ITPA Genotype

About one in three people carries a variant that slows a particular cleanup enzyme inside their cells. Most of the time it changes nothing at all. It starts to matter the day someone hands you a prescription for azathioprine, mercaptopurine, or ribavirin.

This test reads the gene itself, so the answer is the same at 30 as at 70 and you only need it once. The point is having it on file before a first dose, not reconstructing it after your blood counts fall.

The Enzyme and What It Cleans Up

Your cells assemble DNA and RNA from a small set of standard building blocks. A few near-copies show up that were never meant to be used, and if they get built into a chromosome they cause trouble. ITPA (inosine triphosphate pyrophosphatase) codes for the enzyme that catches those rogue blocks and knocks them back down to a harmless form first.

The same enzyme handles a byproduct of thiopurine drugs. Azathioprine and mercaptopurine break down along a path that produces a compound called 6-thio-ITP, and the enzyme converts it back to a milder form. When the enzyme is slow, that compound accumulates inside the cell.

Two variants account for most of the reduced activity in the population. A single-letter swap in the coding sequence, written c.94C>A, leaves people with one copy running at roughly a quarter of normal enzyme activity and people with two copies at essentially none. A second change sits just outside the coding region and is gentler: one copy gives about 60 percent of normal activity, and one of each variant together lands near 10 percent. In Asian populations the coding variant turns up on 11 to 19 percent of gene copies, several times its frequency in Europeans.

Thiopurine Toxicity

This is the part with real weight. Azathioprine and mercaptopurine are workhorses in inflammatory bowel disease, autoimmune conditions, and childhood leukemia, and their main hazard is that they suppress bone marrow. Pooling the published studies, people carrying a reduced-activity variant run roughly two and a half times the risk of neutropenia. That is the fall in infection-fighting white cells that forces a dose cut or a full stop.

That figure comes mainly from children treated for leukemia, where thiopurine doses are highest. In inflammatory bowel disease specifically, a pooled analysis of the published cohorts found no clear link between either variant and low white cell counts, and a study in autoimmune hepatitis found the same nothing while a different gene, NUDT15, did predict trouble. So the strength of this signal depends on which illness you are being treated for.

In children treated for leukemia at doses already tailored to a different gene, TPMT (thiopurine methyltransferase), the reduced-activity variant still predicted severe fever with very low white cell counts. Among 262 adults with inflammatory bowel disease, the same variant tracked with low white cells, and liver injury in that group did not track with genotype, though pooled analyses in children on mercaptopurine since have found the coding variant associated with roughly twice the odds of liver toxicity. In a prospective Crohn's study, carriers and people with low TPMT activity were the ones most likely to quit azathioprine early because of side effects. A study in Ethiopian children on mercaptopurine found the same link to the most severe grade of neutropenia and to fever with it.

Unlike TPMT and NUDT15, this gene has not made it into thiopurine prescribing guidelines. So a reduced-activity result is an argument for watching more closely, not for starting at a different dose.

Ribavirin and Hemolytic Anemia

Ribavirin, used against hepatitis C and prescribed far less now than it once was, destroys red blood cells in a meaningful share of the people who take it. Here the reduced-activity variants protect you. In a cohort of 304 treated adults, carriers held their hemoglobin and needed fewer dose reductions. A study of 474 Japanese patients found the same, and so did work in 213 Caucasian patients, where it was the people with normal enzyme activity who ended up needing dose cuts and drug support to rebuild red cells.

The protection holds in people coinfected with HIV and hepatitis C, across hepatitis C genotypes, and in newer regimens that pair ribavirin with direct-acting antivirals instead of interferon. A meta-analysis pooling ribavirin-treated cohorts concluded that genotype predicts who develops hemolytic anemia well enough to be worth knowing before treatment starts.

Whether that protection also means a better cure rate is unsettled. One cohort of 238 patients found the variants prevented anemia without raising the share of people who cleared the virus, and a 2024 study of 170 people on sofosbuvir plus ribavirin reached the same conclusion. A Japanese cohort found higher clearance in carriers, and a Swedish trial linked the variants to less relapse after treatment. Treat the anemia finding as solid and the cure-rate question as open.

Same Variant, Opposite Directions

A variant that shields you from one drug and exposes you on another looks like a contradiction until you look at what the enzyme does. Its job is to clear one class of byproduct out of the cell, and whether clearing it helps you depends on which byproduct and which cell. Inside a red blood cell on ribavirin, letting that byproduct build up appears to steady the cell against destruction. Inside bone marrow on a thiopurine, letting the analogous compound build up is precisely what poisons the cells making your blood.

So this is not a good-number, bad-number test, and there is no direction you should want your result to point. It is a drug-specific result. The only question that makes it useful is which drug you are about to take.

