Fighting COVID-19 spread among nursing home residents even in absence of molecular diagnosis: a retrospective cohort study

Alessio Strazzulla (  alessiostrazzulla@yahoo.it ) Centre Hospitalier de Melun https://orcid.org/0000-0003-0199-3449 Paul Tarteret Centre Hospitalier de Melun Maria Concetta Postorino Centre Hospitalier de Melun Marie Picque Centre Hospitalier de Melun Astrid de Pontfarcy Centre Hospitalier de Melun Nicolas Vignier Centre Hospitalier de Melun Catherine Chakvetadze Centre Hospitalier de Melun Coralie Noel Centre Hospitalier de Melun Cecile Drouin Centre Hospitalier de Melun Zine Eddine Benguerdi Centre Hospitalier de Melun Sylvain Diamantis Centre Hospitalier de Melun


Introduction
Since coronavirus disease 2019 (COVID- 19) worldwide epidemic begun in China at the end of 2019, it caused 9,653,048 new diagnosis and 491,128 deaths by June 27th, 2020 [1]. Among them, 162,936 cases and 29,778 deaths occurred in France [2]. COVID-19 diagnosis is based on the association of symptoms and epidemiology criteria (probable COVID-19 case) with molecular tests (con rmed COVID-19 case) [3]. However, the use of molecular tests is limited by cost, lack of polymerase chain reaction (PCR) detection kits and need of well trained health personnel for correct specimen collection. In particular, the dramatic case escalation in many countries made impossible to perform severe acute respiratory syndrome coronavirus 2 (SARS-Cov2) PCR in any suspected cases. These problems affected high income as well as low income countries. Moreover, serology tests cannot substitute PCR because SARS-Cov2 antibodies are produced over days to weeks after infection and consequently they are recommended only for epidemiologic purposes [4].
The median age at death for COVID-19 was 82 and 80 years in Spain and Italy, respectively [5,6]. It suggested that preventing measures needed to be reinforced in places with high concentration of seniors, namely hospitals and nursing homes [7].
Unfortunately, many outbreaks in nursing homes were reported with disastrous consequences for their residents. In the United States during a twenty-one-daylong period in March 2020, COVID-19 outbreak occurred in two nursing homes at King and Snohomish Counties (Washington) and it was responsible of 101/118 (85.5%) con rmed cases and 34/101 (33.7.2%) deaths among residents [8]. In France 37,995 con rmed or possible cases and 10,488/37,995 (27.6%) deaths among nursing home residents occurred by June 26th, 2020 [2].
On March 13th, 2020, extraordinary measures were applied in a nursing home in the Ile de France region in France, including: obligation of protective mask use by health care personnel, prohibition of visit by patients' families, obligation of in-room meal consumption, and training of health personnel for self-recognition of COVID-19 symptoms. These measures were enforced with the purpose of reducing the risk of contagious from visitors and minimize the risk of intramural transmission. On March 16th, 2020, a member of the health care personnel was tested positive to SARS-Cov2 PCR after three days of COVID-19 symptoms. This event produced the activation of the inner surveillance system and a quick epidemiologic inquiry was conducted. It individuated 21 close contacts, including four symptomatic nursing home residents who were quickly isolated with implementation of contact and air precautions. Only one patient was tested positive while other three patients did not receive test. This choice was motivated by the necessity of SARS-Cov2 PCR sparing because of lack of reagents in the entire French territory and the consequent prioritisation for in-hospital diagnosis. All other contact cases, including health care personnel, were isolated at home. The rst three consecutive symptomatic members of health care personnel were tested positive to SARS-Cov2.
This study describes the evolution of contagion during the COVID-19 outbreak in a French nursing home and all measures enforced to limit COVID-19 morbidity in a context of lack of molecular tests. Moreover, residents' clinical characteristics were compared to individuate possible risk factors for COVID-19 contagion and death.

