I. A Label Is a Legal Object Before It Is a Warning
A prescription-drug label can look like a sheet of paper, a paragraph in a package insert, or a familiar warning on a digital screen. In litigation it is something harder: a legal object shaped by federal approval, state tort duties, dated scientific information, and a record that must show what a manufacturer could lawfully have done at a particular moment. Combs v. Bayer AG, 2026 OK CIV APP 17, makes that point with unusual clarity. The Oklahoma Court of Civil Appeals affirmed summary judgment for the manufacturer on claims centered on the alleged failure to warn of a risk associated with an FDA-approved contrast agent. The court did not decide that every prescription-drug warning claim is barred, that a product is harmless, or that an injured patient cannot prove a warning case. It decided that the plaintiff before it had not supplied the evidence needed to show the manufacturer could have changed its federally approved warning unilaterally before the 2016 injection at issue. That distinction is the whole case. Law often invites broad language—federal rules, public safety, patient risk—but a warning claim cannot survive on atmosphere. It must identify the warning allegedly required, the information that made it required, the legal mechanism for adding it, the date on which that mechanism was available, and the causal path from the absent warning through the treating professional to the claimed harm. In the space between a frightening medical event and a viable civil claim, the record does the work.
II. The Court Treated the Record, Not Fear, as the Threshold
Combs arose from claims concerning Magnevist, an FDA-approved gadolinium-based contrast agent used in an MRI. The plaintiff alleged that the manufacturer failed to provide an adequate warning about gadolinium retention and its asserted consequences in patients with normal kidney function. The court described a record in which the parties did not dispute that gadolinium could remain in the body and that the manufacturer and FDA were aware of retention and reported adverse reactions before the injection. But a litigation record does not become legally sufficient merely because it contains a condition, a concern, an internal discussion, or a sequence of reported events. The court held that the plaintiff had not identified evidence establishing the legally relevant causal association—evidence that could support a unilateral warning change under the governing federal framework. The opinion also rejected unsupported assertions about the manufacturer’s motives and noted that a collection of sealed exhibits did not itself establish the propositions the plaintiff attributed to them. This is not a lesson about minimizing a patient’s experience. It is a lesson about forensic discipline. A competent evaluation preserves the medical chronology, the product identity, the label version, the treating records, adverse-event materials, regulatory communications, and the actual expert opinions. It then asks the narrower question the court had to answer: what did the admissible record show at the time? Civil justice has room for human suffering, but it does not replace proof with the intensity of that suffering. A careful lawyer treats the medical event as a reason to investigate more deeply, not as a substitute for the evidence the governing law requires.
III. The CBE Path Requires Newly Acquired Information
The federal-label question in Combs centered on the changes-being-effected, or CBE, regulation. Ordinarily, a manufacturer cannot simply replace an FDA-approved prescription-drug label with its own preferred version. Federal regulations nevertheless allow specified labeling changes through a supplemental process when the change reflects newly acquired information and adds or strengthens a warning, precaution, contraindication, or adverse reaction for which the evidence meets the applicable causal-association standard. See 21 C.F.R. § 314.70(c)(6)(iii)(A). The related labeling rule explains that warnings should address a clinically significant hazard when there is reasonable evidence of a causal association; a causal relationship need not be definitely established. 21 C.F.R. § 201.57(c)(6)(i). Those propositions can sound generous in the abstract, but they are not a free pass around federal approval. The court explained that newly acquired information means data, analyses, or other information not previously submitted to the FDA, including certain later studies, adverse-event reports, or new analyses of prior data if they reveal a risk of a different type or greater severity or frequency than previously submitted information. The date of that information is therefore not a clerical detail. It is the hinge. A label change that could be supported today may not have been available years earlier; a warning that was legally possible after new evidence emerged cannot automatically be projected backward into an earlier record. The question is not whether a later observer wishes a different warning had existed. The question is whether the manufacturer had newly acquired information that made the CBE route legally available when it mattered.
IV. Retention Was Not the Same as a Causal Association
The most important evidentiary boundary in Combs is the line between a physical observation and the causal association required for a label change. The court recounted evidence that gadolinium could be retained in the body and that regulators were examining the issue. It nevertheless found that the record did not establish reasonable evidence, before the plaintiff’s injection, of a causal association between retention in patients with normal renal function and the asserted clinically significant adverse reactions. In other words, the existence of retention did not by itself answer the separate question whether the retention caused the specific type of serious harm that would support a strengthened warning under the regulation. The opinion’s careful separation of those propositions is a useful safeguard against a common litigation error: collapsing “something was found” into “this thing caused the claimed injury.” That collapse can hurt plaintiffs as much as defendants. It invites an opponent to attack the whole claim as speculative rather than confront a focused, supportable theory. A strong record distinguishes the exposure from the mechanism, the mechanism from the clinical consequence, the consequence from the individual medical history, and the individual medical history from the legal elements. That work may require treating records, retained samples where appropriate, pharmacovigilance material, regulatory history, qualified expert analysis, and a timeline tied to the operative label. Combs does not say that causation evidence can never be developed. It says the plaintiff’s summary-judgment record had not developed the kind of dated causal evidence that would have permitted the unilateral warning change on which her preemption argument depended.
