Premature birth occurs to more than half a million babies in the United States each year, often for reasons not yet understood. The earlier in pregnancy a baby is born, the more health problems it is likely to have. Babies who are born extremely preterm, prior to 28 weeks gestation, are at high risk for breathing problems, brain damage, and digestive problems in the first few days of life, but they are also at risk for a wide range of lifelong health challenges in the form of cerebral palsy, hearing and vision loss, and learning and developmental delays. There is also mounting evidence that babies born more than three months early face a much higher risk of developing autism compared to babies born full-term.

Early last year, researchers from Harvard Medical School and McGill University completed a study involving 91 children where they found that one in four very low birth weight premature infants showed early signs of autism when evaluated before the age of 2.

In June, researchers at the Centers for Disease Control and Prevention released findings of a study comparing 565 children with autism, born in Atlanta between 1986 and 1993, to children without autism. They found that baby girls born under 5.5 pounds had 3.5 times the risk of developing autism than children of normal births. The risk of developing autism increased to 5.4 times for baby girls born more than seven week early. However, neither factor seemed to reflect a significant difference in the risk factor for male babies.

Now, a much larger study shows that one in five children born more than three months prematurely exhibited early signs of autism by age 2, regardless of birth weight. For the latest study, researchers from Boston University and Wake Forest University in Winston-Salem, North Carolina followed 988 U.S. children born between 2002 and 2004 who participated in the ELGAN (Extremely Low Gestational Age Newborn) study, a large, multi-center study that enrolled more than 1,500 infants born at least three months before their due date. At age 2, the toddlers were evaluated using a screening tool referred to as M-CHAT (Modified Checklist for Autism in Toddlers), a questionnaire for caregivers on 23 behaviors. Checking any three items or two of six critical items as "unable to perform" leads to a positive score for autism spectrum disorder (ASD).

Using this tool, the researchers found that 21 percent of the children scored positive for possible ASD, compared to 5.7 percent of children in the general population. Even after excluding children with motor, vision, and hearing impairments, 16 percent still tested positive for autism risk. When additionally excluding children with cognitive impairment, the rate was still about 10 percent. "Children who had no motor, hearing, visual, or cognitive impairments still had double the risk of being M-CHAT positive," said Dr. Karl Kuban, chief of pediatric neurology at Boston Medical Center, who led the study.

Dr. Kuban noted that M-CHAT is not considered a definitive diagnosis, but is used to flag children who might have autism. Since a formal diagnosis typically does not occur until around age 3, further follow-up will be needed to determine how many of these children will ultimately go on to be diagnosed with autism. If the rate is verified as only minimally or moderately increased among this group, "then the M-CHAT might require modification for use in children born at extremely low gestational ages and for children who have physical and special sensory impairment," the researcher noted. However, on the other hand, "if it turns out these children do go on to have autism much more often than we would expect, it does beg the question of why that should be and it offers us an avenue of study," Kuban said, noting that early identification and educational support can improve outcomes for those found to have autism.

In a related editorial, Dr. Samantha Johnson and Dr. Neil Marlow, both of the University College London, stress that because early identification of children with ASD leads to early treatment, screening tests are designed to over-identify those at risk. By following the children as they mature and seeing how many who initially screened positive actually develop ASD, they suggest that useful knowledge may be gained. Dr. Marlow notes that the study is valuable because "it raises our awareness of the difficulties in interpreting screening results," but cautions that further research is needed before drawing conclusions about the direct correlation between preterm birth and ASD.

According to government figures, between one in 500 and one in 150 U.S. children have an autism spectrum disorder. These developmental problems include difficulty with communication and social interaction, repetitive behaviors or obsessive interests and can range from mild to disabling.

The cause of autism continues to be a subject of great debate. Some still feel that autism is caused by the mercury-based preservatives used in vaccines, though numerous studies have dispelled the theory. Many experts say that autism can be caused by a range of factors, both genetic and environmental, but it remains unclear which factor plays a bigger role in the onset of the disorder.

The study appears in the Journal of Pediatrics.