Clinical Characteristics of Chlamydia psittaci Pneumonia Infection in Central South China

In this retrospective, multi-center study, all consecutive patients who were diagnosed with C. psittaci pneumonia and admitted to nine major tertiary hospitals in Central-South China from October 4, 2018, to October 23, 2020, were enrolled. The inclusion criteria were: first, patients met with the diagnostic criteria for community-acquired pneumonia [14]; second, metagenomic next-generation sequencing (mNGS) from an airway sample or blood revealed a specific DNA fragment in C. psittaci. Diagnostic criteria, clinical classification, treatment and discharge criteria for C. psittaci pneumonia cases were based on the Guidelines of the American Thoracic Society on CAP [15]. This study was approved by the institutional review board of the Second Xiangya Hospital (no. luohong201906). Informed consent was obtained from all included patients.

Epidemiologic and clinical data were extracted from electronic medical records. Clinical outcomes were followed up to October 31, 2020. The date of disease onset was defined as the day when the symptoms were noticed by the patient. A high fever was defined as a body temperature > 39.1 °C. Fever duration day after antibiotic use referred to the time from first use of quinolones or tetracyclines to recover from fever for patients who had fever during hospitalization.

Bronchoalveolar lavage fluid (BALF) or blood was collected based on the standard clinical procedure [16]; 0.3 ml BALF or blood samples were separated in a new micro-centrifuge tube, and DNA was extracted using the TIANamp Micro DNA Kit (DP316, Tiangen Biotech) according to the manufacturer’s recommendation. The DNA library was constructed according to the protocol of the BGISEQ-50 sequencing platform. Constructed library was qualified by Agilent2100 (Agilent Technologies, Santa Clara, CA) and Qubit 2.0 (Invitrogen, USA), and a qualified double-strand DNA library was transformed into a single-strand circular DNA library. DNA nanoballs (DNBs) were generated from single-strand circular DNA using rolling circle amplification. DNBs were qualified using Qubit 2.0. Qualified DNBs were loaded into the flow cell and sequenced (50 bp, single end) on the BGISEQ-50 platform.

Patients were grouped into severe/non-severe pneumonia groups according to the ATS guideline on CAP [15]. Furthermore, patients in the severe group were grouped into survival/non-survival groups according to their prognosis. Continuous variables were tested for normality using the skewness-kurtosis test and expressed as mean ± standard deviation (SD) or median (interquartile range [IQR]) depending on normal/non-normal distribution; continuous variables were compared using Student’s t-test or Mann-Whitney U test, as appropriate. Categorical variables were expressed as number and percentage and compared using the chi-square test. Univariate and multivariate logistic regression models were performed to explore the risk factors for severe cases and hospital death. To build a multivariate regression model, variables with p < 0.05 in the univariate model and known related factors based on clinical consideration were included in the original model and screened by backward method, which removed the most insignificant variable with a significance level > 0.05 stepwise. All data were analyzed by Stata (16.0) and Python (3.7). A two-tailed p value < 0.05 was considered statistically significant in all tests.

One hundred and sixteen patients with C. psittaci pneumonia were included in this study. Their demographics and baseline characteristics are shown in Table 1. For all 116 patients, the mean age was 59.7 (SD 12.5; range 26.0–81.0) years, and 79.3% patients were older than 50 years; 72.4% patients were male. Only 18 patients had a history of exposure to parrots or poultry. C. psittaci pneumonia was diagnosed throughout the whole spectrum of seasons, but was more likely to occur in autumn (38.8%) and winter (33.6%). Forty-seven patients had at least one underlying disease including hypertension (19%), diabetes (18.1%), cardiovascular disease (8.6%), digestive disease (7.8%), cerebrovascular diseases (6.9%), etc. The most common symptoms included fever (96.6%), high fever (78.4%), cough (65.5%), dyspnea (46.6%) and fatigue (44.8%). Time from illness onset to first hospital admission was 7.0 (IQR 6.0–9.2; range 1.0–20.0) days. Among them, four medical staff were associated with clusters.

After admission, 63 and 53 patients were categorized into non-severe/severe subgroups as mentioned above. The severe group were of a significantly older age than the non-severe group (63.2 vs. 56.7 years, p < 0.001). There was no difference in gender, exposure history, season or coexisting conditions between the two groups. Subsequently, we divided the 53 patients in the severe subgroup into survival/non-survival groups according to their prognosis. We found that the non-survival group had a higher proportion of having at least one comorbidity (70% vs. 41.9%, p < 0.05), especially for respiratory diseases (20% vs. 0, p < 0.001), compared with the survival group.

Table 2 shows the radiologic and laboratory findings on admission. Chest computed tomography (CT) on admission was performed in 96 patients. Among these 96 patients with available CT results, 53.1% showed bilateral pulmonary involvement on lung CT, and the most common imaging findings on CT at admission included consolidation (99.0%) and pleural effusion (43.8%), while interstitial changes (5.2%) were rare. Figure 1 demonstrates the representative CT findings of four patients with non-severe or severe C. psittaci pneumonia. Severe cases yielded more prominent bilateral pulmonary involvement on CT than non-severe cases (68.2% vs. 40.4%, p < 0.01).

