◈ Nootpedia
Compact view
Research pass: medium Pharmaceutical · Oral SKIP-FOR-NOW HIGH

Roflumilast

Extended Research
◈ Extended Research ◈

Our depth — beyond the mirror

Deeper analysis, verdict reasoning, and per-archetype recommendations from our research team.

▸ Our verdict SKIP-FOR-NOW HIGH

Off-label cognitive use is supported by two real human RCTs (Van Duinen 2018 episodic memory in healthy elderly; Gilleen 2018 schizophrenia adjunct) but the catalytic-site, non-subtype-selective binding produces nausea/diarrhea/weight loss in 10-20% of users — the same problem that killed rolipram. For Dylan-archetype: BPN14770 (allosteric PDE4D NAM, primate-specific binding pocket) is the cleaner version of this exact mechanism, already documented on the wiki, and that's where the cognitive-PDE4 thesis should live until/unless a topical or subtype-selective replacement emerges. Verdict would not change without a healthy-young-adult cognitive RCT showing tolerable risk-benefit at chronic dosing.

Research pass: medium
▸ Decision matrix by user profile Per-archetype
  • Dylan20-30, brain-priority, high cognitive workload (Dylan-archetype)
    SKIP-FOR-NOW

    HIGH confidence. Real cognitive signal in two RCTs but the catalytic-site, non-subtype-selective binding produces nausea/diarrhea/weight loss in 10-20% of users at the cognitively-active dose. Weight loss is mechanism-driven, not solely GI, and a 20yo MMA athlete protecting muscle mass has no use for chronic ~2 kg weight loss. BPN14770 is the cleaner PDE4 expression of this exact mechanism if Dylan ever wants to pursue it. V5 already includes higher-evidence cognitive options (modafinil, semax/NASA, bromantane, Adamax).

  • 30-50, executive maintenance
    SKIP

    Same reasoning; cognitive lift is small and AE cost is meaningful. Topical Zoryve is OPTIONAL for relevant psoriasis/AD indications but doesn't deliver cognitive effect.

  • 50+, mild cognitive decline / pre-MCI
    WATCH-LIST

    PDE4 inhibitors as AD therapy remain perpetually-promising. If patient already tolerates 500 mcg/day for COPD or another indication, the cognitive lift may be a meaningful bonus. Not a primary cognitive intervention given the AE profile. Subgroup with higher-CDR-SB-baseline impairment is a hypothesis worth tracking (echoes BPN14770 PICASSO subgroup story).

  • Anxiety-prone
    AVOID

    Mood/psychiatric signal — could amplify anxiety, low mood. Skip.

  • High athletic load, tested status
    AVOID

    for chronic use. Weight loss is a major problem for athletes. WADA: not on banned list (PDE4 inhibitors are not prohibited), but the AE profile alone disqualifies for cardio/strength athletes who can't afford 2 kg loss.

  • Sleep-disordered
    NEUTRAL-CAUTION

    Insomnia signal in some users; avoid PM dosing. Generally not sleep-relevant if dosed AM.

  • Recovery-focused (post-injury, post-illness)
    SKIP

    Weight loss + nausea are anti-recovery. Wrong tool for this profile.

  • Strength/anabolic-focused
    STRONG SKIP

    Weight loss directly opposes goals.

  • Severe COPD with chronic bronchitis (on-label)
    STRONG INDICATED

    This is what it's approved for. Standard of care discussion with pulmonologist, not biohacking.

  • Plaque psoriasis / atopic dermatitis (topical Zoryve)
    STRONG

    INDICATED for relevant skin conditions. Topical delivers minimal systemic exposure; AE profile dramatically better than oral. Different conversation than oral cognitive use.

▸ Subjective experience (deep)

On-label (COPD patients): Most patients report nausea and loose stools in the first 2-4 weeks, often with weight loss (median ~2 kg over 1 year). For most who continue, GI symptoms attenuate after week 4-6 but never fully resolve. Some patients report mood changes (low mood, anxiety) and insomnia. Cognitive effects are not commonly reported because COPD cohorts are not assessing cognition.

