Surveillance of Rotavirus Strains in Rio de Janeiro, Brazil, from 1997 to 1999

Abstract

One hundred fifty-seven (23%; n = 678) rotavirus-positive stool samples were collected between March 1997 and December 1999 in the cites of Rio de Janeiro and Niterói. Rotaviruses in 143 (91%) samples were genotyped by reverse transcription-PCR for G and/or P specificity. Rotaviruses in the majority of G-P-typeable samples (73.3%; 74 of 101) were identified as having globally common genotypes G1P[8], G2P[4], G3P[8], and G4P[8]. Unusual strains such as G1P[9], G2[P8], G3P[9], and G9P[4] strains were detected in 8.9% (9 of 101) of the samples. Genotypes G9P[8], G9P[6], and a mixture of G9 and other G or P types represented 15.9% (25 of 157) of the isolates. Mixed infections were detected in 25 (15.9%) samples, and rotaviruses in 15 samples (9.6%) were not typed.

Rotaviruses, which are members of the Reoviridae family, are classified into seven groups (rotaviruses A to G). A complete virus particle possesses 11 double-stranded RNA segments surrounded by three concentric protein layers. Since their first detection 3 decades ago, group A rotaviruses have been established as the major cause of severe acute gastroenteritis in infants and young children in both developed and developing countries and are estimated to cause 800,000 or more deaths annually among children <5 years of age in the developing countries (22). The development of a safe and effective rotavirus vaccine has thus been a global public health goal. Extensive epidemiological studies worldwide have demonstrated that genotypes G1P[8], G2P[4], G3P[8], and G4P[8] are responsible for the majority of rotavirus infections (13, 32). Thus, most of the candidate rotavirus vaccines developed so far are designed to cover G and/or P serotypes of epidemiologic importance (22). Recently, rotaviruses bearing G9P[8] or G9P[6] specificity have emerged as globally important (1, 2, 3, 5, 9, 18, 26, 29, 31, 33, 34, 38, 41, 42). There are also indications that in Brazil rotavirus genotype G5P[8] may be of epidemiologic importance (17).

Since 1996 we have been conducting an epidemiological rotavirus strain surveillance in Rio de Janeiro, Brazil, to monitor the distribution and evolution of the rotavirus genotypes (e.g., the emergence of unusual or new rotavirus genotypes) circulating in the community. During the first year of surveillance we detected a high prevalence rate of genotype G5 in the two largest cities in Rio de Janeiro State (Rio de Janeiro and Niterói) (36). We present here the follow-up data of the rotavirus strain surveillance in Rio de Janeiro and Niterói from 1997 to 1999.

One hundred fifty-seven (23%, n = 678) rotavirus-positive stool samples from children under 5 years of age with acute diarrhea were collected between March 1997 and December 1999 in the cities of Rio de Janeiro and Niterói. The presence of rotavirus in those samples was determined by polyacrylamide gel electrophoresis analysis. Fifty-two rotavirus strains that could not be typed directly from stool samples, possibly due to inhibitors present in the sample, were adapted to growth in cell cultures of primary African green monkey kidney cells (BioWhittaker, Walkersville, Md.). Rotavirus double-stranded RNA was extracted from stool samples or infected-cell culture lysates and analyzed by reverse transcription-PCR for determination of rotavirus G and P specificity by using primers specific for genotypes G1 to G6, G8 to G11, P[4], P[6], P[8], P[9], and P[10] (10, 12, 14, 15, 21, 39).

G genotypes were successfully determined for 129 of 157 (82.2%) rotavirus-positive samples. Overall, G1 was the most prevalent (30%; 47 of 157), followed by G2 (14%; 22 of 157), G9 (12.1%; 19 of 157), G3 (10.2%; 16 of 157), and G4 (1.9%; 3 of 157). Rotavirus in 28 (17.8%) samples was nontypeable, and 22 (14%) specimens had mixed G genotype infections (Table 1). P genotypes were determined for 132 (84%) samples, and the P type distribution was as follows: P[8], 45.9% (72 of 157); P[4], 18.5% (29 of 157); P[9], 4.4% (7 of 157); P[6], 1.9% (3 of 157). Twenty-five (15.9%) samples were nontypeable, and 21 (13.4%) samples had mixed P genotype infections (Table 2).

TABLE 1.

