Multicenter Prospective Study on the Burden of Rotavirus Gastroenteritis in Turkey, 2005–2006: A Hospital-Based Study

Abstract

Acute diarrhea in early childhood is among the most important challenges in public health [1] and is considered by the World Health Organization (WHO) to be one of the leading causes of childhood mortality [2]. Among all enteric pathogens that infect young children, rotavirus is the primary cause of severe gastroenteritis worldwide, affecting nearly all children by age 5 years and accounting for ∼39% (range, 29%–45%) of all diarrhea-related hospitalizations [3]. Although the incidence of rotavirus is similar in developing and developed countries, the mortality rate is higher in developing countries, likely because of poor access to hydration therapy and a high prevalence of malnutrition [1]. Global mortality due to rotavirus is estimated to be 527,000 deaths (range, 475,000–580,000 deaths) among children aged <5 years, with >80% of the deaths occurring in the developing countries of South Asia and sub-Saharan Africa [4,5]

Two vaccines (RotaTeq [Merck], licensed in the United States, Europe, and >69 countries worldwide, and Rotarix [GlaxoSmithKline Biologicals], licensed in the United States, Europe, and >100 countries worldwide) have demonstrated good safety, immunological, and efficacy profiles in large clinical trials conducted in western industrialized countries and in Latin America [6–12]. The WHO has given high priority to rotavirus vaccination, and both Rotarix and RotaTeq have been prequalified for use in developing countries [13, 14]

Epidemiological data on rotavirus disease burden are useful for guiding rotavirus vaccination recommendations, including assessment of the need for vaccination and the impact of the vaccine in reducing the burden of rotavirus disease [15]. Numerous reports have been published on the epidemiology of rotavirus infection in various regions of the world, including Europe and Middle East. However, few studies report the clinical and epidemiological features of rotavirus gastroenteritis in Turkey [15, 16], and other publications focus on the laboratory diagnosis of rotavirus disease [17–19]

Turkey is a developing country with a population of 73 million people and a diarrhea-associated mortality rate of 3.84 deaths per 1000 children aged <5 years [20]. A model described by Parashar et al [1] projects 13,371,800 episodes of diarrhea annually in Turkey among children aged <5 years, corresponding to 94,817 hospitalizations and 1,182,046 outpatient visits

The primary objective of this study was to estimate the proportion of diarrhea-related hospitalizations attributable to rotavirus among children aged <5 years who reside in Turkey, with the use of local data as opposed to models extrapolated from other sources. Other objectives included the determination of age and seasonal distribution of rotavirus hospitalizations and the identification of the prevalent rotavirus strains

Materials and Methods

Study designThis prospective, multicenter, hospital-based study was conducted among children aged <5 years who were hospitalized for gastroenteritis in 4 hospitals located in 4 major cities in Turkey (Ankara, Istanbul, Izmir, and Adana); these hospitals serve 37% of the population aged <5 years in Turkey. Data were collected for a 12-month period from 1 June 2005 through 1 June 2006. The study design was based on the WHO 2002 generic protocol for hospital-based surveillance to estimate the burden of rotavirus gastroenteritis among children and was adapted on the basis of the country requirements [21]

The study was conducted in accordance with “good clinical practice,” the 1996 version of the Declaration of Helsinki, and local regulatory requirements and was approved by the local ethics committee of each study hospital. Informed consent was obtained from the parents or guardians of all eligible children. This study was sponsored by GlaxoSmithKline Biologicals, Belgium

Inclusion and exclusion criteriaAll children aged <5 years who were hospitalized for gastroenteritis at the 4 study hospitals were enrolled in the study. Gastroenteritis was defined by diarrhea (⩾3 looser-than-normal stools per day) with or without vomiting (⩾1 episode of forceful emptying of partially digested stomach contents ⩾1 h after feeding per day). Children were ineligible for inclusion if the diagnosis for treatment at the study hospital did not include gastroenteritis or if the onset of gastroenteritis was after hospitalization

Data collectionThe parents or guardians of the children enrolled in the study were asked to complete a questionnaire to collect information regarding the child’s demographic characteristics, medical history, gastroenteritis episode, and area of residence. The severity of gastroenteritis was assessed by the investigators using the Vesikari scale [22]. Stool samples were collected from all children who were able to provide them at the time of recruitment and were tested for the presence of rotavirus by enzyme-linked immunosorbent assay at the GlaxoSmithKline Biologicals laboratory (Rixensart, Belgium). All rotavirus-positive samples were tested by reverse-transcriptase polymerase chain reaction for the VP7 and VP4 genes, followed by a reverse hybridization assay at DDL Diagnostic Laboratory (Delft, the Netherlands) to identify G and P types [23]

