Acute gastroenteritis (AGE) remains a burden among Ukrainian children, with more than 50,000 cases of AGE registered annually [1]. The etiology of the majority of these AGE cases remains unspecified and data concerning rotavirus detection are lacking because diagnosis of rotavirus infection is not routinely performed in medical units. In December 2006, Ukraine joined the World Health Organization’s Global Rotavirus Surveillance Network (WHO GRSN) [2,3] in order to document rotavirus disease burden for future consideration of rotavirus vaccine (RVV) introduction. Data from 2007, the first year of the network, showed that among hospitalized Ukrainian children with AGE, 49% were rotavirus positive in Kyiv and 41% in Odesa [3].

In 2010, the Ministry of Health of Ukraine (MOH) recommended use of monovalent Rotarix (RV1, GlaxoSmithKlein), and it was subsequently offered through the private market. The health benefits of RVV after its incorporation into national immunization schedules elsewhere has been well documented [4–7]. After universal RVV introduction in the Republic of Moldova, the percentage of hospitalization with rotavirus gastroenteritis fell from 45% to 25% in the first post-vaccine year and 14% in the second post-vaccine year, among children less than 5 years of age [8]. Similarly in Armenia in the first and second surveillance years after RVV introduction, the percentage of rotavirus-positive hospital admissions among those aged less than 5 years fell from 38% to 20% and 10%, respectively [9]. In Ukraine, the introduction of RVV as a recommendation and not as an obligatory vaccine in the national immunization program, has led to coverage of only 0.15–0.6% (our unpublished data).

As incorporation of RVV into the national immunization schedule is currently under consideration, our objective is to describe the epidemiology of rotavirus disease among Ukrainian children from sites participating in the WHO GRSN during 2007–2015.

During July 2007–June 2015, 10,784 children younger than 5 years were enrolled in the surveillance programs in Odesa, and during July 2011–June 2015 1771 children were enrolled in Kyiv (Table 1), representing 78% and 88% of eligible children, respectively. Of these enrolled children, 10,616 (98%) and 1734 (98%) had stool specimens collected and tested for rotavirus in Odesa and Kyiv, respectively. During the surveillance period, fluctuations in enrollment at each site were observed. A lower percentage of enrolled children was observed during Jul 2011–Jun 2012 in Kyiv, with 64% of children enrolled, and in Odesa lower enrollment was seen during each of the surveillance years between Jul 2007 – Jun 2012, with between 56–64% of eligible children enrolled.

Overall, rotavirus infection was detected in 5412/12350 (44%) of children; 929/1734 (54%) of those in Kyiv and 4483/10616 (42%) in Odesa. In some years the proportion of rotavirus positive cases was as high as 60% (Kyiv) or as low as 39% (Odesa). In general, the proportion of rotavirus positive children was higher in each year in Kyiv as compared to Odesa.

Rotavirus detection occurred similarly among boys and girls (Table 2). The median age of rotavirus positive children was slightly higher as compared to rotavirus negative cases in both sites (21 vs 18 months in Kyiv and 24 vs 21 months in Odesa, both p < .01). Both in Kyiv and in Odesa AGE was significantly more often due to rotavirus rather than non-rotavirus etiology among children 12–23 months old; 36% vs 29% in Kyiv and 30% vs 25% in Odesa, both p < .01. Among the rotavirus positive cases, 50% occurred among children between 6–23 months of age in Kyiv, and 45% in Odesa.

Similar clinical characteristics were observed in both Kyiv and Odesa. Rotavirus positive children had significantly more episodes of diarrhea as compared to rotavirus negative children, and a significantly higher proportion had experienced vomiting, with higher number of vomiting episodes reported. Rotavirus positive children were also significantly more febrile and scored higher on the severity grouping scores as compared to their negative counterparts.

Treatment differed between the sites, with nearly all children receiving intravenous hydration in Odesa, whether rotavirus was detected or not, whereas in Kyiv, rotavirus positive children received significantly more intravenous hydration than rotavirus negative children (42% vs 23%, p < .01). A protracted disease course was observed more often in cases of gastroenteritis due to rotavirus, with a significantly higher proportion of rotavirus positive children hospitalized for 3–6 days as compared to rotavirus negative children (p < .01) at both sites. There were no deaths among rotavirus positive cases in either site, while 3 deaths occurred among rotavirus negative AGE cases in Odesa, one of which was diagnosed with salmonellosis, while a definite etiology was not found for the remaining 2 cases.

Seasonal fluctuations in the frequency of rotavirus AGE were observed both in Kyiv and in Odesa (Figs. 1a, 1b). Peak rotavirus activity occurred in the winter months in both sites, while the lowest proportion of detections occurred during the summer.

In all years at both sites, 5 predominant genotypes were identified. G4P[8] was most commonly identified at both sites, ranging from 61–78% of detections in Kyiv and 7–69% in Odesa. This was followed by G1P[8], G2P[4], G3P[8], and G9P[8]. All together, these 5 genotypes accounted for 97% of identified strains in Kyiv and 95% in Odesa (Table 3a, 3b).

