Prevalence and distribution of G6PD deficiency: implication for the use of primaquine in malaria treatment in Ethiopia

Glucose-6-phosphate dehydrogenase (G6PD) is an enzyme involved in the pentose monophosphate pathway. Deficiency of this enzyme leads to free radical-mediated oxidative damage to red blood cells, and in turn causes haemolysis. G6PD deficiency is the most common enzymatic disorder of red blood cells, affecting 400 million people worldwide [1]. It is an X-linked disorder with high prevalence particularly in people of African, Asian, and Mediterranean descent [1]. In Africa, the most common G6PD enzyme-deficient variant is A- [2]. Females with G6PD A- heterozygotes have been shown with selective advantage against severe malaria [3‐5]. By selection, this G6PD-deficient trait becomes prevalent (8%) in populations where malaria is endemic [6].

The prevalence of G6PD deficiency is highly relevant to the choice of drug used in anti-malarial treatment [7, 8]. A number of drugs, such as primaquine, dapsone, sulfonamides, quinolones, chloramphenicol, nitrofurantoins (antibiotics), and phenazopyridine (analgesics), have been described as haemolytic trigger that causes haemolytic crisis in G6PD-deficient individuals [9, 10]. Primaquine is the recommended treatment drug to eliminate Plasmodium vivax hypnozoites and Plasmodium falciparum gametocytes, along with the goal to progress towards zero malaria transmission in Africa [11‐14]. It is an ideal agent to be used as primary prophylaxis against P. vivax [6]. However, primaquine can also induce oxidative stress causing a spectrum of haemolytic anaemia ranging from mild to severe haemolysis in G6PD-deficient individuals [15]. The likelihood of developing haemolysis and its severity depends on the level of enzyme deficiency, which in turn is determined by the type of G6PD variant [16‐18]. The risk for haemolytic anaemia is particularly high in patients who are treated for P. vivax malaria because they are usually given a higher dose of primaquine (0.25–0.5 mg in a 14-day treatment regime) compared to those treated for P. falciparum (a single dose of 0.25 mg on the first day of treatment) [19]. A high dose of primaquine (0.5 mg base/kg daily for 14 days) has been previously shown to be more effective than a low dose (0.25 mg base/kg daily for 14 days) in eliminating primary blood infection and preventing relapse episodes in P. vivax patients [11, 20‐22]. However, the lack of G6PD level information, inaccurate methods of screening G6PD deficiency, and the uncertainty in the safety of a single versus long-term primaquine dosage pose risk to malaria patients when treat with primaquine.

The gene encoding the G6PD protein consists of 13 exons and 12 introns [23] and is located on the X chromosome. This gene is highly polymorphic with nearly 160 mutations at the DNA level that are potentially associated with G6PD deficiency [24]. The frequency of these mutations varies among populations and countries. For instance, mutation S188F, sometimes called the Mediterranean mutation, is most prevalent among individuals from the Middle East [25]. Mutations C131G and G487A that were common in Dhaka, Bangladesh appear to be linked to G6PD deficiency by affecting NADP binding or disrupting the protein structure [26]. The G6PD genetic variants were relatively homogeneous in America, Africa, and western Asia compared to those in East Asia and Oceania. In North America, Africa, Yemen and Saudi Arabia, G6PD*A- variant is predominant among populations. By contrast, G6PD variants are highly heterogeneous in East Asia such as China and the Asia–Pacific where no single variant predominates [6, 27].

Ethiopia is one of the few African countries where P. vivax and P. falciparum coexist, and account for 60% and 40% of the malaria cases, respectively [28]. G6PD deficiency was previously estimated to be as high as 17% in southwestern Ethiopia based on the CareStart™ fluorescence spot test [29]. The mutation A376G that constitutes G6PD*A- variant accounted for nearly 23% of the malaria patients. Other mutations including rs782669677 (535 G->A), rs370658483 (485 + 37 G->T), and chrX:154535443 (C->T) were recently observed in the same geographical region through an investigation of a short segment of the G6PD gene [30]. These mutations did not appear to disrupt function or structure of the G6PD protein [30]. In this study, the prevalence and distribution of G6PD mutations were investigated among a large number of febrile patients across broad areas of Ethiopia. For a subset of samples, the association between G6PD genetic mutations and enzyme level was examined. Malaria parasitaemia and demographic features in the G6PD deficient and non-deficient individuals were characterized.

