Objectives: To determine whether a characteristic form of brain damage (encephaloclastic porencephaly) was associated with chest physiotherapy treatment in preterm babies. Methods: A retrospective case-control study was undertaken among 454 infants of birth weight less than 1500 gm cared for during the 3-year period of 1992 to 1994. Thirteen babies of 24 to 27 weeks of gestation who weighed 680 to 1090 gm at birth had encephaloclastic porencephaly. Twenty-six control subjects were matched for birth weight and gestation. Results: The patients received two to three times as many treatments with chest physiotherapy in the second, third, and fourth weeks of life as did control infants (median 79 vs 19 treatments in the first 4 weeks, p < 0.001). Patients also had more prolonged and severe hypotension in the first week than did control subjects (median duration of hypotension 4 vs 0.5 days, p < 0.01), and were less likely to have a cephalic presentation (31% vs 81%, p < 0.01). Since December 1994 no very low birth weight baby has received chest physiotherapy treatment in the first month of life in our nursery, and no further cases have occurred. Conclusions: Encephaloclastic porencephaly may be a previously unrecognized complication of chest physiotherapy in vulnerable extremely preterm infants. (J Pediatr 1998;132:440-4)