- Email: [email protected]
Outbreak of Hepatitis B among injecting drug users in Denmark
Journal of Clinical Virology 22 (2001) 133– 141 www.elsevier.com/locate/jcv
Outbreak of Hepatitis B among injecting drug users in Denmark P.B. Christensen a,*, H.B. Krarup b, H.G.M. Niesters c, H. Norder d, O.B. Schaffalitzky de Muckadell e, B. Jeune f, J. Georgsen a a
Department of Clinical Immunology, Odense Uni6ersity Hospital, 500 Odense, Denmark b Department of Clinical Chemistry, Aalborg Hospital, Aalborg, Denmark c Department of Virology, Uni6ersity Hospital Rotterdam Sophia/Dijkzigt, Rotterdam, The Netherlands d Department of Virology, Smittskyddsinstitutet, Stockholm, Sweden e Department of Medical Gastroenterology S, Odense Uni6ersity Hospital, Odense, Denmark f Institute of Public Health, Uni6ersity of Southern Denmark, Odense Uni6ersity, Odense, Denmark Received 30 January 2001; accepted 7 April 2001
Abstract Background: The incidence of hepatitis B is low in Denmark, but injecting drug users (IDUs) remains a high-risk group for this infection. Objecti6es: The aim of the study was to describe a hepatitis B outbreak among IDUs by comparing existing registers. Additionally, we wanted to analyze the genetic variation of the hepatitis B virus involved in the outbreak. Study design: In the County of Funen, registers of laboratory diagnosis, hospital records and reports from clinicians to the Medical Officer of Health (MOH) were compared between 1992 and 1998. HBsAg positive sera recovered from the epidemic were sequenced and compared to known HBV strains. Results: We identified 648 cases of hepatitis B of which 51% (332) were acute infections. The laboratory database identified 96% (319/332) of these, 45% (150/332) were admitted to hospital and 38% (127/332) were reported to public health. By capture– recapture analysis based on MOH reports and hospital records the estimated total number of acute cases were 334 (95% C.I. 283–385). We sequenced 75 HBsAg positive samples and identified two very similar strains of genotype D (serotype ayw3) among IDUs involved in the outbreak. Conclusions: The current surveillance system did not detect the majority of acute hepatitis B cases in County of Funen. We suggest laboratory-based surveillance of hepatitis B to be implemented at a national level as this may identify new outbreaks faster and more complete than the current surveillance system. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Hepatitis B epidemiology; Hepatitis B surveillance; Outbreak analysis; Molecular epidemiology; Injecting drug users
Abbre6iations: HBV, hepatitis B virus; IDU, injecting drug user; MOH, medical officer of health. * Corresponding author. Tel.: + 45-65-413578; fax: + 45-65-417975. E-mail address: [email protected] (P.B. Christensen). 1386-6532/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 6 - 6 5 3 2 ( 0 1 ) 0 0 1 7 5 - 5
134
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
1. Introduction
2. Materials and methods
Denmark is a country of low endemicity for hepatitis B infection. The annual number of reported cases is 2/100 000 inhabitants per year and has slightly decreasing for the last decade. The reported incidence among injecting drug users (IDUs) is 50– 100 times higher than in the general population, and corresponded to 32% of all cases in the period between 1991 and 1998. In the same period, 30% of reported cases had an unknown route of transmission. Chronic HBV infection was not registered by the public health authorities until April 2000 and the size of this population is currently unknown. In two recent Danish studies of 247 and 347 IDUs, respectively, the prevalence of hepatitis B core antibodies (anti-HBc) was 64 and 68%, HBsAg was found in 5 and 10% (Fuglsang et al., 2000; Christensen, 2000). All health statistics in Denmark are based on the personal identification number (PIN), a unique number assigned to all permanent residents of the country, however IDUs have an unstable contact to health services and morbidity for this group may be underreported. According to the Danish reporting system of communicable diseases a clinician who makes the diagnosis of acute hepatitis B in a patient is obliged to send a case report to the local Medical officer of Health (MOH). The obligation to report HBV infections is limited to the clinician making the diagnosis, laboratories and clinicians involved at a later state of patient care are not required to report. The standard case form is reported with PIN code and contains demographic, clinical and serological information as well as suspected route of transmission. The case form is used for all infectious diseases (except for HIV), that are mandatory to report to MOH, and do not focus on HBV specific questions. The aim of this study was to compare the coverage of registers for hepatitis B cases diagnosed in the laboratory, cases admitted to the hospital, and cases reported to the Medical Officer of Health (MOH) for the County of Funen during a hepatitis B outbreak. Additionally, we wanted to examine the genetic relationship of the HBV strains involved in the outbreak.