To date, 8-aminoquinolines, primaquine and tafenoquine are the only effective drugs that eradicate the dormant liver stages of P. vivax and avoid relapse. However, primaquine and tafenoquine can cause acute haemolytic anaemia in individuals with low G6PD enzyme level. The issue of low G6PD has also impacted the use of primaquine as gametocytocide for P. falciparum malaria. In the Greater Mekong Subregion, G6PD deficiency has been reported to be common in males with a prevalence of 7.3% to 18.8% based on the CareStart™ Fluorescent Spot Test [33]. In Bangladesh, a prevalence of 17.4% (173/995) for G6PD deficiency (< 60% of the adjusted male median G6PD activity) was reported using standard UV spectrophotometry [34]. Using the CareStart™ G6PD test kit, this study identified a prevalence of 4.3% G6PD deficiency among febrile patients in southwestern Ethiopia based on a threshold cutoff of 1.88 U/g Hb (i.e. < 30% of the AMM G6PD activity). This prevalence rate was slightly higher than a previous report of 1.4% G6PD deficiency in other parts of Ethiopia using the same device and cutoff threshold [35], but much lower than those reported in Asia [33, 34]. Such a pattern indicated differences not only at the continental but also regional level in the distribution of G6PD deficiency.

Some earlier studies demonstrated an association of low G6PD level with a low risk of asymptomatic P. falciparum infections [36]; whereas others showed no association of G6PD deficiency with total severe malaria or diseases caused by malaria parasite species [37]. While the potential association between low G6PD level and host immunity is not tested in the present study because of sampling bias towards malaria patients, malaria-infected individuals with normal G6PD enzyme level indicated a slightly but non-significantly higher parasite density than those with low G6PD enzyme. Association analyses did not reveal a significance difference in G6PD enzyme level between males and females, as well as among various age groups. While most of the populations in southwestern Ethiopia belong to either the Amhara or Oromo tribes, ethnicity data of the included patients is not available for us to formally test the association between ethnicity and G6PD enzyme level.

More than 400 allelic variants of the G6PD gene have been reported [10, 18, 26, 38, 39]. Among the 344 Ethiopian samples, three SNPs were detected between exon 3 and exon 11 of the G6PD gene, including codon A376G, one of the most common G6PD mutations with an average frequency of 6.1% across the study sites. This mutation has also been previously reported in in Ethiopia [30, 35, 40]. The other two mutations G267+119C/T and G116A were detected with relatively low frequencies in this study. Based on the analyses of a subset of 184 samples, none of these mutations was associated with low G6PD enzyme level. One possible explanation could be only a small number of low G6PD samples were sequenced. The association between G6PD genotype and phenotype merits further verification with broader samples. Another possibility is that there could be other codons of the G6PD gene that we did not sequence in this study. Also, G6PD phenotype was only measured by one but not multiple biosensors or device for comparison. The CareStart™ G6PD rapid test kit used in this study has the advantages of being small and handy, easy to perform, produce results within a few minutes, and can be used without electricity or specific equipment. It is affordable and can detect G6PD enzyme activity at a very low level [41]. Compared with the gold standard photospectrometric assay, the sensitivity and specificity of the CareStart™ test kit are 90–100% and 84.8–100%, respectively [33, 41, 42], although a low sensitivity was also reported when used on individuals with G6PD enzyme activity < 30% [43, 44]. Though this CareStart™ G6PD test kit represents a significant improvement for quantitative diagnosis of G6PD level over previous models, the accuracy of its measurement still requires further validation before clinical deployment. It is also noteworthy that the samples included in the present study were febrile patients instead of the general population, and that such sampling could bias with a high number of malaria-infected individuals. Thus, this study is limited to inferring the distribution of G6PD level between non-malaria and malaria-infected samples rather than the rate of malaria infection between normal and low G6PD individuals.