This case is one of several thousand involving the promotion, sale, and use of pedicle screw fixation devices to treat spinal injuries. Appellant William Cooper brought this lawsuit against Smith & Nephew, Inc., claiming that its defective device was responsible for his failed back surgeries and the accompanying deleterious side effects. The district court dismissed Cooper's claims after determining that Cooper had no admissible medical evidence indicating that Smith & Nephew's device was the proximate cause of his injuries. Because the district court did not abuse its discretion in excluding the testimony of Cooper's expert on medical causation, we affirm the judgment.

Appellee Smith & Nephew, Inc., ("S&N;"), manufactures the Rogozinski Spinal Rod System. The Rogozinski System is a medical device consisting of connectors, rods, bone screws, and hooks. Orthopedic surgeons use the Rogozinski System to aid them in performing spinal fusion surgeries. This type of surgery stabilizes the spine by fusing two or more vertebrae together into a single bone. The Rogozinski System is used to promote a successful fusion between vertebrae by providing stability for the spine.

Since June 25, 1990, the FDA has cleared the Rogozinski System for sale for some spinal applications. Prior to the initiation of this lawsuit, the Rogozinski System had not been approved for pedicle fixation. The pedicles are two pieces of bone on each side of the spinal vertebrae. Pedicle fixation is the process by which bone screws are inserted through the pedicles in order to provide increased spinal stability, thereby promoting fusion.

Once the FDA has cleared a device for introduction into the stream of commerce, physicians may use the device in any manner they determine to be best for the patient, regardless of whether the FDA has approved the device for this usage. This practice by physicians is known as "off-label" usage. See Buckman Co. v. Plaintiffs' Legal Comm., 121 S. Ct. 1012, 1018 (2001). Although the FDA had not approved the Rogozinski System for pedicle fixation prior to 1998, many physicians engaged in the off-label usage of the Rogozinski System for this purpose. Indeed in 1998, the FDA adjusted its assessment of devices capable of lumbar pedicle screw fixation, stating that "FDA believes that there is reasonable assurance that pedicle screw spinal systems are safe and effective for certain intended uses." See 63 Fed. Reg. 40,025, 40,035-36 (July 27, 1998).

On January 28, 1990, appellant William Cooper slipped on a patch of ice, resulting in lower back and leg pain. An MRI performed on Cooper in March 1990 revealed herniated disks between lumbar vertebrae positions 3 and 4 ("L3-L4"), 4 and 5 ("L4-L5"), and lumbar position 5 ("L5") and sacral region 1 ("S1"). The MRI also revealed a compression of the L4 nerve root sheath. Cooper subsequently underwent two procedures in order to alleviate his pain. The hospital records from these surgeries revealed that Cooper had smoked one pack of cigarettes per day for 25 years.

Following his second surgery, Cooper experienced worsening pain in his lower back and legs. After further diagnosis, an orthopedic surgeon, James Murphy, M.D., determined that there was instability in Cooper's lumbar spine. On June 19, 1991, Dr. Murphy performed spinal fusion surgery on Cooper, inserting the Rogozinski System between the L3-L4, L4-L5, and L5-S1 vertebrae. The goal of the surgery was to obtain a fusion at all three levels of vertebrae.

Following this surgery, Cooper continued to experience back and leg pain. By May 1992, Dr. Murphy believed that there likely was a nonunion at all three levels of vertebrae. Subsequent X rays revealed that the right S1 pedicle screw appeared to have loosened. During this time, Dr. Murphy continued to urge Cooper to stop smoking.

On October 15, 1992, a second fusion surgery on Cooper revealed that fusion had indeed occurred at L3-L4 and L4-L5. Dr. Murphy removed the pedicle screws from these levels. The surgery also revealed, however, that there was a nonunion at the L5-S1 level. Dr. Murphy attempted to refuse L5-S1, using the same bone screws, but attaching new rods and cross-links. After this surgery, Cooper continued to experience pain. An X ray taken on June 15, 1993, revealed that a screw had cracked on the left side of S1. In addition, a fusion failed to occur at L5-S1 following the second surgery. In preparation for yet another surgery, Dr. Murphy wrote in Cooper's medical records: "I think the most important function if we are going to reoperate on [Cooper] is that he stop smoking because that is going to really have a great deal of effect on the success or failure of the procedure."