The split shows up cleanly across three kinds of cohorts.

Sources, in order: Stocco et al. 2008; von Ahsen et al. 2005; Thompson et al. 2010.

If a thiopurine is on the table, a reduced-activity result argues for closer blood counts in the first weeks and for treating an unexplained drop in counts as drug toxicity rather than something to push through. If ribavirin is on the table, the same result is reassurance, not a reason to change anything about the dose.

Two Non-Working Copies

Inheriting a completely non-working copy from both parents is a different situation, and a rare one. It causes a severe infant disease with seizures in the first months of life, a head that grows too slowly, and in many children a failing, enlarged heart. In the largest published series of affected children, small head size present at birth and heart involvement were the features that predicted the worst outcomes. Brain imaging in these children shows a recognizable pattern in the deep white matter.

For an adult reading their own report, this matters in exactly one way. A single non-working copy does not cause that disease; it is the reduced-activity state the rest of this page is about. It matters if a partner carries one too, because then each pregnancy carries a one in four chance of the severe form, and that is a conversation to have before pregnancy rather than during one.

Signals Still Being Worked Out

One cross-sectional study of 102 adults with blood cancers found that lower enzyme activity went along with more mutations in mitochondrial DNA. That is the small separate genome inside the compartments that power your cells. A snapshot like that cannot say which came first. It is a reason the gene is being studied, not a risk figure that applies to you.

Sequencing Versus Measuring the Enzyme

There are two ways to get at this. You can sequence the gene, or you can measure enzyme activity directly in red blood cells with a laboratory separation method. The activity assay has an appeal: it captures how the enzyme is actually performing, whatever the underlying letters say, and it correlates well with genotype in studied populations.

It also drifts. Enzyme activity in red cells is lower in people living with HIV than in uninfected controls, and that difference is not explained by the antiviral drugs they take. A recent blood transfusion measures someone else's cells, not yours. Sequencing the gene avoids both problems and picks up rarer variants that a targeted assay checking two known spots would skip entirely. Which variants a given lab includes is not standardized, and at least one analysis found that a different pairing of markers predicted ribavirin anemia better than the pairing in common use.

A Result You Only Need Once

Your DNA sequence does not move, so there is no trend here to follow and no reason to repeat this test. The only situation that warrants running it again is doubt about the call itself, which is a question of confirming by a second method rather than watching a number over time.

What does need repeating are the companion tests. If you take a thiopurine, your complete blood count and your liver enzymes are the things to track: closely in the first weeks on the drug, and on a regular schedule after that for as long as you stay on it. A reduced-activity genotype is an argument for the tighter end of whatever monitoring interval your prescriber suggests, not the looser end.

What an Unexpected Result Should Change

A reduced-activity result is not read alone. Two other genes govern thiopurine safety: TPMT and NUDT15, and NUDT15 in particular drives toxicity in people of Asian and Hispanic ancestry, where testing it before starting a thiopurine is increasingly expected. If more than one of the three comes back reduced, that combination carries more weight than any single result and is worth raising with whoever is prescribing, usually a gastroenterologist, rheumatologist, or hematologist.

If sequencing returns something unexpected, a variant nobody has characterized or a call the lab flags as low confidence, confirmation by a second method is reasonable, particularly for insertions and deletions, which are the variant types sequencing handles least reliably. For a result suggesting two non-working copies, or for any question about what your result means for children and siblings, a genetic counselor is the right person, not a general search.

The pattern that should actually change behavior is this: a reduced-activity genotype plus falling white cells or platelets in the first months on a thiopurine. That combination points at the drug. Chasing other explanations while the counts keep dropping is the mistake this test exists to prevent.

When a Genetic Result Can Mislead

• Panel coverage: an assay reports the variants it was built to find. Labs test different sets, and roughly one in seven pathogenic variants across genes is hard for sequencing to detect at all. A result showing no variant lowers the odds that you carry one; it does not eliminate them.
• Ancestry: the coding variant appears on 11 to 19 percent of gene copies in Asian populations and is less frequent elsewhere. How much a negative result reassures you depends on which variants are common in your own background.
• Variants of uncertain significance: sequencing sometimes returns a change nobody can yet classify. In healthy people with no symptoms these are especially hard to interpret, and an uncertain call should not be treated as a diagnosis or as a reason to avoid a drug you need.
• A recent transfusion or marrow transplant: a blood sample drawn soon after either can read donor DNA rather than your own. A cheek swab, or a blood draw timed well away from those events, gives your own genotype.
• Genotype is not current enzyme activity: your genes set the ceiling, but measured activity can sit lower for other reasons, including HIV infection. A normal genotype is not a guarantee of normal enzyme function on any given day.