dissenting:
This case boils down to a choice of whether the Federal Drug Administration’s (“FDA”) specification that a potentially deadly medical device, an internally implanted pump, “maintain[ ] the delivery accuracy of ± 15%” (J.A. 93) is a requirement, or whether it is a mere aspirational figure. Because I believe that the accuracy rate specification is indeed a requirement, and because the medical device’s failure to meet that requirement underlies Plaintiff Sherry Walker’s state law tort claims, I would hold that her claims are not preempted under Riegel v. Medtronic, Inc., 552 U.S. 312, 128 S.Ct. 999, 169 L.Ed.2d 892 (2008). I therefore respectfully dissent.
In Riegel, the Supreme Court expressly stated that state tort suits are preempted only to the extent that they impose requirements “different from, or in addition to” the requirements imposed by federal law. 552 U.S. at 321, 128 S.Ct. 999. The Supreme Court affirmed a Second Circuit opinion that emphasized at the outset that
our preemption analysis is quite limited in scope, affecting the small universe of cases resting on claims alleging liability despite a [premarketj-approved device’s adherence to the standards upon which it secured FDA premarket approval. We take care to explain that we do not hold that all state tort claims as to [premarketj-approved devices are preempted. Thus, tort claims that are based on a manufacturer’s departure from the standards set forth in the device’s approved [Pre-Market Approval] applica*582tion — such as the Riegels’ negligent manufacturing claim — are not preempted.
Riegel v. Medtronic, Inc., 451 F.3d 104, 106 (2d Cir.2006) (emphasis added).
The Supreme Court established a two-step analysis for determining if state-law claims are preempted. First, a court must determine whether “the Federal Government has established requirements applicable to” the particular medical device. Riegel, 552 U.S. at 321, 128 S.Ct. 999. Claims involving a Class III medical device with Pre-Market Approval automatically clear step one of the analysis because the Pre-Market Approval process establishes specific requirements applicable to the particular device. Id. at 322-23, 128 S.Ct. 999. In the second step, a court must determine whether the state law claims are based on requirements “different from or in addition to” the federal requirements relating to safety and effectiveness. Id. at 323, 128 S.Ct. 999 (quotation marks omitted).
In my view, Walker’s suit clears both of these hurdles. The Medtronic Syn-chroMed EL Infusion Pump (“Pump”) at issue in this case is undisputedly a Class III device with specific requirements. The first Riegel step is therefore met. Further, the Pre-Market Approval established a plus or minus 15 percent delivery accuracy requirement for the internally-implanted Medtronic SynchroMed EL Infusion Pump (“Pump”). Walker’s suit is based on the Pump’s failure to meet that requirement. Indeed, she proffered evidence, which we must take as true for summary judgment purposes, that her husband, Arnold Walker, died because the Pump delivered an extreme overdose of 258 percent of his prescribed medicine dosage. Walker’s suit therefore does not impose additional or different requirements but rather parallels the pre-existing federal requirements for the Pump. I therefore do not believe that her suit is preempted under Riegel.
The majority casts the Federal Drug Administration’s (“FDA”) requirement that the Pump “maintain[] the delivery accuracy of ±15%” (J.A. 93) as a mere aspirational figure, i.e., what “the device is likely to achieve under optimal circumstances.” Ante at 578. I cannot agree. And, apparently, neither would the FDA. In an amicus brief filed in Horn v. Thoratec Corp., 376 F.3d 163 (2004), in which the Third Circuit addressed whether state claims were preempted, the FDA stated that “the agency’s approval of this [Class III] device through the PMA process does impose specific requirements for its design, manufacturing, performance, labeling, and use.” Brief for the United States as Amicus Curiae at 15, Horn v. Thoratec Corp., 376 F.3d 163 (3d Cir.2004) (No. 02-4597), 2004 WL 1143720, at *15 (emphasis added). Further, “in reviewing a PMA, FDA considers in great depth and detail the performance and design specifications, methods of manufacture, labeling, and indications for use of a proposed medical device.” Amicus Br. at 16. Nowhere in its 31-page brief did the FDA even mention the words “performance standard.” Nevertheless, the FDA repeatedly made clear that it can and does exact performance specifications and requirements in the ordinary course of the Pre-Market Approval process — which is precisely what we have in this case.
