The occurrence and risk factors associated with post‐traumatic stress disorder among discharged COVID‐19 patients in Tianjin, China

Background Post‐traumatic stress disorder (PTSD) is a serious mental health condition that is triggered by a terrifying event. We aimed to investigate the occurrence and risk factors of PTSD among discharged COVID‐19 patients. Methods This study included 144 discharged COVID‐19 patients. PTSD was assessed by using validated cut‐offs of the impact of event scale‐revised (IES‐R, score ≥25). All patients completed a detailed questionnaire survey, and clinical parameters were routinely measured in the hospital. Binary logistic regression models were applied to identify factors associated with PTSD. Results Of the 144 participants with laboratory‐confirmed COVID‐19, the occurrence of PTSD was 16.0%. In multivariable analyses, age above 40 years (adjusted OR [95% CI], 5.19 [2.17–12.32]), female sex (adjusted OR [95% CI], 7.82 [3.18–18.21]), current smoker (adjusted OR [95% CI], 6.72 [3.23–15.26]), and ≥3 involved pulmonary lobes (adjusted OR [95% CI], 5.76 [1.19–15.71]) were significantly associated with a higher risk of PTSD. Conversely, history of hypertension and serum hemoglobin levels were significantly associated with a lower risk of PTSD with adjusted ORs (95% CI) of 0.37 (0.12–0.87) and 0.91 (0.82–0.96), respectively. Conclusion Old age, gender (being female), current smoking, bacterial pneumonia, and ≥3 involved pulmonary lobes were associated with an increased occurrence of PTSD among discharged COVID‐19 patients.


INTRODUCTION
Coronavirus disease 2019 , caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has affected more than 213 countries worldwide, resulting in an ongoing global pandemic.
According to the statistics available in Worldometer (Worldometer, 2020), as of July 20, 2020, COVID-19 had caused more than 14 million cases and nearly 610,000 deaths in the world.
The COVID-19 not only affects physical health but also causes mental health problems (Talevi et al., 2020;Vindegaard & Benros, 2020). One of the common psychological problems is post-traumatic stress disorder (PTSD) arising from exposure to trauma (Pfefferbaum & North, 2020). Bo HX et al. reported that 96.2% out of 714 hospitalized stabile patients suffered from significant posttraumatic stress symptoms associated with the COVID-19 before discharge (Bo et al., 2020). A previous study in Iraq and Afghanistan found that veterans who reported PTSD have greater difficulties in their romantic relationships, less family cohesion, less social support, poorer social functioning, and lower life satisfaction compared to subjects without PTSD (Tsai et al., 2012). Furthermore, PTSD was associated with attempted suicide and resulted in a high economic burden for society (Galea et al., 2005). Therefore, identifying risk factors of PTSD symptoms in confirmed COVID-19 patients is important for the early detection of at-risk subgroups and potential interventions.
Although several recent studies have investigated the factors associated with PTSD in front-line medical staff , the general population , or university students (Tang et al., 2020), few studies have focused on the factors associated with PTSD among confirmed COVID-19 patients.
Thus, the related factors of the occurrence of PTSD in patients with confirmed COVID-19 infection are unknown. To our knowledge, only one small cross-sectional study (n = 41) on PTSD in COVID-19 confirmed patients has been performed, demonstrating a prevalence of PTSD symptoms in 5/41 (12.2%) of participants (Qi et al., 2020). However, only non-severe types of COVID-19 cases were included, and the median time interval between hospitalization and PTSD assessment was less than one month. On the other hand, in addition to the traditional risk factors for PTSD such as age, sex, smoking, and alcohol drinking, factors associated with the severity of COVID-19 (including peripheral blood indicators, hospitalization days, severe pneumonia, hypoalbuminemia, liver injury, and other severe complications) and patient metabolic characteristics (e.g., obesity) might be associated with risk of developing PTSD (Lee et al., 2018;. Furthermore, information on personal and family history of diseases was available in this study. Thus, we conducted this prospective study with a relatively long-term follow-up to investigate the factors potentially involved in PTSD among confirmed COVID-19 patients with various clinical types by including numerous predictors.

Participants
From January 19 to April 30, 2020, a total of 144 participants were recruited from Tianjin Haihe Hospital, which is a designated hospital for COVID-19 in Tianjin, China. This was a convenience sample of survivors. All participants provided written consent and all the protocol of this study was approved by the Institutional Review Board of Tianjin Haihe Hospital.
A SARS-CoV-2 virus-specific real-time reverse transcriptase polymerase chain reaction (RT-PCR) assay of respiratory specimens, including nasal and pharyngeal swabs is routinely used for diagnosis of COVID-19 . We included all participants with laboratory-confirmed SARS-CoV-2 infection. week on a 5-point scale: 0 = "Not at all", 1 = "A little bit", 2 = "Moderately", 3 = "Quite a bit", 4 = "Extremely". Scores ≥25 were considered to be of clinical concern (Rash et al., 2008). The IES-R is the most widely used measure internationally for the psychological impact of trauma, and high reliability and validity were observed in the Chinese version of the IES-R (Wu & Chan, 2003).

