Sociodemographic and Clinical Features of COVID-19 Reinfection Cases Among Healthcare Workers in Turkey

Background: Global pandemic of novel Coronavirus Disease (SARS- COV-2) has spread across all continents and infected almost 80 million people. Since it is a novel disease, unknowns about the disease characteristic, treatment and length of immunity still persist. This study aims to characterize reinfection, personal protective equipment use and disease progress in healthcare workers in İstanbul. Methods: 23 healthcare workers who had conrmed negative PCR results after infection and another positivity later were questioned about both infection progress, their symptoms and treatment through an online questionnaire. Results: While the symptoms during both courses did not change drastically, 73.9% were treated as outpatient during the rst infection while all but one (95.7%) were treated as such during second time around. Median time between two infections were 106 days. All participants were cleared of disease and none had to be treated in intensive care unit. Conclusion: Use of personal protective equipment was found subpar compared to World Health Organization recommendations. This is the rst study from Turkey characterizing reinfected cases in healthcare workers. the swab methodology plays a role in diagnosing reinfections of COVID-19 and also the sensitivity and specicity varies on the swab collector: While in nasopharyngeal samples rates were 7.3%, it went up to 21.4% in oropharyngeal and fecal samples. [11] In our study all swab samples were nasopharyngeal for PCR testing. In a meta-analysis study, the required time elapsed between rst and second positive results are accepted as 35.4 days [13]. However, there is evidence that shedding of SARS-COV-2 RNA last until 53,65 or 83 days. our minimum two


Introduction
COVID-19 pandemic has become a global concern with almost 80 million cases around the World and more than 1.6 million fatalities almost a year after the announcement of the rst case [1]. As it is a novel disease, the clinical and pathological characteristics of the disease are still uncertain and constantly evolving with more studies being undertaken. [2,3] While the diagnosis and symptoms are becoming more apparent, there are still many unknowns such as treatment and immunization efforts. Other characteristics of the disease, possibility of reinfection, although thought not likely in the beginning has changed. Scienti c community is receiving evidence more and more on not just the possibility of reinfection but also mortality and morbidity associated with a second course of disease. Yet, international consensus on de nition and diagnosis of reinfection is still unclear. How long the acquired immunity due to past infection is unknown and it is hypothesized that mild course of infection may cause decrease in antibody levels and lead to reinfection. [4,5] In addition, PCR tests that utilize nasopharyngeal swab samples collection procedures vary and might lead to misdiagnosis, due to sensitivity and speci city of test kit, collection methods or collectors experience issues ,and to false negative results. [6][7][8] For these reasons, it is unsure whether the the positive PCR results of the patients are genuine reinfection or viral shedding of RNA fragments. In addition, there are evidence in the literature showing people of older age or with comorbidities might be reinfected with different strains of Corona Virus, even after two negative PCR results. [9] Our study aims to describe the progression of disease and characteristics of disease inhealthcare workers who are reinfected in İstanbul.

De nitions
Con rmed case: Turkey uses WHO's case de nition of at least one PCR positivity for COVID-19. Cases that have a clinical presentation or radiological signs of COVID-19 are not accepted as con rmed cases and are not included in our study population.
Reinfected case: For our study, reinfected COVID-19 case was de ned as having two consecutive PCR negative results after having a PCR test con rmed COVID-19 infection.

Data Collection
Data was collected from a central COVID-19 database of the country. Study population was healthcare workers who were tested with PCR between August 1,2020 and September 11, 2020 who also had a positive PCR result for COVID-19 in the last six months. Healthcare workers who did not have 2 consecutive negative PCR results after the rst positive result were eliminated. All of the participants were contacted by phone and asked for their consent. Questionnaires were distributed online, through various mediums. The questionnaire included sociodemographic information (age, sex, profession, comorbidities, medication use), rst and second infection durations, progression, severity, symptoms, hospitalization status and contact with COVID-19 patients. In addition, participants were asked questions about the treatment they received, when they went back to work, possible factors that could have affected transmission and if they cared for COVID-19 patients, the interventions (including aerosol producing intervention) and their contact time, place and protection tools used caring for these patients. The questions about frequency of use and practice of hygiene and personal protection were also included.

Ethical Approval
This study is approved by Üsküdar University Non-Interventional Research Studies Ethical Review Board with decree number 61351342/2020-475. All participants who consented to take part were shared the online questionnaire.

