Post–Loper Bright Validation of AEDPA Deference in Strickland Habeas Review (and No Per Se Duty to Retain Defense Experts)

I. Introduction

A Michigan jury convicted Scott Allen DeBruyn of delivering oxycodone to Camille Gesiakowski, causing her death, under Mich. Comp. Laws Ann. § 750.317a. The prosecution theory was straightforward: Gesiakowski repeatedly sought “oxys,” DeBruyn obtained Percocet (oxycodone + acetaminophen), and Gesiakowski died in a hotel room with oxycodone in her blood and acetaminophen in her urine—consistent with Percocet ingestion. The defense contested causation, arguing that in a mixed-drug setting the State failed to prove oxycodone was a “substantial factor,” pointing instead to alternative mechanisms like serotonin syndrome or difluoroethane-induced arrhythmia.

After conviction, DeBruyn pursued ineffective assistance in state post-conviction proceedings via a Ginther hearing, claiming his trial lawyers (1) failed to investigate an “acetaminophen-based defense” (no acetaminophen in blood) and (2) failed to call a defense expert to bolster his causation/alternative-cause theories. The Michigan courts rejected relief on the merits. On federal habeas, the district court denied the petition, and the Sixth Circuit affirmed, emphasizing AEDPA’s constraints and the narrowness of federal review.

The opinion is notable for its explicit rejection of the argument that Loper Bright undermines AEDPA’s constitutionality, and for its insistence that neither expert-retention nor exhaustive investigation of every conceivable defense is constitutionally required—particularly under AEDPA’s deference layered atop Strickland.

II. Summary of the Opinion

The Sixth Circuit affirmed denial of habeas relief because DeBruyn could not satisfy AEDPA’s requirements. The court held:

• AEDPA is constitutional under Article III and is not displaced by Loper Bright Enterprises v. Raimondo; AEDPA does not transfer federal judicial power to state courts.
• Failure to investigate “acetaminophen-based defense”: the Michigan Court of Appeals reasonably applied Strickland v. Washington in finding no deficient performance and no prejudice; the blood/urine acetaminophen evidence was consistent with the prosecution’s timeline because acetaminophen metabolizes faster than oxycodone.
• Failure to call a defense expert: the state court reasonably concluded cross-examination of prosecution experts was an objectively reasonable strategy and that proposed expert opinions were speculative or cumulative, defeating prejudice under Strickland as constrained by AEDPA.

III. Analysis

A. Precedents Cited

1. Governing ineffective-assistance framework

• Strickland v. Washington, 466 U.S. 668 (1984).
  Role in the opinion: The constitutional baseline: deficient performance + prejudice. The panel repeatedly invokes Strickland’s presumption of reasonable professional assistance, its “all the circumstances” lens, and its recognition that counsel need not investigate every conceivable defense. This becomes decisive once AEDPA requires deference to the state court’s Strickland application.
• Harrington v. Richter, 562 U.S. 86 (2011).
  Role: Central to the court’s “double deference” analysis. The opinion relies on Harrington for: (i) the heightened prejudice formulation (“substantial” likelihood, not merely “conceivable”); (ii) the idea that cross-examination is often sufficient and that Strickland does not require “equal and opposite” defense experts; and (iii) the AEDPA standard that relief requires a state decision so wrong it is beyond fairminded disagreement. The panel uses Harrington to reject any per se expert requirement.
• Premo v. Moore, 562 U.S. 115 (2011).
  Role: Supports deference to counsel’s predictions about how trial will proceed. The Sixth Circuit uses Premo to justify treating counsel’s choice to focus on causation (rather than disputing the drug-delivery mechanism) as within reasonable strategic range.
• Cullen v. Pinholster, 563 U.S. 170 (2011).
  Role: Provides two key tools: (i) courts must “affirmatively entertain” possible reasons for counsel’s actions; (ii) cumulative evidence undermines prejudice. The panel applies this in rejecting claims that additional expert testimony on serotonin syndrome would have altered the verdict.
• Knowles v. Mirzayance, 556 U.S. 111 (2009).
  Role: Highlights how general Strickland’s standards are, giving state courts “latitude” under AEDPA. The opinion uses this to explain why DeBruyn must show something close to an indefensible strategy choice.
• Dunn v. Reeves, 141 S. Ct. 2405 (2021) (per curiam).
  Role: Reinforces that counsel need not be “exemplary,” only competent, and that AEDPA forbids substituting a federal court’s preferred approach for a reasonable state-court assessment. The panel leans on Dunn to reject the idea that expert testimony is mandatory whenever expert evidence is presented by the State.
• Shinn v. Kayer, 141 S. Ct. 517 (2020) (per curiam) and Sexton v. Beaudreaux, 585 U.S. 961 (2018) (per curiam).
  Role: Supply the “beyond any possibility for fairminded disagreement” articulation and the caution that “the more general the rule, the more leeway state courts have.” The panel uses these to frame DeBruyn’s burden as extraordinarily high.
• White v. Woodall, 572 U.S. 415 (2014), Brown v. Davenport, 142 S. Ct. 1510 (2022).
  Role: Define “clearly established” law and the needed specificity. The opinion uses these to reject DeBruyn’s attempt to transform general principles about investigation and experts into a specific rule requiring a toxicologist.

