Epidemiology of Rotavirus Infections in Children Less Than 5 Years of Age

Germany, 2001–2008

Abstract

Background: 

Rotavirus (RV) infection is the leading cause of acute gastroenteritis in young children worldwide. In 2006, 2 live-attenuated RV-vaccines became available for use in infants ≤6 months of age. In Germany, a statutory notification system for RV infection has been in place since 2001 to monitor RV epidemiology. Our objective was to assess RV disease burden in German children <5 years of age.

Methods: 

Detailed descriptive analysis of national RV surveillance data in children <5 years of age collected in Germany between January 2001 and December 2008.

Results: 

Between 2001 and 2008, 72% of all notified RV patients were children <5 years of age. The highest annual incidence (approximately, 200/10,000) was in children <2 years of age. In the <5 years age-group, approximately 50% of reported patients were hospitalized and of those, 9% acquired the infection nosocomially. Since 2004, a total of 8 children <5 years of age were reported as RV-associated deaths, and case fatality due to RV infection was <0.01/10,000.

Conclusions: 

The high incidence of RV infection and RV-associated hospitalization in children <5 years of age results in a high disease burden. Routine childhood RV vaccination would be a measure to reduce the burden in this age-group. However, cost-effectiveness analyses specific to the German setting should be considered in the decision-making process. An RV-surveillance system is in place in Germany that could potentially monitor the effect of an RV-vaccination program once implemented.

Rotavirus (RV) infection is the leading cause of infectious diarrhea among infants and young children worldwide.^1,2 It is estimated that each year, RV causes globally 111 million episodes of gastroenteritis in children <5 years of age that require home care, 25 million clinic visits, 2 million hospitalizations, and 440,000 deaths. By the age of 5 years, nearly every child has an episode of RV gastroenteritis.³ In European children <5 years of age, a model suggests that each year 3.6 million episodes of RV disease, 700,000 outpatient visits, >87,000 hospitalizations, and 231 deaths occur.⁴ Although rarely fatal in industrialized countries, RV disease remains a major cause of severe dehydration and hospitalization and contributes significantly to health care costs among young children.⁵

In 2006, 2 RV vaccines, a human vaccine developed by GlaxoSmithKline Biologicals (Rotarix) and a human-bovine vaccine developed by Merck (RotaTeq), were licensed for use in Europe.^6,7 Both vaccines have shown high efficacy and good safety profiles in clinical trials.^8–10 In April 2009, the World Health Organization's (WHO) Strategic Advisory Group of Experts on immunization recommended that RV vaccination of infants should be included in all national immunization programs.¹¹ In addition, it has also recommended that a surveillance system for severe RV gastroenteritis should be in place to monitor the vaccines' impact. In Germany, adoption of RV vaccination in the routine childhood immunization schedule is currently still under consideration by the German Standing Committee on Vaccination.

In 2001, RV infection became a notifiable disease in Germany under the new German Protection against Infection Act.¹² The surveillance system consists of weekly reporting of case-based data on acute RV infections from local via state health authorities to the Robert Koch Institute (RKI), the federal institution responsible for disease control and prevention in Germany.¹³ We analyzed RV-surveillance data reported from January 2001 to December 2008 with the objectives to describe the epidemiology of RV and to assess RV-disease burden focused on hospitalization rates among children <5 years of age in Germany to support vaccine introduction decisions. We assessed whether the current surveillance system for severe RV gastroenteritis is suitable for monitoring the effect of routine RV vaccination, should this be introduced into the German childhood immunization schedule.

MATERIALS AND METHODS

Since 2001, acute RV infection presenting with diarrhea and/or vomiting is notifiable in Germany.¹¹ According to a standardized case definition, RV infection presenting with 1 of the 2 above-mentioned clinical manifestations must be confirmed either by the detection of RV antigen in stool using enzyme immunoassays, polymerase chain reaction methods, or electron microscopy, or the case must be epidemiologically linked to a laboratory-confirmed case of an outbreak.¹⁴ Laboratories routinely report confirmed RV infection to the local health office. If the case meets the case definition, a case-based report form is electronically forwarded by the local health authority within 1 week via the federal state authority to the RKI. A minimal data set includes information on age, sex, onset of symptoms, hospitalization, fatal outcome, and laboratory diagnostic method.

A detailed descriptive analysis of RV-surveillance data was performed to describe the distribution of RV by age, sex, season, and among cases the proportion of hospitalization, nosocomial infections, and fatal outcome for January 2001 to December 2008. An infection was defined as nosocomially acquired if the onset of RV diseases occurred >48 hours after hospital admission. Cases were classified as nosocomial infections by using date of symptoms onset, date of hospital admission, and a medium incubation period for RV infection of 48 hours.

