The main study objective was to determine the number of NGE hospitalizations overall and stratified by age (<1, 1–17, 18–44, 45–64, 65–84 and >85 years) in Germany by using hospital discharge data obtained from DESTATIS. German inpatient data are recorded centrally and provided to DESTATIS. All hospitals are required by law to report patient information (for reimbursement purposes) on age, sex, duration of hospital stay, reason for admission, and discharge based on codes from the International Classification of Diseases, 10th Revision (ICD-10), and the Diagnosis-Related Group (DRG). We requested information on all hospitalizations resulting from AGE (ICD-10 codes A00–A09, representing all infectious intestinal diseases) for the period 2007–2012. In addition, we specifically requested information on hospitalizations with ICD-10 code A08.1 (NGE) as the primary or secondary diagnosis and a DRG code of G67 (esophagitis, gastroenteritis, gastrointestinal bleeding, ulcers, and miscellaneous digestive system disorders). Including the base DRG code G67 enabled us to differentiate between the different DRG codes in the same base DRG group (i.e., G67 A–D), which are linked to different reimbursement amounts. For comparison, we requested similar information on rotavirus hospitalizations (ICD-10 code A08.0). We also requested data on deaths related to the same codes as the primary diagnosis.

We assessed the direct hospitalization costs associated with NGE hospitalizations in Germany (primary diagnosis), which were available for the years 2007–2009. The DRG reimbursement data from DESTATIS are independent of the duration of hospital stay and do not include surcharges based on long-stay DRG or discounts based on short-stay DRG.

To validate the reporting system for NGE cases, we reviewed the clinical files of a random selection of 1,214 patients at the University Medical Centre Mainz (Mainz, Germany) who had been tested for norovirus. Demographic, clinical, and microbiologic information was extracted by using a standardized case report form.

The study protocol was reviewed by the local ethics committee in Mainz (Landesärztekammer Rheinland-Pfalz). Because this study was noninterventional and all data were anonymized and aggregated, a waiver for ethics approval was obtained.

We identified a total of 408,124 hospitalizations with an NGE diagnosis in Germany during 2007–2012, for an annual average of 68,187 hospitalizations (Table 1). The overall number of NGE hospitalizations were consistently higher than the number of rotavirus hospitalizations for all age groups except children <2 years of age. Furthermore, for each of the study years, more deaths among hospitalized patients with an NGE diagnosis occurred than among those with rotavirus AGE (Table 1).

Hospitalizations with NGE as Primary Diagnosis

The annual number of hospitalizations with NGE as the primary diagnosis ranged from 21,442 to 33,440 (average 27,910) (Technical Appendix Table 1). NGE hospitalizations represented 11.5%–15.6% of all hospitalizations with AGE as the primary diagnosis annually. The overall incidence of hospitalization with NGE as the primary diagnosis in Germany during the study period was 3.4/10,000 population (range 2.6–4.1/10,000 population) (Technical Appendix Table 1).

Figure 1

Figure 1. Annual number and incidence of hospitalizations for norovirus gastroenteritis as primary diagnosis, Germany, 2007–2012. A) Annual cases by age group; B) annual incidence (no. cases/10,000 population) by age group. Inset shows...

Among children <1 year of age, annual incidence ranged from 34.8 to 53.1/10,000 population, and among children 1 to <2 years of age, the annual incidence ranged from 41.8 to 52.0/10,000 population (Figure 1). After the age of 2 years, NGE hospitalization incidence rates decreased rapidly with age but increased again for persons >65 years of age, reaching 11.8 to 20.9/10,000 population in persons >85 years of age (Figure 1).

In each study year, the absolute number of NGE hospitalizations was highest among adults 65–84 years of age (range 4,800–9,495/y) and children <5 years of age (range 7,438–9,729/y) (Figure 1). In each study year, the overall rate of NGE hospitalizations was higher in females than males (rate ratio 1.2 for all study years). Among hospitalized patients with NGE, 194–265 deaths (average 203) occurred annually (Table 1), most (96.3%) of which occurred among adults >65 years of age (data not shown). The number of deaths in children <5 years of age ranged from 0 to 5 annually.

Hospitalizations with NGE as Secondary Diagnosis

Figure 2

Figure 2. Annual number and incidence of hospitalizations for norovirus gastroenteritis as secondary diagnosis, Germany, 2007–2012. A) Annual cases by age group; B) annual incidence (no. cases/10,000 population) by age group. Inset shows...

