Letter to the Editor: Variability but not admission or trends in NEWS2 score predicts clinical outcome in elderly hospitalised patients with COVID-19.

Dear Editor,

In a recent article in the Journal, Bruno and colleagues present short-term outcomes in elderly patients with severe COVID-19 disease admitted to a single Italian Infectious Disease unit. The study found that elderly patients are at increased risk of adverse outcomes due to high number of comorbidities and emphasises the need to improve clinical management in these patients. In particular, elderly patients who are likely to deteriorate will need to be rapidly identified. Existing prognostic models for COVID-19 based on clinical, laboratory and radiological variables are at high risk of bias. In the UK, the National Early Warning Score (NEWS) and its updated version NEWS2 – an a priori weighted composition of the patient's observations - is used routinely to monitor patients in hospital and identify early those who may deteriorate.

Compared to other early warning scores, the NEWS Score has a greater ability to discriminate patients at risk of cardiac arrest, unanticipated intensive care unit admission or death. Currently, guidance from the National Institute of Clinical Excellence (NICE) supports the use of the NEWS2 score to risk assess patients with COVID-19 in the community, who may require hospitalisation.

In a recent rapid review, Greenhalgh and colleagues do not recommend using the NEWS2 Score alone for risk assessment of patients with COVID-19 in the community. Blood pressure and oxygen saturation measurements are difficult to take remotely. The score also does not include age or comorbidities, strong independent predictors of survival in patients with COVID-19. The value of the NEWS2 Score in predicting outcome in patients admitted to hospital with COVDI-19 remains uncertain.

We therefore undertook, as part of service evaluation, a prospective pilot assessment of patients with confirmed COVID-19 admitted to a tertiary infectious diseases unit, in the first month of the pandemic reaching the UK.

We studied all patients who had a clinical outcome (either discharged from hospital or died) between 12th March and 2nd April 2020. Clinical (presenting symptoms, comorbidity and the NEWS2 Score throughout hospital stay), laboratory (routine blood tests) and radiological (chest x-ray reports) findings on admission were collated. Our main aim was to examine the utility of the NEWS2 Score in predicting the clinical deterioration of hospitalised COVID-19 patients. Continuous data are expressed as a median (25th - 75th percentiles) and categorical data are expressed as n (%). Independent t-tests and Mann-Whitney U tests were used to compare two continuous variables for normally and non-normally distributed data. The chi-squared test was used to compare proportions between groups.

Overall, 17 patients with COVID-19 had an outcome by 2nd April 2020. The median age of our cohort was 85 years (IQR 83-88 years); 53% were male and 82% were Caucasian. All patients who were unsuitable for escalation to intensive care and received ward-based care as per NICE rapid guidance for COVID-19. The majority of patients died (N=10, 59%). Compared to patients who survived, those who died were more likely to be male, with bilateral consolidation on chest radiographs on admission. Admission SARS-CoV-2 quantitative PCR Ct values on nasopharyngeal swab did not seem to relate to survival. All patients who died required some form of oxygen therapy, ranging from nasal canulae to non-invasive ventilation through continuous positive airway pressure. Less than half of those who survived required oxygen therapy, all of which were delivered via nasal canulae.

Fig. 1 shows the trend in the National Early Warning Score2 (NEWS2) throughout hospitalisation, stratified by severity of NEWS2 on admission and clinical outcome. First, we found that the initial NEWS2 score did not predict mortality. For example, four out of the ten patients (40%) who died presented with a NEWS2 score of 0-3 while three out of seven patients (43%) who survived presented with a NEW2 Score of 5 or above. Secondly, there was no significant difference in the admission NEWS2 score and its components, between patients who died and those who survived, apart from systolic blood pressure. (Table 1 ). Thirdly, examining the NEWS2 scores over time, patients who died had a higher variability in their scores compared to those who survived. Seven out of ten patients (70%) who died had a maximum daily change in NEWS2 score of over 5, while none of those who survived had such dramatic fluctuations. (Fig. 1)

Fig. 1

NEWS2 scores of all patients with COVID-19, stratified by admission NEWS2 score.

Table 1

Baseline characteristics compared those who died vs those who survived.