In February 1995, Dr. Murphy performed a final surgery on Cooper, in which he attempted to fuse L5-S1. Dr. Murphy also removed the broken screw at left S1, and inserted new Rogozinski screws and rods at L4 and S1. Although Cooper suffered some complications, the surgery resulted in a fusion at L5-S1.

On August 12, 1997, Cooper filed a nine-count complaint against S&N; and several other defendants. Cooper asserted claims for fraud, civil conspiracy, concert of action, fraudulent marketing and promotion, negligent misrepresentation, strict liability in tort, liability per se, negligence, and breach of implied warranty of merchantability. Pursuant to 28 U.S.C. S 1407, the suit initially was transferred to the multidistrict litigation in the United States District Court for the Eastern District of Pennsylvania. While in the multidistrict litigation, Cooper's conspiracy claims were dismissed. See In re Orthopedic Bone Screw Prods. Liab. Litig., 193 F.3d 781 (3d Cir. 1999). On August 11, 1999, the case was transferred back to the district court, with only the claims against S&N; remaining.

In the present litigation, Cooper retained William Mitchell, M.D., and Harold Alexander, Ph.D., to present expert testimony that the Rogozinski System was the proximate cause of Cooper's injuries. Dr. Alexander is an expert in the field of biomedical engineering. He is not a medical doctor. For this reason, Dr. Alexander was found unqualified in the multidistrict litigation to opine as to causation in any individual plaintiff or to testify outside the area of orthopedic bioengineering. See In re Orthopedic Bone Screw Prods. Liab. Litig., MDL No. 1014, 1997 WL 39583 (E.D. Pa. Jan 23, 1997).

Cooper retained Dr. Mitchell to serve as his medical expert on specific causation. Dr. Mitchell has been a practicing orthopedic surgeon for approximately 38 years, and regularly examines spinal patients. Dr. Mitchell has performed numerous spinal fusion surgeries throughout his career. While Dr. Mitchell has been trained in the use of spinal implants, he has never implanted them in his patients. It was Dr. Mitchell's opinion that the Rogozinski System was incapable of bearing the stress to which it was subjected in Cooper's body. Dr. Mitchell claimed that, as a result of this defect, the left S1 pedicle screw fractured. This in turn caused the non-union at L5-S1. Dr. Mitchell's ultimate conclusion was that all of Cooper's injuries were the result of the Rogozinski System's failure to provide spinal stability.

On November 20, 2000, the district court granted S&N;'s motion to exclude all of Dr. Mitchell's testimony on medical causation. The court assumed that Dr. Mitchell was qualified as an expert under Fed. R. Evid. 702. Nonetheless, the court agreed with the numerous other district courts that had excluded Dr. Mitchell's testimony as unreliable in similar bonescrew lawsuits. Because all of Cooper's claims required expert medical testimony that the Rogozinski System was the proximate cause of his injuries, and because Dr. Alexander was unqualified to opine as to specific causation, the court granted S&N;'s motion for summary judgment on all claims. Cooper now appeals.

Under Federal Rule of Evidence 702, trial judges act as gatekeepers to "ensure that any and all scientific testimony . . . is not only relevant, but reliable." Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 588 (1993).1 While Rule 702 was intended to liberalize the introduction of relevant expert evidence, courts"must recognize that due to the difficulty of evaluating their testimony, expert witnesses have the potential to `be both powerful and quite misleading.'" Westberry v. Gislaved Gummi AB, 178 F.3d 257, 261 (4th Cir. 1999) (quoting Daubert, 509 U.S. at 595). Therefore, a trial judge, faced with a proffer of expert scientific testimony, must conduct "a preliminary assessment of whether the reasoning or methodology underlying the testimony is scientifically valid and of whether that reasoning or methodology properly can be applied to the facts in issue." Daubert, 509 U.S. at 592-93. The proponent of the testimony must establish its admissibility by a preponderance of proof. See id. at 592 n.10.

