G6PD: the deficiency with a much shorter drug list than its reputation
G6PD deficiency has one of the longest folk lists of forbidden medicines in medicine, and one of the shortest evidence-graded ones. The gap between those two lists is most of what this page is about.
Also known as: glucose-6-phosphate dehydrogenase.
What G6PD does
Red blood cells cannot repair themselves. They carry no nucleus and no machinery to rebuild damaged proteins, so their only defence against oxidative stress is a chemical buffer, and G6PD is the enzyme that keeps that buffer stocked. When the enzyme is deficient and something oxidising arrives, whether a drug, an infection or a plate of fava beans, the buffer runs out and the cells break apart. That is acute haemolytic anaemia: not an allergy, not a slow toxicity, but red cells failing in bulk over a day or two.
The activity groups
CPIC grades G6PD by residual enzyme activity rather than by metabolizer speed: normal, deficient, deficient with chronic non-spherocytic haemolytic anaemia, variable, and indeterminate. The gene sits on the X chromosome, and that geometry drives the odd-looking middle categories. A male has one copy, so a deficient variant leaves him plainly deficient. A female with one deficient copy has a mixed population of red cells, because X-inactivation is random, and the fraction that is deficient differs from person to person. That is what CPIC's variable category means, and it is why the guidance for variable results points at measuring enzyme activity rather than at a rule.
Medications where it matters
primaquine
Primaquine and its relatives are the reason G6PD testing exists as routine practice. The drug is needed to clear the liver stage of Plasmodium vivax and ovale malaria, and in a deficient person a standard anti-relapse course carries a high risk of haemolysis. This is one of the few places where the guidance is a genuine avoid, with named exceptions run under expert malaria protocols.
| If your result is | Published CPIC guidance |
|---|---|
| Deficient | Per CPIC (A): "Avoid primaquine, except in the following cases where established expert consensus guidelines for the treatment of malaria should be followed: (1) Treating Plasmodium vivax or Plasmodium ovale malaria for radical cure of liver-stage infections: 0.75 mg/kg once weekly x8 weeks (WHO) or 45 mg once weekly x8 weeks (CDC) - with close monitoring for hemolysis; (2) Treating Plasmodium falciparum malaria by using primaquine single dose as a gametocytocide at 0.25 mg/kg (WHO) - without need for monitoring for hemolysis." |
| Deficient with CNSHA | Per CPIC (A): "Avoid primaquine" |
| Indeterminate | Per CPIC (A): "To ascertain G6PD status, enzyme activity must be measured. Drug use should be guided per the recommendations based on the activity-based phenotype." |
| Normal | Per CPIC (A): "No reason to avoid based on G6PD status" |
| Variable | Per CPIC (A): "To ascertain G6PD status, enzyme activity must be measured. Drug use should be guided per the recommendations based on the activity-based phenotype." |
rasburicase (Elitek)
Rasburicase is given to bring down uric acid during tumour lysis, often urgently and often before anyone has had time to ask about G6PD. It carries a boxed warning for exactly this reason, and the guidance for a deficient result is unambiguous.
| If your result is | Published CPIC guidance |
|---|---|
| Deficient | Per CPIC (A): "Avoid use" |
| Deficient with CNSHA | Per CPIC (A): "Avoid use" |
| Indeterminate | Per CPIC (A): "To ascertain G6PD status, enzyme activity must be measured. Drug use should be guided per the recommendations based on the activity-based phenotype." |
| Normal | Per CPIC (A): "No reason to avoid based on G6PD status" |
| Variable | Per CPIC (A): "To ascertain G6PD status, enzyme activity must be measured. Drug use should be guided per the recommendations based on the activity-based phenotype." |
nitrofurantoin
Nitrofurantoin is a first-line urinary antibiotic, which makes it the drug on this list a G6PD-deficient person is most likely to be offered on an ordinary Tuesday. CPIC grades it as a medium risk for a deficient result, and the recommendation is caution and monitoring rather than a flat ban.