2.1. Register comparison The Department of Clinical Immunology, Odense University Hospital performs all laboratory testing for hepatitis in the County of Funen (population 471 732) and has stored HBsAg positive samples since 1993. Based on the PIN code all new cases positive for HBsAg by ELISA (OrthoClinical Diagnostics GmbH, Neckargemu¨ nd, Germany) and confirmed by neutralization assay between January 1, 1992 and December 12, 1998 were retrieved from the laboratory database that contains all test results from January 1, 1980 and onwards. Anonymous samples, samples with incorrect PIN, and cases positive for HBsAg prior to January 1, 1992 were excluded. All reports of hepatitis B infection from clinicians to MOH between 1992 and 1998 were reviewed. Records of patients discharged from the hospital with the diagnosis of acute hepatitis B (ICD 9 code 070.02, 070.03 between 1992 and 1993 and ICD10 code B16.0-B16.9 between 1994 and 1998) were retrieved from the county register of hospital admissions (Anonymous, 1986, 1993). Furthermore, ICD10 codes B18.0-B18.9 (chronic viral hepatitis) were retrieved to identify patients with acute hepatitis B that were classified as chronic infections. To identify cases with acute hepatitis B coded under a different diagnosis, all hospital admissions 1 year before and 1 year after serological diagnosis or reports to MOH were identified in a second retrieval. If hepatitis B was found to be an underlying or contributing cause for admission, the case was included. For hospitalized patients and MOH reported cases not HBsAg positive in the laboratory database, all tests for hepatitis B markers 1980–1998 were retrieved. We defined acute hepatitis B as jaundice and the presence of HBsAg and/or anti-HBc IgM. Seroconversion for HBsAg or one sample positive for anti-HBc IgM was required for the definition of cases without known clinical illness. Chronic HBV infection was defined as HBsAg present for more than 6 months or the presence of HBsAg in the absence of anti-HBc IgM. Due to different
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
test strategies among the requiring physician not all samples were tested for anti-HBc IgM, and not all HBsAg positive samples had a follow up test to determine the duration of infection.Vital status at the end of 1998 was determined from the civil register. To estimate the completeness of our data, we performed a capture– recapture analysis using hospital admissions and MOH reports as two independent samples of the target population (Hook and Regal, 1992; Ackman et al., 1996).
2.2. HBV – DNA sequencing We selected samples from all available acute cases and chronic carriers among known IDUs. To construct the phylogenetic tree we included chronic HBV carriers of different geographical origin diagnosed in Funen in the same period but unrelated to the outbreak. Additionally we received samples from other Danish laboratories from acute cases of hepatitis B among IDUs diagnosed during 1996. PCR analysis was performed blinded in two independent reference laboratories by H.G.M.N and H.N as previously described (Yokosuka et al., 1993; Niesters et al., 1998; Norder et al., 1990, 1994). In brief HBV– DNA was extracted and amplified by PCR standard procedures. Nucleotides 56–806, corresponding to the HBsAg gene were amplified and both strands were sequenced (nucleotides 459–806). For phylogenetic analysis an unweighted pair group with aritmethic mean (UPGMA) analysis was performed based on nucleotide 512– 688. The analysis was performed with the program Geneworks software version 2.5 (Oxford Molecular, Oxford, UK). The S-gene (nucleotide 155– 832) of selected samples were sequenced for comparison with previously reported HBV isolates. Data processing and statistical analysis were performed with Epi info version 6.04 (Dean et al., 1997). Non-parametric tests were used for statistical analysis, ( 2 test, Fisher’s exact test or MannWhitney test as appropriate). Level of statistical significance was set to P B 0.05. The study was approved by the Danish Data Protections Agency.
135
3. Results
3.1. Registers The registered number of HBV infections between 1982 and 1998 in the County of Funen is shown in Fig. 1. In 1994, the MOH detected an outbreak of hepatitis B among IDUs in the county (indicated by arrow). From a yearly average below 10 cases (2.1/100 000), the number of reported cases rose to 38 in 1994 (8.1/100 000), and declined slowly during the next 5 years (Anonymous, 1994a,b, 1995). A simultaneous outbreak with 15 cases was reported among IDUs in Esbjerg, a town 150 km west of Funen (Helberg and Gaub, 1995). The figures of hospital admissions and reports from clinicians to MOH were grossly similar but the number of laboratory cases detected were 2– 3 timers higher. Furthermore, the outbreak could be identified in 1991 from the laboratory data when the number of new cases rose to more than twice the annual average of the 5 preceding years. By this definition the outbreak could not be identified from the reports to MOH until 2 years later in 1993. A total of 160 cases of acute hepatitis B infections were reported to the MOH between 1992 and 1998. Of these 80% (127) were concluded to be true cases, 14% (22) were chronic HBV infections and 7% (11) did not have HBV infection. Among the acute cases, 59% (76/127) indicated transmission related to IDU, 11% (14/127) were heterosexual transmission, 7% (9/127) had other known routes of transmission and 22% (28/127) had unknown route of transmission. A total of 187 patients was identified in the hospital database, of whom 150 had acute hepatitis B. Seventy-two percent of acute cases were coded correctly. The 150 acute cases had 180 hepatitis B related hospital admissions with a total of 1533 bed-days (median 8, range 1–65). Four patients developed fulminant hepatitis and the 6 month mortality was 1.3% (2/150). During follow-up additionally five cases with chronic hepatitis B died from hepatic failure: three infected during the epidemic and two with unknown time of transmission.
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
Fig. 1. Hepatitis B in County of Funen, 1983 – 1998.
136
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
We detected 615 new HBsAg positive cases in the laboratory database between 1992 and 1998, of whom 365 were not found in the other databases. In order to evaluate reporting efficacy, all cases were classified as acute or chronic infections, as it was mandatory to report only acute HBV infection. Based on clinical and serological information, 220 cases were classified as acute and 48 as chronic infections. By serological definitions only, 112 cases were classified as acute and 163 as chronic infections. Ninety HBsAg positive cases could not be classified by our case-definition and were excluded from analysis. The overlap of the three populations was only moderate (Fig. 2). Among a total of 648 cases, 332 acute HBV infections were identified corresponding to an average of 10.1/100 000/year. Median age was 30 years, and 70% were men. Among the acute hepatitis B infections 96% (319/332) were detected by laboratory tests, 45% (150/332) were hospitalized, and only 38% (127/332) were reported to the MOH. Reporting efficacy was identical in and outside the hospital. Ten of 220 acute cases (4.5%) reported or admitted to hospital were HBsAg negative and anti-HBc IgM positive. Capture– recapture analysis of MOH reported cases and hospital admissions estimated 112 cases to be unknown by both samples corresponding to a total of 334 cases (95% C.I. 283– 385) (Table 1).