Materials And Methods
A monocentric prospective study was conducted. The study was settled in a public nursing home located in the Ile de France region in France, accounting for 66 beds. The entire population of residents was followed from March 17th, 2020 to June 11th, 2020 after a member of the health care personnel had been tested positive to SARS-Cov2 from nasopharyngeal swab on March 16th, 2020 (positive result received on March 17th, 2020).
The study was conducted in accordance with Declaration of Helsinki and national and institutional standards. According with the current French legislation, approval by the local ethic committee was not demanded because a non-interventional research was conducted; meaning that all procedures and drugs administrated did not differed from daily clinical practice [9,10]. Patients were informed about the inclusion in a non-interventional retrospective study. No patient or legal representative expressed opposition for the inclusion in this study.
After the rst con rmed COVID-19 case among health care personnel, the rst three consecutive symptomatic members of health care personnel were tested and they resulted positive to SARS-Cov2. These results justi ed the prosecution of the extraordinary measures, the surveillance for the rest of nursing home population (residents and health care personnel) and the reinforcement of the nursing staff. Subsequently, test for SARS-Cov2 was not systemically proposed to members of health care personnel and SARS-Cov2 molecular test was only administrated according to medical decision case by case. Probable COVID-19 cases among health care personnel were systemically con ned at home.
During the time-frame analysed, all residents who developed COVID-19 symptoms after June 17th, 2020 were classi ed as "suspected cases" and they were isolated with implementation of contact and air precautions and room con nement, according with recommendations of the Health Regional Agency (Agence Régionale de Santé ou ARS) [11]. All suspected cases were tested negative for in uenza virus infection by rapid u test. Only one suspected case was tested for COVID-19 by nasopharyngeal swab and classi ed as "con rmed case" after positive result of SARS-Cov2 PCR, according with International de nition [3]. The patient was randomly selected among the rst four suspected cases with the purpose of identifying a possible cluster of COVID-19 and justifying the continuation of isolation for suspected cases and surveillance for the rest of nursing home population (residents and health care personnel). No other molecular tests were performed among nursing home residents because of the necessity of SARS-Cov2 PCR sparing as a consequence of the lack of reagents in the entire French territory and prioritisation for in-hospital diagnosis. Once available, serology test was performed retrospectively in the entire resident population at the end of the study. Serology test were performed from June 8th, 2020 to June 11th, 2020 using the Roche® SARS-CoV-2 assay [12].
All COVID-19 suspected cases were re-classi ed as "probable cases" according with the French survey protocol among nursing homes. This protocol de nes probable case not only in presence of acute respiratory illness or fever (and no other aetiology fully explains the clinical presentation) but also in presence of any other symptom compatible with COVID-19 (and no other aetiology fully explains the clinical presentation). According with this protocol all probable cases were signalled to ARS [13]. The following signs and symptoms were used for the de nition of probable case: fever, severe respiratory distress requiring oxygen therapy, cough, conjunctivitis, nasal congestion, abnormal pulmonary noise at chest examination, diarrhoea, abdominal pain, fatigue, fall, neurologic troubles, lymphopeniain the complete blood count [14][15][16][17]. Fall and neurologic troubles were included among symptoms because they could reveal a marked asthenia. Sensitivity and sensibility of diagnosis based only on symptoms were calculated using serology as reference.
At the end of the study period, patients were distributed in groups according to the following criteria: i) presence of COVID-19 symptoms (symptomatic probable COVID-19 case group and asymptomatic resident group); ii) SARS-Cov2 serology results (seropositive resident group and seronegative resident group).
Residents' les were analysed to collect the following personal parameters and co-morbidities: age, gender, grade of autonomy, not normal nutritional status (obesity or underweight), 2019-2020 seasonal u vaccination, arterial hypertension, heart failure, atrial brillation, ischemic heart disease, thromboembolic venous disease, chronic obstructive pulmonary disease, solid or hematologic neoplasia, chronic kidney disease and diabetes. For the grade of autonomy, the autonomy, gerontorology, iso-resources group (autonomie, gerontology groupe iso resources or AGGIR) panel was used because it is the actual legal instrument for evaluating autonomy in aged people in France. According to this panel, the minimal AGGIR grade 1 is given to elderly patients who are con ned at bed and needing a continuous surveillance while the maximal AGGIR grade 6 is given to patients with preserved autonomy [18].
