Effect of neonatal reticulocytosis on glucose 6-phosphate dehydrogenase (G6PD) activity and G6PD deficiency detection: a cross-sectional study

Glucose 6-phosphate dehydrogenase (G6PD) deficiency is the most common hereditary blood disorder, affecting millions of people worldwide [1]. G6PD is a rate-limiting enzyme in the hexose monophosphate shunt (HMP) generating reduced nicotinamide adenine dinucleotide phosphate (NADPH), a coenzyme in glutathione metabolism, to defend against oxidative stress in red blood cells [2]. Patients with defective G6PD enzymes are generally asymptomatic unless they are exposed to infection or stimulants, such as fava beans and the antimalarial drug primaquine [1, 3, 4]. G6PD deficiency manifests as acute hemolytic anemia (AHA), hemoglobinuria and favism [1]. In Asians, severe cases in newborns with delayed diagnosis and treatment can develop cerebral jaundice and hyperbilirubinemia, resulting in life-threatening bilirubin neurotoxicity, kernicterus, and mental retardation [3‐5]. In Thailand, the prevalence of G6PD deficiency in male newborns is 10.7–17.0% and that of female newborns is 1.2–5.8% [6‐8]. Approximately 65% of severe neonatal jaundice and 21.2–22.1% of hyperbilirubinemia were coincidently observed in G6PD-deficient Thai newborns [6, 9]. Thus, the World Health Organization (WHO) has recommended routine screening for G6PD deficiency in newborns living in areas where the prevalence of G6PD deficiency is as high as 3–5% in males to avoid unwanted consequences [10].

G6PD deficiency is an X-linked recessive inborn error commonly observed in hemizygous males, whereas normal, intermediate, or grossly deficient phenotypes depending on a mosaicism of lyonization can be observed in heterozygous females [8]. Heterozygous females with partial G6PD activity may be misdiagnosed. A quantitative G6PD enzymatic activity assay based on a spectrophotometric technique is the reference assay to detect G6PD deficiency and G6PD intermediates [11]. It was suggested by previous studies that the cut-off values for G6PD deficiency and intermediate G6PD should be activity values less than 30% and 30% to 60% of the adjusted male median (AMM), respectively [11‐14]. The WHO has recommended that activity values less than 30% and 30%-80% of the normal median should be used instead [15]. Nevertheless, the differences in mosaicism, study techniques, and the reticulocyte/erythrocyte ratio may affect the interpretation of G6PD deficiency [5, 16, 17]. In particular, high levels of G6PD-rich reticulocytes, known as reticulocytosis, commonly observed in newborns may be an important interfering factor for the detection of G6PD deficiency in newborns, leading to false-negative results [18, 19].

Therefore, the present study aimed to investigate the influence of reticulocytes on newborn G6PD activity determined by different methods, including the fluorescent spot test (FST) and two quantitative assays (standard and automated UV enzymatic methods). The performance of the automated method was compared with that of FST and the standard quantitative method.

A total of 1,015 Thai newborns consisted of 502 males and 513 females. The demographic data of the newborns are summarized in Table 1. The mean gestational age and postpartum age of all subjects were 39.2 ± 18.0 weeks and 2.7 ± 0.8 days, respectively. The mean birth weight of male newborns was 3,179.5 ± 414.1 g, which was significantly higher than that of female newborns (3,061.4 ± 389.4 g; p < 0.001) (Table 1).

The aetiology of haemolytic disease of the foetus and newborn (HDFN) can be either immune or non-immune mediated. Blood group incompatibility between the mother and the fetus is the main cause of immune-mediated HDFN. The causes of nonimmune-mediated HDFN include enzyme abnormalities, RBC membrane problems, thalassemia, and hemoglobinopathies. In this study population, we found 8 newborns with ABO incompatibility, 44 newborns with thalassemia/hemoglobinopathies, and 1 coincidence of ABO incompatibility and thalassemia/hemoglobinopathies. The incidence of G6PD deficiency in the male and female populations of this study using three diagnostic techniques is shown in Table 1. The spectrophotometric assay was used as the gold standard method to verify the results of the other two techniques.

