Extensive pneumatosis intestinalis and portal venous gas mimicking mesenteric ischaemia in a patient with SARS-CoV-2

Abstract

We present the case of a critically ill 47-year-old man diagnosed with SARS-CoV-2 (COVID-19) who developed extensive pneumatosis intestinalis and portal venous gas in conjunction with an acute abdomen during the recovery phase of his acute lung injury. A non-surgical conservative approach was taken as the definitive surgical procedure; a complete small-bowel resection was deemed to be associated with an unacceptably high long-term morbidity. However, repeat computed tomography four days later showed complete resolution of the original computed tomography findings. Pneumatosis intestinalis from non-ischaemic origins has been described in association with norovirus and cytomegalovirus. To our knowledge, this is the first time that this has been described in COVID-19.

Background

This case highlights an unusual association between active SARS-CoV-2 infection and a computed tomography (CT) appearance that mimics extensive mesenteric ischaemia.

Case history

A 47-year-old man presented to the emergency department with a one-week history of fever, dry cough and vomiting. His past medical history was unremarkable except for anxiety, treated with fluoxetine, and obstructive sleep apnoea for which he was on nocturnal continuous positive airway pressure (CPAP). His body mass index (BMI) was slightly elevated at 31.

At presentation, he was hypoxic with an arterial oxygen tension (PaO₂) of 75mmHg on 40% oxygen via a face mask. Chest x-ray showed bilateral patchy opacities (Fig 1). The patient received augmentin and clarithromycin as treatment for a community-acquired pneumonia and was placed in isolation on the ward. Subsequently, a nasopharyngeal swab for SARS-CoV-2 returned positive. The patient deteriorated on day two of his hospital admission with increasing inspired oxygen requirements and required admission to intensive care for mechanical ventilation.

Figure 1.

Chest x-ray showing bilateral pulmonary infiltrates

To facilitate ventilation over the initial 48 hours, paralysis was induced with an atracurium infusion and nebulised prostacyclin was used to minimise hypoxic pulmonary vasoconstriction. A modest negative fluid balance was achieved with the aid of frusemide and acetazolamide. There was no haemodynamic compromise and no vasopressor requirement throughout the intensive care stay. Enteral feeding was established on day 2 of this period and his bowel motions were normal. His respiratory function improved gradually and by day 8 he had largely weaned from the ventilator and required 30% oxygen to maintain saturations above 92%.

However, on day 8 of intensive care, a distended yet soft abdomen was noted with associated diarrhoea. His lactate remained below 2mmol/l and he did not require vasopressor support, although he did develop a sinus tachycardia. He became intolerant of his enteral feed and feculent-like material was aspirated from the nasogastric tube. A contrast-enhanced CT of the abdomen revealed diffuse small-bowel distension with widespread pneumatosis, circumferential mural thickening, free fluid, mesenteric free air and portal venous gas (Fig 2). There was pneumatosis affecting the jejunum, proximal ileum and caecum with skipped normal segments of distal small bowel. Normal blood flow was visualised through the origin of the superior mesenteric artery. CT of the thorax showed bilateral ground glass changes in keeping with his diagnosis of SARS-CoV-2 (Fig 3).

Figure 2.

Coronal computed tomography image showing extensive pneumatosis intestinalis, bowel wall thickening, mesenteric free air and portal venous gas

Figure 3.

Computed tomography of the thorax showing bilateral ground glass changes

A surgical opinion was sought. The initial opinion was that the CT was highly suspicious of extensive mesenteric ischaemia. A decision was made not to proceed to laparotomy because the definitive surgical procedure, if the CT appearance was truly indicative of extensive mesenteric ischaemia, would be a complete small bowel resection, which would be associated with a very poor outcome.

The patient was commenced on an infusion of unfractionated heparin with a targeted activated partial thromboplastin time ratio of 2.0–2.5 and crystalloid fluid boluses were administered. A dobutamine infusion was commenced at 5μg/kg/minute for 12 hours and nasogastric feeding was stopped.

The patient remained haemodynamically stable. His lactate levels never increased above 2mmol/l and vasopressor support was never required. His abdomen remained soft but distended. A repeat CT was performed 4 days later, on day 12 of his intensive care stay. This showed complete interval resolution of pneumatosis intestinalis (Fig 4). Enteral feeding recommenced uneventfully. He was successfully extubated the following day and was discharged to the ward 48 hours later, and subsequently to home.

Figure 4.

Coronal computed tomography of the abdomen showing complete resolution of the pneumatosis intestinalis and the portal venous gas

Discussion

SARS-CoV-2 is known to be present within the bowel and gastrointestinal upset is a common presenting feature in these patients.¹ Direct mucosal damage as opposed to mesenteric ischaemia is likely to have led to the CT findings we describe here. This is supported by the normal levels of lactate, haemodynamic stability and presence of normal blood flow through the superior mesenteric artery. Furthermore, the original SARS virus was described as causing degenerative changes in the lymphoid component of the intestine, and lymphoid follicles showed a marked decrease in lymphocytes. In severe cases only the depleted stromal framework structure remained.² Pneumatosis intestinalis has also been described in association with other viruses, including norovirus and cytomegalovirus,^3,4 and the mechanism is purported to be mucosal disruption.

We propose that atrophy of the lymphoid follicles, caused by SARS-CoV-2, resulted in increased mucosal permeability permitting dissection of the gas into the bowel wall.⁵ To our knowledge, this is the first time that this condition has been described in COVID-19. We suggest that a conservative approach can be considered when these findings occur in the absence of signs of systemic deterioration and that a good outcome is possible.