| |
Perinatal Hepatitis C Virus Transmission: Role of Human Immunodeficiency Virus Infection and Injection Drug Use
|
| |
| |
Gregory L. Armstrong, Joseph F. Perz, and Miriam J. Alter
Division of Viral Hepatitis, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia
To the Editor: In a recent article in the Journal, Resti et al. concluded that current or former injection drug use (IDU) among women infected with hepatitis C virus (HCV) may substantially increase the risk of perinatal HCV
transmission. The study was notable in several respects, particularly in its large sample size (1372 mother-infant pairs) and in its extraordinarily high participation rates (100% enrollment and 92% retention). The findings of this study, performed from 1993 to 1996, appear to support those of a previous study, which was performed by the same authors at the same sites. We are concerned that the study populations may not have been independent.
Additional questions are raised by the results of the analysis in the article. All of the crude odds ratios presented, including those in tables 3 and 5, are either incorrect or incorrectly labeled, as are their 95% confidence intervals.
The
confidence limits (2.2%-7.2%) around the 7.1% overall transmission rate are also clearly incorrect. Together, these inaccuracies raise the question whether other statistics presented in the article are erroneous. The multivariable model
displayed in table 4 includes a nonsignificant risk factor (breast-feeding;
P = .74) that repeatedly has been shown to be not associated with HCV transmission, whereas a variable (the maternal HCV RNA status) that has been shown to be significantly associated with transmission is excluded. Other variables of interest such as maternal HCV viral load, use of internal fetal monitoring, signs of fetal distress (such as meconium staining), status of human immunodeficiency virus (HIV) antiretroviral therapy, and CD4 countwere either unavailable or not shown.
Some of the prior studies cited by Resti et al. in support of their conclusions are themselves inconclusive. For example, one study cited as not finding an association between HCV/HIV coinfection and perinatal transmission of
HCV included only 15 coinfected pregnant women, all of whom were on antiretroviral therapy. Results from another study that reportedly showed no difference, in HCV transmission rates, between current and past IDU found
that the risk of perinatal transmission was 5% (2 of 40) for current IDU and 0% (0 of 7) for past IDU (relative risk undefined).
Given these issues, we are concerned by the conclusion that perinatal HCV transmission is independently associated with IDU and not with HIV infection. An association with current IDU might be plausible, but we question why
transmission would be associated with a history of IDU among women who do not currently engage in IDU. Resti et al. suggest infection of peripheral blood mononuclear cells as a mechanism for increased transmission, but this is
unlikely to explain an association with past IDU. Furthermore, when the results of prior studies are examined in aggregate, they show transmission to be more strongly associated with HIV coinfection than with IDU. It seems more plausible that, among the mother-infant pairs in this study, IDU is a confounder for another, unmeasured risk factor for transmission.
|
|
| |
| |
|
|
|