Statistical Analysis
Continuous variables are presented as median, maximum and minimum while categorical variables are in absolute frequency (n) relative frequency (%). McNemar's test was used to analyze the difference between rst and reinfection characteristics for dependant variables and signi cance level was set at p < 0.005 with %80 power calculation. All analysis are done in SPSS version 22.0. Findings 31 participants were eligible for the study. Three of the participants were unable to be reached and one participant lacked the technical infrastructure to ll the questionnaire online. 23 participants out of 27 accepted to participate and lled the questionnaire (85%).
26.1% (n=6) were working in clinical departments, 17.4% (n=4) were in COVID-19 ward, 13.0% (n=3) were in emergency unit and 13.0% (n=3) were in intensive care unit. Three participants were employees of municipal health directorate, one was from a primary care center, one from surgical ward and one from dental practice.
34.8% (n=8) had a chronic disease (Hypertension, COPD, Diabetes Mellitus or Hepatitis B) and all participants but one with a chronic disease reported using at least one drug for their condition. Smoking prevalence was 43.5% (n=10).
One participant took Plaquenil after contact with a patient. 57.1% of participants (n=12) received Hydroxychloroquine as treatment (Table 1), while two of them did not receive any treatment. 63.6% (n= 13) of treatment receivers started their treatment in the rst two days of diagnosis. 39.1% (n=9) of participants went back to work on 14 th day of diagnosis (min.10 -max.25).
Participants were asked about risk of transmission due to attending to COVID-19 patients. 26.1% (n= 6) treated COVID-19 patients, 39.1% (n=9) had contact face to face with a patient, 26.1% (n= 6) applied aerosol producing intervention to COVID-19 patients and 39.1% (n=9) spent time in COVID-19 patient treatment units ( Table 2). 52.2% of participants (n= 12) indicated that their patients were using a disposable mask during contact and 65.2% (n=15) of the participants used appropriate personal protective equipment (PPE) during contact. 94.1% (n=16) were using masks all the time before diagnosis, 64.7% (n= 11) were using surgical gloves and 17.6% (n=3) were using protective shield or glasses and 23.5% (n=4) were using surgical gown all the time ( Table 2).
76.5% (n=11) of the participants started treatment in two days after the diagnosis. 39.1% (n=6) went back to work on 14 th day (min.1 -max.30). 34.7% (n=8) of the participants used Favipiravir. Six of the participants did not receive any treatment.
56.5% of participants (n= 13) indicated that their patients were using a disposable mask during contact and 69.9% (n=16) of the participants used appropriate personal protective equipment (PPE) during contact. Among respondents for this question, 94.1% (n=16) were using masks all the time before diagnosis, 72.2% (n= 13) were using surgical gloves and 94.4% (n=17) were using protective shield or glasses and 33.6% (n=6) were using surgical gown all the time ( Table 2).