The FDA underscored that “although FDA does not itself design any medical devices, through the PMA approval process it certainly establishes ‘specific requirements’ applicable to a ‘particular device’ because the specifications for that device’s design, performance, manufacture, labeling, and use are approved by the agency based on what the applicant sub*583mits. See Kemp v. Medtronic, Inc., 231 F.3d at 216.” Id. at 16 (emphasis added). And, notably, “FDA does not, and has never, used notice-and-comment regulations ” — and formal performance standards are just such regulations, as the majority recognizes — “to approve individual products or to establish product-specific requirements for manufacture, performance, labeling, and use. Rather, a PMA order is better conceptualized as an individual adjudication that imposes ‘specific requirements’ on the device.” Id. at 23-24 (emphasis added).
Accordingly, the accuracy rate at issue here is a requirement for the Pump per the Pre-Market Approval. The accuracy rate of “delivery accuracy of ±15%” is included twice in the PMA approval, which itself was based on information that Medtronic provided and that included the figure. J.A. 93-94. Medtronic itself argued in a previous appeal before the Sixth Circuit that “the FDA’s [Pre-Market] [Approval and the Conditions of Approval, taken together, ... establish! ] the specific federal requirements for the” device at issue. Kemp v. Medtronic, 231 F.3d 216, 228 (6th Cir.2000) (quotation marks omitted). The Sixth Circuit agreed with Medtronic’s position — and so do I. See id.; see also Horn, 376 F.3d at 171 (“The FDA, when PMA approval is granted, imposes federal requirements based on the highly detailed and prescriptive nature of the PMA process and the approval order that results from it.”). Indeed, even in this litigation, Medtronic conceded that an overinfusion of medication above 15 percent would violate the Pump’s Pre-Market Approval. Medtronic’s corporate representafive for purposes of this case, Patrick L. Johnson, Regulatory Affairs Director for the Medtronic Neurological Business, testified in his deposition as follows:
Q: Okay. Let me rephrase it then. Assuming there are no procedural errors, the physician hasn’t done anything wrong and Mr. Walker’s pump has been used within the environmental conditions that are spelled out in the technical manual and the pump because of a system complication, a system error overinfuses Mr. Walker by more than 15 percent, that would be a violation of the PMA approval for this pump, would you agree with that?
A: That would be a reasonable statement, yes.
S.J.A.1 248 (objection to form omitted). That is precisely what Walker is claiming.2
The majority notes that “neither the FDA’s initial grant of premarket approval nor its later approvals of supplemental applications [for the Pump] were subject to any requirement that [the Pump] comply with a formal performance standard pursuant to § 360d.” Ante at 574. The majority then goes on, however, to assert, without citation to any support, that “only such a performance standard could create the type of binding requirement that would make Walker’s claims impose requirements parallel to, as opposed to more restrictive than, those imposed by the FDA.” Ante at 578. I cannot agree that the absence of a formal performance standard for this Class III device frees the manufacturer from accountability for ensuring that *584its device performs as approved, including with regard to the approved accuracy rate.
It is instructive to consider what a formal performance standard is. “Medical devices are developed individually, but classes of products with similar features can sometimes benefit from a uniform standard of safety and effectiveness. These are the Class II devices.” James T. O’Reilly, 1 Food & Drug Admin. § 18:18 (Sd ed. 2011). Performance standards are a form of special control that may be instituted when “[t]he FDA ... make[s] a finding that a performance standard is necessary to provide reasonable assurance of the device’s safety and efficacy.” Id. (quotation marks omitted).
Notably, however, to obtain Pre-Market Approval, a Class III device manufacturer must always “provide[] reasonable assurance that the product is safe and effective under the conditions of use for which it is labeled....” Id. at § 18:50. This is echoed in 21 C.F.R. § 860.7, the regulation dedicated to the “[d]etermination of safety and effectiveness[,]” which equates the establishment of performance standards for Class II devices with the Pre-Market Approval of Class III devices. The regulation states:
In determining the safety and effectiveness of a device for purposes of classification, establishment of performance standards for class II devices, and premarket approval of class III devices, the Commissioner and the classification panels will consider the following, among other relevant factors:
(1) The persons for whose use the device is represented or intended;
(2) The conditions of use for the device, including conditions of use prescribed, recommended, or suggested in the labeling or advertising of the device, and other intended conditions of use;
(3) The probable benefit to health from the use of the device weighed against any probable injury or illness from such use; and
(4) The reliability of the device.