137
3.2. DNA sequencing A phylogenetic tree of part of the S-gene (nt 512–688) by UPGMA (unweighted pair group with aritmethic mean) is shown in Fig. 3. Two HBV strains were found in the vast majority of IDUs, both were genotype D and serotype ayw3 (serotyping performed by Dr Courouce´ , Paris). One strain was previously found among Scandinavian IDUs (24 cases) (Genbank accession number AF372623). The other (40 cases) was a new variant with a nucleotide change from C to T in position 582 leading to a change in amino acid 143 of HBsAg from serine (S) to leucine (L) (hereafter named L143) (Genbank accession number AF372622). This was the only difference between the two isolates in the entire S-gene. Patients with the two isolates did not differ according to age, gender, year of infection, admission to hospital or mortality. However, the prevalence of L143 was higher in the major city (Odense) compared to the rest of the county (P= 0.037). Sequences different from the two were found among 7% (3/46) of IDUs from Funen compared to 33% (4/12) of sequences from other acute cases (patients with unknown route of transmission and IDUs with acute hepatitis B from other parts of Denmark) (P= 0.028) (Fig. 3). Chronic carriers, who were not IDUs, clustered separately in agreement with their geographical origin.
Fig. 2. Venn diagram of database overlap for cases of acute hepatitis B, Funen, 1992 – 1998.
138
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
Fig. 3. Phylogenetic tree of Danish HBV strains with data of transmission and predicted serotypes.
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
139
Table 1 Capture–recapture estimate of population with acute hepatitis B Admitted to hospital
Reported to medical officer of health
Yes No
Total a
Yes
No
Total
57 93 150
70 112a
127 334a
Calculation: 93*70/57= 112, 150*127/57= 334 (95% C.I. 283–385) (Hook and Regal, 1992; Ackman et al., 1996).
4. Discussion In this study, we found that current public health disease surveillance significantly underestimated the incidence of acute hepatitis B infection in the County of Funen. Among definite cases of acute HBV infection only 38% were reported to MOH, and the reporting was possibly even lower, as 90 HBsAg positive cases could not be classified as acute or chronic infection. It was an unexpected finding that the reporting frequency was equally low among hospitalized patients as these are the more severe cases. Our study did not examine why the underreporting took place, so we can only speculate for the reasons. Cases that were misclassified as other diseases (28% of hospitalized cases) cannot be expected to be reported to MOH as acute hepatitis B infections. There are no economic remuneration to private practitioners to fill out the case reports, and most medical doctors diagnose less than one case per year making it easy to forget the mandatory report. The epidemic was recognized by the national surveillance system in February 1994 (Anonymous, 1994b), but Fig. 1 indicates that the outbreak could have been detected 2 years earlier if a laboratory based surveillance had been implemented. The outbreak continued for 4 years after it had been recognized and a prolonged duration of IDU HBV outbreaks have been reported previously (Lettau et al., 1987; Shattock et al., 1982). In spite of this, the county did not give economic priority to a HBV vaccination campaign among IDUs proposed in 1994 by the authors. The need for vaccination was demonstrated by a 1996 survey in the county where only 2% of IDUs had been vaccinated (Christensen, 2000).
A large proportion of HBsAg positive patients was chronically infected. Chronic HBV infection should be evaluated for treatment, and it is remarkable that 34% (220/648) of new HBsAg positive cases did not have a follow up test performed to determine the outcome of the infection. Due to our study design we could have missed acute cases that were only anti-HBc IgM positive and misclassified acute infections that were positive for HBsAg more than 6 months as chronic infections. The capture– recapture analysis indicated that our retrieval was close to the estimated total number of acute cases, and that the laboratory detected nearly all cases known to the health care system. In Finland the diagnostic laboratories were incorporated into the hepatitis B surveillance system in 1994 and this more than doubled the number of notified cases in the first year of functioning (Pebody et al., 1999). Among adults, 2/3 of acute hepatitis B infections are found to be subclinical and these are unlikely to be detected unless population surveys are applied (McMahon et al., 1985). This suggests that the total number of transmissions during the outbreak could have been in the order of 1000 cases compared to an estimated IDU population in Funen of 1300 persons (Christensen, 2000). This indicates that the majority of the IDU population was infected during the outbreak and it is possible that the epidemic stopped not due to preventive measures, but because there was none left to infect. The ‘ceiling-effect’ as reason for the outbreak termination has to our knowledge not been described in other HBV outbreaks but is supported by a survey in 1996, where 70% of IDUs in the county drug treatment center were anti-HBc positive as sign of previous HBV exposure (Christensen, 2000).
140
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
Cyclic outbreaks of hepatitis B among IDUs have been reported in several countries, and are supposed to reflect the gradual accumulation of susceptibles (Christenson, 1987a,b; Levine et al., 1994). Statistics of drug related deaths indicated that the Funen IDU population grew significantly quicker than in the rest of Denmark during the first part of the 1990s (Christensen, 2000). This is in agreement with previous reports where HBV outbreaks among IDUs have been preceded by an increase in the number of IDUs in the community (Lettau et al., 1987; Shattock et al., 1982). The sequence analysis of HBV showed that several index cases were involved and that two strains dominated. Although individual cases could be classified as related or unrelated to the outbreak, it was not possible to demonstrate individual chains of transmission, because of too little variation between the strains in the investigated region of the genome. The L143 mutation was not found among genotype D reported to Genbank, but it was previously found in other genotypes of HBV. There was no evidence of changed pattern of virulence of L143 compared to ayw3 based on duration or severity of clinical illness, and it is probably not a new mutation. It was found in an IDU who had been HBsAg positive for 18 years, and was also found in samples from outside Funen. The results indicate that, rather than being introduced from outside, the epidemic arose from the resident IDU population of chronic carriers. L143 has been reported from Spain where it was found among three IDUs, but has not previously been found among Scandinavian, British or Dutch IDU isolates. (Dr H. Norder, Dr C.G. Teo, Dr H.G.N. Niesters, personal communication) Recently L143 was described among dialysis patient in Germany indicating that the strain is not confined to IDU populations (Petzold et al., 1999). In conclusion, the hepatitis B outbreak among IDUs was associated with a significant morbidity and mortality. The current surveillance system was unable to identify the majority of cases and we suggest laboratory surveillance for hepatitis B to be implemented in Denmark.
