Erythrocyte disorders:
Anemias due to increased destruction of erythrocytes with enzyme deficiencies

Glucose-6-phosphate dehydrogenase deficiency


Ernest Beutler

[from Williams' HEMATOLOGY ]

Please note that all references that appear in the original paper have been omitted here. This should be corrected in the future. In the meantime, please consult the comprehensive G6PD Deficiency bibliography (about 550 kb).

Definition and history

Glucose-6-phosphate dehydrogenase (G-6-PD) deficiency is a group of hereditary abnormalities in which the activity of the erythrocyte enzyme G-6-PD is markedly diminished. Such a deficiency may result in hemolytic anemia, particularly after the administration of drugs, during infections, possibly during diabetic acidosis, and in the neonatal period.

The recognition of G-6-PD deficiency was the direct result of investigations of the hemolytic effect of the antimalarial drug primaquine, carried out in the early 1950s and described in some detail elsewhere. These early studies defined G-6-PD deficiency as a hereditary sex-linked enzyme deficiency that affected primarily the erythrocytes, older cells being more severely affected than newly formed ones. They showed that this enzyme deficiency was most prevalent in individuals of African, Mediterranean, and Oriental ethnic origins, but that it could be found in virtually any population. The common (polymorphic) forms of G-6-PD deficiency were found to be associated with anemia only under conditions of stress. Functionally more severe forms occurred sporadically and were associated with nonspherocytic hemolytic anemia. Electrophoretic variants of G-6-PD that are valuable in studies of gene regulation and population biology have been found.

Etiology and pathogenesis


The gene for G-6-PD consists of 13 exons and 12 introns spread over a region of 100 kb on the X chromosome. A G-6-PD-like locus, possibly a pseudogene, has been identified on chromosome 17. G-6-PD cDNAs from normal subjects and those with some mutations have been sequenced. A particularly long noncoding sequence is present at the 5' end. Methylation of certain cytidines at the 3' end is believed to have a regulatory function.


The "normal" enzyme is designated as G-6-PD B. It represents the most common type of enzyme encountered in all the population groups that have been studied. G-6-PD deficiency results from the inheritance of any one of a large number of abnormalities of the structural gene that codes the amino acid sequence of the enzyme G-6-PD. Among persons of African descent a mutant enzyme with normal activity is prevalent. Known as G-6-PD A+, it migrates electrophoretically more rapidly than the normal B enzyme. A single amino acid substitution of Asn–Asp has been identified both by "fingerprinting" and by DNA sequencing. In the case of the common deficient African A– and in Mediterranean mutations, the abnormal enzyme formed may be synthesized in normal or near-normal quantity but has decreased stability in vivo. The amount of enzyme antigen in the red cells declines concurrently with enzyme activity. This suggests that the mutant protein in these variants is rendered unusually sensitive to proteolysis in the environment of the erythrocyte. Other mutations also result in the formation of enzyme molecules with decreased enzyme activity and with altered kinetic properties, some of which may render them functionally inadequate. For example, G-6-PD Oklahoma manifests marked decrease in its affinity for the substrates glucose 6-phosphate and NADP, and G-6-PD Manchester and Tripler are abnormally sensitive to the inhibitory effect of NADPH. Various combinations of abnormal properties occur, and over 350 variants of this enzyme, of which over 275 are probably unique, have been described (Table 58-1 [not offered in this online version but may be viewed here ]). Detailed biochemical characteristics have been tabulated.

Figure 58-1 is a semischematic representation of the biochemical properties of two of the more common variants. The A– type of G-6-PD is the most common clinically significant type of abnormal G-6-PD in the American black population. The red cells contain only 5 to 15 percent of the normal amount of enzyme activity. The mobility of the enzyme present is rapid and is indistinguishable from that of the A+ variant in conventional electrophoretic systems. The fact that these two electrophoretically rapid variants are common in African populations is not a coincidence. Sequence analysis of G-6-PD A+ and G-6-PD A– have shown that these two mutations have a common nucleotide substitution at nucleotide 376 that evidently accounts for the rapid electrophoretic mobility (see Fig. 58-3 [not offered in this online version]).

Most samples with G-6-PD A– manifest an additional mutation at nucleotide 202 that accounts for its in vivo instability. Less commonly, the additional mutation is at a different site. Thus, it is evident that G-6-PD A– arose in an individual who already had the G-6-PD A+ mutation.