Our analysis focused on children <5 years of age because this group is reported to be mainly affected by RV and is of most interest as the licensed RV vaccines target infants and young children. Surveillance data of RV infection was compared with data of other notifiable acute gastroenteritis (AGE) (notifiable pathogens include: Shigella, Salmonella, Yersinia, Giardia, norovirus, Campylobacter, E. coli, and enterohemorrhagic E. coli), which are collected in a similar manner using standardized case definitions for laboratory confirmed symptomatic infections. Between 2001 and 2008, the average German population added up to 82,378,825 inhabitants (range: 82,002,356–82,536,680), the mean number of the annual birth cohort accounted for 699,301 infants (range: 673,132–735,755) and the mean number of children <5 years of age was 3,633,421 (range: 3,445,172–3,892,984).¹⁵ Incidence rates (per 10,000 population) were based on the number of children in the respective age-group during the period of surveillance.

Data analysis was conducted using MS Excel and [email protected]. [email protected] is the electronic reporting system established at RKI.¹³

RESULTS

More than two-third (72%) of the total 442,199 RV cases reported between 2001 and 2008 in Germany were diagnosed in children <5 years of age. With 316,874 notified cases and a mean annual incidence of 115/10,000 children (range: 82–138), RV infection was the most frequently reported disease in Germany in children <5 years of age between January 2001 and December 2008 followed by infections with norovirus (38/10,000; range: 4–1,001), Salmonella (35/10,000; range: 22–50), and Campylobacter (14/10,000; range 11–15).

Information on age, sex, and week of notification were available for 99%, on hospitalization for 97%, and on disease outcome for 96% of cases. The majority (96.1%) of the 316,874 reported RV cases fulfilled the clinical case definition and was laboratory confirmed, whereas 3.9% fulfilled the clinical case definition and were epidemiologically linked to a laboratory-confirmed case. RV infection was confirmed in 98.6% by Ag-enzyme immunoassay, in 1.0% by polymerase chain reaction, and in 0.4% by electron microscopy.

Seasonality and Annually Reported Number of RV Infections

Between 2001 and 2008, the annually reported number and incidence of RV infection in children <5 years of age ranged from a minimum of 28,134 cases (incidence, 82/10,000) in 2004 to a maximum of 47,619 cases (incidence, 138/10,000) in 2008 (average annual incidence: 115/10,000) (Table 1). Years with the highest reported number of RV infections were all found after 2004, the year when the German reimbursement system for hospitalization was changed. The number of RV infection from 2004 to 2008 was significantly higher than from 2001 to 2003 (P < 0.01).

TABLE 1: Age-Specific Incidence and Total Number of All Notified Rotavirus Infections and Rotavirus Hospitalizations in Age Groups of Children Aged <5 Years; Germany, 2001 to 2008

RV infections were reported over the entire observation period but there was a clear seasonal trend with a dominant late winter and early spring peak from January to May, during which 83% of all cases were reported (Fig. 1). The maximum of weekly reported RV cases was observed in March (week 10: median 2201 cases; range: 1263–3501) and the minimum in August (week 33: median 98 cases; range: 54–123) (Fig. 2).

FIGURE 1.: Monthly number of notified RV infections for outpatients and hospitalized patients in children <5 years of age; Germany, 2001 to 2008.

FIGURE 2.: Median (minimum, maximum) number of notified RV infections in children <5 years of age by week; Germany, 2001 to 2008.

Age and Sex Characteristics of Reported RV Cases

Average annual RV incidences in children <5 years of age between 2001 and 2008 were highest in infants (198/10,000) and 1-year-old children (206/10,000) (Table 1). In 2-year-old children (98/10,000) RV infection was half as frequent as in the youngest but twice as high as in 3-year-old children (48/10,000) and 3 times as high as in 4-year-old children (28/10,000). In contrast, incidences in age groups ≥5 years were all below 10/10,000. In children <5 years of age RV incidence was slightly higher in males than in females (121/10,000 vs. 108/10,000; P < 0.001).

