hollander v. sandoz pharmaceuticals … v. sandoz pharmaceuticals corp.1193 cite as 289 f.3d 1193...

1193HOLLANDER v. SANDOZ PHARMACEUTICALS CORP.Cite as 289 F.3d 1193 (10th Cir. 2002)

on predictions about the scope of the stateproceedings, the possibility of delay orprocedural inadequacy in the state pro-ceedings, the possibility that another fed-eral action will be time-barred should theinstant suit be dismissed, and any otherappropriate factor.

V. CONCLUSION

This court concludes that the districtcourt’s decision to apply Brillhart andwithhold jurisdiction over this declaratoryjudgment action was not an abuse of dis-cretion. It is thus unnecessary to resolvewhether the district court erred in dismiss-ing under the Colorado River doctrine.The district court, however, should consid-er whether the preferable remedy is tostay the federal proceedings. The judg-ment of the District Court for the Districtof New Mexico is therefore VACATEDand the case is REMANDED for furtherproceedings consistent with this opinion.

,

Dee HOLLANDER and Don Hollander,Plaintiffs–Appellants,

v.

SANDOZ PHARMACEUTICALS COR-PORATION, a New Jersey corpora-tion; Sandoz, Ltd., a foreign corpora-tion; and HCA Health Services ofOklahoma, Inc., an Oklahoma corpo-ration, d/b/a/ Presbyterian Hospital,Defendants–Appellees.

No. 00–6135.

United States Court of Appeals,Tenth Circuit.

May 10, 2002.

Following removal of products liabili-ty action against drug manufacturer, andother defendants, the United States Dis-

trict Court for the Western District ofOklahoma, Ralph G. Thompson, J., 95F.Supp.2d. 1230, dismissed suit againstmanufacturer’s parent corporation and en-tered judgment in favor of manufacturer,and patient appealed. The Court of Ap-peals, Henry, Circuit Judge, held that: (1)alleged deficiencies in removal proceduredid not divest the district court of subjectmatter jurisdiction where federal jurisdic-tional requirements were met at the timejudgment was entered; (2) patient’s experttestimony regarding the causal connectionbetween prescription drug and intracere-bral hemorrhages was not sufficiently reli-able to be admissible; (3) court lackedpersonal jurisdiction over Swiss parentcorporation; and (4) dismissal of Swisscorporation from products liability actionshould have been without prejudice.

Affirmed in part and remanded.

1. Removal of Cases O94Alleged deficiencies in removal proce-

dure did not divest the district court ofsubject matter jurisdiction where federaljurisdictional requirements were met atthe time judgment was entered.

2. Evidence O508, 555.2Reliability of expert testimony is de-

termined by assessing whether the reason-ing or methodology underlying the testi-mony is scientifically valid, and relevancedepends upon whether that reasoning ormethodology properly can be applied tothe facts in issue. Fed.Rules Evid.Rule702, 28 U.S.C.A.

3. Federal Courts O823District court’s application of Daubert

to exclude expert opinion evidence is re-viewed for an abuse of discretion. Fed.Rules Evid.Rule 702, 28 U.S.C.A.

4. Evidence O555.2Under Daubert’s reliability prong for

determining admissibility of expert testi-mony, an inference or assertion must be

1194 289 FEDERAL REPORTER, 3d SERIES

derived by the scientific method and mustbe supported by appropriate validation—i.e. ‘‘good grounds,’’ based on what isknown. Fed.Rules Evid.Rule 702, 28U.S.C.A.

5. Evidence O555.10Patient’s expert testimony regarding

the causal connection between prescriptiondrug and intracerebral hemorrhages wasnot sufficiently reliable to be admissible inproducts liability action against drug man-ufacturer; finding of bromocriptine’s simi-larity to other ergot alkaloids constitutedan unreliable basis on which to concludethat the drug caused vasoconstriction andensuing adverse effects like patient’sstroke, and it was not unreasonable for thedistrict court to characterize the case re-ports as unreliable evidence of causation,given the large number of women whotook the drug and the variety of possiblecauses for many of their injuries. Fed.Rules Evid.Rule 702, 28 U.S.C.A.

6. Products Liability O82.1Absent expert testimony on causal

connection between prescription drug andpatient’s stroke, drug manufacturer couldnot be held liable under Oklahoma law inproducts liability action.

7. Products Liability O46.1Under Oklahoma law, a plaintiff seek-

ing recovery for an injurious side effectfrom a properly manufactured prescriptiondrug must prove that the drug caused theinjury and that the manufacturer breacheda duty to warn of possible detrimentalreactions.

8. Evidence O584(1)Under Oklahoma law, a plaintiff must

introduce expert testimony if the fact inissue is not within the realm of ordinaryexperience of mankind.

9. Federal Civil Procedure O1837.1

Federal Courts O86

District court lacked personal jurisdic-tion over Swiss parent corporation whichsold drug manufacturer to American sub-sidiary prior to patient suffering strokeallegedly as a result of taking the drugwhere Swiss corporation completely ceaseddoing business in the United States afterthe sale; however, dismissal of Swiss cor-poration from products liability actionshould have been without prejudice to fil-ing in an appropriate forum since dismissaldid not address the merits of patient’sallegations as to Swiss corporation.

Steven R. Hickman (James E. Frasierwith him on the briefs), of Frasier, Frasier& Hickman, LLP, Tulsa, OK, for thePlaintiffs–Appellants.

Grant J. Espisito of Mayer, Brown &Platt, New York, NY; Joe G. Hollings-worth (Katherine R. Latimer and Kirby T.Griffiths of Spriggs & Hollingsworth,Washington, DC, for Defendant–AppelleeSandoz Pharmaceuticals Corporation;Richard M. Eldridge and Thomas E. Stei-chen of Eldridge Cooper Steichen &Leach, P.L.L.C., Tulsa, OK, for Defen-dants–Appellees Sandoz PharmaceuticalsCorporation and Sandoz, Ltd., with themon the brief).

David A. Branscum (Glenn D. Huff withhim on the brief) of Foliart, Huff, Ottaway& Bottom, Oklahoma City, OK, for theDefendant–Appellee HCA Health Servicesof Oklahoma, Inc.

Before EBEL and HENRY, CircuitJudges, and ROGERS, District Judge.*

* The Honorable Richard D. Rogers, UnitedStates District Judge for the District of Kan-

sas, sitting by designation.


HENRY, Circuit Judge.

Dee and Don Hollander filed this prod-ucts liability action in the District Courtfor Oklahoma County alleging that Parlo-del, a drug manufactured by Sandoz Phar-maceuticals Corporation (‘‘Sandoz’’), nowknown as Novartis Pharmaceuticals Cor-poration, and distributed by HCA HealthServices of Oklahoma, Inc., doing businessas Presbyterian Hospital (‘‘PresbyterianHospital’’), caused Ms. Hollander to sufferan intracerebral hemorrhage shortly aftershe gave birth to the Hollanders’ secondchild. After the Oklahoma County DistrictCourt dismissed the Hollanders’ claimagainst Presbyterian Hospital, the remain-ing defendants removed the case to thefederal district court.

The federal district court denied theHollanders’ motion to remand the case tostate court. It rejected the Hollanders’arguments that it lacked jurisdiction overthe remaining claims and that the defen-dants’ removal petition was untimely.Subsequently, the federal district courtdismissed the defendant Sandoz, Ltd. withprejudice, reasoning that the holding com-pany had its principal place of business inSwitzerland and that the court lacked per-sonal jurisdiction over it.

Finally, applying Daubert v. MerrellDow Pharmaceuticals, Inc., 509 U.S. 579,113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), thefederal district court ruled that the Hol-landers’ expert testimony regarding thecausal connection between Parlodel andintracerebral hemorrhages lacked the nec-essary reliability and was therefore inad-missible. See Hollander v. SandozPharms. Corp., 95 F.Supp.2d 1230, 1238–39 (W.D.Okla.2000). As a result, the courtgranted summary judgment to Sandoz.

The Hollanders now appeal those rul-ings, arguing that: (1) the federal districtcourt lacked subject matter jurisdictionand therefore erred in denying their mo-tion to remand the case to the Oklahoma

state court; (2) the court erred in dismiss-ing their claim against the defendant Pres-byterian Hospital; (3) the court abused itsdiscretion in ruling that the testimony oftheir experts was not sufficiently reliableto be admissible; (4) the court erred ingranting summary judgment to Sandoz;and (5) the district court erred in dismiss-ing their claim against Sandoz, Ltd., withprejudice.

For the reasons set forth below, weconclude that the federal district court hadsubject matter jurisdiction. We furtherhold that the court did not abuse its dis-cretion in finding that the Hollanders’ ex-pert testimony was not sufficiently reliableand that the court did not err in grantingsummary judgment to Sandoz. However,we agree with the Hollanders that thefederal district court should have dis-missed their claim against Sandoz, Ltd.,without prejudice. In light of these con-clusions, we do not address the Hollanders’challenge to the dismissal of their claimagainst Presbyterian Hospital.

Accordingly, we affirm the districtcourt’s judgment against the Hollandersand in favor of Presbyterian Hospital andSandoz. We remand the Hollanders’ claimagainst Sandoz, Ltd., so that it may bedismissed without prejudice.

I. BACKGROUND

On July 23, 1990, Ms. Hollander gavebirth by cesarean section to a healthy babyboy at Presbyterian Hospital in OklahomaCity. Because Ms. Hollander did not wantto breast feed her son, her obstetricianprescribed a fifteen day course of Parlodel,to be taken in two 2.5 mg doses per day.

Parlodel is manufactured by Sandoz.The drug’s active ingredient is bromocrip-tine mesylate, a compound derived fromergot (a naturally occurring substancemade from a fungus that attacks cerealgrains). The compound blocks the pro-


duction of prolactin, a hormone that trig-gers the secretion of milk in postpartumwomen. The Federal Drug Administra-tion (FDA) approved Parlodel for the sup-pression of post-partum lactation in 1980,and approximately 9 million women in theUnited States have taken it for that pur-pose. See Siharath v. Sandoz Pharms.Corp., 131 F.Supp.2d 1347, 1349 (N.D.Ga.2001) (discussing the history of Parlodel).Parlodel is also prescribed for severalother disorders, including acromegaly (adisease caused by hypersecretion of thepituitary growth hormone), Parkinson’sdisease, and various diseases involving theexcessive production of prolactin.

Ms. Hollander received her first dose ofParlodel at 6:00 p.m. on July 23, 1990.About two hours later, her blood pressureincreased sharply to 180/90. On the fol-lowing day, she received her second andthird doses of Parlodel, and her bloodpressure returned to the normal range.She continued to take two 2.5 mg doses ofthe drug each day. Presbyterian Hospitaldischarged her on July 27, 1990.

