Figure 1

Figure 1. Resected cystic pentastomid lesions extracted from patients during abdominal surgery, Sankuru District, Democratic Republic of the Congo, 2014–2015. A) Six abdominal cysts resected from patient 3, who was found to be...

From 188 patients seen during the study period, we identified 7 (3.7%, of whom 4 were female) with incidental visceral fibrotic lesions, cystic lesions, or both (Table). Age (not determined for 4 patients) ranged from 35 to 42 years. All patients had undergone surgery for non–pentastomid-related conditions. A total of 23 lesions (1–6 individual cysts per patient) were resected. Multiple lesions (2 to 6) were resected in 5 (2.7%) of the 188 participants (Figure 1). All 7 patients with cystic lesions had pentastomid larvae as the underlying cause, as identified by pentastomid 18S rDNA PCR, Armillifer cox PCR, or both. In all but 3 lesions, at least 1 PCR was positive and gave a pentastomid sequencing result directly from the ethanol fixed specimens or from the formalin-fixed, paraffin-embedded tissue (tissue slice or whole block). In 2 of these 3 lesions, ovarial tissue was detected by histology (patient 6, cysts 2 and 3); in 1 necrotic pentastomid lesion, no positive PCR result was obtained (patient 1, cyst 4) (Table).

Pentastomids identified after sequencing (98%–100% homology) were A. armillatus larvae in 10 (43%) of 23 lesions, A. grandis larvae in 9 (39%), and Raillietiella sp. larva in 1 (0.04%). The Raillietiella sequence obtained was identical to 2 unspecified Raillietiella sp. GenBank entries (accession nos. EU370434 and AY744887). In total, we detected A. armillatus in all 7 patients. Co-infections with A. armillatus and A. grandis larvae were identified in 3 (43%) patients (patients 1–3); patient 1 had a triple pentastomid species infection, including a Raillietiella sp. larva.

Figure 2

Figure 2. Histologic and immunohistochemical analyses of resected pentastomid lesions from patients in Sankuru District, Democratic Republic of the Congo, 2014–2015. A) Typical necrotic pentastomid lesion from patient 5. Internal structures of the...

Histology was performed from 12 cysts (including the 2 diagnosed as ovarian cysts) resected from 5 patients. Necrotic pentastomids were identified in 5 cysts, from a total of 3 patients, and the decay-refractory shed exuvia surrounded by fibrosis was found (Figure 2, panels A, B). In 4 of these 5 necrotic pentastomid cysts, PCR and sequencing results were positive, even from tissue slices. We found total necrosis of histologically unknown etiology in an additional 5 cysts from 2 patients; in all of them, we obtained positive pentastomid PCR and sequencing results (Table).

Immunohistochemical analysis showed B cells and T cells clustering locally at the fibrous capsule of necrotic pentastomid lesions (Figure 2, panel C). Immunostaining for TGF-β showed a slight expression of TGF-β in fibrous tissue surrounding the larvae (Figure 2, panel D).

Figure 3

Figure 3. Adult Armillifer and Raillietiella parasites found in snakes at local markets in Kole, Sankuru District, Democratic Republic of the Congo, 2014–2015. A) Adult A. grandis in respiratory tract of a local...

Examination of 4 large snakes (2 Gaboon vipers [Bitis gabonica], 1 rhinoceros viper [B. nasicornis], and 1 African rock python [Python sebae]) offered for consumption at markets near Kole during the study period showed pentastomid infections of the trachea and lungs and the mesenteric membrane. A. grandis was detected morphologically and molecularly (99%–100% sequence homology) in the 2 Gaboon vipers (8 and 10 parasites, respectively) and in the rhinoceros viper (10 parasites). The python harbored 6 adult Raillietiella sp. as identified by PCR (identical to the 2 unspecified Raillietiella sp. sequences in GenBank) (Figure 3).

Figure 4

Figure 4. Molecular phylogenetic analysis of Armillifer spp. sequences obtained from humans and snakes in Sankuru District, Democratic Republic of the Congo, 2014–2015. Parasite cytochrome oxidase subunit I gene sequences from abdominal (Abd)...

Phylogenetic analyses based on cox results obtained from patients’ samples and snake specimens showed sequence variations within the A. grandis and the A. armillatus branch, indicating the presence of multiple local parasite strains. The Armillifer sequences found in 1 patient did not all cluster directly together, suggesting multiple infection events or ingestion of multiple pentastomid species or strains at the same time. Some sequences from patients’ larval Armillifer parasites were highly similar to sequences from adult Armillifer parasites found in snakes (Figure 4). The Raillietiella sp. 18S rDNA sequence from the patient showed 100% nt identity with the sequence from the snake’s Raillietiella.

After reports of heavily symptomatic eye infections (7,8) and incidentally found asymptomatic abdominal infections (12) with pentastomids in the remote tropical and forested Sankuru District of DRC, we conducted a short-term systematic study that sheds light on visceral pentastomiasis as an emerging infectious disease in this area. The previously described series of ocular infections with A. grandis larvae in the region was believed to indicate a local problem (8) because ophthalmologic infections are severely symptomatic, whereas visceral (abdominal) cases usually go unnoticed (1). The prevalence of visceral pentastomiasis we detected was not totally surprising because an earlier report of 2 patients had suggested that emerging situation (12). However, multispecies pentastomid co-infections as described here were not anticipated.

In this investigation, co-infection was detected molecularly by application of established pentastomid 18S rDNA PCRs and the newly developed Armillifer-specific cox PCRs. Positive PCR results were obtained from all but 1 resected pentastomid lesions, even from necrotic pentastomid cysts and necrotic lesions of histologically unidentifiable etiology. By histologic examination only, determining the pentastomid species, even in nonnecrotic lesions, is nearly impossible and would rely on counting the body annulations in arbitrary section planes. Molecular tools, as shown in this investigation, can facilitate diagnosis and are able to determine the causative organism to the species level, even in necrotic or calcified tissue samples.

We detected A. armillatus and A. grandis in roughly similar frequencies in the cysts we tested (43.5% and 39.1%, respectively). Whether the co-infections in 3 patients represent superinfections or simultaneous infection events remains unclear. The finding of highly similar sequences of A. grandis from local snakes and patients underscores local transmission, although the snake A. armillatus sequence was more divergent from patient A. armillatus sequences. A. armillatus is geographically more widely distributed; its preferred final hosts are mainly pythons (1,3,13), whereas vipers are usually the hosts for the more geographically restricted A. grandis (1). We did not find co-infected snakes, but co-infections occur in reptile hosts as well (14).

Our finding of a third pentastomid species responsible for a human infection, Raillietiella, was completely unexpected. Earlier work in 1954 proposed that R. hemidactyli, a pentastomid using lizards as final hosts and coprophageous insects, such as cockroaches, as intermediate hosts, might cause a dermatologic condition similar to creeping eruption in Vietnam (15). The authors described the syndrome in 3 patients who had eaten live lizards, similar to a report from 1952 (16). However, Raillietiella was not proven as the causative agent in the reported disease (and human infection with this parasite in general) because the parasite (adult or larval) was not found in the skin; speculation was based on the finding of local lizards harboring this parasite (15). This investigation confirms that Raillietiella can cause human infection, similar to Armillifer parasites. We also showed that snake meat consumption might factor into transmission of this pentastomid parasite because a Raillietiella sp. was found in a local snake in the district with identical 18S rDNA sequences to those from the human infection. In contrast, the consumption of lizards, a reptile host in which Raillietiella is common, is not a habit in this region.

In the human host, encapsulated pentastomid larvae can live for a few years, after which they calcify (which might be seen as C-shaped calcifications on radiology films [1]). Dying larvae are thought to release antigens, provoking an immune reaction (1). We detected an accumulation of B cells and T cells around the disintegrating pentastomid lesions. TGF-β, a fibrosis-inducing cytokine, was expressed in the fibrous capsule surrounding the parasite lesion. Because patients sometimes harbor hundreds of larvae without overt clinical symptoms (1), the host–parasite interface and possibly locally produced host cytokines or factors released by living pentastomids deserve future research attention. A local or systemic immunosuppression induced by secretion of soluble factors by pentastomid larvae might be possible.

The patients in our study regularly ate snake meat, as is the habit in this region. At local markets, we found adult A. grandis and A. armillatus parasites in snake meat on several occasions and once adult Raillietiella parasites. The presence of adult pentastomids in snakes is well known to the local population. The local residents believe that Armillifer pentastomids are internal structures of snakes, “spiral springs,” that help in the serpent’s locomotion. Residents are unaware of the infective nature of these structures and, during consumption of undercooked snake meat, spit out adult parasites. This habit seems highly risky, leading to ingestion of pentastomid ova released during chewing. Simple hygiene instructions, such as removal of any parasites visible in slaughtered serpents, hand washing after handling snake meat, and proper cooking of snake meat, is likely to reduce the risk.

Our study also draws attention to the medical consequences of increasing bushmeat consumption in the Congo basin (17). Extensive deforestation and reduced availability of birds and mammals have resulted in consumption of snake and other reptile meat (8). Although these phenomena are not limited to Sankuru District, the geographic extent and incidence of snakeborne pentastomiasis as an emerging infectious disease is likely to increase. Currently, only scarce and outdated information exists about the prevalence of human pentastomiasis: radiology studies in the Congo region in the 1950s (<1% prevalence [18]), in Nigeria in the 1990s (1.4% [19]), a serosurvey in Côte d'Ivoire in the 1980s (4.2% [20]); and autopsy studies in the Congo region in the 1910s and 1930s (12%–22.5% [21,22]), in Cameroon in the 1910s (7.8%–12.6% [23,24]), Nigeria in the 1960s (33% [25]), and in indigenous people in Malaysia in the 1960s (45.4% [26]).

In summary, we undertook a surgery-based study of patients in a region where snake meat consumption is widely practiced to investigate etiology of abdominal cystic lesions for the presence of pentastomid larvae. Seroprevalence studies using crude parasite antigens (ELISA and immunoblot [3]) in the affected region will follow. Our findings indicate that the rate of visceral pentastomiasis as an incidental finding during surgery indicates endemicity, most likely fueled by consumption of bushmeat and animal exploitation.