Background
With more than 1.4 million women estimated to give birth annually and 8.2% HIV prevalence rate at antenatal clinics (ANC), approximately 122,000 HIV-infected women deliver annually in Tanzania. Assuming a 40% transmission rate in the absence of any intervention, an estimated 48,000 children will become HIV-infected each year. In 2003, the number of children (0–15 years) living with HIV/AIDS in Tanzania was estimated to be between 85,000 and 230,000[
1,
2]. A large number of infected children have been associated with significant increase in morbidity, hospitalization and mortality [
2].
With increased access to antibiotics and antifungal agents hepatitis viruses, especially hepatitis B and C, are emerging as the leading causes of morbidity and mortality among children on ART [
3].
Relatively little is known regarding HCV or HBV co-infection in HIV infected children in Tanzania. Thus, not surprisingly the current national guidelines for management of HIV disease in children do not include screening and management of hepatitis viral co-infection.
We conducted this study to determine the seroprevalence and risk factors for HBV and HCV among HIV infected children aged 18 months to 17 years attending the paediatric HIV Care and Treatment Center (CTC) at MNH, Dar-es-Salaam, Tanzania.
Results
A total of 167 children were recruited between April and august 2005. Among them, 88(52.7%) were males and 79(47.3%) were females. Most of the children (49.7%) were aged 6–10 years. The overall prevalence of hepatitis viral co-infection among the HIV infected children was about 15% (25/167). The prevalence of hepatitis viral co-infection was significantly higher among girls (21.5%) than boys (9.0%). There was no association between age and hepatitis viral co-infection. Two children (1.2%) were co-infected with HBsAg and 23(13.8%) were co-infected with HCV (Table
1). The prevalence of HCV was significantly higher among girls (20.3%) than boys (8%) (P = 0.02) (Table
1). However there were no sex differences in the occurrence of HBsAg P = 0.93 (Table
1). The occurrence of HCV was highest (28.5%) in children aged less than two years and in children older than 10 years (Table
1). The same table shows that HBsAg were only seen in two children aged 6–10 years. However the age differences were not statistically significant, with P-values for HBsAg and HCV being 0.72 and 0.69, respectively (Table
1).
Table 1
Seroprevalence of HBsAg and HCV infection by age and sex among HIV infected children
Male | 88 | 1(1.1) | 7(8.0) |
Female | 79 | 1(1.3) | 16(20.3) |
| |
P = 0.93
|
P = 0.02
|
Age group (yrs)
| | | |
<2 | 7 | 0 (0) | 2 (28.5) |
2–5 | 37 | 0 (0) | 4 (10.8) |
6–10 | 83 | 2 (2.4) | 10 (12.0) |
11–15 | 34 | 0 (0) | 6 (17.6) |
>15 | 6 | 0 (0) | 1 (16.7) |
| |
P = 0.72
|
P = 0.69
|
Total
|
167
|
2 (1.2)
|
23 (13.8)
|
There were 148 (88.6%) children with history of injection, 26 (15.6%) had blood transfusion, 48 (28.7%) had uvulectomy and 2 (1.2%) were sexually abused. Among those with history of injection 1(0.7%) had HBsAg co-infection and 19(12.8%) had HCV. However, there was no association between these risk factors and the occurrence of hepatitis viral co-infection (Table
2).
Table 2
Association between risk factors and HBsAg and HCV co-infection in HIV infected children.
Injection
| | | |
Yes | 148 | 1(7) | 19(12.8) |
No | 19 | 1(5.3) | 4(21.1) |
| |
*P = 0.21
|
*P = 0.30
|
Blood transfusion
| | | |
Yes | 26 | 0(0) | 5(19.2) |
No | 141 | 2(1.4) | 18(12.8) |
| |
*P = 1.00
|
P = 0.36
|
Uvulectomy
| | | |
Yes | 48 | 1 (2.1) | 9 (18.8) |
No | 119 | 1 (0.8) | 14 (11.8) |
| |
*P = 0.49
|
P = 0.32
|
Sexual abuse
| | | |
Yes | 2 | 0 (0) | 0 (0) |
No | 165 | 2 (1.2) | 23 (13.8) |
|
*P = 1.00
| |
*P = 1.00
|
Among 167 HIV infected children 59 (35.3%) had no evidence of immunosuppression, 59 (35.3%) had moderate immunosuppression and 49 (29.3%) had severe immunosuppression. There was no statistically significant association between immune status and hepatitis viral co-infection.
None of the investigated children had HBV and HCV dual infection.
Among 167 children, 15 (9%) had elevated ALT. The prevalence of elevated ALT among children co-infected with hepatitis B or C virus was 20.0% as compared with 7.0% among those without viral co-infection. Using logistic regression analysis, children co-infected with hepatitis had 4 times more risk of elevated ALT than those without hepatitis (OR 3.99, 95% CI 1.1–14.2 (Table
3).
