G6PD Assay: Procedure and Interpretation of Results

A G6PD assay measures the activity of glucose-6-phosphate dehydrogenase, an enzyme inside red blood cells that protects them from oxidative damage. The test comes in several forms, from a simple visual screening that takes minutes to a precise laboratory spectrophotometry method that gives a numerical result in international units per gram of hemoglobin. The choice of method, the timing of the blood draw, and a handful of biological variables all affect the number you get back, which makes interpreting results less straightforward than a simple “normal or deficient” label might suggest.

Why the Test Is Ordered

The most common reason clinicians order a G6PD assay is to determine whether a patient can safely take certain medications. The 8-aminoquinoline antimalarials, particularly primaquine and the newer tafenoquine, can trigger severe destruction of red blood cells in people who lack sufficient G6PD activity. Testing for G6PD deficiency is mandatory before prescribing tafenoquine.¹ Other drugs on the risk list include dapsone, certain sulfonamide antibiotics, and rasburicase.

G6PD deficiency is also a major risk factor for severe newborn jaundice and kernicterus, a form of permanent brain damage caused by dangerously high bilirubin levels.² The World Health Organization recommends screening all newborns for G6PD activity in countries where the deficiency is common.³ Beyond medication safety and newborn screening, clinicians sometimes order the assay to investigate unexplained hemolytic anemia, particularly when a patient reports episodes triggered by infections, fava beans, or oxidative stress.

The Fluorescent Spot Test

The most widely used screening method worldwide is the fluorescent spot test, or FST. A drop of blood is mixed with reagents that let G6PD do its job: the enzyme converts a molecule called NADP into its reduced form, NADPH. NADPH glows under ultraviolet light. If the blood spot fluoresces brightly, the person has normal G6PD activity. If it stays dark, activity is severely reduced. That is the entire procedure in basic terms, and it can be performed in settings with minimal laboratory infrastructure.

The FST performs well at identifying people with severe deficiency. At a threshold of 30% of normal enzyme activity, one study in neonates found the test had about 91% sensitivity and 97% specificity.⁴ A comparative analysis similarly reported robust sensitivity and specificity at and above that 30% threshold.⁵

The weakness is in the middle. Because the FST gives a binary yes-or-no answer, it struggles with intermediate enzyme levels. When activity hovers near the 30% cutoff, the degree of fluorescence is ambiguous, and the person reading the result has to make a subjective call.⁶ When the same neonatal study raised the threshold to 60% of normal activity, the FST’s sensitivity plummeted to about 29%, meaning it missed roughly seven out of ten individuals with intermediate deficiency and classified them as normal.⁴ That matters clinically, because people with intermediate activity can still experience hemolysis under the right conditions.

Quantitative Spectrophotometry

When a numerical result is needed, the standard laboratory method is UV spectrophotometry. The principle mirrors the FST but replaces the human eye with a machine. The enzyme in the blood sample converts NADP to NADPH, and a spectrophotometer measures the rate of change in light absorbance at 340 nanometers. The faster the absorbance rises, the more active the enzyme. That rate is then converted to international units per gram of hemoglobin (U/g Hb) using the patient’s hemoglobin concentration.⁷

A commonly cited normal range at 37°C is roughly 6.4 to 18.7 U/g Hb, though this can vary by assay kit and population.⁷ In practice, two commercial kits dominate the field: the Trinity Biotech assay and the Randox assay. A systematic review found that the median normal activity for the Trinity assay was about 10.0 U/g Hb, while Randox reported about 8.3 U/g Hb.⁸ The same review revealed a striking problem: across 13 studies using the Trinity assay, the standardized measure of “100% normal activity” varied from 5.7 to 12.6 U/g Hb.⁸ That kind of spread means that a result labeled “30% of normal” in one lab could correspond to a very different enzyme activity than “30% of normal” in another. It also means labs ideally need to establish their own population-based reference value rather than simply borrowing a textbook number.

When two commonly used spectrophotometric kits (Trinity and Pointe Scientific) were tested head-to-head on the same samples, their results correlated closely, with a mean difference of only 0.1 U/g Hb.⁹ So the kits themselves are reasonably interchangeable, but the variation in normal controls across different populations and laboratory conditions remains a genuine concern.

