点击显示 收起
Division of Viral Hepatitis, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia
Background.
The genetic relatedness of hepatitis A virus (HAV) isolates was determined to identify possible infection sources for case patients in the Sentinel Counties Study of Acute Viral Hepatitis.
Methods.
A 315-nucleotide segment of the VP1-P2 region of the HAV genome was amplified and sequenced from serum of case patients and analyzed together with risk-factor data.
Results.
Of 508 HAV-RNApositive case patients, 449 (88.4%) were interviewed, and 255 (50.1%) reported 1 risk factor. Some 123 unique nucleotide sequence patterns (UNSPs) were identified77 (62.6%) from only 1 case patient and the rest in 299 persons. Among international travelers, a single person was more often infected with a single type of UNSP (17/54 [31.5%]), compared with other case patients (48/393 [12.2%]; P < .001). UNSPs from travelers to Mexico (33/37 [89.2%]) clustered with those from Hispanic children (47/49 [95.9%]). Of 119 men who had sex with men, 96 (80.7%) had the same or similar UNSPs, which were also found in 37 men and 10 women with no identified infection source.
Conclusion.
HAV is often transmitted within networks of persons with similar risk factors, which may be the infection source for others in the community.
In the United States, hepatitis A is a relatively common infectious disease, with 25,00035,000 symptomatic cases reported each year during most of the past several decades [1]. Of persons identified with hepatitis A disease, 50% report a potential source of infection. Household or other close contact with a person with hepatitis A disease is the most frequently reported source; other potential sources include being a man who has sex with men (MSM), travel to a country where hepatitis A virus (HAV) is endemic, use of illicit drugs, or exposure to a day-care center. For 40%50% of cases, the source of infection remains unknown even after thorough investigation [2]. Some of these latter cases may be due to transmission from persons with subclinical infection, particularly children, or could be ascribed to unacknowledged risk factors [13].
HAV, an RNA virus and a member of the Picornaviridae family, displays a high degree of antigenic and genetic conservation [47]. However, enough diversity exists in the capsid genome region to define genotypes and identify differences between HAV isolates. HAV genotypes are defined by 15% nucleotide variation, and subgenotypes have 0%7.5% nt variation [7]. Although an RNA virus, HAV appears to exhibit a low rate of mutation over time [8].
Since 1981, the Centers for Disease Control and Prevention (CDC) has conducted the Sentinel Counties Study of Acute Viral Hepatitis, to determine trends and risk factors for infection with hepatitis viruses [3, 9]. We combined information on the genetic relatedness of HAV isolated from all reported cases of hepatitis A disease in the 6 counties participating in the study during 19961997 with epidemiologic case investigation data, to determine the molecular epidemiology of transmission patterns and sources of HAV disease. Specimens from these years would represent the epidemiology of HAV before the implementation of the first recommendation for the use of vaccine.
PATIENTS, MATERIALS, AND METHODS
Serum specimens were obtained during 1996 and 1997 through the 6 counties participating in the Sentinel Counties Study of Acute Viral Hepatitis: Denver, CO; Jefferson, AL (including Birmingham); Pinellas, FL (including St. Petersburg); Pierce, WA (including Tacoma); Multnomah, OR (including Portland); and Contra Costa, CA (including Oakland). Suspected cases of hepatitis A disease were reported to their respective health departments. For the purposes of the study, cases of hepatitis A disease, identified by laboratory and clinician reporting, were defined as the presence of IgM antibody to HAV (IgM anti-HAV). Most cases also included an illness with acute onset and signs and symptoms compatible with hepatitis. In each county, a trained study nurse attempted to contact all case patients and, after informed consent was obtained, conducted an interview in a private setting using a standardized questionnaire that included questions on potential sources of infection during the 26 weeks before illness onset [3]. Case patients who reported >1 risk factor were considered to have all acknowledged risk factors; no attempt was made to select a single most-important risk factor. When a case patient could not be contacted, the patient's health-care provider was asked to provide limited demographic information about the patient. Serum specimens from case patients were analyzed at the CDC for markers of infection with hepatitis viruses. All specimens that tested positive for IgM anti-HAV were later tested for HAV RNA.
Amplification and sequencing of HAV RNA.
