Renal involvement in henoch-schonlein purpura.

non-ipregnant cervix in vitro have a tendency to contract in response to PGF2,, and to relax with PGE2.9 It is therefore possible that PGE2 is less likely to cause or allow such damage to occur, though it must be said that these experiments have not been performed using strips of pregnancy cervix in vitro, nor have such observations been made in vivo. That cervicovaginal fistulae have also been reported following abortion induced with intra-amniotic hypertonic saline'0 11 and intra-amniotic hypertonic ureal2 suggests that the primary abortifacient agent is not a contributory factor. The relationship of route of administration and type of prostaglandin used is therefore equivocal. Sinilarly the incidence of fistulae does not appear to be related to short abortion times'-4 and by inference excessive uterine activity per se. Before this can be conclusively stated, however, individual intrauterine pressure records will require critical analysis. The problem probably lies in developing levels of uterine contractility in the presence of an inadequately dilating cervix. The existing hormone balance may be of importance. It has been shown in sheep at termn that induced uterine contractility does not produce cervical dilatation if progesterone levels are maintained." In prostaglandin and saline terminations in the humnan the progesterone level declines during the contractile phase, though in some cases the significant decline does not occur until shortly before abortion." It is possilble therefore that if hormonal balance does not favour cervical effacement and dilatation in some individuals cervical fistulae may result.-We are, etc.,