Near-fatal bleeding, senna, and the opposite of lettuce

Abstract

In February, 2006, a 45-year-old woman went to her local hospital with diff use abdominal pain. When the hospital realised that she might have a serious illness, she was brought by ambulance to our hospital, for insurance reasons. By the time she arrived, the pain had been present for 15 h. Her aortic valve had been replaced in 1999, but she had had no abdominal problems until mid-2005, when she developed a tendency to constipation. She had therefore taken 1–2 tablets of a senna-based laxative, once or twice a week. She had seen no need to tell her doctors about the laxative. For 3 weeks preceding her admission she had taken several tablets of laxative a day, and passed loose, watery stools up to three times a day. The day before the pain started, she had taken many tablets, after passing no stool for 2 days. She had then passed three watery stools, followed by three bloody stools— some blood was fresh and red, and some old and black. She had no other medical history of note; notably, she had no psychiatric disorder. She was on no drugs other than warfarin, which she had taken ever since her operation. She took her warfarin regularly, and her international normalised ratio (INR) had been in the therapeutic range for years. Her diet, including her consumption of green vegetables, varied little. She took no herbal medications. 25 days before admission, her INR had been 2·3. She was hypotensive and tachycardic. Her abdomen was distended, with widespread tenderness and guarding. Rectal examination revealed fresh blood. Blood tests showed a haemoglobin concentration of only 84 g/L; the platelet count was normal, at 164×109/L, but the INR was 11·9; the activated partial thromboplastin time was 92·4 s (control time 29·3 s). The concentrations of bilirubin and aspartate aminotransferase were slightly high, at 33·5 μmol/L and 48 U/L, respectively, but the results of liver function tests were otherwise normal. CT of the abdomen showed extensive haematoma in the pelvic, paracolic, and subhepatic regions. We admitted the patient for several days, during which we gave her vitamin K, fresh frozen plasma, and packed red cells. She subsequently resumed taking warfarin, and stopped taking laxatives in excess; her INR rapidly returned to the target range, where it has remained ever since. When last seen, in November, 2007, she was well. Vitamin K is an essential cofactor for γ-carboxylation, the post-translational modifi cation of coagulation factors II, VII, IX, and X (fi gure). Although vitamin K is also a cofactor for the γ-carboxylation of protein C and protein S, lack of vitamin K reduces coagulation. Green leafy vegetables have especially high con centrations of vitamin K, but the vitamin K in vegetable oils and soya may be more bioavailable. Many patients on warfarin are aware—as, indeed, was our patient—that their dietary consumption of vitamin K should remain fairly constant. However, not all are aware that diarrhoea can reduce absorption of vitamin K, and increase the risk of bleeding. Although some physicians recommend halving the dose of warfarin when the patient has diarrhoea, many doctors seem unaware of the increased risk of bleeding—and many textbooks, formularies, and patient-information leafl ets fail to mention the increased risk, or do so only in passing. Cases like ours will continue to occur, until doctors and patients are better informed.