A person with mild hemophilia A can get anywhere from a quarter to more than triple the standard response to the same dose of desmopressin, depending on which single spelling change they carry in one gene.
That range comes from a study of 1,441 people published July 27 in Research and Practice in Thrombosis and Haemostasis ↗. It is the largest look yet at why a decades-old, cheap peptide drug works well for some patients and barely at all for others.
What desmopressin does
Desmopressin ↗, sold as DDAVP, Stimate, and other brands, is a synthetic version of vasopressin, the hormone that tells the kidneys to hold water. Ferring Pharmaceuticals built it in the 1960s by tweaking the natural molecule. Besides its water-retention job, it has a second use. It pushes the cells lining the blood vessels to dump their stored clotting factor VIII and von Willebrand factor into the blood. That spike is often enough to control or prevent bleeding in people with mild forms of hemophilia A, the inherited disorder where factor VIII runs low.
The appeal is practical. A shot or a nasal spray of desmopressin costs a fraction of an infusion of manufactured factor VIII and can be given at home. The catch has always been that the size of the factor VIII bump varies widely from patient to patient, which is why bleeding-disorder centers give a test dose first and measure the response before anyone relies on it for surgery or an emergency.
What the gene does
Hemophilia A is caused by errors in F8, the gene that codes for factor VIII. A team applied a machine-learning method called SHAP to a population pharmacokinetic model, a way of teasing apart which patient traits drive a drug response, using desmopressin test data from 1,441 people who between them carried 55 different F8 missense variants (single-letter swaps in the protein).
Ten of those variants moved the needle enough to matter. Three blunted the response. Carrying Tyr450Asn, Gly477Val, or Pro149Arg left patients with 24 percent, 25 percent, and 56 percent of the reference factor VIII rise. Seven amplified it. Arg717Leu carriers got 336 percent, more than triple, with Thr74Met, Ser2030Asn, and four others ranged in between. Same drug, same dose, a roughly fourteenfold spread set by genotype.
The delivery route mattered too. Compared with an intravenous dose (100 percent), a subcutaneous shot delivered 88 percent of the peak and the intranasal spray 73 percent. The two convenient options give up some of the effect.
Why it is not a substitute for testing
The honest reading is narrower than "a gene test replaces the test dose." This is a retrospective model built on data already collected, not a prospective trial. The reference group it measures everything against lumps together most variants plus patients whose variant was unknown, a mixed baseline. Only 10 of the 55 variants had enough carriers to reach significance. The rest are silent for lack of numbers, not proven neutral.
And the model does not make the uncertainty disappear. Adding genotype, route, and other patient traits cut the unexplained spread in factor VIII response from 141.6 percent to 78.1 percent. That is a real improvement, close to half, but still a wide band. A carrier of a known good-responder variant could still fall short on the day it counts. The authors keep the guideline intact: test the individual before you trust the drug.
What the study adds is a prior. If a patient carries Tyr450Asn, the odds that desmopressin alone will carry them through a procedure are poor before the first vial is drawn, and the plan can account for that early rather than discover it in the operating room.
Platform note
Desmopressin is one of the oldest approved peptide drugs still in daily use, a V2 vasopressin-receptor (avpr2 ↗) agonist that predates the entire GLP-1 era. Its story here is a reminder that whether a peptide works is often less about the molecule than about the genome reading it.