|Education + Advocacy = Change|
The agreement between reporting to the NNDSS and to the VHSP does not necessarily measure the completeness of reporting from a particular state, since not all cases may be reported to the NNDSS and the two systems have different reporting criteria. The increasing discrepancy between the two systems has resulted in differences in the relative proportions of types of viral hepatitis reported. Before 1990, the proportions of reported cases by type were similar between the two surveillance systems. Since then, the proportion of hepatitis cases reported as hepatitis A to the two systems have remained similar, but the proportion of cases reported as hepatitis B have been discrepant: 24% to 27% of the total VHSP cases were reported as hepatitis B, compared with 35% to 38% of total cases reported to NNDSS. The VHSP also received reports on smaller proportions of the total number of NANB hepatitis cases (7% to 8% of total cases) than did NNDSS (up to 13% of total cases).
These differences in proportions of cases are partly due to the fact that VHSP excludes cases that do not meet the case definition (VHSP eliminated 15% of reported cases as non-cases in 1993). In addition, because of strict adherence to the case definition, VHSP classified a larger proportion of reported cases as nonspecific hepatitis: 15% of cases were classified as hepatitis unspecified by the VHSP during 1993 compared with 1.5% of cases reported to NNDSS. Beginning with data collected in 1995, hepatitis cases that have type unspecified are no longer requested or printed in the MMWR.
The VHSP excludes reported cases that do not meet the case definition for acute viral hepatitis (see “Case Definition” section, below), including cases that seem to be due to chronic infections. Some responses to the VHSP questionnaires are incomplete, and the information is insufficient to verify the case as an acute infection, or to confirm the serologic type of hepatitis, even though partial testing may have been done. Cases may also be reported too late to be included in the analysis. The latest date for submitting case reports to the VHSP for the calendar year is March 31 of the following year.
Use of Serologic Tests for Diagnosis
Serologic testing for the diagnosis of hepatitis, beginning with hepatitis B surface antigen (HBsAg) in 1972, immunoglobulin-M antibody to hepatitis A virus (IgM anti-HAV) in 1981, and IgM antibody to hepatitis B core antigen (IgM anti-HBc) in 1984, has been critical in distinguishing the types of viral hepatitis. Serologic testing for any marker using one or more tests has increased from 60% in 1983 to 94% in 1993 (Figure 1). By 1993, only 6% of reported cases were diagnosed on the basis of the HBsAg test alone. However, there has been a decline in the number of cases reported in which testing for both hepatitis A and hepatitis B was done. In 1989, 76% of physicians reported using tests for both types (the highest percentage reached); this declined to 70% in 1990, 68% in 1991, 63% in 1992, and to 55% in 1993. At the same time, the number of cases reported in which testing only for hepatitis A was done increased over this period, from 15% in 1989 to 28% in 1993. The reliance on testing for hepatitis A alone for these cases may be related to the higher incidence of hepatitis A in community-wide outbreaks since 1989.
Figure 1. Serologic Tests to Diagnose Hepatitis, 1983-1993
Among persons less than 15 years of age, hepatitis A remained the most frequent of the types reported; hepatitis B and NANB hepatitis were reported in small numbers of persons in this age-group (Table 2). The percentage of NANB hepatitis cases among patients 60 years old and older (8.8%) was the highest of the three types. However, most persons who acquire any type of viral hepatitis are between the ages of 20 and 39: approximately 45% of hepatitis A, 63% of hepatitis B, and 61% of NANB hepatitis are reported among persons in this age span.
From 1992 to 1993, the number of hepatitis A cases among patients 20-39 years of age decreased 10%; hepatitis B cases, 21%; and NANB hepatitis cases, 19%. Demographic factors for all types showed patterns consistent with those of previous years (Table 2).
The male-to-female case ratios were similar to previous years: for hepatitis A, the male-to-female ratio was 1.2:1; for hepatitis B, 1.5:1; and for NANB hepatitis, 1.4:1.
