DENNIS JACOBS, Circuit Judge:

Dr. Jesse Polansky appeals from a partial final judgment of the United States District Court for the Eastern District of New York (Cogan, J.), dismissing his False Claims Act ("FCA") and state analog causes of action. We have appellate jurisdiction by virtue of certification. Polansky alleges that in and after 2002: his former employer, defendant Pfizer, Inc. ("Pfizer"), improperly marketed Lipitor, a popular statin, as appropriate for patients whose risk factors and cholesterol levels fall outside the National Cholesterol Education Program Guidelines ("NCEP Guidelines" or "Guidelines"); that the Guidelines are incorporated into and made mandatory by the drug's label; and that Pfizer thus induced doctors to prescribe the drug, pharmacists to fill the prescriptions, and federal and state health care programs to pay for "off-label" prescriptions. Judge Cogan dismissed the claims because he determined that the FDA's approval of Lipitor was not dependent upon compliance with the Guidelines.1 We affirm.

BACKGROUND

A

The Food, Drug and Cosmetic Act ("FDCA") forbids pharmaceutical manufacturers from marketing or selling a drug until the Food and Drug Administration ("FDA") has approved it as safe and effective for its intended use or uses (the drug's "indications"). See 21 U.S.C. § 355(a), (d); United States v. Caronia, 703 F.3d 149, 152-53 (2d Cir.2012); see also 21 U.S.C. § 393(b)(2)(B). The exact wording of the drug's "label" must be approved by the FDA, and thereafter generally cannot be altered without further approval. Wyeth v. Levine, 555 U.S. 555, 568, 129 S.Ct. 1187, 173 L.Ed.2d 51 (2009); see 21 U.S.C. § 355(b)(1)(F), (d); 21 C.F.R. §§ 314.105(b), 601.12. The label (which can be quite lengthy) must include, inter alia, the drug's indications, contra-indications, limitations of use, use by specific populations, and dosage instructions. 21 C.F.R. § 201.57.

"Once FDA-approved, prescription drugs can be prescribed by doctors for both FDA-approved and -unapproved uses; the FDA generally does not regulate how physicians use approved drugs." Caronia, 703 F.3d at 153; see also 21 U.S.C. § 396 (principle of non-interference with the practice of medicine). "Indeed, courts and the FDA have recognized the propriety and potential public value of unapproved or off-label drug use." Caronia, 703 F.3d at 153 (citing cases and FDA draft guidance). However, pharmaceutical manufacturers are generally prohibited from promoting off-label uses of their products if the off-label marketing is false or misleading, or if it evidences that a drug is intended for such off-label use and is therefore "misbranded."2

Polansky contends that prescriptions written for off-label uses are generally not reimbursable by federal and state health care programs. Federal reimbursement for prescription drugs under Medicare and Medicaid is generally limited to drugs prescribed for FDA-approved (on-label) uses or for certain purposes included in any of three drug compendia. See 42 U.S.C. § 1396r-8(k)(2), (3), (6); id. § 1395w-102(e)(1), (4). State Medicaid programs "may exclude or otherwise restrict coverage" if a drug is prescribed off-label unless included in any of those compendia. Id. § 1396r-8(d)(1)(B)(i).

B

Lipitor (atorvastatin calcium) is a popular statin, a drug that lowers cholesterol levels by blocking enzymes essential to cholesterol production. Broadly speaking, Lipitor is approved for treatment of elevated cholesterol, and for prevention of cardiovascular disease. During the time period relevant to this case, Lipitor was approved for five "indications" relating to treatment of elevated cholesterol.3

Polansky alleges that Lipitor's approved use is more narrow than these specific indications: that it is approved only when the patient's risk factors and cholesterol levels fall within a framework outlined in the NCEP Guidelines, and that any use by a patient outside that framework is unapproved and off-label. He further alleges that Pfizer widely marketed Lipitor for outside-Guidelines use,4 causing physicians to write Lipitor prescriptions for patients whose risk factors and cholesterol levels fell outside the Guidelines framework, and causing those prescriptions to be submitted for reimbursement by federal and state health care programs. Because government health care programs generally do not reimburse prescriptions for off-label use, see supra, he contends that these requests for reimbursement impliedly certified (falsely) that the prescription was for an on-label use, and thus constituted false claims under the FCA and state laws.

The Guidelines were promulgated in 2001 by the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults, under the aegis of the National Heart, Lung, and Blood Institute of the National Institutes of Health.5 The Guidelines were "recommendations for cholesterol testing and management," J.A. 32, and cautioned that they were advisory only:

J.A. 33. The full Guidelines report is nearly 300 pages long.