Point-of-Care and Biosensor Devices

Spectrophotometry requires a functioning laboratory, trained technicians, cold-chain reagents, and time. That rules it out in many of the tropical and subtropical regions where G6PD testing matters most. Newer point-of-care devices aim to close that gap by offering quantitative or semi-quantitative results from a finger-prick blood sample in minutes, without specialized equipment.¹⁰

The CareStart rapid diagnostic test, one of the earliest lateral-flow options, was found to be about 90% sensitive for severe deficiency and roughly 85% sensitive for moderate-to-severe deficiency compared to laboratory spectrophotometry. Its negative predictive value, the likelihood that a “normal” result truly means normal, was about 98%, which is reassuring when the goal is to identify people safe to take primaquine.¹¹ Biosensor-based devices from the same manufacturer provide a quantitative readout rather than a simple pass/fail, and pilot studies in malaria-endemic areas support their utility for screening, although researchers have cautioned that more head-to-head comparison with spectrophotometry is needed before widespread rollout.¹²

One advantage of handheld biosensors is reproducibility. A study comparing the CareStart Biosensor with spectrophotometry found no significant difference in results between different devices, different operators, or different testing sites with the biosensor. Spectrophotometry, by contrast, showed significant variation between sites even when the labs followed the same standardized procedures and used the same controls.¹³ For field programs that need consistent results across dozens of remote clinics, that kind of between-device consistency is a major practical benefit.

Interpreting Results and the WHO Classification

Once you have a number in U/g Hb, the next question is what it means. Until recently, the WHO grouped G6PD variants into five classes. Class I was the most severe (chronic hemolysis even without a trigger). Class II meant less than 10% of normal activity, and Class III meant 10 to 60%. In practice, the boundary between Class II and Class III turned out to be clinically unhelpful. Newborns carrying supposedly “milder” Class III variants like G6PD A- still developed dangerous jaundice, and heterozygous females with Class III variants were at real risk, too. The WHO has since merged the old Class II and Class III into a single category, now called Class B, which covers variants with less than 45% of normal activity.¹⁴ This reclassification better reflects the actual clinical risk.

In practical terms, most laboratories report results as a percentage of the local adjusted male median, or AMM, which represents what 100% normal activity looks like in that specific lab with that specific assay. A result below 30% of the AMM is generally considered deficient. A result between 30% and 70% is intermediate. Above 70% is normal. These thresholds matter especially for drug prescribing. Research in Indonesian women, for instance, found that setting the safety threshold at 70% of normal activity correctly identified over 90% of genetically confirmed G6PD-deficient females while excluding relatively few normal women.¹⁵

Why Female Results Are Harder to Read

The G6PD gene sits on the X chromosome. Males have one copy, so they are either fully deficient or fully normal. Females have two copies, and in each cell, one X chromosome is randomly switched off. The result is that a heterozygous woman ends up with a patchwork of normal and deficient red blood cells, and her overall measured enzyme activity can fall anywhere on a spectrum from fully deficient to fully normal.¹⁶ The exact mix depends on the randomness of X-inactivation. One study even identified a woman who was genetically heterozygous for a deficiency-causing variant but whose enzyme activity tested completely normal, presumably because her cells had almost entirely inactivated the X chromosome carrying the deficient gene.¹⁶

This is why the fluorescent spot test is particularly unreliable in women. A heterozygous female might produce enough NADPH in her normal red cells to generate a faint glow, earning a “normal” label despite having a substantial population of deficient cells that are vulnerable to hemolysis. Quantitative assays do better but still report a single averaged number that can mask the underlying mosaicism. A Thai study comparing multiple methods found that a cytochemical assay, which stains individual red blood cells to count the fraction that lack G6PD activity, came closest to the genetic-testing gold standard in prevalence estimates, detecting about 13% deficiency versus the 14% found by molecular analysis, while the fluorescent spot test caught only about 6%.¹⁷

Factors That Throw Off Results

Even a perfectly performed quantitative assay can give misleading results if certain conditions are present at the time of blood collection. Four well-recognized interferences deserve attention:

• Low hemoglobin: Because the result is expressed per gram of hemoglobin, severely anemic blood (hemoglobin below about 7 g/dL) can artificially pull the enzyme activity number down, making a normal person look deficient.
• Recent transfusion: Donor red blood cells carry donor-level G6PD activity, so a deficient patient who recently received blood may test falsely normal. The effect can persist for up to 60 days.
• High reticulocyte count: Young red blood cells contain more G6PD than older ones. After a hemolytic episode, the bone marrow floods the circulation with reticulocytes, which can push the measured activity into the normal range even in a deficient person.
• Very high white blood cell count: White blood cells and platelets have their own G6PD, and if they are not removed from the sample, they inflate the red-cell enzyme reading.