Hospitalization of RV Cases

From 2001 to 2008, almost half (45%) of the children <5 years of age reported with RV infection have been treated in hospital or had acquired RV nosocomially; this proportion decreased with increasing age from 52% in infants <1 year of age to 37% in 4-year-old children. During this observation period, an average of 17,636 hospitalizations (incidence: 51/10,000) attributable to RV were annually reported in children <5 years of age (Table 1); case number and incidence ranged from a minimum of 12,383 cases (incidence 36/10,000) in 2004 to a maximum of 23,263 hospitalized cases (incidence 67/10,000) in 2008. For the years 2004 to 2008 the proportion of cases hospitalized (48%) (2004: reimbursement system for hospitalization changed) was significantly higher compared with the years 2001 to 2003 (40%; P < 0.01). Three-fourth (75%) of hospitalized children <5 years of age with RV infection were ≤2 years old. Figure 3 is illustrating the distribution of hospitalized RV cases by age in months. Already in newborns hospitalization due to RV infections was high (incidence, 76/10,000). The average annual number of RV cases steadily increased with age, reached a plateau in 6- to 12-month-old children, and continuously declined in >12-month-old children. From 2001 to 2008, the average annual age-specific incidence of RV hospitalization in infants below 12 months of age was 103/10,000, in 1-year-old children 87/10,000, in 2-year-old children 38/10,000, and in children >2 years of age was <20/10,000 (Fig. 4).

FIGURE 3.: Average annual number of hospitalized RV infections in children <5 years of age by age in months; Germany, 2001 to 2008.

When comparing cases before and after 2004 (the year when the German reimbursement system for hospitalization was changed), the mean annual incidence of reported hospitalizations among all RV notifications in children <5 years of age increased from 42/10,000 to 60/10,000 (P < 0.01). The median duration of hospitalization in community acquired RV infection was 3 days for cases <2 years of age and 2 days for cases aged 2 to 4 years.

In comparison to other notifiable AGE (Shigella, Salmonella, Yersinia, Giardia, norovirus, Campylobacter, E. coli, and enterohemorrhagic E. coli), RV was the most frequently diagnosed pathogen among hospitalized children <5 years of age in Germany. RV was responsible for 70% (average annual incidence: 51/10,000) of all reported hospitalizations due to one of the notifiable AGE, followed by norovirus with 13% (9.7/10,000), Salmonella with 11% (7.7/10,000), and finally Campylobacter and E. coli both with 2% (1.6/10,000). Among all AGE hospitalizations with notifiable pathogens, the proportion of hospitalization due to RV infection was highest in <1-year-old children (76%) (Table 2). This proportion declined with increasing age. In 4-year-old children, half of hospitalized AGE cases (49%) were still caused by RV.

TABLE 2: Mean Annual Incidence of Hospitalizations Stratified by Age for Acute Gastroenteritis (AGE) and Rotavirus Infection in Children <5 Years of Age; Germany, 2001 to 2008

Nosocomially Acquired RV Infections

For the classification of nosocomially acquired RV infection, information on date of symptom onset and date of hospital admission was available for 72% of all notified hospitalized RV cases <5 years of age. Nosocomial infections were significantly more common among children <2 years of age when compared with children aged 2 to 4 years (10.7% vs. 4.8% of hospitalized RV cases, P < 0.01). The proportion of nosocomially acquired RV infections significantly decreased over the observation period for both age-groups: In children <2 years of age, the proportion declined from 13.3% to 8.6%, and in 2- to 4-year-old children from 7.3% to 3.6% (for both groups, χ² test for trend, P < 0.001). Among children <2 years of age, the median age of nosocomial cases was lower than that of hospitalized cases with community-acquired RV infection (6 and 12 months, respectively).

RV-associated Mortality

Since 2004, RKI has systematically verified fatal cases reported to be associated with RV. Between 2004 and 2008, 8 deaths attributable to RV infection were reported in children <5 years of age (case fatality rate: <0.001%); of these, 7 in children <1 year of age and 1 fatal case in a 1-year-old child).

DISCUSSION

The German RV surveillance system provides robust data for a comprehensive and differentiated monitoring of RV disease since its implementation in 2001. The findings of this analysis indicate that RV infection is the leading cause of diarrhea in children <5 years of age and constitutes a substantial public health problem in Germany. Our data are in agreement with results of previous studies and mathematical models estimating RV-associated disease burden.^3,16,17 The high proportion of RV diagnoses among hospitalized AGE cases <5 years of age documents the effect of RV on the health of children and suggests a high financial burden for the German health care system.^4,18–21

The true burden of RV disease is likely to be substantially underestimated by the German surveillance system. One reason for the underestimation is that mild RV infections are often managed at home and are not presented to a physician. In population-based studies of infectious gastroenteritis in Europe, only 5% to 20% of those who develop symptoms of AGE consult a primary care physician or other health care professionals.²² Furthermore, there is little incentive for physicians to collect stool samples from AGE patients because the detection of RV in stool does not influence medical treatment.²³ Therefore, a high number of mild or moderate RV diseases will not be detected by a passive surveillance system. As most mild or moderate RV infections do not require hospitalization, it can be assumed that underestimation is in particular a problem for the assessment of the outpatient disease burden. Still, we believe that the reporting of outpatient data is strength of our RV surveillance system, since nationwide long-term outpatient RV data are lacking in the scientific literature but are important for the economic evaluation of an RV vaccination program. Therefore, our outpatient data must be regarded as a lower-limit estimate for number of RV cases presenting at outpatient clinics and can be used as a baseline to monitor future trends.