The Guidelines recommended a focus on lowering LDL (low-density lipoprotein) cholesterol. Patients were grouped on the basis of their risk for coronary heart disease events. Each of the three risk categories was accorded (1) an LDL cholesterol therapeutic "goal"; (2) an LDL level at which to initiate therapeutic lifestyle changes; and (3) an LDL "cutpoint" at which to consider drug therapy. (The particulars are in the margin.6)

The Lipitor label changed somewhat during the time between the Guidelines' promulgation in 2001 and the filing of the operative complaint in 2010; but the only difference material to the outcome of this case is one made in June 2009. Pre-2009 Lipitor labels included a table summarizing the recommendations of the NCEP Guidelines (see footnote 6); but this table does not appear anywhere in the 2009 label. Notwithstanding that change, the 2009 label is substantively the same as its pre-2009 iteration7: any substantive modifications had to be specifically disclosed in the new, 2009 label, see 21 C.F.R. § 201.57(a)(5); no such modifications were listed in the 2009 label, which was submitted to and approved by the FDA; and it was agreed by the parties during district court proceedings that the 2009 label did not effect any substantive changes.8

Pre-2009 labels referred to the NCEP Guidelines in two sections: "Indications and Usage" and "Dosage and Administration." The "Indications and Usage" section enumerated the drug's five indications relating to treatment of elevated cholesterol, and then added:

E.g., J.A. 718-19 (2005 label). The Guidelines summary table followed.

The "Dosage and Administration" section of pre-2009 labels contained four-to-six patient subcategories (depending on the date); in one of these patient subcategories, the labels referenced the Guidelines: "The starting dose and maintenance doses of LIPITOR should be individualized according to patient characteristics such as goal of therapy and response (see NCEP Guidelines, summarized in Table [above])." E.g., J.A. 729 (2005 label).

In the 2009 label, the summary table does not appear. The Guidelines are not mentioned at all in the "Indications and Usage" section (even though that section includes a restriction based on LDL level for patients between the ages of ten and seventeen, as did pre-2009 labels). The "Dosage and Administration" section retains the same parenthetical reference to the Guidelines as did earlier labels. The Guidelines appear nowhere else in that label.

C

The FCA is an anti-fraud statute; accordingly, Polansky must plead fraud with particularity pursuant to Federal Rule of Civil Procedure 9(b). See Gold v. Morrison-Knudsen Co., 68 F.3d 1475, 1476 (2d Cir.1995) (per curiam). Judge Korman, to whom this case was originally assigned, reviewed the previous version of the complaint, and dismissed the FCA and state analog claims for failure to satisfy Rule 9(b)'s requirements. United States ex rel. Polansky v. Pfizer, Inc., 04-CV-0704 (ERK), 2009 WL 1456582, at *5-10 (E.D.N.Y. May 22, 2009) ("Polansky I").

Polansky then filed the operative complaint, seeking to cure the defects Judge Korman had identified. Judge Cogan, to whom the case was reassigned, again dismissed these claims, but on the different ground that Pfizer had not engaged in off-label marketing as a matter of law and therefore could not have caused false claims to be submitted. United States ex rel. Polansky v. Pfizer, Inc., 914 F.Supp.2d 259, 266 (E.D.N.Y.2012) ("Polansky II"). Judge Cogan determined that the premise underlying Polansky's FCA claims — that the label required compliance with the Guidelines — was implausible. See Bell Atl. Corp. v. Twombly, 550 U.S. 544, 570, 127 S.Ct. 1955, 167 L.Ed.2d 929 (2007) (complaint must provide "enough facts to state a claim to relief that is plausible on its face" to survive motion to dismiss).

DISCUSSION

As relevant here, the FCA imposes liability on any person who "knowingly presents, or causes to be presented, a false or fraudulent claim for payment or approval" to the U.S. government; or who "knowingly makes, uses, or causes to be made or used, a false record or statement material to a false or fraudulent claim." 31 U.S.C. § 3729(a)-(b). Polansky's theory of FCA liability necessarily rests on the contentions that (1) the NCEP Guidelines were incorporated into Lipitor's FDA-approved label and made mandatory, and (2) requests for reimbursement of Lipitor prescriptions impliedly certified (falsely) that the prescription was for an on-label use.9

Pfizer urges that we can affirm on any of several alternative grounds: that the Guidelines are not incorporated into the Lipitor label, as held by Judge Cogan in Polansky II; that the operative complaint failed to cure the Rule 9(b) particularity defect identified by Judge Korman in Polansky I; or that no false claims have been alleged because the complaint fails to plausibly allege that requests for reimbursement impliedly certified on-label use. We expressly endorse and adopt Judge Cogan's carefully considered and thorough analysis, and affirm on that basis.