The VP1-P2A junction (28703381 nt) of the HAV genome was amplified from serum specimens (100 L) by nested reverse-transcription polymerase chain reaction (PCR), and PCR products were sequenced by use of an automated sequencer (ABI model 373 or 377; Applied Biosystems), as described elsewhere [10]. Preliminary sequence analysis was performed by use of ABI software (version 1.0.1; ABI), and further analysis was performed by use of GCG software (version 10.3; Wisconsin Package) [11].
Statistical methods.
Statistical analysis was performed with SAS software (version 9.1; SAS Institute). Differences in proportions were compared by 2, Fisher's exact, or Cochran-Mantel-Haenszel test. A logistic-regression model was used to evaluate the independence of determinants of HAV RNA detection among the case patients.
Multiple sequence analysis [11] was used to determine the genetic relatedness of the nucleotide sequence of each case patient and to categorize these into genotypes and subgenotypes. Closely related unique nucleotide sequence patterns (UNSPs) observed within the subgenotypes were considered to be clusters. Each UNSP found in the case patients in the study has been identified under GenBank accession numbers AY753408AY753530.
RESULTS
Study Participants
During 19961997, a total of 799 cases of hepatitis A disease were reported in the 6 participating counties, 750 (93.9%) in case patients who also had symptoms of hepatitis. Most individuals were from Denver, Multnomah, and Pierce Counties (table 1). Study nurses at the various sites were able to administer the questionnaire to 612 (76.6%) of these people. Case patients were predominantly male, 1544 years old, and either non-Hispanic white or Hispanic (table 1).
Detection of HAV RNA
Serum specimens were available from 554 (69.3%) case patients, and HAV RNA was amplified from 508 (91.7%). Case patients whose specimens were positive for HAV RNA did not differ from the overall sample of 799 case patients with respect to demographic variables after stratification by county (table 1) (P > .05 for all comparisons).
The likelihood of detecting HAV RNA in serum increased with higher alanine aminotransferase (ALT) levels and decreased with the time elapsed between symptom onset and serum collection (table 2). RNA was found in 431 (93.9%) of 459 samples collected within 27 days of symptom onset and in 26 (63.4%) of 41 samples collected after 27 days. The longest time from symptom onset to an HAV RNApositive sample was 55 days. The RNA detection rate was highest (95.8%) among persons with an ALT level at least 7 times the upper limit of normal in samples that were collected within 27 days of symptom onset. In a logistic regression model, RNA detection was independently associated with increased ALT level and shorter duration between symptom onset and sample collection but not with demographic or risk-factor variables (data not shown).
Sequence Distribution
A PCR-generated 315-bp product from the VP1-P2A region was available for analysis in all 508 samples with detectable RNA. Most were genotype 1a (n = 495); the remainder were genotype 1b (n = 11) and 3a (n = 2). Among the genotype 1a samples, the sequences varied by 025 nt (0%7.9%). The subtype 1a sequences differed from the 1b sequences by 1739 nt (5.4%12.4%), and both differed from the subtype 3b sequences by 6373 nt (20.0%23.2%).
Multiple sequence alignments of the VP1-P2A region identified 123 UNSPs, many of which were found in samples from multiple case patients. Seventy-seven UNSPs were found in only 1 case patient, 24 in 24 case patients (n = 57), 9 in 59 case patients (n = 70), 8 in 1019 case patients (n = 108), 4 in 2026 case patients (n = 97), and 1 (UNSP 9) in 99 case patients.
Relationship of NSPs to Risk Factors
Of the 508 case patients for whom a UNSP was available, 449 (88.4%) had a completed epidemiologic questionnaire. Most cases of hepatitis A disease (411 [80.9%]) occurred in Denver, Multnomah, and Pierce Counties. Within each county, case patients with identical or similar UNSPs often shared common risk factors (figure 1), such as being an MSM (e.g., UNSP 9), injection drug use (e.g., UNSP 103), or recent travel to Mexico (e.g., UNSP 45). Other UNSPs, such as 53 and 107, showed no predominant risk factor. In addition, case patients with an identical UNSP were frequently found in >1 county (e.g., UNSP 9, Denver, Multnomah, Pierce, and Contra Costa; UNSP 78, Denver, Multnomah, Jefferson, and Pinellas).
