Outbreak of COVID-19 in a nursing home in Madrid

Dear Editor,

We read with interest the study by Wang and colleagues recently reporting a high proportion of severe to critical cases associated to a high mortality in elderly hospitalized patients with COVID-19, what is in line with other reports.1, 2, 3, 4, 5, 6, 7 In nursing homes it is of paramount importance to know the situation of the residents and staff members, which would allow health care workers and surviving residents to be presumed as "protected" or "exposed” to the disease. There are only a few COVID-19 outbreaks reported in nursing homes.8, 9, 10 We evaluated the status against SARS-CoV-2 of people either residing or working at a privately run nursing home located at Madrid area (Spain) that was severely affected by an outbreak of COVID-19.

The 94 residents and staff members who consented to participate in the study were sampled on the 18th of April 2020 for nasopharynx PCR determination (GeneXpert^Ⓡ; Xpress SARS-CoV-2, Cepheid) and for finger stick whole blood and venepuncture IgG/IgM antibodies detection (All Test, Hangzhou All Test Biotech Co., LtD; Hangzhou, China). COVID-19 cases were proven (a patient with signs and symptoms and PCR-positive) or probable (a patient with signs and symptoms in the absence of PCR results). Clinical situation at the time of the study was active infection (PCR-positive), past infection (presence of antibodies in PCR-negative participants), and naïve “susceptible” population (no previous history of COVID-19 in both PCR- and antibody-negative participants). The qualitative variables are presented with their frequency distribution and the quantitative variables in mean and standard deviation or median and interquartile range in case of asymmetry. Categorical variables were compared using the Chi-square test or Fisher's test. In the case of quantitative variables, non-parametric methods were used (median test). The statistical significance was established at p<0.05. For the statistical analysis, the software SPSS Statistics for Windows, Version 21.0 (IBM Corp, Armonk, NY, USA) was used. The study was approved by the Hospital Ethics Committee (MICRO.HGUGM.2020–019).

The 84-available-bed facility had 79 beds occupied at the beginning of March 2020. The first case occurred on the 15th of March and preceded the additional 26 residents who died (34%) in the forthcoming 15 days what shrank the nursing home population to 52 survivors. All 27 (12 proven and 15 probable COVID-19 cases, respectively) residents presented with diarrhea and progressed to rapid deterioration with respiratory failure, shock, and death. Two residents died of other reasons. The clinical situation of the survivors in the prior month was no evidence of disease in 20 (40%), probable COVID-19 in 21 (42%), or proven COVID-19 in 9 (18%) who required hospital admission. Six staff members had proven COVID-19 (the PCR-positive result dated back on the 3rd of March in one of them) and 11 had probable disease. Twenty out of the 44 staff workers had been on sick leave due to COVID-19 in the last month.

On the day of the study, none of the 50 survivors was acutely ill (Table 1 ). Virtually all residents had at least one underlying condition and a median Charlson comorbidity index of 7 (IQR 5–8). Only one (2%) resident could be considered strictly immunocompromissed. Functional self-sufficiency measured by the Barthel index was a median of 35 (IQR 10 and 75). Out of the 50 residents, 30 (60%) were still PCR-positive and had detectable antibodies in serum samples (Table 2 ). Sixteen out of the 20 (80%) PCR-negative residents were seropositive. Thus, 46/50 (92%) residents had data suggesting active or past disease. Accepting a potential universal exposure dated between the 15th and 22nd of March, all residents had a presumed time period of contact with the disease of more than three weeks. In the case of the 44 staff members, eight were men, and had ages ranging from 37 to 51 (median of 43); none of them had relevant underlying diseases. At the time of the study, five were PCR-positive (11.4%); 21 were found to be seropositive (45.4%) including the five PCR-positive cases. Of the 94 participants, 32 (34%) serum samples were IgM-positive and all but one were also IgG-positive; 14 patients (43.7%) were PCR-positive. In contrast, PCR was positive in 20 (32.25%) out of the 62 IgM-negative patients (P = 0.18). In the 66 IgG-positive participants, 35 were PCR-positive (53%) while of the 28 IgG-negative participants all were PCR-negative (P<0.001). When the performance of the different serological techniques was compared to establish the criterion of seropositivity, the determination was positive in serum samples in 67/94 (71.3%) and in finger stick in 60/94 (63.8%). Concordance between finger stick and venepuncture samples was high though performance of the test was better when venepuncture samples were tested (Table 2).