On the evening of July 28, 1990, Ms.Hollander complained of a severe head-ache. By the following morning, she couldneither speak nor move her right side. AtPresbyterian Hospital, a CT scan revealedthat Ms. Hollander had suffered an intra-cerebral hemorrhage in the left basalganglia area of her brain. Ms. Hollander’streating physicians were puzzled as to thecause. One of them noted that her strokeresembled those caused by hypertensionbut added that Ms. Hollander had no his-tory of the disorder. Clinical informationrevealed no pregnancy-related disorders

involving hypertension, such as eclampsiaor preeclampsia.1

On August 1, 1990, Ms. Hollander’scondition deteriorated. As a result, herphysicians performed an emergency leftfrontal craniotomy and removed an in-tracerebral hematoma. Ms. Hollanderslowly improved. She remained in thehospital for over three weeks and thentransferred to a rehabilitation center.

The Hollanders filed this action againstSandoz, Sandoz, Ltd., and PresbyterianHospital in May 1995 in the District Courtfor Oklahoma County. They alleged thatMs. Hollander’s stroke was caused by Par-lodel, that the drug was unreasonably dan-gerous to the ordinary consumer whenused as a lactation suppressant, and thatSandoz, Sandoz, Ltd., and PresbyterianHospital had failed to warn of the dangersof the drug. They further alleged thatMs. Hollander had suffered permanent in-juries.

Presbyterian Hospital filed a motion todismiss the Hollanders’ claims, arguingthat a hospital could not be held strictlyliable for providing a drug prescribed by adoctor. The Oklahoma County DistrictCourt granted Presbyterian Hospital’s mo-tion to dismiss in an oral ruling at anAugust 25, 1995 hearing. It issued a writ-ten ruling on May 10, 1996.

Sandoz filed a notice of removal on May10, 1996. The Hollanders then filed amotion to remand the case to the Okla-homa state court, which the district courtdenied. See Aplt’s App. vol. I, at 116–17(District Court Order, filed June 12, 1996).Subsequently, the court dismissed theHollanders’ claims against Sandoz, Ltd.,

1. Preeclampsia involves the ‘‘[d]evelopmentof hypertension with albuminuria or edemabetween the 20th week of pregnancy and theend of the 1st week postpartum.’’ The MerckManual, § 18 at 2057 (17th ed.1999). ‘‘Al-buminuria’’ refers to ‘‘the presence in the

urine of serum albumin.’’ Dorland’s Illustrat-ed Medical Dictionary at 42 (28th ed.1994).Eclampsia involves ‘‘[c]onvulsive seizures orcoma without other etiology occurring in thesame time period.’’ The Merck Manual, § 18at 2057.


reasoning that it was a holding companyincorporated in Switzerland with its princi-pal place of business there, that it had nooffice, manufacturing, distribution, or salesfacilities in the United States, and that itdid not advertise here. As a result, thecourt concluded that it lacked personaljurisdiction over Sandoz, Ltd., and it dis-missed with prejudice the claims againstthe company. See id. at 354 (DistrictCourt Order, filed Dec. 17, 1996).

The dispute between the Hollanders andSandoz involves issues that have beenraised in other litigation as well as inregulatory proceedings. See Kuhn v. San-doz Pharms. Corp., 270 Kan. 443, 14 P.3d1170, 1174 (2001) (describing a ‘‘decade—long disagreement between Sandoz andthe [FDA] concerning the use of Parlodelfor the prevention of physiologic lacta-tion’’). In 1984 (four years after first ap-proving the drug as a lactation suppres-sant), the FDA reported that ‘‘the labelingof Parlodel (bromocriptine) is being re-vised to reflect reports of postpartum hy-pertension, seizures, and cerebrovascularaccidents.’’ Aplt’s App. vol. IV–B, at2401–02 (FDA Drug Bulletin, vol. 14, no.1, at 3–4). The FDA explained that it hadreceived seven reports of hypertensionalone, seven reports of seizures, and threecases of cerebrovascular accidents (includ-ing one fatality). Because approximately500,000 women had used Parlodel to sup-press postpartum lactation, however, thesignificance of those reports was difficultto assess. The FDA expressly acknowl-edged that ‘‘[a] cause and effect relation-ship has not been established.’’ Id.

Sandoz eventually modified the Parlodelpackage insert to include informationabout these cases. However, the companynoted that hypertension, seizures, strokes,and myocardial infarctions regularly occur

in postpartum women who are not treatedwith bromocriptine. Thus, it maintained,‘‘the number of cases reported to Sandoz isless that one would expect even in theabsence of any drug effect.’’ Id. at 2448(Letter from Sandoz’s Executive Directorof Sales, Aug. 20, 1987).

Over the next few years, the FDA con-tinued to receive reports of adverse reac-tions to Parlodel. Sandoz commissioned astudy by Epidemiologic Resources, Inc.,regarding the relationship between Parlo-del and strokes and seizures (the ‘‘ERIstudy’’). See Aplt’s App. vol. II–D, at1361–1532 (Kenneth Rothman, et al., ‘‘AnEpidemiologic Evaluation of the PossibleRelation Between Bromocriptine, Puerper-al Seizures and Strokes,’’ (Sept. 30,1988)); 2 Siharath, 131 F.Supp.2d at 1356–57 (discussing the ERI study). Althoughthe ERI study did not find a causal con-nection between strokes and seizures, theFDA concluded that the study failed toallay concerns regarding the drug’s associ-ation with seizures and involved too fewindividuals to adequately characterize therisk of stroke.

The FDA further concluded that thepossibility that Parlodel might cause seri-ous adverse reactions in some patients out-weighed the limited benefits associatedwith its use. As a result, it requested allmanufacturers to remove the indication forlactation suppression from the Parlodel la-bel. Initially, Sandoz refused to complywith the FDA’s request, arguing that Par-lodel should not be used routinely butshould be available in specific circum-stances recommended by physicians. Notsatisfied with this position, the FDA initi-ated formal procedures for withdrawing itsprior approval for the labeling of Parlodel.The FDA explained its position as follows:

2. The purpureum is ‘‘the period from the endof the third stage of labor until the involutionof the uterus is complete, usually lasting 3 to

6 weeks.’’ Dorland’s Illustrated Medical Dic-tionary at 1386.


FDA now has new information suggest-ing that therapeutic use of bromocrip-tine for the prevention of physiologicallactation may lead to serious adverseexperiences, including death and paraly-sis, in a small but significant number ofpatients. Patients at high risk of expe-riencing these serious adverse experi-ences cannot be adequately predeter-mined. In light of the limited benefit ofusing bromocriptine for the preventionof lactation, and the effectiveness andlack of serious adverse effects of conser-vative treatments such as breast bindingwith or without mild analgesics, the riskthat bromocriptine may cause a seriousadverse effect in a postpartum woman isunacceptable.

Accordingly, the Director concludesthat the potential risks associated withthe use of bromocriptine for the preven-tion of physiological lactation outweighits limited benefits and bromocriptine isno longer shown to be safe for use inpreventing physiological lactation.

59 Fed.Reg. 43347, 43351 (Aug. 24, 1994).Sandoz then agreed to FDA’s proposal towithdraw the indication for the suppres-sion of postpartum lactation.

Following the FDA’s approval of Parlo-del as a lactation suppressant, professionalmedical journals began to publish reportsregarding women who had suffered heartattacks and strokes after taking the drug.For example, one of the Hollanders’ expertwitnesses described two patients who hadsuffered from cardiac dysfunction, sei-zures, and cerebral vasospasm. See Ken-neth Kulig, ‘‘Bromocriptine AssociatedHeadache: Possible Life ThreateningSympathomimetic Intersection,’’ Obstetricsand Gynecology, 72: 941(1991) (Aplt’sApp. vol. IV–B, at 2444–46). Anotherexpert published case histories concerningwomen who had suffered from heart at-tacks. See, e.g., Leslie Iffy, et al., ‘‘AcuteMyocardial Infarction in the Puerperiumin Patients Receiving Bromocriptine,’’

American Journal of Obstetrics and Gyne-cology, vol. 155, No. 2, at 371–72 (1986)(Aplt’s App. vol. IV–D, at 2976–77).Medical researchers also published numer-ous articles reporting the effects of bromo-criptine in animals. See Siharath, 131F.Supp.2d at 1366–69 (discussing animalstudies), Glastetter v. Novartis Pharms.Corp., 107 F.Supp.2d 1015, 1037–1041(E.D.Mo.2000) (same), aff’d, 252 F.3d 986,991 (8th Cir.2001). Some of the studiesinvolved dogs, rats, and pithed animals.See id.; see also Aplt’s App. vol. I–A, at369–376 (Sandoz’s statement of materialfacts, filed July 15, 1999) (discussing ani-mal studies). Some researchers concludedthat, contrary to the Hollanders’ allega-tions regarding the effect of bromocriptineon Ms. Hollander, the drug actually de-creases blood pressure. See, e.g., SaadLahlou & Pierre Demenge, ‘‘Contributionof Spinal Dopamine Receptors to the Hy-potensive Action of Bromocriptine inRats,’’ Journal of Cardiovascular Phar-macology, vol. 18, 317–323 (1991) (Aplt’sApp. vol. II–E, at 1646–54).

As the discussion in the scientific litera-ture continued, the controversy over Par-lodel made its way to the courts. In 1994,a Kentucky jury awarded $968,512 in com-pensatory damages and $1,000,000 in puni-tive damages to a woman who alleged thatParlodel had caused her stroke. SeeAplt’s App. vol. IV–A, at 2171 (Judgmentin Roberts v. Betts, no. 89–CI–653–V,25th Judicial Dist., Pulaski Cir. Ct., July20, 1994). A Kentucky appellate court af-firmed that judgment in an unpublishedopinion. See id. at 2175–78. In contrastto the Roberts case, several recent deci-sions have rejected claims that Parlodelhas caused strokes and heart attackswhen prescribed as a post-partum lacta-tion suppressant, concluding that the sci-entific evidence supporting these claimswas not sufficiently reliable under Dau-


bert.3 However, several other decisionshave reached the opposite conclusion.4

See generally Mark Hansen, ‘‘When Ex-pert Testimony Fails the Test: DistrictCourts Disagree on what Defines Causa-tion Evidence in Drug Disability Cases,’’88 ABA Journal 22 (Jan.2002) (statingthat ‘‘[an] Alabama magistrate’s decisionbrought to eight the number of productsliability suits over Parlodel that have sur-vived a so-called Daubert challenge to theadmissibility of the plaintiffs’ causation ev-idence [b]ut [an] Illinois judge’s ruling—tantamount to an order of summary judg-ment for the defense—marked the sev-enth trial or appellate decision to excludesuch evidence’’).

In support of their contention that Par-lodel caused Ms. Hollander’s stroke, theHollanders relied primarily on the testimo-ny of three experts: (1) Dr. Kenneth Ku-lig, a physician who is board-certified intoxicology and emergency medicine andwho has served as the Chairman of thePharmacy and Therapeutics Committeeand Director of the Porter Regional Toxi-cology Center at Porter Adventist Hospitalin Denver, Colorado, and as an associateclinical professor in the Division of Emer-gency Medicine and Trauma in the De-partment of Surgery at the University ofColorado Health Sciences Center; (2) Dr.Leslie Iffy, M.D., a professor in the De-partment of Obstetrics and Gynecology ofthe University of Medicine and Dentistryof New Jersey; and (3) Dr. Pedro A. Jose,M.D., Ph.D., a Professor of Pediatrics,Physiology, and Biophysics at Georgetown

University and an expert on the role ofdopamine and dopaminergic drugs on thedevelopment of hypertension.