Chronic kidney disease was de ned for estimated glomerular ltration rate < 60 ml/min [19]. For nutritional status, obesity and underweight were de ned for values of body mass index (BMI) > 24.9 and BMI < 18.5, respectively [20]. Also, data about the following co-medications were collected: loop diuretics, thiazide diuretics, α-blocker antihypertensive drugs, calcium channel blockers, angiotensin converting enzyme (ACE) inhibitors, angiotensin II receptor blockers, β-blockers, insulin, metformin and other anti-diabetic drugs.
The following analysis were performed: i) chronologic description of new probable cases and deaths from March 17th, 2020 to June 11th, 2020 among residents and health care personnel; ii) sensitivity and speci city calculation for diagnosis based only on symptoms; iii) description of COVID-19 symptoms in probable cases among residents; iv) between-group (probable COVID-19 cases among residents vs asymptomatic residents) comparison of personal parameters, co-morbidities and co-medications; v) betweengroup (seropositive resident group vs seronegative resident group) comparison of personal parameters, co-morbidities and comedications.
The following statistic tests were performed: χ 2 -test for qualitative variables and Student's T-test for quantitative variables. Quantitative variables were presented in the text as mean values. R, the language for statistical computing (Vienna, Austria, https://www.r-project.org/) was used to perform all statistical analysis. Nominal statistical signi cance was set at p < 0.050.

Discussion
This study showed that COVID-19 outbreak in nursing homes can be managed even in absence of molecular diagnosis. Indeed, early detection of respiratory and not respiratory symptoms and the consequent preventive measures achieved in limit contagion at 58% and deaths at 6%. Besides, the between group analysis showed that that COVID-19 was associated with older age and heart diseases (atrial brillation).
Death rates in the entire resident population (6%) and among probable COVID-19 cases (9%) were low when compared with death rates among nursing home residents in France (27% among COVID-19 con rmed or probable cases) [2]. Moreover, the rate of symptomatic patients was limited at 49% and prevalence of seropositive patients maximized at 58%. We can infer that the quick isolation of probable COVID-19 cases limited the spread of SARS-Cov2 infection among other residents.
In our study, the most frequent symptom was fever (46.8%) and only 25% of symptomatic residents developed cough and 12.5% required oxygen support. These data showed that COVID-19 is not only a respiratory disease but rather a "systemic" disease and that COVID-19 patients are more susceptible to develop not speci c symptoms rather than a respiratory disease, at least in the early symptomatic period [17]. It is critically important to consider not respiratory COVID-19 symptoms with the purpose of ensuring quick health care of ill patients and contagion containment. In particular, digestive symptoms should be monitored because patients with presence of SARS-Cov2 in faeces can be infectious and diarrhoea (18.8% in our study) can facilitate the spread of the infection [16].
We applied the French strategy for limiting COVID-19 outbreak in nursing homes. This strategy had two main points: i) enforcement of extraordinary measures (obligation of protective mask use by all health care personnel, prohibition of visit by patient families, obligation of in-room meal consumption) with the purpose of limiting contagion from health care personnel and asymptomatic patients or visitors; ii) de nition of probable COVID-19 cases not only in presence of fever and respiratory distress but also in presence of not respiratory COVID-19 symptoms. The consequent early isolation of symptomatic patients balances the loss of microbiology diagnosis due to the SARS-Cov2 PCR limiting politics currently applied in French nursing homes [13]. Notwithstanding, it persists the problem of the asymptomatic SARS-Cov2 carriers (health care personnel and residents) who is impossible to punctually detect even in case of extended use of SARS-Cov2 PCR [21]. In this context, the use of SARS-Cov2 serology can be useful only for retrospective epidemiological analysis [22].
