The history of warfarin

It was a Saturday afternoon in February 1933. ¹ Despite the fierce blizzard raging outside, it was quiet in the laboratory at the University of Wisconsin campus in Madison. Professor Karl Paul Link and his colleagues were working undisturbed when suddenly Ed Carlson, a young farmer, burst through the door in a flurry of wind and snow. He was agitated and upset. Several of his cows had died in the preceding days and he had driven 190 miles with a dead heifer, a milk churn full of her uncoagulated blood and a pile of mouldy straw. He was looking for the state veterinarian but somehow stumbled upon Link’s research laboratories.

Cover photo. Karl Paul Link (right), professor of biochemistry, with Mark A Stahmann. University of Wisconsin-Madison, 1940. Courtesy of University of Wisconsin-Madison Archives. S00094.

A few years earlier, several unexplained cattle deaths in the area had been attributed to a haemorrhagic disease. Investigations by local veterinary scientists had traced the source to mouldy hay, discovering that something in the feed was preventing clotting by reducing prothrombin activity.^2,3 The problem was thought to be due to the fermentation of a particular strain of sweet clover, but no active agent was identified. The disease was reversible if the feed was stopped in time and, in some cases, transfusions from healthy cattle also arrested the disease. Farmers were advised to avoid feeding contaminated hay to their cattle, but the farming community, struggling through the depressed economic circumstances of the 1930s, had few other options. They were sceptical of the science, refusing to believe the sporadic deaths were due to the feed they had used for generations.

In Link’s laboratory they were studying sweet clover, but not the haemorrhagic disease. Clover contains coumarin, a colourless substance with a vanilla odour, responsible for the sweet smell of newly mown hay. It has a bitter taste, and cattle and rabbits avoid it given other choices. Link was working with the university geneticists, R Alexander Brink and WK Smith to isolate a clover that was low in coumarin and would suit the Wisconsin growing conditions. But Carlson’s arrival that Saturday afternoon switched the focus of their research. They suddenly realised that advising the impoverished farmers to change feeds was useless and the disease would continue unless they could identify the problem and find a solution.

Of obvious interest to farmers and veterinarians, the subsequent research was also closely followed by the medical community. Heparin had recently been incorporated into clinical medicine as prophylaxis and treatment for deep vein thrombosis and, while its benefits were clear, it was expensive. ⁴ It also required regular intravenous injections, a complicated procedure before the invention of plastic intravenous catheters. An oral anticoagulant was therefore eagerly anticipated.

The breakthrough came in June 1939 when Harold Campbell, one of Link’s assistants, isolated a white crystalline substance later named dicoumarol. ⁵ Over the next few months, they were able to synthesise dicoumarol, and many other substances with similar properties. ¹ Importantly, they established that dicoumarol was the degradation product of coumarin present in the spoiled hay.

The Wisconsin Alumni Research Foundation (WARF) at the University of Wisconsin, helped the research team secure patents on many of the compounds they synthesised. ⁶ Link then made dicoumarol available to medical teams at Wisconsin General Hospital and the Mayo Clinic. Hugh Butt and his colleagues published their initial findings from the Mayo Clinic in 1941. ⁷ Having established that dicoumarol prolonged prothrombin time in dogs with no ill effects, they extrapolated the dose to humans and administered it to six people, including some volunteers, for several days with similar results.^7,8 The following year they published a much bigger study where 368 patients were given oral dicoumarol for either treatment or prevention of thromboembolic disease. They concluded that ‘studies strongly suggest the value of administering dicoumarin in preventing intravascular thrombosis’. ⁹ They also counselled that blood transfusion was an appropriate and effective treatment if patients developed haemorrhagic complications but ‘synthetic vitamin K has little or no effect’. ⁹

This last statement, repeated by many early researchers, was a great source of frustration to Karl Link. ¹ Even before they isolated dicoumarol, the scientists at his laboratory had conclusively established that feeding concentrated alfalfa hay to affected cattle reversed the haemorrhagic disease. ¹⁰ Alfalfa hay contains high levels of vitamin K and Link subsequently demonstrated that vitamin K reversed the effects of dicoumarol on prothrombin time. While physicians understood that low vitamin K levels in liver disease prolonged prothrombin time, they did not believe this could be corrected by synthetic vitamin K administration. Link attributed this to less sensitive tests, insufficient vitamin K and ‘meager experimental trials’. ¹⁰ Even though Shapiro subsequently showed that vitamin K was an effective reversal agent in humans, ¹¹ it was some years before this was clinically accepted.

This was partly because research slowed considerably in the mid-1940s as resources and personnel were diverted to the war effort. Then, in 1945, Karl Link became unwell with recurrent tuberculosis. ¹ Bored during his protracted sanatorium stay and frustrated by his stagnant laboratory, Link began reviewing the data from earlier experiments, discovering that one substance they had created, coumarin 42, was particularly selective for rats. Once back at work, he proposed developing a rodenticide. There were some reservations from the university but coumarin 42, named warfarin by Link (combining WARF, the research foundation and -marin from coumarin) was an instant commercial success with farmers. Existing rodenticides, such as strychnine, were hazardous and required multiple dosing, but a single dose of warfarin-impregnated grain eliminated rats efficiently with little risk to other domestic animals.

Meanwhile, dicoumarol was proving to be an unpredictable drug and the medical community was seeking an alternative. Several other substances, like Tromexan ¹² and Marcumar ¹³ , were produced and marketed, particularly in Europe, but none met all the desired criteria. ¹⁴ Link’s ongoing research all pointed to warfarin being the ideal medication, but it was difficult to convince the medical community, ‘the transition to a substance originally promoted to exterminate rats and mice was a bit more than they could accept with enthusiasm’. ¹ Perversely, it was the attempted suicide of a young army recruit that provided the impetus for warfarin research in man. After taking warfarin for 5 days, the recruit presented with significant haemorrhagic disease, which was successfully treated with blood transfusion and vitamin K. ¹⁵ As a direct result of the information provided by the treating team, human trials commenced with an intravenous form of warfarin, and later oral preparations.^16,17 In 1955, Byron Pollock and colleagues stated ‘warfarin sodium possesses properties that make it more nearly an ideal anticoagulant than the other agents now available’. ¹⁷

By 1964, warfarin was being recommended as long-term outpatient therapy for all thromboembolic disease. ¹⁸ Its growth was exponential from there and by 2006, it was estimated that at least 1% of the UK population, and 8% of those over 80, were taking warfarin. ¹⁹ However, in the last 10 years, there has been a dramatic shift in warfarin prescribing with the introduction of direct oral anticoagulants, like rivaroxaban and apixaban. ²⁰