The parties offered deposition testimo-ny, affidavits, expert reports, and tran-scripts of testimony from other cases in-volving Parlodel. In general, the experts’theory was that in certain women, Parlodelcauses vasoconstriction (a narrowing of theblood vessels) and hypertension (highblood pressure). Vasoconstriction and hy-pertension, the experts reasoned, can thencause strokes, as they did in the case ofMs. Hollander.

In his written report, Dr. Kulig ex-plained that he had reviewed Ms. Hol-lander’s medical records, medical litera-ture regarding bromocriptine and otherergot alkaloids, FDA documents, andmarketing, promotional and research ma-terial complied by Sandoz. He concludedthat Ms. Hollander suffered ‘‘an intracere-bral hemorrhage secondary to ergot in-duced vasospasm resulting in blood vesselrupture in her brain.’’ Aplt’s App. vol.II–A, at 652 (Dr. Kulig’s Sept. 22, 1998report, at 3). He added, ‘‘It is my opinionwith a reasonable degree of medical cer-tainty that Mrs. Hollander’s stroke wascaused by the drug bromocriptine, andhad the patient not been taking the drug,she would not have had a stroke.’’ Id. at653. According to Dr. Kulig, the fact thatbromocriptine could cause strokes waswell know to Sandoz at the time that Ms.Hollander began taking the drug. Id. at652.

3. See Glastetter v. Novartis Pharms. Corp., 252F.3d 986, 989 (8th Cir.2001); Caraker v. San-doz Pharms. Corp., 172 F.Supp.2d 1046(S.D.Ill.2001); Siharath, 131 F.Supp.2d 1347(N.D.Ga.2001); Brumbaugh v. Sandoz Pharm.Corp., 77 F.Supp.2d 1153, 1155, (D.Mont.1999).

4. See Brasher v. Sandoz Pharms. Corp., 160F.Supp.2d 1291 (N.D.Ala.2001); Globetti v.Sandoz Pharms. Corp., 111 F.Supp.2d 1174

(N.D.Ala.2000); Aplt’s App. vol. V, at 3184–3213 (Tr. of unpublished ruling in Kittelson v.Sandoz Pharms. Corp., No. 98–2277 (D.Minn.March 2, 2000)) (denying motion to excludescientific testimony as unreliable); Kuhn, 14P.3d at 1179–85 (applying the test for admis-sibility set forth in Frye v. United States, 293F. 1013 (D.C.Cir.1923) and concluding thatthere were genuine issues of material fact asto whether Parlodel caused a patient’s death).


In an affidavit in another case involvingParlodel, Dr. Kulig provided a more de-tailed explanation as to how he hadreached his conclusions. See Aplt’s App.vol. II–E, at 1742–59 (Dr. Kulig’s affidavitin Railey v. Novartis PharmaceuticalsCorp., 94–1440 (C.D.Ill.)). He noted thatbromocriptine is an ergot, ‘‘a class of drugswith known molecular structures andmany common properties,’’ including thetendency to cause vasoconstriction. Seeid. at 1745 (stating that ‘‘[o]ne needs onlyto look at the package inserts from Sandozregarding other ergot alkaloids it manufac-tures to understand that vasoconstrictionis indeed the core property of ergot alka-loids’’). Bromocriptine differs from thenaturally occurring ergot alkaloid alphaergocriptine only in that the molecule hasan additional bromine atom attached to thesecond carbon atom of the basic ergot ring.See id.

Dr. Kulig explained the significance ofthe structural differences between bromo-criptine and other ergot alkaloids as fol-lows:

Although the adding of a bromine atomto the core nucleus makes the drug insome patients a vasodilator (the firstdose may cause a precipitous fall inblood pressure, another fact that makesthe drug unsafe in the post-partum peri-od), it is misleading and inaccurate tosuggest that the drug can never causevasoconstriction or hypertension in anyperson. While the addition of a bromineatom to a very large organic moleculemay change the pharmacokinetics andpharmacological dynamics of a drug, itwould be unlikely to change the coreproperties of an ergot alkaloid from be-ing a vasoconstrictor to a vasodilator inall cases. Clinical studies, epidemiologicevidence and adverse drug reaction ex-perience with this drug indicates thatvasoconstriction with bromocriptine un-questionably occurs, as would be expect-

ed based on the fact that it is an ergotthat can cause ergotism.

Id. at 1745–46.

Dr. Kulig also discussed several othercategories of evidence on which he hadrelied in forming his opinion. These in-cluded the pharmacological and toxicologi-cal properties of bromocriptine, studies ofthe relationship between bromocriptineand hypertension, and case reports con-cerning adverse reactions.

As to pharmacology, Dr. Kulig ex-plained that bromocripitne is ‘‘a dopamineagonist.’’ Id. In other words, the drugstimulates the release of dopamine, a neu-rotransmitter. Dr. Kulig added that bro-mocriptine is also a ‘‘dopamine–1 antagon-ist,’’ inhibiting the effects of dopamine atspecific ‘‘D–1’’ receptors. Id. According toDr. Kulig, dopamine is a well known vaso-constrictor. However, activation of the‘‘D–1’’ dopamine receptors results in vaso-dilation. Thus, the fact that bromocrip-tine stimulates the release of dopamineand that it inhibits the D–1 dopamine re-ceptors is consistent with the drug causingvasoconstriction.

Moreover, Dr. Kulig reported that bro-mocriptine ‘‘has a very long beta elimina-tion half-life of about 50 hours.’’ Id. at1747. That means that a steady state isnot achieved in someone taking the drugtwice a day until about ten days afterleaving the hospital. As a result, one wouldnot expect women who have taken thedrug to suppress post-partum lactation todevelop hypertension and vasospasm untilafter their discharge from the hospital.

With regard to hypertension, Dr. Kuligreferred to three studies. In the first,commissioned by Sandoz, nineteen percentof patients demonstrated increases inblood pressure after taking bromocriptine.In the second study, published by the FDAin 1984, six out of seven patients whodeveloped hypertension while on bromo-


criptine regained normal blood pressureafter they stopped taking the drug. Final-ly, a study by Dr. Dorothy Watson 5 foundthat women with pregnancy-induced hy-pertension had a higher incidence of post-partum hypertension after taking the drugthan those women not receiving bromo-criptine. See id. at 1748–49 (stating that‘‘[t]his case control study provides impor-tant evidence that bromocriptine causespost-partum hypertension in women whohad pregnancy-induced hypertension priorto delivery’’).

Finally, Dr. Kulig discussed the adversereactions to bromocriptine that had beenspontaneously reported to the FDA andSandoz. These reactions included hyper-tension, seizures, strokes, and myocardialinfarction. Although he acknowledgedthat these reports did not establish causa-tion, he presented them as an importantfactor to be considered along with theother scientific evidence.

In addition to Dr. Kulig, the Hollandersalso relied on the opinion of Dr. LeslieIffy. See Aplt’s App. vol. II–A, at 733–739(Dr. Iffy’s written report, July 25, 1997).Dr. Iffy concluded that there was ‘‘an over-whelming probability’’ that Ms. Holland-er’s stroke was caused by bromocriptine.Id. at 738. He listed five factors thatsupported his opinion: (1) Ms. Hollanderhad normal blood pressure during a priorpregnancy but displayed an episode of hy-pertension when she took Parlodel duringher first childbirth; (2) she suffered epi-sodes of hypertension during her secondpregnancy and immediately after the birthof her second child; ‘‘[t]his being the case,she was predisposed for the hypertensiveeffect of bromocriptine’’; (3) when shetook Parlodel after the second pregnancy,her blood pressure increased, ‘‘in all prob-ability a bromocriptine effect’’; (4) ‘‘[t]he

time of occurrence of the cerebral hemor-rhage (sixth day postpartum) was highlycharacteristic of bromocriptine related cat-astrophic side effects’’; and (5) ‘‘[a]partfrom her moderate smoking habit, the pa-tient had no identifiable predisposing fac-tors for cerebral hemorrhage[;][t]he ab-sence of evidence of congenital defect atthe site of the hemorrhage further empha-sizes the lack of predisposing factors onthe part of the patient.’’ Id. Dr. Iffy add-ed that, in his view, Sandoz was aware ofthe dangers of Parlodel at the time thatMs. Hollander suffered her stroke. Seeid. at 739.

The third expert on whom the Holland-ers relied—Dr. Pedro Jose-set forth amore specific theory of causation. In hiswritten report, Dr. Jose stated, ‘‘I thinkthat the hypertension (which caused thestroke) is probably due to Parlodel; thehypertension would not have happened ifParlodel was not prescribed.’’ Aplt’s App.vol. II–A, at 842 (Dr. Jose’s report, Jan.23, 1998). Dr. Jose acknowledged thatbromocriptine often decreases blood pres-sure. However, in deposition testimony heexplained that in instances in which extra-cellular fluid volume is increased—as is thecase in pregnancy—and in which the activ-ity of the sympathetic nervous system isdecreased, bromocriptine has the ‘‘para-doxical effect of increasing blood pres-sure.’’ Id. at 801.

After conducting discovery, Sandoz fileda motion to exclude the opinion testimonyoffered by the Hollanders’ experts on thegrounds that the testimony was insuffi-ciently reliable under Daubert. In thesame motion, Sandoz requested the courtto grant summary judgment in its favor.It argued that, in the absence of the unre-liable opinion testimony, the Hollanders

5. Watson, D.I., et al., ‘‘Bromocriptine Mesy-late for Lactation Suppression: A Risk forPostpartum Hypertension?’’ Obstetrics and

Gynecology, 74(4): 573–576 (1989) (Aplt’sApp. vol. II–D, at 1596–97).


could not demonstrate that there werecontroverted issues of material fact on theissue of whether Parlodel caused Ms. Hol-lander’s stroke.

The federal district court granted San-doz’s motion to exclude the Hollanders’expert testimony as well as its motion forsummary judgment. See Hollander, 95F.Supp.2d at 1235–39. The court set forthfour reasons in support of its evidentiaryruling.

First, the court observed that the Hol-landers’ experts were unable to explainthe physiological mechanism by whichParlodel caused vasoconstriction and ensu-ing hypertension and strokes. The courtexplained that ‘‘Dr. Kulig could only listpossible mechanisms for Parlodel causinghypertension,’’ that ‘‘Dr. Jose could notcite any studies or tests that proved hishypothesis that bromocriptine might causehigh blood pressure,’’ and that ‘‘Dr. Iffyalso classified his opinion that Parlodelcaused Mrs. Hollander’s stroke as being ahypothesis, which is not held by a medicaldegree of certainty.’’ See id. at 1235–36(internal quotation marks omitted).

Second, the court reasoned that thekinds of case reports on which the Hol-landers relied have been repeatedly reject-ed as a scientific basis for establishingcausation. It stated: ‘‘The problems withcase reports and adverse drug experiencereports were acknowledged by Dr. Iffy—because they are not controlled studiesand do not eliminate confounding varia-bles, the reported effect or injury could bedue to some other cause than Parlodel.’’See id. at 1237.