After the Kolmogorov–Smirnov and Shapiro–Wilk tests for normality, G6PD activity of neonatal males as measured by spectrophotometric and automated UV enzymatic assays showed a bimodal distribution (p < 0.001) and that of neonatal females showed a trimodal distribution of homozygous and heterozygous G6PD deficient and homozygous normal samples (p < 0.001) (Fig.1a-d). The median ± IQR of G6PD activity of males (10.1 ± 2.5 U/g Hb) and females (10.0 ± 2.7 U/g Hb) obtained from the spectrophotometric assay were not significantly different (p = 0.848 by the Mann–Whitney test). The results from the automated UV enzymatic method demonstrated that there was no significant difference in median G6PD activity between males (25.9 ± 7.8 U/g Hb) and females (25.6 ± 9.0 U/g Hb) (p = 0.375 by Mann–Whitney test). The G6PD reference values in newborn subjects determined by spectrophotometric assay were as follows: deficiency < 3.03 U/g Hb, intermediate 3.03—8.07 U/g Hb, and normal > 8.07 U/g Hb, whereas that determined by the automated UV enzymatic assay was as follows: deficiency < 7.78 U/g Hb, intermediate 7.78—20.75 U/g Hb, and normal > 20.75 U/g Hb (Fig. 1a-d and Table 2). The median values of G6PD activity of deficiency, intermediate, and normal males and females obtained from both assays are shown in Table 3.

G6PD deficiency is an important risk factor for severe hyperbilirubinemia and bilirubin neurotoxicity in newborns [6]. The prevalence of G6PD deficiency differs between geographic regions and ethnic groups. For example, the prevalence of G6PD-deficient males was 7.3% in Thailand, 8.1% in Lao PDR, 15.8% in Myanmar, 18.8% in Cambodia, and 8.9% in Vietnam [23]. In this study, our spectrophotometric analysis recommended by the WHO as a gold standard method revealed that the frequency of G6PD deficiency in Thai male newborns was 10% and 2.3% for males and females, respectively. The incidence in male newborns was in agreement with that reported by Bancone et al. [23], but the incidence in female newborns was lower than that reported by Nuchprayoon et al. [6]. However, G6PD intermediate was more prevalent in female newborns than in male newborns. The possible reasons for the differences might be random X inactivation of female heterozygotes. According to the molecular analysis of the G6PD genotype, the most prevalent mutation in the study population was G6PD Viangchan^G871A (MAF = 0.066). This finding was consistent with other reports [9, 11]. The hemizygous males and homozygous females for G6PD Viangchan^G871A had severe to moderate deficiency, whereas the G6PD Viangchan^G871A heterozygous female phenotype varied from moderate deficiency to normal. Additionally, frequent mutations in Southeast Asia and China, including G6PD Canton^G1376T, G6PD Kaiping^G1388A, G6PD Mahidol^G487A, G6PD Songklanaklarind^T196A, G6PD Chinese-4^G392T, G6PD Union^C1360T, G6PD Chinese-5^C1024T, G6PD Valladolid^C406T, and G6PD Aures^T143C were found in the study population. Both homozygous females carrying G6PD Kaiping^G1388A and G6PD Chinese-5^C1024T had deficient phenotypes, whereas heterozygous mutants were found in intermediate subjects. As the inheritance of G6PD is X-linked, homozygous females and hemizygous males fully express the deficiency phenotype, whereas heterozygous females present partial expression as a result of random X inactivation.

Screening for G6PD deficiency in newborns has been encouraged in many countries with high incidence to prevent jaundice and possible consequences and to raise parental awareness about their children’s condition. It has been suggested that the presence of a large number of reticulocytes may interfere with the diagnosis of G6PD deficiency since reticulocytes have more G6PD enzyme activity than mature erythrocytes [4]. They are released into the bloodstream in response to G6PD deficiency-induced hemolytic anemia [24]. Moreover, the reticulocyte count is higher at birth than at any other point in a healthy life [25]. This led us to investigate the effect of reticulocytosis on G6PD activity in Thai newborns using the automated platform. We found that the mean reticulocyte count in this study population was 4.87 ± 1.36% (range 0.98–12.38), which is higher than that in adults (generally 0.5%-2.5%). It is interesting that the reticulocyte counts in newborns with G6PD deficiency were significantly higher than those in normal newborns in both males and females. After adjusting for thalassemia and hemoglobinopathies commonly observed in this region, we found that G6PD deficiency, G6PD Viangchan^G871A, female gender, neonatal anemia (Hb < 15 g/dL), ABO incompatibility, and postpartum age were independent factors promoting reticulocytosis in newborns. These novel findings indicate that increased reticulocyte status in newborns with G6PD deficiency and intermediate is a result of the compensatory effect of reduced enzymatic activity. However, our results demonstrated that the levels of G6PD activity in normal newborns were highly positively correlated with the percentage of reticulocytes obtained from both assays. Even though the results supported an increase in reticulocytes in G6PD deficiency compared to G6PD normal, there was no association between G6PD activity levels and the percentage of reticulocytes in newborns with G6PD deficiency. This finding agreed with a previous report [26] that high reticulocytosis in neonatal blood samples did not cause a false negative diagnosis for G6PD deficiency. G6PD deficiency is a genetic disorder caused by mutation in the G6PD gene affecting G6PD deficiency in all cell types, including reticulocytes. Although there are numerous reticulocytes in the circulation during the hemolytic crisis as a result of G6PD deficiency, their G6PD activity cannot increase the total blood G6PD activity to the normal level. Hence, elevated levels of G6PD activity in reticulocytes only interfere with the detection of G6PD activity in normal blood samples. In newborns with intermediate levels of G6PD, the levels of G6PD activity were highly negatively correlated with the percentage of reticulocytes. This might result from the presence of a mixed population of G6PD deficiency and normal reticulocytes as a result of heterozygosity.

