Septic pulmonary embolism in China: clinical features and analysis of prognostic factors for mortality in 98 cases

Septic pulmonary embolism (SPE), unlike the much more common thrombotic pulmonary embolism, is a rare disorder in which a bacterial embolus containing microorganisms originating from the primary extrapulmonary infectious foci [1] (such as liver abscesses [2], peritonsillar abscesses [3], infection of skin and bone [4], infective endocarditis [5, 6], periodontal disease [7], intravascular devices and catheters [3]) obstructs the small pulmonary vasculature, causing sepsis and a focal abscess in the lung. SPE is associated with a high mortality rate and remains a diagnostic challenge in clinical practice due to its insidious onset, nonspecific clinical manifestations (such as fever, cough, dyspnea, chest pain and hemoptysis), and life-threatening complications [8]. Therefore, early diagnosis and appropriate treatment play important roles in reducing mortality [9]. However, histopathologic examination is rarely available in clinical practice, and there is a lack of unified diagnostic criteria and prognostic factors for SPE based on clinical, radiological and laboratory features [3]. The present study aimed to not only characterize the clinical features of SPE but also analyze the prognostic factors for mortality.

The present study retrospectively analyzed a large cohort of 98 SPE patients. The results showed that fever was the most common clinical symptom and MSSA was the most common pathogen in blood culture. Moreover, bilateral multiple peripheral nodules were seen on chest CT scans, and echocardiography was performed in some of the patients, suggesting vegetations. These results agree well with previous studies reported in the literature [3, 4, 6, 13]. However, the clinical characteristics of the patients included in the present study differed from those of patients in previous reports, including the most common infectious foci and comorbidities. Furthermore, bronchoscopy examination of some SPE patients demonstrated acute bronchitis, suppurative pneumonia and organizing pneumonia, which were not reported previously. Finally, the present study found that the independent risk factors for SPE mortality were age, hypotension and ineffective or delayed EAT, whereas drainage treatment was a protective factor against mortality. In contrast to a previous study, tachypnea (respiratory rate > 22/min) was not found to be an independent risk factor for mortality [14].

Historically, SPE was known as a deadly complication in the pre-antibiotic era following septic pelvic thrombophlebitis due to either septic abortion or postpartum uterine infection [15]. In the 1970s, the epidemiology of the disease was changed with the use of catheters, and it was discovered that many patients suffering from SPE had infective endocarditis due to intravenous drug abuse [8] [15]. In the 2000s, Cook et al. reported that Lemierre syndrome and central venous catheter infection were the common causes of SPE [3]. Moreover, two studies from Taiwan and one from Japan reported on SPE cases caused by liver abscesses [16, 17] and periodontal disease [7], respectively. Recently, several studies demonstrated that bone, skin, and soft tissue infections are the most common source of infection [4, 6, 14]. Similarly, the present study also found that skin and soft tissue infection (30.6%) was the most common foci of primary infection, followed by infective endocarditis (IE, 20.4%). Unfortunately, the present study did not study the management of these primary infections in detail since diagnosis and management of the primary infections was carried out by consultants from corresponding departments.

In the present study, almost half of the patients had a diabetes diagnosis (41.8%), which was much higher than the previously reported incidences of 16.7, 17.1 and 28.6% [4, 6, 7]. This could be related to the increased incidence of diabetes in China in the recent years. Interestingly, we found that the incidence of diabetes was as high as 82.4% in patients with SPE caused by liver abscess, similar to two other studies from Taiwan, which reported incidences of 66.7 and 85.7% [2, 17]. However, no studies have reported diabetes as an independent risk factor for mortality in SPE, probably because most patients with diabetes can receive timely and effective treatment to keep their blood glucose under control after diagnosis.

Echocardiography could aid the diagnosis of IE which can eventually lead to SPE. A total of 39 (39.8%) patients received echocardiography in this study, which was lower rate than other similar studies (57.1–100%) [3, 4, 6, 13, 14]. The reasons for this were as follows: 1. Some patients did not undergo echocardiography because SPE and their primary infection foci were clearly diagnosed and there was no reason to suspect IE and comorbidity of heart disease; 2. Some patients were initially misdiagnosed as other diseases without undergoing echocardiography, but later improved on antibiotic treatment against IE and because of the lack of basic heart disease, echocardiography was not performed; 3. Some patients refused to undergo echocardiography for financial reasons (some medical insurance does not cover this examination) despite recommendation by doctors.