Table 3
Association between elevated ALT, immune status, ART usage and hepatitis viral co-infection expressed as odds ratios with 95% confidence intervals
Immunosuppression
| | | | | |
No evidence
| | 1.0 | - | 1.0 | |
Moderate
| | 0.74 | 0.16–3.45(P = 0.69) | 0.61 | 0.12–2.96(P = 0.54) |
Severe
| | 2.68 | 0.79–9.52(P = 0.13) | 2.48 | 0.68–9.02(P = 0.94) |
ART
| | | | | |
Not on ART | 3(7.1) | 1.0 | - | 1.0 | |
On ART | 12(9.6) | 1.38 | 0.37–5.15(P = 0.09 | 1.95 | 0.24–3.77(P = 0.94) |
HBsAg
| | | | | |
No | 14(93.3) | 1.0 | - | | |
Yes | 1(6.7) | 10.79 | 0.74–8.81(p = 0.14) | | |
HCV
| | | | | |
No | 11(73.3) | 1.0 | - | 0.1 | |
Yes | 4(26.7) | 2.55 | 1.02–10.65(P = 0.04) | 2.83 | 0.75–10.62(P0.12) |
Hepatitis B/C viral co-infection
| | | | | |
No | 0(0.0) | 1.0 | | 1.0 | |
Yes | 5(20.0) | 3.25 | | 3.99 | 1.12–14.19(P = 0.03) |
One hundred and twenty five (74.9%) children were on ART and among them 12(9.6%) had elevated ALT compared with 7.1% among those who were not on ART (Table
3). However, there was no statistically significant association between elevated ALT and being on ART (P = 0.94) (Table
3).
Discussion
This study showed a high prevalence (15%) of hepatitis co-infection, which is approximately one out seven HIV-infected children. The prevalence of HCV alone observed in this study was 13.8%, which is significantly higher than 7.1%, which was reported by Kitundu
et al (Jesse Kitundu – Post-transfusion hepatitis C seroprevalence in Tanzanian children. 2001) [
7] in post-transfused Tanzanian children. The difference between these studies could be due the nature of the studied populations. In the present study, all children were HIV infected and were therefore more prone to hepatitis viruses, which share modes of transmission. Secondly, the methods used were different, in this study antibodies to HCV was detected using an ELISA technique while in this study agglutination test was used which is relatively less sensitive and less specific due to potential cross reactions [
8]. However, it is important that these ELISA based results are confirmed with more sensitive technique such as HCV-RNA by PCR, since it has been shown that some positive ELISA results becomes negative when confirmed by PCR [
9].
The seroprevalence of HBsAg alone in this study was 1.2%, which is significantly lower than 12% observed by Kitundu
et al (Prevalence of Hepatitis B and C among children transfused with anti-HIV negative donor blood at the same hospital [unpublished observation]. A similar study conducted in Kenya by Rana
et al [
10] found a prevalence of hepatitis B to be 4%, among African children infected with HIV.
In this study the prevalence of HCV/HBV was not associated with age (18 months to 17 years). One limitation of this study was that children below 18 months were not investigated due to methodological limitations associated with the use of IgG based ELISA[
9]. This limitation may, at least in part, have obscured age trends.
Regarding HCV, females were more affected 17(21.5%) than males 8(9%) a finding that is consistent with a study done by Kitundu et al (Prevalence of Hepatitis B and C among children transfused with anti-HIV negative donor blood at MNH hospital in 2000) [unpublished data] in the same setting but the reason for the differences in sex is unclear.
In this study, 148 (88.6%) children had history of injection, 26 (15.6%) had Blood transfusion, 48 (28.7%) had uvulectomy and 2 (1.2%) were sexually abused. However none of these potential risk factors was associated with occurrence of the hepatitis viruses, which in another study conducted in the same hospital by Kitundu et al (Prevalence of Hepatitis B and C among children transfused with anti-HIV negative donor blood at MNH hospital [unpublished data]. This finding coupled with the young age of infection, may indicate that possibly these children acquired these viral infections vertically rather than horizontally.
We found no association between HBV and HCV, a finding that is in keeping with observations by Matee
et al [
8] and Wadell
et al [
11] among blood donors in the same hospital. Collectively these studies indicate that the epidemiology of these viruses is different in our setting [
8].
The higher prevalence of raised ALT (20.0%) was among children co-infected with hepatitis virus compared with 7.0% without hepatitis co-infection (OR 3.99, 95% CI 1.12–14.19) (Table
3) is expected since hepatitis viruses are known to elevate liver enzyme [
12].
We found no association between ART and elevated ALT, which is contrary to the findings of other studies [
13‐
15]. Children recruited in this study were on first line regimen, which does not contain protease inhibitors, which are associated with more risk of elevated ALT than other antiretrovirals [
15]. There has also been reports that different genotype of hepatitis C virus differ in their ability to cause hepatotoxicity, with genotype 3 being more hepatotoxic than genotypes 1, 2, and 4. It would be interesting to perform genotyping of the HCV strains to better understand the association with liver pathology.
The high prevalence of hepatitis co-infection among HIV infected children indicates the need to revisit the current guidelines for the clinical management of HIV and AIDS. The regimen (zidovudine, lamivudine, and nevirapine for children < 3 years, with efaviranz replacing nevirapine for children > 3 years and didanosine, abacavir and ritonavir boosted lopinavir as second line) does not take care of the hepatitis co-infection. Although lamivudine works for both HBV and HIV, prolonged lamivudine therapy can result in drug-resistant HBV mutants and has been associated with hepatitis flares [
16,
17].
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SPT designed the study and supervised interviews, clinical examination as well as data collection. MIM helped in design of the study and supervised laboratory work. EKM participated in design of the study and supervised clinical work. Finally, all authors participated in preparation of the manuscript read and approved the final manuscript.