Immediate- vs Delayed-Intensive Statin in Acute Cerebral Ischemia: INSPIRES

What this is

INSPIRES tested the simplest possible question in acute atherothrombotic stroke care: when you have decided to give a high-intensity statin, does the first dose have to land inside the 72-hour window, or can it wait three days? Across 6100 randomized Chinese patients with mild stroke or high-risk TIA, immediate atorvastatin 80 mg matched a 3-day-delayed start on the only outcome that mattered prespecified — recurrent stroke at 90 days — and matched it precisely (8.1% vs 8.4%; HR 0.95). The 90-day functional-outcome signal (OR 0.83) is interesting but cannot be acted on after a negative primary. The headline is the absence of harm and the absence of incremental efficacy from “rushing the statin.”

1. Shadow Audit

The framing of the paper is “we tested whether immediate is better than delayed.” Look at the design: both arms ended up on atorvastatin 40 mg daily from day 22-90, so what was actually randomized was a 3-day head start on 80 mg — plus 17 extra days of 80 mg vs 40 mg. The “delayed” arm still got intensive statin; it just got a placebo first. That makes this a comparison of loading-dose intensity in the first 3 weeks, not of statin vs no statin. The trial cannot speak to whether early intensive statin helps compared with no statin at all in the acute window — only whether adding an immediate high-dose push on top of an already-planned statin prescription changes 90-day outcomes. The number of LDL-C measurements is also under-reported in the abstract; if the immediate arm achieved meaningfully lower day-7 LDL-C, the negative primary would imply that LDL trajectory in the first 3 weeks simply does not drive 90-day recurrence in this population.

2. Inversion Engine

For the opposite conclusion — that immediate-intensive statin does reduce recurrent stroke at 90 days — to be true, the loading-dose intensity would need to capture a pleiotropic or plaque-stabilizing effect that is large enough to shift an 8% recurrence rate. The point estimate is HR 0.95, with the 95% CI excluding any effect better than a 20% relative reduction. To overturn this with a new trial of similar size, one would need either (a) a population with much higher baseline LDL-C and lower background statin use, where the absolute LDL delta between arms would be larger, or (b) a much higher event rate to power smaller effects. A trial in untreated LDL-C > 190 mg/dL patients with 15% recurrence would have detectable a 25% relative reduction at 80% power with ~5000 patients; the INSPIRES cohort, with median LDL in a Chinese secondary-prevention population already partly treated, cannot.

3. Second-Order Catalyst

The protocol and stroke community already default to early high-intensity statin (within days, often within 24 h) for ischemic stroke of presumed atherosclerotic origin. INSPIRES will not change that default — and arguably reinforces it (delayed arm did fine). The protocol-level change at large comprehensive stroke centers is minimal: keep the existing in-patient loading order, but stop racing to give the first dose in the ED if the patient is admitted overnight. The subgroup signal that might matter — statin × diabetes interaction favoring immediate treatment in non-diabetics on functional outcome — is exactly the kind of exploratory finding that a follow-up trial in non-diabetic atherosclerotic stroke could probe, but no such trial is currently registered. The more important downstream effect is on the antiplatelet arm of the same 2 × 2 factorial: that result (clopidogrel+aspirin vs aspirin alone) will determine whether INSPIRES shifts any practice at all.

4. Asymmetric Leverage

This is a small-effect, large-denominator result: 6100 patients, 501 primary events, and a hazard ratio indistinguishable from 1. There is no asymmetric leverage here — even if true, the effect size is below the threshold most clinicians would notice in practice. The genuinely asymmetric finding is on the secondary endpoint: 49 fewer poor-outcome patients per 3050 (an absolute difference of 1.6 percentage points; NNT ~62) with a credible mechanism (pleiotropic neuroprotection, faster plaque stabilization), but exactly the wrong epistemic status to act on. The lever is upside-asymmetric only because of the functional-outcome signal, and that asymmetry cuts both ways: it could be real benefit, or it could be the play of chance on a negative primary.

5. Paradigm Destroyer

The reflex this paper kills: “give atorvastatin 80 mg stat in the ED for every TIA/stroke admission.” That reflex has been widespread for a decade based on SPARCL, TST, and observational data, but it always assumed an efficacy gradient across the first 72 hours that INSPIRES did not find. The two-sentence bedside update: continue high-intensity statin in patients with mild ischemic stroke or high-risk TIA of atherosclerotic origin; the first dose within 24 hours versus 72 hours does not change 90-day stroke risk, and is no longer a time-sensitive quality metric.

MVP — Minimum Viable Proof

The minimum statement that, if true, would change bedside practice: In an adult with mild ischemic stroke or high-risk TIA of presumed atherosclerotic origin, the first dose of high-intensity statin given at hour 4 produces the same 90-day stroke recurrence as the first dose given at hour 72. INSPIRES comes close to but does not strictly meet this MVP because the trial’s “delayed” arm still gave placebo rather than no treatment; the equivalence is between “loading-dose intensity” arms, not between a true early-vs-late comparison. The MVP still passes in the relevant clinical sense: a statin-naïve patient can wait until the next inpatient medication round without measurable harm.

Best Combination

INSPIRES pairs naturally with (a) SPARCL (atorvastatin 80 mg vs placebo in stroke/TIA, 2006), which established the high-intensity-vs-nothing benefit; (b) TST (Treat Stroke to Target, 2020), which established that LDL-C targets matter more than statin dose per se; and (c) the upcoming INSPIRES antiplatelet arm, which will determine whether the dual-pathway approach (intensive statin + intensive antiplatelet) is where the residual benefit lies. The combined reading: SPARCL showed statins work; TST showed targets matter; INSPIRES shows the first 3 days don’t. The remaining lever is LDL-C trajectory over months, not statin timing over hours.

Overvalue Warning

Two specific things readers might overinterpret from INSPIRES. First, that the secondary functional-outcome OR of 0.83 is actionable — it is not, given the negative primary and the explicit author statement that it must be viewed as hypothesis-generating. Second, that “delayed intensive statin” is equivalent to “no acute statin” — it is not; both arms received high-intensity atorvastatin within 21 days, so the trial does not license withholding statin in the acute window, only deprioritizing the first-dose timing.