High frequency and diversity of parechovirus A in a cohort of Malawian children

Members of the genus Parechovirus (PeV) within the family Picornaviridae are small single-stranded RNA viruses. PeVs belonging to the species Parechovirus A, previously known as “Human parechovirus”, infect humans and can cause a variety of symptoms, including gastrointestinal and respiratory symptoms. Currently, 19 types of PeV-A have been distinguished, named PeV-A1-19. PeV-A3 in particular is a known cause of severe neurological disease such as meningitis and encephalitis, mainly in young children [1, 2]. For assigning new clinical strains to a type, several rules and methods have been proposed: typing based on the VP1 sequence with a 75% nucleotide (nt) sequence identity threshold [3], typing based the VP1 sequence with a 77% nt sequence identity threshold and an 87% amino acid (aa) sequence identity threshold [4], and typing based on the VP3/VP1 junction region sequence with a 82% nt and 92% aa sequence identity threshold [5]. Over the last decades, PeVs have been shown to be highly prevalent around the world. While in Europe and the USA, studies usually find a PeV prevalence (i.e. prevalence of viral RNA or infectious virus in clinical or surveillance samples) between 1 and 7% [6‐11], PeV prevalence in Asia has been reported to be as high as 25% [12‐15]. The most prevalent types in all of these geographical regions are PeV-A1 and -A3 [2, 7‐10, 12‐15]. Data on PeV circulation in Africa are scarce; only three studies have been conducted – in Kenya, Côte d’Ivoire and Ghana – finding very divergent prevalences (2%, 5.2% and 24% respectively) [16‐18]. The aim of this study was to contribute to the little knowledge available on PeV circulation in Africa. For this, we tested samples collected in a cross-sectional study in Malawi for the presence of PeV.

We found a remarkably high PeV prevalence (57%) within a cohort of Malawian children between 2002 and 2004. This prevalence is much higher than the PeV frequencies found in Asia, Europe and North America [6‐10, 12‐15, 24‐31]. Recently, a PeV prevalence of 24% was found in a cohort of Ghanaian children [18], pointing towards a higher PeV prevalence in Africa than in other continents. The same seems to be the case for enteroviruses, for which the reported prevalence in this and other cohorts in Africa is higher than elsewhere in the world [32‐35]. We speculate that poorer hygiene than in more developed countries is a possible explanation for the high prevalence. Alternatively, other factors may have contributed to the high prevalence of PeV found in our study. Higher prevalence of PeVs was previously reported in the rainy season compared to the dry season in Ghana [18]. The fact that the vast majority of our samples were collected during the rainy season, when malaria mainly occurs, might have contributed to the high prevalence. Furthermore, PeVs are known to circulate primarily in very young populations [1], and overall, studies that included children aged up to 5 years found higher frequencies of PeV than studies that included older children and/or adults [10, 14, 17, 18, 24, 25, 30, 31]. The fact that our population consisted solely of children between 6 and 60 months of age, will therefore have contributed to the high PeV prevalence. As there was no association between PeV positivity and inclusion groups, we consider it unlikely that hospital-acquired infections or the presence of severe anemia biased the results.

Although PeV is known to cause cases of severe disease [36, 37], PeV infection is predominantly subclinical. Seroepidemiological studies have shown that the majority of children are positive for PeV neutralizing antibodies by the age of 5 years [38‐41]. In line with this, we did not find a significant association between PeV infection and clinical symptoms. PeV3 in particular is known to cause severe disease, such as meningitis, encephalitis and sepsis-like illness, mainly in children under three months of age [36, 37]. However, the 11 PeV3-positive participants in our study, all above the age of six months, did not show signs of CNS infection or sepsis.

We performed typing using different, frequently used methods: NJ phylogeny based on the VP1 sequence, ML phylogeny based on the VP1 sequence, ML phylogeny based on the VP3/VP1 junction region sequence, and typing by BLAST analysis. Typing using NJ and ML phylogenetic trees identified 15 different PeV genotypes in our study; all genotypes except PeV-A13, -A15, and -A18 were found. Typing using BLAST resulted in 14 different PeV genotypes – all genotypes except PeV-A13, -A15, -A17, and -A18– while five strains had ambiguous results and were labeled indeterminate. Strain P02-4058 remained untypable by all methods. However, a similar strain had recently been submitted to the Picornavirus Study Group and was classified as PeV-A19 after minor reorganizations of the PeV-A classification (Roland Zell, personal communication, December 2018). Strain P02-4058 was therefore typed as PeV-A19. Strain P04-4393 was typed as PeV-A14 based on the VP1-sequence and as PeV-A8 based on the VP3/VP1 junction sequence. Although recombination events in the structural parts of PeV genomes are rare, they do occur and could possibly explain the different typing results obtained by different methods [42]. Our data are in line with findings in Ghana, where all genotypes were identified except PeV-A11, -A13, -A16 and -A19 [18]. We speculate that this may reflect regional differences, with a wider variety of PeV genotypes circulating in Africa than in Europe, North America and Asia, where types other than PeV-A1-6 are rarely seen [2, 7‐10, 12‐14]. Of the genotypes found in our study, PeV-A1, -A2 and -A3 were most prevalent. While PeV-A1 is known to circulate extensively around the world, PeV-A2 is a relatively rare genotype [9, 10, 14, 27, 29]. The high prevalence of this genotype in our study is therefore notable. While PeV-A3 is also highly prevalent worldwide [9, 10, 14, 27, 29], our results are in contrast with studies conducted in Ghana and Côte d’Ivoire, where no PeV-A3 was reported [17, 18]. Since PeV-A3 circulates more widely in children under the age of 3 months [9, 26‐28], we speculate that the proportion of PeV-A3-positive samples in this study would have been even higher if children under the age of 6 months had been included.

Different rules and methods to type PeV strains have been proposed in recent years, and currently, there is no consensus regarding which specific method to use. As a result, typing is performed using different regions and lengths of the viral genome [3‐5] and by using different methods, such as BLAST, NJ phylogeny and ML phylogeny [13, 14, 17, 25, 27, 30, 43]. We have shown that these methods can result in different typing results for the same viral strain, leading to inconsistent and possibly incorrect typing of PeV strains. We believe that consensus on a genotyping framework, preferably based on distinct clustering in a specialized phylogenetic analysis, is needed and would provide more accurate and consistent typing in further studies.

In conclusion, we found a high frequency of PeV circulation in a population of Malawian children. We saw multiple inconsistencies in typing of strains when comparing BLAST and phylogenetic methods. However, with all methods, we found a wide variety of genotypes, with PeV-A1, -A2 and -A3 being the most prevalent types. The presence of the higher-numbered genotypes (PeV-A7-12, -A14, -A16, -A17 and -A19) and the high prevalence of PeV-A2 are especially notable. Further studies and surveillance are needed to elucidate the impact of the high prevalence and diversity of PeV and its clinical relevance on this continent. Moreover, in the future, a consensus on a typing method may be required to avoid inconsistent typing.