Non-Hispanic whites accounted for the majority of all types reported, including 57% of hepatitis A, 54% of hepatitis B, and 65% of NANB hepatitis (Table 2). However, the proportion of each type of hepatitis reported as non-Hispanic white declined. Non-Hispanic blacks in 1993 continued to represent disproportionately higher percentages of hepatitis B, accounting for 31% of all hepatitis B cases. Among black patients with any type of hepatitis, hepatitis A was the predominant type in 1993, accounting for 55% of all cases. This represents a shift from 1989, when 53% of all cases among blacks were hepatitis B cases. Data from a large population-based seroprevalence study confirm that the prevalence of HBV infection is more than four times higher among blacks than among whites (1). The percentage of blacks among NANB hepatitis patients increased from 12% in 1989 to 22% in 1992, but decreased to 18% in 1993. In 1989, Hispanic patients accounted for 9% of reported hepatitis A cases. While this percentage increased to 12% by 1993, the absolute number of Hispanic cases declined, as was true for other racial/ethnic groups. When the percentages of Hispanic cases were examined for both old reporting forms and newly revised forms for the 1990 data, there was no evidence that the coding of ethnicity separately from race affected reporting of such cases.
Analysis of Risk Factor Data
The analysis of epidemiologic data for 1993 took into consideration the changes in both incidence and reporting practices. Reporting was analyzed by groups of states to determine if significant biases existed in the data when reports from all participating states were included for analysis. Criteria for good reporting states (“core” states) included adequate serologic testing of reported cases (at least 80% of reported cases tested for IgM anti-HAV or HBsAg), and reporting to the VHSP of a high proportion of cases reported to NNDSS (at least 50% of total cases reported to NNDSS also reported to VHSP). In addition, core states were further subdivided into those with rates above the national average for each type, and those with rates below the national average, and comparisons were made between these subgroups. Trends in these core states were then compared to trends in the remaining states for evidence of consistency and potential bias.
For hepatitis A, analysis of the core group of states showed that trends were very similar between the core states and all reporting states, and between the high-rate and low-rate subgroups. In the trend analyses that follow, hepatitis A risk factors were based on reported cases from all reporting states, and trends were analyzed by using absolute numbers of cases. For hepatitis B and C/NANB, a core group of 15 states were selected using the same reporting criteria and high levels of serologic testing for HBV during 1983-1993. These states accounted for approximately 30% of all cases of hepatitis B reported to the VHSP in this period.
For hepatitis B and C/NANB hepatitis, artifactual changes in reporting levels resulted in significant differences between the trends for all VHSP states and the trends in the core states, although there were no differences between high- and low-incidence states. For hepatitis B and hepatitis C/NANB hepatitis, trends in risk factors were analyzed by using absolute numbers of cases from the core states only.
Table 3 presents crude frequencies of the potential sources of infection reported by patients with viral hepatitis. The same questionnaire was used for all patients with hepatitis, regardless of type. Although questions about selected risk factors associated primarily with hepatitis A have not always been asked for hepatitis B and NANB hepatitis, and vice versa, cases reported in 1993 have shown an improvement in this respect. Patients may also give a positive response to more than one factor; therefore, the data listed in Table 3 are not mutually exclusive.
Personal contact with a hepatitis A patient continued to be the predominant source of infection among persons with hepatitis A in 1993. The crude frequency that this potential source was reported, 34%, was similar to the rates in previous years. Many persons reported two or more potential sources of infection. Of those patients who were associated with day-care centers, 41% also reported personal contact with a hepatitis A patient, and 7% were part of a suspected foodborne or waterborne outbreak. Of those reporting contact with a hepatitis A patient, 6% also reported being part of a suspected foodborne or waterborne outbreak.
Since hepatitis A has an average incubation period of 30 days and is transmitted by the fecal-oral route, the characteristics reported by persons with hepatitis A as having occurred in the 6 weeks to 6 months prior to illness (Table 3) are generally not applicable to transmission of this virus (2). Although homosexual men are considered at increased risk of acquiring hepatitis A (3), the frequency with which homosexual activity was reported by persons with hepatitis A (3.6%) may be understated, since only 46% of the patients were asked the question in 1993. However, this percentage has increased in recent years. The frequency with which injection drug use was reported by patients with hepatitis A may be more reliable than in the past, since over 70% of patients with hepatitis A answered this question in 1993. These improvements lend greater validity to these data than in previous years.
