Clinical Characteristics of G6PD Deficiency in Infants WithMarked Hyperbilirubinemia

Yi-Hao Weng, MD* and Ya-Wen Chiu, PhDw

Summary: This study analyzes the clinical features of glucose-6-phosphate dehydrogenase (G6PD) deficiency in infants withmarked hyperbilirubinemia. We retrospectively assessed a cohort of413 infants with peak total serum bilirubin (TSB) level Z20mg/dLfrom 1995 to 2007. The prevalence of G6PD deficiency wasproportional to the level of peak TSB: 21.1% (81/383) in 20mg/dLto 29.9mg/dL, 45.5% (10/22) in 30mg/dL to 39.9mg/dL, and100% (8/8) in Z40mg/dL. Male sex was more common in G6PDdeficiency (75.8%). When compared with G6PD-normal infants,those with G6PD deficiency tended to have extreme hyperbilir-ubinemia (peak TSB level Z25mg/dL) and hemoglobin value<13 g/dL (P<0.001). Furthermore, mortality rate was signifi-cantly higher in G6PD-deficient infants (3.0%) than in the G6PD-normal counterparts (0.0%). Among 58 of the G6PD-deficientinfants who were followed for more than 12 months, 4 developedthe classic neurologic manifestations of kernicterus (6.6%). Thesefindings show that G6PD deficiency is an important risk factor ofextreme hyperbilirubinemia, death, and kernicterus.

Key Words: glucose-6-phosphate dehydrogenase deficiency, neonatal

hyperbilirubinemia, kernicterus

(J Pediatr Hematol Oncol 2010;32:11–14)

Glucose-6-phosphate dehydrogenase (G6PD) deficiencyis a worldwide hereditary disorder with the potential

for causing neonatal hyperbilirubinemia (NH).1–4 Destruc-tion of red blood cells by the contact with oxidative agents,such as mothballs, has been labeled as a factor leadingto NH.5,6 In addition, there is increasing evidence thatG6PD-deficient infants are at great risk for NH even in anenvironment free from agents that can induce hemolysis.4,7

The pathogenesis may involve genetic interactions.8,9 NHcarries a substantial threat for deleterious complications,including death and long-term neurologic impairments. Thetotal serum bilirubin (TSB) value has been used as a surro-gate index to evaluate the risk of irreversible sequences.10

Although there is no distinguished threshold of a safe TSBlevel to adopt, most physicians work with the assumptionthat infants with a TSB level >20mg/dL are vulnerable tosequelae.11 With the advent of therapeutic intervention,significant complications have become rare in recent years.

In this cohort study, the spectrums of G6PD-deficientinfants in relation to marked NH are evaluated. We illus-trate G6PD deficiency in severe NH by comparing itsprevalence within different levels of peak TSB. AlthoughNH with G6PD deficiency has been widely investigated,very little of the outcome has been studied.12 This surveyhas identified the risk factors contributing to irreversiblesequelae in infants with G6PD deficiency.

MATERIALS AND METHODS

Patients and Laboratory AnalysisPermission to collect the data was obtained from

the Institutional Review Board of Chang Gung MemorialHospital. Medical charts of infants admitted to theneonatal intensive care units at Chang Gung Children’sHospital from 1995 to 2007 with marked hyperbilirubine-mia (peak TSB value Z20mg/dL) were reviewed. Infantswith a direct bilirubin value/TSB value r15% wereincluded. Those with the following conditions that couldconfuse the neurologic outcomes were excluded: gestationalage less than 34 weeks, birth weight less than 2000 g,perinatal asphyxia, and congenital disorders of the centralnervous system. TSB values were measured in a clinicallaboratory with a Unistat bilirubinometer (CambridgeInstruments, Buffalo, NY). As described in an earlierstudy,4 the G6PD activity of red blood cells was determinedspectrophotometrically at 340 nm by the reduction ofNADP+ in the presence of glucose-6-phosphate. G6PDdeficiency was confirmed with the enzyme activity below12.5U/gm Hb.

