Heart failure as a cause of hospitalization in IPF: comparison of the HFpEF, HFrEF, and PH

Abstract


Introduction
Idiopathic pulmonary brosis (IPF) is a speci c form of chronic, progressive and brosing lung disease of unknown etiology and poor prognosis [1].The natural course of IPF is highly variable [2].Many patients with IPF require hospitalization related to respiratory and/or cardiovascular disease [3,4].
Heart failure (HF) is an important health problem worldwide.The prevalence of HF is estimated to be 1%-2% of the adult population in developed countries [5].About 50% of patients with HF have a normal or nearly normal left ventricular ejection fraction (LVEF) [6][7][8].The prevalence of this type of HF, termed HF with preserved ejection fraction (HFpEF), is increased in older populations [9].Because many patients with IPF are elderly, HFpEF may affect a high proportion of IPF patients with HF.Additionally, some patients with IPF have pulmonary hypertension (PH) that occasionally leads to right heart failure [10].
Hospitalization, which is sometimes required for patients with IPF, has recently been recognized as an important clinical outcome [11,12].HF can account for some of the hospitalizations of patients with IPF.
The clinical features and outcomes of patients with IPF and HF, including the HF subgroups HFpEF, HF with reduced ejection fraction (HFrEF), and HF due to PH (HFPH) are unknown.This study aimed to clarify the details of the clinical features and outcomes of HF in patients with IPF who needed nonelective hospitalization.

Patients
We retrospectively reviewed the clinical records of all patients with IPF who required nonelective hospitalization for HF to the Kindai University hospital from January 2008 to December 2018.The diagnosis of IPF was based on the eligibility criteria adopted in the INPULSIS trial [13].Brie y, patients who had a histologically con rmed usual interstitial pneumonia (UIP) pattern on a surgical lung biopsy (SLB) specimen were included.In the absence of a SLB specimen used for con rmation of UIP, a highresolution computed tomography (HRCT) scan showing honeycombing and/or a combination of reticular abnormalities and traction bronchiectasis without atypical features of UIP was required for study inclusion.

Diagnosis of HF
The diagnosis of HF at the hospitalization of the study patients was mainly based on the Framingham study, as follows : 1) clinical evidence of HF such as dyspnea on exertion, paroxysmal nocturnal dyspnea or orthopnea, or peripheral edema; 2) chest radiography or HRCT revealing pulmonary venous congestion, cardiomegaly, or pleural effusion; 3) level of B-type natriuretic peptide (BNP) ≥ 100 pg/mL recorded at admission; 4) deterioration not explained by other causes such as acute exacerbation of IPF, pulmonary infection, pulmonary embolism, mediastinal emphysema and/or pneumothorax, or anemia [14,15].The types of HF were as follows: HFpEF was de ned as HF with LVEF ≥ 50% and HFrEF as HF with LVEF< 50%.Furthermore, patients with an echocardiographic right ventricular systolic pressure (RVSP) >50 mm Hg, were classi ed with HF due to PH (HFPH) [16].Finally, we divided the IPF patients with HF into 3 groups, as follows: HFpEF (LVEF ≥ 50% and RVSP < 50 mm Hg), HFrEF (LVEF < 50% and RVSP < 50 mm Hg), and HFPH (RVSP ≥ 50 mm Hg).

Pulmonary function tests
The pulmonary function tests that were performed within 1 year prior to the hospitalization were used for baseline pulmonary function.Spirometry and single-breath measurements of diffusing capacity for carbon monoxide were performed by the CHESTAC-8800 spirometer (Chest, Tokyo, Japan), according to current international best practices [17,18].Results were expressed as percentages of Japanese normal predicted values [19,20].

Echocardiography
The LV end-systolic and end-diastolic diameters, and left atrial diameter were measured by M-mode echocardiography.LVEF was measured according to the modi ed Simpson method [21].The RVSP was estimated based on a Doppler assessment of peak velocity of tricuspid regurgitation [22].In our study, we used the data from the echocardiographic assessment performed at the time of admission.