Among Caucasian populations the most common variant s enzyme was originally designated G-6-PD Mediterranean. The enzyme activity of the red cells of individuals who have inherited this abnormal gene is barely detectable. Many similar but distinct variants exist in the Mediterranean region.


The life-span of G-6-PD-deficient red cells is shortened under many circumstances, such as during drug administration, infection, and the neonatal period. The exact reason for this is not known. Drug-induced hemolysis in G-6-PD-deficient cells is generally accompanied by the formation of Heinz bodies, particles of denatured hemoglobin and stromal protein (see Chap. 30), formed only in the presence of oxygen. The mechanism by which Heinz bodies are formed and become attached to red cell stroma has been the subject of considerable investigation and speculation. Exposure of red cells to certain drugs results in the formation of low levels of hydrogen peroxide as the drug interacts with hemoglobin. In addition, some drugs may form free radicals that oxidize GSH without the formation of peroxide as an intermediate. The formation of free radicals of GSH through the action of peroxide or by the direct action of drugs may be followed by either oxidation of GSH to the disulfide form (GSSG) or complexing of the glutathione with hemoglobin to form a mixed disulfide. Such mixed disulfides are believed to form initially with the sulfhydryl group of the b 93 position of hemoglobin. The mixed disulfide of GSH and hemoglobin is probably unstable and undergoes conformational changes exposing interior sulfhydryl groups to oxidation and mixed disulfide formation. Chain separation into free a and b chains also occurs. Phenylhydrazine-like drugs have also been shown to form a hemochromogen directly with hemoglobin, a complex forming between the iron of ferriheme and the nitrogen bound to the benzene ring of the drug. Once such oxidation has occurred, hemoglobin is irreversibly denatured and will precipitate as Heinz bodies. Normal red cells can defend themselves to a considerable extent against such changes by reducing GSSG to GSH and by reducing the mixed disulfides of GSH and hemoglobin through the glutathione reductase reaction. However, the reduction of these disulfide bonds requires a source of NADPH. Since G-6-PD-deficient red cells are unable to reduce NADP+ to NADPH at a normal rate, they are unable to reduce hydrogen peroxide or the mixed disulfides of hemoglobin and GSH. When such cells are challenged by drugs, they form Heinz bodies more readily than do normal cells. Cells containing Heinz bodies encounter difficulty in traversing the splenic pulp and are relatively rapidly eliminated from the circulation. The metabolic events that may lead to red cell damage and eventually destruction are summarized in Fig. 58-2 [not offered in this online version].

The formation of methemoglobin frequently accompanies the administration of drugs that have the capacity to produce hemolysis of G-6-PD-deficient cells. The heme groups of methemoglobin become detached from the globin more readily than do the heme groups of oxyhemoglobin. It is not clear whether methemoglobin formation plays an important role in the oxidative degradation of hemoglobin to Heinz bodies or whether formation of methemoglobin is merely an incidential side effect of oxidative drugs.

Icterus neonatorum in G-6-PD deficiency is probably due both to some shortening of red cell life-span and to inadequate processing of bilirubin by the immature liver of G-6-PD-deficient infants. Anemia does not appear to be present in these infants, and this suggest that liver dysfunction plays a greater role. Limited data available on liver G-6-PD in deficient adults suggest that a considerable degree of deficiency may be present, and if such a deficiency is also present in infants it may play a role in impairing the borderline ability of even normal infant liver to catabolize bilirubin.

The mechanism of hemolysis induced by infection or occurring spontaneously in G-6-PD-deficient subjects is even less well understood. It has been suggested that the generation of hydrogen peroxide by phagocytizing leukocytes may play a role in this type of hemolytic reaction.

Substances capable of destroying red cell GSH have been isolated from fava beans. Favism occurs only in G-6-PD-deficient subjects, but not all individuals in a particular family may be sensitive to the hemolytic effect of the beans. Nonetheless, some tendency toward familial occurrence has suggested the possibility that an additional, genetic factor may be active. The observation of increased excretion of glucaric acid has led to the suggestion that a defect in glucuronide formation might be present. Immunologic factors do not seem to play a role in favism. Increased levels of red cell calcium and consequent "cross-bonding" of membrane so that facing inner membrane surfaces are tightly bonded together may play a role in the destruction of red cells.