2. Limits on what counts as “clearly established” law under AEDPA

• Glebe v. Frost, 574 U.S. 21 (2014) (per curiam) and Marshall v. Rodgers, 569 U.S. 58 (2013) (per curiam).
  Role: The court uses these to reject reliance on Sixth Circuit decisions as “clearly established” Supreme Court law and to forbid “sharpening” Supreme Court generalities into new specific rules.

3. Fact-finding deference and credibility/interpretation of competing evidence

• Shoop v. Twyford, 142 S. Ct. 2037 (2022).
  Role: Provides the “substantially higher threshold” for challenging state fact determinations. The panel applies this to reject DeBruyn’s claim that the state court unreasonably credited the State’s toxicology interpretation.
• Anderson v. City of Bessemer City, 470 U.S. 564 (1985) and Hill v. Shoop, 11 F.4th 373 (6th Cir. 2021) (en banc).
  Role: Reinforce that choosing between two reasonable interpretations of evidence is not “unreasonable.” The opinion uses these to uphold the state court’s acceptance of metabolism testimony (acetaminophen clears faster than oxycodone).

4. Expert-witness decisions and what Hinton does (and does not) require

• Hinton v. Alabama, 571 U.S. 263 (2014) (per curiam).
  Role: The court narrows DeBruyn’s reading of Hinton. It explains Hinton was about an attorney’s non-strategic mistake regarding funding availability—not a categorical requirement to hire an expert whenever the State uses experts.
• Burt v. Titlow, 571 U.S. 12 (2013).
  Role: Cited via Dunn for the proposition that deficient performance requires an approach “no competent lawyer” would take; supports the deference afforded to counsel’s cross-examination strategy.

5. State post-conviction procedure and Michigan law framing

• People v. Ginther, 212 N.W.2d 922 (Mich. 1973).
  Role: Establishes the mechanism for developing an ineffective-assistance record in Michigan. The Sixth Circuit uses the Ginther hearing testimony as the evidentiary basis to evaluate prejudice and the plausibility of proposed expert theories.
• People v. DeBruyn, No. 352274, 2022 WL 981281 (Mich. Ct. App. Mar. 31, 2022) (per curiam); People v. DeBruyn, 978 N.W.2d 836 (Mich. 2022) (mem.).
  Role: The state adjudications to which AEDPA deference attaches; the Sixth Circuit repeatedly measures whether their applications of Strickland and their fact determinations were “unreasonable.”

6. Sixth Circuit’s emerging post–Loper Bright AEDPA-constitutionality line

• Sanders v. Plappert, --- F.4th ---, 2026 WL 593932 (6th Cir. Mar. 3, 2026).
  Role: The principal authority used to reject DeBruyn’s Article III attack on AEDPA and to explain why Loper Bright has no application to AEDPA’s habeas limits.
• Bowling v. Parker, 882 F. Supp. 2d 891 (E.D. Ky. 2012) (Thapar, J.).
  Role: Cited for the proposition that Congress may restrict the availability of habeas relief to state prisoners without violating Article III.
• Miles v. Floyd, No. 24- 1096, 2025 WL 902800 (6th Cir. Mar. 25, 2025).
  Role: Reinforces the court’s view that Loper Bright does not destabilize AEDPA.
• Loper Bright Enterprises v. Raimondo, 144 S. Ct. 2244 (2024).
  Role: Treated as inapposite. The Sixth Circuit distinguishes the APA’s demand for independent judgment about agency statutory authority from Congress’s power to set standards governing issuance of habeas relief to state prisoners.

