In 1993, an unusual brain lesion was recognized in very pre- term infants cared for at National Women’s Hospital. The only report in the literature of similar lesions was from Birmingham (UK) describing encephaloclastic porencephaly (ECPE) in 15 preterm infants. 1 At the time of publication, ECPE was thought to be due to an as yet unrecognized post-natal event. The authors of that paper were contacted and stated that they thought that the lesion was linked to the condition of the baby at birth, early hypotension and chest physiotherapy. A case controlled study of the 13 affected babies at National Women’s Hospital revealed a statistical association between ECPE and non-vertex presentation, early hypotension and the number of chest physiotherapy treatments in weeks 2–4 after birth. 2 The present retrospective audit was designed to identify any cases of ECPE occurring in the 7 years prior to the first known case in 1992, to document its disappearance since 1995 and to investigate the temporal relationship between the availability and use of chest physiotherapy and the occurrence of ECPE. METHODS Cerebral ultrasound scan reports were located from radiology and clinical records of all babies with a birth weight ≤ 1500 g who were admitted to the neonatal unit from 1985 to 1991. In addition, data from the previous study of babies cared for in 1992–94 2 and data collected prospectively from 1995 onwards were included. For babies born before 1992, a neonatologist reviewed all ultrasound scan reports. If those reports described any parenchymal brain abnormality, the original films were retrieved and reviewed by a radiologist (RLT) unaware of the reports. If the original ultrasound films were unavailable, the clinical record was reviewed in detail. For babies who did not have late cerebral ultrasound scans, physiotherapy and follow- up records from the hospital’s Child Development Unit were searched. Ultrasound films of babies born between 1992 and 1994 had been reviewed previously. 2 From 1995 onwards, the original reports were used, because radiology staff were then familiar with the appearances of ECPE. Details of all chest physiotherapy treatment given were available from Physiotherapy Department records from 1985 to mid-1990. In addition, the clinical records of all babies < 30 weeks gestation and ≤ 1500 g born in the 12 month period from November 1993 to October 1994 were reviewed to docu- ment the chest physiotherapy actually given to each baby. This 12 month period was chosen because it was the time period during which ECPE occurred when chest physiotherapy was most readily available (staffing was sufficient to give each baby up to six treatments per day). Data for blood pressure (the lowest recorded mean pressure in the first 12 h, using indwelling arterial catheter readings J. Paediatr. Child Health (2001) 37, 554–558 Chest physiotherapy and porencephalic brain lesions in very preterm infants DB KNIGHT, CJ BEVAN, JE HARDING, RL TEELE, CA KUSCHEL, MR BATTIN and RSH ROWLEY National Women’s Hospital, Auckland, New Zealand Objective: National Women’s Hospital is one of two hospitals to report a destructive brain lesion, namely encephaloclastic porencephaly (ECPE), in extremely preterm infants. It has been associated with non-cephalic presentation, early hypo- tension and the number of chest physiotherapy treatments in the first month. The aim of the present study was to determine the temporal relationship between ECPE and chest physiotherapy use in very low-birth weight (VLBW) infants in our unit. Methodology: Cerebral ultrasound scan reports, post-mortem reports, clinical and physiotherapy records and, if indicated, original ultrasound films were reviewed for all VLBW babies admitted between 1985 and 1998. Results: Over the 14 year period in question, 2219 babies with a birth weight ≤ 1500 g were admitted. Encephaloclastic porencephaly was found in only the 13 previously reported babies born between 1992 and 1994. Encephaloclastic poren- cephaly was excluded in 1564 (70%) babies. In 621 (28%) babies who did not have late ultrasound scans, ECPE was thought to be unlikely either because the babies never had any chest physiotherapy (n = 479) or because they had chest physio- therapy but were known to be neurodevelopmentally normal on follow up (n = 142). Data were incomplete for 21 babies (0.9%). The number of chest physiotherapy treatments per baby decreased from a median of 95 prior to 1989 to 38 and the age of starting treatment increased from 5 to 8 days after 1990. The use of chest physiotherapy ceased in 1995. Conclusions: Encephaloclastic porencephaly emerged as a problem at a time when the use of chest physiotherapy had decreased. The cluster of cases seen between 1992 and 1994, although associated with the number of chest physiotherapy treatments given, began to appear because of some other factor. Key words: chest physiotherapy; porencephaly; preterm. Correspondence: Dr DB Knight, Newborn Services, National Women’s Hospital, Claude Road, Auckland, New Zealand. Fax: +64 9 6309 753; email: davidk@adhb.govt.nz Accepted for publication 29 March 2001.