Dapsone-Induced Methemoglobinemia with a 'Normal' G6PD Result

The clinical threshold is met: symptomatic methemoglobinemia at 24% with a normal PaO2 is ordinarily an indication for methylene blue 1-2 mg/kg IV, repeated once if needed.

The pharmacology that makes this a judgment call: methylene blue is not itself the reductant. It is reduced by NADPH-methemoglobin reductase (NADPH-cytochrome b5 reductase pathway) to leukomethylene blue, which then non-enzymatically reduces methemoglobin. NADPH supply depends on the hexose monophosphate shunt, i.e., on G6PD. In severe G6PD deficiency the drug is ineffective and, being an oxidant, can precipitate frank hemolysis. Doses above ~7 mg/kg cause methemoglobinemia themselves.

Why a 'normal' 6.8 U/g Hb does not settle the question:

1. Reticulocytosis (6.1%) - young red cells are enzyme-rich, so a hemolytic episode can mask a deficient genotype and produce a falsely normal value.
2. Transfusion or recent hemolysis can shift the measured activity.
3. The result is close to the lower reference limit, and reference intervals are not calibrated for this decision.

Judgment: treat the hypoxia, but do not treat blind. Give methylene blue only with a low threshold to stop if hemolysis worsens; if deficiency is a real possibility and the patient is unstable, use exchange transfusion (removes both oxidant drug and methemoglobin) or hyperbaric oxygen rather than repeated methylene blue doses, and consider ascorbic acid as an adjunct (slow, weak). Stop dapsone and benzocaine immediately. Recheck G6PD after reticulocytes normalise.

Anemia mechanism. Dapsone is N-hydroxylated (CYP3A4/2C9) to dapsone hydroxylamine, which co-oxidises hemoglobin iron to Fe3+ while generating reactive oxygen species; the same metabolites cause oxidative damage to the red cell membrane, Heinz body formation, and extravascular hemolysis. Low haptoglobin, raised LDH and bilirubin, and 6.1% reticulocytes indicate a hemolytic process, not simply a dyshemoglobinemia. Benzocaine is an additional oxidant and the combination is additive. In G6PD deficiency the antioxidant (NADPH/glutathione) reserve is low, so hemolysis occurs at lower oxidant load - which is why the G6PD result deserves skepticism, not dismissal.

Pulse oximetry. The device compares absorbance at two wavelengths and assumes only oxy- and deoxyhemoglobin are present. Methemoglobin absorbs similarly at both wavelengths, driving the ratio toward ~85% regardless of true saturation; the reading is therefore a fixed artifact, not a measurement. Only multi-wavelength co-oximetry quantifies methemoglobin. The normal PaO2 with the low saturation is the classic 'saturation gap.'

Practical point: the degree of anemia here is itself a reason to stop both oxidant drugs and to treat the methemoglobinemia aggressively rather than relying on drug cessation alone.

Recommendation: do not simply restart at 100 mg/day. The prior episode was clinically significant (24% methemoglobin plus hemolysis) and a rechallenge carries a recurrence risk, amplified if G6PD status is truly deficient.

Sequence of reasoning:

1. Establish G6PD status definitively once hemolysis has resolved and reticulocytes have normalised; genotype or repeat quantitative testing is preferable to a single borderline value.
2. If deficiency is confirmed, dapsone is relatively contraindicated; pursue alternatives with the dermatologist (e.g., sulfasalazine/sulfapyridine are also oxidants, so the alternatives are not automatically safer).
3. If G6PD is normal, a lower dose with cimetidine (which inhibits the CYP-mediated hydroxylamine formation) is a recognised strategy, but it is an off-label risk-management maneuver, not a licence to abandon monitoring.

Monitoring if rechallenged: baseline CBC, reticulocyte count, bilirubin, LDH, haptoglobin and methemoglobin; repeat at 1-2 weeks, then periodically; patient education about cyanosis, dyspnea, dark urine, and fatigue; and removal of all other oxidants including topical benzocaine and nitrite-containing products. Document the reasoning - 'normal G6PD' during a hemolytic episode is not a defensible clearance.