| If your result is | Published CPIC guidance |
|---|---|
| Deficient | Per CPIC (A): "Use at standard doses with caution and with close monitoring for anemia" |
| Deficient with CNSHA | Per CPIC (A): "Avoid" |
| Indeterminate | Per CPIC (A): "To ascertain G6PD status, enzyme activity must be measured. Drug use should be guided per the recommendations based on the activity-based phenotype." |
| Normal | Per CPIC (A): "No reason to avoid based on G6PD status" |
| Variable | Per CPIC (A): "If deemed necessary to ascertain G6PD status, enzyme activity must be measured. Drug use should be guided per the recommendations based on the activity-based phenotype." |
sulfamethoxazole / trimethoprim (co-trimoxazole, Bactrim)
This one is on the page to correct a widely repeated claim. Sulfa antibiotics head most folk lists of drugs that G6PD-deficient people must never take. CPIC's graded assessment puts co-trimoxazole in the low-to-no risk group at standard doses for a deficient result. The row below is what the published guidance actually says, which is not what the internet says.
| If your result is | Published CPIC guidance |
|---|---|
| Deficient | Per CPIC (C): "No reason to avoid based on G6PD status at standard doses" |
| Deficient with CNSHA | Per CPIC (C): "Use drug cautiously; if used, close monitoring for acute exacerbation of chronic hemolysis is recommended" |
| Indeterminate | Per CPIC (C): "To ascertain G6PD status, enzyme activity must be measured. Drug use should be guided per the recommendations based on the activity-based phenotype." |
| Normal | Per CPIC (C): "No reason to avoid based on G6PD status" |
| Variable | Per CPIC (C): "No reason to avoid based on G6PD status at standard doses" |
Is this in your raw DNA file?
G6PD is not part of the standard chip or variant-file report. It is a Deep Read gene, read from aligned reads in a BAM or CRAM file, because hundreds of different variants cause deficiency and grading them needs sequence rather than a chosen handful of positions. There is a second limit that no genetic method removes: CPIC's own guidance says that for a variable or indeterminate result the way to establish G6PD status is to measure enzyme activity in blood. Genotype narrows the question. Only the enzyme assay answers it, and that is a test to ask a clinician for.
A genotyping chip reads only a few hundred thousand pre-selected positions, about 0.02% of your genome, chosen for common variation. It does not sequence the rest, so it cannot find the rare or novel pathogenic variants the clinical and carrier modules look for: a “no finding” from chip data means “this chip never looked”, far more so than with whole-genome sequencing. Chip data suits common-variant traits, pharmacogenomic tag SNPs, and haplogroup ancestry, not clinical or carrier screening.
Common questions
Is G6PD in my 23andMe or AncestryDNA raw data?
A few G6PD positions may be on the array, but deficiency is caused by hundreds of different variants across the gene and a chip tests only what it was designed to test. Aimosti reads G6PD from aligned reads in the Deep Read panel, and reports an unresolved genotype as such rather than as normal.
Does a genetic result replace the enzyme test?
No, and CPIC says so directly. For a variable or indeterminate result the guidance is to measure enzyme activity. Genotype tells you which question to ask; the blood test answers it. This matters most for women, where one deficient copy produces a mixed cell population that genotype alone cannot quantify.
So which medicines actually have to be avoided?
Far fewer than the circulating lists suggest. In CPIC's grading the high-risk group for a deficient result is small and specific, and most of the drugs people are told to fear sit in the low-to-no risk group at standard doses. The tables above show the graded position drug by drug. That said, if you have had a haemolytic episode on a particular medicine, that history outranks any table.
What about fava beans?
Favism is real and it is not a drug question, so it sits outside CPIC's remit and outside this table. It is one of the reasons the deficiency was identified in the first place, and avoiding fava beans remains standard advice for people known to be deficient.
I am a woman with one deficient copy. Am I affected?
Possibly, and the honest answer is that genotype alone will not tell you how much. Random X-inactivation means the proportion of your red cells that are deficient varies between individuals, which is why CPIC has a variable category at all and why it points to enzyme measurement.