Third, the court found that the fact thatbromocriptine belongs to a class of com-pounds (ergot alkaloids) that have beenshown to cause hypertension did not con-stitute reliable causation evidence. Thecourt observed that the Hollanders hadfailed to refute Sandoz’s evidence that‘‘[t]he chemical diversity of ergot alkaloids

corresponds to the diversity of the biologi-cal activities of these compounds.’’ Id. at1238 (internal quotation marks and empha-sis omitted).

Fourth, the animal studies on which theHollanders’ experts relied were too dissim-ilar to the facts of this case. ‘‘The studiesrelied upon involved different drugs, didnot test the systemic effect of the drug,some of the animals were anaesthetized,and they were neither pregnant nor post-partumTTTT Doctors Jose and Kulig wereunaware of any controlled animal studiesin which bromocriptine caused an increasein blood pressure.’’ Id. at 1238 (internalcitations omitted).

In light of its conclusion that the Hol-landers’ opinion testimony was insufficient-ly reliable under Daubert, the districtcourt briefly assessed the state of the rec-ord absent that testimony. It concludedthat, without expert opinion testimony, theHollanders could not demonstrate thatParlodel caused Ms. Hollander’s stroke.The court therefore granted Sandoz’s mo-tion for summary judgment.

II. DISCUSSION

In this appeal, the Hollanders primarilychallenge the district court’s assessment oftheir evidence that Parlodel caused Ms.Hollander’s stroke. In particular, theychallenge both the district court’s applica-tion of Daubert to exclude their expertopinion testimony and the district court’sassessment of the record absent that testi-mony. They maintain that ‘‘[w]hether ornot the experts are allowed to give theirultimate opinion, there is sufficient evi-dence that a jury could find in [their] favoron the issue of causation.’’ Aplt’s Br. at28. The Hollanders also present severalother challenges to the district court’s rul-ings.

We begin our analysis by addressingthe Hollanders’ argument that the district


court lacked subject matter jurisdictionand therefore erred in denying their mo-tion to remand the case to the Oklahomastate court. Because we conclude thatthe federal district court had subject mat-ter jurisdiction, we then turn to the Hol-lander’s arguments regarding scientificevidence. Finally, we address their argu-ment that the district court erred in dis-missing their claims against Sandoz, Ltd.,with prejudice.

A. Jurisdiction

[1] Focusing on several alleged defectsin the removal procedure, the Hollandersargue that the federal district court erredin denying their motion to remand. As aresult, they contend, that court lacked jur-isdiction to adjudicate the case, and thejudgment in favor of the defendants shouldbe vacated so that the case may be heardin the Oklahoma state court.

In light of the Supreme Court’s decisionin Caterpillar, Inc. v. Lewis, 519 U.S. 61,117 S.Ct. 467, 136 L.Ed.2d 437 (1996), weneed not address the Hollanders’ specificarguments. In Caterpillar, the Court heldthat ‘‘a district court’s error in failing toremand a case improperly removed is notfatal to the ensuing adjudication if federaljurisdictional requirements are met at thetime judgment is entered.’’ Id. at 64, 117S.Ct. 467; see also Feichko v. Denver &Rio Grande W. R.R. Co., 213 F.3d 586,590–91 (10th Cir.2000) (discussing Cater-pillar ). The Court reasoned that, despitedeficiencies in the removal process, ‘‘[to]wipe out the adjudication postjudgment,and return to state court a case now satis-fying the federal jurisdictional require-ments, would impose an exorbitant cost onour dual court system, a cost incompatiblewith the fair and unprotracted administra-

tion of justice.’’ Caterpillar, 519 U.S. at77, 117 S.Ct. 467.

Here, as the defendants observe, com-plete diversity existed between the remain-ing parties at the time that the federaldistrict court entered judgment: the Hol-landers resided in Arkansas and the defen-dant Sandoz was a Delaware corporationwith its principal place of business in NewJersey. Thus, the alleged deficiencies inthe removal procedure do not divest thedistrict court of subject matter jurisdic-tion. We therefore proceed to the merits.

B. Admissibility of the Hollanders’Experts’ Testimony Under

Daubert

Rule 702 of the Federal Rules of Evi-dence provides:

[i]f scientific, technical, or other special-ized knowledge will assist the trier offact to understand the evidence or todetermine a fact in issue, a witness qual-ified as an expert by knowledge, skill,experience, training, or education, maytestify thereto in the form of an opinionor otherwise, if (1) the testimony isbased upon sufficient facts or data, (2)the testimony is the product of reliableprinciples and methods, and (3) the wit-ness has applied the principles andmethods reliably to the facts of the case.

Fed.R.Evid. 702.In Daubert, the Supreme Court conclud-

ed that Rule 702 superseded the ‘‘generalacceptance’’ standard for the admissibilityof scientific evidence first set forth inFrye, 293 F. at 1014.6 Under Daubert,when faced with a proffer of expert scienti-fic testimony, a district court ‘‘must deter-mine at the outset, pursuant to [Fed.R.Evid.] 104(a), whether the expert is pro-posing to testify to (1) scientific knowledgethat (2) will assist the trier of fact to

6. The Frye test required district courts to ex-clude evidence when the underlying scientificprinciples were not ‘‘sufficiently established

to have gained general acceptance in the par-ticular field in which [they belong].’’ Frye,293 F. at 1014.


understand or determine a fact in issue.’’Daubert, 509 U.S. at 592, 113 S.Ct. 2786.Thus, under Daubert, the district courtperforms an important gatekeeping role inassessing scientific evidence. See Macsen-ti v. Becker, 237 F.3d 1223, 1230–34 (10thCir.2001) (discussing the district court’sgatekeeping function under Daubert ).

[2] The Daubert standard ensures thatthe proffered evidence is both ‘‘reliable’’and ‘‘relevant.’’ See Daubert, 509 U.S. at589, 113 S.Ct. 2786. Reliability is deter-mined by assessing ‘‘whether the reason-ing or methodology underlying the testi-mony is scientifically valid.’’ Id. at 592–93,113 S.Ct. 2786. Relevance depends upon‘‘whether [that] reasoning or methodologyproperly can be applied to the facts inissue.’’ Id. at 593, 113 S.Ct. 2786.

[3] We review the district court’s ap-plication of Daubert to exclude expertopinion evidence for an abuse of discretion.See General Electric v. Joiner, 522 U.S.136, 143, 118 S.Ct. 512, 139 L.Ed.2d 508(1997); Mitchell v. Gencorp Inc., 165 F.3d778, 780 (10th Cir.1999). Thus, we mustafford substantial deference to the districtcourt’s application of Daubert. See Kum-ho Tire Co. v. Carmichael, 526 U.S. 137,152, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999)(‘‘the trial judge must have considerableleeway in deciding in a particular case howto go about determining whether particu-lar expert testimony is reliable’’); Joiner,

522 U.S. at 143, 118 S.Ct. 512 (noting thatthe court of appeals ‘‘failed to give the trialcourt the deference that is the hallmark ofabuse-of-discretion review’’). Under theabuse of discretion standard, ‘‘a trialcourt’s decision will not be disturbed un-less the appellate court has a definite andfirm conviction that the lower court madea clear error of judgment or exceeded thebounds of permissible choice in the circum-stances.’’ McEwen v. City of Norman,Okla., 926 F.2d 1539, 1553–54 (10th Cir.1991); see also Summers v. Missouri Pa-cific R.R. System, 132 F.3d 599, 603 (10thCir.1997) (stating that, under the abuse ofdiscretion standard, ‘‘[w]e will not disturbthe trial court’s determination ‘absent adistinct showing it was based on a clearlyerroneous finding of fact or an erroneousconclusion of law or manifests a clear errorof judgment’ ’’).

Here, the Hollanders focus on the dis-trict court’s application of the ‘‘reliability’’prong of the Daubert inquiry. They in-voke two methods of causation analysis:one promulgated by Sandoz’s Drug Moni-toring Center and the other set forth by aprofessor of medical statistics, Sir AustinBradford Hill. See Aplt’s App. vol. IV–B,at 2378–79 (Sandoz’s classifications of evi-dence of causation); vol. IV–D, at 2949–52 (Bradford Hill, ‘‘The Environment andDisease: Association or Causation?,’’ Pro-ceedings of the Royal Society of Medicine,vol. 58 no. 5 (May 1965)).7

7. Under the Sandoz scheme, there are fourmain categories of causation: (a) not related;(b) remote; (c) probable; and (d) definite.The Hollanders focus on the standard for‘‘probable’’ causation.

An adverse reaction is considered to be‘‘probably related to a drug’’ if the reaction:(1) ‘‘occurs within a reasonable time intervalfollowing the administration of the drug’’; (2)‘‘could not readily be attributed to the pa-tient’s clinical condition/underlying disease,concomitant therapy, or to environmental ortoxic factors’’; (3) ‘‘follows a known patternof response to the drug’’; and (4) ‘‘disappears

or improves following cessation of treatmentor dose reduction.’’ See Aplt’s App. vol. IV–B, at 2378–79.

Sir Bradford Hill sets forth nine factors thatshould be considered ‘‘before deciding thatthe most likely interpretation [of the associa-tion] is causation.’’ These factors are:strength, consistency, specificity, temporality,dose response, biological plausibility, coher-ence, experimental evidence and analogy.See Aplt’s App. vol. IV D, at 2949–52 (Brad-ford Hill paper); vol. II–E, at 1751–55 (Dr.Kulig’s affidavit). Dr. Kulig applies the Brad-


Relying primarily on Dr. Kulig’s testi-mony, the Hollanders maintain that hisopinion that Parlodel caused Ms. Holland-er’s stroke comports with these generalstandards of causation analysis. Accord-ingly, they reason, Dr. Kulig’s testimony issufficiently scientific to be admissible un-der Daubert. See Aplt’s Br. at 24 (‘‘[I]tcannot be gainsaid that Dr. Kulig’s meth-odology—how he takes the informationavailable and analyzes it in a scientificmanner—is good science and would behelpful to a jury of laymen.’’).

Additionally, the Hollanders note thatDr. Kulig criticized some of Sandoz’s ownevidence on scientific grounds. They pointto his testimony that the studies invokedby Sandoz do not demonstrate that thereis an increased risk of stroke in the post-partum period generally. As a result, theymaintain, the evidence is disputed as towhether Ms. Hollander’s stroke may beexplained by this generally increased riskduring the post-partum period, as Sandozcontends, rather than by her taking Parlo-del.

Finally, the Hollanders contend thatDrs. Kulig, Iffy, and Jose applied themethodology generally employed by ex-perts in the relevant fields. Thus, by rely-ing on scientific principles, professionalpublications, animal studies, differential di-agnoses, and case reports, these expertsdid what Daubert requires: they groundedtheir conclusions in ‘‘the methods and pro-cedures of science’’ rather than ‘‘subjectivebelief or unsupported speculation.’’ Dau-bert, 509 U.S. at 590, 113 S.Ct. 2786.

In order to assess the Hollanders’ argu-ment, we begin with an overview of thestandards for reliability under Daubertand its progeny. Then we turn to anexamination of the scientific opinion evi-dence at issue here.