A total of 148 case patients had the 13 UNSPs that formed cluster A (figure 1), including the 96 (64.9%) men who reported being an MSM. Cluster A was found predominantly in Denver and Multnomah Counties, with only 7 cases occurring elsewhere (i.e., Contra Costa, Pierce, and Jefferson Counties). Of the 52 case patients in this cluster who did not report being an MSM, 42 (80.8%) were male case patients 15 years old, and 37 (88.1%) did not acknowledge any risk factors for infection. Ten female subjects were in this cluster, and 7 reported risk factors that included recent close contact with a person with hepatitis A disease (n = 5), injection drug use (n = 1), and recent travel to Mexico (n = 1). The UNSPs in this cluster were found throughout the study period in Denver County; they appeared in Multnomah County in mid-1996 and peaked in mid-1997 (figure 2).
Fourteen case patients had the UNSPs in cluster B (figure 1). Twelve were in Denver County, with 1 case each in Pierce and Jefferson Counties. As with cluster A, cases in cluster B occurred predominantly in men (13 [92.9%]), of whom most (9 [64.3%]) identified being an MSM as their risk factor for infection.
Injection drug use.
Injection drug use was reported as a risk factor by 35 case patients for whom nucleotide sequences were available. Ten of these case patients also reported recent contact with another person with hepatitis A disease, and 5 were men who also reported being an MSM.
Three closely related UNSPs (103, 111, and 112) were found in just over one-half (n = 18) of case patients who reported injection drug use. These UNSPs (figure 1) were found predominantly in Denver (111 and 112), Multnomah (103 and 111), and Pierce (103 and 111) Counties.
Of the case patients with UNSPs 103, 111, or 112 (n = 58), all were at least 15 years old, 62% were male, and 80% were white. Other potential risk factors reported by these case patients included contact with another person with hepatitis A disease (n = 10), employment in a day-care facility or other close contact with a child in day care (n = 5), and being an MSM (n = 3).
International travel.
Thirty-six UNSPs were found in case patients who reported recent travel outside of the United States (n = 54). Most of these case patients had traveled to Mexico (n = 45) or another South or Central American country (n = 4). A UNSP from international travelers was more likely to be observed in only 1 case patient (17/54 [31.5%]) than were those from persons with no recent history of international travel (48/393 [12.2%]; P < .001, 2 test). Of the 19 UNSPs observed in >1 case patient, 3 (33, 70, and 119) were found exclusively in international travelers and were found in 2 case patients each (figure 1). The remaining 16 UNSPs were also found in case patients without a history of international travel. When matched by UNSP, the onset of illness in case patients with no history of international travel (n = 12) preceded the onset of illness in case patients with a history of international travel.
Of 37 case patients with a history of recent travel to Mexico and no other acknowledged risk factors, 33 (89.2%) had UNSPs that all fell within a cluster that included UNSPs 2598, excluding UNSP 47, and represented 215 case patients. This cluster also included 47 (95.9%) of 49 Hispanic children <15 years old, 12 (25.5%) of whom reported international travel.
DISCUSSION
The present population-based study provides new information on HAV transmission in case patients by combining epidemiologic and HAV relatedness information. It indicates that individual HAV strains (UNSPs) are often transmitted within networks of persons who have similar risk factors for infection. Most striking was the predominance of certain UNSPs among MSM, including some that were found in multiple geographic locations. This suggests connections among transmission networks. The predominance of certain UNSPs among MSM suggested that other persons infected with these strains may have been connected to this population, through homosexual, bisexual, or close personal contact.
The tracing of UNSPs also identified international transmission patterns of HAV. Certain strains circulated primarily among Hispanic children and recent travelers to Mexico, which suggests that these infections and their resultant chains of transmission were introduced into Mexican American communities in the United States by returning travelers.
The clustering of certain UNSPs with particular risk factors was identified only through the molecular epidemiologic analysis performed for the present study. For example, cases in MSM who shared common UNSPs (e.g., clusters A and B) occurred over a 2-year period in Denver County and a 1-year period in Multnomah County, and this linkage to one another was not evident from conventional history taking for risk factors. Similar UNSPs that differed by 1 nt were found in 12 of 17 men during an outbreak of hepatitis A disease among MSM in Oslo, Norway, during October 1997March 1998 [12]. In this latter outbreak, the remaining 5 men had a UNSP identical to UNSP 61 in cluster B.
