Impact of pulmonary hypertension on outcomes of influenza pneumonia patients: A nationwide analysis

Introduction

Despite many significant medical advances, community-acquired pneumonia (CAP) significantly contributes to global morbidity and mortality.¹ The influenza virus accounts for 15–20% of CAP cases,² and severe influenza infections cause pneumonia in >50% of affected patients, leading to multiple organ dysfunction.³ Influenza pneumonia (IP) has mortality rates similar to that of patients with pneumonia caused by bacterial or other viral pathogens.⁴ Between September 2018 and February 2019, the Centers for Disease Control and Prevention in the United States reported a weekly mortality rate of 5.5% to 7.4% attributed to pneumonia and influenza infection, indicating the significant impact on the population and resources every season.⁵ IP patients frequently develop complications during hospitalization and can rapidly develop acute lung injuries requiring mechanical ventilation.⁶ Pulmonary heart disease and diseases of the pulmonary circulation (PHDPC) encompasses a wide range of conditions, including pulmonary embolism, various types of pulmonary hypertension, and diseases of pulmonary vessels.⁷ They can be a significant cause of morbidity and mortality for IP patients and the utilization of resources. However, in IP patients, there is a lack of data on the role of PHDPC on mortality and other outcomes, including the need for mechanical ventilation and infectious complications. Hence, we used the multicentric national inpatient sample database for those admitted for IP to compare all-cause in-hospital mortality, in-hospital complications, and resource utilization between patients with and without PHDPC.

Methods

Data selection and study population

We extracted influenza pneumonia (IP) hospitalizations using the ICD-10-CM codes (Extended data: Supplementary Table 1⁹) in any disease diagnosis field. We used relevant ICD-10-CM codes in the secondary diagnosis fields to extract patients with PHDPC (Extended data: Supplementary Table 1⁹). Next, we evaluated the IP hospitalizations for the 1st and 99th percentile distribution, which included hospitalizations for patients aged 22 to 90 years. Our two cohorts (Figure 1) comprised the study arm, which included patients with PHDPC, and a control arm with patients without underlying pulmonary heart disease and diseases of the pulmonary circulation (non-PHDPC).

Figure 1. Patient selection and study design.

Results

Of the 121,097,410 weighted discharges in the NIS datasets 2016–2019, 353,460 influenza pneumonia-related hospitalizations were found between 2016 and 2019 for ages 22 years to 90 years based on the 1st and 99th percentile age distribution of IP. Of these, 6.5% (n = 23,000) had PHDPC. Table 1 details the baseline characteristics between the two cohorts. The PHDPC cohort was older (mean age, 72.3 years vs. 67.5 years), had more females (59.0% vs. 52.2%) and patients of African American (AA) race (16.2% vs. 13.3%). Medicare was the primary expected payer in both cohorts, and Medicare enrollees were significantly higher in the PHDPC cohort (77.4.% vs. 65.4%). IP patients with PHDPC were more likely to be admitted to large bed-size hospitals (48.6% vs. 46.4%) and urban teaching hospitals (69.6% vs. 64.1%) than the non-PHDPC cohort. Amongst comorbidities, complicated hypertension, diabetes with chronic complications, heart failure, atrial fibrillation or flutter, valvular heart disease, peripheral vascular disease, dyslipidemia, obesity, obstructive sleep apnea, chronic pulmonary disease, renal failure, hypothyroidism, autoimmune conditions, non-metastatic solid tumors, and tobacco use was significantly more frequent within the PHDPC cohort.