Since the number of reticulocytes in normal samples affected the measurement of G6PD activity, the efficacy of the automated assay was assessed. Our results revealed that the frequencies of G6PD-deficient males and females obtained qualitatively by using the FST and quantitatively by using the automated UV enzymatic method were not different from those obtained by using the standard spectrophotometric method, indicating the reliability of the automated method. Although the FST method is a simple and rapid qualitative method for screening G6PD deficiency, as recommended by the International Committee for Standardization in Hematology (ICSH) [23, 27], it does not well serve the purpose of separating G6PD intermediate from G6PD normal. The quantitative automated UV enzymatic method is a new robotic spectrophotometric system that is capable of quantifying G6PD activity in 50 blood samples simultaneously. The bimodal and trimodal distribution patterns of G6PD activity values observed in this study population coincided with the characteristics of hemizygous G6PD deficiency and normal males and heterozygous and homozygous G6PD deficiency and homozygous normal females, respectively. Based on previous recommendations, this study used 30% and 80% of normal median activity as cut-off points for G6PD deficiency and intermediate, respectively [15]. In our study, the detection of G6PD deficiency in females at 30% of normal activity by the automated UV enzymatic assay was close to the estimated frequency of G6PD-deficient females reported previously by Domingo GJ. et al. 2019 [22]. The results from ROC analysis revealed that the optimal cut-off value of the automated UV enzymatic method for G6PD deficiency in newborns was less than 7.8 U/g Hb, which was comparable to the value reported previously [26], and for those with intermediate levels of G6PD activity, it was 7.8 to 20.8 U/g Hb. The diagnostic performance of the automated UV enzymatic method was practically faultless for G6PD deficiency, while the effectiveness of intermediate screening was reasonable.

Our results also revealed that the normal reference values of males and females determined by the automated UV enzymatic method were not significantly different. We also found that the reference values of the automated UV enzymatic method for the identification of deficiency and intermediates among newborns were higher than those of adults reported previously [11]. This finding is in agreement with previous reports that G6PD activity in normal newborn blood is higher than that in adult blood [28, 29]. In contrast to the spectrophotometric assay, our reference values for G6PD-deficient and intermediate newborns were comparable to those of adults [11]. Although the G6PD activity values from both methods showed a strongly positive correlation (Fig. 3a-c), we are aware that G6PD activity values obtained from the automated UV enzymatic assay are excessive by a factor of almost three. We plan to introduce corrections in our automated assay until the results are in line with those that have been long established. Both techniques are based on the same principle in measurement the absorbance of NAPDH production at 340 nm and 37 °C. The high values of G6PD activities from the automated UV enzymatic assay may be explained by two reasons: (i) the differences in sample preparation for each assay and (ii) Hb measurement for normalization. The packed red cells were used for the automated UV enzymatic assay, whereas the whole blood was used for the spectrophotometric assay. The packed red cell samples were prepared by centrifugation and removal of plasma and the buffy coat containing the white blood cells. The enrichment of reticulocytes in packed red cell samples may result in high G6PD activity values, compared to whole blood samples in which all cellular contents including reticulocytes were diluted with blood plasma. In this study, Hb values used to normalize the G6PD activity values from both assays were obtained from CBC analysis of whole blood. Hence, we propose that Hb values suitable for normalization of G6PD activity from the automated UV enzymatic method should be derived from the packed red cell hemolysate, rather than from the whole blood.

In summary, reticulocytosis, commonly observed in newborn G6PD deficiency, did not interfere with the diagnosis of G6PD deficiency and intermediate. However, G6PD activity in the normal group was positively correlated with the percentage of reticulocyte count. This affects the activity of G6PD as measured by the automated UV enzymatic method. Nevertheless, the robotic quantitative method using reticulocyte-enriched packed red cells shows a strong diagnostic capability to identify G6PD deficiency and intermediates with high cutoff values. A quantitative spectrophotometric assay is suitable to quantify newborn G6PD activity and detect G6PD deficiency in the prevention of hemolytic anemia of G6PD deficiency.