Early detection of possible COVID-19 patients considering respiratory and non respiratory symptoms had modest sensitivity and speci city (65% and 78%, respectively) when compared to serology. However, it is di cult to extrapolate our results to other settings because COVID-19 prevalence of disease can signi cantly change according to geographical area and time period.
In this study, COVID-19 was associated with increased age and heart diseases (atrial brillation). These data are in line with the current knowledge of SARS-Cov2 infection [23,24]. The lower prevalence of ACE inhibitor intake among COVID-19 seropositive patients does not support the hypothesis that these molecules could make patients more susceptible to SARS-Cov2 infection. However, literature data still remains controversial [25,26].
It could be argued that some patients with in uenza virus infection could have been classi ed as probable COVID-19 cases because of a bias in our classi cation of COVID-19 cases. However, this hypothesis is unlikely because: i) rates of seasonal u vaccination were high and not signi cantly different at the between-group analysis; ii) all probable COVID-19 cases were tested negative by rapid u test by well trained health personnel at the beginning of symptoms [27]; iii) during the study period, the u epidemic in France was clearly reducing its rate of new contagion and, for this reason, the Ile de France region was declared in "post-epidemic" phase early in March [28,29]; the retrospective SARS-Cov2 serology analysis con rmed that a COVID-19 outbreak occurred during the study period.
This study presents several limitations: i) its monocentric design limits the comparison with studies performed in other settings; ii) the small population size limited statistical analysis; iii) the use of serology to con rm the diagnosis of COVID-19 was inaccurate because current serology tests are not the gold standard and they have a not optimal sensitivity (60-90%) [30]; iv) absence of systematic administration of nasopharyngeal swab for SARS-Cov2 PCR prevented microbiologic diagnosis. However, SARS-Cov2 PCR has a high false negative rate, especially in the elderly, and for this reason adjunctive criteria are needed to diagnose COVID-19, such as radiological and clinical criteria [31][32][33]. Indeed, purpose of the study was to show the effectiveness of a strategy based on diagnosis of probable COVID-19 cases only by the detection of respiratory and not respiratory symptoms and the consequent enforcement of preventing measures and standard of care.

Conclusions
The COVID-19 outbreak in nursing homes can be managed even in absence of molecular diagnosis. Indeed, in this study contagion were limited at 58% and deaths at 6%. These results are satisfactory when the worldwide health crisis in nursing homes is considered. The early diagnosis of probable COVID-19 cases based on recognition of respiratory and not respiratory symptoms (in both health care personnel and residents) was the key factor for the containment of SARS-Cov2 outbreak.
In case of lack of molecular tests, we propose a strategy of systematic con nement and early diagnosis based on the early detection of respiratory and not-respiratory symptoms. This strategy can reduce mortality and make COVID-19 a controllable disease rather than an inevitable doom for nursing home residents, even in absence of molecular diagnosis. According with the current French legislation, approval by the local ethic committee was not demanded because a non interventional research was conducted, meaning that all procedures and drugs administrated did not differed from daily clinical practice. A written consent form was not proposed to patients because the non interventional nature of the study required only the absence of patients' opposition, according to the French law. No patients or legal representative expressed opposition for the inclusion in this study.

Funding sources
This research did not receive any funding from public, commercial or not-for-pro t organisations.
Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interest
The authors declare that they have no competing interests Figure 1 Evolution of new probable COVID-19 cases and deaths among health care personnel and residents from March 17th, 2020 to June 11th, 2020. NOTE: i) among health care personnel, 6/19 (31.5%) probable COVID-19 cases were tested for COVID-19 by nasopharyngeal swab and classi ed as "con rmed cases" after positive result of SARS-Cov2 PCR; ii) all probable COVID-19 cases were tested negative for in uenza virus infection by rapid u test; iii) two patients developed COVID-19 symptoms early in March (March 2nd and 3rd, respectively) before the start of extraordinary measures and they were retrospectively classi ed as probable cases; iv) among residents only one probable case was tested for COVID-19 by nasopharyngeal swab and classi ed as "con rmed case" after positive result of SARS-Cov2 PCR; v) among dead residents,1/4 (25%) was asymptomatic.