What this is

The ETLAS-2 trial tested whether a daily 20 mg dose of the PDE-5 inhibitor tadalafil could be tolerated by patients with cerebral small vessel disease (CSVD) and whether it would slow white matter damage or improve cognition over three months. It is a small (n = 76), rigorous phase II trial from Copenhagen — and it is, on its primary endpoint, a negative feasibility study with a near-significant mechanistic signal buried inside.

1. Shadow Audit

The most underweighted number in this trial is the 46% female discontinuation rate versus 16% male, against a backdrop of 76% vs 36% overall adverse events. The authors interpret the trial’s failure as a “feasibility threshold” problem — the 20 mg dose was simply not tolerable enough — but the sex-specific discontinuation pattern is the real mechanistic story. PDE-5 inhibitors have known sex-differential pharmacokinetics and tolerability profiles (women experience more headache and flushing at equivalent exposures), and the trial enrolled 20/76 (26%) women. If the 46% female dropout were reproducible in a larger trial, the dose-finding question would have a sex-specific answer: women should start at 5 or 10 mg, not 20 mg. The shadow is that ETLAS-2’s feasibility failure is actually a dose-selection failure disguised as a drug-class failure.

2. Inversion Engine

For the opposite conclusion to hold — that PDE-5 inhibition is futile in CSVD — you would need a much larger, longer, dose-finding trial in which even the most responsive subgroup (e.g., per-protocol completers with severe baseline WMH burden) failed to show a WMH signal. The per-protocol P = 0.054 sits in a zone where the absence of evidence is not yet evidence of absence. PASTIS (sildenafil in vascular cognitive impairment) and OxHARP (sildenafil in small vessel disease, both 2022/2024) reported mixed signals on related endpoints. The PDE-5 mechanism — boosting NO-cGMP signaling, which is documented impaired in CSVD — is biologically plausible. A replication at 5–10 mg in a more tolerant cohort (lower AE burden, longer follow-up) could produce a cleaner signal.

3. Second-Order Catalyst

If the per-protocol WMH signal is real, the second-order effect is the opening of a non-antithrombotic vascular-reactivity drug class for CSVD. The current CSVD pharmacological pipeline is dominated by antiplatelet agents (LACI-2: isosorbide mononitrate + cilostazol) and BP control (SPRINT3).). A PDE-5 strategy at lower doses would expand the therapeutic landscape beyond antithrombotic + antihypertensive to include endothelial-function-targeted vasodilators. This is a paradigm shift in CSVD drug development that LACI-1/LACI-2 only partially address.

4. Asymmetric Leverage

The asymmetric payoff is the per-protocol WMH volume signal (relative change 0.939, 95% CI 0.881–1.001, P = 0.054). If tadalafil delivers a 6% relative reduction in WMH volume over 3 months in completers, longer follow-up could plausibly amplify this — WMH progression in CSVD is on the order of 5–10% per year in observational studies. A 2-year trial at a tolerable dose could shift the rate of WMH progression by 10–15% relative, which is clinically meaningful. The leverage is in dose-finding at 5–10 mg + 12-month follow-up, which would be a cheaper trial than ETLAS-2 and might succeed where ETLAS-2’s 20 mg dose failed.

5. Paradigm Destroyer

What does this paper kill? It kills the assumption that a simple “high-dose PDE-5 inhibitor for CSVD” strategy is feasible. PDE-5 inhibitors are well-tolerated in erectile dysfunction and pulmonary hypertension populations — but those populations are younger, male- predominant, and on chronic (not first-3-month) dosing. CSVD patients are older (mean 68), often female, and have lower acute tolerability to PDE-5 side effects. The 20 mg daily dose that works for erectile dysfunction does not transfer. ETLAS-2 quietly establishes that dose-finding studies in CSVD need a sex-stratified design — a methodological point that will reappear in future CSVD drug development.

MVP — Minimum Viable Proof

For the bedside clinician: tadalafil is not yet a CSVD treatment. The ETLAS-2 trial failed its feasibility endpoint; the mechanistic WMH signal is hypothesis-generating only. Continue standard CSVD management: BP control (target < 130/80 per AHA/ESO guidance), antithrombotic therapy per stroke subtype, lifestyle modification (smoking cessation, exercise), and consideration of cilostazol/isosorbide mononitrate per LACI-2 results. Watch for a lower-dose PDE-5 inhibitor trial.

Best Combination

Pair this with: (1) the LACI-2 trial (isosorbide mononitrate + cilostazol in CSVD — JAMA Neurology 2023), which tested a similar vasoactive strategy with a positive imaging signal and reasonable tolerability; (2) the PASTIS trial (sildenafil in vascular cognitive impairment), which tested a related PDE-5 mechanism at a different dose with mixed results; (3) the OxHARP trial (sildenafil in small vessel disease, Circulation Research 2024), which reported cerebrovascular reactivity improvements with sildenafil. Together: PDE-5 inhibitors have mechanistic rationale for CSVD but the dosing question is unresolved; LACI-2’s cilostazol/ISMN combination is currently the leading vasoactive strategy for CSVD.

Overvalue Warning

Three things to not over-read. First, the P = 0.054 WMH volume finding is from a per-protocol subset of a feasibility trial — the sample size is small, the subgroup is selected (tolerant completers), and the P-value would not survive multiple-comparison correction across all secondary endpoints. Second, the 76% vs 36% adverse event rate is a clear drug-class signal — PDE-5 inhibitors at 20 mg daily are not well-tolerated in elderly CSVD patients, and lower-dose trials must show substantially better tolerability before any efficacy signal can be trusted. Third, female-specific tolerability is not a footnote — the 46% female dropout is the single most actionable finding in the trial and should drive dose-selection in any future PDE-5 CSVD program.