Clinical and laboratory data were collected byexamining the medical charts. Any other possible etiologiescausing NH—such as infants of diabetic mothers, poly-cythemia, congenital hypothyroidism, spherocytosis, bac-terial infection (sepsis, urinary tract infection, omphalitis),gastrointestinal obstruction, breast feeding, cephalohema-toma, bruise, ABO incompatibility (defined as any bloodgroup A or B newborn of group O mother), and Rhincompatibility (defined as Rh-positive infants born to Rh-negative mothers)—were recorded. G6PD enzymatic activ-ity was examined in all infants with NH. Other routinelaboratory examinations included complete blood count,reticulocytes, blood smear, total and direct bilirubin value,blood and urine cultures, blood type, and direct Coombstest. Phototherapy was applied when peak TSB Z12mg/dLat 24 to 47 hours old, Z14mg/dL at 48 to 71 hours old,Z15mg/dL at 72 to 119 hours old, and Z17mg/dL atZ120 hours old. In addition, exchange transfusion (ET)was carried out as indicated: peak TSB Z15mg/dL at 24to 47 hours old, Z20mg/dL at 48 to 95 hours old, andZ25mg/dL at Z96 hours old. A routine examination forTSB values was carried out at 1, 4, and 12 hours after ET.Copyright r 2010 by Lippincott Williams & Wilkins

Received for publication April 3, 2009; accepted July 26, 2009.From the *Department of Pediatrics, Chang Gung Memorial Hospital,

Chang Gung University College of Medicine, Taoyuan; andwDivision of Health Policy Research and Development, Institute ofPopulation Health Sciences, National Health Research Institutes,Miaoli, Taiwan.

Supported in part by the National Health Research Institutes.Reprints: Yi-Hao Weng, MD, Department of Pediatrics, Chang Gung

Children’s Hospital, 5 Fuxin Street, Guishan 333, Taoyuan,Taiwan, ROC (e-mail: yihaoweng@adm.cgmh.org.tw).

ORIGINAL ARTICLE

J Pediatr Hematol Oncol � Volume 32, Number 1, January 2010 www.jpho-online.com | 11

Mortality associated with NH was defined as deathoccurring within 7 days after the development of markedhyperbilirubinemia without other known fatal causes.Kernicterus was defined as having 2 or more of thefollowing symptoms after follow-up for more than 1 year:(1) athetoid cerebral palsy, (2) gaze impairment, especiallyof upward gaze, (3) delayed developmental milestones, and(4) auditory disturbances.10

Statistical AnalysesThe statistical analyses were conducted using a

commercially available program (SPSS for Windows,version 12.0). Categorical variables were analyzed usingthe w2 test or Fisher exact test. Significance was definedas P<0.05. The TSB values related to ET were calculatedby the number of procedures rather than the number ofinfants.

RESULTSA total of 413 infants had peak TSB Z20mg/dL

during the 13-year study period. G6PD deficiency wasconfirmed in 99 cases. The prevalence of G6PD deficiencyin infants with peak TSB Z20mg/dL was 24.0% (99/413).Figure 1 illustrates the prevalence of G6PD deficiency bydifferent levels of peak TSB. The prevalence of G6PDdeficiency in the general population was adopted from apublished cohort of 42,110 neonates born in the samehospital between 1994 and 2001.4 In general, the prevalenceof G6PD deficiency increased proportionally with thelevel of peak TSB: 21.1% in 20 to 29.9mg/dL, 45.5% in30 to 39.9mg/dL, and 100% in Z40mg/dL. Among maleinfants, the prevalence of G6PD deficiency was 3.5% in thegeneral population, 23.9% in infants with peak TSB level of20 to 24.9mg/dL, 37.5% in 25 to 29.9mg/dL, 57.1% in 30to 39.9mg/dL, and 100% in Z40mg/dL. Among femaleinfants, the prevalence was 1.2% in the general population,12.7% in those with 20 to 24.9mg/dL, 15.2% in 25 to29.9mg/dL, 25% in 30 to 39.9mg/dL, and 100% in thosewith Z40mg/dL.