V. Impossibility Preemption Is Narrow but Demanding
Federal preemption is often described in one word, as if it ends the analysis the moment it is spoken. The doctrine is more exacting than that. Under Wyeth v. Levine, 555 U.S. 555 (2009), state failure-to-warn claims involving prescription drugs are not categorically preempted merely because the FDA regulates drug labeling. But a state claim may be barred where it would have been impossible for the manufacturer to comply with both state law and federal law. The court in Combs explained that the preemption question is one for the court, citing Merck Sharp & Dohme Corp. v. Albrecht, 587 U.S. 299 (2019). In the setting before it, the plaintiff needed to show that a different warning was one Bayer could have added under the CBE regulation. Without that threshold showing, the court held that the warning-based claims were preempted. The word “impossibility” should not conceal the fact-specific work underneath it. The analysis can depend on the precise drug, the approved label, the date of use, the risk asserted, what had been submitted to the FDA, what later evidence actually showed, and the warning change proposed. It is therefore unsafe to treat Combs as a broad immunity decision for all medical-product manufacturers. It is equally unsafe to treat it as a purely technical federal rule that can be ignored in an Oklahoma injury case. The decision places the work where it belongs: on the documents, scientific evidence, and regulatory history that establish whether a state-law duty could have been carried out without violating federal law.
VI. Oklahoma’s Presumption Is a Separate Proof Layer
The court gave an independent reason for affirmance that deserves its own attention. Under 76 O.S. § 57.2(A), an Oklahoma product manufacturer or seller may invoke a rebuttable presumption of non-liability for injury caused by a product’s formulation, labeling, or design when the applicable mandatory federal safety standard or regulation governed the risk and the product complied with or exceeded that requirement. Subsection B identifies ways a claimant may rebut the presumption, including by establishing that the relevant federal standards were inadequate to protect the public from unreasonable risks or that the manufacturer withheld or misrepresented material information relevant to the federal determination. In Combs, the court concluded that the plaintiff had not made the argument or proof necessary to rebut the presumption on summary judgment and therefore had waived the argument she later pursued on appeal. This feature of the opinion matters because it shows that federal approval and federal preemption are not interchangeable concepts. A lawyer may have to address both, along with the underlying elements of product defect, warning adequacy, and proximate cause. The statute itself also has boundaries: it does not extend to manufacturing flaws or defects merely because a manufacturer complied with federal quality-control practices, and its treatment of sellers is separately defined. A case should therefore begin with product identification and a theory map, not a slogan. Was the asserted defect a label, a design, a manufacturing departure, a distribution act, a professional decision, or some combination? Which federal standard is actually invoked? What proof is available to meet or rebut the statutory presumption? These are questions to answer from the record before a motion makes them urgent.
VII. The Prescriber Still Matters to Causation
Even a legally viable warning theory must still prove causation. Prescription-drug litigation carries a familiar but demanding feature: the warning ordinarily runs through a prescribing or treating professional rather than directly from the manufacturer to the patient. Combs treated that point through the learned-intermediary framework and cited Eck v. Parke, Davis & Co., 256 F.3d 1013 (10th Cir. 2001). The court assumed the plaintiff could have the benefit of a rebuttable presumption that an adequate warning would have been read and heeded by the physician. But the manufacturer offered testimony that the treating physician would not have changed the course of treatment even with later knowledge about the alleged risk. The court held that the plaintiff had not produced evidence creating a material factual dispute on that causation question. This part of the decision should guide investigation, not discourage it. A warning case needs more than a comparison between the old label and a hypothetical new paragraph. It needs evidence about the actual medical decision: why the procedure was ordered, what alternatives were reasonably available, what information the clinician had at the time, what the clinician would have done with an additional warning, and whether a different decision would likely have avoided the claimed harm. Deposition testimony can be decisive, but it must be tested against contemporaneous records, standard practice, the patient’s condition, the stated purpose of the procedure, and the scope of the proposed warning. The causal chain is not a formality. It is how a court separates an arguable information deficit from a compensable injury legally caused by that deficit.