Acknowledgements Kirsten Kronborg, Medical Officer of Health, County of Fynen provided the data of reported cases of acute hepatitis in the County of Funen. Sera from IDUs with acute hepatitis B outside Funen were provided by Mads Buhl, Department of Infectious Diseases, Marselisborg Hospital, Anders Siboni, Department of medicine, Esbjerg hospital and Svend E.H. Jacobsen, Department of Clinical Immunology, Sønderborg Hospital. Dr A. M. Courouce´ , Institute National de la Transfusion Sanguine, Paris performed serological subtyping of the two strains of HBV detected in the outbreak.
References Ackman DM, Birkhead G, Flynn M. Assessment of surveillance for meningococcal disease in New York State, 1991. Am J Epidemiol 1996;144:78 – 82. Anonymous. International Statistical Classification of Diseases and Health Related Problems. 8th revised ed. National Board of Health, Copenhagen, 1986. Anonymous. International Statistical Classification of Diseases and Health Related Problems. 10th ed. National Board of Health and Munksgaard, Copenhagen, 1993. Anonymous. Hepatitis B 1993. Epi-news, 50. Copenhagen, Statens Serum Institut, 1994a. Anonymous. Hepatitis B outbreak among injecting drug uses in County of Funen. Epi-news, 8. Copenhagen, Statens Serum Institut, 1994b. Anonymous, Hepatitis B. Epi-news, 41. Copenhagen, Statens Serum Institut, 1995. Christensen PB. Viral infections in Funen. University of Southern Denmark, Odense, 2000. Christenson B [Transmission of hepatitis B — a model for the evaluation of future HIV epidemic outbreaks](in Swedish). Lakartidningen 1987a, 84: 4037 – 4041. Christenson B. Epidemiology of hepatitis B in Sweden. J Inf 1987b;15:269 – 77. Dean AG, Dean JA, Columbier D, Brendel DA, Smith DC, Burton AH, Dicker RC, Sullivan K, Fagan RF, Arner TG. Epi Info Version 6.04: a word-processing, database, and statistics program for epidemiology on microcomputers. Atlanta, GA, 1997. Fuglsang T, Fouchard J, Ege P. Prevalence of HIV, hepatitis B, and C among drug users in the City of Copenhagen. Ugeskr Laeger 2000;162:3860 – 4. Helberg JC, Gaub J. [Hepatitis epidemic amongst intravenous drug-users in Esbjerg, July 1993 to May 1994] (in Danish). Ugeskr Laeger, 1995, 157/41: 5701 – 5703.
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141 Hook EB, Regal RR. The value of capture-recapture methods even for apparent exhaustive surveys. The need for adjustment for source of ascertainment intersection in attempted complete prevalence studies. Am J Epidemiol 1992;135:1060– 7. Lettau LA, McCarthy JG, Smith MH, Hadler SC, Morse LJ, Ukena T, Bessette R, Gurwitz A, Irvine WG, Fields HA, et al. Outbreak of severe hepatitis due to delta and hepatitis B viruses in parenteral drug abusers and their contacts. N Engl J Med 1987;317:1256 –62. Levine OS, Vlahov D, Nelson KE. Epidemiology of hepatitis B virus infections among injecting drug users: seroprevalence, risk factors, and viral interactions. Epidemiol Rev 1994;16:418 – 36. McMahon BJ, Alward WL, Hall DB, Heyward WL, Bender TR, Francis DP, Maynard JE. Acute hepatitis B virus infection: relation of age to the clinical expression of disease and subsequent development of the carrier state. J Infect Dis 1985;151:599 –603.1 Niesters HG, Honkoop P, Haagsma EB, de-Man RA, Schalm SW, Osterhaus AD. Identification of more than one mutation in the hepatitis B virus polymerase gene arising during prolonged lamivudine treatment. J Infect Dis 1998;177:1382 – 5.
.
141
Norder H, Courouce´ A-M, Magnius LO. Complete genomes, phylogenetic relatedness, and structural proteins of six strains of the Hepatitis B virus, four of which represent two new genotypes. Virology 1994;198:489 – 503. Norder H, Hammas B, Magnius LO. Typing of hepatitis B virus genomes by a simplified polymerase chain reaction. J Med Virol 1990;31:215 – 21. Pebody RG, Ruutu P, Nohynek H, Leinikki P. Changing epidemiology of hepatitis B infection in Finland. Scand J Infect Dis 1999;31:251 – 4. Petzold DR, Tautz B, Wolf F, Drescher J. Infection chains and evolution rates of hepatitis B virus in cardiac transplant recipients infected nosocomially. J Med Virol 1999;58:1 – 10. Shattock A, Kelly M, Fielding J, Arthurs Y. Epidemic Hepatitis B with delta-antigenaemia among Dublin drugabusers. Ir J Med Sci 1982;151:334 – 8. Yokosuka O, Tagawa M, Omata M. PCR Detection of Hepatitis B Virus. In: Persing DH, Smith TF, editors. Diagnostic Molecular Microbiology, Principles and Applications. American Society for Microbiology, Washington DC, 1993. p. 322 – 326.