Human group A rotavirus G genotypes detected in the cities of Rio de Janeiro and Niterói, Brazil, from 1997 to 1999

Yr	No. (%) of rotavirus isolates	No. (%) of genotypea:					No. (%) of:
		G1	G2	G3	G4	G9	Nontypeable strains	Mixed G infections
1997	67	14	12	14	0	1	16	10
1998	36	8	3	1	1	5	10	8
1999	54	25	7	1	2	13	2	4
Total	157 (100)	47 (30.0)	22 (14.0)	16 (10.2)	3 (1.9)	19 (12.1)	28 (17.8)	22 (14.0)

^aMixed rotavirus infections are not included in the G1 to G4 and G9 totals.

TABLE 2.

P genotypes of human group A rotavirus detected in the cities of Rio de Janeiro and Niterói, Brazil, from 1997 to 1999

Yr	No. (%) of rotavirus isolates	No. (%) of genotypea:				No. (%) of:
		P[8]	P[4]	P[6]	P[9]	Nontypeable strains	Mixed P infections
1997	67	24	16	0	7	9	12
1998	36	15	5	0	0	10	6
1999	54	33	8	3	0	6	3
Total	157 (100)	72 (45.9)	29 (18.5)	3 (1.9)	7 (4.4)	25 (15.9)	21 (13.4)

^aMixed rotavirus infections are not included in the P[4], P[6], P[8], and P[9] totals.

Both G and P genotypes were determined for 101 of 157 (64.3%) rotavirus strains. The most frequent G and P combination was G1P[8] (39.6%; 40 of 101), followed by G2P[4] (20.8%; 21 of 101), G9P[8] (14.8%; 15 of 101), G3P[8] (9.9%; 10 of 101), G9P[6] (3.0%; 3 of 101), and G4P[8] (3.0%; 3 of 101) (Table 3). Unusual combinations such as G1P[9], G2[P8], and G9P[4] were observed at low frequency (1%; one each), and G3P[9] strains represented 5.9% (6 of 101) of the isolates. Fifteen of 157 (9.6%) specimens remained nontypeable for both G and P genotypes, and 16 of 157 (10.2%) specimens were typed for the G or P genotype only. Twenty-five of 157 (15.9%) samples presented a mixture of G and/or P types (Table 4).

TABLE 3.

Human group A rotavirus G-P genotype combinations identified in the cities of Rio de Janeiro and Niterói during rotavirus seasons from 1997 to 1999

G-P combination	No. of isolates in:			Total (%)
	1997	1998	1999
G1P[8]	11	7	22	40 (39.6)
G1P[9]	1	0	0	1 (1.0)
G2P[4]	11	3	7	21 (20.8)
G2P[8]	1	0	0	1 (1.0)
G3P[8]	9	1	0	10 (9.9)
G3P[9]	5	0	1	6 (5.9)
G4P[8]	0	1	2	3 (3.0)
G9P[8]	1	5	9	15 (14.8)
G9P[4]	0	0	1	1 (1.0)
G9P[6]	0	0	3	3 (3.0)
Total	39	17	45	101 (100)

TABLE 4.

Mixed G and/or P rotavirus infections detected in Rio de Janeiro and Niterói from 1997 to 1999

Mixed G-P combination	No. of isolates in:			Total
	1997	1998	1999
G1P[8] + P[6]	0	0	2	2
G1 + G2P[8] + P[4]	1	0	0	1
G1 + G3P[8]	1	0	0	1
G1 + G3P[9]	2	0	0	2
G1 + G3P[4] + P[9]	1	0	0	1
G1 + G3P?	1	2	0	3
G1 + G9P[8]	0	0	1	1
G1 + G9P[9]	0	1	0	1
G1 + G9P?	0	0	1	1
G1 + G9P[8] + P[6]	0	1	0	1
G2 + G3P[4]	1	0	0	1
G3P[8] + P[4]	1	0	0	1
G3P[8] + P[9]	1	0	0	1
G3P[8] + P[4] + P[9]	1	0	0	1
G3 + G9P[9]	0	1	0	1
G9 + G?P[9] + P?	0	1	0	1
G?P[8] + P[4]	2	2	0	4
G?P[8] + P[4] + P[9]	1	0	0	1
Total	13	8	4	25