Of patients reporting personal contact with a hepatitis A patient, 10% reported sexual, 45% reported household, and 45% reported other contact. Of those reporting other than sexual or nonsexual household contact, none had reported day-care-related exposures, but 8% reported being a part of a suspected outbreak.
To better define patterns of hepatitis A virus transmission, patients who reported more than one potential source of infection were assigned to only one group on the basis of their most probable source. These mutually exclusive groups are shown in Table 4. Contact with another person with hepatitis A was the risk factor most frequently cited. Association with a day-care center and international travel were the two risk factors next in importance.
The frequency with which the various risk factors were reported was influenced by the age of the patient. Contact with another person with hepatitis A was the most frequently reported risk factor for all age-groups, although the percentage of patients reporting this risk factor decreased with increasing age. For persons less than 15 years old, being a child in a day-care center was the next most frequently reported risk factor. For those aged15-39, contact with a day-care child or employee was the next most important risk factor. Reporting of injection drug use as a risk factor for hepatitis A dropped by 1993 to low levels (4%) for this age-group. For persons over 40 years of age, international travel and being a part of a suspected foodborne or waterborne outbreak were the next most frequent risk factors.
International travel was reported in 6% of hepatitis A cases in 1993. South and Central America were the locations visited most frequently (67% of travel-related cases in 1993). Destinations in Asia and the South Pacific were visited next most often (10% of cases in 1993). The duration of stay was 1-3 days in 17% of cases with international travel as a risk factor, 4-7 days in 15%, and more than 7 days in 68%. Among patients reporting short stays (1-3 days), over 90% reported visits to South/Central America.
Race and ethnicity were examined among hepatitis A patients with international travel as a risk factor. Hispanic patients accounted for 47%, non-Hispanic whites accounted for 43%, and Asian/Pacific Islanders for 8% of cases. Non-Hispanic blacks accounted for less than 2% of travel-related cases in 1993. There was an association between race and location visited: 92% of Hispanic patients with travel-related hepatitis A visited South/Central America, while 75% of non-Hispanics did so. Among Asian/Pacific Islander patients, 85% visited Asian/South Pacific destinations, while 0% to 7% of other races or ethnic groups visited these locations.
Because the total number of hepatitis A cases reported has changed over the years, the absolute numbers of cases for each risk factor show more accurately the trends over time for hepatitis A. The numbers of cases associated with personal contact with another hepatitis A patient during 1983-1993 have exhibited the greatest variation (Figure 2), with an increase of over 100% occurring from 1983 to 1989, followed by a comparable decrease from 1989 to 1993. Day-care-related cases increased more slowly during this period, but peaked in 1989 also, followed by a drop of 47%. The numbers of cases attributable to drug use increased steadily between 1983 and 1989, and declined rapidly to their present low level. Cases related to homosexual activity remained at low levels from 1983 through 1987. By 1989, however, there was a 3.6-fold increase in cases of hepatitis A among homosexual men, and outbreaks of hepatitis A in this population subgroup were reported. Cases among homosexual men have remained at higher levels through 1993. Foreign travel and foodborne outbreak-associated cases peaked in 1988 and declined overall since then.
Jaundice characterized an average of 85% of the reported hepatitis A cases in 1993. Although this frequency was similar across age-groups, jaundice and other symptoms are uncommon among young children infected with hepatitis A virus. Thus, reported cases substantially underestimate the infection burden among the youngest age-group. The rate of hospitalization of patients with hepatitis A has remained steady in 1993, and continues to increase with increasing age. The case-fatality rate for hepatitis A patients also increased with age, and showed a slight increase with time as well for those aged 15 and over in 1993.
Based on crude frequencies of reported risk factors, contact with another hepatitis B patient, injection drug use, and having multiple sex partners were the three most frequently reported potential sources of infection for hepatitis B patients in 1993 (Table 3). In 1993, having multiple sex partners was the most frequent potential source of infection reported. Homosexual preference was reported by 7% of hepatitis B patients during 1993. As with other types of hepatitis, several possible sources of infection were often reported for the same patient.