To identify the unique characteristics of G6PD defici-ency, we incorporated 10 variables (birth place, sex, birthweight, gestational age, delivery mode, peak TSB level, ageat peak TSB, hemoglobin value, management, and acuteoutcome) for the univariate analytic model (Table 1). Theanalysis showed discrepancies between G6PD-deficientand G6PD-normal infants in 5 categories: sex, hemoglobinvalue, peak TSB level, management, and acute outcome.Males were more common among G6PD-deficient infants(P<0.001). In addition, infants with G6PD deficiency weremore likely to have a hemoglobin value less than 13 g/dLcompared with infants without G6PD deficiency (P<0.001).Three among the G6PD-deficient infants (3.0%) died. Incontrast, no fatalities were seen in G6PD-normal infants.The mortality rate was significantly higher in G6PD-deficienct infants (P<0.05). Furthermore, G6PD-deficientinfants tended to have a high peak TSB level (P<0.001)and, therefore, required more ET (P<0.01) than G6PD-normal counterparts.

Among 99 G6PD-deficient infants, 43 had combinedfactors that may enhance hyperbilirubinemia. These factorsincluded breast feeding (n=20), blood group incompat-ibility (n=15), urinary tract infection (n=3), omphal-itis (n=2), cephalohematoma (n=5), gastrointestinalobstruction (n=2), congenital hypothyroidism (n=1),

polycythemia (n=1), spherocytosis (n=1), and massivebruise (n=1). There were no significantly different demo-graphic characteristics between infants with and withoutcombined factors (data not shown).

Double-volume ET was carried out in 37 G6PD-deficient infants, of which 6 required a second procedure.The TSB values after ET were not available in 2 patientsbecause of their death, leaving 41 procedures to be ana-lyzed. Before ET, the TSB level was 30.4±7.6mg/dL. At1 hour after ET, the TSB level significantly reduced to18.7±4.3mg/dL. Thereafter, at 4 (n=39) and 12 hours(n=36), there was no significant decrease in the TSB level(data not shown).

Table 2 displays the demographic and laboratorydata of G6PD-deficient infants by different levels of peakTSB. Death was more common in infants with a peak TSBlevel Z40mg/dL than in infants with a peak TSB level<40mg/dL (P<0.001). In addition, there was a significantcorrelation of peak TSB level with hemoglobin value(P<0.05). Infants with a higher peak TSB level were morelikely to have hemoglobin value less than 13 g/dL thaninfants with a lower peak TSB level. Other characteristics—including sex, birth place, birth weight, gestational age,delivery mode, G6PD activity, presence of combined factors,and age at peak TSB—carried no significant differencesbetween each level of peak TSB.

100

(1/1

)

(7/7

)

90

80

70

60 (8/1

4)

50

40 (18/

48)

30

20

(42/

17h6

)

(5/3

3)

(2/8

)

10

inci

denc

e of

G6P

D d

efic

ienc

y (

%)

(6/1

26)1

0

Peak total serum bilirubin (mg/dL)

FIGURE 1. Prevalence of glucose-6-phosphate dehydrogenase(G6PD) deficiency in different levels of peak total serum bilirubinand general population. Solid bars indicate female population.Empty bars indicate male population. n/n = number of G6PD-deficient infants/number of total infants.

Weng and Chiu J Pediatr Hematol Oncol � Volume 32, Number 1, January 2010

12 | www.jpho-online.com r 2010 Lippincott Williams & Wilkins

Among 58 G6PD-deficient infants who were followedfor more than 12 months, 4 were confirmed having theclinical manifestations of kernicterus. To determine the riskfactors for the development of irreversible sequelae (deathor kernicterus), we analyzed 10 variables—birth place, sex,birth weight, gestational age, delivery mode, G6PD activity,peak TSB level, age at peak TSB, hemoglobin value, andpresence of combined factors. The results showed thatinfants with sequelae (n=7; 3 death and 4 kernicterus) weremore likely to have a hemoglobin value less than 13 g/dL(P<0.05) and peak TSB value Z25mg/dL (P<0.01) thanthose without sequelae (n=54). In addition, all infantswith sequelae were born outside our hospital, which wasmore common than those without sequelae (P<0.05).

