Background
Malaria is endemic in 75% of Ethiopia. For this parasitic disease, the predominant species is
Plasmodium falciparum followed by
Plasmodium vivax with regional variations. In Jimma zone, Southwestern Ethiopia, transmission is seasonal and unstable with a peak of infection after the rainy season in autumn. A study conducted by this group showed that 36–70% of the malaria infections in Jimma zone were caused by
P. vivax, depending on the study location [
1].
After the worldwide implementation of artemisinin-based-combination therapy (ACT), a general decline in falciparum malaria and a relative increase of vivax malaria was reported [
2]. For a long time,
P. vivax malaria was considered a harmless form of malaria compared to
P. falciparum [
3]. However, over the past few years, an increase of fatal and severe disease outcomes were described [
4]. As a single vivax infection can result in multiple episodes due to relapses caused by the liver stages (hypnozoites), treatment of the liver stage is essential for controlling the disease. Additionally, increasing resistance of
P. falciparum to ACT resulted in a containment strategy including the treatment of gametocytes to stop spread of resistance [
5]. This strategy goes along with elimination efforts. The only recommended treatment by the World Health Organization (WHO) against
P. vivax hypnozoites and
P. falciparum gametocytes is the 8-aminoquinoline primaquine (PQ). A full 14-day course of primaquine (0.25 mg/kg/day) is required for radical cure of
P. vivax, whereas a single-low-dose regimen is recommended for the elimination of
P. falciparum gametocytes. The United States Food and Drug Administration (FDA) has now approved, under priority review, a single-dose regimen of tafenoquine for the radical cure of
P. vivax in patients ≥ 16 years. Tafenoquine is a long-acting analogue of primaquine and as such an 8-aminoquinoline as well [
6]. Unfortunately, the use of 8-aminoquinolines is compromised by the potential haemolysis in individuals with the enzyme deficiency of glucose-6-phosphate dehydrogenase (G6PD). In consequence, enzyme activity testing of G6PD is essential before the administration of 8-aminoquinolines.
Affecting over 400 million people, it is one of the most common enzyme deficiencies in the world that is probably protective against malaria especially severe malaria outcomes [
7]. G6PD is an essential enzyme in the pentose phosphate pathway assuring normal oxidizing processes in the red blood cells (RBC). Although the inherited enzyme deficiency is mainly asymptomatic, it can cause neonatal hyperbilirubinaemia and chronical anaemia. The exposure to oxidative stress can provoke destruction of red blood cells and haemolysis, triggered by even small amounts of oxidants like primaquine, dapsone, paracetamol, fava beans and many more [
8].
The size of the
G6PD gene is 18.5 kB. It consists of 13 exons, 12 introns and is located on the long arm of the X-chromosome on the genloci Xq28. Over 217 Mutations have been characterized to date [
9,
10]. Most of them are single base substitutions. The most common mutations described in Africa are A+, A− (202A) and Mediterranean. Non-synonymous mutations can cause different degrees of enzyme deficiency by activity or stability loss of the enzyme. As the gene is located on the X-chromosome, women are either homozygous deficient, heterozygous deficient (one gene encoding a normal enzyme and one gene encoding a deficient enzyme) or homozygous normal. The phenotype of heterozygous females varies between normal and deficient due to the X-chromosome inactivation in females expressing only one gene copy in each cell (mosaicism). Males are either hemizygous normal (normal phenotype) or hemizygous deficient (deficient phenotype) [
11]. A prevalence of 0.5–2.9% of G6PD deficiency in the male hemizygote population was generally assumed for Ethiopia [
12,
13]. However, a recent study in Southwestern Ethiopia reported higher prevalence of G6PD deficiency of up to 14.3% depending on the ethnic group of the patient [
14]. A genetic study from Jimma detected A+ in 23.3% of the individuals and other uncommon mutations [
15].
PQ is not part of the national Ethiopian policy as G6PD testing is not implemented at healthcare facility level. Population screening is recommended in regions with a G6PD deficiency prevalence ≥ 3–5% to guide decision-making for 8-aminoquinoline treatment in order to minimize the risk of drug-induced complications but population-based G6PD studies in Ethiopia are patchy [
16].
This small study intended to collect first data on G6PD deficiency measured by spectrophotometry in combination with the analysis of associated genetic variants in healthy participants in Jimma zone. Test methods, procedures and logistics were additionally evaluated.