The Supreme Court in Daubert identified several factors that may bear on a judge's determination of the reliability of an expert's testimony. See id. at 592-94. Those factors include: (1) whether a theory or technique can be or has been tested; (2) whether it has been subjected to peer review and publication; (3) whether a technique has a high known or potential rate of error and whether there are standards controlling its operation; and (4) whether the theory or technique enjoys general acceptance within a relevant scientific community. See Daubert, 509 U.S. at 592-94.

In Kumho Tire Co. v. Carmichael, 526 U.S. 137, 150 (1999), the Supreme Court noted that the factors discussed in Daubert were neither definitive, nor exhaustive. The Court explained that particular factors may or may not be pertinent in assessing reliability, depending on the nature of the issue, the expert's particular expertise, and the subject of his testimony. See id. The Court further emphasized that the objective of Daubert's gatekeeping requirement is to "make certain that an expert . . . employs in the courtroom the same level of intellectual rigor that characterizes the practice of an expert in the relevant field." See id. at 152.

Courts of appeals apply an abuse of discretion standard when reviewing a trial court's decision to admit or exclude expert testimony. See General Electric Co. v. Joiner, 522 U.S. 136, 138-39 (1997). The Supreme Court also has emphasized that"the trial judge must have considerable leeway in deciding in a particular case how to go about determining whether particular expert testimony is reliable." See Kumho Tire, 526 U.S. at 152.

As an initial matter, S&N; claims that Dr. Mitchell was not qualified to offer opinions as an expert in this case. Like the district court, we assume, without deciding, that Dr. Mitchell was qualified to testify regarding the medical cause of Cooper's injuries. However, we must still review for abuse of discretion the district court's determination that Dr. Mitchell's testimony was unreliable and, therefore, inadmissible.

Cooper contends that Dr. Mitchell conducted a differential diagnosis to determine the cause of his injuries. Cooper argues that since this court has approved differential diagnosis as a reliable methodology, see Westberry, 178 F.3d at 263, the district court abused its discretion in refusing to admit Dr. Mitchell's testimony.

In Westberry, this court explained that differential diagnosis is "a standard scientific technique of identifying the cause of a medical problem . . . by determining the possible causes for the patient's symptoms and then eliminating each of these potential causes until reaching one that cannot be ruled out or determining which of those that cannot be excluded is the most likely." 178 F.3d at 262. A reliable differential diagnosis typically, though not invariably, "is performed after physical examinations, the taking of medical histories, and the review of clinical tests, including laboratory tests." Id. (internal citations omitted). Cooper is correct that Westberry held that a reliable differential diagnosis provides a valid foundation for an expert opinion under Rule 702. See id. at 263. However, the district court in turn correctly held that Dr. Mitchell's differential diagnosis was not reliable.

"A reliable expert opinion must be based on scientific, technical, or other specialized knowledge and not on belief or speculation, and inferences must be derived using scientific or other valid methods." Oglesby v. General Motors Corp., 190 F.3d 244, 250 (4th Cir. 1999). In this case, Dr. Mitchell asserted what amounted to a wholly conclusory finding based upon his subjective beliefs rather than any valid scientific method. Dr. Mitchell has never implanted a pedicle screw device in his patients' spines because he believes them to be inherently dangerous. His position conflicts with that of the FDA and the majority of his colleagues in the American Academy of Orthopedic Surgeons who believe that the use of spinal instrumentation, including the use of pedicle screws, is the standard of care in the profession. See, e.g., 60 Fed. Reg. 51,946, 51,947 (Oct. 4, 1995) ("By mid-1992, FDA discovered that the use of pedicle screw spinal systems outside of approved IDE studies was widespread, and that pedicle screw fixation was considered to be the standard of care by the surgical community.").

