Enteric involvement in hospitalised patients with COVID-19 outside Wuhan.

Early studies1, 2 of coronavirus disease 2019 (COVID-19) reported that the proportion of patients presenting with gastrointestinal symptoms was low. However, evidence for enteric involvement is emerging.3, 4, 5, 6 Recently, a multicentre study in Hubei province (China) indicated that gastrointestinal symptoms were common in patients with COVID-19. We examined the intestinal symptoms of patients with COVID-19 from multiple medical centres located in and outside of Hubei province.

We retrospectively analysed data from 232 patients who were positive for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) RNA admitted to 14 hospitals (two hospitals in Guangdong province, two in Hubei province, and ten in Jiangxi province) between Jan 19, 2020, and March 6, 2020. Most patients were admitted because of fever, cough, dyspnoea, and chest CT findings consistent with COVID-19 pneumonia. Diagnosis of COVID-19 was based on positive SARS-CoV-2 RNA tests. Two patients with pre-existing digestive diseases were excluded from our analysis. The analysis was approved by the institutional review boards of Sun Yat-sen University and the participating hospitals. Full details of the methods used are in the appendix (p 1).

The clinical and demographic characteristics of the 230 patients analysed are shown in the appendix (p 2). There were 129 men and 101 women; median age was 47·5 years (range 7–90). The most common symptoms at onset of illness were fever (193 [84%] patients), cough (159 [69%] patients), and sputum production (98 [43%] patients). Diarrhoea was observed in 49 (21%) patients. Patients with diarrhoea were older and were more likely to have comorbidities than patients without diarrhoea (table ). A greater proportion of patients admitted to hospital had diarrhoea as the outbreak progressed: nine (43%) of 21 patients admitted between Feb 12 and March 6, 2020, had diarrhoea versus 40 (19%) of 209 patients admitted between Jan 19 and Feb 11, 2020.

Table

Demographic and clinical characteristics of patients with COVID-19 with and without diarrhoea

		With diarrhoea (n=49)	Without diarrhoea (n=181)	p value
Age, years		55 (40–65)	46 (36–57)	0·017
Sex		..	..	0·87
	Male	27 (55%)	102 (56%)	..
	Female	22 (45%)	79 (44%)	..
Any comorbidity		19 (39%)	39 (22%)	0·017
Date of admission to hospital		..	..	0·022
	Jan 19–Feb 11, 2020	40 (82%)	169 (93%)	..
	Feb 12–March 6, 2020	9 (18%)	12 (7%)	..
Severe COVID-19 disease		26 (53%)	35 (19%)	<0·0001
Oxygen supplementation		44 (90%)	143 (79%)	0·10
Ventilatory support		6 (12%)	3 (2%)	0·0036
Intensive care		15 (31%)	20 (11%)	0·0015
Died		4 (8%)	2 (1%)	0·020
White blood cell count, ×109 cells per L		5·6 (2·0)	5·6 (3·0)	1·0
Lymphocyte count, ×109 cells per L		1·0 (0·6)	1·1 (0·5)	0·28
Neutrophil count, ×109 cells per L		3·4 (1·8)	3·9 (2·9)	0·34
Alanine aminotransferase, U/L		37·9 (27·4)	34·0 (24·4)	0·36
Aspartate aminotransferase, U/L		39·3 (27·0)	34·7 (17·9)	0·19
Total bilirubin, μmol/L		12·5 (7·3)	11·5 (5·9)	0·36
Activated partial thromboplastin time, s		30·5 (9·5)	31·0 (7·8)	0·74
D-dimer, mg/L		1·0 (1·8)	1·6 (5·4)	0·48
Procalcitonin, ng/mL		0·29 (0·69)	0·19 (0·36)	0·25
Erythrocyte sedimentation rate, mm/h		40·7 (30·0)	23·8 (18·7)	0·0002
C-reactive protein, mg/L		40·5 (52·0)	30·0 (38·3)	0·16
Antibiotics		36 (73%)	138 (76%)	0·71
Antiviral treatment		49 (100%)	180 (99%)	1·0

Data are median (IQR), n (%), or mean (SD). p values comparing patients with and without diarrhoea were calculated by use of Fisher's exact test, Mann–Whitney U test, or Student's t test. COVID-19=coronavirus disease 2019.

More patients with diarrhoea showed severe symptoms of pneumonia than those without diarrhoea, with an adjusted (for age, sex, and comorbidities) odds ratio (OR) of 4·95 (95% CI 2·14–11·70, p=0·0002). Although there was no correlation between diarrhoea and oxygen supplementation, a greater proportion of patients with diarrhoea required ventilator support (adjusted OR 6·52 [95% CI 1·44–35·86], p=0·018) and required admission to intensive care (3·19 [1·25–8·16], p=0·015). Most laboratory test results did not differ significantly between patients with and without diarrhoea (table).

A similar proportion of patients with and without diarrhoea were given antibiotics and antiviral medications (table). Less frequent gastrointestinal symptoms included bloody stool (ten [4%] of 230 patients) and abdominal pain (three [1%] of 230 patients). No patients presented with gastrointestinal symptoms alone, which might reflect the recommendation for patients with mild symptoms to stay at home.

Our observations of patients with COVID-19 from medical centres within and outside of Hubei province (appendix pp 3–5) are consistent with other reports3, 4, 5, 6, 7, 8, 9 showing gastrointestinal involvement in COVID-19. Many of these studies3, 4, 5, 6, 7, 8, 9 reported patients with COVID-19 who only had gastrointestinal symptoms during the course of the illness, and another study reported patients with COVID-19 who presented with diarrhoea 1–2 days before the development of fever and dyspnoea.

We noted correlations between the presence of diarrhoea and the severity of respiratory symptoms, and found that more patients with diarrhoea required ventilator support and were admitted to intensive care than those without diarrhoea. Similarly, during the SARS outbreak in 2003, more patients with SARS who had diarrhoea required ventilator support and were admitted to intensive care than those who did not have diarrhoea. Understanding the mechanism underlying these correlations could help to identify novel intervention strategies to reduce or prevent severe outcomes of COVID-19.

The proportion of patients with COVID-19 who had gastrointestinal symptoms in our analysis (21% with diarrhoea) is higher than that observed in earlier reports (2–3% with diarrhoea1, 11). Similarly, a greater proportion of patients admitted to hospital had diarrhoea as the outbreak progressed. This difference could be due to a heightened awareness of the gastrointestinal symptoms of COVID-19 among clinicians. Additionally, the passing of time could have allowed the virus to accumulate mutations that might have affected its tropism for the intestines. The difference could also reflect geographic differences between populations studied (ie, early studies from Wuhan vs our patients from four other cities). The increasing number of patients with COVID-19 who have gastrointestinal symptoms warrants further study.

A potential limitation of our analysis is the absence of data on the prevalence of asymptomatic COVID-19. It is likely that more asymptomatic patients existed, leading to an overestimation of the proportion of patients with gastrointestinal symptoms. Another possible confounding factor is that the antivirals, lopinavir and ritonavir, are known to cause diarrhoea. However, we did not observe a correlation between diarrhoea and use of antiviral medicine or antibiotics.