Laboratory measurements
Venous blood samples were routinely drawn from each patient in the fasting state at the earliest time points possible upon hospitalization.

CURB-65 score
Data for the calculation of CURB-65 were recorded at hospital admission. The CURB-65 severity score (range 0-5 points) was calculated by attributing one point to each of the following clinical features present: confusion, urea > 7 mmol/L, respiratory rate ≥30/min, systolic blood pressure < 90 mmHg, and/or diastolic blood pressure 60 mmHg, and age 65 years (Lim et al., 2003).

Thin-section CT imaging and scoring
Thin-section computed tomography (CT) scans (Aquillion Prime 128, Canon Medical Systems, Otawara, Japan) were performed in all patients. Thin-section CT was performed from the lung apices to the adrenal glands at full inspiration, and the process was later repeated at full expiration. The CT scanning parameters were: 64 × 0.5 mm collimation, 120 kV, automatic tube current modulation, 0.5 s gantry rotation time. Images were reconstructed with a hybrid iterative reconstruction algorithm (AIDR 3D) at mild settings, and contiguous inspiratory thin-section CT images were obtained with a thickness of 1 mm at 0.8 mm intervals. CT scans were interpreted at window settings that were optimal for lung parenchyma (window level, −600 HU; window width, 1500 HU) and soft tissue (window level, 400 HU; window width, 40 HU). Patients were scanned while lying supine and during full inspiration and expiration. All CTs were scored by two experienced radiologists blinded to laboratory values and the medical history of the participants. Definitions of radiological terms like ground-glass opacity, reticulation, honeycombing, parenchymal bands, consolidation, air trapping, and bronchiectasis were based on the recommendations of the Nomenclature Committee of the Fleischner Society (Hansell et al., 2008).

Measurement of other variables
Age, sex, alcohol drinking status (never or drinking), history of diseases (hypertension, diabetes mellitus, and heart disease), hospitalization days, and oxygenation index (OI) were obtained from a detailed questionnaire survey and electronic medical records.
Data on smoking status were collected by a question: "Do you usually smoke?" with two response options: yes (current) or no (former/never). Height and body weight were measured with participants wearing light clothes and no shoes. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared (kg/m 2 ).

Complication definitions
Severe pneumonia was defined by the 2007 Infectious Disease Society of America/American Thoracic Society (IDSA/ATS) guidelines (Mandell et al., 2007). Two major criteria include the need for invasive mechanical ventilation and septic shock with the need for vasopressors.
The normal range of serum albumin was 35-55 g/L. Based on pre-  creatinine, and urea nitrogen, separately. These variables were potential factors associated with PTSD (Lee et al., 2018;.

Statistical analysis
Stepwise logistic regression analysis with inclusion and exclusion criteria of p < .20 was performed to screen statistically significant the factors associated with PTSD (Chowdhury & Turin, 2020). No significant multicollinearity was found when the variance inflation factor was assessed. All tests were two-tailed and statistical significance was set as p < .05.

RESULTS
Of the 144 participants with laboratory-confirmed COVID-19, the incidence of PTSD in all subjects was 16.0% (23/144

DISCUSSION
To our knowledge, this is the first prospective study to examine factors associated with PTSD in patients with confirmed COVID-19 infection.
The results showed that older age, gender female, current smoking, and the number of involved pulmonary lobes (≥3), are risk factors for PTSD, while hypertension and high hemoglobin levels are inversely associated with the risk of PTSD.
In this study, 16.0% of COVID-19 patients had PTSD 3 months after infection. This incidence rate is higher than that reported for 41 COVID-19 patients at less than 1 month after symptom onset, where 12.2% had PTSD symptoms (Qi et al., 2020). A possible explanation might be that patients who were severely ill were excluded and PTSD was detected over the relatively short term in that study. However, the rate presented in our study was similar to that among nurses exposed to COVID-19 in Hubei, China, which showed that the incidence of PTSD was 16.83% . By contrast, a previous study reported that PTSD prevalence was 2.7% in home-quarantined college students 1 month after the outbreak of the COVID-19 epidemic (Tang et al., 2020). This difference corresponding to previous studies indicated that the prevalence of PTSD is higher among persons who are directly exposed to a disaster (Neria et al., 2008).
In the present study, older age and gender female are the main risk factors associated with PTSD in confirmed COVID-19 patients.
In line with our results, a 4-year follow-up study indicated that survivors of SARS with PTSD were more likely to be older and female (Hong et al., 2009). In addition, a previous study conducted in confirmed COVID-19 patients indicated that COVID-19 patients with mental health problems were more likely to be female (Qi et al., 2020). Note: Continuous variables were described as medians (interquartile ranges), and categorical variables were described as counts (percentages). Abbreviations: BMI, body mass index; CRP, C-reactive protein, CT, computed tomography; PTSD, post-traumatic stress disorder. Two additional studies also found that being female was a strong predictor of PTSD symptoms (Gonzalez-Sanguino et al., 2020;Liu et al., 2020). Similarly, animal studies have demonstrated different mood changes in rats related to sex, with female rats showing stronger emotional changes than their male counterparts (Yang et al., 2019). Collectively, these findings highlight that old age and gender female are the main risk factors associated with PTSD in confirmed COVID-19

