Routine Epidemiologic Data

The study was conducted from the 14th calendar week of 2007 through the 13th calendar week of 2008. Acute hepatitis A has been a notifiable disease in Germany for many years. Since 2001, laboratories have reported infections (test results indicative of acute HAV infection, detection of specific IgM in serum, or detection of HAV RNA in serum or feces by PCR) to local health departments. These departments collect information on case-patients (age, sex, travel history, date of onset, clinical symptoms, duration of hospitalization), take preventive measures to avoid further spread (including recommending vaccination of contacts, barring infected food handlers from working), and report standard case information electronically to state health authorities and RKI in a form in which names and addresses are removed. Data can be grouped on all levels to indicate outbreaks.

Additional Epidemiologic Data

To transcend standard information obtained for each reported infection and add a virologic perspective, we collected additional case information on all HAV infections in Germany over a 1-year period. The 16 state health departments in Germany were requested to participate and coordinate distribution and collection of questionnaires to all local public health departments. These departments recorded additional case information obtained during routine case investigations on paper forms (additional case information sheets). Data included details of travel, concurrent health conditions, and potential migration background. These investigations consisted of telephone interviews with patients (rarely with their physicians as proxies). Completed forms were sent to RKI without names and addresses but did contain case codes. Accommodations other than hotels were defined as those presumably involving closer contact with the local population or exposure to food prepared under potentially suboptimal hygienic conditions (e.g., private accommodations, hostels, or campgrounds).

Persons with a migration background were defined as those who moved to Germany after 1949 as non-German nationals, children born in Germany to non-German nationals, or children born to at least 1 parent belonging to either of these groups. Adults were defined as persons >18 years of age.

Laboratory Data

We obtained serum samples from >10% of persons with HAV infections diagnosed in Germany during the study. To facilitate sequencing and phylogenetic analysis of a representative selection of HAV strains causing infections in Germany during the study, we asked >120 large private laboratories, university clinics, and hospitals in Germany to provide serum samples for patients for whom IgM against HAV was detected at their facilities. Samples were either sent to RKI immediately or stored at –20°C and sent in larger batches.

Collation of Data Sources

Information recorded on paper forms was entered into a database and matched to electronically transmitted routine data by case codes. All symptomatic cases reported from participating states were analyzed. Clinical specimens were matched to questionnaire and surveillance data according to anonymous patient information (year and month of birth, sex, crude area of residence, date of blood sampling) provided by the laboratories.

Isolation and Sequencing of HAV RNA

HAV RNA was isolated from serum samples by using a Viral RNA Mini Kit (QIAGEN, Hilden, Germany). Reverse transcription and first-round amplification of the capsid protein (VP1)/2A junction region of HAV were performed by using a Onestep RT-PCR Kit (QIAGEN) and primers (Table 1).

Nested PCR was performed by using the HotStarTaq Master Mix Kit (QIAGEN). Purified products of the nested PCR (forward and reverse strands) were sequenced by using a 3130x ABI Prism Genetic Analyzer and a BigDye Terminator version 3.1 Cycle Sequencing Kit (PE Applied Biosystems, Weiterstadt, Germany).

Statistical Analysis

Sequences were processed by using Lasergene SeqMan Pro software (DNASTAR, Inc., Madison, WI, USA), aligned by using the ClustalW algorithm (7), and optimized manually. Phylogenetic trees were constructed by using all available HAV sequences from obtained HAV IgM–positive samples (including HAV sequences from cases not reported). Sequence statistical and phylogenetic analyses were conducted by using MEGA4 (8). Sequences obtained are referenced in GenBank under accession nos. EU416232–EU416273 and EU825848–EU825918. For statistical analysis, we used Microsoft Excel 2003 (Microsoft, Redmond, WA, USA), SPSS version 15.0 (SPSS Inc., Chicago, IL, USA), and STATA version 10.1 (StataCorp, College Station, TX, USA).

