786 F.2d 859
54 USLW 2508, Prod.Liab.Rep.(CCH)P 10,960
Maria MULLIGAN, Appellee,
v.
LEDERLE LABORATORIES, a DIVISION OF AMERICAN CYANAMID
COMPANY, Appellant.
No. 85-1111.
United States Court of Appeals,
Eighth Circuit.
Submitted Oct. 18, 1985.
Decided March 20, 1986.
Alston Jennings, Little Rock, Ark., for appellant.
Charles Phillip Boyd, Jr., Little Rock, Ark., for appellee.
Before LAY, Chief Judge, ARNOLD and FAGG, Circuit Judges.
ARNOLD, Circuit Judge.
Lederle Laboratories brings this appeal to challenge jury verdicts of $50,000 in compensatory and $100,000 in punitive damages awarded to Maria Mulligan. Mrs. Mulligan, a medical laboratory technician, brought this action in 1979, alleging that Varidase, a drug manufactured by Lederle and given to her in 1960, caused her to develop several chronic health problems, including mouth sores in 1962, microscopic hematuria in 1967, and red cell cast, indicating kidney disease, in 1976.
The defendant makes two arguments on appeal. First, it claims the District Court1 incorrectly refused to hold that the three-year statute of limitations had run on Mrs. Mulligan's cause and, instead, submitted the limitations question to the jury. Second, Lederle argues that there was insufficient evidence to support an award of punitive damages, and that the District Court should have either refused to submit the question to the jury or granted defendant's motion for judgment notwithstanding the verdict. Plaintiff responds to the merits of defendant's claims but also relies on the District Court's ruling that Lederle failed to renew its motion for a directed verdict at the close of its case, and, under Fed.R.Civ.P. 50, has therefore waived its right to ask for judgment notwithstanding the verdict or to challenge the factual findings made by the jury. The defendant replies that it substantially complied with the dictates of Rule 50. We assume for purposes of this appeal that Lederle's motions were properly preserved, and affirm on the merits.
I.
Varidase was a biological drug product developed by Lederle about 1949 for the treatment of inflammation and to break up deposits of clotted blood as well as accumulations in tissues and tissue surfaces of fluid, cells, and cellular debris which have escaped from blood vessels. The product was made of a mixture of the enzymes streptokinase and streptodornase, which had been extracted from streptococcus bacteria. During the 1950's, Lederle first began producing Varidase to be applied directly to clots and inflammation in topical form; later, an intramuscular product was developed, and finally buccal tablets were produced to be held in the mouth and dissolved.2 Although the precise scientific mechanisms have never been understood, streptokinase has been recognized for some time as an activator of plasminogen, capable of breaking up blood clots.
Maria Mulligan, then 27 years old, was given Varidase by intramuscular injection in February and March 1960 to break up blood clots caused by the episiotomy performed during the delivery of her first child on February 21, 1960, in Lake Village, Arkansas. She received a total of seven injections and was released from the hospital on March 4. In 1962, Mrs. Mulligan developed oral sores or lesions, which have recurred since that time. Treatment of the sores, known as aphthous stomatitis, has been successful only very briefly; they disappeared completely only when Mrs. Mulligan was again pregnant. In 1967, when doing her own urinalysis, Mrs. Mulligan first noticed microscopic hematuria, red blood cells that could be seen only by microscope. This condition persisted and went undiagnosed, despite visits to various physicians, as well as frequent conversations with the doctors at the Lake Village clinic where she worked. She testified that she assumed the problem was related to her bladder until 1976, when her urinalysis revealed red blood cell cast, indicating glomerulonephritis, a kidney condition. Also in 1976, Mrs. Mulligan came across an article on drug-induced nephritis, which led her to connect her condition to the Varidase she was given in 1960. In response to a letter drafted by Mrs. Mulligan and signed by her physician, a Lederle official said the company had no knowledge of any possible link between the drug and nephritis or her other symptoms.3 Continued investigation by plaintiff led her to the conclusion that such a link was possible, and she filed suit in June 1979.4
At trial, plaintiff introduced evidence showing that in the 1950s, batches of Varidase were unpredictably pyrogenic, that is, fever-causing. Both internal Lederle documents and company letters to physicians inquiring about apparent patient reactions to Varidase indicate that a majority of those given Varidase suffered some sort of fever, and others also suffered mouth ulcers, nausea, hematuria, cystitis, nephritis, muscle aches, and anaphylactic shock. Plaintiff's expert witnesses testified that the defendant's description in the 1960 Physician's Desk Reference (PDR) of the drug's possible side effects were not "full proper disclosure," Tr. at 204, and "grossly inadequate," id. at 396. The internal documents also suggested problems with the product's purity. One expert criticized the way Lederle tested for pyrogenicity and toxicity, and both witnesses suggested the drug, which was removed from the market in 1981, had never been effective.
As to plaintiff's particular circumstances, Mrs. Mulligan's 1960 hospital records were used to suggest that she had had a febrile reaction shortly after being given Varidase, although her experts disagreed on whether such a reaction had in fact occurred. Dr. Russell Steele, professor of pediatrics and director of the clinical immunology center for the University of Arkansas for Medical Sciences, said of Varidase and plaintiff's condition that "there is a tie-in, this is a streptococcal antigen, with a production of kidney disease that's called nephritis or glomerulonephritis, ... and I feel it is, the most probable precipitating event for her development of hematuria and red cell cast." Tr. at 468-469. As support for the theory that Varidase caused plaintiff's condition, Dr. Steele, who was described at trial as the only physician-clinical immunologist in Arkansas, cited a variety of articles and studies, dating from both before and after 1960, including a reference in a 1971 book written by Lederle's director of medical communications.