Discussion
The median activity in the study population was 6.1 U/g Hb (range 4.3–10.8 U/g Hb). No activity below 70% or 4.27 U/g Hb could be detected. A prevalence of 0.5–2.9% of G6PD deficiency in the male hemizygous population is generally assumed for Ethiopia. In comparison, prevalence of G6PD deficiency in the male population in other African countries are reported as follows: Somalia: 0.5–2.9%, Sudan: 7.0–9.9%, Uganda, Kenya: 10.0–14.9%, Tanzania, Congo, Gabon and Cameroon: 15.0–26.0% [
12,
13]. As Tsegaye et al. published before, the prevalence of G6PD deficiency in Gambella, Ethiopia, varied between ethnic groups. Overall prevalence of G6PD deficiency measured by CareStart™ G6PDd screening test (Access Bio, Inc., New Jersey, USA) was 7.3%. Gambella is low land area (300–500 m above sea level) characterized by high malaria transmission. They reported high prevalence among the indigenous group of Nuer (14.3%) and Anuak (12.0%) in rural areas and no deficiency in 46 patients from Oromiya [
14]. This corresponds to the 0% prevalence in the presented study. Patients from other “highland” regions like Amhara, Hadya and Tigre showed also no deficiencies [
14]. According to genetic studies, the Oromo and Amhara people appear more similar to Europoids (particularly to the South Arabians) and quite different from the Negritic people. Intermixture of Europoid and Negritic populations seem to have resulted in the present day Cushitic (and Semitic)-speaking group of eastern Africa [
20].
As there are no other studies published from Ethiopia measuring the G6PD activity by spectrophotometry, a local comparison of the median level in U/g Hb had to be omitted. A study from the US examining participants with Afro-American origin reported a median G6PD activity of 7.30 U/g Hb. Male median was unadjusted 7.06 and adjusted 7.18, female median was 7.70 U/g Hb [
19]. The female median was higher than the male. This is in accordance to the results here. Measured in U/g Hb, women (6.4 U/g Hb) had significantly higher values of enzyme activity than men (6.1 U/g Hb; p = 0.0315) as the haemoglobin value differed significantly (p < 0.0001) between the sexes, with women having lower Hb levels than men. There was no difference between the sexes when the enzyme activity was expressed in μmol metabolic rate/10
11 erythrocytes/min.
In another study from Bangladesh with 999 participants, the adjusted male median derived by spectrophotometry was 7.03 U/g Hb [
21]. Compared to the two studies above, the overall median in the presented study was lower. Several reasons are possible. The conversion formula used is an approximation only. Furthermore, G6PD testing in the field poses many challenges to the study team. Loss of enzyme activity during the process cannot be excluded. As Ley and colleagues point out, the duration between blood collection and spectrophotometry and the storage conditions in between are the crucial factors to maintain a stable enzyme activity [
7].
G6PD enzyme activity and degradation are temperature dependent, EDTA blood should be cooled at least at 4–8 °C. The possible duration without loss of activity is unknown [
7]. One study reports a loss of a maximum of 5% of activity after 21 days storage at 4 °C [
22]. However, others report loss of activity of 15–21% after 21 days, 6% after 7 days and 40% after 7 days at 4 °C [
7]. Cryopreservants and storage at − 80 °C [
23] or replacing plasma with an additive [
22] are recommended.
The samples of this study were constantly cooled at 4 °C until shipment. Transport via DHL express (4–7 days) or by study members (2 days) was at ambient temperature. Shipment via DHL was problematic due to logistical problems and customs issues and cannot be safely recommended. Overall mean storage time was 12.2 days (SD 5.2, range 4–21 days). Storage duration and G6PD activity levels showed no evident correlation to indicate a systematic significant loss of activity over time. Haemolysis and significant loss of enzyme activity was only suspected in two samples with overall low activities for nine measured control enzymes and a storage time of 14 and 19 days. Median G6PD activity in samples stored 14 and 19 days was 6.5 and 6.1 U/g Hb, respectively.
Assuming a normal male median of 7.0 U/g Hb [
19,
21] a median of 6.1 U/g Hb would mean a loss of 13% in enzyme activity. The real absolute values in the study could be slightly higher than the presented values. On the other hand, the haemoglobin level was relatively high with 17.1 g/dl for males and 14.3 g/dl for females. Twenty percent of the study population had elevated haemoglobin levels. All participants, though originating from different villages, were living in Jimma town that is situated at 1780 m above sea level. Higher haemoglobin levels result in lower G6PD values as the calculation with the formula used is haemoglobin-dependent. The two studies cited above presented no haemoglobin levels for comparison but were conducted with participants living at sea level [
19,
21]. In conclusion, activity loss cannot be excluded but relevant loss seems unlikely; relative values in per cent should not be affected in any way.