Study Population

A total of 1,213 HAV infections were reported. Among them, 952 (78.5%) were in patients with clinical symptoms consistent with acute hepatitis A. Of the 16 states in Germany, 13 participated in the intensified surveillance and contributed 1,037 (85.5%) of all reported infections and 816 (86%) of all reported symptomatic cases (maximal denominator for analysis). Additional case information sheets were available for 571 (70%) symptomatic cases. Serum samples positive for IgM against HAV were available for 189 (23.2%) cases; 95 (11.6%) were PCR positive.

A total of 74.6% of the cases were reported as single cases; the remainder were in recognized clusters. Among case-patients, 47.1% were male (Table 2); age range <1 to 90 years (median 32 years). Among nonadults, 81.1% were reported to have had jaundice. Among adults, the proportion of persons with jaundice decreased with age to 40.6% in persons >60 years of age.

Of case-patients, 46.4% were hospitalized for a median of 6 days (range 1–28 days). Among those with jaundice, no clear trend for age was found in 50.2% who were hospitalized. Among those without jaundice, the proportion of those hospitalized increased among persons >60 years of age (56.5%) when compared with that of younger persons (31.3%, relative risk [RR] 1.81, 95% confidence interval [CI] 1.35–2.41). Among those employed, absences from work ranged from 2 to 32 work days (median 6 days).

Overall, 43.6% of the infections were imported (i.e., infection acquired while traveling outside Germany). Nonadults (60.6% imported; p<0.001) and male patients (48.1% imported; p = 0.018) were overrepresented.

A migration background was reported by 42.2% of case-patients (78.8% of nonadults and 19.1% of adults), of whom 64.8% had been born in Germany. Among migration background, Turkey was reported most frequently (48.5%), followed by the former Yugoslavia (11.9%), southern and Southeast Asia (9.7%), and the former USSR (8.8%). Nonadults with a migration background lived with a larger number of household members (range 1–11 persons in addition to the case-patient, median 4) than those without a migration background (range 1–6 persons, median 3). Among adult case-patients, 5.2% were professional food handlers.

Imported Infections and Comparison with Autochthonous Infections

Among known destinations, Turkey (35.6%) was reported most frequently, followed by the former Yugoslavia, Egypt, and Spain (Table 3). Most affected case-patients (63.9%), especially nonadults (82%), had not traveled for vacation or business but had visited friends or family. Median duration of travel preceding infection was 29 days (range 1–180 days). Accommodations other than hotels (e.g., staying with friends or family) predominated (73.1%), especially among those visiting Turkey (91.9%) and the former Yugoslavia (94.1%). A total of 92.9% of those infected in Turkey and 76.2% of those infected in countries of the former Yugoslavia had matching migration backgrounds; only 3 (16.7%) of those infected in Egypt had matching migration backgrounds. Of case-patients with migration backgrounds who had become infected in these ancestral countries, 75.9% from Turkey, 87.5% from the former Yugoslavia, and 100% from Egypt were nonadults.

Figure 1

Figure 1. Reported cases of hepatitis virus A infection (n = 679) by month of onset and patient age group (y), Germany, 2007–2008. A) Imported cases. B) Nonimported cases.

Imported infections were most likely to cause disease from August through October, and a prolonged wave of autochthonous infections then followed from October through March (Figure 1). Among imported infections, children and persons <20 years of age were overrepresented (70.2% from August through October vs. 33.9% in other months; p>0.001). Among autochthonous cases, incidence in children and persons <20 years of age increased 62% from September through February compared with March through August. Autochthonous cases in adults >40 years of age were almost evenly spread throughout the year.

Figure 2

Figure 2. Age distribution (y) of persons with hepatitis A virus (HAV) infection by migration background and country where HAV infection was acquired (n = 520 with all 3 factors known), Germany, 2007–2008.

Case-patients with autochthonous infections were older than case-patients with imported infections and less likely to have migration backgrounds (Figure 2). Almost all patients >60 years of age had autochthonous infections.