During 2007–2012, an annual average of 40,278 (range 32,259–57,561) hospitalizations occurred with an NGE ICD-10 code A08.1 listed as a secondary diagnosis (Figure 2, panel A). In any study year, the number of hospitalizations with NGE as secondary diagnosis was 1.3–1.7 times higher than for those with NGE as a primary diagnosis. In contrast to NGE as a primary diagnosis, in which the highest rates were in children, the rate of hospitalizations with NGE as a secondary diagnosis was highest in adults >85 years of age (range 33.6–59.1/10,000 population) (Figure 2, panel B).

Seasonal and Regional Distribution of NGE Hospitalizations (Primary Diagnosis)

Figure 3

Figure 3. Monthly number of hospitalizations for rotavirus and norovirus gastroenteritis as primary diagnosis among all age groups (A) and among children 1 to <2 years of age (B), Germany, 2009–2012.

Figure 4

Figure 4. Regional distribution of norovirus hospitalizations among all age groups, by federal state, Germany, 2010. Numbers in parentheses indicate no. cases/10,000 population. Map template obtained from http://www.presentationmagazine.com/editable-maps/page/3.

NGE hospitalizations showed strong seasonality; the highest case numbers occurred during December–March for each of the study years (Figure 3). In 4 out of 5 study years, hospitalizations with a NGE diagnosis peaked before rotavirus hospitalizations. Regional distribution of NGE hospitalizations across Germany showed marked differences: >3-fold variations in the calculated region-specific incidence rates. In 2010, the year with the most NGE hospitalizations, the incidence of NGE hospitalizations was as low as 2.2/10,000 population in Hamburg/Schleswig-Holstein and as high as 7.2/10,000 population in Saxony-Anhalt. However, we found no consistent geographic pattern (Figure 4).

We used national hospitalization statistics to estimate the overall burden of NGE hospitalizations in Germany. During 2007–2012, a total of 408,124 NGE-related hospitalizations occurred. On average, >60,000 hospitalizations occurred annually, including 27,910 hospitalizations with NGE as a primary diagnosis and 40,278 hospitalizations with NGE as a secondary diagnosis, as well as up to 265 NGE-related deaths at a hospital per year.

As reported in other studies of NGE epidemiology (20,21), the incidence of NGE hospitalizations by age follows a U-shaped curve, representing the highest rates in young children and the elderly. Among children <5 years of age, an annual average of >8,000 hospitalizations occurred with NGE as a primary diagnosis; among adults >65 years of age, the annual average was >10,000 hospitalizations.

We described hospitalizations with NGE as a primary or secondary diagnosis separately because we believe they represent different aspects of the burden caused by norovirus infections. The hospitalizations with NGE as a primary diagnosis are most likely representative of community-acquired infections, whereas the secondary NGE diagnoses are more likely to represent nosocomial infections. This assumption about secondary diagnoses might be wrong if patients entered the hospital with NGE occurring simultaneously with another disorder, where the NGE was judged to be a secondary cause of the hospitalization. This scenario might be especially true for older patients with co-occurring conditions that might have been exacerbated as a result of AGE. Also, several patients might have been hospitalized for reasons other than AGE but had onset of signs of NGE in the first 48 hours of hospitalization, which would indicate the patient acquired the infection before hospitalization.

Our estimate of the rate of hospitalization with NGE as a primary diagnosis is higher than the rate for community-acquired NGE estimated by Spackova et al. (14), based on German NGE reports made to RKI. That study estimated that 8%–19% of the 856,539 total NGE hospitalization cases over an 8-year period were community-acquired, corresponding to ≈8,500–20,000 community-acquired NGE hospitalization cases annually. The German Health Insurance Medical Service (Medizinischer Dienst der Krankenkassen [MDK]) (22) carries out a stringent quality-control program of DRG and ICD-10 codes reported by hospitals, as required by law. Because our results are based on data abstracted from this rigorously controlled reimbursement system, we believe our results are more likely to represent the actual burden of all NGE hospitalizations in Germany. Moreover, our validation of the test results in a subset of patients with AGE found a high reliability for the NGE ICD-10 code compared with actual testing, suggesting little overattribution of NGE in this setting. This finding is further supported by data provided in a study by Bernard et al. (13), which estimated norovirus cases to be underreported to RKI by a minimal underreporting factor of 1.7. Our estimates of the NGE hospitalization rate are also higher than previously reported estimates from countries such as the Netherlands (23), the United States (24), and Canada (25) but are comparable to a recent study in England (26,27); (Table 3). Our estimates of annual NGE hospitalizations resulting in death (average 203) were in the same range as previous RKI estimates for Germany (192 annually during 2004–2008) (14,20).