DISCUSSIONThis study depicts the clinical spectrum of G6PD-

deficient infants with peak TSB level Z20mg/dL. Bycomparing with G6PD-normal infants, we have disclosedthe unique properties of G6PD deficiency with marked NH.Our data showed that male sex was more common in G6PDdeficiency. This is not surprising as the G6PD gene isencoded in the X chromosome. In addition, we showed thatmore G6PD-deficient infants had a lower hemoglobinvalue, an index of hemolysis. Although genetic interactions

are alleged to induce NH in G6PD deficiency,8,9 destructionof red blood cells has still been observed in some circum-stances.6,13,14 For instance, G6PD-deficient infants whowere born outside medical centers were more susceptible tohemolysis.15,16 Moreover, our study identified that infantswith G6PD deficiency tended to have extreme NH, definedas a peak TSB value Z25mg/dL.17–19 We also furtherverified that lower hemoglobin value was closely related tohigher peak TSB level in G6PD-deficient infants. Thesefindings lead to the suggestion that extreme NH in G6PDdeficiency is mediated, at least in part, by hemolysis.

To our knowledge, this study is the first to illustratethat the prevalence of G6PD deficiency is proportionalto the level of peak TSB. In this study, more than onehalf of infants with peak TSB level Z30mg/dL wereG6PD-deficient. In the past, isoimmune hemolytic diseaseowing to blood group mismatch was the major cause ofextreme NH. With the introduction of immunoglobulin,the incidence of severe NH by blood group incompatibilityhas declined.20 In contrast, prevention of G6PD deficiency-related NH by Taiwan’s newborn screen has been fruit-less.21 Recent studies also reveal that G6PD deficiency ismore common in infants with NH.4,7,12,22 Their dataand ours highlight the fact that G6PD deficiency is animportant etiology of severe NH.

TABLE 2. Clinical Characteristics of 99 G6PD-deficient Infants byDifferent Levels of Peak TSB

Peak TSB Level (mg/dL)

Number (%)

20-24.9

n=58

25-29.9

n=23

30-39.9

n=10

40-50

n=8 P

Sex 0.770Male 42 (72.4) 18 (78.3) 8 (80.0) 7 (87.5)Female 16 (27.6) 5 (21.7) 2 (20.0) 1 (12.5)

Birth place 0.152Inborn 21 (36.2) 5 (21.7) 3 (30.0) 0 (0.00)Outborn 37 (63.8) 18 (78.3) 7 (70.0) 8 (100)

Delivery mode 0.203Cesarean section 19 (32.8) 7 (30.4) 0 (0.00) 2 (25.0)Vaginal delivery 39 (67.2) 16 (69.6) 10 (100) 6 (75.0)

Gestational age (wk) 0.702<37 16 (27.6) 2 (8.7) 3 (30.0) 0 (0.00)37-42 42 (72.4) 21 (91.3) 7 (70.0) 8 (100)

Birth weight (g) 0.357<2500 8 (13.8) 4 (17.4) 0 (0.00) 0 (0.00)Z2500 50 (86.2) 19 (82.6) 10 (100) 8 (100)

Combined factors 0.626With 35 (60.3) 12 (52.2) 4 (40.0) 5 (62.5)Without 23 (39.7) 11 (47.8) 6 (60.0) 3 (37.5)

G6PD activity(U/g Hb)

0.418

<1.25 18 (31.0) 3 (13.0) 3 (30.0) 2 (25.0)1.25-12.4 40 (69.0) 20 (87.0) 7 (70.0) 6 (75.0)

Age at peak TSB (d) 0.063<7 36 (62.1) 11 (47.8) 2 (20.0) 3 (37.5)7-30 22 (37.9) 12 (52.2) 8 (80.0) 5 (62.5)

Hemoglobin (g/dL) <0.05<13 20 (34.5) 11 (47.8) 5 (50.0) 7 (87.5)13-22 38 (65.5) 12 (52.2) 5 (50.0) 1 (12.5)

Acute outcome <0.001Death 0 (0) 0 (0) 0 (0) 3 (37.5)Survival 58 (100) 23 (100) 10 (100) 5 (62.5)

G6PD indicates glucose-6-phosphate dehydrogenase; TSB, total serumbilirubin.