1. Scientific Reliability Under Daubert

[4] Under Daubert’s reliability prong,‘‘an inference or assertion must be derivedby the scientific method TTT [and] must besupported by appropriate validation—i.e.‘good grounds,’ based on what is known.’’Id. The Supreme Court listed four nonex-clusive factors that the trial court mayconsider in assessing reliability: (1) wheth-er the opinion at issue is susceptible totesting and has been subjected to suchtesting; (2) whether the opinion has beensubjected to peer review; (3) whetherthere is a known or potential rate of errorassociated with the methodology used andwhether there are standards controllingthe technique’s operation; and (4) whetherthe theory has been accepted in the scien-tific community. See id.

The list is not exclusive, and districtcourts applying Daubert have broad dis-cretion to consider a variety of other fac-tors. See Kumho Tire, 526 U.S at 150, 119S.Ct. 1167 (‘‘[W]e can neither rule out, norrule in, for all cases and for all time theapplicability of the factors mentioned inDaubert, nor can we now do so for subsetsof cases categorized by category of expertor by kind of evidence. Too much dependsupon the particular circumstances of theparticular case at issue.’’). Generally, thedistrict court should focus on the experts’methodology rather than the conclusionsthat they generate. See Daubert, 509 U.S.at 595, 113 S.Ct. 2786. However, the ex-perts’ conclusions are not immune fromscrutiny: ‘‘A court may conclude thatthere is simply too great an analytical gapbetween the data and the opinion prof-fered.’’ Joiner, 522 U.S. at 147, 118 S.Ct.512 (‘‘[N]othing in either Daubert or theFederal Rules of Evidence requires a dis-trict court to admit opinion evidence that isconnected to existing data only by the ipsedixit of the expert.’’). Regardless of the

ford Hill criteria in reaching his opinion that Parlodel caused Ms. Hollander’s stroke.


specific factors at issue, the purpose of theDaubert inquiry is always the same: ‘‘[t]omake certain that an expert, whether bas-ing testimony upon professional studies orpersonal experience, employs in the court-room the same level of intellectual rigorthat characterizes the practice of an expertin the relevant field.’’ 8 Kumho Tire, 526U.S. at 152, 119 S.Ct. 1167.9

2. The Hollanders’ Scientific Evidence

[5] We assess the Hollanders’ chal-lenge to the district court’s Daubert rulingby examining the opinions of their threeprimary experts: Drs. Kulig, Iffy, andJose. As to each expert, we must assessthe grounds that they provide for theiropinion that Parlodel causes stroke, askingwhether those grounds involve ‘‘the meth-ods and procedures of science,’’ Daubert,509 U.S. at 590, 113 S.Ct. 2786, and ‘‘thelevel of intellectual rigor of the expert inthe field.’’ Kumho Tire, 526 U.S. at 152,119 S.Ct. 1167.

In doing so, we note that the scope ofour review is quite narrow: we may re-verse the district court’s ruling only if weconclude that it abused its discretion inapplying Daubert to exclude opinions ofthe Hollanders’ experts. Because the dis-trict court has discretion to consider a

variety of factors is assessing reliabilityunder Daubert, and because, in light ofthat discretion, there is not an extensivebody of appellate case law defining thecriteria for assessing scientific reliability,we are limited to determining whether thedistrict court’s application of the Daubertmanifests a clear error of judgment orexceeds the bounds of permissible choicein the circumstances. See McEwen, 926F.2d at 1553–54 (discussing appellate re-view for an abuse of discretion). Thus,when coupled with this deferential stan-dard of review, Daubert’s effort to safe-guard the reliability of science in thecourtroom may produce a counter-intui-tive effect: different courts relying on theessentially the same science may reachdifferent results. See generally FederalJudicial Center, Reference Manual onScientific Evidence 27 (2d ed.2000) (ob-serving that, in light of the abuse of dis-cretion standard of review for Daubert de-terminations of reliability, ‘‘in theoryjudges are free to select different proce-dures and apply different factors to a par-ticular expert or type of expertise thantheir colleagues do in the same district orcircuit’’ and that ‘‘[a]s a consequence, sim-ilar cases could be resolved differently onthe basis of inconsistent determinations

8. As Justice Breyer has observed, the require-ment that the district court assess reliabilityand relevance under Daubert ‘‘will sometimesask judges to make subtle and sophisticateddeterminations about scientific methodologyand its relation to the conclusions an expertwitness seeks to offer-particularly when acase arises in an area where the science itselfis tentative or uncertain, or where testimonyabout general risk levels in human beings oranimals is offered to prove individual causa-tion.’’ Joiner, 522 U.S. at 147–48, 118 S.Ct.512 (Breyer, J., concurring). Even thoughjudges usually do not have the formal scienti-fic training to assist them in making thesedecisions, there are Rules of Evidence andCivil Procedure that may assist them in mak-ing the necessary determinations. ‘‘Amongthese techniques are an increased use of pre-

trial conference authority [pursuant to Fed.R.Civ.P. 16] to narrow the scientific issues indispute, pretrial hearings where potential ex-perts are subject to examination by the court,and the appointment of special masters andspecially trained law clerks.’’ Id. at 149, 118S.Ct. 512 (citations omitted).

9. Judge Posner has expressed a similar view.‘‘When the Supreme Court in Daubert toldjudges to distinguish between real and court-room science, it was not with the object ofdiscovering the essence of ‘‘science,’’ if thereis such an essence. The object TTT was tomake sure that when scientists testify in courtthey adhere to the same standards of intellec-tual rigor that are demanded in their profes-sional work.’’ Rosen v. C G Corp., 78 F.3d316, 318 (7th Cir.1996).


about admissibility’’); see also Brasher,160 F.Supp.2d at 1298 n. 17 (observingthat the Eighth Circuit’s decision in Glas-tetter, 252 F.3d at 989–92, affirming theexclusion of Parlodel evidence as unrelia-ble ‘‘does not necessarily [establish] thatan inconsistent holding by this courtwould constitute an abuse of discretion’’).10

a. Similarity to other ergot alkaloids

We begin our analysis with Dr. Kulig’stestimony that bromocriptine is an ergotalkaloid. According to Dr. Kulig, this factsupports his theory that bromocriptinecauses vasoconstriction.

As the district court observed, neitherDr. Kulig nor the Hollanders’ other ex-perts disputed the fact that bromocriptinehas a different chemical structure thanergot alkaloids known to cause vasocon-striction. Moreover, neither Dr. Kulig northe Hollanders’ other experts disputed thescientific literature stating that small dif-ferences in chemical structure may pro-duce substantial differences in physiologi-cal effects.

In light of these considerations, severalcourts have agreed with the district court’sconclusion that the fact that bromocriptineis an ergot alkaloid does not constitutereliable scientific evidence that it causesvasoconstriction and associated adverse re-actions like heart attacks and strokes. SeeGlastetter, 252 F.3d at 990 (‘‘[T]his genericassumption that bromocriptine behaveslike other ergot alkaloids carries little sci-entific value. Even minor deviations inmolecular structure can radically change a

particular substance’s properties and pro-pensities.’’); 11 Brumbaugh, 77 F.Supp.2dat 1156 (‘‘Testimony extending generalconclusions about similar drugs does notmeet Daubert’s requirement of reliabili-ty.’’); Siharath, 131 F.Supp.2d at 1363–65(finding a lack of reliable evidence thatbromocriptine acts like other ergot alka-loids). Moreover, in a similar case, thiscircuit has rejected the argument that onechemical’s resemblance to another knownto have deleterious effects constituted reli-able causation evidence. See Mitchell, 165F.3d at 782 (‘‘Missing from [the plaintiff’sevidence] is additional testimony explain-ing what these similarities are and how thesimilarities cause the human body to re-spond to Defendant’s chemicals in a man-ner similar to benzene.’’).

These decisions support the districtcourt’s analysis. Accordingly, the districtcourt did not abuse its discretion in findingthat bromocriptine’s similarity to other er-got alkaloids constituted an unreliable ba-sis on which to conclude that the drugcauses vasoconstriction and ensuing ad-verse effects like Ms. Hollander’s stroke.

b. Pharmacology of bromocriptine

In important respects, the Hollanders’experts’ discussion of the specific pharma-cological properties of bromocriptine issimilarly speculative. For example, Dr.Kulig’s affidavit refers to the drug’s knowneffects on dopamine and serotonin, andnotes that these neurotransmitters areknown to trigger vasoconstriction and va-

10. Conflicting decisions in the district courtsregarding the reliability of opinion testimonyabout Parlodel further illustrate this point.Compare Siharath, 131 F.Supp.2d 1347 (evi-dence regarding Parlodel’s adverse effects un-reliable); and Brumbaugh, 77 F.Supp.2d 1153(same); with Brasher, 160 F.Supp.2d 1291(Parlodel evidence reliable); and Globetti v.Sandoz Pharms., Corp., 111 F.Supp.2d 1174(N.D.Ala.2000) (same).

11. The Eighth Circuit also noted that ‘‘oneleading treatise on medical toxicology con-cludes that bromocriptine has no vasocon-strictive properties.’’ Glastetter, 252 F.3d at990 (emphasis in original) (citing Matthew J.Ellenhorn, Ellenhorn’s Medical Toxicology:Diagnosis and Treatment of Human Poisoning1879, table 74–23 (2d ed.1997)).


sospasm. However, although he statesthat the effect of Parlodel on serotoninreceptors ‘‘has been demonstrated in com-pany studies,’’ Aplt’s App. vol. II E, at1747, Dr. Kulig provides no details on themethodology or conclusions of these stud-ies. Moreover, the mere fact that Parlodelacts on serotonin and dopamine receptorsdoes not establish that Parlodel itself, asopposed to some other agent that triggerseither the release of these neurotransmit-ters or some other physiological mecha-nism, causes vasoconstriction, hyperten-sion, and stroke.

The testimony of Dr. Jose (the experton peripheral dopamine receptors) revealssimilar deficiencies. Although he present-ed an elegant theory of the way in whichbromocriptine might have the paradoxicaleffect of causing vasoconstriction and in-creased blood pressure, he acknowledgedthat there were a number of animal stud-ies that concluded that bromocriptine de-creases blood pressure. Dr. Jose at-tempted to distinguish these studies byobserving that the specific conditions thathe had posited as necessary to trigger theparadoxical effects of bromocriptine werenot present. However, he acknowledgedthat his thesis had not been tested. SeeAplt’s App. vol. II–A, at 802 (‘‘I havebits and pieces proving that bromocriptinecan increase sodium reabsorption, bitsand pieces that can show that bromocrip-tine can decrease blood flow; but to putall of them together [to demonstrate thatbromocriptine can cause high blood pres-sure] no, that has not been tested.’’). Ac-cordingly, the district court did not exer-cise manifestly unreasonable judgment inconcluding that the opinions of the Hol-landers’ experts did not meet the Daubert

reliability standard insofar as those opin-ions were based on the pharmacology ofbromocriptine.

c. Studies of hypertension

Both Dr. Kulig and Dr. Iffy relied onstudies regarding the relationship betweenbromocriptine and hypertension. Dr. Ku-lig referred to a 1981 study commissionedby Sandoz in which women received thedrug to treat amennorrhea-galactorrheasyndrome.12 Dr. Kulig states that nine-teen percent of those women reported in-creases in blood pressure. Both expertsinvoke a study by D.I. Watson, which con-cluded that women with pregnancy-in-duced hypertension who then took bromo-criptine after giving birth had a higherincidence of postpartum hypertension thanwomen who did not receive the drug.13

Again, it is not an abuse of discretion toconclude that there is ‘‘simply too great ananalytical gap’’ between these studies andthe experts’ conclusion that Parlodelcaused Ms. Hollander’s stroke. See Join-er, 522 U.S. at 146, 118 S.Ct. 512. Thestudies in question do not directly addressthe relationship between Parlodel andstroke. Moreover, the Hollanders pre-sented no expert analysis as to how onemight extrapolate from bromocriptine’s ef-fect on a small group of women with amen-norrhea-galactorrhea syndrome to deter-mine the effect that the drug would haveon women like Ms. Hollander who took thedrug as a postpartum lactation suppres-sant.