A migration background was more likely among persons with imported infections; 52.8% of case-patients had a background of migration from Turkey. However, 23% of persons with autochthonous infections had a migration background (Table 4). Patients with autochthonous infections were older; 56.4% persons with nonimported cases were >40 years of age and 31.1% were >60 years of age, compared with only 24% and 7.5%, respectively, of persons of the same ages with cases of imported infections. Case-patients with imported infections were also slightly more likely to be male and part of case clusters than were autochthonous case-patients.

Close contacts of 58.4% of the patients received prophylactic vaccinations (range 1–150 persons, median 3). Contacts were more frequently vaccinated in response to cases in nonadults (RR 1.594, 95% CI 1.374–1.849).

No large outbreaks were reported during the study. In the 13 participating states, 9 clusters with >5 infections were detected. The largest (13 symptomatic and 2 asymptomatic infections) was in a school with multiple generations of infection. The ultimate source of infection could not be elucidated. The school index case-patient had not traveled. However, the implicated HAV strain was genetically similar to strains from Turkey. In 4 of the 8 outbreaks with 5–7 infected persons, >1 travel-associated index cases were recognized, leading to 1–4 secondary infections in Germany.

Detection of HAV RNA

HAV RNA was detected in 95 (50.3%) of 189 samples matching a reported case of symptomatic HAV infection. The likelihood of detecting HAV RNA in serum depended on clinical characteristics of patients (Table 5). Frequency of RNA detection increased with number of symptoms reported (p<0.02) and was negatively associated with patient age (p<0.001). Although this frequency was similar for nonadults and adults <39 years of age (mean 65%), RNA was detected in only 10% of symptomatic patients >60 years of age. No correlation was seen between duration from symptom onset to day of blood sampling (maximum 29 days) and positive results for HAV RNA.

Molecular Epidemiology

Figure 3

Figure 3. Neighbor-joining phylogenetic tree of a 348-bp section of the viral capsid protein 1//2A junction region of hepatitis A virus (HAV) constructed by using the Kimura 2-parameter distance model. Place of infection...

A PCR-generated 348-bp fragment was available for analysis of isolates with detectable HAV RNA from 126 patients. Of these patients, 73 (57.9%) had genotype IB strains, 36 (28.6%) had genotype IA strains, and 17 (18.6%) had genotype IIIA strains. Sequences differed in <8.9%, <7.5%, and <4.9% of positions within genotype IB, IA, and IIIA strains, respectively. Sequence variability was reflected in the diversity of countries or regions from which sequences originated (Figure 3).

Phylogenetic analysis showed that imported strains clustered according to region or country. Imported genotype IB strains were isolated from patients who traveled to Turkey, the Middle East, and Africa. Strains acquired in South Africa formed a small but distinct subcluster. Patients infected with genotype IA strains reported traveling to eastern Europe, Asia, South America, and northwestern Africa. The small number of imported genotype IIIA strains was obtained from patients who had traveled to eastern and southeastern Europe and Central and southern Asia. Three strains imported from Spain (2 genotype 1A and 1 genotype 3A) were genetically more distant from each other than to strains from other regions (e.g., northern Africa).

Strains from patients with known migration backgrounds (Figure 3) but no travel histories clustered with imported strains of the respective region. This finding was evident in a clade containing 35 highly related genotype IB strains, of which 14 were imported from Turkey. Of 23 persons with a background of migration from Turkey and for whom sequence data were available, 18 (78%) were infected with strains that were found within this clade. Furthermore, among persons who traveled to Turkey and available information on possible migration background (n = 12), all also had a background of migration from Turkey. Strains isolated from patients with autochthonous infections but without known migration backgrounds were nearly as genetically diverse as imported strains. Individual autochthonous strains, however, were frequently very homogeneous to individual imported strains.

Most (53.2%) HAV sequences obtained were found only once. Another 18.3% were isolated from smaller outbreaks of hepatitis A, which had already been detected through routine surveillance (Figure 3). Remaining sequences (28.6%) were each seen 2 or 3 times without any known epidemiologic connection. The exception was a strain found in 12 patients who represented sporadic cases and smaller outbreaks. This strain is part of the clade from Turkey mentioned earlier; among autochthonous cases with this strain, no clustering of time or place of infection was apparent.