Together with other data on the epidemiology of HAV, the findings of the present study indicate that HAV transmission occurs primarily within local networks of persons with common risk factors. When one of these networks acquires a sufficient number of susceptible members, the introduction of HAV into the network results in spread of that particular strain until transmission is no longer sustainable, presumably because of a reduced number of susceptible members resulting from either infection or immunization. In addition, HAV strains may be transmitted from one network to another, as was observed with UNSP 9, which has circulated among networks of MSM in several counties and on at least 2 continents.
The pattern of HAV transmission within population networks is consistent with the observation that the incidence of hepatitis A disease is cyclical and that this incidence often peaks in different counties in the United States at different times [1]. Presumably, this reflects the random introduction of virus into different networks at different times. This pattern of transmission also explains why single HAV NSPs are often found in localized outbreaks among persons who have specific risk factors. For example, UNSPs were found in outbreaks among MSM in Ohio in 1998, methamphetamine users in Florida in 2001, methamphetamine users in Washington state in 2002 (CDC, unpublished data), and children in Stanislaus County, CA [13].
The present study suggests how children can serve as key links for HAV transmission in local networks, because it was conducted before extensive childhood HAV vaccination [14]. During the prevaccination era, the incidence of infection in children was many-fold higher than that in adults, and most infections in children were inapparent [15]. Children with asymptomatic infection have been shown to be an important source of hepatitis A disease for adults who have no recognized risk factors during communitywide hepatitis A outbreaks [2]. The present study does not allow for a robust characterization of the molecular epidemiology of HAV in children, because it was conducted during a period when outbreaks in adults predominated. Nonetheless, HAV strains from Hispanic children, most of whom lacked an identified source of infection, clustered with recent travelers to Mexico. Mexico is a country where HAV infection is highly endemic, and these findings suggest the significance of travel to Mexico, both among children and adults, as an important source of HAV.
Networks of children and adults with common risk factors clearly provide sources of HAV transmission for individuals outside of the networks. Although specific UNSPs were found predominantly in MSM, these strains were also found in women. Several sequence patterns were found disproportionately among injection drug users, but many case patients with these patterns were not drug users. It is probable that transmission outside of the primary networks is so inefficient as to be unsustainable and results in certain UNSPs not being found more widely among case patients. Whether the latter is true or not, interrupting transmission within the major transmission networks will be necessary to interrupt HAV transmission in the United States.
The Sentinel Counties Study of Acute Viral Hepatitis has previously provided epidemiologic information on HAV and other types of viral hepatitis that has been found to be applicable to the entire country [3, 9]. For this reason, our conclusions about HAV transmission are most likely applicable to the entire country. However, although it might be tempting to assume that Denver County was the source of UNSP 9 for Multnomah County, this sequence pattern most likely has a wider geographic distribution and may have been introduced in both counties from other locations in the United States. In addition, hepatitis A disease notifications to public-health departments probably represent only 1 in 5 HAV infections in adults and 1 in 20 infections in children [15]. Thus, the case patients in the present study represent a small proportion of all HAV infections in the 6 counties. This may account, at least in part, for the large proportion of case patients singly infected with a unique HAV strain.
Current US guidelines call for HAV vaccination of adults with known risk factors for infection and routine immunization of children in those parts of the country with historically high disease rates [1, 14]. Achieving high HAV vaccination coverage rates in communities of high-risk adults has the potential to interrupt HAV transmission networks. Although no data are available on HAV vaccination coverage among high-risk adults, such as MSM, hepatitis B virus (HBV) immunization coverage among MSM is low, which suggests that the same situation exists for HAV [16]. Studies conducted in MSM communities undergoing outbreaks of hepatitis A disease have indicated that most patients have a regular source of medical care and would accept HAV vaccination if it were offered by their health-care provider [17]. Recent population-based demonstration projects have shown the feasibility of vaccinating high-risk adults for HBV [18]. High HAV vaccination coverage rates in these populations have the potential to interrupt transmission networks. However, there are no sustained regional or national programs to vaccinate adults at high risk for hepatitis A disease.
