July 30, 2012 — Harvard School of Public Health (HSPH) researchers have found an effective and cost-efficient strategy for monitoring the health and development of children whose HIV-infected mothers took antiretroviral (ARV) drugs during pregnancy, and for estimating the health risks of their in utero exposure to these drugs.
The researchers found that keeping an eye on “trigger” events—certain development, motor, or cognitive problems—was a highly effective way to find those children who may have adverse effects related to ARVs. By analyzing data in a large, long-term study of children born to HIV-infected mothers, known as the SMARTT study, the researchers found that using a trigger-based design made it easier to find children who might have serious problems and focus on understanding their problems.
The study was published April 6, 2012 in the American Journal of Epidemiology. Lead author was Paige Williams, HSPH senior lecturer on biostatistics and senior statistician for SMARTT.
Scientists are concerned that exposure to ARVs in the womb may lead to problems in children such as impaired growth, higher-than-average body mass index (BMI), seizures, hearing problems, or cognitive or motor delays. But research on the subject has been inconclusive.
A streamlined study design
The five-year-old SMARTT (Surveillance Monitoring of ART Toxicities) study is part of the Pediatric HIV/AIDS Cohort Study, which was established in 2005 to both evaluate the long-term safety of fetal and infant exposure to ARV therapy as well as the effects of perinatally acquired HIV infection in adolescents. The trigger-based design was chosen for SMARTT because confirming any potential adverse reaction to ARVs is costly and may require either a biopsy or other invasive tests, and this type of design meant not having to conduct these types of invasive tests on every enrolled child.
“This design helps keep the study visits shorter, which helps keep the children involved—almost 98% have stayed in the study so far,” said Williams. “And it makes the study less expensive.”
Using this method, Williams and her colleagues studied data on nearly 2,300 children up to age 12, all born to HIV-infected mothers, who had enrolled in SMARTT as of January 2011. The researchers found that problems with metabolism, language, and hearing are common among young children born to mothers infected with HIV. Of the roughly 2,200 children who participated in at least one study visit, 900 (41%) met at least one trigger. For example, 365 (27%) had high BMI—they weighed a lot relative to their height. These 365 children were given blood tests; results showed that 45% had high cholesterol or other metabolic problems. Given this data, the researchers were then able to estimate that 12.1% of all SMARTT children had metabolic problems.
From a statistical perspective, the trigger-based design also has two important benefits, Williams said. First, it yields more precise estimates of the percent of children with metabolic problems than giving blood tests to 365 children chosen randomly from the full cohort of 2,200 children. Secondly, it has high power for identifying associations with ARV exposures.
“Ethical mandate” for long-term monitoring
Williams said it’s important to continue to follow the children through adulthood and their own childbearing years to learn more about the long-term effects of ARVs on children, noting that in the past, some drugs given to pregnant women—like DES—were later found to be harmful to their children. DES was a synthetic estrogen prescribed to women from 1938 to 1971 in the mistaken belief that it would help prevent miscarriages and other pregnancy problems, but which was found in 1971 to cause cancer in girls and women who’d been exposed to the drug in utero.
“Some of these ARVs cause mutations or cancer in animal studies and that makes us a bit worried about them,” said Williams. “We know that most of these drugs cross through the placenta, so the babies are getting exposed. We have an ethical mandate to continue to follow these babies.”
She added, “In the United States, because of ARVs, hardly any children are born with HIV any more. Over the past 15 years, the HIV transmission rate from mother to children has been cut to less than 1%. We know this is a huge benefit, but there are so many different antiretroviral drugs being used and there is a lot that is unknown about their effect on children if their mothers take them during pregnancy.”
Knowing more about the relative health risks to children of different types of ARVs could help clinicians choose the safest ones to prescribe to HIV-infected pregnant women, Williams said.
Other HSPH authors of the study included George Seage, professor of epidemiology; Raymond Griner, statistical programmer, Center for Biostatistics in AIDS Research (CBAR); Katherine Tassiopoulos, research scientist, Department of Epidemiology; Cenk Yildirim, statistical programmer, CBAR; Yanling Huo, biostatistician, CBAR; and Denise L. Jacobson, senior research scientist, Department of Biostatistics.
— Karen Feldscher
photo: Aubrey LaMedica