Data Collection
We reviewed and recorded the clinical characteristics of all the study patients, which included the physical examination and standard laboratory tests upon admission, usage of long-term oxygen therapy as outpatient, and IPF treatment before the admission.

Survival assessments
We evaluated the hospital, 30-day, and 90-day mortality of the patients.All deaths were obtained from review of the hospital charts.

Recurrence of HF
Recurrence of HF in patients with IPF was reviewed.The time to recurrence was de ned as the number of days from the date of hospitalization for the rst episode of HF during the study period until the date of hospitalization for a recurrence of HF.

Statistical analysis
Continuous variables are presented as means ± standard deviation.Categorical variables are presented as frequencies.Comparisons between categorical variables were performed by the Fisher exact test.Comparisons between parameters for the types of HF (HFpEF, HFrEF, and HFPH) were performed by oneway analysis of variance, followed by the Bonferroni correction for multiple comparisons.For all tests, P < 0.05 was considered statistically signi cant.Analysis was performed by Stat ex ver.6 software (Artech, CO., Ltd., Osaka, Osaka, Japan).

Results
Figure 1 shows the study inclusion owchart.Thirty-seven patients with IPF were hospitalized because of HF from January 2008 through December 2018.Three were excluded because they did not undergo echocardiography.Eventually, a total of 34 patients were included in this study.Among the 34 patients, 17 (50.0%)were classi ed with HFpEF, 6 (17.6%) with HFrEF, and 11 (32.3%) with HFPH.The baseline characteristics of the patients before admission are shown in Table 1.The differences between the parameters of the characteristics of the 3 study groups were not signi cant, except for the numbers of patients receiving corticosteroid therapy for IPF.
The clinical data of patients determined at admission are shown in Table 2. Signi cant differences were observed between the values for BNP, LV end-systolic diameter, LV end-diastolic diameter, LVEF, and RV end-systolic pressure (P=0.01,P<0.0001, P=0.0005, P<0.0001 and P<0.0001, respectively).Intergroup comparisons found that patients with HFpEF had signi cantly lower BNP levels than patients with HFrEF and HFPH (P=0.01 and P=0.0004, respectively).Patients with HFrEF had signi cantly higher values for LV end-systolic and end-diastolic diameters than patients with HFpEF (P<0.0001 and P=0.01, respectively) and HFPH (P<0.0001 and P=0.0004, respectively).Patients with HFpEF had a signi cantly higher value for LV end-diastolic diameter than patients with HFPH (P=0.01).Patients with HFrEF had signi cantly lower values for LVEF than patients with HFpEF and HFPH (P<0.0001 and P<0.0001, respectively).Patients with HFPH had signi cantly higher values for RV end-systolic pressure than patients with HFpEF and HFrEF (P<0.0001 and P=0.0002, respectively).
The outcomes of patients during the rst hospitalization are shown in Table 3.The 30-day mortality rates of patients with HFpEF, HFrEF, and HFPH were 0%, 16.6%, and 36.3%, respectively, with a signi cantly better survival for the patients with HFpEF (P=0.02).The 90-day mortality rates were 11.7%, 16.6%, and 45.4% showing survival rates from best to worst for patients with HFpEF, HFrEF, and HFPH, respectively; although the differences were not signi cant.Differences between recurrence rates for the 3 patient groups were not signi cant.However, the recurrence rate of patients with HFpEF was approximately 2-to 3-fold higher than the recurrence rates of the 2 other patient groups.

