Background

Psittacosis, also known as parrot fever and ornithosis, is a zoonotic infectious disease caused by Chlamydia psittaci (C. psittaci) that is aerobic gram-negative bacterium [1]. Birds are natural hosts of C. psittaci that could spread through feces and respiratory secretions [2]. Humans become infected with C. psittaci through direct contact with infected birds or by inhaling aerosols or dust containing C. psittaci [3]. Furthermore, human-to-human transmission of C. psittaci has been documented [1, 4]. All birds and humans of all ages are susceptible to C. psittaci; however, it is more common in adults, especially those who have contact with pet birds and poultry including pet bird owners, pet store employees, poultry slaughter workers, and veterinarians [5]. Parrot fever is sporadic worldwide, such as the USA [6], Europe [7, 8], Australia [9], and Japan [10]. In China, multiple cases of parrot fever have been increasingly reported in recent years [11, 33]. In our study, we observed highest prevalence of C. psittaci in December, January, and February, which is winter in China. However, parrot fever has been likely to occur in spring and summer in the Netherlands [34]. Another study in the Netherlands further showed that parrot fever occurred significantly higher in spring than that in other seasons [25]. The seasonality may be further studied. Additionally, multiple case reports found that parrot fever might be more common in southwestern China and southern China [12, 35]. Our findings provided similar evidence; however, it might be biased by the difference in the capacity and implementation of mNGS in clinical practice across regions in China. Therefore, it warrants further epidemiological investigation for more evidence.

Moreover, we explored the clinical characteristics of parrot fever. The majority (87.5%) of the parrot fever cases were admitted to the departments of respiratory medicine and intensive care medicine, suggesting they had severe infection. However, in all patients included in our study, complicated or atypical pulmonary infection differed by severity. Pulmonary infection is principal manifestation of parrot fever, so BALF may be recommended for examination. In our study, prevalence of C. psittaci was 2.9% in BALF and 1.1% in sputum, suggesting BALF may be the optimal specimen for examination of C. psittaci. Furthermore, peripheral blood specimens may be optional when respiratory specimens are not available in patients with pulmonary infection [15]. Cell-free DNA of C. psittaci in the cells can be released into peripheral blood after apoptosis in the lungs [36]. Previous study revealed that DNA copies of C. psittaci in BALF were significantly higher than in blood specimens [17]. In our study, we did not find the correlation in the reads of C. psittaci between various types of specimens in 14 cases with both peripheral blood and respiratory specimens. Number of reads is usually related to the collection time and sites of specimens, and possible interaction by co-infection [37]. Nevertheless, detection of C. psittaci in peripheral blood remains crucial for clinical diagnosis, regardless of the number.

Additionally, co-infection of multiple microorganisms was common among the patients with complicated or atypical pulmonary infection. We found Candida albicans was more likely to exist in BALF and peripheral blood specimens in parrot fever cases, while other principal pathogens were likely to be tested in non-parrot fever cases including Streptococcus pneumoniae, Klebsiella pneumoniae, Haemophilus parainfluenzae, Staphylococcus epidermidis, Human herpesvirus 1, and Acinetobacter baumannii complex. However, they did not significantly differ between parrot fever cases and non-parrot fever cases. It might be attributable to the fact that these pathogens are very common in pulmonary infection. Notably, we found a higher co-infection prevalence of Epstein-Barr virus and Staphylococcus epidermidis in peripheral blood specimens of parrot fever cases. Epstein-Barr virus is a human herpesvirus that causes systemic infection by infecting B lymphocytes [38]. Staphylococcus epidermidis is a normal bacterial community in the skin mucosa of the body, which might be detected due to contamination during sampling, instead of real blood infection. If multiple pathogens are detected in peripheral blood, it would indicate serious bloodstream infection. Thus, our findings proved parrot fever cases may have multiple co-infections of pathogens, which would exacerbate the disease. Clinical treatment should be tailored based on the co-infection of C. psittaci and other pathogens [39].

This study also had some limitations. First, we included the patients with complicated or atypical pulmonary infection across 14 provinces and municipalities in China. Limited study regions and implementation of mNGS might result in selection bias. Furthermore, complicated or atypical pulmonary infection might differ by severity across the departments and hospitals, due to variation in the judgment of disease severity. Second, we collected only respiratory or peripheral blood specimens in the parrot fever patients. It could not achieve the comparison between various types of specimens for examination of C. psittaci. In addition, we did not detect C. psittaci in pleural fluid, oral and nasolaryngeal secretion, lung tissue, or cerebrospinal fluid, which might underestimate the prevalence. Third, we included basic demographics of the patients, while did not collect more information such as other laboratory testing. Nevertheless, our study had the strength. Compared with case reports, this study illustrated a scenario of parrot fever epidemiology in China based on a 2.5-year observational study of a moderate sample size.

Conclusion

Prevalence of parrot fever remains low and sporadic in China. It was significantly associated with senior age, winter-spring, and certain regions such as southwestern China and central and southern China. Application of mNGS showed an optimal performance for clinical diagnosis in the detection of C. psittaci, particularly in BALF. Moreover, parrot fever cases might have diverse co-infection of other principal pathogens, such as Candida albicans, Epstein-Barr virus, and Staphylococcus epidermidis. It warrants further studies on the influence of co-infection on the disease severity.