Bayesian Analysis of Blood Transfusion in Dracula

M O’Connor, N Dunbar

Dartmouth-Hitchcock Medical Center, Lebanon, USA


Bram Stoker’s novel Dracula was one of the first works of popular fiction to feature a case of blood transfusion. The book was published prior to the development of ABO cross-matching, so the recipient of the transfusion was fortunate to avoid a hemolytic transfusion reaction. Analyzing this case from a modern perspective allows one to review the history of transfusion medicine, the science of ABO incompatibility, and basic principles of probability. Bayesian reasoning can be used to conclude that Dracula’s victim likely had type A blood.

Bram Stoker’s novel Dracula is well known for spawning a wildly successful genre of vampire literature, but it also offers a glimpse into the history of transfusion medicine as it was one of the first works of popular fiction to feature a case of blood transfusion. Stoker may have learned about the procedure from one of his brothers, three of whom were physicians including one who eventually became president of the Royal College of Surgeons in Ireland. Here we present an analysis of this fictional case of blood transfusion from Dracula as a way of reviewing basic principles of transfusion medicine and probability.

Let us begin with a case description. The patient is a previously healthy young woman named Lucy who is attacked at night by the Transylvanian vampire Count Dracula. Lucy quickly becomes severely anemic, and her worrisome appearance prompts a friend to summon Abraham van Helsing, a man said to know as much about obscure diseases as anyone in the world. The doctor recognizes Lucy’s vampire-induced anemia and declares, “She will die from sheer want of blood!” He arranges for her to receive four blood transfusions, each from an unrelated donor. She tolerates all four transfusions without any side effects, regaining some vitality. Unfortunately, Dracula returns, and Lucy eventually succumbs to the vampire and joins the ranks of the undead. Reading this story from a modern perspective, one notices that Stoker makes no mention of cross-matching the donated blood. This apparent omission reflects the state of medical knowledge at the time. When Dracula was published in 1897, no one had started cross-matching donors and recipients. Blood transfusion was a risky undertaking, and life-threatening reactions were common and poorly understood. James Blundell, an English obstetrician who performed many of the first human-to-human transfusions starting in the first half of the nineteenth century, noted that many of his patients “suffered fever, backache, headache, and passed dark urine” after receiving donated blood.1 Some of the patients died, and Blundell did not have an explanation.

In 1900, three years after the publication of Dracula, the puzzle was solved by Austrian physician Karl Landsteiner, who realized that there are four ABO blood types and that not all blood types are compatible. It is now known that one’s blood type is determined by a single gene. This gene is called ABO and encodes an enzyme that adds a sugar molecule to a carbohydrate chain on red blood cells. The gene has three alleles. The A and B alleles encode enzyme variants that add different sugars, and the O allele is the A allele with a frameshift mutation rendering the protein non-functional. At a very early age, people develop IgM antibodies to A or B antigens not found on their own red blood cells because these antigens resemble polysaccharides from gastrointestinal bacteria. These naturally occurring antibodies cause red blood cells with foreign ABO antigens to agglutinate, leading to hemolytic transfusion reactions such as those described by Blundell.

Let us now return to the case from Dracula. Lucy receives four units of blood without cross-matching from four unrelated donors. For the sake of this analysis, let us assume that any incompatible blood would have caused a reaction. This assumption is likely safe given Lucy’s young age, but it is worth noting that there are some patients with low antibody titers who can receive incompatible transfusions without symptoms.2 Lucy’s risk of having a hemolytic transfusion reaction can be calculated based on the distribution of the four blood types in the European population. In this group, 44% of people are type O, 43% are A, 9% are B, and 4% are AB.3 If we assume that Lucy has a certain blood type, her probability of finding four compatible donors is simply the probability of finding one compatible donor for that blood type raised to the fourth power. Taking the weighted average of these four values weighted by the initial probability of each blood type gives the overall probability that all four transfusions are successful. Using this approach, one finds that the overall chance of a European such as Lucy tolerating four non-matched transfusions from European donors is only 31%. Lucy is fortunate to have avoided a hemolytic transfusion reaction.

The fact that she does avoid a reaction allows us to make inferences about her blood type. This type of thinking is known as Bayesian reasoning. To understand the idea, consider someone who has tolerated one hundred non-matched transfusions from random donors. Given this transfusion history, it is highly likely that the recipient has type AB blood, the universal recipient. Applied to Lucy’s case, Bayes’ theorem can be used to generate four conditional probabilities, one for each possible blood type. Her chance of having type O blood, the most common blood type in the population but also the least compatible on the receiving end of a transfusion, falls from 44% to 6%. The probability of her having AB blood rises but only to 13% given the low prevalence of AB blood in the population. Her probability of having type A blood rises from 43% to 79%, and her chance of having type B blood falls from 9% to 2%. Thus, it is possible to conclude that Lucy, the first victim in Bram Stoker’s classic novel Dracula, likely had type A blood.

Working through this analysis provides an opportunity to review basic principles of transfusion medicine and Bayesian reasoning. The case has been used successfully as a teaching tool for medical students and residents. It prompts one to remember the basics of ABO compatibility and to think about Bayes’ theorem, which is the basis for all discussions of pre- and post-test probability. Lucy tolerates all four transfusions, but given her low pre-test probability of being a universal recipient, her post-test probability of having type AB blood is still low. And as a final aside, this analysis might even have implications for the genre of vampire literature as a whole. Some bloodsucking creatures such as mosquitoes are known to preferentially seek out certain blood types when it is time to feed.4 It is reasonable to speculate that vampires might exhibit preferential feeding behavior as well. If so, perhaps those with type A blood should watch out.

Correspondence: M O’Connor

Dartmouth-Hitchcock Medical Center, One Medical Center Drive, Lebanon, NH 03755, USA

Email: mark.j.o’


  1. Learoyd P. The history of blood transfusion prior to the 20th century–part 2. Transfus Med. 2012 Dec;22:372-6.
  2. Janatpour KA, Kalmin ND, Jensen HM, Holland PV. Clinical outcomes of ABO-incompatible RBC transfusions. Am J Clin Pathol. 2008 Feb;129:276-81.
  3. Reid M, Lomas-Francis C. The Blood Group Antigen Facts Book. 2nd ed. New York: Elsevier Academic Press; 2004
  4. Anjomruz M, Oshaghi MA, Pourfatollah AA, Sedaghat MM, Raeisi A, Vatandoost H, Khamesipour A, Abai MR, Mohtarami F, Akbarzadeh K, Rafie F, Besharati M. Preferential feeding success of laboratory reared Anopheles stephensi mosquitoes according to ABO blood group status. Acta Trop. 2014 Dec;140:118-23.

Page 362