Off-label cognitive (self-experimenter reports):

  • Days 1-3: Often nausea, sometimes diarrhea. Headache common. Take with food helps but doesn't eliminate. Some users report mild mental "fog" the first few days that lifts.
  • Days 4-14: GI tolerance improves but doesn't disappear. Subjective cognitive lift starts emerging — "stickier memory," better word retrieval, faster reading comprehension. Consistent with the Van Duinen signal in young/middle-aged users.
  • Weeks 3-8: Cognitive lift, when present, plateaus. ~20-30% of users report no benefit at all (consistent with the small effect size of acute studies). Weight loss starts to show up at the higher end (~1-2 kg over 8 weeks at 500 mcg/day).
  • Discontinuation: No reported withdrawal; cognitive effect (subtle) fades over 1-2 weeks consistent with the half-life (~17h roflumilast, ~30h N-oxide active metabolite).

Honest expectation-setting for Dylan-archetype: This is not a "feel something" compound for cognition. The Van Duinen signal was modest even in optimized acute conditions. The chronic experience for healthy young users is "real but small cognitive benefit, real and uncomfortable GI/weight cost" — and the size of the cognitive benefit does not generally justify the size of the GI cost when better options exist. For someone specifically prioritizing the PDE4 cognitive-enhancement mechanism, BPN14770 (or its allosteric class) is the cleaner bet.

▸ Tolerance + cycling deep dive
  • Tolerance buildup: Not well-characterized for the cognitive endpoint. The anti-inflammatory effect in COPD is durable over 1-year trials. Mechanistically, chronic PDE4 inhibition could theoretically downregulate CREB-pathway responsiveness, but this is not documented.
  • Recommended cycle: No cycling protocol established. COPD use is chronic. For off-label cognitive use, 4-8 weeks on / 4 weeks off would parallel the BPN14770 protocol but has no published rationale.
  • Reset protocol: Discontinue 4 weeks. No documented reset adjuncts. Mood/GI baseline restoration within 1-2 weeks given half-life.
▸ Stacking deep dive

Synergistic with (theoretical, no clinical data)

  • None recommended for cognitive use at the supplemental level. The dose-limiting toxicity is GI/cAMP-load related, not deficiency of any specific cofactor.

Avoid stacking with

  • bpn14770 — same mechanism (PDE4 inhibition); redundant and additive emetic risk. Pick one. BPN14770 is the cleaner pick if the PDE4 mechanism is wanted.
  • rolipram, apremilast, ibudilast — same class concern; cumulative GI and area-postrema cAMP load
  • High-dose theophylline, pentoxifylline — non-selective PDE inhibitors, additive cAMP elevation
  • forskolin — directly stimulates adenylyl cyclase (cAMP from production side); roflumilast blocks degradation (cAMP from breakdown side). Combined cAMP elevation could amplify GI / nausea / cardiovascular signaling. Theoretical; not a typical stack but mechanistically caution-flagged.
  • CYP1A2 inhibitors (fluvoxamine, ciprofloxacin, high-dose caffeine in slow metabolizers) — raise roflumilast plasma levels; amplifies AEs. Fluvoxamine roughly doubles roflumilast exposure.
  • Strong CYP3A4 inhibitors (ketoconazole, clarithromycin, ritonavir, grapefruit at high intake) — modest exposure increase; caution in chronic combination.
  • Strong CYP1A2 + CYP3A4 inhibitor combinations (e.g., enoxacin) — substantially raise exposure; avoid.
  • Hepatic impairment — Roflumilast is contraindicated in moderate-to-severe hepatic impairment (Child-Pugh B/C).