Mode of inheritance

The gene determining the structure of G-6-PD is carried on the X chromosome: Inheritance of G-6-PD deficiency is sex-linked. For this reason, the defect is fully expressed in affected males and is never transmitted from father to son, but only from mother to son. In females, only one of the two X chromosomes in each cell is active (see Chap. 9). Consequently, female heterozygotes for G-6-PD deficiency have two populations of red cells; deficient cells and normal cells. The proportion of deficient to normal cells may vary greatly. Some heterozygous females appear to be entirely normal; others appear to be fully affected. The marked variability of expression of G-6-PD deficiency of heterozygotes is the result of certain features of the X-inactivation process. Because the process of inactivation is random, in some instances more of the maternally derived or more of the paternally derived X chromosomes may escape inactivation. More important, perhaps, is the fact that the clones of cells in which the maternally derived X is active, on the one hand, or in which the paternally derived X is active, on the other, may have a proliferative advantage. In the many cell generations between the time of X inactivation and maturity, even a small selective advantage of one set of clones over the other would result in marked disparity between the number of normal and deficient cells. The marked variability of the ratio of G-6-PD-deficient to normal red cells in the circulation of female heterozygotes, then, accounts for the marked differences in expression of the deficiency in such individuals.


The prevalence of G-6-PD deficiency among white populations ranges from less than 1 in 1000 among northern European populations to 50 percent of the males among Kurdish Jews. G-6-PD deficiency is also found among certain Chinese populations and in southeast Asia. Several variants appear to be common in Asian populations. G-6-PD deficiency of the A– type is very common in West Africa, and in the United States the incidence among black males is approximately 11 percent. Approximately 16 percent of American black males carry the nondeficient G-6-PD A+ gene. The distribution of G-6-PD deficiency among various population groups has been presented in detail elsewhere.

The high-frequency of G-6-PD-deficient genes in many populations implies that G-6-PD deficiency confers some sort of selective advantage. The suggestion that resistance to malaria could account for the frequency of G-6-PD deficiency was supported by studies in heterozygotes for G-6-PD A– that showed a higher degree of infestation of G-6-PD-sufficient cells than of G-6-PD-deficient cells. In vitro, too, G-6-PD deficiency inhibits that growth of the malaria parasite. The advantage conferred by G-6-PD deficiency seems to be limited to female heterozygotes. It has been proposed that a plasmodial G-6-PD is induced after the parasite has passed through several cycles in G-6-PD-deficient cells.

It has been suggested that a higher prevalence of G-6-PD deficiency in individuals with sickle cell disease than in the general black population reflects a favorable effect of the enzyme deficiency on the clinical course of the sickling disorders. However, it seems that the increased prevalence of G-6-PD deficiency in patients with sickle disease may merely result from the markedly heterogeneous genetic composition of American blacks; those with more African genes are more likely to inherit sickle hemoglobin and G-6-PD A–. Similar factors may be responsible for the slight excess of G-6-PD deficiency observed among sicklers in Arab populations.

Clinical features

Most G-6-P-deficient persons never suffer any clinical manifestations from this common genetic trait. The major clinical consequence of G-6-PD deficiency is hemolytic anemia. Usually the anemia is episodic, but some of the unusual variants of G-6-PD may cause nonspherocytic congenital hemolytic disease (see below). In general hemolysis is associated with stress, most notably drug administration, infection, the newborn period, and, in certain individuals, exposure to fava beans.


A large number of drugs and other chemicals that may have the capacity to precipitate hemolytic reactions in G-6-PD-deficient individuals are listed in Table 58-2. Some drugs, such as chloramphenicol, may induce mild hemolysis in people with severe, Mediterranean-type G-6-PD deficiency but not in those with the milder A– or Canton type of deficiency. Furthermore, it appears that different individuals with the same G-6-PD variant experience a difference in the severity of their reaction to the same drug. For example, red cells from a single G-6-PD-deficient individual were hemolyzed in the circulation of some recipients who were given thiazolsulfone but their survival was normal in the circulation of others. Sulfamethoxazole, which was clearly hemolytic in experimental studies, does not appear to be a common cause of hemolysis in a clinical setting. Undoubtedly, individual differences in the metabolism and excretion of drugs influence the extent to which G-6-PD-deficient red cells are destroyed.