Outbreak of Hepatitis B among injecting drug users in Denmark P.B. Christensen a,*, H.B. Krarup b, H.G.M. Niesters c, H. Norder d, O.B. Schaffalitzky de Muckadell e, B. Jeune f, J. Georgsen a a
Department of Clinical Immunology, Odense Uni6ersity Hospital, 500 Odense, Denmark b Department of Clinical Chemistry, Aalborg Hospital, Aalborg, Denmark c Department of Virology, Uni6ersity Hospital Rotterdam Sophia/Dijkzigt, Rotterdam, The Netherlands d Department of Virology, Smittskyddsinstitutet, Stockholm, Sweden e Department of Medical Gastroenterology S, Odense Uni6ersity Hospital, Odense, Denmark f Institute of Public Health, Uni6ersity of Southern Denmark, Odense Uni6ersity, Odense, Denmark Received 30 January 2001; accepted 7 April 2001
Abstract Background: The incidence of hepatitis B is low in Denmark, but injecting drug users (IDUs) remains a high-risk group for this infection. Objecti6es: The aim of the study was to describe a hepatitis B outbreak among IDUs by comparing existing registers. Additionally, we wanted to analyze the genetic variation of the hepatitis B virus involved in the outbreak. Study design: In the County of Funen, registers of laboratory diagnosis, hospital records and reports from clinicians to the Medical Officer of Health (MOH) were compared between 1992 and 1998. HBsAg positive sera recovered from the epidemic were sequenced and compared to known HBV strains. Results: We identified 648 cases of hepatitis B of which 51% (332) were acute infections. The laboratory database identified 96% (319/332) of these, 45% (150/332) were admitted to hospital and 38% (127/332) were reported to public health. By capture– recapture analysis based on MOH reports and hospital records the estimated total number of acute cases were 334 (95% C.I. 283–385). We sequenced 75 HBsAg positive samples and identified two very similar strains of genotype D (serotype ayw3) among IDUs involved in the outbreak. Conclusions: The current surveillance system did not detect the majority of acute hepatitis B cases in County of Funen. We suggest laboratory-based surveillance of hepatitis B to be implemented at a national level as this may identify new outbreaks faster and more complete than the current surveillance system. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Hepatitis B epidemiology; Hepatitis B surveillance; Outbreak analysis; Molecular epidemiology; Injecting drug users
Abbre6iations: HBV, hepatitis B virus; IDU, injecting drug user; MOH, medical officer of health. * Corresponding author. Tel.: + 45-65-413578; fax: + 45-65-417975. E-mail address: [email protected] (P.B. Christensen). 1386-6532/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 6 - 6 5 3 2 ( 0 1 ) 0 0 1 7 5 - 5
134
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
1. Introduction
2. Materials and methods
Denmark is a country of low endemicity for hepatitis B infection. The annual number of reported cases is 2/100 000 inhabitants per year and has slightly decreasing for the last decade. The reported incidence among injecting drug users (IDUs) is 50– 100 times higher than in the general population, and corresponded to 32% of all cases in the period between 1991 and 1998. In the same period, 30% of reported cases had an unknown route of transmission. Chronic HBV infection was not registered by the public health authorities until April 2000 and the size of this population is currently unknown. In two recent Danish studies of 247 and 347 IDUs, respectively, the prevalence of hepatitis B core antibodies (anti-HBc) was 64 and 68%, HBsAg was found in 5 and 10% (Fuglsang et al., 2000; Christensen, 2000). All health statistics in Denmark are based on the personal identification number (PIN), a unique number assigned to all permanent residents of the country, however IDUs have an unstable contact to health services and morbidity for this group may be underreported. According to the Danish reporting system of communicable diseases a clinician who makes the diagnosis of acute hepatitis B in a patient is obliged to send a case report to the local Medical officer of Health (MOH). The obligation to report HBV infections is limited to the clinician making the diagnosis, laboratories and clinicians involved at a later state of patient care are not required to report. The standard case form is reported with PIN code and contains demographic, clinical and serological information as well as suspected route of transmission. The case form is used for all infectious diseases (except for HIV), that are mandatory to report to MOH, and do not focus on HBV specific questions. The aim of this study was to compare the coverage of registers for hepatitis B cases diagnosed in the laboratory, cases admitted to the hospital, and cases reported to the Medical Officer of Health (MOH) for the County of Funen during a hepatitis B outbreak. Additionally, we wanted to examine the genetic relationship of the HBV strains involved in the outbreak.
2.1. Register comparison The Department of Clinical Immunology, Odense University Hospital performs all laboratory testing for hepatitis in the County of Funen (population 471 732) and has stored HBsAg positive samples since 1993. Based on the PIN code all new cases positive for HBsAg by ELISA (OrthoClinical Diagnostics GmbH, Neckargemu¨ nd, Germany) and confirmed by neutralization assay between January 1, 1992 and December 12, 1998 were retrieved from the laboratory database that contains all test results from January 1, 1980 and onwards. Anonymous samples, samples with incorrect PIN, and cases positive for HBsAg prior to January 1, 1992 were excluded. All reports of hepatitis B infection from clinicians to MOH between 1992 and 1998 were reviewed. Records of patients discharged from the hospital with the diagnosis of acute hepatitis B (ICD 9 code 070.02, 070.03 between 1992 and 1993 and ICD10 code B16.0-B16.9 between 1994 and 1998) were retrieved from the county register of hospital admissions (Anonymous, 1986, 1993). Furthermore, ICD10 codes B18.0-B18.9 (chronic viral hepatitis) were retrieved to identify patients with acute hepatitis B that were classified as chronic infections. To identify cases with acute hepatitis B coded under a different diagnosis, all hospital admissions 1 year before and 1 year after serological diagnosis or reports to MOH were identified in a second retrieval. If hepatitis B was found to be an underlying or contributing cause for admission, the case was included. For hospitalized patients and MOH reported cases not HBsAg positive in the laboratory database, all tests for hepatitis B markers 1980–1998 were retrieved. We defined acute hepatitis B as jaundice and the presence of HBsAg and/or anti-HBc IgM. Seroconversion for HBsAg or one sample positive for anti-HBc IgM was required for the definition of cases without known clinical illness. Chronic HBV infection was defined as HBsAg present for more than 6 months or the presence of HBsAg in the absence of anti-HBc IgM. Due to different
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
test strategies among the requiring physician not all samples were tested for anti-HBc IgM, and not all HBsAg positive samples had a follow up test to determine the duration of infection.Vital status at the end of 1998 was determined from the civil register. To estimate the completeness of our data, we performed a capture– recapture analysis using hospital admissions and MOH reports as two independent samples of the target population (Hook and Regal, 1992; Ackman et al., 1996).