VIII. What the Court Did Not Decide
Every appellate opinion has edges, and Combs should be read inside its own. The court affirmed summary judgment on the warning-based theories before it. It did not find that the plaintiff’s medical experience was unreal, decide the general medical status of every condition discussed in the filings, determine that a different drug or label could never present a viable warning claim, or resolve the merits of an unbriefed design-defect theory. It also did not turn a defendant’s regulatory compliance into a universal answer to every product case. The opinion noted that the plaintiff’s remaining legal theories were not preserved in the summary-judgment response or on appeal, which is a procedural boundary rather than a substantive finding on every conceivable claim. These limits are not technical footnotes. They explain why lawyers must avoid borrowing a result from one product, one label, one record, and one procedural posture as a prediction for another. A current label may be different from a historical label; a manufacturer’s FDA submissions may differ; a later study may reveal a risk of different severity or frequency; a prescriber may give different testimony; and a plaintiff may preserve arguments that the litigant in Combs did not. The disciplined position is neither automatic optimism nor automatic defeat. It is a fact-specific inquiry that identifies what the court actually held, what remains open, and what evidence would decide the next case. Restraint is not weakness in legal analysis. It is the form of precision that keeps an argument credible when the record is examined under oath.
IX. A Warning Claim Needs a Dated Evidence Map
The practical value of Combs is its insistence on a dated evidence map. The map should begin with the product and treatment event: the precise product name, lot information if available, date of administration, medical indication, consent materials, and the label or prescribing information in effect at that time. It should then identify the claimed risk and build backward from the legally relevant date. What peer-reviewed research, adverse-event reports, safety communications, FDA submissions, internal analyses, regulatory correspondence, or expert opinions existed before that date? What did each item actually say? Was it new information or a new analysis of prior information? Did it reveal a different type of risk, or a greater severity or frequency? Did it establish reasonable evidence of a causal association rather than merely identify a question for future research? The map must also preserve the causation branch: the treating professional’s reasoning, alternatives, likely response to the proposed warning, and the medical basis for linking the decision to the claimed injury. An evidence map does not manufacture a claim. It exposes the claim’s real architecture early enough to preserve what can be preserved and to avoid investing years in a theory that cannot satisfy the federal and state proof layers. This is the opposite of a cookie-cutter checklist. It is a chronology of authority, science, medical judgment, and proof—one that tells a lawyer where the record is strong, where it is uncertain, and what discovery is needed before a confident conclusion is possible.
X. Preservation Starts Before the Litigation Narrative Hardens
Once a medical-product injury is being evaluated, the basic records should be preserved before the case story turns into a fixed narrative. The starting point is usually the unedited medical record: ordering notes, imaging or procedure documentation, medication-administration records, informed-consent materials, discharge instructions, subsequent care, laboratory or diagnostic results, and the product identification in the facility’s files. It may also include pharmacy or distributor records, communications about an adverse event, copies of contemporaneous labeling, and any available regulatory material that existed at the relevant time. The purpose is not to turn every medical concern into a lawsuit. It is to prevent a later fight over what happened, which product was used, what the clinician knew, or what information was available when the choice was made. The same discipline protects a defense as well as a claim: a manufacturer, hospital, clinician, and patient are all better served when the record is preserved in its original sequence. In a plaintiff-side review, the task is to investigate whether the law recognizes a remedy without flattening medicine into accusation. That means separating confirmed medical facts from allegations, treating a case-specific expert opinion as distinct from a regulatory conclusion, and refusing to promise an outcome before the evidence is complete. Combs reminds us that litigation may turn on a record assembled years earlier. The person facing a possible claim needs the truth of that record, not a premature slogan about what it must prove.
XI. The Law Does Not Excuse Anyone From Looking Closely
The human temptation after an adverse medical event is to search for a simple answer: someone should have warned, someone should have acted, someone should pay. The legal system is not entitled to answer that question with indifference, but it cannot answer it honestly without the patient’s actual record, the product’s actual label, the science available at the relevant time, and the causal chain that connects the alleged omission to the injury. Combs supplies no shortcut. It instead explains why proof matters at every step: a manufacturer’s ability to change a label, the effect of federal regulation, the Oklahoma statutory presumption, and the treating professional’s decision. For lawyers, the decision is a warning against vague pleading and after-the-fact reconstruction. For families, it is a reason to preserve documents and seek a careful evaluation before evidence disappears. For courts, it is a reminder that the difference between a compelling narrative and a legally sufficient case may rest on a dated document, a defined scientific proposition, and a witness who can explain what would have changed. Hicks Law Firm evaluates serious Oklahoma injury matters from the governing law and the available record. This article is general legal information, not medical or legal advice, and it does not predict the result of any individual case. A right without proof can vanish in silence; proof without disciplined analysis can mislead. The work is to hold both together.
If a serious injury may involve a medication, medical device, or other product, preserve the treatment records, product information, labeling, and communications while their timeline can still be verified. Contact Hicks Law Firm at (405) 759-0515 or through the contact page. Do not send confidential details until an attorney-client relationship has been established. Every matter depends on its own facts, medical record, parties, evidence, and governing law.