Neutral / safe co-administration

  • V4 stack components: citicoline, DHA, magnesium glycinate/threonate, NAC, phosphatidylserine, rhodiola, theanine, glycine, D3+K2, beta-alanine, vitamin C — no plausible interaction at supplemental doses. Note: Dylan's caffeine ramp is a CYP1A2 substrate (not inhibitor), so no induction concern, but in CYP1A2 slow metabolizers with high caffeine intake, additive CYP1A2 burden could marginally raise roflumilast exposure.
  • Modafinil — no documented interaction. Theoretical cardiovascular additive load (mild).
  • bromantane, semax, n-acetyl-semax-amidate, adamax — no documented interaction; orthogonal mechanisms.
  • Hormonal contraceptives — no significant interaction documented.
▸ Drug interactions deep dive
  • CYP1A2 substrate: Major. Roflumilast is metabolized to its active N-oxide by CYP1A2 and CYP3A4. CYP1A2 inhibitors raise levels significantly:
    • Fluvoxamine (strong CYP1A2 inhibitor) ≈ 2× exposure increase
    • Ciprofloxacin / enoxacin — significant exposure increases
    • Cimetidine — modest
    • High-intake fluoroquinolone or fluvoxamine co-therapy is the textbook DDI to avoid
  • CYP3A4 substrate: Moderate. Strong CYP3A4 inhibitors (ketoconazole, clarithromycin, ritonavir) modestly raise exposure; grapefruit at heavy intake could contribute. Strong CYP3A4 inducers (rifampin, carbamazepine, St John's wort, phenytoin) lower exposure and may reduce efficacy.
  • Theophylline: Both PDE-related; pharmacokinetic interaction not major but pharmacodynamic additive (cAMP). Caution.
  • Hormonal contraceptives: No significant interaction; safe to co-administer.
  • Anticoagulants: No documented signal (unlike forskolin's mild antiplatelet effect via cAMP).
  • Alcohol: No PK interaction; alcohol's GI irritation may compound nausea risk in early dosing — avoid heavy intake during first 4 weeks.
  • Smoking: Smokers have induced CYP1A2 activity → modestly lower roflumilast exposure. COPD-relevant; not relevant for Dylan (non-smoker).
▸ Pharmacogenomics
  • *CYP1A2 (rs762551, 1F allele) — primary pharmacogenomic axis. Slow metabolizers (homozygous *1F/*1F or other slow alleles) may have higher roflumilast exposure and amplified AEs at standard 500 mcg/day. Fast metabolizers (induced or genetic) may need higher doses to reach efficacy thresholds — this is the same axis that explains caffeine sensitivity variability.
  • CYP3A4/5: Moderate contributor. Polymorphisms (CYP3A5*3 — common in Dylan's Nordic/British ancestry, low-CYP3A5 expressor) modestly affect exposure. Practical effect minor at standard doses.
  • PDE4D / PDE4B polymorphisms: Some studies link PDE4D variants to working-memory baseline differences and stroke risk; no published roflumilast response data stratified by PDE4D genotype.
  • Dylan-specific: Once 23andMe results land (~June 2026), pull rs762551 (CYP1A2 *1F). Results would predict roflumilast tolerability — slow metabolizers should not start at 500 mcg if they ever pursued this. Still a future-direction consideration; not actionable today.
▸ Sourcing deep dive
Path Vendor Cost Reliability Notes
Rx (US) Daliresp brand 500 mcg × 30 ~$300-400/month brand high Insurance often required for PA; expensive cash-pay
Rx (US) Generic roflumilast 500 mcg × 30 ~$50-150/month with discount programs (GoodRx etc.) high Generic available since approx 2024; supply expanding
Rx (UK/EU) Daxas (AstraZeneca) similar to US brand high EU-approved 2010
Rx (India / international) Generic roflumilast (Cipla, Lupin, Sun Pharma) $20-60/month medium-high Common for off-label cognitive self-experimenters; verify reputable pharmacy
Topical Zoryve cream/foam (Arcutis) — different formulation ~$700-1000 brand 60 g cash; insurance often covers high Approved 2022 (psoriasis), expanded for AD/seb derm. Negligible systemic exposure → does NOT deliver the cognitive effect
Research-chem Various — analytical grade for in-vitro use $50-200 / mg high (analytical) Not designed for personal use; price-prohibitive

Recommendation for Dylan: Don't source. The cleaner mechanism (BPN14770 if PDE4 thesis pursued, modafinil/semax/bromantane for general cognitive enhancement) is already on V5/wiki with substantially better evidence and tolerability for his archetype.

▸ Biomarkers to track (deep)

If running roflumilast off-label (Dylan should not, but for completeness):

Baseline (before starting)

  • Weight + body composition
  • Resting HR + blood pressure
  • Mood baseline (PHQ-9, GAD-7) — particularly important for psychiatric AE watch
  • Cognitive battery — N-back, One Card Back, CNS Vital Signs full panel; verbal memory (RAVLT or NIH Toolbox)
  • LFT panel (ALT, AST, ALP, bilirubin)
  • CBC
  • Once 23andMe lands: CYP1A2 rs762551 status

During use

  • Weight weekly (early warning for mechanism-driven loss)
  • Daily VAS: nausea, diarrhea, headache, mood, sleep, appetite
  • Cognitive battery week 2, 4, 6, 8
  • Mood (PHQ-9, GAD-7) week 4 and week 8
  • LFTs at week 8 (cheap and prudent)

Post-cycle

  • Weight at 2 and 4 weeks post-stop (look for rebound)
  • Cognitive battery week 1 and week 4 post-stop
  • Mood at week 4 post-stop
▸ Controversies / open debates Live debate

  1. Is the cognitive signal real or artifact? Van Duinen 2018 was a single-dose acute study with episodic memory as outcome — small N (20), strong design (crossover double-blind placebo-controlled). The signal looks real. Gilleen 2018 in schizophrenia is consistent. Whether the signal translates to chronic dosing in healthy young adults at tolerable doses is genuinely unknown because no such RCT exists. This is the central evidence gap.