Genotyping of the
G6PD gene revealed two mutations in the small sub-population. The G6PD-A+ mutation (A376G) was found in one male (2.9%). Enzyme activity was 4.3 U/g Hb (70.5%). He reported an intolerance to norfloxacin that was not further specified but no haemolysis was reported in the medical history. G6PD-A−, the combination of A376G and G202A, and C563T (G6PD-Mediterranian) are common variants in Africa, the Mediterranean region and the Middle East but G202A and C563T could not be found in this study. This is in accordance with another study from Southwestern Ethiopia, Malo, genotyping 555 samples for G202A and C563T. None of the variants were detected [
24]. A recent paper reported G6PD genetic variants in malaria patients from Jimma zone [
15]. They found G6PD-A+ in 23.3% of the 86 individuals whereas G202A and C563T were absent. They reported other mutations: G535A in two females and two additional mutations in the intronic region in one male (485 + 37 G → T, rs370658483, chrX: 154535131).
While G6PD-A+ is usually regarded as variant with normal enzyme function, G6PD-A− was associated with a milder degree of enzyme deficiency. However, recent reports suggest that G6PD-A− was associated with severe illness after administration of primaquine [
25,
26]. Primaquine sensitivity has only been reliably characterized for the variants Mediterranean and Mahidol with severe and moderate degrees of sensitivity to primaquine, respectively [
27]. The second mutation that was observed in this study was the substitution G445A. The male participant showed an enzyme activity of 75.4%. He originated directly from Jimma town. This variant was not observed before.
Considering the results presented, would there be any consequences concerning radical cure of
P. vivax for the subjects of the study? A generally adopted method and its corresponding cut-off for G6PD deficiency would simplify comparison and implementation of standards. The recommended thresholds concerning malaria treatment are indicated in percent. PQ treatment for radical cure in vivax malaria (0.25–0.5 mg/kg/day for 14 days) should not be given in individuals with G6PD activity below 30% of the adjusted male median (AMM). This level excludes severe deficiency and seems safe to discriminate homozygous females and hemizygous men from heterozygous females or variants causing only moderate deficiency [
27]. For the long-acting tafenoquine, the recommended threshold in the approval studies was 70% of the AMM for all participants and 90% for anaemic females (7–10 mg/dL) [
28] to exclude moderate deficiencies.
According to the WHO, heterozygote females are classified in activity levels of < 30%, 30–80% and > 80% correlating in a deficient, intermediate or normal phenotype, respectively. Sensitivity to primaquine seems possible in heterozygote females with the intermediate phenotype and unlikely in heterozygote females with the normal phenotype [
29]. Eleven patients showed activities between 70 and 80% of the male median in this study: three females and eight men with the normal genetic variant and two hemizygote men with the stated mutations above. Both showed a borderline intermediate phenotype. Sensitivity to 8-aminoquinolines seems at least possible.
If the above-mentioned thresholds of 30% for PQ and 70% for tafenoquine are used, both regimens could be administered to the study population. However, the males with the G6PD variants should be treated with caution and closely monitored. There is few or no information about 8-aminoquinoline sensitivity concerning these mutations.
From a clinician’s perspective, a cut-off of 80% would prevent any possible, probably mild harm in two cases (1.0%) but nine healthy patients (4.4%) would miss the valuable radical cure. From a public health point of view, this approach lacks efficiency and a lower cut-off should be considered, but a general recommendation to treat without testing in this population would be irresponsible as the study population is not representative enough to safely exclude severe deficiency. However, in absence of a better drug the low prevalence of G6PD deficiency should encourage further malaria elimination efforts in this area, as the chances of an effective and broadly applicable use of 8-aminoquinolines are high.
No parasitaemia was found by RDT or microscopy in this study. Recruitment was between March and May during and after the smaller rainy season. A recent study from rural Southwestern Ethiopia conducted in a similar mountainous region with seasonal malaria found asymptomatic sub-microscopic parasitaemia with
P. falciparum in 5.2%,
P. vivax in 4.3% and mixed infection in 0.2% of 555 individuals by PCR [
24]. Microscopy and RDT were negative in all samples. Therefore, sub-microscopic parasitaemia cannot be excluded in the underlying study group although probability seems lower as all participants lived in Jimma town in contrast to rural areas. Microscopy and RDT (Binax NOW
® Malaria, Alere, USA) correlated well so far.
Evaluation of the G6PD RDT (Binax NOW® G6PD, Alere, USA) was impossible due to the small sample size in combination with zero prevalence of deficiency. Discrimination of slightly diminished enzyme activity (70–80%) and normal activity was impossible as expected.