He was initially found to be anaemic with haemoglobin (Hb) of 9.8 mg/dl with a normochromic normocytic picture. This was noted to be chronic. His iron studies showed saturation of 9% with normal B₁₂, folate and coagulation profile. His gastroscopy, haemolytic and celiac screen were negative. During the work up of his anaemia, he started to develop lower grade fever between 38 and 38.5 °C. Investigations, including chest radiograph and urine culture, came back negative. However, blood cultures grew coagulase negative staphylococcus on three different occasions. Subsequently, transoesophageal ECG demonstrated vegetations in keeping with endocarditis. He was started on vancomycin and rifampicin as per microbiological advice. The inflammatory markers fell with treatment. However, his anaemia was noted to be getting worse and required blood transfusion. He was not well enough to have colonoscopy and his faecal occult blood test was positive on three different samples. There was a noted episode of melaena. Aspirin was stopped and subsequent CT of the abdomen and carcinoembryonic antigen (CEA) levels was normal.

In the context of his endocarditis and anaemia, his haemolytic screen was repeated and was negative. However, his coagulation profile was quite deranged with high prothrombin time, high partial thromboplastin time, normal fibrinogen and international normalised ratio (INR) of 4.8, which was not the case on admission. Unfortunately, he did not have d-dimer as part of the work-up. The blood picture support a diagnosis of vitamin K deficiency rather that disseminated intravascular coagulation (DIC) as his platelets and fibrinogen level were normal. Liver function tests were normal and liver texture appeared normal on a CT scan. The patient was started on supplement vitamin K and his rifampicin was stopped. Unfortunately, the patient did not have vitamin K measurement as supplement treatment was started immediately.

Both phylloquinone and menaquinone are used by the human liver;¹ therefore, anyconditions affecting either type can result in vitamin K deficiency. Such conditions include malnutrition, common bile duct obstruction (stones, primary biliary cirrhosis and cholangiocarcinoma), malabsorption syndromes, parenchymal liver disease (cirrhosis) and other conditions such as lupus, multiple myeloma and nephrotic syndrome. It has been hypothesised that antibiotics could cause vitamin K deficiency through two mechanisms: either direct inhibition of hepatic utilisation of vitamin K or antibiotic mediated eradication of vitamin K producing intestinal micro flora resulting in iatrogenic vitamin K deficiency and, possibly, life-threatening coagulopathy.^{2 3}

Our patient had deranged coagulopathy within 2 weeks of being on antibiotics, which could suggest subclinical deficiency prior to the antibiotics usage. Subclinical vitamin K deficiency is difficult to assess through routine blood tests and may require special tests to diagnose (see below). It has been reported in experiments that rifampicin can cause vitamin K deficiency and the effect depends on the dose and duration.⁴

In the critical care settings, there is increased awareness of vitamin K deficiency in this group of patients but it is still a debatable subject. One study showed that vitamin K deficiency is common in the intensive care units setting with up to 25% of admitted patients (using functional coagulation factor II to Echis factor II ratio <0.70 as a method of diagnosis).⁵ However, only 5% of these patients developing coagulopathy required medical attention. On the other hand, long-term antibiotics, like endocarditis, will put patients under increase risk of coagulopathy; hence, this category of patients needs to be monitored especially if they are on rifampicin.