TA B L E 1 Characteristics of study participants
patients.
To our knowledge, no studies have investigated how smoking influences PTSD. In the current prospective study, we observed that smoking was positively associated with PTSD in COVID-19 patients. Previous studies also suggested that smoking (nicotine) could cause poor mental health through neurotransmitter pathways (Picciotto et al., 2002;Plurphanswat et al., 2017). Moreover, evidence has suggested that quitting smoking was associated with a decrease in depressive symptoms (Rodriguez-Cano et al., 2016), which are strong risk factors for PTSD (Momma et al., 2014). Thus, the integration of stop smoking services into the PTSD care pathway would be a good way to improve mental health.
Multivariable logistic regression analysis indicated that the COVID-19 patients with ≥3 involved pulmonary lobes were more likely to have PTSD. Previous studies have suggested that COVID-19 patients with ≥3 involved pulmonary lobes had more severe illness (Dudoignon et al., 2020). This might be explained by the fact that a high incidence of PTSD was associated with the severity of exposure to disaster (Lee et al., 2018).
In our study, we found that a high hemoglobin level may act as a protective factor for PTSD in COVID-19 patients. This aligns with previous studies which documented that low hemoglobin level is a risk factor for postpartum PTSD (Sentilhes et al., 2017). These works also suggested that low baseline hemoglobin level strongly predicts incident depression (Trevisan et al., 2016). Taken together, this information suggests that anemia may contribute to PTSD in COVID-19 patients. Indeed, the current study also showed anemia was a risk factor for PTSD, with adjusted OR = 1.59 (95% CI: 0.48-4.59). Our findings suggest that discharged COVID-19 patients with high hemoglobin level had a low possibility of developing PTSD.
Interestingly, our study showed that baseline hypertension was inversely associated with the risk of PTSD. In contrast, a previous study demonstrated that a history of chronic illness was significantly associated with higher IES-R scores among the general population in China . Furthermore, an inverse association between hypertension and depression has been previously reported (Briggs et al., 2016). Nevertheless, the inverse association between hypertension and PTSD should be interpreted with caution as the proportion of COVID-19 patients with hypertension was small, resulting in low statistical power for this finding. Future research should investigate the role of hypertension in the development of PTSD using higher powered prospective studies.
To the best of our knowledge, there is no prospective study to identify the risk factors that develop PTSD in confirmed COVID-19 patients. In this study, we collected many variables at baseline, such as smoking status, drinking status, medical history, complications, clinical parameters, and blood parameters, which have allowed us to prospectively investigate the factors associated with PTSD in COVID-19 patients. Nevertheless, the present study has some limitations. First, PTSD was assessed by telephone follow-up. However, the follow-up was performed by a trained psychiatrist using a standardized interview technique. Second, although this is a prospective study, the baseline PTSD was not assessed. Third, we did not collect data on dietary intake, physical activity, and the history of psychiatry (depression, anxiety, schizophrenia, etc.), which might have affected the incidence of PTSD (Hall et al., 2015). In addition, either COVID-related or non-COVIDrelated interventions such as clinical and/or follow-up social support were not assessed in the study, so the residual confounding existed.
Therefore, future research should be conducted to assess the effect of nutrition and exercise on PTSD. Fourth, our study had a relatively small size (144 patients) and a short follow-up time. Therefore, future largescale prospective study with longer follow-up time is needed. Finally, our study sample was from a hospital in Tianjin, China, potentially limiting generalizability to hospital settings elsewhere.
In conclusion, the present study showed that the incidence of PTSD was not rare among patients with confirmed COVID-19 infection.
We also found that older age, female sex, current smoking, and ≥3 involved pulmonary lobes were associated with increased risk of PTSD in COVID-19 patients, whereas hypertension and high hemoglobin level were inversely associated with risk of PTSD in COVID-19 patients.

ACKNOWLEDGMENTS
We thank all the people who have participated in this study.