use.

Outcome

The outcome was the rate of all-cause mortality determined using the combined VA vital status file (97.6% exact agreement with the National Death Index) (7). Risk time accrued after the 90-day window that was used to establish exposure status. Exposure and outcome assessment did not temporally overlap in an effort to avoid immortal time bias.

Statistical Analysis

We used Cox proportional hazards models to estimate associations between H2RA use and mortality. In limited models, we adjusted for age, sex, race, and body mass index. In fully adjusted models, we also accounted for participants’ markers of socioeconomic status and health behaviors. In separate models, we further adjusted for comorbid medical conditions or comedication use.

To account for confounding by indication, analyses were repeated in a restricted cohort comparing participants who used H2RAs with those who used proton pump inhibitors. In a second restricted cohort, propensity scores were used to match H2RA users with nonusers. Analyses were repeated in cohorts limited to participants with a diagnosis of heart failure or chronic obstructive pulmonary disease (COPD).

Careful phenotyping by World Health Organization group was not feasible; however, PAWP was available for all participants. PAWP was evaluated as an effect modifier of the relationship between H2RA use and mortality. A typical cutoff of 15 mm Hg and a more stringent cutoff of 12 mm Hg to exclude left-heart disease were used.

Analyses were performed using SAS 9.4 (SAS Institute) and R 3.3.1 (R Project for Statistical Computing).

Results

Participant characteristics are included in Table 1. A total of 589 H2RA users died in 4,719 person-years (12.5 deaths per 100 person-years) and 6,341 nonusers died in 46,129 person-years (13.7 deaths per 100 person-years). H2RA use within 90 days of RHC was associated with a 10% lower risk for all-cause mortality (adjusted hazard ratio, 0.90; 95% confidence interval, 0.83–0.98; P = 0.02; the proportional hazard assumption was not violated). This relationship was slightly stronger when we accounted for comedication use, accounted for comorbidity, compared H2RA users with users of proton pump inhibitors, and when we compared H2RA users with propensity-matched nonusers (Table 2). When associations were evaluated in cohorts limited to veterans with COPD or heart failure, estimates were similar to those obtained in the full cohort but less precise.

Table 1.

Characteristics of the Study Sample

		H2RA Users (n = 1,518)	Non-H2RA Users (n = 16,011)	PS Matched Nonusers (n = 1,518)
Age, yr		65.3 ± 12.1	66.4 ± 9.3	66.4 ± 9.8
Male, %		95.8	96.7	95.0
Race, %
White		80.6	76.8	80.9
African American		17.5	21.1	16.7
Other		1.8	2.1	2.4
Body mass index, %
Underweight, <18.5 kg/m2		0.8	0.7	0.8
Normal, 18.5–25 kg/m2		15.2	17.1	15.6
Overweight, 25–30 kg/m2		30.4	28.8	30.3
Obese or severely obese, ≥30 kg/m2		53.6	53.4	53.3
Income, $10,000		49.0 ± 16.7	49.8 ± 17.4	49.1 ± 16.2
Marital status, %
Single		13.7	13.6	14.5
Married		50.1	48.4	49.9
Divorced or widowed		36.2	38.0	35.5
Documented current/previous alcohol abuse, %		10.0	10.4	9.4
Documented current or previous smoking, %		61.9	60.7	60.5
Comorbidity, %
Diabetes mellitus		55.7	51.7	56.1
End-stage renal disease/dialysis		35.3	33.2	34.4
Cirrhosis		6.3	7.2	5.8
Obstructive sleep apnea		14.4	14.6	14.0
Chronic obstructive pulmonary disease		40.6	37.3	40.5
Asthma		6.9	5.6	6.8
Interstitial lung disease		0.6	0.5	0.7
Prior MI, PCI, or CABG		48.7	39.9	47.3
Congestive heart failure		67.1	68.1	66.9
Valvular heart disease		40.6	38.7	41.2
Congenital heart disease		0.5	0.6	0.7
Atrial fibrillation/flutter		30.1	33.1	31.1
Hemodynamics at right-heart catheterization
Systolic systemic blood pressure, mm Hg		130 ± 15	131 ± 16	132 ± 16
Diastolic systemic blood pressure, mm Hg		73 ± 10	74 ± 10	74 ± 10
Right atrial pressure, mm Hg		12 ± 6	12 ± 6	12 ± 6
Pulmonary artery systolic pressure, mm Hg		53 ± 14	53 ± 14	53 ± 14
Pulmonary artery diastolic pressure, mm Hg		24 ± 7	24 ± 7	24 ± 7
Mean pulmonary arterial pressure, mm Hg		35 ± 9	35 ± 9	35 ± 9
Pulmonary artery wedge pressure, mm Hg		22 ± 7	22 ± 8	22 ± 8
Cardiac output, L/min		5.2 ± 1.7	5.2 ± 1.9	5.1 ± 1.6
Pulmonary vascular resistance, Wood units*		2.8 ± 2.0	3.0 ± 2.1	3.0 ± 2.2

Definition of abbreviations: CABG = coronary artery bypass graft; H2RA = H2 receptor antagonist; MI = myocardial infarction; PCI = percutaneous coronary intervention; PS = propensity score.