The study years (2007–2012) preceded introduction of universal rotavirus vaccination in Germany in 2013. Considering all age groups, the number of hospitalizations caused by norovirus is markedly higher than the number caused by rotavirus. Although the highest incidence of NGE hospitalizations was in children <2 years of age, within this age group, rotavirus AGE hospitalizations continued to be more frequent.

Our study has several strengths, foremost being the data collection and recording of norovirus-related hospitalization data in Germany. We have taken advantage of the unique system in Germany, which has built-in financial incentives for norovirus testing of inpatients with AGE and mandatory reporting of laboratory-confirmed norovirus, to estimate the burden of NGE hospitalizations. This system provides a structure that, we believe, more comprehensively captures the role of norovirus in nonspecific AGE diagnoses than is possible in many other national databases. Our study used national hospitalization data, which provided age-specific information on length of hospital stay and medical costs. We used data from a 6-year study period, which enabled us to estimate NGE burden across multiple norovirus seasons.

We could not distinguish between community-acquired and nosocomial norovirus infections. An earlier study in Germany estimated 49% of all hospitalized NGE cases to be nosocomial in origin (14), whereas a study in Denmark put this estimate at 63% (28). We found that hospitalization with NGE as a secondary diagnosis represented 59% of all hospitalizations with NGE in our study, suggesting that these cases most likely represent nosocomial infections. The higher rate of secondary diagnoses among older adults compared with younger age groups further suggests that these cases actually represent a different type of hospitalization episode. The inverse age relationship might be related to poorer underlying health and longer hospital stays resulting in potentially greater exposure to a nosocomial infection. We could not exclude the possibility that some patients with NGE as a secondary diagnosis were given an unspecified AGE diagnosis as a primary diagnosis while awaiting the norovirus test results and should have had their infection counted as community-acquired instead of nosocomial. We therefore made an additional request to DESTATIS to provide us with the number of such cases and found that this combination occurred in <0.4% of patients who were given NGE as a secondary diagnosis. This finding, the consistency between our ratio of primary to secondary diagnoses to those reported by other investigators, and the age distribution support our assumption that secondary diagnoses reported by DESTATIS are most likely nosocomial NGE cases.

We were unable to include deaths in cases coded with NGE as a secondary diagnosis because we could not ascertain that NGE was the cause of death. The number of deaths attributable to NGE might thus have been underestimated.

Although the reliance upon the reporting of NGE cases to DESTATIS is another potential limitation of our study, the finding that 95.2% of the coded cases in a sample of patients had a documented positive norovirus result suggests this reliance is unlikely to have led to a major overestimation of the actual burden of NGE hospitalizations in Germany. Our findings in a single medical center might not be generalizable to the entire country. We do not expect, however, that testing rates would differ substantially between hospitals given the financial incentive and the strict quality controls carried out by MDK (22). From an economic perspective, our assessment of NGE hospitalization costs only considered the reimbursable fraction and did not take into account indirect or societal costs, which would have led to substantially higher cost estimates.

On the basis of our validation assessment of ICD-10 diagnostic codes compared with norovirus laboratory results, our estimated burden of NGE hospitalizations according to ICD-10 diagnostic coding might have underestimated the actual number of NGE hospitalizations. We had, a priori, assumed that all patients with AGE caused by norovirus are tested because of financial reimbursement incentives. However, our finding that 12.4% of patients with positive norovirus results did not have a corresponding NGE ICD-10 diagnostic code in the patient’s medical records suggests that underreporting of NGE hospitalization cases based on diagnostic codes has occurred. We also observed seasonality among the cases coded as unspecified AGE (ICD codes A08.3, A08.4, A08.5, and A09); peaks in winter and early spring suggest that a proportion of these cases might also be attributable to norovirus.

During the study period, most testing for norovirus in Germany was conducted by using an ELISA test. Compared with the standard reverse transcription PCR test, the sensitivity of ELISA has been estimated to be 77% and the specificity 96% (29). Correcting for these estimates would reduce the proportion of AGE hospitalizations caused by norovirus slightly, from an average of 13.8% to 13.2% (30).

In summary, by using national hospitalization discharge and cost data, we found that NGE resulted on average in 27,910 hospitalizations with NGE as a primary diagnosis and an additional 40,278 hospitalizations with NGE as secondary diagnosis annually in Germany, a number substantially higher than previously assumed. The associated reimbursed annual medical cost for community-acquired NGE was up to €42 million annually. These data could aid in the identification of target groups for future norovirus vaccination policies and could be used to assess the effects of vaccination on the NGE hospitalization burden after implementation.