The present study was, to our best knowledge, the first to describe bronchoscopy findings of SPE patients in detail. The purpose of bronchoscopy and TBLB was to determine the pathogen in infectious pulmonary diseases and to exclude other pulmonary diseases. However, it would be unethical to perform this procedure in every single patient not every patient because it is a high-risk procedure for SPE patients, with risks such as fever, dyspnea, hemoptysis and pneumothorax, and the procedure is therefore only reserved for the following patients: (1) who did not reveal culture results from blood and primary foci, (2) who did not present with typical infectious manifestation, (3) to exclude other pathology such as pulmonary TB and cancer with lung to lung metastasis, (4) who did not have fever, pulmonary hypertension, coagulopathy, organ dysfunction and other contraindications, (5) TBLB was not performed on lesions through chest CT scan, including cavitary septic foci or abscess, that could lead to pneumothorax [18]. In our study, only 24 patients with atypical manifestation underwent bronchoscopy and 4 patients underwent TBLB. The results from differential cell counts and microbial culture of BALF as well as lung biopsy were consistent with deep airway infectious disease. Besides, CT-guided or ultrasound-guided percutaneous needle biopsy have also been reported to be effective diagnostic methods for peripheral pulmonary lesions (PPLs), with a high diagnostic rate of 81.8–97% and acceptable complication rate of 8–17% [19, 20]. In this study, 4 patients with PPLs refused the above two examinations because of the results of enhanced CT and ultrasound, which showed abundant blood vessels and high risk of hemoptysis. Furthermore, endobronchial ultrasound (EBUS)-guided transbronchial biopsy (TBB) has a high diagnostic accuracy of 61.8–77% [21, 22]. Unfortunately, our hospital only had short experience with this procedure and lacked sufficient clinical data. Increased use of interventional examinations could improve the diagnosis rate of SPE patients without typical infection manifestation in the future.

Some studies have suggested that septic shock and heart failure are both independent risk factors of IE [23, 24]. However, among the patients who had both IE and heart disease in the present study, the hypotension seen in some patients could not be explained by either infection or heart failure and could lead to septic shock and heart failure. Therefore, hypotension was evaluated as a risk factor for SPE mortality. Consistent with the previous results, the present study showed a significant increase in the risk of mortality if patients remained hypotensive after appropriate antibiotics and therapy for heart failure.

Antimicrobial therapy plays an important role in SPE treatment. It is well known that inappropriate initial antibiotic treatment, including delayed administration of antibiotics and ineffective EAT, increases mortality in sepsis or septic shock cases [25‐27]. Similarly, our analysis showed that SPE patients who received ineffective or delayed EAT had a seven-fold increased risk of mortality. Although there is a lack of standardized SPE guidelines, we believe that appropriate antibiotic treatment should be given as soon as possible, similar to the recommendation for sepsis patients with or without septic shock [28‐30], because SPE is essentially a pulmonary infection caused by bacteremia resulting from another primary infectious site. Although our findings showed that the most common pathogen was MSSA, the positive rates of Gram-negative bacteria (Klebsiella pneumonia and Escherichia coli) as well as fungal organisms were also high. These results suggest that both Gram-positive and Gram-negative bacteria should be covered with broad-spectrum antibiotics when choosing the initial EAT and antifungal therapy should also be administered. The choice of EAT can be based on a patient’s history, the anatomic site of infection, the patient’s immunosuppression status and local microbiology prevalence [31‐34] and adjusted according to results of microbial culture if EAT is ineffective. However, there are no clear guidelines on the optimal duration of the course of antibiotic therapy for SPE patients. In the present study, the mean duration of total antibiotic therapy was 48.3 ± 22.6 days, which was consistent with the 4–8 weeks reported in previous publications [3, 4, 6, 13]. The decision to stop antibiotics can be based on clinical improvement, subsequent culture results, levels of inflammatory indicators and radiological findings [35].