2.2. HBV – DNA sequencing We selected samples from all available acute cases and chronic carriers among known IDUs. To construct the phylogenetic tree we included chronic HBV carriers of different geographical origin diagnosed in Funen in the same period but unrelated to the outbreak. Additionally we received samples from other Danish laboratories from acute cases of hepatitis B among IDUs diagnosed during 1996. PCR analysis was performed blinded in two independent reference laboratories by H.G.M.N and H.N as previously described (Yokosuka et al., 1993; Niesters et al., 1998; Norder et al., 1990, 1994). In brief HBV– DNA was extracted and amplified by PCR standard procedures. Nucleotides 56–806, corresponding to the HBsAg gene were amplified and both strands were sequenced (nucleotides 459–806). For phylogenetic analysis an unweighted pair group with aritmethic mean (UPGMA) analysis was performed based on nucleotide 512– 688. The analysis was performed with the program Geneworks software version 2.5 (Oxford Molecular, Oxford, UK). The S-gene (nucleotide 155– 832) of selected samples were sequenced for comparison with previously reported HBV isolates. Data processing and statistical analysis were performed with Epi info version 6.04 (Dean et al., 1997). Non-parametric tests were used for statistical analysis, ( 2 test, Fisher’s exact test or MannWhitney test as appropriate). Level of statistical significance was set to P B 0.05. The study was approved by the Danish Data Protections Agency.
135
3. Results
3.1. Registers The registered number of HBV infections between 1982 and 1998 in the County of Funen is shown in Fig. 1. In 1994, the MOH detected an outbreak of hepatitis B among IDUs in the county (indicated by arrow). From a yearly average below 10 cases (2.1/100 000), the number of reported cases rose to 38 in 1994 (8.1/100 000), and declined slowly during the next 5 years (Anonymous, 1994a,b, 1995). A simultaneous outbreak with 15 cases was reported among IDUs in Esbjerg, a town 150 km west of Funen (Helberg and Gaub, 1995). The figures of hospital admissions and reports from clinicians to MOH were grossly similar but the number of laboratory cases detected were 2– 3 timers higher. Furthermore, the outbreak could be identified in 1991 from the laboratory data when the number of new cases rose to more than twice the annual average of the 5 preceding years. By this definition the outbreak could not be identified from the reports to MOH until 2 years later in 1993. A total of 160 cases of acute hepatitis B infections were reported to the MOH between 1992 and 1998. Of these 80% (127) were concluded to be true cases, 14% (22) were chronic HBV infections and 7% (11) did not have HBV infection. Among the acute cases, 59% (76/127) indicated transmission related to IDU, 11% (14/127) were heterosexual transmission, 7% (9/127) had other known routes of transmission and 22% (28/127) had unknown route of transmission. A total of 187 patients was identified in the hospital database, of whom 150 had acute hepatitis B. Seventy-two percent of acute cases were coded correctly. The 150 acute cases had 180 hepatitis B related hospital admissions with a total of 1533 bed-days (median 8, range 1–65). Four patients developed fulminant hepatitis and the 6 month mortality was 1.3% (2/150). During follow-up additionally five cases with chronic hepatitis B died from hepatic failure: three infected during the epidemic and two with unknown time of transmission.
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
Fig. 1. Hepatitis B in County of Funen, 1983 – 1998.
136
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
We detected 615 new HBsAg positive cases in the laboratory database between 1992 and 1998, of whom 365 were not found in the other databases. In order to evaluate reporting efficacy, all cases were classified as acute or chronic infections, as it was mandatory to report only acute HBV infection. Based on clinical and serological information, 220 cases were classified as acute and 48 as chronic infections. By serological definitions only, 112 cases were classified as acute and 163 as chronic infections. Ninety HBsAg positive cases could not be classified by our case-definition and were excluded from analysis. The overlap of the three populations was only moderate (Fig. 2). Among a total of 648 cases, 332 acute HBV infections were identified corresponding to an average of 10.1/100 000/year. Median age was 30 years, and 70% were men. Among the acute hepatitis B infections 96% (319/332) were detected by laboratory tests, 45% (150/332) were hospitalized, and only 38% (127/332) were reported to the MOH. Reporting efficacy was identical in and outside the hospital. Ten of 220 acute cases (4.5%) reported or admitted to hospital were HBsAg negative and anti-HBc IgM positive. Capture– recapture analysis of MOH reported cases and hospital admissions estimated 112 cases to be unknown by both samples corresponding to a total of 334 cases (95% C.I. 283– 385) (Table 1).