	Died (N=10)	Survived (N=7)	P value (Died vs Survived)
Demographics
Age	86 (83-88)	83 (82-88)	0.59
Sex (male), N (%)	7 (70)	2 (29)	0.09
Ethnicity, N (%)			0.68
Caucasian	8 (80)	6 (86)
Asian Indian	1 (10)	1 (14)
Black	1 (10)	0
Number of comorbidities, N (%)
0	0	1 (14)	0.46
1-2	3 (30)	2 (29)
3+	7 (70)	4 (57)
Ischaemic heart disease, N (%)	1 (10)	0	0.39
Peripheral vascular disease, N (%)	1 (10)	2 (29)	0.32
Stroke, N (%)	2 (20)	1 (14)	0.76
Hypertension, N (%)	6 (60)	1 (14)	0.06
Atrial fibrillation, N (%)	1 (10)	2 (29)	0.32
Heart failure, N (%)	1 (10)	4 (57)	0.04
Diabetes, N (%)	2 (20)	1 (14)	0.76
Dementia, N (%)	2 (20)	1 (14)	0.76
Chronic kidney disease, N (%)	1 (10)	0	0.39
Cancer, N (%)	4 (40)	1 (14)	0.25
COPD, N (%)	2 (20)	3 (43)	0.31
Obstructive sleep apnoea, N (%)	2 (20)	0	0.21
Clinical frailty scale	6 (5-7)	6 (4-7)	0.65
Observations on admission
NEWS2	5 (0-8)	3 (2-6)	0.92
Respiratory rate (breaths/min)	21 (18-28)	24 (20-28)	0.69
O2 saturation (%)	96 (92-97)	96 (93-98)	0.96
FiO2 (%)	28 (21-100)	21 (21-21)	0.09
Systolic blood pressure (mmHg)	129 (124-133)	146 (141-157)	0.03
Diastolic blood pressure (mmHg)	74 (58-84)	81 (66-98)	0.26
Heart rate (beats per minute)	81 (74-95)	84 (79-86)	0.53
Temperature (degree celcius)	37.5 (36.8-38.5)	37.4 (36.8-38.1)	0.70
Presenting complaint
Fever, N (%)	8 (80)	2 (29)	0.03
Cough, N (%)	3 (30)	5 (71)	0.09
Breathlessness, N (%)	5 (50)	4 (57)	0.77
Diarrhoea, N (%)	0	3 (43)	0.02
Confusion, N (%)	2 (20)	2 (29)	0.68
Falls, N (%)	1 (10)	0	0.39
Blood tests on admission
White cell count (x109/L)	8.5 (3.7-15.9)	7.3 (2.3-9.1)	0.38
Neutrophils (x109/L)	6.8 (2.8-13.6)	5.1 (1.8-8.0)	0.33
Lymphocyte (x109/L)	0.99 (0.63-1.48)	0.63 (0.31-1.07)	0.14
Haemoglobin (g/L)	125 (113-142)	116 (102-119)	0.22
Platelets (x109/L)	173 (114-260)	206 (135-261)	0.38
C-reactive protein (mg/L)	65 (33-156)	31 (9-110)	0.24
Sodium (mmol/L)	133 (131-137)	136 (131-138)	0.46
Potassium (mmol/L)	4.7 (4.2-5.1)	3.9 (3.4-4.4)	0.03
Urea (mmol/L)	9.2 (7.0-15.9)	7.2 (3.0-10.5)	0.28
Creatinine (µmol/L)	122 (87-168)	91 (57-119)	0.12
Chest radiographs (CXR - findings on admission)
Total number during admission	2 (1-3)	1 (1-1)	0.12
Clear	3 (30)	4 (57)	0.007
Unilateral consolidation	0	3 (43)
Bilateral consolidation	7 (70)	0
Ct Value (nasopharyngeal swab)	22.51 (17.96-27.46)	25.12 (18.54-31.04)	0.40
Treatment
Oxygen therapy, N (%)	10 (100)	3 (43)	0.006
Nasal canulae, N (%)	9 (90)	3 (43)	0.04
Non-rebreathe mask, N (%)	9 (90)	0	<0.001
High-flow oxygen, N (%)	3 (30)	0	0.11
Non-invasive ventilation, N (%)	2 (20)	0	0.21
Length of stay (days)	6 (4-13)	9 (7-10)	0.13

In our small pilot of elderly patients admitted to hospital with COVID-19, admission NEWS2 scores did not seem to be useful in predicting clinical outcomes. For some patients, death occurred regardless of admission NEWS scores and without a prior deteriorating trend. Originally, the NEWS score was developed using data from 35585 acute hospital admissions, most of whom would have had an underlying diagnosis of sepsis. Sepsis is a clinical syndrome caused by overwhelming systemic bacterial infection. Clinical deterioration is seen within days in hospital. However, COVID-19 is caused by SARS-CoV-2, a coronavirus which predominately appears to affect the respiratory system as an initial viral pneumonitis. In China, a fifth of all COVID-19 inpatients rapidly became critically ill with hypoxia and respiratory failure. The weighting of the NEWS2 score does not account for the degree of supplemental oxygen (FiO2) a patient may require, thus limiting its utility to identify early deterioration in patients with COVID-19. In our cohort, patient 8 had a NEWS2 score of 2 on day 2 and 3 despite requiring a large increase in FiO2 (from room air to 60%). A more sensitive early warning score for COVID-19 needs to be urgently developed and validated.