Clinical experience suggests that it is difficult to achieve control and improve the symptoms of systemic infection by administering antibiotics if patients do not receive timely and effective control of the primary infection locally, such as the drainage of an abscess, debridement of infected necrotic tissue, and removal of a potentially infected device [16, 36‐38]. Some studies have suggested that drainage therapy for infectious foci should be given together with antibiotics, but no further analysis of the effect of drainage on SPE mortality has been reported [2, 3, 6, 17]. The present study has now confirmed that drainage of the infectious foci is a protective factor against mortality in SPE. Patients who received drainage had a lower risk of death. Therefore, we recommend drainage of abscesses as soon as possible once a specific primary infection is identified.

Some studies suggested that cardiac surgery is an independent protective factor that improves the prognosis in IE patients [23, 24, 39]. However, due to the small sample size in the present study, we were unable to assess the relationship between cardiac surgery and SPE mortality. Among the six patients who received cardiac surgery in the present study, the survival rate was relatively high (66.7%). However, a prospective study with a larger sample size is required to more closely assess this relationship.

In addition, we speculated that deep fungal infection (particularly candidemia), infections caused by MDRPs and severe coagulation disorders could also be risk factors for SPE mortality, because these conditions are more difficult to control clinically. The results of univariate analysis of candidemia and coagulation disorders between the mortality and survival groups were consistent with this hypothesis. However, these two factors were not included in the logistic regression analysis due to the small sample size. Deep fungal infection and infection caused by MDRPs were on the other hand included in the logistic regression analysis, because they were shown to differ significantly between the mortality and survival groups in the univariate analysis. However, they were not identified as independent risk factors. Nevertheless, we recommend that SPE patients with risk factors for invasive fungal infection, such as immunosuppressed status, long-term use of broad-spectrum antibiotics, prolonged hospital/ICU stay, prolonged invasive vascular catheters, receipt of parenteral nutrition and use of H2 blockers should receive treatment with antifungal drugs empirically to avoid a poor prognosis [40, 41]. At the same time, attention should be paid to the rational use of antibiotics to reduce the incidence of MDRP infections.

MDRP is a growing issue globally that endangers the lives of many patients. Twenty-five cases of MDRP were found in the present study. No relationship between MDRP and the type of primary infection was found. The occurrence of MDRP is increased in the presence of the following risk factors: prolonged hospital stay, recent use of antimicrobial agents, prior hospitalization, and prior colonization or infection with MDRP [35]. These risk factors should be considered in the choice of empirical antibiotic therapy. For examples, if there are risk factors for MDRP in patients with skin and soft tissue infection and IE, in which the Gram-positive bacteria is most commonly the pathogenic agent, MRSA infection should be considered and empirical anti-MRSA agent, such as vancomycin, linezolid or teicoplanin, could be used when the condition is critical.

The present study has several limitations. First, this study was a single-center retrospective study, which inevitably results in bias. Second, the positive rate of blood cultures was much lower than those reported in developed countries (70–90%) [3, 4]. This was most likely because many of the patients had already received antibiotics before admission to the hospital where blood culture was performed. It was therefore difficult to gather enough data to conduct a good statistical analysis. Third, TTE was only performed in a limited number of SPE patients, and none underwent transesophageal echocardiography, which could have led to some missed IE cases. Finally, the present study lacked follow-up data. Only in-hospital mortality was analyzed as it was not possible perform a survival analysis and explore the impact of prognostic factors on overall mortality.

In conclusion, the present study demonstrated that SPE should be suspected in high-risk patients who present with typical clinical manifestations (fever, dyspnea, chest pain) and radiologic features (particularly multiple nodules in bilateral lobes on chest CT). Undergoing blood culture before the use of antibiotics and identifying primary infection foci are important to make early diagnosis of SPE. Gram-positive bacteria are responsible for a higher proportion in SPE patients with skin and soft tissue infection and infective endocarditis, while Gram-negative bacteria are mainly found in SPE patients with liver abscess. Bronchoscope is strongly recommended to exclude noninfectious diseases, if patients present with atypical infectious symptoms. Extra attention should be given to SPE patients over 56 years old, those who are hypotensive and those who received ineffective or delayed EAT, as they could be at high risk for in-hospital mortality. Timely drainage treatment of infectious foci could reduce mortality.