Seventy-two percent of the persons with hepatitis B were asked about potential risk factors commonly associated with hepatitis A that occurred within the 2 to 6 weeks prior to illness. Although these factors are generally not associated with the transmission of HBV because the incubation period is too short, health-care workers interviewing patients with hepatitis are encouraged to obtain from each patient information on all types of risk factors, both to detect newly emerging problems (as occurred with injection drug use and hepatitis A) and to ensure a complete exposure history when cases are serologically classified.
Events or conditions reported within the 6 months prior to hepatitis B illness -- such as history of dental work, surgery, acupuncture, tattooing, or other percutaneous exposures -- are not considered likely sources of sporadic infection, but are primarily useful in identifying clusters of cases at the local level.
Of three patients reported with acute hepatitis B and evidence of having responded to the hepatitis B vaccine, all three were also reported to have coinfections with acute hepatitis A. After follow-up with the reporting health department, none of these cases were found to be true candidates for breakthrough infections.
Persons who reported multiple risk factors for hepatitis B were assigned to mutually exclusive groups (2,4-6) (Table 5). As a percentage of all cases, being heterosexually active with multiple partners has replaced injection drug use as the predominant risk factor for acquisition of hepatitis B. Personal contact with another hepatitis B patient was the third most common risk factor. Of personal contacts in 1993, 68% were sexual, and 17% were nonsexual household contacts. The remaining 15% of personal contacts, classified as “other”, are unclear as to specific sources because information was insufficient to determine how transmission occurred. Employment in the medical or dental field, blood transfusions, and dialysis accounted for less than 5% of cases. For those patients employed in a medical, dental or other field involving contact with human blood, 23% reported frequent blood contact in 1993, down from 36% in 1992.
Transfusion as a source for HBV has remained at a low level (0.8%) because of routine screening of blood donors for HBsAg and anti-HBc, and because of donor selection and deferral procedures. Screening for HBsAg has been mandatory since 1972. Smaller improvements in preventing post-transfusion hepatitis B occurred in the mid-1980s, with self-exclusion of high-risk donors related to the prevention of human immunodeficiency virus (HIV) infection and later anti-HBc screening. Hepatitis B among children younger than 15 years old is associated primarily with personal contact with another infected person. The percentage for 1993, 26%, is somewhat higher than the 22% reported in 1992. None of these patients reported injection drug use, while 6% reported multiple sex partners in 1993 as their primary risk factor. The percentages of persons reporting no known source of infection in the youngest and oldest age-groups were similar to those reported in 1992.
To ensure that possible biases owing to artifactual decreases in reporting were minimized, the analysis of trends in hepatitis B risk factors for 1983 to 1993 was restricted to the absolute numbers of cases reported in the core states only. For these states during 1989-1993, decreases occurred in the numbers of cases attributed to injection drug use (an 83% decrease), personal contact with a hepatitis B patient (73% decrease), and multiple sexual partners (35% decrease).
The trends in risk factors associated with hepatitis B in the core states, among men and women separately, are shown in Figures 3 and 4. Among men, injection drug use has shown the largest change from 1983 to 1993. After an increase of 116% from 1983 to 1989, the numbers of cases among men attributed to injection drug use decreased by 85% (Figure 3). Safer needle-using practices, or changes in the types of drugs used (injection to noninjection) are possible reasons for this reduction. The numbers of cases among men attributable to personal contact with another hepatitis B patient has been more stable, showing a gradual decline from 1989 to 1993. For these male patients, 52% to 67% of contacts were sexual, while 13% to 20% were household contacts. Homosexual activity, the second most commonly reported risk factor, declined to its lowest level in 1993. Declines in the other reported risk factors -- health-care employment and blood transfusion -- continued through 1993.
Risk factors for women with hepatitis B displayed some of the same trends presented for men, with injection drug use as a risk factor increasing from1983 to a peak in 1989 (Figure 4), followed by a drop to pre-1983 levels. However, among women, contact with another hepatitis B patient increased more dramatically than among men and since 1990, was reported with a higher frequency than injection drug use. As with men, the majority of contacts associated with such cases have been sexual, reaching 72% in 1993, while only 11% have been household contacts.
The decrease in the percentage of female patients reporting medical and dental employment as a risk factor during 1983-1993 has been more pronounced than that for men. This decline is most probably attributable to immunization of health-care workers with hepatitis B vaccine. The percentage of cases attributable to blood transfusions has remained at low levels since 1988. The same trends in both men and women have been observed in the Sentinel Counties study(19).
