Gastrointestinal sequelae 90 days after discharge for COVID-19

Huang and colleagues recently reported that as many as 76% of patients discharged after hospitalisation for COVID-19 had at least one symptom persisting 6 months after disease onset,¹ including fatigue or muscle weakness (63%), sleep difficulties (26%), and anxiety or depression (23%). Additionally, more than 50% of the patients had abnormal chest CT images indicating impaired pulmonary function.

Although SARS-CoV-2 mainly affects the lungs, many other organs are also affected. Enteric symptoms are common in COVID-19, and gastrointestinal symptoms can be the only symptom, or can be present before respiratory symptoms.² The cellular receptor for SARS-CoV-2, ACE2, is highly expressed in the gut, and SARS-CoV-2 has been observed in the colonic tissue³ and faeces⁴ of patients with COVID-19. Therefore, we examined the long-term gastrointestinal sequalae of SARS-CoV-2 infection in patients who were admitted for COVID-19 to 12 hospitals in the Hubei and Guangdong provinces, China, between Jan 16 and March 7, 2020, and subsequently discharged (appendix pp 1–3).

117 patients with COVID-19 who had been discharged (53 [45%] of whom were aged 60 years or older) completed one return visit (usually 1 month after discharge) and a telephone interview around 90 days after discharge; their baseline characteristics are shown in the appendix (pp 4–5). The most common symptoms on admission were fever (79 [69%] of 114 patients) and cough (77 [66%] of 117 patients); 20 (17%) of 117 patients presented with dyspnoea. Gastrointestinal symptoms were recorded for 15 (13%) of 117 patients on admission and for 49 (42%) of 117 patients during hospitalisation. Median length of hospital stay was 19 days (IQR 16–23), during which most patients (102 [87%] of 117) required supplemental oxygen; 24 (22%) of 111 patients exhibited decreased blood oxygen saturation; 33 (28%) of 117 were severely ill; and 28 (24%) of 117 required admission to the intensive care unit.

Gastrointestinal sequelae were defined as gastrointestinal symptoms that presented after discharge but were not present within the month before onset of COVID-19. 52 (44%) of 117 patients reported gastrointestinal symptoms after discharge at the 90 day telephone interview, of whom 51 patients had gastrointestinal symptoms at 90 days after discharge, and one had gastrointestinal sequelae that had resolved by the 90 day follow-up. The most common gastrointestinal sequelae in 117 patients were loss of appetite (28 [24%] patients), nausea (21 [18%]), acid reflux (21 [18%]), and diarrhoea (17 [15%]); less common gastrointestinal sequelae included abdominal distension (16 [14%] patients), belching (12 [10%] patients), vomiting (11 [9%]), abdominal pain (eight [7%]), and bloody stools (two [2%]). None of the 65 patients without gastrointestinal sequelae at 90 days had gastrointestinal symptoms on admission or during hospitalisation. Of the 52 patients with gastrointestinal sequelae after discharge, 15 (29%) had gastrointestinal symptoms on admission and during hospitalisation, 34 (65%) had such symptoms during hospitalisation, and three (6%) had such symptoms only after discharge.

Patients with gastrointestinal sequelae at 90 days were similar in age, sex, body-mass index, and incidence of comorbidities to those without gastrointestinal sequelae, and had similar lengths of hospital stay (appendix pp 4–5). Blood test results on admission showed that white blood cell count, neutrophil count, and procalcitonin concentration were higher in patients with gastrointestinal sequelae at 90 days, although values in both groups were in the normal range (appendix p 6). C-reactive protein concentrations were higher in patients with gastrointestinal sequelae at 90 days than in those without gastrointestinal sequelae. 90 days after discharge, blood tests showed that alanine aminotransferase was higher in patients with gastrointestinal sequelae (appendix p 7). No other differences were noted; procalcitonin and C-reactive protein were not tested at 90 days.

Compared with patients without gastrointestinal sequelae at 90 days, patients with gastrointestinal sequelae more frequently presented with dyspnoea (23% vs 12%) and myalgia (17% vs 11%) on admission, although these differences were not significant. Patients with gastrointestinal sequelae were less frequently severely ill than were those without gastrointestinal sequelae (17% vs 37%; p=0·021), although after adjustment for confounding factors, this difference was not significant (p=0·051). Patients with gastrointestinal sequelae at 90 days exhibited a lower frequency of supplemental oxygen therapy (79% vs 94%; p=0·016), and a trend of lower frequency of intensive care unit admission during hospitalisation (appendix p 4). Patients with gastrointestinal sequelae at 90 days were treated more often with proton pump inhibitors (PPIs) and corticosteroids and were less frequently treated with enteral nutrition than were patients without such sequelae (appendix pp 4–5).

Gastrointestinal sequelae including loss of appetite, nausea, acid reflux, and diarrhoea are common in patients 3 months discharge from hospitalisation due to COVID-19. Persistent gastrointestinal symptoms have important implications for proper management of patients and health-care resources. Our data highlight the importance of gastrointestinal care and nutritional support for patients discharged post-COVID-19 hospitalisation.

That severe illness during hospitalisation was not associated with post-discharge gastrointestinal sequelae was unexpected. Decreased blood oxygen saturation, a symptom closely related to severe pneumonia, was found to be associated with gastrointestinal sequelae. This association might involve hypoxia mediating the multi-organ injuries frequently observed with COVID-19.⁵ It is important to note that hypoxia not only occurs in patients with COVID-19 with dyspnoea, but also in many patients without dyspnoea.⁶ Asymptomatic hypoxaemia could explain the apparent discrepancy in our observation that decreased blood oxygen saturation, but not severe illness, is associated with gastrointestinal sequelae.

PPIs were often used for aspiration prophylaxis in patients admitted for suspected COVID-19 receiving anaesthesia, parenteral nutrition, or other acid-related treatments. Acid rebound is a known consequence of deprescribing PPIs,⁷ so it might be expected that PPI treatment during hospitalisation was associated with acid reflux after discharge.

Nutritional interventions during hospitalisation appeared to be associated with a lower incidence of subsequent gastrointestinal sequelae. As many as 78% of patients hospitalised for COVID-19 report lack of appetite.⁸ In addition to lack of appetite, other common gastrointestinal symptoms such as nausea, vomiting, and diarrhoea during hospitalisation5, 8 can lead to malnutrition, which has been linked to increased mortality in patients with COVID-19.9, 10 Nutritional support could be critical in sustaining patients' lives. Reduced gastrointestinal sequelae is probably a part of the overall benefit from nutritional support.

Our study has several limitations. Being retrospective, data were missing for several relevant blood biochemical tests, such as blood markers for inflammation and serum titres of SARS-CoV-2. Another important limitation was that only 71% of discharged patients had sufficient data to be included in this analysis, which might lead to biased conclusions. Reliance on patient recall of symptoms during follow-up is another potential limitation. A prospective study will allow a better understanding of the longer-term gastrointestinal sequelae of COVID-19.