Discussion
IPF is a progressive brosing interstitial pneumonia that increases in prevalence with advanced age [1].The median age for the diagnosis of IPF is about 70 years [23,24].Some patients with IPF require hospitalization for HF during the clinical course of the disease [3].We classi ed the patients with HF into 3 groups, those with HFpEF, HFrEF, or HFPH, and compared the groups.HFpEF was the most frequent type of HF.The patients with HFpEF showed the most favorable survival, but also recurred at the highest rate.
Approximately 50% of patients with HF have been reported to have a normal or nearly normal LVEF [6-8].
Old age, female sex, hypertension, high body mass index, smoking, and diabetes were reported to be risk factors for HFpEF [9,25].Patients with HFpEFalso can have a respiratory comorbidity such as chronic obstructive pulmonary disease [26,27], which leads to increased risk of mortality [27].
Although only a few studies have reported on the relationship between interstitial lung disease and left HF, it has been suggested that some patients with IPF have HFpEF, or diastolic dysfunction [28][29][30].Our study focused on LV function in patients with IPF who were hospitalized because of HF.We found that some patients with IPF had HFpEF when they were hospitalized.
We de ned patients with HFPH as those patients with an RVSP ≥ 50 mm Hg on echocardiography.The patients with HFPH showed the highest mortality rate.We could not clarify whether or not the PH we identi ed was a new condition that began at the time of hospitalization.Some patients with IPF and existing PH also need hospitalization.In those cases, the hospitalization may be associated with the existing PH and related right HF [10].Patients who have IPF and chronic, stable PH have a poor prognosis [31,32].Although several clinical trials for patients with IPF and PH have been conducted, there is no effective treatment [33][34][35][36][37][38].This might account for the poor survival of our hospitalized study patients with acute HF and PH coinciding with IPF.
A preventative treatment against the acute worsening of HFpEF remains unknown.Myocardial brosis is reported to be important in the pathophysiology of HFpEF [39].Transforming growth factor beta (TGF-β) has a key role in both the development of myocardial brosis and IPF [40,41].Based on the hypothesis that HFpEF and IPF share common molecular pathways, anti brotic agents are candidates for the treatment of patients with HFpEF and also for the prevention of acute worsening.Actually, pirfenidone, which was approved for patients with IPF, might be effective for patients with HFpEF, given that pirfenidone attenuates the pro brotic effects of TGF-β in human lung broblasts and inhibits cardiac brosis in animal models [42][43][44][45][46].A phase II trial of pirfenidone for patients with HFpEF is now ongoing [47].
Nintedanib is a tyrosine kinase inhibitor that has also been approved for the treatment of IPF [48,49].
Nintedanib has been shown to inhibit TGF-β-induced transformation of broblasts to myo broblasts [50].However, the effect of nintedanib on cardiac brosis has not been proven.
Our study has several limitations.First, the study was conducted in a retrospective fashion.Second, the study was performed at a single center with a small number of patients.Third, both the underdiagnosis and overdiagnosis of HF might have occurred, although we diagnosed HF based on the criteria proposed in the Framingham study [14].IPF and HF share a number of common signs/symptoms such as cough, dyspnea, fatigue, and reduced exercise tolerance in patients with chronic stable disease [51][52][53].In the acute setting, an accurate diagnosis is even more di cult to obtain.Fourth, PH was diagnosed by echocardiography, although right heart catheterization (RHC) is required for the de nitive diagnosis of PH [22].However, performing RHC for all patients with suspected PH who are hospitalized is unpractical because it is an invasive and inconvenient procedure.Fifth, patients with HFpEF or HFrEF associated with PH were classi ed as a single group of patients with HF due to PH [54].Further studies are needed that focus on hospitalized patients with IPF and PH in order to determine an accurate classi cation.And sixth, the LV function and RVSP of patients before hospitalization were unknown.We could not clarify if the HF and PH at hospitalization were due to the worsening of chronic HF and PH or the new onset of these conditions.A prospective study is needed to clarify the times of onset.
In conclusion, HFpEF is the most common type of HF that requires nonelective hospitalization in patients with IPF.Patients with HFpEF survived longer than patients with HFrEF and HFPH.

Figures
Figures