Rates of outcomes and regression analysis (before and after matching by EB) are depicted in Table 2. The PHDPC cohort had a significantly higher rate of in-hospital mortality (8.9% vs. 5.8%; P <0.00, in-hospital complications that included sepsis (4.8% vs 3.9%, P = 0.004), septic shock (11.3% vs 8.8%, P <0.001), cardiogenic shock (1.9 vs 0.8%, P <0.001) and need for mechanical ventilation (18.6% vs 12.7%, P <0.001). Moreover, patients with PHDPC had a higher need of mechanical ventilation for 24 to 96 hours (7.6% vs 4.9%, P <0.001) and more than 96 hours (9.5% vs 6.6%, P <0.001). However, although the need for MV for less than 24 hours, secondary pneumonia, and complications of mechanical ventilation showed a higher trend in PHDPC, there were no statistically significant differences between the two cohorts. After matching by EB and multivariate regression analysis, the PHDPC cohort had higher adjusted odds of in-hospital mortality (aOR 1.4, 95% CI: 1.21–1.61; P <0.001), sepsis (aOR 1.3, 95% CI: 1.08–1.57), septic shock (aOR 1.3, 95% CI: 1.11–1.44), cardiogenic shock (aOR 1.7, 95% CI: 1.25–2.31), need for mechanical ventilation in overall (aOR 1.4, 95% CI: 1.27–1.58), need for mechanical ventilation for 24–96 hours (aOR 1.3, 95% CI: 1.14–1.56) and need for mechanical ventilation for more than 96 hours (aOR 1.4, 95% CI: 1.19–1.60). The PHDPC cohort had a higher comorbidity index for the risk of all-cause 30-day readmission (5.0 vs. 4.1, P <0.001) than the non-PHDPC cohort. In addition, the mean length of hospital stay was longer in the PHDPC cohort (8.7 days vs. 6.8 days, IRR 1.2, 95% CI: 1.12–1.20; P <0.001), with a higher associated mean cost of stay (113501.7 USD vs. 87530.4 USD, IRR 1.2, 95% CI: 1.13–1.25; P <0.001). In addition, statistically significant differences in hospital disposition were also appreciated, with PHDPC patients requiring frequent transfers to other facilities or needing home health care (54.5% vs. 40.8%, P <0.001).

Discussion

In this multicentric retrospective cohort study of IP patients comparing patients with and without PHDPC, we derived the following significant findings, which were found to be significant both before and after matching by EB: 1) Patients with PHDPC had a 38% higher risk of in-hospital mortality as compared with non-PHDPC patients; 2) PHDPC was associated with a higher risk of in-hospital complications including sepsis, septic shock, cardiogenic shock, and need for mechanical ventilation for more than 24 hours compared with non-PHDPC; 3) the comorbidity index for the risk of all-cause 30-day readmission was higher in PHDPC than in non-PHDPC patients; 4) PHDPC was associated with higher resource utilization (longer LOS, higher cost of hospital stay, higher transfers to skilled nursing facilities or Intermediate Care Facilities, higher need of home health care) compared with non-PHDPC.

Influenza is most common in the young,¹⁴ the elderly, the pediatric population, and those with underlying medical conditions and they are most at risk for hospitalization and severe complications of pneumonia due to influenza.¹⁵ We included only adult hospitalized IP patients in our study and had a higher proportion of female and AA patients in the PHDPC cohort. AA patients and females are a relatively vulnerable population for venous thromboembolism¹⁶ and pulmonary embolism due to hypercoagulable conditions like pregnancy, hereditary factor V Leiden and hormone replacement therapy, and predominance of disorders like idiopathic pulmonary hypertension in females.¹⁷ Hence, consistent with the findings in our study, females, and AA make up a relatively higher proportion of the PHDPC cohort. The PHDPC cohort has a higher burden of comorbidities, a higher Elixhauser comorbidity index, and higher adjusted odds of several complications than the non-PHDPC cohort. As evident in our study and reported by previous studies, the risk of hospitalization, poorer outcomes, and death due to IP increases in the presence of other comorbidities.^18,15,19 Hence, these comorbidities do predict poorer outcomes; still, even after adjusting the comorbidities and demographics on multivariate regression analysis, PHDPC was an independent predictor for worse outcomes.

Influenza pneumonia is a notorious disease with poorer outcomes in patients with comorbidities.¹⁵ We found significantly higher rates and odds of mortality in PHDPC patients compared to the ones who did not have PHDPC. Influenza infection has a detrimental effect on pulmonary circulation by causing pulmonary parenchymal inflammation and edema, interfering with alveolar gas exchange, resulting in ventilation/perfusion imbalance and hypoxemia. This hypoxia and carbon dioxide retention will cause the reflex spasm of pulmonary blood vessels and increase pulmonary circulation pressure.²⁰ However, in patients with preexisting resistance to flow due to pulmonary hypertension²¹ or chronic thromboembolism,²² the pulmonary and systemic circulation is already compromised, and a superimposed influenza infection will burden the already compromised pulmonary circulation and increase RV overload. In addition, multiple studies reported that influenza infection is independently associated with increased atherosclerosis and acute cardiovascular events.^23–25 Therefore, various mechanisms cumulatively result in higher mortality and complications associated with influenza pneumonia.