TABLE 58-2: Drugs and chemicals that should be avoided by persons with G-6-PD deficiency

Acetanilid Phenylhydrazine
Doxorubicin Primaquine
Furazolidone (Furoxone) Sulfacetamide
Methylene Blue Sulfamethoxazole (Gantanol)
Nalidixic acid (NeGram) Sulfanilamide
Naphthalene Sulfapyridine
Niridazole (Ambilhar) Thiazolesulfone
Nitrofurantoin (Furadantin) Toluidine blue
Phenazopyridine (Pyridium) Trinitrotoluene (TNT)

Typically, an episode of drug-induced hemolysis in G-6-PD-deficient individuals begins 1 to 3 days after drug administration is initiated. Heinz bodies appear in the red cells, and the hemoglobin concentration begins to decline rapidly. As hemolysis progresses, Heinz bodies disappear from the circulation, presumably as they or the erythrocytes that contain them are removed by the spleen. In severe cases abdominal or back pain may occur. The urine may turn dark – even black. Within 4 to 6 days, there is generally an increase in the reticulocyte count, except in instances in which the patient has received the drug in treatment of an active infection. Because of the tendency of infections and certain other stressful situations to precipitate hemolysis in G-6-PD-deficient individuals, many drugs have been incorrectly implicated as a cause. Other drugs, such as aspirin, have appeared on many lists of proscribed medications because very large doses had the capability of slightly reducing the red cell life-span. It is important to recognize that such drugs, listed in Table 58-3 , do not produce clinically significant hemolytic anemia. Advising patients not to ingest these drugs not only may deprive patients of potentially helpful medications but also will weaken their confidence in the advice that they have received. Most G-6-PD-deficient patients, after all, have taken aspirin without untoward effect and are likely to distrust an advisor who counsels them that the ingestion of aspirin would have catastrophic effects.


TABLE 58-3: Drugs which can probably safely be given in normal therapeutic doses to G-6-PD-deficient subjects without nonspherocytic hemolytic anemia

Acetaminophen (paracetamol, Tylenol, Tralgon, hydroxyacetanilide) Phenylbutazone
Acetophenetidin (phenacetin) Phenytoin
Acetylsalicylic acid (aspirin) Probenecid (Benemid)
Aminopyrine (Pyramidon, amidopyrine) Procain amide hydrochlonde (Pronestyl)
Antazoline (Antistine) Pyrimethamine (Daraprim)
Antipyrine Quinidine
Ascorbic acid (vitamin C) Quinine
Benzhexol (Artane) Streptomycin
Chloramphenicol Sulfacytine
Chlorguanidine (Proguanil, Paludrine) Sulfadiazine
Chloroquine Sulfaguanidine
Colchicine Sulfamerazine
Diphenhydramine (Benadryl) Sulfamethoxypyridazine (Kynex)
Isoniazid Sulfisoxazole (Gantrisin)
L-Dopa Trimethoprim
Menadione sodium bisulfite (Hykinone) Tripelennamine (pyribenzamine)
Menapthone Vitamin K
p -Aminobenzoic acid

In the A– type of G-6-PD deficiency, the hemolytic anemia is self-limited because the young red cells produced in response to hemolysis have near-normal G-6-PD levels and are relatively resistant to hemolysis. The hemoglobin level may return to normal even while the same dose of drug that initially precipitated hemolysis is administered. In contrast, hemolysis is not self-limited in the more severe Mediterranean type of deficiency.


Anemia has often developed rather suddenly in G-6-PD deficient individuals within a few days of onset of a febrile illness. The anemia is usually relatively mild, with a decline in the hemoglobin concentration of 3 or 4 g/dl. Hemolysis has been noted particularly in patients suffering from pneumonia and in those with typhoid fever. The fulminating form of the disease occurs particularly frequently among G-6-PD-deficient patients who are infected with Rocky Mountain spotted fever. Jaundice is not a prominent part of the clinical picture, except where hemolysis occurs in association with infectious hepatitis. In that case it can be quite intense. Presumably because of the effect of the infection, reticulocytosis is usually absent, and recovery from the anemia is generally delayed until after the active infection has abated.

Diabetic ketoacidosis has usually been considered a cause of hemolysis in G-6-PD deficiency, but a review of 36 episodes of diabetic ketoacidosis in G-6-PD-deficient subjects yielded only 10 in whom hemolysis occurred, and these were all associated with infection or drug ingestion. It has been suggested that hypoglycemia may precipitate hemolysis.


Favism is potentially one of the gravest clinical consequences of G-6-PD deficiency. It occurs much more commonly in children than in adults. The onset of hemolysis may be quite sudden, having been reported to occur within the first hours after exposure to fava beans. More commonly the onset is gradual, hemolysis being noticed 1 to 2 days after ingestion of the beans. Occasional hemolysis has been reported to occur after ingestion of other foodstuffs such as unripe peaches. The urine becomes red or quite dark, and in severe cases shock may develop within a short time.


