Sir,Belfort et al. 1 conclude in their paper that headache in women with pre-eclampsia is strongly associated with the presence of abnormal cerebral perfusion pressure (CPP). They suggest also that this information may be of use in clinical management. Unfortunately, their paper contains signi®cant errors, which do not allow the above conclusions to be drawn.One of the main results of the study by Belfort et al. 1 is the ®nding of more frequent abnormal CPP in pre-eclamptic women with headache (34/42; 88%) than in those without headache (18/ 37; 49%). The abnormal ± low or high ± CPP values were de®ned as being above the 95% or below 5% limits of the normative data, respectively. The normative CPP data are based on a longitudinal study in 63 normal pregnant women examined by the same research group. The normal data have obviously not been published in an original paper, as the authors refer to two abstracts of presentations at scienti®c meetings 2,3 . The abstract by Belfort et al. 2 gives only the mean values of CPP on gestational age. The abstract by West et al. 3 does not give full background data on the group of normal pregnant women, however, a plot of all CPP values and the predicted means and 95% con®dence intervals are presented in a ®gure. The discrepancy between the plotted individual CPP values in the abstract 3 and the normative data published in the British Journal of Obstetrics and Gynaecology 1 (Fig. 4) shows clearly that the authors have erroneously considered the con®dence limits of the estimated mean values to be the 5 th and 95 th percentiles of the normal population CPP values.According to the referred abstract 3 , 43 normal pregnant women were examined at 36 weeks of gestation, i.e., during the period of gestation in which the majority of pre-eclamptic women of the present paper were studied. For the normal women, a mean CPP value of 72.1 mm Hg was found with standard error of mean (SE) 2.9 mm Hg. This corresponds to one standard deviation (SD) of 19.02 mm Hg, mean 2 2 SD of 34.1 mm Hg and mean 1 2 SD of 110.1 mm Hg. For 16 weeks of gestation, following data are given: n 43, mean CPP 47.6 mm Hg and SE 3.0, giving SD of 19.67 mm Hg, mean 1 2 SD 86.9 mm Hg and mean 2 2 SD 8.3 mm Hg. Considering these normal limits, according to Fig. 4 in the paper, no CPP values of the pre-eclamptic women without headache had abnormal CPP (0/37; 0%). Of the 42 pre-eclamptic women with headache, approximately 4 had CPP values above the upper normal limit and one below the lower normal limit. (The numbers were estimated from Fig. 4.) Thus, in none of the pre-eclamptic women without headache abnormal CPP was found, which is not signi®-cantly different from the ®nding of abnormal CPP in 12% of preeclamptic women with headache (OR 0; 95% CI 0 to 1.22; P 0.12).Belfort et al. 1 did not ®nd any signi®cant difference between the CPP values of pre-eclamptic women with and without headache ( Table 2). This is not surprising considering the overlap of the CPP values shown in Fig. 4. By calculating the multiples of standar...