Chest radiographs as predictors of length of stay in right-sided infective endocarditis

Objective: Right-sided infective endocarditis (IE) occurs less frequently than left-sided IE and is usually caused by intravenous drug use or intravascular device-related infection. Septic pulmonary embolism can cause an abnormal chest radiograph (CXR), possibly raise pulmonary artery pressures, and may influence clinical and hospital outcomes. Methods: We conducted a retrospective chart review of patients diagnosed with rightsided IE from January 2000 to December 2011. Clinical parameters were collected and analyzed to define patients’ characteristics and their association with clinical outcomes, including length of stay (LOS). Results: 208 eligible patients had a diagnosis of IE; 19 (9.1%) had right-sided IE. Most were related to intravenous drug use (73.7%) and presented with dyspnea and fever (55.6%). 78.9% of patients had abnormal CXR (parenchymal involvement, cardiomegaly, or pleural effusion). Echocardiography revealed tricuspid valve involvement in all patients; the median mean pulmonary artery pressure (MPAP) by transthoracic echocardiographic estimation was 26.4 mm Hg. Patients with abnormal CXR had significantly longer LOS than those with normal CXR (21.4 vs. 7.5 days, p = 0.008). MPAP was not associated with LOS (p = 0.72). Conclusions: Right-sided IE is often associated with intravenous drug use. The majority of these patients have mild pulmonary hypertension, which could be due to hyperdynamic circulation and probable septic emboli. CXR with pulmonary involvement may be useful in predicting the length of hospital stay in these patients and identifying patients with more complications.


IntroductIon
Right-sided infective endocarditis (IE) comprises 10% of IE cases and is caused by Staphylococcus aureus infections in 70% of the cases. 1,24][5] Patients with right-sided IE can have septic emboli in the lung and develop septicemic pneumonia, pulmonary infarcts, lung abscesses, bilateral pneumothoraces, and empyema. 1,5,6The clinical presentations, chest radiograph (CXR), and/or echocardiographic parameters may predict and /or influence the clinical outcomes in these patients.In particular, the size of vegetations influences prognosis. 7We speculated that embolic events and vascular changes associated with vegetations would increase pulmonary artery pressures, which, in turn, would predict hospital outcomes.

Methods
We conducted a retrospective review of the medical records of patients diagnosed with infective endocarditis between 2000 and 2011identified by ICD-9 code of 421.1.Only patients with right-sided IE were included in the study.Demographic and clinical data were collected and analyzed.Radiographic reports and images have been reviewed by the authors.Pulmonary artery systolic pressure (PASP) was tabulated from echocardiographic reports and used to calculate to mean pulmonary artery pressure (MPAP) according to equation: MPAP = 0.61 x PASP +2.Clinical outcomes included in-hospital complications (heart failure, septic or cardiogenic shock, pulmonary infarction, and acute kidney injury), length of stay, and mortality.Statistical analysis was performed using ttests, analysis of variance (ANOVA), chi-square tests, Pearson's correlation and non-parametric correlation, and Mann-Whitney U tests using IBM SPSS Statistics version 20.0 (Armonk, NY).P-values < 0.05 were considered statistically significant.This study was approved by the Institutional Review Board at Texas Tech University Health Sciences Center, Lubbock, Texas.
There was no correlation between PASP and length of stay (r= -0.11, p-value = 0.72).Patients with abnormal CXRs had longer lengths of stay than those with normal CXRs (mean length of stay ± standard deviation = 21.4 ± 15.3 days vs. 7.5 ± 2.4 days, p = 0.008).There was also no association between length of stay and age, ethnicity, body mass index, presence of comorbidities, or number of comorbidi-ties in this study.Mean PASP in patients with normal CXR and with parenchymal involvement group were 29.2 ± 25.1 mmHg and 25.8 ± 4.2 mmHg, respectively (p=0.73).

Patient Characteristics
Frequency Percent

dIscussIon
The incidence of right-sided IE is significantly lower than left-sided IE and usually ranges between 5-10 percent of all IE cases. 8Possible explanations include the lower rate of right-sided congenital heart disease, the lower pressure and shear stress on rightsided valvular structures, and the lower oxygen content of venous blood.However, among intravenous drug users, the proportion of right-sided IE increases to 86% and is even higher in those who have concurrent HIV infection with low CD4+ lymphocyte count. 1,9ven though IVDU is the classic cause of right-sided IE, intravenous catheters and intracardiac devices are also important causes.
The relationship between IE and pulmonary hypertension has not been clearly elucidated.Our patients had mild pulmonary hypertension in the majority of cases.However, accurate MPAP measurements from right-sided heart catheterization were not performed, resulting in a limited analysis between actual MPAP and clinical parameters in our study.Hyperdynamic circulation from sepsis, pulmonary septic emboli, pulmonary granulomatosis from injection of "stabilizer" from drug abuse, and HIV infection can cause pulmonary hypertension.In our series, however, HIV infection is not a factor since only one patient was HIV-positive.Pulmonary septic emboli and infiltrates seen on the chest radiographs probably do not explain pulmonary hypertension either since there was no significant difference in MPAP between the normal chest radiograph and abnormal CXR (parenchymal involvement) groups.Although there was no definite evidence, subclinical septic embolism to the lungs could be a plausible cause of pulmonary hypertension in these patients.
Parenchymal involvement was common in our series and was also associated with increased length of stay.This reflected the importance of pulmonary involvement in right-sided IE.The pathology of parenchymal involvement is likely complex and includes septic embolism, parenchymal infarction, and community-acquired and health-care associated pneumonia.These possibilities would increase the length of hospital stay.
The role of septic pulmonary embolism in pulmonary hypertension is uncertain.Studies of thrombotic pulmonary embolism revealed that pulmonary embolism must occlude more than 30% of pulmonary vasculature to cause pulmonary hypertension. 10In addition to mechanical acute obstruction, there are also changes in the small vessels downstream from the emboli including medial hypertrophy, fibrointimal hyperplasia, microvascular thrombosis, and plexiform lesion. 11Thrombus formation and thrombolysis also release pro-inflammatory cytokines and vasoactive factors contributing to those changes. 12Septic pulmonary embolism can also cause mechanical obstruction and elicits inflammation.The mechanisms of development of pulmonary hypertension are probably similar to those of chronic thromboembolism with possibly more inflammation.
The limitations of our findings include the retrospective, single-center study design, small number of patients in the analysis, and incomplete or missing data.Moreover, none of the patients included had right heart catheterization done.Abnormality on the CXR may indicate other concomitant acute or chronic disease processes, which may not directly related to right-sided IE. conclusIon Right-sided IE often occurs in Staphylococcal septicemia and is usually associated with intravenous drug use.The majority of these patients have pulmonary hypertension, possibly from septic emboli.However, there was no association between PASP and length of stay.CXR can help predict the length of hospital stay in these patients.Additional studies with more patients and better statistical power are needed to provide more information on the importance of echocardiographic parameters on the clinical course in patients with right-sided IE.