Dr. Mitchell's unconventional views about the overall safety and efficacy of the use of pedicle screws in spinal surgeries is reflected in his testimony in this particular case. He opined that since the Rogozinski System is designed to prevent nonunion, it must stop motion in the spine 100 percent of the time or else it is defective. According to Dr. Mitchell, if a fusion results after implantation surgery, the device is not defective; if a fusion does not result, then the device is defective. Since a nonunion occurred in Cooper's case, Dr. Mitchell contended that the Rogozinski System therefore was defective and caused all of Mr. Cooper's subsequent injuries.

During his deposition, Dr. Mitchell explained his opinion to S&N;'s counsel as follows:

Q: Purely retrospective[ly] looking back,[Cooper] didn't get a fusion, the screw broke, therefore it's defective; right?

A: Right.

. . .

Q: Let's assume in patient A we do not get a fusion, one of the screws loosens and breaks. Okay?

A: Right.

Q: Let's assume in patient B we get a fusion and there's no loosening and no breakage of the screw. Now, in your opinion the Rogozinski device in patient A would be defective?

A: Right.

Q: But the Rogozinski device in patient B would not be defective; right?

A: Correct.

. . .

Q: Your opinion is that a screw in a lumbar pedicle must stop motion 100 percent of the time or it's defective.

A: Yes. . . .

Daubert requires more than an assertion that if there is a lack of surgical success, there is ipso facto a product defect, and hence causation is established. Dr. Mitchell's methodology simply failed to provide any medical evidence as to what caused Cooper's specific injuries. Rather, Dr. Mitchell seems to have inferred causation from the existence of a nonunion alone. However, this basis for finding causation makes little sense in light of Dr. Mitchell's own concession that a non-union is a well-known risk of spinal fusion regardless of whether instruments are used or not.2 In addition, Dr. Mitchell acknowledged that the more levels a physician attempts to fuse, the less likely it is that the patient obtains a successful fusion. In fact, Dr. Mitchell himself estimated Cooper's chances of a successful fusion at all three levels to be only 25 percent.

Furthermore, Dr. Mitchell failed to address S&N;'s claim that the fractured pedicle screw actually resulted from the nonunion rather than vice-versa. The package insert for the Rogozinski System warns that if a fusion fails to develop, the resulting motion in the spine may cause the screw or other components to break. And Dr. Mitchell stated that fusion usually occurs within three to six months. In this case, the pedicle screw fractured more than six months after Cooper's second fusion surgery, and approximately two years after the screw was first implanted. Dr. Mitchell's causation theory failed to address the possibility that the screw fracture was caused by the nonunion. The district court properly determined that Dr. Mitchell essentially had a set of boilerplate objections to the use of spinal implantation devices which he asserted regardless of the specific facts of the particular case. This unscientific approach is not a reliable basis for expert testimony.

Furthermore, S&N;'s experts provided several alternative explanations for the nonunion in Cooper's case. Primary among these was Cooper's smoking. The medical literature in peer-reviewed journals indicates that smoking increases the likelihood of a nonunion. The district court held that Dr. Mitchell's diagnosis was unreliable in part because Dr. Mitchell "summarily rejects evidence that Cooper's long history of smoking caused nonunion, even though Cooper's medical records contain repeated references to Cooper's smoking habit." Cooper contends that the district court erroneously focused on Dr. Mitchell's ultimate conclusion that smoking was not a cause rather than his methodology.

We disagree. As the Supreme Court has recognized,"conclusions and methodology are not entirely distinct from one another." Joiner, 522 U.S. at 146. A medical expert's opinion based upon differential diagnosis normally should not be excluded because the expert has failed to rule out every possible alternative cause of a plaintiff's illness. See Westberry, 178 F.3d at 265 (citations omitted). In such cases, the alternative causes suggested by a defendant normally affect the weight that the jury should give the expert's testimony and not the admissibility of that testimony. See id. at 265 (citations omitted). However, a "differential diagnosis that fails to take serious account of other potential causes may be so lacking that it cannot provide a reliable basis for an opinion on causation." Id. at 265 (citations omitted). Thus, if an expert utterly fails to consider alternative causes or fails to offer an explanation for why the proffered alternative cause was not the sole cause, a district court is justified in excluding the expert's testimony. See id. at 265-66 (citations omitted); see also In re Paoli R.R. Yard PCB Litig., 35 F.3d 717, 758 n.27 (3d Cir. 1994).