137
3.2. DNA sequencing A phylogenetic tree of part of the S-gene (nt 512–688) by UPGMA (unweighted pair group with aritmethic mean) is shown in Fig. 3. Two HBV strains were found in the vast majority of IDUs, both were genotype D and serotype ayw3 (serotyping performed by Dr Courouce´ , Paris). One strain was previously found among Scandinavian IDUs (24 cases) (Genbank accession number AF372623). The other (40 cases) was a new variant with a nucleotide change from C to T in position 582 leading to a change in amino acid 143 of HBsAg from serine (S) to leucine (L) (hereafter named L143) (Genbank accession number AF372622). This was the only difference between the two isolates in the entire S-gene. Patients with the two isolates did not differ according to age, gender, year of infection, admission to hospital or mortality. However, the prevalence of L143 was higher in the major city (Odense) compared to the rest of the county (P= 0.037). Sequences different from the two were found among 7% (3/46) of IDUs from Funen compared to 33% (4/12) of sequences from other acute cases (patients with unknown route of transmission and IDUs with acute hepatitis B from other parts of Denmark) (P= 0.028) (Fig. 3). Chronic carriers, who were not IDUs, clustered separately in agreement with their geographical origin.
Fig. 2. Venn diagram of database overlap for cases of acute hepatitis B, Funen, 1992 – 1998.
138
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
Fig. 3. Phylogenetic tree of Danish HBV strains with data of transmission and predicted serotypes.
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
139
Table 1 Capture–recapture estimate of population with acute hepatitis B Admitted to hospital
Reported to medical officer of health
Yes No
Total a
Yes
No
Total
57 93 150
70 112a
127 334a
Calculation: 93*70/57= 112, 150*127/57= 334 (95% C.I. 283–385) (Hook and Regal, 1992; Ackman et al., 1996).
4. Discussion In this study, we found that current public health disease surveillance significantly underestimated the incidence of acute hepatitis B infection in the County of Funen. Among definite cases of acute HBV infection only 38% were reported to MOH, and the reporting was possibly even lower, as 90 HBsAg positive cases could not be classified as acute or chronic infection. It was an unexpected finding that the reporting frequency was equally low among hospitalized patients as these are the more severe cases. Our study did not examine why the underreporting took place, so we can only speculate for the reasons. Cases that were misclassified as other diseases (28% of hospitalized cases) cannot be expected to be reported to MOH as acute hepatitis B infections. There are no economic remuneration to private practitioners to fill out the case reports, and most medical doctors diagnose less than one case per year making it easy to forget the mandatory report. The epidemic was recognized by the national surveillance system in February 1994 (Anonymous, 1994b), but Fig. 1 indicates that the outbreak could have been detected 2 years earlier if a laboratory based surveillance had been implemented. The outbreak continued for 4 years after it had been recognized and a prolonged duration of IDU HBV outbreaks have been reported previously (Lettau et al., 1987; Shattock et al., 1982). In spite of this, the county did not give economic priority to a HBV vaccination campaign among IDUs proposed in 1994 by the authors. The need for vaccination was demonstrated by a 1996 survey in the county where only 2% of IDUs had been vaccinated (Christensen, 2000).
A large proportion of HBsAg positive patients was chronically infected. Chronic HBV infection should be evaluated for treatment, and it is remarkable that 34% (220/648) of new HBsAg positive cases did not have a follow up test performed to determine the outcome of the infection. Due to our study design we could have missed acute cases that were only anti-HBc IgM positive and misclassified acute infections that were positive for HBsAg more than 6 months as chronic infections. The capture– recapture analysis indicated that our retrieval was close to the estimated total number of acute cases, and that the laboratory detected nearly all cases known to the health care system. In Finland the diagnostic laboratories were incorporated into the hepatitis B surveillance system in 1994 and this more than doubled the number of notified cases in the first year of functioning (Pebody et al., 1999). Among adults, 2/3 of acute hepatitis B infections are found to be subclinical and these are unlikely to be detected unless population surveys are applied (McMahon et al., 1985). This suggests that the total number of transmissions during the outbreak could have been in the order of 1000 cases compared to an estimated IDU population in Funen of 1300 persons (Christensen, 2000). This indicates that the majority of the IDU population was infected during the outbreak and it is possible that the epidemic stopped not due to preventive measures, but because there was none left to infect. The ‘ceiling-effect’ as reason for the outbreak termination has to our knowledge not been described in other HBV outbreaks but is supported by a survey in 1996, where 70% of IDUs in the county drug treatment center were anti-HBc positive as sign of previous HBV exposure (Christensen, 2000).
140
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141
Cyclic outbreaks of hepatitis B among IDUs have been reported in several countries, and are supposed to reflect the gradual accumulation of susceptibles (Christenson, 1987a,b; Levine et al., 1994). Statistics of drug related deaths indicated that the Funen IDU population grew significantly quicker than in the rest of Denmark during the first part of the 1990s (Christensen, 2000). This is in agreement with previous reports where HBV outbreaks among IDUs have been preceded by an increase in the number of IDUs in the community (Lettau et al., 1987; Shattock et al., 1982). The sequence analysis of HBV showed that several index cases were involved and that two strains dominated. Although individual cases could be classified as related or unrelated to the outbreak, it was not possible to demonstrate individual chains of transmission, because of too little variation between the strains in the investigated region of the genome. The L143 mutation was not found among genotype D reported to Genbank, but it was previously found in other genotypes of HBV. There was no evidence of changed pattern of virulence of L143 compared to ayw3 based on duration or severity of clinical illness, and it is probably not a new mutation. It was found in an IDU who had been HBsAg positive for 18 years, and was also found in samples from outside Funen. The results indicate that, rather than being introduced from outside, the epidemic arose from the resident IDU population of chronic carriers. L143 has been reported from Spain where it was found among three IDUs, but has not previously been found among Scandinavian, British or Dutch IDU isolates. (Dr H. Norder, Dr C.G. Teo, Dr H.G.N. Niesters, personal communication) Recently L143 was described among dialysis patient in Germany indicating that the strain is not confined to IDU populations (Petzold et al., 1999). In conclusion, the hepatitis B outbreak among IDUs was associated with a significant morbidity and mortality. The current surveillance system was unable to identify the majority of cases and we suggest laboratory surveillance for hepatitis B to be implemented in Denmark.