During the 3-year surveillance, 157 rotavirus strains were detected and analyzed, and the majority of the specimens whose G and P types were identified (73.3%; 74 of 101) had the globally common genotypes G1P[8], G2P[4], G3P[8], and G4P[8]. Unexpectedly, 15.9% (25 of 157) of the strains were typed as either G9P[8], G9P[6], G9P[4], or a mixture of G9 and other G or P types. Before 1997 in Brazil, a single G9 rotavirus strain was detected in humans (24) whereas two G9 strains were detected in diarrheal piglets in 1991 (37). In the present survey, only one G9 strain was detected in 1997; however, the relative frequency of the G9 genotype increased abruptly to 25% (9 of 36 strains) in 1998 and to 27.7% (15 of 54 strains) in 1999. It will be important to monitor the evolution of the G9 genotype in this and other regions of this country in coming years. Unusual strains such as those with G1P[9], G2[P8], G3P[9], and G9P[4] genotypes represented 5.9% (9 of 101) of the isolates. That the G5 genotype was not detected was particularly intriguing since this G genotype represented 57% of the isolates in 1996 (17, 36). Whether its disappearance is related to the emergence of the G9 genotype is unknown. However, it is possible that the evolution of the rotavirus G5 genotype in Brazil may resemble that of the rotavirus G9 genotype in the United States, where it was initially detected in Philadelphia, Pa., in 1983 to 1984 in 9.2% of infants with rotaviral disease (7) and then disappeared for about 1 decade. The G9 genotype reemerged in the same city with an incidence of 56% in rotavirus diarrhea in 1995 to 1996 (H. F. Clark, personal communication) and at a lower frequency in other U.S. cities as well (18, 34).

Surveillance of rotavirus genotype distribution has been conducted regularly in Brazil since 1982 (2, 6, 16, 19, 20, 23, 25, 27, 28, 35, 36, 38, 40). These studies were performed in various locations in the country and provided a rich database on rotavirus genotype distribution. Overall, the relative frequencies and distribution of rotavirus G and P genotypes in Brazil are similar to those of other countries, with genotypes G1, G2, G3, G4, P[8], and P[4] being the most prevalent. Nevertheless, other rotavirus genotypes have also been detected. Rotavirus genotype epidemiology in Brazil has been shown to have three characteristic features: (i) a great diversity of G and P genotypes circulating simultaneously; (ii) a frequent occurrence of unusual G and/or P genotypes as well as unusual G-P combinations, and (iii) a high proportion of mixed rotavirus infections, which appear to facilitate genetic reassortment between strains and the subsequent appearance of new G-P combinations.

For example, in 1994 Gouvea and colleagues first demonstrated the occurrence of rotavirus genotype G5 among Brazilian children with diarrhea in São Paulo (16). This G genotype was later shown to be widely distributed in the country, almost exclusively in association with the P[8] genotype (17). More recently, one single strain bearing G5P[6] genotype specificity has been described (28). The frequency of rotavirus type G5 infections increased progressively until 1996, when it was detected in 57% of rotavirus-positive fecal specimens analyzed in Rio de Janeiro (17, 36). After 1996, detection of the G5 strains decreased, showing that they undergo a periodic fluctuation in prevalence in a community, as has been described for G1 to G4 strains (4, 11, 29, 30). However, the circulation of G5 rotavirus among humans was shown to be restricted no longer to Brazil, as infections with such strains appeared in Argentina and Paraguay, suggesting a possible spread of these viruses across South America (4, 8). Interestingly, the decreased frequency of G5 strains coincided with the emergence of the rotavirus G9 genotype in Brazil (2, 35, 38). The genotype G9 strains were usually detected in combination with genotype P[8], although a few G9 strains bearing P[4] or P[6] specificity were also detected (2, 38). The frequency of rotavirus genotype G9 detection appears to be increasing throughout the world in the last few years, and G9 could be regarded as the fifth globally important genotype epidemiologically (1, 2, 3, 5, 9, 18, 26, 29, 31, 33, 34, 38, 41, 42). With regard to P genotypes in Brazil, even though P[8] and P[4] have been shown to be the most common P genotypes in the country, other P genotypes such as P[6] (often) and P[9] (less frequent) have also been detected among the Brazilian populations (2, 23, 38, 40). The need for continuous surveillance of rotavirus strains in communities in Brazil is emphasized.