In order to determine whether CPO2 measurements on the anterior chest of infants weighing <1,000 g accurately reflect PaO2, 7 infants with birth weights 670 g to 990 g (BW 835 ± 120 g) and gestational age 25-28 wk (EGA 26 ± 1.5 wk) were studied at mean postnatal age 9 ± 5.5 days with two cPO2 sensors (44°C, Litton) placed simultaneously on upper chest and abdomen. In all 7, cPO2 abdomen (CPO2-A) correlated well with umbil- ical artery PaO2 (r = 0.947); in 5 of the 7, cPO2-A was greater than CPO2 chest (CPO2-C) (mean difference = 50% ± 13%). Simultaneous measure- ments of right upper arm and abdom- inal cPO2 showed no difference. All infants were normotensive; however, CPO2-C tracings in the 5 infants with CPO2-A > CPO2-C showed evidence of peripheral hypoperfusion, ie, cPO2 fell towards zero after application and failed to rise, and frequent marked drops of CPO2-C were seen unrelated to changes in cPO2-A. In 2 infants (925 g, 770 g) who showed these pat- terns, simultaneous measurements of skin blood flow by relative heating energy indicated poor perfusion under the chest electrode. When 2 infants were later monitored after a 400-500 g increase from birth weight, the chest- abdomen CPO2 differences were ab- sent. It therefore appears that inade- quate local perfusion of the anterior chest skin makes cPO2 approximation of PaO2 subject to underestimation in the VLBW infant. This could increase the risk of inappropriate elevation of FIO2 and the risk of retrolental fibro- plasia.