  2. Mechanism vs tolerability tradeoff is intrinsic. Roflumilast's binding profile (catalytic site, all four PDE4 isoforms) means the cognitive effect and the emetic effect are coupled — both flow from raising cAMP via PDE4 inhibition. The whole BPN14770 design philosophy was built on the recognition that this coupling is breakable only via allosteric / subtype-selective approaches. Roflumilast is the molecule that proves the rule.

  3. Topical Zoryve doesn't substitute for oral. A natural question is whether the topical formulation could deliver cognitive effects with better tolerability. The answer is no — systemic exposure from topical Zoryve is far below the threshold for CNS PDE4 inhibition. Topical works for skin because it doesn't need to cross the BBB.

  4. Generic roflumilast and the "cheap PDE4 inhibitor" bet. Now that generic roflumilast is available (~2024 expansion, $20-60/month from international pharmacies, $50-150 with US discount programs), the cost barrier to off-label cognitive use has dropped dramatically. This will likely produce more self-experimenter logs over the next 1-2 years. If those reports skew positive, the verdict could shift to a weak OPTIONAL for users specifically motivated by the PDE4 thesis. Currently the AE profile is the limiter, not cost.

  5. Mood/suicidality signal — real or artifact of COPD population? COPD patients have high baseline rates of depression and anxiety. Disentangling drug-induced vs population-baseline psychiatric AEs has been contentious. The conservative read is: real but small signal, watch carefully in patients with mood-disorder history, do not initiate without psychiatric baseline.

  6. Why isn't this a STRONG SKIP? Because the cognitive signal is genuinely real (two RCTs is more than most "cognitive enhancers" can claim) and the molecule has been validated in humans for the exact mechanism that BPN14770 is still trying to prove. For a user willing to accept 10-20% GI cost and 2 kg weight loss for a small but real cognitive lift, the molecule has a coherent use case. For Dylan-archetype, the cost-benefit is wrong. For an older user with insufficient cognitive options and tolerable GI baseline, it could be defensible. Hence SKIP-FOR-NOW (Dylan-specific) rather than SKIP-PERMANENT (universal).

▸ Verdict change log
  • 2026-05-06 — Initial verdict: SKIP-FOR-NOW, HIGH confidence. Real human cognitive signal (Van Duinen 2018, Gilleen 2018) but catalytic-site, non-subtype-selective PDE4 binding produces nausea/diarrhea/weight loss in 10-20% of users at the cognitively-active dose. Mechanism-driven weight loss disqualifies for combat-sport athlete profile. BPN14770 (allosteric PDE4D NAM) is the cleaner expression of this exact mechanism and is already on the wiki at WATCH-LIST. V5 already includes higher-evidence cognitive options. Verdict could shift if a healthy-young-adult chronic cognitive RCT emerged showing tolerable risk-benefit, or if topical/subtype-selective replacements expanded the option set.
▸ Open questions / gaps Open
  • Healthy-young-adult chronic cognitive RCT. Single-dose acute (Van Duinen) and 8-day (Gilleen) trials exist; no 8-12 week RCT in healthy young adults at 250-500 mcg/day with cognitive battery as primary outcome. This is the missing study that would settle the verdict.
  • CYP1A2 pharmacogenomic stratification. Slow metabolizers should theoretically have amplified AEs at standard doses; this hasn't been formally tested in cognitive trials.
  • Dose-response below 250 mcg. The Van Duinen 100 mcg arm showed cognitive benefit. Could a sub-clinical 100-150 mcg/day chronic dose deliver cognitive effect with substantially lower AE profile? Untested.
  • Combination with anti-emetic prophylaxis. Could ondansetron / domperidone co-therapy in the first 4 weeks open a tolerability window for chronic cognitive use? No published data.
  • Subtype-selective successors. PDE4B-selective inhibitors (less emetogenic than pan-PDE4) are in development for inflammatory indications. Whether any reach cognitive testing is the watch-item.
  • Long-term safety beyond 1 year in cognitive (lower-dose) regimens — no chronic data outside COPD.
▸ Sources (full, with our context)
Back to compact view