One institution’s protocol flagged any result as “provisional” when one of these interferences was present, meaning it required retesting under cleaner conditions before clinical decisions were made.¹⁸

The white-cell interference problem has been recognized for decades. A landmark study showed that after complete removal of leukocytes and platelets, the measured G6PD activity of truly deficient red cells dropped to levels far below what most labs typically report. Many of the biochemical properties historically attributed to the deficient enzyme in red cells actually belonged to contaminating white-cell and platelet enzyme.¹⁹ In modern practice, laboratories working with patients who have extremely high white cell counts (such as those with leukemia) are advised to deplete the white cells before running the assay. Testing of pre- and post-depleted samples has shown that the depletion step reduces measured G6PD activity by roughly 4 to 5 U/g Hb, enough to flip a truly deficient sample from “normal” to “deficient.”²⁰

Sample Storage and Dried Blood Spots

In resource-limited settings, dried blood spots on filter paper are sometimes used instead of fresh blood samples. These are cheap, easy to transport, and do not need immediate refrigeration, but the enzyme in a dried spot degrades over time, and how fast it degrades depends heavily on temperature and humidity. Samples stored in a refrigerator or freezer with desiccant retained over 90% of their initial enzyme activity for months. But at 37°C with high humidity, samples lost about 10% of activity within three days, and by 30 days under extreme conditions the enzyme had degraded enough to fall below quality-control thresholds.²¹ For tropical field programs, this means keeping dried blood spots cool and dry is not optional. Samples that sit in a hot, humid field bag for a week before reaching the lab may give falsely low results.

When Genetic Testing Adds Value

Enzyme activity testing tells you how much G6PD is working right now, but it cannot tell you which genetic variant is responsible, and it can be fooled by all the interferences discussed earlier. Genetic testing, which looks for known mutations in the G6PD gene, gives a permanent answer that does not change with transfusions, reticulocyte counts, or sample handling. However, it has its own limitation: if a patient carries a rare or novel variant not covered by the test panel, genetic testing can miss it entirely.

In research settings, combining both methods has proven valuable. A study looking at the relationship between G6PD status and malaria susceptibility found that classifying women by genotype alone missed the point, because many genetically heterozygous females had normal enzyme activity (thanks to favorable X-inactivation patterns) and were not truly at clinical risk. When the researchers restricted their analysis to women who were deficient by both genotype and enzyme activity, a protective association with malaria emerged clearly.²² The takeaway for clinical practice is that neither test alone captures the full picture in women, and using both in combination gives the most accurate classification.

The Geographic and Evolutionary Context

G6PD deficiency is not randomly distributed around the world. It is most common in populations with historical exposure to malaria, which makes evolutionary sense: red blood cells with low G6PD activity are harder for the malaria parasite to thrive in. A large study across Chinese ethnic groups confirmed this geographic pattern. The frequency of G6PD-deficient gene variants correlated with lower altitude and lower latitude, mirroring the historical range of malaria transmission. Populations living below 600 meters had significantly higher frequencies of deficiency than those above 1,000 meters.²³

The specific variants also differ by geography. In Thai populations, for example, the most common deficiency-causing variant was G6PD Viangchan, accounting for over 80% of cases, followed by G6PD Mahidol and G6PD Union.¹⁷ In African populations, G6PD A- predominates. In Mediterranean populations, G6PD Mediterranean is most prevalent. Knowing which variant is common in a region helps laboratories choose appropriate genetic test panels and can also influence how aggressively clinicians monitor patients, since some variants carry greater hemolytic risk than others despite producing similar enzyme-activity numbers.

What Happens in Deficient Red Blood Cells

Understanding why the test matters requires a brief look at what goes wrong inside a G6PD-deficient red blood cell when it encounters an oxidative trigger. The enzyme’s normal job is to regenerate NADPH, which in turn keeps glutathione in its reduced, protective form. Glutathione acts as a chemical shield, neutralizing reactive oxygen species before they can damage the cell membrane. Without enough G6PD, that shield runs out.

Research into the fava-bean compound divicine demonstrated the cascade precisely. The oxidative compound rapidly consumes glutathione in a one-to-one reaction. In normal cells, G6PD quickly restores the glutathione supply. In deficient cells, it cannot. Once glutathione is exhausted, a series of downstream events follow: protein sulfhydryl groups are oxidized, the cell membrane stiffens, and the spleen recognizes the damaged cells and destroys them.²⁴ The clinical result is hemolytic anemia, which can range from mild and self-limiting to life-threatening depending on the severity of the enzyme deficiency and the potency of the trigger.

Stored blood from G6PD-deficient donors shows measurable signs of this vulnerability even without an acute trigger. Compared to normal donor blood, deficient red cell concentrates accumulated significantly more markers of oxidative damage during routine storage, including higher concentrations of malondialdehyde, a byproduct of membrane lipid breakdown.²⁵ This has implications for blood banking: red cells donated by G6PD-deficient individuals may not store as well and could perform differently after transfusion, a question that blood services in high-prevalence regions are actively working through.