Thrombectomy Widens: Mild Deficits, Late Window: Biplane Weekly No. 7

MILD-MT randomised 304 patients (300 in the modified intention-to-treat analysis) with anterior-circulation large-vessel occlusion, NIHSS below 6, a small core and a large perfusion-mismatch volume, within 24 hours of onset or last known well. Thrombectomy plus medical management was compared with medical management alone.

Excellent outcome (mRS 0–1 at 90 days) occurred in 69.4% with thrombectomy and 50.0% with medical management alone (OR 2.32, 95% CI 1.41–3.81). Symptomatic intracranial haemorrhage within 48 hours occurred in three patients in each group. Serious procedure-related complications were recorded in 19 patients (13.2%) in the thrombectomy arm.

This is randomised evidence in a population where practice has rested on equipoise. The trial is open-label with blinded outcome assessment and was run in Chinese centres. The gain is 19.4 percentage points in mRS 0–1; the price is procedural, with serious procedure-related complications in 13.2% of thrombectomy patients and no difference in symptomatic haemorrhage.

LATE-MT randomised 336 patients (168 per group) with anterior-circulation large-vessel occlusion and imaging-identified salvageable tissue who presented 24–72 hours after onset or last known well; median time to randomisation was 38.5 hours. Thrombectomy plus standard medical care was compared with standard medical care alone.

The 90-day mRS distribution favoured thrombectomy (common OR 1.74, 95% CI 1.19–2.55). The procedure carried a cost: symptomatic intracranial haemorrhage occurred in 4.8% with thrombectomy versus 0% with control, and 90-day case fatality was 6.0% versus 1.8%.

The design is open-label with blinded outcome assessment, in Chinese comprehensive stroke centres, and patients were chosen by imaging. The trial extends randomised evidence beyond the 6–24-hour window of DAWN, and the higher rates of symptomatic haemorrhage and 90-day death belong in any consent discussion alongside the shift in disability.

This updated cumulative meta-analysis included nine randomised trials (2,828 participants) of intra-arterial fibrinolytic given after successful or near-successful thrombectomy; the mRS 0–1 and mortality results each pool eight. Excellent recovery (mRS 0–1) was more frequent with adjunctive therapy (RR 1.27, 95% CI 1.14–1.41; prediction interval 1.05–1.53).

Mortality was not resolved. The pooled estimate was RR 1.04 (0.82–1.31), compared with RR 0.98 (0.80–1.21) before CHOICE-2 was added; in that trial, as cited by the authors, mortality was 12.1% with intra-arterial alteplase versus 6.4% with thrombectomy alone. In the authors' reverse-fragility analysis, changing the outcome of 24 patients would make the mortality estimate significant for harm (RR 1.17, 1.00–1.38); they state this does not establish excess mortality. Certainty for haemorrhagic outcomes was low.

The analysis follows the ANGEL-TNK one-year result covered on 6 October. The functional benefit is consistent across trials (I² 0%; fragility index 15), so the unresolved question is mortality, not efficacy.

In this prospective multicentre diagnostic study within the French OPTICH cohort, 277 adults (median age 71) with non-traumatic intracerebral haemorrhage had MRI with arterial spin labelling within seven days. Angiography was performed selectively; digital subtraction angiography confirmed an arteriovenous shunt in 24 (8.7%), of which 23 were AVMs and one a dural AVF.

Against the final adjudicated aetiology, ASL had sensitivity 95.8% (95% CI 79.8–99.3), specificity 94.5% (90.9–96.7) and a negative predictive value of 99.6% (98.1–99.9). Positive predictive value was 62.2%. Every shunt visible on initial angiography was detected, and three ASL-positive shunts were occult on the initial angiogram.

Because angiography was selective, the negative predictive value rests partly on adjudicated aetiology rather than on angiography in every patient, and the sensitivity interval is wide with only 24 shunts and no external validation. Within those limits, a negative ASL map could support deferring angiography in patients who resemble this cohort.

This meta-analysis pooled six studies (723 patients; 332 treated with drug-coated balloons, 391 with stent angioplasty) in intracranial atherosclerotic stenosis. Most were non-randomised.

Drug-coated balloons were associated with less restenosis (RR 0.28, 95% CI 0.18–0.44) and fewer recurrent ischaemic events (RR 0.30, 0.22–0.42), with no clear difference in periprocedural complications (RR 0.78, 0.39–1.55). Follow-up was short.

The estimates are large and consistent but come largely from comparative cohorts, where operators chose the device. The 16 October network meta-analysis favoured medical therapy over bare-metal stenting; this analysis does not answer whether coated balloons beat medical therapy.

Two retrospective comparisons of embolic agent in middle meningeal artery embolisation for chronic subdural haematoma appeared in the same week. The single-centre series (147 patients, 194 procedures) found repeat surgical evacuation in 15 of 118 procedures with Squid 12 (13%) and 3 of 76 with Magic Glue (4%); the unadjusted RR for glue was 0.31 (0.10–0.96), but glue was not independently associated with fewer reoperations on exploratory multivariable analysis, and haematomas were thicker at baseline in the Squid group (median 18 versus 12 mm).

The four-centre cohort (144 patients; 79 with n-hexyl cyanoacrylate, 65 with EVOH) recorded nine repeat evacuations in each arm (11.4% versus 13.8%; HR 0.68, 0.27–1.72; adjusted OR 0.92, 0.32–2.63).

Taken together, the first result does not survive adjustment and the second is compatible with no difference. The randomised MMAE trials used EVOH-type liquids, so choosing glue rests on experience, not trial data.

This systematic review pooled 14 studies (eight randomised, six observational) with 3,698 patients with cerebral venous thrombosis: 1,669 on a DOAC and 2,029 on a vitamin K antagonist.

No difference was detected. Recurrent venous thromboembolism was RR 0.92 (95% CI 0.55–1.53) overall, major bleeding RR 0.74 (0.44–1.25), and good functional outcome (mRS 0–2) RR 1.02 (0.93–1.11). In the randomised studies alone the intervals were wide, for example recurrence RR 1.06 (0.35–3.24) and mortality RR 1.15 (0.28–4.68), because events were few.

This extends the registry analysis covered on 24 October with pooled trial data. It is consistent with DOACs being a reasonable alternative, but the intervals do not exclude a meaningful difference either way.