Laboratory examination data are also given in Table 2. On admission, all patients underwent metagenomic sequencing with airway sample (91.4%) or/and blood sample (14.7%). The median DNA sequence number of airway/blood samples for C. psittaci was 100.0 (IQR 23.8–372.8). This number was higher in the severe and non-survival groups compared with the non-severe and survival groups (113.0 vs. 56.0, p < 0.05; 593.9 vs. 100.0, p = 0.165). The combined detection rates of bacteria, fungi and virus were 16.4%, 18.1% and 10.3%, respectively, and there was no difference in the detection rates for these pathogens between different groups.

Interestingly, most patients presented with increased neutrophils and significantly decreased lymphocyte percentage. In general, compared with the non-severe and surviving groups, the severe group and non-survive group had a higher neutrophil-to-lymphocyte ratio and fewer lymphocytes (all p < 0.05). Biochemistry tests showed elevated alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels in most patients, with median ALT and AST values of 63.5 and 100.1 U/l (IQR 47.8–99.3; 57.9–146.1). However, no obvious difference in ALT and AST values was found between severe and non-severe subgroup patients. For infection biomarkers, erythrocyte sedimentation rate (ESR) and high-sensitivity C-reactive protein (h-CRP) were increased in > 90% of patients. Moreover, compared with the non-severe group, the severe group had a higher level of h-CRP, procalcitonin, d-dimer and lactate dehydrogenase and lower oxygenation index (all p < 0.05).

Treatment and clinical outcomes in 116 patients with C. psittaci pneumonia are shown in Table 3. During hospitalization, 68 (58.6%) of 116 patients received respiratory support therapy, including oxygen (35.3%), high-flow nasal cannula (7.8%), non-invasive ventilation (6.9%) and invasive ventilation (18.1%). More severe cases received mechanical ventilation (non-invasive: 15.1% vs. 0%; invasive: 39.6% vs. 0%, p < 0.01) compared with non-severe cases. During hospitalization, the percentages of patients being admitted to the ICU, requiring mechanical ventilation and death were 50.0%, 20.7% and 8.6%, respectively. The length of hospital stay ranged from 4 to 57 days in discharged patients. Comparison between survival and non-survival in all 106 included patients are shown in Table S2.

As for antibiotics treatment, 98 patients with detailed antibiotic usage data were further analyzed, and the results are shown in Table 4. Ninety-eight patients were sub-grouped according to antibiotic usage. The group with use of tetracycline or quinolones only had no statistically significant association with severe pneumonia, ICU admission or hospital death. However, length of hospital stay (13.8 vs. 18.8 day, p < 0.001) and fever duration days after antibiotic use (3 vs. 4, p < 0.05) were significantly reduced in the quinolone group compared with the non-quinolone group, which was not observed in the tetracycline group.

Univariate logistic analysis was performed to explore the risk factors for severe pneumonia and hospital death, respectively, and the results are shown in Table S1 in Supplementary Appendix. To adjust the confounding effect, multivariate logistic regression models were built for severe pneumonia and hospital death following the previously mentioned method, and the results are shown in Table 5 and 6. In the logistic model for severe pneumonia, we initially included the following variables: age, APACHE II, PSI, WBC, neutrophil percentage, lymphocyte percentage, neutrophil-to-lymphocyte ratio, hemoglobin, albumin, globulin, total/direct bilirubin, lactate dehydrogenase, ESR, procalcitonin, d-dimer, partial carbon dioxide pressure, cardiovascular diseases, any complications, bilateral pulmonary involvement and respiratory therapy except oxygen. The final model revealed that higher globulin concentration (OR 0.838, 95% CI 0.704–0.998, p = 0.048) was a protective factor for survival, and higher APACHE II score (OR 38.342, 95% CI 1.118–1.59, p = 0.001) and respiratory therapy except oxygen (OR 8.304, 95% CI 1.874–36.783, p = 0.005) were risk predictors of severe pneumonia. Respiratory therapy except oxygen refers to all other respiratory therapy except oxygen therapy, including high-flow nasal cannula, non-invasive ventilation and invasive ventilation. In the regression model for hospital death, the included variables were APACHE II, PSI, neutrophil percentage, lymphocyte percentage, platelet count, d-dimer, partial pressure of carbon dioxide, cardiovascular diseases, any complications, bilateral pulmonary involvement and respiratory therapy except oxygen. The final model showed that higher lymphocyte percentage (OR 0.696, 95% CI 0.514–0.942, p = 0.019) was the only protective factor for survival, and higher PSI (OR 5.984, 95% CI 1.812–19.759, p = 0.003) and respiratory therapy except oxygen (OR 10.077, 95% CI 1.361–74.579, p = 0.002) were risk predictors of hospital death.