Jaundice as a clinical characteristic of hepatitis B is a common symptom in patients over 10 years of age (Table 5); 82% of all patients were reported with jaundice, regardless of age. As with hepatitis A, jaundice and other symptoms were notably less frequent for young children, suggesting more extensive under-representation of this age-group among reported cases. Overall hospitalization rates remained stable, showing little change since 1988, but the rates of hospitalization for patients 40 years old and older dropped slowly but steadily, from 50% in 1985 to 36% in 1993. Death as a result of hepatitis B was reported in approximately 1% of patients in 1993.
Nationwide, the incidence of hepatitis B increased by 67% from 1978 to 1985 and then declined to its lowest incidence since 1974. Since its original licensing in 1981, hepatitis B vaccine has been used in increasing quantities each year. However, the role of the vaccine in the decline of the incidence of hepatitis B varies across risk groups. From 1985 to 1989, hepatitis B among homosexual men declined more rapidly than among other risk groups, not because of vaccine use but because of behavioral changes resulting from awareness of acquired immunodeficiency syndrome (AIDS)(7). Hepatitis B also declined among health-care workers during this period, who were the largest users of hepatitis B vaccine. From 1989 to 1993, hepatitis B among injection drug users declined by 46% despite the low levels of vaccine usage in this risk group. Hepatitis B among heterosexuals decreased during this period also, possibly due to wider use of vaccine.
Vaccination programs and vaccine usage have been focused primarily on three risk groups: health-care workers who are exposed to blood, staff and residents of institutions for the developmentally disabled, and staff and patients in hemodialysis units (9). For health-care and public safety workers, the Department of Labor in 1991 issued regulations that require employers to offer hepatitis B vaccine to persons at occupational risk of infection. However, the ability to immunize the groups that account for most of the HBV infections is severely limited for several reasons: the failure both of health-care providers and of the target populations to recognize the specific groups at high risk for infection; the difficulty in identifying persons with these high-risk behaviors before they become infected; and the difficulties in reaching these groups for the delivery of vaccine and at the appropriate time for vaccination (7).
Adults in general, and groups such as injecting drug users in particular, are extremely difficult to access for delivery of vaccine (11). In addition, once persons begin the lifestyles associated with a high-risk group, they may become infected before vaccine can be given. Thus, the major obstacles to reducing the incidence of HBV infection in the United States have been the difficulties in identifying persons before they become infected and vaccinating them promptly. To overcome these problems, the Immunization Practices Advisory Committee recommended in 1991 a program of routine vaccination of all infants (9). In 1995 the same committee recommended the expansion of this program to cover 1) vaccination of all unvaccinated children aged <11 years who are Pacific Islanders or who reside in households of first-generation immigrants from countries where HBV is of high or intermediate endemicity; and 2) vaccination of all 11- to 12-year-old children who have not previously received hepatitis B vaccine (9).
Hepatitis C/NonA, NonB Hepatitis
Based on the crude frequencies with which risk factors were reported, injection drug use was the risk factor most commonly reported by hepatitis C/NANB patients (Table 3). Many of these persons also reported more than one potential source of infection. Of those reporting contact with another person with hepatitis C/NANB, 25% also reported injection drug use and 5% reported employment in a medical or a dental field. Of those reporting multiple sex partners, 35% also reported injection drug use.
The behaviors commonly associated with hepatitis A that were reported by persons with hepatitis C/NANB to have occurred within 6 weeks of illness are generally not applicable to the transmission of hepatitis C/NANB (Table 3). Since transmission of NANB hepatitis by the fecal-oral route has not been demonstrated in this country, reporting an association with a foodborne or a waterborne outbreak represents misclassification of the source.
As with hepatitis B, potential exposures associated with dental work, surgery, acupuncture, tattooing, and other percutaneous procedures are not judged to be probable sources of sporadic infection (12). Hepatitis C/NANB patients with no known source of infection reported these exposures at rates no different from those of the general population.