The record in this case reveals that the medical literature is replete with evidence that smoking can cause nonunions to occur. Cooper's own medical records indicate that he has been a pack-a-day smoker for approximately twenty-five years. And in July 1991, a month after his first fusion surgery, Cooper's medical records indicate that his smoking habit had increased to two packs a day. Dr. Murphy, Cooper's treating physician, wrote after the second fusion surgery that Cooper's "continued smoking history [ ] is probably the main etiology of his non-union." Dr. Murphy also wrote: "I think the most important function if we are going to re-operate on [Cooper] is that he stop smoking because that is going to really have a great deal of effect on the success or failure of the procedure."

In the face of the medical literature and Cooper's own records, Dr. Mitchell categorically dismissed any suggestion that Cooper's smoking was the cause of the nonunion. Dr. Mitchell stated that he rejected the medical articles as unpersuasive after he read just two of them. When asked if he should have considered more of the articles, Dr. Mitchell replied:

No, for the simple reason, after I read those two articles years ago, they didn't affect my prior judgment, and they still haven't, that smoking doesn't have anything to do with healing in spinal fusion. So even if there were ten more articles, I'm not going to change my mind about it.In this case, Dr. Mitchell did not identify specifically how he ruled out smoking and other potential causes of the nonunion. Simply asserting that he read two articles on smoking and rejected them as unpersuasive is insufficient. Such a practice offers no solid grounds for rejecting smoking as a cause of the nonunion, and renders his opinion that the Rogozinski System was the actual culprit little more than speculation. As the Supreme Court has repeatedly held, "nothing in either Daubert or the Federal Rules of Evidence requires a district court to admit opinion evidence that is connected to existing data only by the ipse dixit of the expert." Kumho Tire, 526 U.S. at 157 (quoting Joiner, 522 U.S. at 146).

In addition, Dr. Mitchell failed to conduct a physical examination of Cooper and did not speak with any of Cooper's treating physicians. In certain circumstances, a physician may reach a reliable differential diagnosis without personally performing a physical examination. See Westberry, 178 F.3d at 262; Kannankeril v. Terminix Int'l, Inc., 128 F.3d 802, 807 (3d Cir. 1997). However, Kumho Tire emphasizes that the purpose of Rule 702's gatekeeping function is to"make certain that an expert . . . employs in the courtroom the same level of intellectual rigor that characterizes the practice of an expert in the relevant field." 526 U.S. at 152.

Dr. Mitchell admitted that his evaluation of Cooper was not consistent with the diagnostic methodology he employs in his own medical practice. With his own patients, Dr. Mitchell insists upon a physical examination. As Dr. Mitchell explained, "you can't go anywhere ifyou don't know what is wrong with the body. So you have to do a hands-on exam to find out what is wrong." Dr. Mitchell had access to Cooper, and yet, for purposes of this litigation, chose to deviate from his traditional method of evaluation. By itself, Dr. Mitchell's failure to conduct a physical examination may not be grounds to exclude his methodology as unreliable. However, the fact that Dr. Mitchell did not employ in the courtroom the same methods that he employs in his own practice provides further support for the district court's ultimate conclusion that his testimony was unreliable.

In light of these facts, the district court properly exercised its discretion in excluding Dr. Mitchell's testimony. And without Dr. Mitchell's testimony, Cooper could not make the requisite showing of causation.3 The district court therefore properly granted S&N;'s motion for summary judgment on all claims.4

For the foregoing reasons, the judgment of the district court is

AFFIRMED.