As to the Watson study, Dr. Iffy admit-ted that it did not claim a ‘‘high degree ofreliability.’’ Aplt’s App. vol. II–A, at 698.

12. Amenorrhea refers to an ‘‘absence or ab-normal stoppage of the menses.’’ See Dor-land’s Illustrated Medical Dictionary 55. Ga-lactorrhea is the ‘‘excessive or spontaneousflow of milk’’ or the ‘‘persistent secretion ofmilk irrespective of nursing.’’ Id. at 672.

13. D.I. Watson, et al., supra, Obstetrics andGynecology, 74(4): 573–576 (Aplt’s App. vol.II D, at 1596 97).


Dr. Iffy acknowledged that, under thestudy’s criteria for determining whichwomen had pregnancy-induced hyperten-sion, Ms. Hollander herself would not qual-ify. Id. Thus, she was not in the class ofpatients whose blood pressure increasedafter taking Parlodel.

d. Animal Studies

According to the Hollanders’ experts,there are certain animal studies that alsoprovide evidence that bromocriptine maycause vasoconstriction, hypertension, andstroke. These studies included those per-formed on isolated veins, on animals thatwere unconscious, and on pithed animals.Many involved large doses of bromocrip-tine, relatively much greater than the dos-es taken by Ms. Hollander.

Several recent decisions consideringthese studies have agreed with the districtcourt’s analysis. The studies suggest onlythat bromocriptine may act as a vasocon-strictor in very specific circumstances incertain kinds of animals; the studies donot constitute reliable evidence that bro-mocriptine causes strokes. See Glastetter,252 F.3d at 991 (noting that one of theplaintiff’s experts concluded that ‘‘not asingle animal study had ever concludedthat [intracerbral hemorrhage] was associ-ated with bromocriptine’’); Caraker v.Sandoz Pharms. Corp., 172 F.Supp.2d1046, 1050–51 (S.D.Ill.2001) (noting that‘‘some [studies] involved animals that hada steel rod injected down their spinal cordto destroy it so the animal has no intactnervous system, some involved bromocrip-tine’s reaction locally (e.g., in a single iso-lated vein of an animal) as opposed to asystemic administration; and some werepoorly documented’’); Siharath, 131F.Supp.2d at 1367–69 (discussing three an-imal studies in detail and concluding thatthey did not constitute a reliable basis forexperts’ opinions that Parlodel caused theplaintiff’s stroke).

In light of these characteristics of theanimal studies, the district court’s conclu-sion that they were unreliable does not‘‘exceed[ ] the bounds of permissible choicein the circumstances.’’ McEwen, 926 F.2dat 1553–54. We therefore discern noabuse of discretion in the court’s analysis.

e. Case studies and differential diag-nosis

The next methodologies employed bythe Hollanders’ experts present a closerquestion. ‘‘Differential diagnosis’’ refersto the process by which a physician‘‘ ‘rule[s] in’ all scientifically plausiblecauses of the plaintiff’s injury. The physi-cian then ‘rules out’ the least plausiblecauses of injury until the most likely causeremains.’’ Glastetter, 252 F.3d at 989 (8thCir.2001). The remaining cause is the ex-pert’s conclusion. Id. In conducting a dif-ferential diagnosis, physicians often usecase reports—‘‘a doctor’s account of a par-ticular patients’ reaction to a drug or otherstimulus, accompanied by a description ofthe relevant surrounding circumstances.’’Id.

Here, Drs. Kulig, Iffy, and Jose per-formed a differential diagnosis, reviewingMs. Hollander’s medical history and medi-cal records, excluding other causes of herstroke, and then attributing the stroke toParlodel. They relied in part on case re-ports, both those filed with the FDA andthose published in the professional litera-ture. Dr. Kulig expressed the view thatthe onset of Ms. Hollander’s initial symp-tom (i.e., developing a headache severaldays after giving birth) and the timing ofher stroke (several days after her dis-charge from the hospital) fit a generalpattern seen in patients suffering adversereactions to Parlodel and was also consis-tent with the pharmokinetics of the drug.

With regard to differential diagnoses,courts have reached contrasting conclu-sions as to reliability under Daubert.


Compare Westberry v. Gislaved GummiAB, 178 F.3d 257, 262–66 (4th Cir.1999)(holding that ‘‘[a] reliable differential diag-nosis provides a valid basis for an expertopinion on causation’’ and concluding thatthe district court did not abuse its discre-tion in admitting a physician’s opinion tes-timony based on differential diagnosis)with Glastetter, 252 F.3d at 989 (holdingthat a district court did not abuse of dis-cretion in excluding a differential diagnosisthat was ‘‘scientifically invalid’’); see alsoFederal Judicial Center, Reference Manu-al on Scientific Evidence 34 (2d. ed.2000)(noting that ‘‘[j]udges disagree on whethera physician relying on the methodology ofclinical medicine can provide adequateproof of causation in a toxic tort action’’).Courts have also reached contrasting con-clusions as to the reliability of case re-ports. Compare Glaser v. ThompsonMed. Co., 32 F.3d 969, 975 (6th Cir.1994)(holding that the district court abused itsdiscretion in excluding physician’s opiniontestimony based in part on case reports)with Casey v. Ohio Med. Prods., 877F.Supp. 1380, 1385 (N.D.Cal.1995) (statingthat ‘‘case reports are not reliable scienti-fic evidence of causation, because they sim-ply described reported phenomena withoutcomparison to the rate at which the phe-nomena occur in the general population orin a defined control group; do not isolateand exclude potentially alternative causes;and do not investigate or explain the mech-anism of causation’’); see generally Feder-al Judicial Center, Reference Manual onScientific Evidence 475 (noting that‘‘[c]ausal attribution based on case studiesmust be regarded with caution’’ but that

‘‘such studies may be carefully consideredin light of other information available, in-cluding toxicological data,’’ and citing casesreaching contrasting conclusions on theiradmissibility under Daubert ).14 Our cir-cuit does not appear to have addressed thereliability of differential diagnosis and casereports under Daubert.

In the instant case, the district courtmade short shrift of the Hollanders’ ex-perts’ differential diagnoses and relianceon case reports. The court stated that‘‘[b]ecause of their limitations, case reportshave been repeatedly rejected as a scienti-fic basis for a conclusion regarding causa-tion.’’ Hollander, 95 F.Supp.2d at 1237.

Because the Daubert reliability inquiryis case-specific, we need not address, ingeneral terms, the reliability of differentialdiagnoses and case reports. See KumhoTire, 526 U.S. at 150, 119 S.Ct. 1167. In-stead, we must only decide whether thedistrict court abused its discretion by char-acterizing the specific diagnoses and casereports at issue here as unreliable underDaubert.

Again, we conclude that the districtcourt did not abuse its discretion. Inmany of the decisions in which a differen-tial diagnosis has been deemed reliable,the party relying on the diagnosis hasoffered independently reliable evidencethat the allegedly dangerous drug or sub-stance had harmful effects. See, e.g., Zu-chowicz v. United States, 140 F.3d 381,385–87 (2d Cir.1998) (affirming admissionof differential diagnosis based in part onscientific articles regarding the effects of adrug); Kennedy v. Collagen Corp., 161

14. The conflicting views of the reliability ofdifferential diagnosis are apparent in the Par-lodel cases too. Compare Brasher, 160F.Supp.2d at 1296 (concluding that differen-tial diagnosis constitutes a reliable methodol-ogy under Daubert ) and Globetti, 111F.Supp.2d at 1178 (characterizing differentialdiagnosis as ‘‘a well-recognized and widely-

used technique relied on by medical cliniciansworldwide to identify and isolate the causes ofdisease’’) with Glastetter, 107 F.Supp.2d 1015(concluding that differential diagnosis andcase reports did not establish reliable proof ofcausation), and Siharath, 131 F.Supp.2d at1361–63 (same).


F.3d 1226, 1228–30 (9th Cir.1998) (holdingthat the district court abused its discretionin excluding expert opinion based on dif-ferential diagnosis when the diagnosis wassupported by scientific and clinical studiesregarding the connection between collagenand autoimmune disorders). That is notthe case here. In order to ‘‘rule in’’ Parlo-del as a scientifically plausible cause of Ms.Hollander’s stroke, the Hollanders’ expertswould need to present reliable evidencethat the drug can cause strokes, and forthe reasons we have discussed, the districtcourt did not abuse its discretion in con-cluding that the experts did not do so. SeeGlastetter, 252 F.3d at 989 (affirming thedistrict court’s exclusion of a differentialdiagnosis); cf. Siharath, 131 F.Supp.2d at1362–63 (‘‘[A] fundamental assumption un-derlying this method is that the final, sus-pected ‘cause’ remaining after this processof elimination must actually be capable ofcausing the injury. That is, the expertmust ‘rule in’ the other suspected cause aswell as ‘rule out’ other possible causes.And, of course, expert opinion on this issueof general causation must be derived fromscientifically valid methodology.’’) (internalquotation marks omitted).

We take a similar view of the case re-ports regarding other women sufferingvarious injuries after taking Parlodel.Many of these case reports contain onlylimited information regarding the medicalhistories of the patients and the nature ofthe injuries they suffered. In addition,given the large number of women whotook Parlodel and the variety of possiblecauses for many of these injuries, it wasnot unreasonable for the district court tocharacterize the reports as unreliable evi-dence of causation. See Siharath, 131F.Supp.2d at 1361 (noting that ‘‘case re-ports TTT do not isolate and exclude po-tentially alternative causes; and do not in-vestigate or explain the mechanism ofcausation’’ and that, as to Parlodel, ‘‘themodest number of case reports associating

the drug with stroke or even postpartumhypertension is not what would be expect-ed if there was a significant increasedrisk ’’) (internal quotation marks omitted).