21 C.F.R. § 860.7(b) (emphasis added).
Stated differently, as part of every PreMarket Approval for a Class III device (in contrast to a Class II device, where these concerns come into play specifically in establishing formal performance standards), the FDA must evaluate “safety and effectiveness” by considering, among other things, “[t]he reliability of the device” and the device’s potential injury versus potential benefit. Id. Given that even the FDA equates a safety and effectiveness determination, including consideration of device reliability, in Pre-Market Approvals for Class III devices with performance standards for Class II devices,31 cannot agree with the majority that “[b]ecause the plus or minus 15 percent specification here was not a formal performance standard, ... Walker cannot claim that ongoing adherence to it was a requirement of the [Pump’s] premarket approval.” Ante at 579. Indeed, it seems illogical that the FDA would place more exacting performance requirements on Class II devices such as “surgical drapes” (Riegel, 552 U.S. at 316, 128 S.Ct. 999) than on Class III devices such as “replacement heart valves, implanted cerebella stimulators, and pacemaker pulse generators,” for which “a less stringent classification would provide [no] reasonable assurance of safety and effectiveness .... ” Id. at 317, 128 S.Ct. 999.
*585Nor do I believe that Medtronic’s inclusion of various warnings in its Pre-Market Approval materials neuters the “maintaining the delivery accuracy of ±15%” requirement. Those warnings address, e.g., potential component failure, conditions under which the device might malfunction (such as at high altitudes), and human error. Those general warnings do not change the fact that the specific plus or minus 15 percent accuracy rate requirement is twice stated in the FDA’s two-page Pre-Market Approval letter, which, per Medtronic, establishes a device’s requirements. See Kemp, 231 F.3d at 228. Further, Medtronic has not alleged that any number of the things addressed by the warnings are in play here (e.g., Medtronic does not contend that Arnold Walker went hiking in the Himalayas).
To be sure, it cannot be said that, under my analysis, device manufacturers must achieve perfection. Indeed, “the delivery accuracy of ±15%” itself recognizes that manufacturers may be given some margin for error. The FDA accepted that margin, based on Medtronic’s Pre-Market Approval application, to be plus or minus 15 percent. Perhaps potential liability for incidents that occur outside of that 30 percent margin for error is simply a cost of doing business. The ramifications of holding anything else are indeed serious. If manufacturers are not held accountable for having their devices perform as approved by the FDA, they will have little incentive to ensure that potentially deadly devices function properly. For example, if Walker’s factual contentions are true (and we must presume that they are in reviewing summary judgment against her (Nat'l City Bank of In. v. Turnbaugh, 463 F.3d 325, 329 (4th Cir.2006))), the Pump, which was approved with a dosage accuracy rate of plus or minus 15 percent, instead infused her husband with 258 percent of the appropriate medication dosage, and this extreme overdose killed him. I fail to see why — in the absence of clear congressional instruction to the contrary — the Walkers alone should bear the burden of this malfunction.
My view is bolstered by the general presumption against preemption of state law claims. As this Court articulated in a 2008 opinion, “we begin our consideration with the basic premise that ‘Congress did not intend to displace state law.’ Maryland v. Louisiana, 451 U.S. 725, 746, 101 S.Ct. 2114, 68 L.Ed.2d 576 (1981).... The presumption against preemption has particular force in the areas of public health and safety that have traditionally been regulated by the states. See Pinney v. Nokia, Inc., 402 F.3d 430, 457 (4th Cir.), cert. denied, 546 U.S. 998, 126 S.Ct. 551, 552, 163 L.Ed.2d 499 (2005).” City of Falls Church v. Fairfax Cnty. Water Auth., 272 Fed.Appx. 252, 256 (4th Cir. 2008) (unpublished). Particularly when viewed through this lens, I fail to see how Walker’s claims are preempted.4
Of course, whether or not Walker can prevail on her claims remains a very open question. But at stake today is not whether Walker should win or Medtronic should be held liable. The question before us is simply whether Walker may pursue her *586state law tort claims in light of the FDA’s regulation of the Pump. I believe that she should be able to do so. Accordingly, I respectfully dissent.
. Certain case materials containing confidential medical information relating to Arnold Walker were filed under seal. While this deposition excerpt is found in the sealed joint appendix, it does not pertain to confidential medical information.
. I also note that even Medtronic’s appellate counsel conceded at oral argument that delivery of more than plus or minus 15 percent of the appropriate dosage of medicine would constitute a Pump "malfunction.”
. I of course recognize that the FDA ‘‘may " (and therefore necessarily need not) establish a performance standard for Class III devices "as a condition to premarket approval....” 21 C.F.R. § 861.1(b) (emphasis added).
. The majority appears to turn the presumption against preemption on its head when it asserts "That a specification is included in a device's premarket approval application, however, does not indicate that the specification imposes a binding requirement for the life of the device — at least not without clear indication by the FDA through promulgation of a formal performance standard.” Ante at 579 n. 5. Under the majority’s analysis, then, in clear contradiction to precedent, we assume preemption, i.e., additional requirements via state tort suits, absent clear indication to the contrary. With this, I cannot agree.