As Judge Cogan explained, "guidelines" usually provide advice and (unsurprisingly) guidance, "not mandatory limitation." Polansky II, 914 F.Supp.2d at 262. The NCEP Guidelines themselves expressly disclaimed prescriptive force: Their "general guidance" "need not hold for individual patients" and "should not override" a physician's clinical judgment about appropriate treatment of a particular patient. J.A. 33.

The district court drew the proper inferences. "Once the doctor's clinical judgment is introduced as the determinative factor in the decision making process, it must be apparent that this data serves as a recommendation, not a limitation or prohibition." Polansky II, 914 F.Supp.2d at 264-65. We "cannot accept plaintiff's theory that what scientists at the National Cholesterol Education Program clearly intended to be advisory guidance is transformed into a legal restriction simply because the FDA has determined to pass along that advice through the label." Id. at 265.

Where the label imposes cholesterol-level restrictions, it does so only for pediatric patients. This express restriction makes "more conspicuous" the absence of a similar restriction for adults, id. at 263, and shows how easily the FDA could have mandated compliance with the NCEP Guidelines with respect to all patients if it wanted to do so.

This conclusion is reinforced by the 2009 label. Notwithstanding that it is substantively identical to the prior version (as a matter of administrative procedure), it omits the Guidelines table, makes no more than fleeting reference to the Guidelines, and fails to mention them at all in the "Indications and Usage" section of the label, which is where a limitation on approved "usage" would be expected to appear. "A person reading the `Indications and Usage' section of the 2009 label must come away with one clear meaning; the drug is to be used if a physician believes his patient should lower his cholesterol. That is the drug's essential purpose as defined by the label — to lower cholesterol." Id. For further particulars of the analysis, the reader is referred to Judge Cogan's opinion, which we adopt.

Because we affirm on that basis, we need not wade into the circuit split regarding whether, to satisfy Rule 9(b), an FCA relator alleging a fraudulent scheme must provide the details of specific examples of actual false claims presented to the government (which Polansky does not do). (That split is detailed in the margin.10) Nor need we decide whether Polansky has adequately alleged that a request for prescription-drug reimbursement is an implied certification of on-label use.

We are skeptical, however, that even under Polansky's theory of the case, anyone could be identified who actually submitted a false claim. "[T]he FDA does not prohibit physicians, who are free to do so, from prescribing Lipitor for patients with normal cholesterol." Polansky I, 2009 WL 1456582, at *10. Accordingly, it is unclear just whom Pfizer could have caused to submit a "false or fraudulent" claim: The physician is permitted to issue off-label prescriptions; the patient follows the physician's advice, and likely does not know whether the use is off-label; and the script does not inform the pharmacy at which the prescription will be filled whether the use is on-label or off. We do not decide the case on this ground, but we are dubious of Polansky's assumption that any one of these participants in the relevant transactions would have knowingly, impliedly certified that any prescription for Lipitor was for an on-label use. Cf. id. at *7 ("[B]ecause the FDA has expressly advised physicians that, `unlabeled uses may be appropriate and rational in certain circumstances, and may, in fact, reflect approaches to drug therapy that have been extensively reported in medical literature,' and because physicians `commonly exercise professional medical judgment and prescribe drugs for uses not within the indications articulated by the FDA,' the entities to which reimbursement claims are made could hardly be understood to have operated on the assumption that the physician writing the prescription was certifying implicitly that he was prescribing Lipitor in a manner consistent with the Guidelines." (citations omitted)).

"The False Claims Act, even in its broadest application, was never intended to be used as a back-door regulatory regime to restrict practices that the relevant federal and state agencies have chosen not to prohibit through their regulatory authority." Polansky II, 914 F.Supp.2d at 266. It is the FDA's role to decide what ought to go into a label, and to say what the label means, and to regulate compliance. We agree with Judge Cogan that there is an important distinction between marketing a drug for a purpose obviously not contemplated by the label (such as, with respect to Lipitor, "to promote hair growth or cure cancer") and marketing a drug for its FDA-approved purpose to a patient population that is neither specified nor excluded in the label. Id. at 265. An FCA relator alleging off-label marketing might be able to satisfy Rule 9(b) and surmount the impediment of implied certification in a case in which it would be obvious to anyone that the use promoted is one that is not approved; but this is emphatically not such a case.

For these reasons, the partial judgment is AFFIRMED.