Statistical analysesThe proportion of rotavirus gastroenteritis cases among all diarrhea-related hospitalizations of children aged <5 years overall and by study center was computed with exact 95% confidence intervals by using the total number of children aged <5 years who were hospitalized for gastroenteritis as the denominator. Statistical analyses were performed using SAS, version 8.2 (SAS) [24]

Results

A total of 411 children were hospitalized for gastroenteritis in the 4 study hospitals and were assessed for rotavirus infection. Of these children, 47 were excluded because they declined to participate or did not sign the consent form, 2 were aged >5 years, and 1 received a diagnosis for treatment at the study hospital that did not include gastroenteritis. The median age of the 361 children enrolled in the study was 12 months (range, 0–60 months), 60.9% were male, and 91.7% lived in the same city as the hospital to which they were admitted

Stool samples were not available for 2 children. Of the 359 children with a stool sample, 15 had samples that were untested or were insufficient in quantity, and 6 samples were lost, yielding 338 children with stool samples available for analysis. The overall proportion of rotavirus gastroenteritis among all children hospitalized for gastroenteritis who were enrolled in the study and were tested for rotavirus was 53% (179 of 338), with the lowest proportion observed in Adana (32.4%) and the highest proportion observed in Izmir (67.4%) (Table 1)

Table 1

Proportion of Rotavirus Gastroenteritis among 338 Children Hospitalized for Gastroenteritis and Tested for Rotavirus

Open in new tabDownload slide

Of the 179 children with rotavirus-positive stool samples, 83.7% were aged <2 years, 43.5% were aged <1 year, and 11.7% were aged <6 months (Figure 1). Rotavirus gastroenteritis was most frequently observed between October and April and peaked in February (Figure 2). Of the 179 rotavirus-positive stool samples that were available for typing, wild-type G1P[8] was the most widely prevalent strain (76% of strains) (Figure 3), followed by G2P[4] (12.8%). G9P[8] was reported in samples from 3.9% of children

Figure 1

Distribution of children by age and by rotavirus (RV) status among the 179 RV-positive children and 159 RV-negative children

Open in new tabDownload slide

Figure 2

Seasonal distribution of 361 gastroenteritis (GE) cases. RV, rotavirus

Open in new tabDownload slide

Figure 3

Rotavirus (RV) strain distribution among 179 children with RV-positive gastroenteritis (GE)

Open in new tabDownload slide

Before hospitalization, 161 (89.9%) of 179 rotavirus-positive children and 133 (83.6%) of 159 rotavirus-negative children were reported to have had a “severe” episode of gastroenteritis, which was assessed using the 20-point Vesikari scale. Vomiting was reported in 165 (92.2%) of 179 rotavirus-positive children and 131 (82.4%) of 159 rotavirus-negative children. Severe dehydration was observed in 23 (12.8%) of 179 rotavirus-positive children and 13 (8.2%) of 159 rotavirus-negative children

The median duration of hospitalization was 3 days, regardless of rotavirus status. During hospitalization, 164 (91.6%) of 179 rotavirus-positive children and 145 (91.2%) of 159 rotavirus-negative children received intravenous rehydration. Also, 16 (8.9%) of 179 rotavirus-positive children and 7 (4.4%) of 159 rotavirus-negative children had ongoing gastroenteritis during discharge from the hospital. One rotavirus-negative child, who received a diagnosis of juvenile myelomonocytic leukemia and febrile neutropenia, died during the study

Discussion

We estimated the burden of rotavirus among young children in Turkey, using a hospital-based surveillance study. Of the 338 children with samples available for laboratory confirmation, 53% tested positive for rotavirus, which is higher than previously reported proportions for rotavirus gastroenteritis in Turkey (39.8%, reported by Kurugol et al [15], and 36.8%, reported by Karadag et al [16]). Although this may reflect an increasing burden of rotavirus gastroenteritis in Turkey in recent years, the proportion of acute gastroenteritis hospitalizations attributable to rotavirus is dependent on the prevalence of other causes of diarrhea-related hospitalizations, which may vary over time. However, our estimates are comparable to results of more-recent hospital-based surveillance studies in Europe [25, 26]. In the Rotavirus Gastroenteritis Epidemiology and Viral Types in Europe Accounting for Losses in Public Health and Society (REVEAL) study, conducted in 7 European countries in 2004–2005, rotavirus accounted for 53.1%–68.8% of hospitalizations for acute gastroenteritis among children aged <5 years [25]. In the Surveillance for Hospitalised Rotavirus Infections in Kids (SHRIK) study, conducted in 12 hospitals in France, Germany, Italy, Spain, and the United Kingdom from February 2005 through August 2006, 56.2% of children aged <5 years who were hospitalized with acute gastroenteritis were found to be positive for rotavirus (range by country, 33.2%–64.4%) [26]