TABLE 1. Demographic and Clinical Data in 99 G6PD-deficientand 314 G6PD-normal Infants With Peak TSB Value Z20 mg/dL

G6PD Status

Deficient,

n (%)

Normal,

n (%) P

Sex <0.001Male 75 (75.8) 170 (54.1)Female 24 (24.2) 144 (45.9)

Birth place 0.267Inborn 29 (29.3) 111 (35.4)Outborn 70 (70.7) 203 (64.6)

Delivery mode 0.173Cesarean section 28 (28.3) 68 (21.7)Vaginal delivery 71 (71.7) 246 (78.3)

Gestational age (wk) 0.498<37 21 (21.2) 57 (18.2)37-42 78 (78.8) 257 (81.8)

Birth weight (g) 0.135<2500 12 (12.1) 23 (7.3)Z2500 87 (87.9) 291 (92.7)

Peak TSB level (mg/dL) <0.00120-24.9 58 (58.6) 244 (78.7)25-29.9 23 (23.2) 58 (18.5)30-39.9 10 (10.1) 12 (3.8)40-50 8 (8.1) 0 (0)

Age at peak TSB (d) 0.128<7 52 (52.5) 192 (61.1)7-30 47 (47.5) 112 (38.9)

Hemoglobin (g/dL) <0.001<13 43 (43.4) 78 (24.8)13-22 56 (56.6) 236 (75.2)

Exchange transfusion <0.01None 62 (62.6) 249 (79.3)Once 31 (31.3) 57 (18.2)More than once 6 (6.1) 8 (2.5)

Acute outcome <0.05Death 3 (3.0) 0 (0)Survival 96 (97.0) 314 (100)

G6PD indicates glucose-6-phosphate dehydrogenase; TSB, total serumbilirubin.

J Pediatr Hematol Oncol � Volume 32, Number 1, January 2010 Neonatal Hyperbilirubinemia With G6pd Deficiency

r 2010 Lippincott Williams & Wilkins www.jpho-online.com | 13

Our study illustrated that infants with G6PD defi-ciency were at increased risk of mortality. We identifiedpeak TSB value Z40mg/dL as a risk factor for mortality,which suggests that the leading cause of death was bilirubinencephalopathy, not G6PD deficiency itself. The resultswere consistent with the known vulnerability of G6PD-deficient infants to irreversible sequelae.12,22–24 More-over, we further extended such inquiries by showing thatextreme NH and low hemoglobin value were 2 relevantrisk factors for irreversible sequelae. Our recent studyshowed similar findings in blood group incompatibility.20

It is well recognized that high TSB level is a risk factorfor kernicterus.5,6 Nevertheless, how anemia causes braininjury remains incompletely understood; probably becauseproducts from the destruction of red blood cells potentiatebilirubin encephalopathy.25

Our study has some limitations. One could questionthat only 60% of infants were followed for more than12 months. We believe they were representative, becausetheir demographic and laboratory data were similar withthose who were not followed regularly. Furthermore, wedid not exclude infants with any combined factors thatcould aggravate anemia, NH, or kernicterus. But ouranalysis has shown that infants with and without combinedfactors had similar background features.

In conclusion, G6PD deficiency is an important riskfactor for the development of severe NH and kernicterus.High TSB level and low hemoglobin value are 2 crucial riskfactors for deleterious consequences. Management shouldbe prompt to avoid irreversible sequelae for infants withG6PD deficiency.

REFERENCES

1. Kaplan M, Abramov A. Neonatal hyperbilirubinemia asso-ciated with glucose-6-phosphate dehydrogenase deficiency inSephardic-Jewish neonates: incidence, severity, and the effectof phototherapy. Pediatrics. 1992;90:401–405.

2. Tan KL. Glucose-6-phosphate dehydrogenase status andneonatal jaundice. Arch Dis Child. 1981;56:874–877.

3. Valaes T. Severe neonatal jaundice associated with glucose-6-phosphate dehydrogenase deficiency: pathogenesis and globalepidemiology. Acta Paediatr. 1994;394:58–76.

4. Weng YH, Chou YH, Lien RI. Hyperbilirubinemia in healthyneonates with glucose-6-phosphate dehydrogenase deficiency.Early Hum Dev. 2003;71:129–136.

5. Dennery PA, Seidman DS, Stevenson DK. Neonatal hyper-bilirubinemia. N Engl J Med. 2001;344:581–590.

6. Dhillon AS, Darbyshire PJ, Williams MD, et al. Massive acutehaemolysis in neonates with glucose-6-phosphate dehydrogenasedeficiency. Arch Dis Child Fetal Neonatal Ed. 2003;88:F534–F536.