During 9 years of active AGE surveillance, rotavirus infection was the predominant etiology for AGE hospitalizations among children younger than 5 years of age in two large Ukrainian cities. These rotavirus AGE hospitalizations followed a typical seasonality seen in similar geographical settings, with increased activity in cooler months [12]. In both Kyiv and Odesa, the proportion of rotavirus AGE among hospitalized children was high and remained elevated during the surveillance period. In Odesa, when surveillance activities increased in 2013 to daily from 3 days per week, there was an increase in the total numbers of rotavirus positive children, though the proportion that were rotavirus positive remained between 39–42% for these subsequent years. In Ukraine the proportion of rotavirus diarrhea in hospitalized children younger than 5 years of age was the highest among countries of the newly independent states included in the Global Rotavirus Surveillance Network [2,3]. Data from the newly independent states indicates that in Ukraine, among children younger than 5 years hospitalized with diarrhea, rotavirus was detected in 45% of all cases. Before rotavirus vaccine introduction, 38% of such detections were observed in Tajikistan, 40% in Georgia, 38% in Armenia, and 45% in Republic of Moldova [3,8,9]. With its large birth cohort, estimated at 475,052 for 2017 [13], in comparison to these countries, the potential benefits of rotavirus vaccine introduction in Ukraine could be substantial.

We found that rotavirus positive children had a median age between 21 and 24 months and that of all the rotavirus positive children enrolled in the surveillance systems, cumulatively 57% and 50% had been infected by rotavirus by 2 years of age in Kyiv and Odesa, respectively. This is slightly older than that seen in other regions in the pre-vaccine era. In a South African literature review including 11,682 cases, 75% of rotavirus infections were seen by 1 year of age [14]. In a multi-year US analysis of hospitalization discharge data, which saw an average of 184,097 AGE hospitalizations peryear, 75% were seen by 2 years of age [15]. Finally, a Latin American review including 1949 children reported 90% infections by 2 years of age [16]. However, similar numbers were seen in pre-vaccine data reported from Armenia and the Republic of Moldova [8,9]. One possible explanation for this variation in age distribution could be that overall fewer children in the youngest age categories were enrolled into this study. Despite the variance in age groups in comparison to other areas, these results reveal the substantial burden of rotavirus infection for young children in two of the most populated cities in Ukraine.

Rotavirus AGE was characterized by more severe symptoms than non-rotavirus AGE, as has been reported elsewhere [17,18]. A higher proportion of rotavirus positive children experienced 6 or more episodes of diarrhea in a 24 h period compared to their negative counterparts, as well as more reports of fever, vomiting and number of vomiting episodes in a 24 h period. A higher proportion of the rotavirus positive children were admitted for 3–6 days as compared to their negative counterparts, who were admitted for shorter stays. Overall a significantly higher proportion of rotavirus positive children received the highest scores on the 2- and 3-level calculated severity groups. Nearly all children in Odesa received treatment with IV hydration, while a smaller proportion did in Kyiv, though the rotavirus positive children received it more often that the rotavirus negative children. This was likely due to intravenous hydration being the standard of care in Odesa, but not in Kyiv.

Five main genotypes – G1P[8], G2P[4], G3P[8], G4P[8] and G9P[8] were associated with rotavirus gastroenteritis. Overall these genotypes were detected in 97% and 95% of cases in Kyiv and Odesa respectively, during the surveillance period. In Kyiv G4P[8] remained the leading genotype in all years, while in Odesa there was variation between years among these leading genotypes. Similar findings were reported from a study of 210 stool samples from children with diarrhea from different regions of Ukraine during 2006–2009 [19]. Rotaviruses with these four genotypes G1P [8], G2P[4], G3P[8], and G9P[8] were commonly reported during 2000–2009 in Asia [20]. The predominant genotypes from June 2006 through December 2008 were G1P [8], G2P [4], and G9P [8] for GRSN-participating countries in Africa [21].

This analysis was subject to several limitations. Enrollment practices changed over time in Odesa, with initially thrice weekly enrollment, changing to daily enrollment by 2013. From 2011 onwards, enrollment at the 2 sentinel sites was quite similar though Odesa experienced lower enrollment in 2010 because the facility was undergoing repairs. Despite these factors affecting enrollment, however, we were able to demonstrate a predictable seasonality. Another limitation in this analysis was that information was not systematically collected on all variables (recorded temperature, dehydration status) for the Vesikari score, as such a modified scale was used, which may not capture dehydration status as accurately. Another variable not recorded was the vaccine status of these children, however, given the very low national coverage reported, this likely did not affect our analysis. Finally, this surveillance is hospital-based, and as such is not representative of the epidemiology of milder cases not requiring hospitalization; however, as rotavirus vaccine is effective against severe rotavirus disease requiring hospitalizations, this surveillance does capture the children at highest risk of severe complications from rotavirus.

Active epidemiologic surveillance of AGE in hospitalized children younger 5 years in Kyiv and Odessa reveals a significant burden of rotavirus infection. These data provide the scientific background supporting the implementation of universal vaccination against rotavirus infection in Ukraine.