The German RV surveillance system produces robust data regarding the number of severe RV infections in young children requiring hospitalization. It is a population-based passive surveillance system collecting data of laboratory confirmed cases, performed throughout the calendar year over a prolonged period. The surveillance data of 2001 to 2008 show that on an average 17,635 (range: 12,383–23,263) children <5 years of age are hospitalized due to RV infection each year. These numbers are similar to that predicted by a mathematical model estimating the annual number of RV cases in children <5 years of age in European countries.⁴ In this model, it is estimated that 13,573 hospitalizations due to RV infection might occur annually in Germany in children <5 years of age.

The observed increase in RV notifications since 2004 might be related with changes in the German reimbursement system for hospitalization charges. In 2004, diagnosis-related groups—a system to classify hospitalized patients—were implemented in German hospitals and act as a basis for payment of services provided by the hospitals. As hospitals receive higher payments for each AGE case with a confirmed causative agent than for microbiologically unconfirmed AGE, a higher number of stool samples might have been tested for RV in German hospitals and might have resulted in a higher number of notified RV cases since 2004.

There is no evidence that the incidence of RV disease in Europe has fallen over the past years and efforts and recommendations for an improved hygiene and sanitation management has not helped to reduce the number of RV infections.²⁴ However, our data analysis revealed that the nosocomial infection rate in children <5 years of age significantly declined during the observation period from 2001 to 2008. For these developments different factors may play a role. On the one hand, it could be possible that improved hygiene management like strict cohort isolation at an early stage of symptomatic AGE might have contributed to this trend. On the other hand, modifications of the German reimbursement system for hospitalization charges, as mentioned earlier, might have generated a higher number of diagnosed RV infections in inpatients and therefore the proportion of nosocomial cases could be reduced.

Studies suggest that RV vaccination protects from severe disease, even if the genotype of the causative RV differs from those in available vaccines.²⁵ After the introduction of RV vaccination in the United States for routine immunization in children under the age of 6 months in February 2006, the onset of the RV season 2007/2008 was delayed by 2 to 4 months and diminished in magnitude by 50%.²⁶ Also in Austria the beneficial effect of adopting RV vaccination in the childhood immunization schedule was shown.¹⁷

In 2007, WHO strongly recommended to introduce RV vaccination into the national immunization programs, where vaccine efficacy suggests a significant public health impact, the vaccination program can be sustained, and where adequate surveillance is in place.²⁷ Based on an updated review of evidence, the WHO recommendations were expanded and the inclusion of RV vaccination into all national immunization programs was suggested in 2009.¹¹ In Germany, the introduction of routine RV vaccination in infancy has the potential to reduce a large number of RV-associated hospitalizations and emergency consultations, and to save direct and indirect costs that are associated with RV disease in young children. Cost-effectiveness analyses and modeling of RV disease burden for Germany, that are independently performed from vaccine manufacturers and that have taken into consideration various scenarios of vaccination coverage and vaccine prices are lacking or have not been published in peer-reviewed journals yet. Data collected in the national surveillance system and presented here are fundamental to inform such models.

In conclusion, the established RV surveillance system demonstrates that RV causes a substantial disease burden in children aged <5 years in Germany mainly due to a high overall incidence in this age-group and high number of hospitalizations. Based on these data, implementation of routine RV vaccination might be considered for Germany but additional analyses on the cost-effectiveness of routine RV vaccination and the potential effect on disease transmission dynamics should be taken into consideration as well. If RV vaccination should be adopted in the German childhood immunization schedule in the future, a stable RV surveillance system will be already in place to provide pre- and postimplementation period data on RV epidemiology, especially in respect to severe cases which require hospitalization. Based on these data and in combination with an already established molecular surveillance system, the effect of RV vaccination could be monitored sufficiently.

FIGURE 4.: Annual age specific incidence of all RV infections for outpatients and hospitalized cases by age for children <5 years of age; Germany, 2001 to 2008.

ACKNOWLEDGMENTS

The authors thank Susanne Behnke for her support in data management and Manuel Dehnert and Ole Wichmann for reviewing the article.