Data are mean ± SD, unless otherwise specified.

^*Pulmonary vascular resistance [(mean pulmonary artery pressure − wedge pressure)/cardiac output] was determined for each participant and then averaged.

Table 2.

Relationship of H2RA Use within 90 Days of Right-Heart Catheterization to Mortality in Veterans with Pulmonary Hypertension (n = 17,529)

	Adjusted Risk of Mortality in H2RA Users Relative to Nonusers
	Hazard Ratio	95% CI	P Value
Full cohort
Unadjusted	0.90	0.83–0.98	0.02
Limited adjustment*	0.91	0.83–0.99	0.03
Full adjustment†	0.90	0.83–0.98	0.02
Full adjustment† + comedication use‡	0.88	0.80–0.96	0.003
Full adjustment† + comorbidity§	0.86	0.79–0.94	0.001
Restricted cohorts
Restricted to H2RA users and PPI users|| (n = 7,967)	0.86	0.79–0.94	0.001
Restricted to PS matched participants|| (n = 2,982)	0.82	0.74–0.92	0.001
Restricted to participants with a diagnosis of COPD|| (n = 6,594)	0.89	0.79–1.00	0.06
Restricted to participants with a diagnosis of CHF|| (n = 11,927)	0.91	0.83–1.00	0.06

Definition of abbreviations: CHF = congestive heart failure; CI = confidence interval; COPD = chronic obstructive pulmonary disease; H2RA = H2 receptor antagonist; PPI = proton pump inhibitor; PS = propensity score.

^*Limited adjustment accounts for age, sex, race/ethnicity, and body mass index.

^†Full adjustment accounts for the limited model and also includes income, tobacco use, alcohol abuse, and marital status.

^‡Comedication use included aldosterone antagonists, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, β-receptor antagonists, calcium channel blockers, digoxin, diuretics, lipid-lowering medication, medications for asthma or chronic obstructive pulmonary disease, medications for diabetes mellitus, oral anticoagulants, platelet inhibitors, systemic vasodilators, and proton pump inhibitors.

^§Comorbidity included the presence or absence of end-stage renal disease/dialysis, diabetes mellitus, cirrhosis, sleep-disordered breathing, chronic obstructive pulmonary disease or asthma, interstitial lung disease, prior myocardial infarction, prior percutaneous coronary intervention, prior coronary artery bypass graft, congestive heart failure, valvular heart disease, congenital heart disease, atrial fibrillation, and/or atrial flutter.

^||Participants in the restricted cohorts were considered in models with full adjustment.

There was no evidence that PAWP modified relationships between H2RAs and mortality using either conventional (15 mm Hg; P value for the interaction = 0.75) or conservative (12 mm Hg; P = 0.25) thresholds for left-heart disease.

Discussion

We observed that H2RA use was associated with a 10% lower risk for all-cause mortality in a large cohort with PH confirmed by RHC. Similar relationships were seen in unadjusted and adjusted analyses, and in restricted cohorts with specific participant characteristics.

Conventional treatment for PH focuses on the underlying cause of the PH (e.g., COPD or left-heart failure) and includes the use of pulmonary vasodilators in pulmonary arterial hypertension (2). VA-CART is not a pulmonary arterial hypertension cohort. Prior work suggested that H2RAs may exert a benefit by targeting the ventricle itself, rather than the pulmonary vasculature. We observed similar survival rates with H2RA use in veterans with COPD, those with heart failure, and when PAWP was above or below key thresholds. This finding is unusual for PH therapies, as treatment often improves mortality in selected groups with precapillary PH but may worsen mortality in those with postcapillary PH (2).

This report is consistent with converging lines of evidence concerning left-heart failure, which suggest that histaminic signaling may be important for cardiovascular health. A previous single-center, open-label, randomized trial demonstrated improvement in cardiac morphology, b-type natriuretic peptide, and functional class for participants with left-heart failure and reduced ejection fraction treated with the H2RA famotidine (8). In addition, H2RA use was associated with a reduced incidence of heart failure in a previous observational study (9). These studies are reinforced by results from animal models suggesting plausible mechanisms for benefit. For example, modulating histaminic signaling in animal models can abrogate heart failure from tachycardia, doxorubicin, and aortic banding (3–5). Furthermore, H2 receptor activation has been shown to contribute to mitochondrial permeability and myocardial fibrosis in experimental models (10).

This study has several limitations. Confounding can complicate inference in observational studies, and occult left-heart disease or an imbalance of H2RA use between incident and prevalent cases may have contributed to our findings. The consistency of results across a range of adjustments and in several restricted cohorts is reassuring but not definitive. Generalizability will require analyses in sex-balanced populations. In addition, all-cause mortality, while expedient, relatively insensitive to misclassification, and inherently patient-oriented, is not “cause specific.” Cause-specific mortality was not available, and the results could reflect outcomes from noncardiovascular causes. Finally, H2RAs are available over the counter and are poorly enumerated in an integrated medical record. It is likely that misclassification of H2RA use was present. True associations between H2RA use and mortality may be diluted by misclassification.

H2RAs are inexpensive and safe, and have a favorable side-effect profile. If our results are confirmed in prospective, randomized studies, H2RAs may represent an attractive treatment strategy for patients with PH.