Icterus neonatorum with no evidence of immunologic incompatibility occurs in some infants with G-6-PD deficiency. The jaundice may be quite severe and, if untreated, may result in kernicterus. Thus G-6-PD deficiency is a preventable cause of mental retardation, and this aspect of the disorder has considerable public health significance. An increased incidence of neonatal icterus has been observed in Mediterranean infants with G-6-PD deficiency, and among the Chinese. It seems to occur quite rarely among neonates with the A– type Of enzyme deficiency in the United States, but some cases have been reported in G-6-PD-deficient infants in Africa. The cause of the difference is unknown, but it may be related to some environmental factor such as vitamin E intake. As noted above, anemia is mild or absent in these infants, and hepatic dysfunction may play a major role in the pathogenesis of the jaundice.


Some of the rare types of G-6-PD deficiency are associated with hereditary nonspherocytic hemolytic anemia. Occasionally, patients with the common, Mediterranean type of defect have also been found to have this disorder. The reason some individuals with the Mediterranean type of enzyme deficiency have chronic hemolysis while the majority have hemolysis only under conditions of stress is not clear; it is possible that hemolysis is due to some as yet undefined associated abnormality.

Hereditary nonspherocytic hemolytic anemia due to G-6-PD deficiency is usually first noted during infancy or childhood. In some instances, neonatal jaundice has been present. Hemolysis is often exacerbated by febrile illnesses or by the administration of drugs. Splenomegaly is commonly present.


In the common variants of G-6-PD, such as G-6-PD A– and Mediterranean, and even in most of the severely deficient variants, there is usually no demonstrated defect in leukocyte number or function. However, there have been reports of isolated instances of leukocyte dysfunction associated with rare, severely deficient variants of G-6-PD. Patients with G-6-PD deficiency do not have a bleeding tendency, and studies of platelet function have yielded conflicting results. Occasionally, cataracts have been observed in patients with variants of G-6-PD that produce nonspherocytic hemolytic anemia, and it has even been suggested that the incidence of senile cataracts may be increased with G-6-PD deficiency. Although claims have been made that an association exists between various kinds of G-6-PD deficiency and cancer the data are not convincing, and a detailed investigation of hematologic malignancies in patients with G-6-PD Mediterranean show no effect.

Laboratory features

In the absence of hemolysis, the light-microscopic morphology of G-6-PD-deficient red cells appears to be normal. Differences in the texture of the stroma have been observed under the electron microscope. Even during hemolytic episodes, morphologic changes are not striking. When a hemolytic drug has been administered, Heinz bodies (Chap. 30) develop in the erythrocytes immediately preceding and in the early phases of hemolysis. If the hemolytic anemia is very severe, spherocytosis and red cell fragmentation may be seen in the stained film. Although "bite cells" have been noted in the blood of some patients undergoing drug-induced hemolysis, they were not G-6-PD-deficient. Varying degrees of hyperbilirubinemia may be evident. As the hemoglobin level of blood falls, reticulocytosis occurs and polychromasia is seen on the stained smear. No consistent changes occur in platelets or white cells.

Diagnosis of G-6-PD deficiency depends on the demonstration of decreased enzyme activity through either a quantitative assay or a screening test. Assay of the enzyme is generally carried out by measuring the rate of reduction of NADP+ to NADPH in an ultraviolet spectrophotometer. Several acceptable visual screening tests have been described, and the prepared reagents for carrying out some of these procedures are commercially available. The fluorescent procedure described in Chap A11 is the most satisfactory screening test. Although detection of enzyme deficiency in the healthy, fully affected (hemizygous) male presents no problem and can be achieved readily through either assay or screening tests, difficulties arise when a patient with G-6-PD deficiency of the A– type has undergone a hemolytic episode. As the older, more enzyme-deficient cells are removed from the circulation and are replaced by young cells, the level of the enzyme begins to increase toward normal. Under such circumstances, suspicion that the patient may be G-6-PD-deficient should be raised by the fact that enzyme activity is not increased, even though the reticulocyte count is elevated. Centrifugation of the blood followed by testing of the most dense (oldest) red cells has been employed as a means for the detection of G-6-PD deficiency in persons with the A– defect who have recently undergone hemolysis. It is helpful to carry out family studies or to wait until the circulating red cells have aged sufficiently to betray their lack of enzyme.