Table 1.

Comparison of PCR-positive and PCR-negative residents.

Residents	Total N = 50	PCR + N = 30	PCR - N = 20	P
Median age in years (IQR)	87.0 (81.7–91.0)	88.0 (82.7–92.2)	86.5 (81.0–91.0)	0.34
Sex (%)
Male	13 (26.0)	8 (26.7)	5 (25.0)	1.00
Female	37 (74.0)	22 (73.3)	15 (75.0)
Underlying conditions (%)
Myocardial infarction	2 (4.0)	0 (0.0)	2 (10.0)	0.15
Congestive heart failure	8 (16.0)	5 (16.6)	3 (15.0)	1.00
Central nervous system disease	15 (30.0)	8 (26.7)	7 (35.0)	0.54
Chronic obstructive pulmonary disease	7 (14.0)	3 (10.0)	4 (20.0)	0.41
Renal dysfunction	3 (6.0)	3 (10.0)	0 (0.0)	0.26
Diabetes mellitus	17 (34.0)	10 (33.3)	7 (35.0)	1.00
Peptic ulcer disease	14 (28.0)	8 (26.6)	6 (30.0)	1.00
Neoplastic disease	16 (32.0)	6 (20.0)	10 (50.0)	0.03
Dementia	34 (68.0)	21 (70.0)	13 (65.0)	0.76
Charlson, median (IQR)	7 (5.0–8.0)	6.0 (5.0–7.2)	7.0 (5.0–8.0)	0.30

Table 2.

PCR and serum determination results of samples taken from residents and staff members of the nursing home.

People sampled	Positive PCR	Total positive antibodies IgG/IgM (serum)	Total positive antibodies IgG/IgM (Fingerstick)	IgG positive		IgM positive
				Serum	Fingerstick	Serum	Fingerstick
Staff (n = 44)	5	21	17	20	17	9	3
Residents (n = 50)	30	46	43	46	43	23	7
Total	35	67	60	66	60	32	10

We classified the residents in three groups: 30 (60%) residents who still had detectable viral RNA and, therefore, may be "potential" transmitters, 16 (32%) non-excreting but seropositive residents who could probably already lead freedom of movement, and four (8%) naïve susceptible residents at risk of acquiring COVID-19 who should be especially protected. In fact, the four naïve susceptible residents had unlimited mobility and two of them shared a room with PCR-positive patients. Our study highlights the extraordinary risk of lethal spread of SARS-CoV-2 infection in nursing homes, the very rapid transmission of the infection among residents and the high degree of infection in staff members. The presence of IgG antibodies with simultaneous PCR data determination poses a model of classification of residents and staff that allows for organizational decisions.

Disclaimer

None of the authors declare any conflict of interest for the development of this work

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Senex-COVID-19 study group

Adán (Iván), Adán (Javier), Alcalá (Luis), Alonso (Roberto), Álvarez-Uría (Ana), Bouza (Emilio), Catalán (Pilar), Escribano (Pilar), Estévez (Agustín), Fernández-Del Rey (Rocío), Galar (Alicia), García de Viedma (Darío), González (Pedro), Gómez-Núñez (Ana), Guinea (Jesús), Herránz (Marta); Kestler (Marta), Machado (Marina), Moure (Zaira), Marín (Mercedes), Muñoz (Patricia), Olmedo (María), Ortiz (Javier), Palomo (María), Pérez-Granda (María Jesús), Pérez (Laura), Pescador (Paula), Rincón (Cristina), Rodríguez (Sara), Rodríguez (Belén), Ruiz-Serrano (María Jesús), Serra (José Antonio), Valerio (Maricela), Vidán (Maite).