7. Sixth Circuit precedent invoked and cabined

• Rayner v. Mills, 685 F.3d 631 (6th Cir. 2012).
  Role: Confirms AEDPA applies when a state court adjudicates an ineffective-assistance claim on the merits.
• Richey v. Bradshaw, 498 F.3d 344 (6th Cir. 2007).
  Role: DeBruyn relies on it for failure-to-investigate; the panel rejects it as effectively overtaken, citing Kendrick v. Parris, 989 F.3d 459 (6th Cir. 2021). The court further explains that, even on its own terms, Richey is distinguishable because DeBruyn’s counsel did investigate and challenge the scientific causation premise through cross-examination.
• Kendrick v. Parris, 989 F.3d 459 (6th Cir. 2021).
  Role: Used to explain why older Sixth Circuit habeas decisions that do not faithfully apply AEDPA deference are unreliable; also used to characterize certain expert-witness discussion as dicta when a decision refused to rest on that ground.
• Stermer v. Warren, 959 F.3d 704 (6th Cir. 2020).
  Role: DeBruyn cites it to argue a duty to retain an expert; the panel rejects it as not “clearly established” Supreme Court law, warns against using circuit decisions to create new rules, and notes the discussion is dicta under Kendrick.
• Clardy v. Pounds, 126 F.4th 1201 (6th Cir. 2025).
  Role: Used to contrast Hinton-type mistakes of law with strategic decisions about experts; supports treating DeBruyn’s counsel as exercising strategy, not misunderstanding legal constraints.

B. Legal Reasoning

1. The “double deference” structure (Strickland + AEDPA)

The opinion’s architecture is classic modern habeas: (1) identify Strickland’s already-deferential performance and prejudice standards; (2) overlay AEDPA’s additional requirement that the state court’s application be not merely wrong but objectively unreasonable; (3) deny relief if “fairminded disagreement” is possible. This framing does more than set the tone—it determines outcomes. DeBruyn’s theories (acetaminophen timing, tolerance re-acquisition, difluoroethane volatility, seizure risk) might generate debate in de novo review, but AEDPA converts debate into defeat unless the state court’s rejection is beyond the permissible range.

2. AEDPA’s constitutionality after Loper Bright

DeBruyn argued AEDPA violates Article III by “transferring” federal judicial power to state courts, and that Loper Bright’s insistence on independent judgment undermines § 2254(d). The Sixth Circuit rejects both: AEDPA does not empower state courts to exercise federal judicial power; it limits federal remedial authority to issue habeas relief to state prisoners. The court reads Loper Bright as addressing how federal courts review federal agency action under the APA, not Congress’s authority to define the conditions for collateral relief. Citing Sanders v. Plappert and Miles v. Floyd, the panel treats this as settled: Article III permits Congress to restrict the writ’s availability in this context.

3. Failure to investigate the “acetaminophen-based defense”

On performance, the court views counsel’s strategic choice as reasonable: the record strongly supported that Gesiakowski consumed oxycodone in Percocet form (requests for “oxys,” DeBruyn’s purchase of 40 Percocet pills, photo of “A333,” oxycodone in blood plus acetaminophen in urine). With that factual backdrop, counsel’s choice to focus on the more contestable element—whether oxycodone was a “substantial factor” in a mixed-drug death—fell within the range Strickland protects, especially given Premo and Cullen’s instruction to credit counsel’s predictive judgments and possible reasons for not pursuing other avenues.

On prejudice, the opinion turns to pharmacokinetics: acetaminophen can metabolize and clear more quickly than oxycodone, meaning absence in blood at death is consistent with ingestion earlier and subsequent excretion (with acetaminophen still detectable in urine). Because the state court credited expert testimony supporting this metabolism timeline, the federal court—bound by § 2254(d)(2) and § 2254(e)(1)—refuses to reweigh the battle of experts. Without a clear and convincing showing that the state factual determination was unreasonable, DeBruyn cannot show a substantial likelihood of acquittal had counsel pursued the acetaminophen theory.

4. Failure to call a defense expert to contest “substantial factor” causation

The court treats the expert decision as paradigmatically strategic under Harrington v. Richter: cross-examination can be sufficient, and Strickland does not require mirroring the prosecution’s experts. The panel underscores that DeBruyn’s counsel actively and knowledgeably cross-examined on serotonin syndrome, mixed-drug uncertainty, and alternative mechanisms—thus this was not a case of abandonment or ignorance.