Coincident with increasing immunization coverage rates among children has been a decline in overall HAV rates to the lowest ever observed in the United States. However, the decline in disease incidence among adults has not been as great as that observed for children, and the incidence in adults now exceeds that in children [19]. The results of the present study indicate that improvements in immunization coverage rates among high-risk adults will be necessary to continue to reduce the incidence of hepatitis A disease in the United States.
References
1. Centers for Disease Control and Prevention. Prevention of hepatitis A through active and passive immunization: recommendations of the Advisory Committee on Immunization Practices. MMWR Morb Mortal Wkly Rep 1999; 48:137. First citation in article
2. Staes CJ, Schlenker TL, Risk I, et al. Sources of infection among persons with acute hepatitis A and no identified risk factors during a sustained community-wide outbreak. Pediatrics 2000; 106:E54. First citation in article
3. Bell BP, Shapiro CN, Alter MJ, et al. The diverse patterns of hepatitis A epidemiology in the United States: implications for vaccination strategies. J Infect Dis 1998; 178:157984. First citation in article
4. Provost PJ, Ittensohn OL, Villarejos VM, Arguedas JA, Hilleman MR. Etiologic relationship of marmoset-propagated CR326 hepatitis A virus to hepatitis in man. Proc Soc Exp Biol Med 1973; 142:125767. First citation in article
5. MacGregor A, Kornitschuk M, Hurrell JG, et al. Monoclonal antibodies against hepatitis A virus. J Clin Microbiol 1983; 18:123743. First citation in article
6. Crevat D, Crance JM, Chevrinais AM, et al. Monoclonal antibodies against an immunodominant and neutralizing epitope on hepatitis A virus antigen. Arch Virol 1990; 113:958. First citation in article
7. Robertson BH, Jansen RW, Khanna B, et al. Genetic relatedness of hepatitis A virus strains recovered from different geographical regions. J Gen Virol 1992; 73:136577. First citation in article
8. Bower WA, Nainan OV, Han X, Margolis HS. Duration of viremia in hepatitis A virus infection. J Infect Dis 2000; 182:127. First citation in article
9. Alter MJ, Hadler SC, Margolis HS, et al. The changing epidemiology of hepatitis B in the United States: need for alternative vaccination strategies. JAMA 1990; 263:121822. First citation in article
10. Hutin YJ, Pool V, Cramer EH, et al. A multistate, foodborne outbreak of hepatitis A. National Hepatitis A Investigation Team. N Engl J Med 1999; 340:595602. First citation in article
11. Devereux J, Haeberli P, Smithies O. A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res 1984; 12:38795. First citation in article
12. Stene-Johansen K, Jenum PA, Hoel T, Blystad H, Sunde H, Skaug K. An outbreak of hepatitis A among homosexuals linked to a family outbreak. Epidemiol Infect 2002; 129:1137. First citation in article
13. Robertson BH, Averhoff F, Cromeans TL, et al. Genetic relatedness of hepatitis A virus isolates during a community-wide outbreak. J Med Virol 2000; 62:14450. First citation in article
14. Centers for Disease Control and Prevention. Prevention of hepatitis A through active and passive immunization: recommendations of the Advisory Committee on Immunization Practices. MMWR Morb Mortal Wkly Rep 1996; 45:138. First citation in article
15. Armstrong GL, Bell BP. Hepatitis A virus infections in the United States: model-based estimates and implications for childhood immunization. Pediatrics 2002; 109:83945. First citation in article
16. MacKellar DA, Valleroy LA, Secura GM, et al. Two decades after vaccine license: hepatitis B immunization and infection among young men who have sex with men. Am J Public Health 2001; 91:96571. First citation in article
17. Cotter SM, Sansom S, Long T, et al. Outbreak of hepatitis A among men who have sex with men: implications for hepatitis A vaccination policy. J Infect Dis 2003; 187:123540. First citation in article
18. Centers for Disease Control and Prevention. Hepatitis B vaccination among high-risk adolescents and adultsSan Diego, California, 19982001. MMWR Morb Mortal Wkly Rep 2002; 51:61821. First citation in article
19. Centers for Disease Control and Prevention. Hepatitis surveillance [report 58]. Available at: http://www.cdc.gov/ncidod/diseases/hepatitis/resource/PDFs/hep_surveillance_59.pdf (Accessed 28 October 2003). First citation in article