We conducted a parasitologic survey of 226 human stool samples in July 2023, during the dry season in Ngounié Province, Gabon. We collected samples from persons in 6 forest-edge communities situated within a tropical savanna climate zone (Köppen classification Aw) (Appendix Figure 1).

Figure 1

Figure 1. Light microscopy images of soil-transmitted helminths from human stool samples, Gabon. A, B) Hookworm eggs; scale bars indicate 25 µm. C) Strongyloidesspp. eggs; scale bar indicates 25 µm....

Upon collection, we preserved stool specimens in 10% formalin and 70% ethanol and shipped them to Australia for analysis. Using formalin ethyl acetate sedimentation microscopy (7), we identified hookworm eggs in 15 samples, S. f. fuelleborni roundworm eggs in 6 samples, and Strongyloides spp. roundworm larvae in 1 sample (Figure 1). A total of 20 samples were helminth-positive, and some involved co-infections (Appendix Table 1).

We performed metabarcoding on DNA extracts from those 20 samples by targeting the mitochondrial cytochrome c oxidase subunit I (cox1) gene and the hypervariable region IV (HVR-IV) of 18S rDNA, 2 well-established genetic markers for helminth species identification (8). We targeted a 217-bp region of cox1 to identify helminth species (8). Then, to further characterize Strongyloides species and genotypes, we conducted a second metabarcoding assay targeting 18S rDNA HVR-IV (≈255 bp) (8) on the 7 Strongyloides spp.–positive samples. For both assays, we performed sequencing on a MiSeq platform by using MiSeq Reagent Nano Kit v2 (both Illumina, https://www.illumina.com) and 500 cycles for 250-bp paired-end reads. We used Geneious Prime version 2024.0.4 (https://www.geneious.com) to analyze sequence data, then used a custom workflow incorporating read quality control, contig assembly, and haplotype assignment. We conducted phylogenetic analyses of MUSCLE-aligned (https://www.ebi.ac.uk/Tools/msa/muscle) cox1 sequences by using maximum-likelihood (MEGA 11, https://www.megasoftware.net) and Bayesian inference (MrBayes, https://github.com/NBISweden/MrBayes) methods and applying the general time-reversible nucleotide substitution model.

Eighteen of the 20 samples yielded cox1 amplicons. Upon sequencing, 3 samples were dominated by reads from co-infecting Ascaris lumbricoides roundworms, but we did not detect sequences for hookworms or Strongyloides spp. roundworms. The other 15 samples yielded sequences assigned to Necator spp. (n = 11), S. f. fuelleborni (n = 3), or Ancylostoma spp. (n = 1) helminths (Appendix Table 1). Among the 11 Necator spp.–positive samples, 10 harbored Necator americanus hookworms, and 4 contained a Necator sp. hookworm with cox1 sequences that had 100% identity to GenBank accession no. AB793562, a species previously detected in researchers from Europe who were infected in the Central African Republic (CAR); that species was later morphologically identified as N. gorillae (3). Three of the 4 N. gorillae–positive samples had N. americanus co-infection (Appendix Table 1).

Figure 2

Figure 2. Maximum-likelihood phylogeny of zoonotic soil-transmitted helminths from human infections, Gabon. A) Necator spp. hookworms; color-coded groups are labeled per nomenclature by Hasegawa et al. (4). B) ...

Analysis of cox1 sequence data revealed 15 haplotypes of N. americanus and 2 of N. gorillae. Maximum-likelihood and Bayesian inference phylogenetic analyses placed N. americanus sequences (217-bp) from this study within a clade containing previously published sequences for that species (Figure 2, panel A). The N. gorillae sequences clustered with isolates from NHPs from CAR and Gabon and with isolates recovered from the infected researchers from Europe (Figure 2, panel A). We also detected 1 cryptic Ancylostoma sp. hookworm that we could not confidently assign to any known species based on available data. At the cox1 locus, that Ancylostoma sp. hookworm clustered basally to Ancylostoma caninum (GenBank accession no. AP017673) and another unidentified Ancylostoma sp. hookworm (GenBank accession no. MK434228) identified in dogs from Australia (Figure 2, panel A). We identified 4 cox1 haplotypes of S. f. fuelleborni roundworms, all of which fell within the African clade of that species (Figure 2, panel B). An attempt to sequence the cox1 of the S. stercoralis–positive sample was unsuccessful.

We obtained 18S rDNA HVR-IV sequences of 255–258-bp length from 6 Strongyloides-positive samples; we identified 5 as S. f. fuelleborni and 1 as S. stercoralis (Appendix Table 1). Haplotyping analysis assigned the S. stercoralis–positive sample to HVR-IV haplotype A, previously found in humans, dogs, cats, and NHPs (8,9). S. f. fuelleborni–positive samples harbored haplotypes K, L, O, or a combination thereof, previous found in NHPs from Africa and humans (10,11) (Appendix Figure 2).

Among sampled communities in Gabon, we found one third (4/12) of hookworm infections were attributable to N. gorilla. Multiple N. gorillae cox1 haplotypes suggest several separate infection events and might be related to higher exposure in certain occupations or other factors, but those data were not available.

Human Necator spp. hookworm infections other than N. americanus were previously identified on the basis of ITS and cox1 haplotyping and phylogenetic analysis on samples from 2 researchers returning to Europe from CAR (4). Adult worms expelled from 1 researcher were morphologically identified as N. gorillae (3). Subsequent molecular work in CAR (12), Gabon (6), and Cameroon (13) similarly reported a zoonotic Necator sp. hookworm sharing an identical ITS haplotype (II) with those from the researchers from Europe, thus presumably representing N. gorillae. Our findings, together with those reports, indicate that N. gorillae hookworm infections could be more widespread than currently recognized in certain human communities in Central Africa. Future surveillance for hookworm infections in Central Africa should use species-specific molecular tools to differentiate between human-specific and zoonotic hookworm species.

We do not know whether the novel Ancylostoma sp. cox1 haplotype we identified represents a zoonotic infection from NHPs or another animal, but detection of hookworm eggs in a fecal sample excludes transient passage of ingested DNA. Further investigations using longer read genotyping targets combined with morphologic analysis of harvested adult hookworms could provide more definitive speciation.

Our findings also suggest that S. f. fuelleborni roundworm infection is common among human populations in Central Africa. However, little S. f. fuelleborni roundworm infection epidemiologic surveillance has been performed in humans in Africa since 1980, when surveys of diagnostic specimens submitted to a hospital in Lusaka, Zambia, reported a 1.0% diagnostic prevalence of S. f. fuelleborni roundworm infections over a 7-month period (14). A 2024 molecular survey conducted in Asia identified a 3.0% (4/134) infection prevalence in some Bangladesh communities (15).

In our study, S. f. fuelleborni sequences clustered closely with isolates from Central Africa at both the cox1 and 18S rDNA HVR-IV loci, supporting the hypothesis of geographic clustering for this species (10,11,15). The S. stercoralis 18S rDNA HVR-IV haplotype we identified is consistent with previous reports of that species in humans from Africa (8,10).

In summary, we report human infections with N. gorillae hookworms and S. f. fuelleborni roundworms in Gabon in Central Africa. Those infections occurred in a forest-edge region where localized environmental disturbance and anthropogenic activities, such as hunting and foraging in the adjacent forest, bring villagers into direct contact with NHP habitats, increasing exposure to NHP STHs (2). To determine the extent of human infections with zoonotic primate STH in areas where populations overlap and to define the clinical effects and most appropriate treatment strategies for infected persons enhanced STH surveillance is needed.