In holding that the district court did notabuse its discretion in excluding the partic-ular differential diagnoses and case re-ports submitted by the Hollanders, we em-phasize that, in other litigation, there maywell be differential diagnoses and case re-ports that do not suffer from the samedeficiencies noted by the district courthere. For example, if the case reports inthis record contained more detailed infor-mation about other women who sufferedstrokes and heart attacks after taking Par-lodel, and if there were a substantiallylarger number of such detailed reports,that information might have provided sup-port for the theories of Drs. Kulig, Iffy,and Jose. In that instance, a district courtmight well be justified in finding opiniontestimony like that of Drs. Kulig, Iffy, andJose reliable under Daubert. See Caraker,172 F.Supp.2d at 1050 (stating that ‘‘anoverwhelming amount of case reports of atemporal proximity between a very specificdrug and a very specific adverse eventmight TTT be enough to make a generalcausation conclusion sufficiently reliable’’but adding that ‘‘[i]n this case, however,we have a scant number of case reports’’).

Moreover, as the Eighth Circuit haswritten:

[W]e do not believe that a medical ex-pert must always cite published studieson general causation in order to reliablyconclude that a particular object causeda particular illness. The first severalvictims of a new toxic tort should not bebarred from having their day in courtsimply because the medical literature,which will eventually show the connec-tion between the victims’ condition andthe toxic substance, has not yet beencompleted. If a properly qualified medi-cal expert performs a reliable differen-


tial diagnosis through which, to a rea-sonable degree of medical certainty, allother possible causes of the victims’ con-dition can be eliminated, leaving only thetoxic substance as the cause, a causationopinion based on that differential diag-nosis should be admitted.

Turner v. Iowa Fire Equip. Co., 229 F.3d1202, 1209 (8th Cir.2000) (internal quota-tion marks omitted); see also Westberry,178 F.3d at 262 (holding that a reliabledifferential diagnosis alone may provide avalid foundation for a causation opinion,even when no epidemiological studies,peer—reviewed published studies, animalstudies, or laboratory data are offered insupport of the opinion). However, in lightof the deficiencies of the particular differ-ential diagnoses and case reports in thisrecord, the district court’s analysis was notunreasonable.

f. Rechallenge and dechallenge reports

Ms. Hollanders’ experts also presentedseveral accounts of rechallenge and dechal-lenge. Rechallenge occurs when a patientis exposed to the same drug thought tohave previously caused an adverse reac-tion. Dechallenge occurs when the drug isremoved. As the Eighth Circuit has not-ed, rechallenge and dechallenge resemblecontrolled experiments in some ways, andthus may be more valuable than typicalcase reports. Glastetter, 252 F.3d at 990–91. Occasionally the results may appearquite dramatic. For example, one Sandoz

employee described a rechallenge-dechal-lenge report as ‘‘ ‘the smoking gun’ wehave looked for so diligently.’’ Aplt’s App.vol. IV–B, at 2364.15

Nevertheless, by the Hollanders’ ac-count, there were only three such reports.See Aplt’s Br. at 31. Of these three, onlyone involved vasoconstriction of the cere-bral blood vessels.16 Moreover, as to thatincident, the only information to which theHollanders’ have directed us is a second-hand account by a Sandoz physician.Thus, the district court conclusion here—that there were too few of these reportsfor them to constitute reliable evidence ofcausation under Daubert—was not unrea-sonable. Cf. Glastetter, 252 F.3d at 990(finding rechallenge and dechallenge datato be more potent proof of causation thandid the district court but further conclud-ing that the district court did not abuse itsdiscretion in excluding it).

g. The Risks of Stroke in the Postpar-tum period

A linchpin of Sandoz’s defense was thatthere is an increased risk of stroke in thepostpartum period generally. Sandoz re-lied heavily on a study concluding thatthere is a 28.3 relative risk of stroke forwomen in the postpartum period, as com-pared with non-pregnant women. See Ste-ven J. Kittner, et al., Pregnancy and theRisk of Stroke, New Eng. J. Med. 768–74(1996) (Aplt’s App. vol. II–C, at 1335–41).17 Thus, according to Sandoz, Ms. Hol-

15. A physician reported to Sandoz that hehad induced vasoconstriction of the cerebralblood vessels by administering small doses ofParlodel to a woman who had previously suf-fered a stroke after taking the drug. SeeAplt’s App. vol. IV B, at 2364–65 (memoran-dum from Dr. William F. Westin, dated Sept17, 1987). There appears to be no furtherdiscussion of this report in the record.

16. One of the incidents involved the vasocon-striction of a coronary artery and anotherinvolved hypertension.

17. One authority explains the concept of ‘‘rel-ative risk’’ as follows:

[Relative risk] is defined as the ratio of theincidence rate (often referred to as inci-dence) of disease in exposed individuals tothe incidence rate in unexposed individuals.TTTT

The incidence rate of disease reflects thenumber of cases of disease that developduring a specified period of time divided bythe number of persons in the cohort understudy. Thus, the incidence rate expresses


lander’s stroke could well have been theresult of the increased risk to which allwomen are exposed during the postpartumperiod rather than an adverse reaction toParlodel.

In the district court proceedings, theHollanders challenged this argument pri-marily through the testimony of Dr. Kulig,who stated that the Kittner study (andothers reaching similar conclusions aboutthe risks of the postpartum period) did notcontrol for bromocriptine use or for specif-ic conditions that increase the risk ofstroke, such as eclampsia. According toDr. Kulig, when these factors are exclud-ed, the Kittner study does not establishthat there is an increased risk of stroke inthe postpartum period.

The district court acknowledged Dr. Ku-lig’s criticisms of the study. However, thecourt found that ‘‘the postpartum incidenceof stroke is a factor that should be consid-ered.’’ See Hollander, 95 F.Supp.2d at1238 n. 21.

The Hollanders now argue that the dis-trict court erred in its qualified affirmationof the Kittner study in the face of Dr.Kulig’s critique of it. If this case required

Sandoz to prove that there was an in-creased risk of stroke during pregnancy,we might agree. However, no such bur-den is imposed on Sandoz here. Instead,it is the Hollanders who have the burdenof demonstrating the harmful effect ofParlodel. Accordingly, it was not unrea-sonable for the district court to concludethat Dr. Kulig’s attack on the Kittnerstudy did not constitute reliable evidencethat Parlodel caused Ms. Hollander’sstroke.

In summary, we agree with the court’sassessment in Siharath: the Hollandershave done the best they could with theavailable data and the scientific literature.See 131 F.Supp.2d at 1373. The data onwhich they rely might well raise seriousconcerns in conscientious clinicians seekingto decide whether the benefits of the drugoutweigh its risks. However, in derivingtheir opinions that Parlodel caused Ms.Hollander’s stroke from the varioussources we have outlined, Drs. Kulig, Iffy,and Jose all made several speculativeleaps. As a result, the district court didnot abuse its discretion in excluding theirtestimony under Daubert.

the risk that a member of the populationwill develop the disease within a specifiedperiod of time.

For example, a researcher studies 100individuals who are exposed to an agentand 200 who are not exposed. After oneyear, 40 of the exposed individuals are diag-nosed as having a disease, and 20 of theunexposed individuals are also diagnosedas having the disease. The relative risk ofcontracting the disease is calculated as fol-lows:

-The incidence rate of disease in the ex-posed individuals is 40 cases per year per100 persons (40/100), or 0.4

-The incidence rate of disease in theunexposed individuals is 20 cases per yearper 200 persons (20/200), or 0.1.

A relative risk of 4.0 indicates that therisk of disease in the exposed group is fourtimes as high as the risk of disease in theunexposed group.

TTTT

Although a relative risk is a straightfor-ward concept, care must be taken in inter-preting it. Researchers should scrutinizetheir results for error. Error in the designof the study could yield an incorrect relativerisk. Sources of bias and confoundingshould be examined. Whenever an associa-tion is uncovered, further analysis shouldbe conducted to determine if the associa-tion is real or due to an error or bias.Similarly, a study that does not find anassociation between an agent and diseasemay be erroneous because of bias or ran-dom error.

Federal Judicial Center, Reference Manual onScientific Evidence 348–49.

Thus, in the instant case, the study cited bySandoz concluded that women in the postpar-tum period were 28.3 times more likely thannon-pregnant women to have suffered astroke.


C. Grant of Summary Judgmentto Sandoz

[6] In a related argument, the Hol-landers maintain that, even if the districtcourt did not abuse its discretion in exclud-ing their experts’ testimony, the court nev-ertheless erred in granting summary judg-ment to Sandoz.

According to the Hollanders, the follow-ing evidence demonstrates that there arecontroverted issues of material fact as towhether Parlodel caused her stroke: (1)the FDA’s determination that Parlodel hadnot been shown to be safe when prescribedas a postpartum lactation suppressant; (2)the judgment entered against Sandoz in acase in Kentucky involving Parlodel; (3)incidents of dechallenge and rechallenge;(4) case reports; (5) studies of hyperten-sion; (6) the fact that bromocriptine is anergot; (7) animal studies; and (8) epidemi-ological studies.

We engage in de novo review of thedistrict court’s summary judgment ruling,applying the same standard as the districtcourt under Fed.R.Civ.P. 56(c). See Adlerv. Wal–Mart Stores, Inc., 144 F.3d 664,670 (10th Cir.1998). Summary judgmentis appropriate ‘‘if the pleadings, deposi-tions, answers to interrogatories, and ad-missions on file, together with the affida-vits, if any, show that there is no genuineissue as to any material fact and that themoving party is entitled to a judgment as amatter of law.’’ Rule 56(c). We view thefacts and the reasonable inferences to bedrawn from them in the light most favor-able to the nonmoving party. Adler, 144F.3d at 670.

[7] Under Oklahoma law, which we ap-ply in this diversity case, see Wood v. EliLilly & Co., 38 F.3d 510, 512 (10th Cir.1994), ‘‘[a] plaintiff seeking recovery for an

injurious side effect from a properly manu-factured prescription drug must prove thatthe drug caused the injury and that themanufacturer breached a duty to warn ofpossible detrimental reactions.’’ McKee v.Moore, 648 P.2d 21, 23 (Okla.1982). Cau-sation is established if ‘‘in a natural andcontinuous sequence, unbroken by an inde-pendent cause’’ the drug produces an inju-ry that would not have occurred if it hadnot been administered. See Gaines v.Providence Apartments, 750 P.2d 125,126–27 (Okla.1987) (defining proximatecause).

We need not address the Hollanders’argument in detail. We have alreadyruled that five of the eight categories ofevidence on which they rely did not consti-tute sufficiently reliable grounds underDaubert for their experts’ opinions.18 As aresult, these categories of evidence do notraise questions of fact on issues of causa-tion.

[8] Moreover, under Oklahoma law, aplaintiff must introduce expert testimony if‘‘the fact in issue is not within the realm ofordinary experience of mankind.’’ Strub-hart v. Perry Mem’l Hosp. Trust Auth.,903 P.2d 263, 274 (Okla.1995). Here, thealleged effect of Parlodel is not within therealm of ordinary experience: in order toassess the arguments regarding the al-leged effects of the drug, the factfinderwould be required to assess the wide vari-ety of scientific evidence that we havediscussed here. As a result, the Holland-ers cannot prove their claim without ex-pert testimony.

