From May 1992 to February 1993, 22 cases of hypertensive putaminal haemorrhage (HPH) treated at our hospital were serially measured with transcranial Doppler (TCD) sonography. Among them, 13 patients underwent surgical intervention (3 stereotaxic surgery and 10 craniotomies), and 9 were conservatively treated. Most of the patients of the two operative groups had larger haematomas and developed clinical and/or neurological deterioration, which was the indication for subsequent surgery. Therefore the groups represent different clinical and physiological entities. On admission, the peak MCA velocities (Vs) in the surgical group (stereotaxic and craniotomy) were significantly lower than those in the conservative group (mean +/- S.E.M.: 38.33 +/- 4.26 and 42.00 +/- 2.62 cm/sec vs. 57.22 +/- 3.23 cm/sec; p < 0.005, respectively). The surgical group also had significantly lower diastolic (Vd) and mean (Vm) velocities than those of the conservative group (p < 0.001). Rather, the admission pulsatility indices (PI = (Vs-Vd)/Vm) in the surgical group were significantly higher than those of the conservative group (mean +/- S.E.M.: 1.42 +/- 0.04 and 1.31 +/- 0.09 vs. 0.95 +/- 0.01; p < 0.005, respectively). Time course velocity curves reached a peak around the 3rd hospital day in all the 3 groups. The Glasgow coma scale (GCS) scores positively correlated with the mean MCA velocities (n = 22; r = 0.63, p < 0.005; y = 2.04 x + 8.74), but negatively with PI values on admission (n = 22; r = -0.53, p < 0.05; y = 1.68-0.053 x). On the 7th hospital day, 2 patients with peak MCA velocities below 50 cm/sec had an unfavourable outcome. All the 3 patients in the stereotaxic group had higher peripheral resistance, as compared with those in conservative craniotomy groups (mean +/- S.E.M.: 1.28 +/- 0.13 vs. 0.99 +/- 0.07 and 0.87 +/- 0.06; p < 0.05, respectively). Our study supports TCD as a safe and valid monitoring method in patients with HPH. "Compromised cerebral haemodynamic status" (Vs < 50 cm/sec, Vd < 15 cm/sec, Vm < 25 cm/sec, PI > 1.15) may offer an aid in the decision for surgical intervention in HPH. Postoperatively, patients who made a favourable recovery had a significant increment in the MCA velocities in contrast to those severely disabled, whose MCA velocities remained low.