Acknowledgements Kirsten Kronborg, Medical Officer of Health, County of Fynen provided the data of reported cases of acute hepatitis in the County of Funen. Sera from IDUs with acute hepatitis B outside Funen were provided by Mads Buhl, Department of Infectious Diseases, Marselisborg Hospital, Anders Siboni, Department of medicine, Esbjerg hospital and Svend E.H. Jacobsen, Department of Clinical Immunology, Sønderborg Hospital. Dr A. M. Courouce´ , Institute National de la Transfusion Sanguine, Paris performed serological subtyping of the two strains of HBV detected in the outbreak.
References Ackman DM, Birkhead G, Flynn M. Assessment of surveillance for meningococcal disease in New York State, 1991. Am J Epidemiol 1996;144:78 – 82. Anonymous. International Statistical Classification of Diseases and Health Related Problems. 8th revised ed. National Board of Health, Copenhagen, 1986. Anonymous. International Statistical Classification of Diseases and Health Related Problems. 10th ed. National Board of Health and Munksgaard, Copenhagen, 1993. Anonymous. Hepatitis B 1993. Epi-news, 50. Copenhagen, Statens Serum Institut, 1994a. Anonymous. Hepatitis B outbreak among injecting drug uses in County of Funen. Epi-news, 8. Copenhagen, Statens Serum Institut, 1994b. Anonymous, Hepatitis B. Epi-news, 41. Copenhagen, Statens Serum Institut, 1995. Christensen PB. Viral infections in Funen. University of Southern Denmark, Odense, 2000. Christenson B [Transmission of hepatitis B — a model for the evaluation of future HIV epidemic outbreaks](in Swedish). Lakartidningen 1987a, 84: 4037 – 4041. Christenson B. Epidemiology of hepatitis B in Sweden. J Inf 1987b;15:269 – 77. Dean AG, Dean JA, Columbier D, Brendel DA, Smith DC, Burton AH, Dicker RC, Sullivan K, Fagan RF, Arner TG. Epi Info Version 6.04: a word-processing, database, and statistics program for epidemiology on microcomputers. Atlanta, GA, 1997. Fuglsang T, Fouchard J, Ege P. Prevalence of HIV, hepatitis B, and C among drug users in the City of Copenhagen. Ugeskr Laeger 2000;162:3860 – 4. Helberg JC, Gaub J. [Hepatitis epidemic amongst intravenous drug-users in Esbjerg, July 1993 to May 1994] (in Danish). Ugeskr Laeger, 1995, 157/41: 5701 – 5703.
P.B. Christensen et al. / Journal of Clinical Virology 22 (2001) 133–141 Hook EB, Regal RR. The value of capture-recapture methods even for apparent exhaustive surveys. The need for adjustment for source of ascertainment intersection in attempted complete prevalence studies. Am J Epidemiol 1992;135:1060– 7. Lettau LA, McCarthy JG, Smith MH, Hadler SC, Morse LJ, Ukena T, Bessette R, Gurwitz A, Irvine WG, Fields HA, et al. Outbreak of severe hepatitis due to delta and hepatitis B viruses in parenteral drug abusers and their contacts. N Engl J Med 1987;317:1256 –62. Levine OS, Vlahov D, Nelson KE. Epidemiology of hepatitis B virus infections among injecting drug users: seroprevalence, risk factors, and viral interactions. Epidemiol Rev 1994;16:418 – 36. McMahon BJ, Alward WL, Hall DB, Heyward WL, Bender TR, Francis DP, Maynard JE. Acute hepatitis B virus infection: relation of age to the clinical expression of disease and subsequent development of the carrier state. J Infect Dis 1985;151:599 –603.1 Niesters HG, Honkoop P, Haagsma EB, de-Man RA, Schalm SW, Osterhaus AD. Identification of more than one mutation in the hepatitis B virus polymerase gene arising during prolonged lamivudine treatment. J Infect Dis 1998;177:1382 – 5.
.
141
Norder H, Courouce´ A-M, Magnius LO. Complete genomes, phylogenetic relatedness, and structural proteins of six strains of the Hepatitis B virus, four of which represent two new genotypes. Virology 1994;198:489 – 503. Norder H, Hammas B, Magnius LO. Typing of hepatitis B virus genomes by a simplified polymerase chain reaction. J Med Virol 1990;31:215 – 21. Pebody RG, Ruutu P, Nohynek H, Leinikki P. Changing epidemiology of hepatitis B infection in Finland. Scand J Infect Dis 1999;31:251 – 4. Petzold DR, Tautz B, Wolf F, Drescher J. Infection chains and evolution rates of hepatitis B virus in cardiac transplant recipients infected nosocomially. J Med Virol 1999;58:1 – 10. Shattock A, Kelly M, Fielding J, Arthurs Y. Epidemic Hepatitis B with delta-antigenaemia among Dublin drugabusers. Ir J Med Sci 1982;151:334 – 8. Yokosuka O, Tagawa M, Omata M. PCR Detection of Hepatitis B Virus. In: Persing DH, Smith TF, editors. Diagnostic Molecular Microbiology, Principles and Applications. American Society for Microbiology, Washington DC, 1993. p. 322 – 326.