On prejudice, the court reviews four proposed expert topics and repeatedly characterizes them as speculative or cumulative:

• Residual tolerance / re-tolerance: general propositions without case-specific quantification of tolerance to 180 ng/mL oxycodone.
• Serotonin syndrome: largely cumulative, and counsel already confronted an expert’s suggestion that oxycodone affects serotonin using studies indicating “does not affect serotonin levels.”
• Difluoroethane: volatility makes levels at death unknown; proposing lethality at 5.3 µg/mL does not establish Gesiakowski had that level, leaving only a “theoretical possibility.”
• Seizure: rests on stacked assumptions (why Keppra was prescribed; whether it was abruptly stopped), again producing only a theoretical possibility.

Under Harrington’s prejudice standard and AEDPA’s deference, theoretical possibilities do not amount to a substantial likelihood of a different verdict.

C. Impact

1. Habeas practice: strengthening the post–Loper Bright AEDPA wall

By squarely rejecting the claim that Loper Bright destabilizes AEDPA, and by relying on Sanders v. Plappert, the Sixth Circuit reinforces a developing, publication-worthy line: arguments that AEDPA violates Article III (or that Loper Bright mandates de novo review of state-court merits determinations) will fail in this circuit absent a new Supreme Court directive. This narrows the strategic space for habeas petitioners seeking to avoid § 2254(d) deference via structural-constitutional theories.

2. Ineffective assistance: no per se “expert required” rule

The opinion reiterates that neither “technical” subject matter nor the prosecution’s use of experts automatically triggers a constitutional duty to retain a defense expert. Instead, the question remains whether counsel’s approach was one “no competent lawyer” would choose and whether omitted expert testimony creates a substantial likelihood of acquittal—both filtered through AEDPA. For future cases, petitioners must develop concrete, case-specific prejudice showings rather than general scientific possibilities.

3. Mixed-drug causation defenses: speculation versus proof

The court’s prejudice analysis signals skepticism toward alternative-cause defenses that rely on what might have happened (unknown difluoroethane level at death, possible seizure, possible rapid re-tolerance) without anchoring those theories to record-supported probabilities. In practice, effective post-conviction presentations in mixed-drug deaths may need tighter factual linkages (timeline evidence, dosage evidence, medical history evidence) to avoid being labeled “theoretical.”

IV. Complex Concepts Simplified

• AEDPA (28 U.S.C. § 2254(d)): A federal court generally cannot grant habeas relief just because it believes the state court was wrong. Relief is available only if the state court’s decision was contrary to, or an unreasonable application of, clearly established Supreme Court law, or rested on an unreasonable fact determination.
• “Clearly established Federal law”: Only Supreme Court holdings count (not lower-court cases, and not Supreme Court dicta). Petitioners cannot convert broad statements into new, specific rules.
• “Double deference”: Strickland already defers to counsel’s strategic choices; AEDPA then adds deference to the state court’s decision that counsel was adequate. The combined effect is highly restrictive.
• Ginther hearing: A Michigan post-conviction evidentiary hearing used to build a factual record for ineffective-assistance claims.
• “Substantial factor” causation (Michigan delivery-causing-death): The State must prove not only delivery of the controlled substance but that it substantially contributed to death. In mixed-drug cases, defendants often argue causation cannot be reliably isolated.
• Pharmacokinetics (metabolism/excretion): Drugs move through the body at different rates. Here, acetaminophen can clear from blood faster than oxycodone, so its absence in blood at death does not necessarily mean it was never ingested—especially if it appears in urine.
• Serotonin syndrome: A potentially fatal condition caused by excessive serotonin activity; certain drug combinations (e.g., tramadol + fluoxetine) can raise risk. A key dispute was whether oxycodone contributes to serotonin elevation.

V. Conclusion

DeBruyn v. Douglas is a published reaffirmation of two powerful constraints on federal habeas relief. First, AEDPA’s deferential scheme survives Article III attack and is unaffected by Loper Bright Enterprises v. Raimondo. Second, under AEDPA’s “fairminded disagreement” standard layered atop Strickland, counsel’s decisions not to pursue an additional toxicology theory and not to retain a defense expert will rarely warrant habeas relief where counsel mounted an active cross-examination and the omitted evidence is speculative or cumulative. The decision thus fortifies the Sixth Circuit’s post–2024 habeas landscape: constitutional challenges to AEDPA’s structure and generalized “expert required” ineffective-assistance theories face exceptionally steep odds.