7. Kaplan M, Hammerman C. Glucose-6-phosphate dehydrogen-ase deficiency: a potential source of severe neonatal hyper-bilirubinaemia and kernicterus. Semin Neonatol. 2002;7:121–128.

8. Weng YH, Chou YH, Cheng ML, et al. Increased hemeoxygenase-1 expression in glucose-6-phosphate dehydrogenasedeficient human fibroblasts. Pediatr Res. 2002;51:328A.

9. Kaplan M. Genetic interactions in the pathogenesis of neo-natal hyperbilirubinemia: Gilbert’s Syndrome and glucose-6-phosphate dehydrogenase deficiency. J Perinatol. 2001;21:S30–S34.

10. American Academy of Pediatrics Subcommittee on Hyper-bilirubinemia. Management of hyperbilirubinemia in thenewborn infant 35 or more weeks of gestation. Pediatrics.2004;114:297–316.

11. Watchko JF. Vigintiphobia revisited. Pediatrics. 2005;115:1747–1753.

12. Katar S. Glucose-6-phosphate dehydrogenase deficiency andkernicterus of South-East anatolia. J Pediatr Hematol Oncol.2007;29:284–286.

13. Kaplan M, Hammerman C, Vreman HJ, et al. Acute hemolysisand severe neonatal hyperbilirubinemia in glucose-6-phosphatedehydrogenase-deficient heterozygotes. J Pediatr. 2001;139:137–140.

14. Samanta S, Kumar P, Kishore SS, et al. Donor blood glucose 6-phosphate dehydrogenase deficiency reduces the efficacy ofexchange transfusion in neonatal hyperbilirubinemia. Pediatrics.2009;123:e96–e100.

15. Weng YH, Chou YH. Glucose-6-phosphate dehydrogenasedeficiency with hyperbilirubinemia in neonates born inside andoutside a tertiary hospital. Clin Neonat. 2003;10:50–53.

16. Chen SH, Lin KS, Lee TY, et al. Glucose-6-phosphatedehydrogenase deficiency and neonatal hyperbilirubinemia inChinese. Acta Paediatr Sinica. 1986;27:286–293.

17. Salas AA, Mazzi E. Exchange transfusion in infants withextreme hyperbilirubinemia: an experience from a developingcountry. Acta Paediatr. 2008;97:754–758.

18. Tiker F, Gulcan H, Kilicdag H, et al. Extreme hyperbilir-ubinemia in newborn infants. Clin Pediatr. 2006;45:257–261.

19. Newman TB, Xiong B, Gonzales VM, et al. Prediction andprevention of extreme neonatal hyperbilirubinemia in a maturehealth maintenance organization. Arch Pediatr Adolesc Med.2000;154:1140–1147.

20. Weng YH, Chiu YW. Spectrum and outcome analysis ofmarked neonatal hyperbilirubinemia with blood group incom-patibility. Chang Gung Med J. 2009;32:400–408.

21. Chuu WM, Lin DT, Lin KH, et al. Can severe neonataljaundice be prevented by neonatal screening for glucose-6-phosphate dehydrogenase deficiency? A review of evidence.Acta Paediatr Sinica. 1996;37:333–341.

22. Manning D, Todd P, Maxwell M, et al. Prospective surveillancestudy of severe hyperbilirubinaemia in the newborn in theUK and Ireland. Arch Dis Child Fetal Neonatal Ed. 2007;92:F342–F346.

23. Bhutani VK, Johnson L. Kernicterus in late preterm infantscared for as term healthy infants. Semin Perinatol. 2006;30:89–97.

24. Nair PA, Al Khusaiby SM. Kernicterus and G6PD deficiency:a case series from Oman. J Trop Pediatr. 2003;49:74–77.

25. Dawodu AH, Sutherland JM, Becker LE, et al. Influence ofheme on the kinetics of bilirubin binding. Pediatr Res. 1977;11:532A.

Weng and Chiu J Pediatr Hematol Oncol � Volume 32, Number 1, January 2010

14 | www.jpho-online.com r 2010 Lippincott Williams & Wilkins