Based on assignment to mutually exclusive categories, persons with hepatitis C/NANB reported injection drug use most frequently, accounting for 23% of cases during 1993 (Table 6). Blood transfusion accounted for 2% of cases, declining from 6% in 1990; contact with another infected person accounted for 5%, and health-care employment for 4%. Of those patients reporting health-care employment, the percentage reporting frequent (several times weekly) blood contact dropped over 1990 to 1993. Fifty-seven percent of patients employed in health-care reported frequent blood contact in 1990. By 1993, the percentage dropped to 17%. Patients classified as having multiple (2 or more) sex partners as their most likely source of infection accounted for 7% of the patients with hepatitis C/NANB; in a case-control study, this risk factor was associated with acquiring disease (12). Overall, 58% of persons reported no known source for their infection. This percentage varied by age, with > 70% of persons younger than 15 years old or 40 years old and older reporting no known source for their infection, compared with 50% for persons 15-39 years of age. Among persons less than 15, 13.6% had a history of blood transfusion.
Among persons 15 to 39 years of age, injection drug use was reported by 28% of all cases during 1993, unchanged from 1992 (Table 6). Ten percent reported multiple sex partners, 7% reported contact with another infected person, 4% reported health-care employment, and 1% reported blood transfusions. Of reported contacts with another infected person, an average of 59% were sexual contacts, 16% were household nonsexual contacts, and 25% were other (unspecified) types of contact. In prior years, persons 40 years old and older reported a history of blood transfusion most frequently among their risk factors (in 1990, 16%), but this percentage declined substantially to 4% by 1993. Injection drug use is now the most frequent risk factor for this age-group (Table 6).
Because total numbers of cases of hepatitis C/NANB have declined, trends in the distribution of risk factors are more accurately reflected by trends in the absolute numbers of cases attributed to each factor. In the core states, hepatitis C/NANB cases attributable to drug use have declined rapidly since 1988, showing a more than 62% decrease (Figure 5). A similar decrease of over 50% was seen in the Sentinel Counties Study (14).
The numbers of hepatitis C/NANB cases attributable to blood transfusions have decreased even more dramatically, dropping by 94% from 1985 to 1993. The significant decline in transfusion-associated cases, which began in the mid-1980s, resulted from a series of events: changes in the blood donor population caused by self-exclusion of high-risk donors, as part of efforts to prevent HIV infection (15,16); the introduction of screening blood donors for alanine aminotranferase and anti-HBc as surrogate markers for hepatitis C/NANB in 1986 and 1987; and use of first- and second-generation anti-Hepatitis C Virus markers for screening donors in 1990 to the present.
Jaundice was reported as a clinical symptom in 67% of reported hepatitis C/NANB patients in 1993 (Table 6). Hospitalization and case-fatality rates were higher in hepatitis C/NANB patients than in patients with hepatitis A or B. Those 40 years old and older experienced the highest rates.
The majority of NANB hepatitis cases in this country are caused by the hepatitis C virus (14); the remainder are probably due mostly to other bloodborne hepatitis agents. Outbreaks of hepatitis E, an enterically transmitted form of hepatitis NANB, have been reported in rural Mexican villages (17), as well as in Asia and North and West Africa (18), but no outbreaks have been reported in this country (19). In the United States and other countries where hepatitis E outbreaks have not been documented to occur, rare hepatitis E cases have been reported, primarily among travelers returning from HEV endemic regions (20). No secondary transmission to family members or other persons in association with these cases has been reported. In the United States, hepatitis E cases have been reported with no history of travel to HEV endemic areas; however, the mode of HEV transmission for these cases has not been determined.
Viral hepatitis surveillance in 1993 revealed several important changes from earlier years. First, total cases reported to the VHSP declined more than 51% from 1990 to 1993, as a result of both real declines in the incidence of hepatitis A and B, and a number of states that previously reported now submitting fewer or none of their cases to the VHSP. Second, the use of serologic tests to diagnose the specific type of hepatitis has declined, with fewer reported cases being diagnosed on the basis of tests for both hepatitis A and B. Third, analysis of trends in risk factors for the acquisition of the different types of hepatitis indicated that injection drug use has declined dramatically for hepatitis A, B, and hepatitis C/NANB. Finally, more widespread use of hepatitis B vaccine