The proportion of rotavirus gastroenteritis varied between the study areas, ranging from 32.4% in Adana to 67.4% in Izmir. These regional differences may be because of differences in health care systems or cultural practices—for example, in some areas, children may have been more likely to be treated at home and not admitted to the hospital, which may have caused the proportion of rotavirus gastroenteritis cases to decrease [25, 26]. Because rotavirus gastroenteritis is treated mainly with oral rehydration, children with rotavirus gastroenteritis may be more likely to be treated at home than children with bacterial gastroenteritis, which may require antibiotic therapy

Most of the rotavirus-positive children were reported to be aged <2 years in this study, which is consistent with the previous literature [15, 20, 25–27]. As in most countries with a temperate climate, rotavirus gastroenteritis was observed to have a clear seasonal pattern, peaking in the winter months between October and April and at the lowest levels during June and July

In this study, wild-type G1P[8] was the most widely prevalent strain (76%), followed by G2, G4, and G9. These findings are similar to those reported for 2000–2001, in which G1 was the most common type (75.1%), followed by G2, G3, and G4 [15]. A review of 124 studies conducted from 1989 through 2004 in 52 countries found that strains with types G1, G3, and G4 in combination with type P[8] and strains with type G2 in combination with type P[4] were responsible for >88% of all rotavirus infections worldwide [28]. In contrast to previous local studies [15], G9 (3.9% of strains) is reported to be an emerging type in Turkey. This is supported by the global review that detected type G9 in every continent, with an overall percentage of 4.1% [28]. Another recent Turkish study in Ankara conducted from September 2004 through December 2005 showed the overall prevalence of G9 to be 17.2% [29]. The authors described a shift of frequency of detection of G9 from 5.3% in 2004 to 17.4% in 2005, and 75% of the enrolled children in that study were outpatients [29]. Our data for inpatients, also in Ankara, showed that 57.4% were rotavirus positive, and only one G9P[8] strain was identified

This study is the first, to our knowledge, to provide comprehensive information on the rotavirus disease burden for 4 major cities in Turkey. Its strengths include standardized study procedures, including enrollment of children at all participating hospitals. A potential limitation is the relatively low number of children hospitalized for gastroenteritis (411) and the fact that only 338 of these children provided stool samples for rotavirus analysis. It is possible that this may have introduced selection bias and that these 338 children may have had a different proportion of laboratory-confirmed rotavirus, compared with all children hospitalized for gastroenteritis. The overall proportion of rotavirus gastroenteritis calculated for all children hospitalized for gastroenteritis in this study (enrolled and nonenrolled children) was 43.6% (179 of 411)

Nevertheless, results of this study serve as good baseline information on the rotavirus disease burden in Turkey. The incidence of rotavirus disease is similar in both industrialized and developing countries, which suggests that improvements in water supply, hygiene, and sanitation may not adequately control the disease [3]. Effective immunization against rotavirus is an important public health measure and should be a priority in countries where disease burden, mortality, and morbidity are high. The introduction and universal application of rotavirus vaccine could provide a measurable and visible benefit to the health of a population within 2–3 years [30, 31] and will make a potentially positive contribution to the United Nations Millennium Declaration—which is to reduce the mortality rate among children aged <5 years by two-thirds by 2015 [32]. Turkey, where diarrheal diseases are the second most common infectious cause of childhood mortality [33], has the infrastructure and now has the supportive data to include a rotavirus vaccine in the national immunization program, in accordance with WHO recommendations

Acknowledgments

We thank Ozlem Ercetin, for local monitoring support; Kusuma Gopala and Priya Rajagopal, for their support in statistical analysis; Taara Madhavan, for assistance in preparing the manuscript; and Slavka Baronikova and Manjula K, for editorial support (all from GlaxoSmithKline Biologicals). We also thank DDL Diagnostic Laboratory (Delft, the Netherlands) for genotyping all rotavirus-positive samples

References