On cross-examination, Dr. Steele acknowledged that the medical literature indicated that the longest reported reaction to Varidase was three months in duration and that there were no conclusive data to show plaintiff's mouth sores were caused by intramuscular Varidase.5 Defendant presented two physicians as experts to rebut plaintiff's theory. One, the director of transplantations at the University of Arkansas Medical Center, testified that he examined Mrs. Mulligan in 1976 and found no evidence of glomerulonephritis. Both expert witnesses said they believed there was no connection between Mrs. Mulligan's condition and her 1960 treatment with Varidase. Defendant also emphasized the unsubstantiated nature of physician complaints about the drug, as well as the fact that each lot of Varidase was approved by the federal Bureau of Biologics.
II.
Mrs. Mulligan was injected with Varidase in 1960; this lawsuit was filed in 1979. The parties agree that the applicable limitations period is contained in the Arkansas Product Liability Act of 1979, Ark.Stat. Secs. 34-2801 et seq., which requires that such a cause of action be "commenced within three (3) years after the date on which the death, injury, or damage complained of occurs." Ark.Stat. Sec. 34-2803. The question here is whether the limitations period started to run prior to June 6, 1976, three years before the suit against Lederle was filed. The District Court denied defendant's motion asserting that the statute had run as a matter of law and instead instructed the jury to find for Lederle if Mrs. Mulligan's injuries were "manifested" or if she received an "informed diagnosis" before the critical date. Tr. at 737-738. In addition, the court told the jury a statutory limitations period is tolled if a defendant committed "some positive act of fraud which would conceal a plaintiff's cause of action." Id. at 743. Plaintiff argues before this Court that the jury's rejection of the limitations defense could be upheld on either of the grounds presented to it by the District Court. Defendant, however, claims that the limitations period began either in 1960 when Mrs. Mulligan was given Varidase, or at the latest in 1967, when she first found microscopic hematuria in her urine.
We begin by looking at the statute itself. In product-liability cases, Section 34-2803 supplants Ark.Stat. Sec. 37-206, which requires a lawsuit be filed "within three (3) years after the cause of action shall accrue." There appears to be no particular reason for the change in language to "after the date on which the death, injury, or damage complained of occurs." One commentator called the "failure ... to use the same language ... puzzling," Comment, The Arkansas Product Liability Act of 1979, 35 Ark.L.Rev. 364, 369 (1981), and suggested Sec. 34-2803 represents either a codification of the case law developed around Sec. 37-206 or perhaps an enlargement of the previous limitations period since "damage complained of" could include later injuries developed more than three years after the original harm, but within three years of the current action. The statute clearly rejects the concept found in some older Arkansas cases that a cause of action accrues when the tort is complete; for example, in an automobile accident case "at the moment the car was turned over," Faulkner v. Huie, 205 Ark. 332, 336, 168 S.W.2d 839, 841 (1943). Obviously, the plaintiff must suffer some harm before the new limitations period starts to run. This also contrasts sharply with the wording of the two-year limitations period for medical malpractice, which begins at "the date of the wrongful act complained of, and no other time." Ark.Stat. Sec. 37-205 (emphasis added).6
Except for its rejection of the principle of immediate accrual, Sec. 34-2803 appears to represent an acceptance of Arkansas case law under the earlier statute. We are particularly persuaded by the fact that in its only opinion dealing with the new products liability limitations period, the Arkansas Supreme Court without comment cited one of these earlier "accrual" cases. Spickes v. Medtronic, Inc., 275 Ark. 421, 423, 631 S.W.2d 5, 7 (1982), citing Field v. Gazette Publishing Co., 187 Ark. 253, 59 S.W.2d 19 (1933). Our interpretation of Sec. 34-2803 will therefore be governed by the case law developed to deal with Sec. 37-206, by both the Arkansas Supreme Court and this Court.
When the three-year period of limitations starts is a complicated question. Spickes, the Arkansas Supreme Court's most recent opinion on this issue, dealt with a plaintiff whose defective pacemaker had to be surgically removed and replaced in 1977; his lawsuit against the manufacturer was brought three years and three weeks later.7 Mr. Spickes contended that the action was timely, based on his 1980 discovery of defendant's 1977 failure to warn him of possible defects in his pacemaker and his resulting fear that the company might again neglect to notify him of defects. The court rejected this theory.
Medtronic committed no tortious act in 1980; Spickes simply found out more about what had been done three years earlier. The statute of limitations begins to run when the negligent damage occurs, not from the time the full extent of the injury is ascertained. Field v. Gazette Publishing Co., 187 Ark. 253, 59 S.W.2d 19 (1933).
275 Ark. at 423, 631 S.W.2d at 6-7. The first sentence of this passage seems to indicate that the critical moment is the tortious act. However, the second sentence, referring to the language of the 1979 statute, returns the focus of the inquiry to the damage suffered by plaintiff. Although Field itself spoke of the "date of the negligent act and not ... the time the full extent of the injury may be ascertained," 187 Ark. at 256, 59 S.W.2d at 20 (emphasis added), then-District Judge Henley has persuasively demonstrated that Field should be taken to refer to the manifestation of the harm alleged.8 Schenebeck v. Sterling Drug Co., 291 F. Supp. 368, 374 (E.D.Ark.1968), aff'd, 423 F.2d 919 (8th Cir.1970). The difficulty, then, is determining what level of manifestation is sufficient to start the limitations period and then distinguishing that from "the full extent of the injury."
Schenebeck gives considerable guidance on this question.9 The plaintiff in that case had from 1958 to 1963 taken a prescription arthritis medication known to produce temporary blurring of vision. Her suit against the manufacturer was not filed until 1966, when she learned after much consultation with a number of physicians that her eyes had been permanently damaged. In holding for the plaintiff, