We identified RVH in a porcine intestinal sample (RVH/Pig-wt/USA/AR7.10-1/2012/GXP[X]) submitted from a farm in Arkansas in 2012. Subsequently, we rescreened 204 available RVA-, RVB-, and/or RVC-positive porcine samples collected during 2006–2009 from 16 US states for RVH. The samples were from 5 different age groups of pigs: 1–3 days (21 samples), 4–7 days (23), 8–20 days (19), 21–55 days (110), and >55 days (9); 22 samples were from pigs of unknown age. Sample selection, histologic examination, extraction of genomic material, reverse transcription PCR (RT-PCR) amplification, sequencing of viral protein (VP) 6 gene, and statistical and sequence analysis are described in the online Technical Appendix (http://wwwnc.cdc.gov/EID/article/20/7/14-0034-Techapp1.pdf).

Figure 1

Figure 1. Epidemiologic and molecular distribution of porcine rotavirus H (RVH) strains, United States, 2006–2009A) Geographic distribution of RVH-positive porcine samples/total number of samples testedPink indicates states containing positive samples; green indicates states...

We identified RVH in 30 (15%) of the 204 samples, including sample AR7.10-1 (online Technical Appendix Table). RVH strains were identified in samples from 10 US states (Figure 1, panel A). The first US sample was identified on November 7, 2006. Of samples from age groups in which we detected positive results, most (20/111, 18%) were from 21–55-day-old pigs; RVH was not detected in 1–3-day-old piglets. We also detected RVH-positive samples in 4–20-day-old (5/42, 12%) and >55-day-old (5/9, 56%) pigs. The number of positive and negative samples differed significantly between age groups (p = 0.036, Fisher exact test). The odds of 21–55 day-old pigs being RVH positive was not significant (odds ratio [OR] 1.63, p = 0.36); however, in the >55-day group, the odds of being RVH positive was significant (OR 5.92, p = 0.031), compared with odds for the 4–20-day group. The trend for increased RVH positivity by age group was not significant (p = 0.94, Wald χ² test).

Although we identified only 5 samples with RVH in pigs co-infected with RVA and RVB, co-infections with RVH and RVA, RVB, both RVA and RVC, or both RVB and RVC (1 sample each) also were identified but did not differ significantly (p>0.05, Fisher exact test) (Figure 1, panel B). We did not identify RVH co-infected with only RVC. Most RVH samples (21 [70%]) were identified from pigs co-infected with RVA, RVB, and RVC, which was significantly higher from any other RVH co-infections with RVA, RVB, RVC, RVAB, RVAC, or RVBC (p<0.001, Fisher exact test). Of these 21 RVA, RVB, RVC, and RVH co-infected samples, 15 were from 21–55-day-old pigs (Figure 1, panel B).

Figure 2

Figure 2. A) Nucleotide neighbor-joining phylogenetic tree of rotavirus (RV) A–D and F–H viral protein (VP) 6 sequencesBlue strains are from the United States; green strains are from Brazil; and the red strain...

The US porcine RVH VP6 sequences (GenBank accession nos. KF757260–KF757289) exhibited 91%–100% nt identity with each other and shared 89%–92% nt identity with Japan porcine strain SKA-1 and 85%–87% nt identity with Brazil porcine strains BR63, BR60, and BR59 (Table 1). The US porcine and human RVH VP6 sequences shared 70%–73% nt identity. The US porcine RVH VP6 sequences were 97%–100% aa identical with each other and 97%–98% and 96%–98% aa identical with the Japan and the Brazil porcine strains, respectively. The US porcine and human RVH VP6 sequences were 75.3%–76.8% aa identical (Table 1). The nucleotide and amino acid pairwise identity charts (Figure 1, panels C and D) and phylogenetic trees (Figure 2, panel A) suggest the existence of at least 2 distinct RVH VP6 (I) clusters/genotypes containing human and porcine strains, respectively.

Compared with other RV species, the US RVH VP6 sequences shared the highest nucleotide and amino acid identities with RVG (51%–53% and 39%–41%, respectively) and RVB (47%–52% and 34%–39%, respectively) (Table 2). In the RV VP6 phylogenetic tree, The RVH, RVG, and RVB VP6 sequences clustered in 1 large branch, whereas the RVA, RVC, RVF, and RVD sequences clustered separately in another large branch (Figure 2, panel A). The RVH evolutionary rate (substitution/site/year) from BEAST (http://tree.bio.ed.ac.uk/) was estimated at 2.6 × 10⁻³ (95% CI 5.83 × 10⁻⁴ to 4.46 × 10⁻³). On the basis of the estimate of the time from the most recent common ancestor for the VP6 gene segment, we believe that US RVH strains circulated in US swine for at least a decade and possibly much longer (the time from the most recent common ancestor 1963–2002, 95% highest posterior density [HPD]) (Figure 2, panel B). The US and Japan RVH VP6 sequences diverged during 1955–1993, 95% HPD, and the estimated divergence of the Brazil RVH VP6 sequences from the US and Japan RVH VP6 sequences was 1832–1991, 95% HPD.

Our data indicate that RVH is widespread in US swine herds. Although the samples analyzed already were known to be positive for RV species A, B, and/or C, our identification of RVH in 15% of samples is remarkable. In the United States, piglets are weaned at 21 days of age and then mixed with other piglets from different production sites, which may explain the higher rate of RV co-infections in 21–55-day-old pigs (10,11). These findings suggest that RVH is underdiagnosed in US swine herds and requires further surveillance.

Our phylogenetic analysis indicates that the RVH strains circulating in US swine is evolutionarily distinct from that found in humans, as well as from swine in Brazil and Japan. Although our low sample number and sequencing of a single gene (VP6) makes the genetic diversity of RVH in US swine herds difficult to fully assess, the lack of spatial structure in the tree indicates extensive gene flow of RVH between swine herds in different US regions. Inferring the circulation of RVH in US swine herds is difficult because of the small sample size, although our time-structured phylogenetic analysis indicates at least 1 decade of circulation. Although US swine are routinely transported to South America, the phylogeny indicates that the VP6 gene of US swine RVH viruses is more closely related to that of Japan strain SKA-1 than to those of the 3 Brazil strains included in this analysis.

In conclusion, we identified RVH in 30 samples from pigs co-infected with RVA, RVB, and/or RVC in the United States, which indicates that RVH has been circulating in US swine for at least 1 decade and perhaps for longer. The human and porcine RVH VP6 sequences clustered into separate branches in the phylogenetic tree, but the presence of RVH in swine clearly raises the possibility of interspecies transmission. Because the swine samples were co-infected with RVA, RVB, and/or RVC, the role of RVH in pathogenesis remains unknown but this circumstance illustrates the need for molecular epidemiologic studies.