Discussion
De ning the criteria of reinfection is important to understand the disease progress and create epidemiological and clinical control and treatment protocols. In our study with 23 participants, the median time between two infections was 106.4 days. In a guideline report by European Center for Disease Control (ECDC) about the COVID-19 infection characteristics, false PCR positivity, or decreased sensitivity due to swab methodology, contamination due to transport or analysis are shown as possible factors for wrong diagnosis. In addition, it is possible to misdiagnose the length of time between two infections that is only due to suppressed immunity or low levels of antibody as a reinfection. Genome sequencing can alleviate this issue by providing information about phylogenetic differences in infecting agent. [10] In addition, many reinfection cases are associated with false negative PCR results, continuing viral shedding or ceasing the treatment due to no symptoms in patients leading to increased viral replication. [11] PCR diagnosis, still regarded as the cornerstone of COVID-19 diagnosis, has been used for all participants in our study to con rm the episodes. Only 6 participants stayed asymptomatic after the diagnosis of COVID-19 rst time while 4 participants stayed asymptomatic after the reinfection was con rmed. Most prevalent symptom during both courses were headache, fatigue and sore throat. Less commonly participants experienced chills, sweating and eye pain/ blurriness.
It is likely that the participants who were not experiencing any symptoms during testing for the rst time could have only tested due to contact with a high-risk patient. In addition, only half of asymptomatic participants developed any symptoms after diagnosis.
While 6 participants were treated as inpatients in the rst infection, most of them were treated as outpatients during reinfection, even though symptoms did not change between two episodes. It could be that mortality and morbidity was estimated higher due to COVID-19 being a emerging disease with a lot unknowns or using novel drugs that have higher e cacy. In addition, using any treatment during reinfection was lower. While these can be interpreted as less severe disease progression during reinfection, it is also possible that more we understand the disease, treatment and follow-up protocols change.
Majority of reinfection cases are diagnosed after being tested for having symptoms in the literature. In our study population, there were participants who were tested without any symptoms likely due to a contact with a high-risk patients or sporadic institutional screenings. For estimating the population level reinfections, testing of asymptomatic persons is important. [12] Studies show the swab methodology plays a role in diagnosing reinfections of COVID-19 and also the sensitivity and speci city varies on the swab collector: While in nasopharyngeal samples rates were 7.3%, it went up to 21.4% in oropharyngeal and fecal samples. [11] In our study all swab samples were nasopharyngeal for PCR testing.
In a meta-analysis study, the required time elapsed between rst and second positive results are accepted as 35.4 days [13]. However, there is evidence that shedding of SARS-COV-2 RNA can last until 53,65 or 83 days. [14,15 ] In our study, minimum time between two infections was 27 days while half of the patients were reinfected before day 102. Studies with larger sample sizes with phylogenetic and RNA analysis of SARS-COV-2 infections are needed to understand the underlying mechanism and discriminate between recurring infections, prolonged viral clearance and reinfections.
World Health organization recommends medical masks, surgical gloves, long sleeved gowns and eye protection (shields or glasses) for aerosol producing procedures for healthcare workers treating COVID-19 patients. [16]. The use of PPE's was less common during the rst infection than reinfection. It could mean that being infected lead to more caution in our group. However, use of disposable mask and surgical gloves stayed around 60% both before two episodes. WHO recommends use of risk appropriate PPE since 2014 with additional recommendations of N95 masks, long sleeved gowns, eye protection and surgical gloves for healthcare personnel working with COVID-19. [17,18] Using face masks are also shown to protect against infections both in healthcare workers and the population, while addition of eye protection provides extra protection. [19] Therefore, the recommendations also include non-healthcare personnel (cleaning, kitchen, security personnel) in the hospital to use the same protection protocols. [20] In our study usage of all protection equipment was below desired levels for all profession groups.
While participants were questioned about their contact with possible transmission, nature and length of these were not questioned, a risk strati cation therefore cannot be made. However, it is known that physical distancing of 2 meters is very effective in preventing transmission. [21] Half of the participants had clinical duties as doctor, nurse or technician which put them into closer contact with COVID-19 patients, while others likely only shorter and more distant contact as cleaning personnel or health o cer. Due to the nature of work, it is highly di cult for healthcare workers to maintain that distance.

Limitations
This study enrolled a small sample of healthcare workers, who are usually under higher risk than normal population for both COVID-19 and reinfection due to nature of their work. In addition, the questionnaire was done online, leading to multiple participants being unable to ll, decreasing the sample. Data collected were through a centralized platform, however no genome sequencing, viral culture or sub genomic RNA detection was done. Since the diagnostic criteria for reinfection are still limited, it is possible that some of the participants were having a recurrent or reactivated disease rather than a reinfection.

Conclusion
This study aims to describe disease characteristics and progression of 23 reinfected healthcare workers. Median time between two infections was 106.4 days. 73.9% of participants experienced symptoms during the rst infection while this rate went u to 82% during reinfection. In this relatively young population, it should also be noted that immune suppression and extended exposure due to the nature of the work and increasing workload due to pandemic could have played an important role. Our study population seems to have experienced rst infection more severely than the reinfection, considering almost all were treated as outpatient during their reinfection while less received any treatment. Even though the rates of reinfection in general population is unknown, increasing the capacity for testing, including asymptomatic persons could lead to increased rates. Genome testing is also vital to con rm whether cases are actual reinfections. While there are still much to learn about COVID-19 pandemic, hygiene, physical distancing and protective equipment are still the best measures for preventing cases around the world.

Declarations Funding
There was no funding for this study.

Con icts of Interest / Competing Interests
The authors declare that they have no con ict of interest.

Ethics approval
This study is approved by Üsküdar University Non-Interventional Research Studies Ethical Review Board with decree number 61351342/2020-475.

Consent to Participate
All participants were asked for oral consent to participate.

Consent for Publication
All participants were asked for oral consent to participate.

Availability of data and material
Data used in this study is not publicly available.
Code availability Not applicable. [2] Among respondents.  [i] Legend for Table 3 Hospitalization