Even greater difficulties are encountered in attempting to diagnose the heterozygous enzyme deficiency state: The presence of a population of normal red cells coexisting with the deficient cells (see Chap. 9) may mask the enzyme deficiency when screening tests are used. Even enzyme assays on heterozygous females may frequently be in the normal range. Here methods that depend upon histochemical demonstration of individual red cell enzyme activity may be useful. In addition, the ascorbate-cyanide test, in which screening is carried out on a whole cell population rather than on a lysate, may be more sensitive than the other screening procedures.

The identification of specific G-6-PD variants requires the use of relatively sophisticated biochemical techniques. The enzyme must be partially purified and then its Km for NADP+ and glucose 6-phosphate, utilization of substrate analogs, pH optima, and electrophoretic mobility must be determined in standard systems. Detailed characterization of G-6-PD variants is of value chiefly in appraising the role of G-6-PD deficiency in the etiology of nonspherocytic hemolytic anemia. For example, if a black male with chronic hemolysis is found to have a common A– variant of G-6-PD, it is necessary to seek elsewhere for a source of hemolysis. On the other hand, if an unusual thermolabile variant is found, it is more likely that the G-6-PD deficiency plays an etiologic role in the hemolytic process.

Differential diagnosis

Drug-induced hemolytic anemia due to G-6-PD deficiency is similar in its clinical features and in certain laboratory features to drug-induced hemolytic anemia associated with unstable hemoglobins (see Chap. 61). Other enzyme defects affecting the pentose-phosphate shunt, such as a deficiency of GSH synthetase, may also mimic G-6-PD deficiency (see Chap. 59). The hemoglobinopathies can be ruled out by performing a stability test and hemoglobin electrophoresis (see Chaps. A7 and A8). Both of these are normal in G-6-PD deficiency. Some of the screening tests, particularly the ascorbate-cyanide test, may give positive results in the above-named disorders, but a G-6-PD assay or the fluorescent screening test will be positive only in G-6-PD deficiency.


G-6-PD-deficient individuals should avoid drugs that might induce hemolytic episodes (see Table 58-2).

If hemolysis occurs as a result of drug ingestion or infection, particularly in the milder A– type of deficiency, transfusion is not usually required. If, however, the rate of hemolysis is very rapid, as may occur, for example, in favism, transfusions of whole blood or packed cells may be useful. Good urine flow should be maintained in patients with hemoglobinuria to avert renal damage. Infants with neonatal jaundice due to G-6-PD deficiency may require exchange transfusion; in areas where G-6-PD deficiency is prevalent, care must be taken not to give G-6-PD-deficient blood to such newborns. Patients with hereditary nonspherocytic hemolytic anemia usually do not require any therapy. Splenectomy is generally ineffective, although some improvement has occasionally been reported following removal of the spleen. In most cases the anemia is not very severe, but in some instances frequent transfusions have been necessary. The antioxidant proper-ties of vitamin E have been tested in G-6-PD-deficient subjects, and it has been reported that a slight but statistically significant reduction in hemolysis was observed. These results could not be confirmed in other studies.

Course and prognosis

Hemolytic episodes in the A– type of deficiency are usually self-limited, even if drug administration is continued. This is not the case in the more severe Mediterranean type of deficiency. In patients with hereditary nonspherocytic hemolytic anemia due to G-6-PD deficiency, gallstones may occur. During periods of infections or drug administration anemia may increase in severity. Otherwise, the hemoglobin level of affected subjects remains relatively stable.

Nearly all patients with drug- or infection-induced hemolysis recover uneventfully. Favism must be considered, by comparison, a relatively dangerous disease. Prior to the institution of modern hospital therapy, fatalities from favism were not uncommon.

In one large population study, a decreasing incidence of G-6-PD deficiency was noted with increasing age of the population, but no such change was observed in another. While age stratification might represent evidence of a shorter life-span for individuals with the A– deficiency, other factors are more likely explanations. For example, in black American society the more economically deprived segment of the population manifests both a higher incidence of African genes and a shorter life-span. This would cause a decreasing incidence of G-6-PD deficiency with advancing age, but there would be no cause-and-effect relation between G-6-PD deficiency and life-span. Examination of the health records of over 65,000 U.S. Veterans Administration males failed to reveal a higher frequency of any illness in G-6-PD-deficient subjects as compared to the frequency in G-6-PD-nondeficient subjects. In view of the benign nature of the common types of G-6-PD deficiency, community-based population screening is not to be recommended. However, screening for G-6-PD deficiency of all patients admitted to the hospital may be useful in anticipating hemolytic reactions and in understanding them if they occur.