Finally, we consider briefly the threecategories of evidence that we have not yetaddressed—the FDA determination, thejudgment in the Roberts case, and theepidemiological studies. None of this evi-

18. These categories of evidence are as fol-lows: incidents of dechallenge and rechal-lenge; case reports; studies of hypertension;

the fact that bromocriptine is an ergot; andanimal studies.


dence provides sufficient support for theHollanders’ claims.

As to the FDA determination, this cir-cuit has noted that differing standards mi-litate against applying regulatory actionsto the elements of tort law. See Mitchell,165 F.3d at 783 n. 3 (10th Cir.1999). Inassessing a district court’s application ofDaubert, we discounted a state agency’sclassification of a substance as a carcino-gen, stating that the methodology em-ployed by the agency ‘‘results from thepreventive perspective that the agenciesadopt in order to reduce public exposure toharmful substances,’’and that ‘‘[t]he agen-cies’ threshold of proof is reasonably lowerthan that appropriate in tort law.’’ Id.(internal quotation marks omitted).

Moreover, several courts have concludedthat this specific FDA ruling about Parlo-del is not relevant to the causation ques-tion. See Glastetter, 252 F.3d at 991 (not-ing that the FDA ruling is not reliableevidence of causation for two reasons: (1)because the FDA ‘‘balanced Parlodel’s pos-sible harm against its limited beneficialuse,’’ an irrelevant consideration in theDaubert inquiry; and (2) because ‘‘[t]heFDA will remove drugs from the market-place upon a lesser showing of harm to thepublic than the standards used to assesstort liability’’); Siharath, 131 F.Supp.2d at1366 (rejecting the plaintiff’s reliance on

the FDA ruling). Moreover, the languageused in the FDA ruling regarding thewithdrawal of Parlodel indicates that theagency did not make a determination thatParlodel causes seizures and strokes.19

Accordingly, we conclude that the FDAruling does not establish that there arecontroverted factual issues as to whetherParlodel caused Ms. Hollanders’ stroke.20

As to the judgment in the Roberts case,the Hollanders fail to explain its relevance.The case was decided under Kentucky law,and our decision is governed by differentlaw and different facts. Moreover, in lightof the district court’s broad discretion inthese matters, the fact that differentcourts reach difference conclusions as toreliability under Daubert does not estab-lish that a legal error has been made byone or the other. See Brasher, 160F.Supp.2d at 1299 n. 17 (noting that incon-sistent rulings Daubert rulings do not nec-essarily establish an abuse of discretion).

Finally, the epidemiological studies inquestion do not support the Hollanders’claim. The district court accurately ob-served that the Hollanders’ own expertsdid not rely on these studies. Moreover,as the district court further noted, ‘‘[a]t-hough several studies have been conductedregarding Parlodel and stroke, none hasshown a statistically significant link be-tween them.’’ Hollander, 95 F.Supp.2d at

19. The FDA ruling states that the evidencereceived by the FDA ‘‘calls into question bro-mocriptine’s safety,’’ that bromocriptine ‘‘maybe an additional risk factor in patients whoare already at risk for seizures and stroke,’’and that the FDA had obtained new evidence‘‘suggesting that therapeutic use of bromocrip-tine for the prevention of physiological lacta-tion may lead to serious adverse experi-encesTTTT’’ 59 Fed.Reg. 43348, 43351 (Aug.24, 1994) (emphasis added).

20. Our conclusion about the FDA’s decisionto withdraw the indication for Parlodel as alactation suppressant should not be read tosuggest that, as a general rule, regulatory

decisions lack the intellectual rigor necessaryunder the Daubert reliability inquiry. Indeed,some authorities view the review process inthe regulatory area as typically ‘‘far morecareful and systematic’’ than the peer reviewprocess employed by scientific journals. SeeKenneth R. Foster & Peter W. Huber, JudgingScience: Scientific Knowledge in the FederalCourts 174 (1997). ‘‘Regulators require thatdocuments submitted to them contain farmore detail than is typically found in paperssubmitted to professional journals TTTT [and]administrative reports-not peer reviewed jour-nals—may provide parties with the solidestavailable data.’’ Id. at 174–75.


1236. See also Siharath, 131 F.Supp.2d at1356–59 (discussing four studies finding nostatistically significant association).21

Accordingly, we conclude that the dis-trict court properly granted Sandoz’s mo-tion for summary judgment.

D. Dismissal of Sandoz, Ltd.

[9] Prior to ruling on the Daubert andsummary judgment motions, the districtcourt dismissed the Hollanders’ claimsagainst the defendant Sandoz, Ltd., a com-pany incorporated in Switzerland with itsprincipal place of business there. Prior toJanuary 1, 1990, Sandoz, Ltd. sold Parlo-del in bulk to the defendant Sandoz Corpo-ration. However, after that date, Sandoz,Ltd. became a holding company, conduct-ing no advertising in the United States andowing no manufacturing, distribution, orsales facilities here. In granting Sandoz’smotion to dismiss, the district court rea-soned that Sandoz, Ltd. ‘‘does not have abank account, a telephone number or anyemployees, officers or directors in thiscountry; and that it does not actively ad-vertise in the United States.’’ Aplt’s App.vol. I, at 344 (Dist. Ct. Order, filed Dec. 10,1996, at 1). Accordingly, the court con-cluded that it lacked personal jurisdictionover Sandoz, Ltd., and it dismissed theclaims against Sandoz, Ltd. with prejudice.

On appeal, the Hollanders argue that,because Sandoz, Ltd. sold Parlodel to itsUnited States subsidiary (Sandoz) prior toJanuary 1, 1990, Sandoz, Ltd. should bedeemed to have continued to do businessin the United States through July 1990,when Ms. Hollander suffered her stroke.

They contend that it is unlikely that San-doz, Ltd. would have completely ceaseddoing business in the United States in theseven month period beginning in January1990 (when it became a holding company)and July 1990 (when Ms. Hollander tookParlodel). Additionally, the Hollanders ar-gue that the district court erred in dis-missing the claims against Sandoz, Ltd.with prejudice.

The Hollanders’ challenge to the court’sjurisdictional ruling raises a legal questionthat we review de novo. See Wenz v.Memery Crystal, 55 F.3d 1503, 1505 (10thCir.1995). Their argument is underminedby precedent that imposes the burden ofproof on the party asserting jurisdiction.See id. Because they have offered noevidence to rebut Sandoz, Ltd.’s evidencethat it did not do business in the UnitedStates after January 1, 1990, we concludethat the district court properly held that itlacked jurisdiction over the company.

However, we further conclude that thedistrict court should not have dismissedthe Hollanders’ claim against Sandoz, Ltd.with prejudice. Its jurisdictional rulingdid not address the merits of the Holland-ers’ allegations as to Sandoz, Ltd., and, asa result, the claim against Sandoz, Ltd.should have been dismissed without preju-dice to filing in an appropriate forum. SeePosner v. Essex Ins. Co., Ltd., 178 F.3d1209, 1221 (11th Cir.1999) (concluding thatthe district court erred in dismissingclaims against a party with prejudice onjurisdictional grounds and instructing thedistrict court to dismiss the claims withoutprejudice); Arrowsmith v. United Press

21. The Hollanders also suggest that a totalityof the circumstances approach establishesthat there are controverted issues of materialfact. In essence they maintain that eventhough each individual category of evidencemay be insufficient, all of the evidence consid-ered as a whole raises factual questions as towhether Parlodel caused her stroke. TheHollanders cite no legal authority in support

of this approach, and in our view, this argu-ment is inconsistent with Daubert. To suggestthat those individual categories of evidencedeemed unreliable by the district court maybe added to form a reliable theory would beto abandon ‘‘the level of intellectual rigor’’ ofthe expert in the field. Kumho Tire, 526 U.S.at 152, 119 S.Ct. 1167.


Int’l, 320 F.2d 219, 221 (2d Cir.1963) (‘‘Adismissal for lack of jurisdiction TTT doesnot preclude a subsequent action in anappropriate forum.’’).

III. CONCLUSION

This case illustrates the continuing im-portance of the Supreme Court’s observa-tion in Daubert:

[T]here are TTT differences between thequest for truth in the courtroom and thequest for truth in the laboratory. Scien-tific conclusions are subject to perpetualrevision. Law, on the other hand, mustresolve disputes finally and quickly.The scientific project is advanced bybroad and wide-ranging consideration ofa multitude of hypotheses, for those thatare incorrect will eventually be shown tobe so, and that in itself is an advance.Conjectures that are probably wrongare of little use, however, in the projectof reaching a quick, final, and bindinglegal judgment often of great conse-quence about a particular set of eventsin the past. We recognize that, in prac-tice, a gatekeeping role for the judge, nomatter how flexible, inevitably on occa-sion will prevent the jury from learningof authentic insights and innovations.That, nevertheless, is the balance that isstruck by Rules of Evidence designednot for the exhaustive search for cosmicunderstanding but for the particularizedresolution of legal disputes.

Daubert, 509 U.S. at 596–97, 113 S.Ct.2786. Thus, in Judge Posner’s words, ‘‘the

courtroom is not the place for scientificguesswork, even of the inspired sort. Lawlags science; it does not lead it.’’ Rosen,78 F.3d at 319.

Here, the district court characterizedthe Hollanders’ evidence as such guess-work, and that characterization was notunreasonable. The Hollanders’ evidenceprovided support for the FDA’s decision towithdraw the indication for Parlodel as apostpartum lactation suppressant, as wellas for the decisions of experienced clini-cians that the apparent risks of Parlodeloutweighed the limited benefits of pre-scribing the drug as a lactation suppres-sant. However, the district court did notabuse its discretion in ruling that the Hol-landers’ evidence did not satisfy the Dau-bert standard of reliability.

Additionally, the district court did noterr in denying the Hollanders’ motion toremand the case to the Oklahoma statecourts or in dismissing the claim againstSandoz, Ltd. However, the court did err indismissing the claim against Sandoz, Ltd.with prejudice.

Accordingly, we AFFIRM the judgmentof the district court in all respects EX-CEPT that we REMAND the Hollanders’claim against Sandoz, Ltd. with instruc-tions to dismiss that claim without preju-dice.22

,

22. In light of our conclusion that the federaldistrict court did not abuse its discretion inconcluding that Hollanders’ expert testimonywas unreliable under Daubert, as well as ourconclusion that the court did not err in grant-ing summary judgment to Sandoz, we neednot address the Hollanders’ challenge to thedismissal of their products liability claimagainst Presbyterian Hospital. Even assum-ing that the dismissal was improper, the Hol-landers have failed to explain why the samereliability problems noted by the federal dis-

trict court do not defeat the Hollanders’ prod-ucts liability claim against Presbyterian Hos-pital.

We do note, as Presbyterian Hospital ob-serves in its response brief, that an over-whelming majority of jurisdictions have re-fused to apply strict liability principles toclaims against hospitals and physicians in-volving the distribution of allegedly danger-ous drugs or medical devices. See, e.g, Royerv. Catholic Med. Ctr., 144 N.H. 330, 741 A.2d

hollander v. sandoz pharmaceuticals … v. sandoz pharmaceuticals corp.1193 cite as 289 f.3d 1193...

Documents