Infection and pneumonia due to the influenza virus significantly interact with the immune system, and this can result in sepsis directly or indirectly secondary to a bacterial infection.²⁶ Sepsis has been previously reported with influenza infection.^27,28 A 2009 national study by Jain et al.²⁹ reported a rate of 18% sepsis on admission in patients with influenza pneumonia. In our study, we report sepsis at a rate of 4.8% vs. 3.9% and septic shock at 11.3% vs. 8.8% among the two cohorts. The difference in rates can be due to multiple reasons, as sepsis is recorded according to clinical judgment and may not have adhered to strict definitions written in critical care guidelines³⁰ and the coding errors with sepsis and septic shock. Sepsis is a broader term for life-threatening organ dysfunction caused by a dysregulated host response to infection.³⁰ At the same time, septic shock is a subset of sepsis in which particularly profound circulatory, cellular, and metabolic abnormalities are associated with a greater mortality risk than with sepsis alone.³⁰ We found that the PHDPC cohort has significantly higher odds of sepsis and septic shock (both before and after PS matching), which compromised pulmonary circulation can explain. In our study, the rate of septic shock reported is more than double the rate of sepsis, suggesting the significantly high degree of severity and influence of the virus on the immune system and inflammatory response of the body.

In addition, Influenza virus infection is highly associated with acute myocarditis and pericarditis.³¹ Persistent inflammation, as in the case of sepsis, causes depression in myocardial function and increases myocardial oxygen demand.³² The endotoxins and cytokines from infection and inflammation cause left ventricular dilatation and depressed ejection fraction, causing sepsis-induced cardiomyopathy.³³ Furthermore, an increase in sympathetic nervous system activity, which is a primary response to inflammation, causes increased heart rate and vascular resistance; this causes a decrease in cardiac output and coronary perfusion of the heart.³² Together, these mechanisms lead to the development of cardiogenic shock in IP patients, although rare but previously reported.^34,35 To the best of our knowledge, our study is the first to report the rate of cardiogenic shock in IP hospitalizations. We found a meager percentage of IP patients developing cardiogenic shock, 1.9% in PHDPC vs. 0.8% in non-PHDPC. We also found that those with PHDPC have significantly higher odds of developing cardiogenic shock, both before and after matching by EB.

IP patients frequently require admission to the intensive care unit (ICU), and most need mechanical ventilation.^36–38 Moreover, acute respiratory distress syndrome can develop in severe influenza infection, leading to developing or exacerbating pulmonary hypertension due to vessel obliteration, pulmonary vasoconstriction, and microthrombosis due to hypoxia, hypercapnia, and an imbalance in vasoactive mediators.³⁹ In our study, the rate of mechanical ventilation was 18.6% vs. 12.7% in patients with and without PHDPC. Previous studies by Piroth et al.⁴⁰ and Ludwig et al.³⁷ have reported rates of 4% and 6%, respectively. However, our study’s higher rates of mechanical ventilation could be because the above studies have included the pediatric population. In contrast, our study has an adult population only and a higher comorbidity burden among them. In addition, we found significantly higher odds of need for mechanical ventilation for more than 24 hours in the PHDPC cohort. However, there was no significant difference in the need for mechanical ventilation for less than 24 hours between the two cohorts in our study, and this can be explained as IP admission in the ICU requires much longer mechanical ventilation for eight days with an additional two days needed for cleaning, maintenance, and other such functions for a total of 10 days.⁴¹

Our study reported higher resource utilization with a higher risk of all-cause 30-day readmission, longer LOS, and higher cost of hospital stay and discharge or transfer to skilled facilities for patients with PHDPC. Given the burden of pneumonia in our population,^42–44 our findings have important implications. Clinicians need to realize the importance of pre-existing PHDPC in patients with IP and exercise appropriate clinical alertness for their timely recognition of complications. Moreover, health officials need to increase efforts to optimize influenza vaccination rates among the elderly and those with chronic pulmonary conditions to reduce the incidence of pneumonia in these high-risk groups, ultimately reducing the healthcare facility burden. The prevention and optimal management of these patients may significantly reduce the burden of death associated with IP.

Conclusions

In a retrospective cohort study of IP patients from NIS, patients with PHDPC had a higher risk of in-hospital mortality and in-hospital complications, including sepsis, septic shock, cardiogenic shock, and need for mechanical ventilation for more than 24 hours compared with non-PHDPC patients. Moreover, PHDPC was associated with a higher comorbidity index for the risk of all-cause 30-day readmission and higher resource utilization than non-PHDPC. Although our study does not include stratification for vaccination status for the outcome, primary care physicians, cardiologists, and pulmonologists should pro-actively educate such patients on preventive strategies during the flu season.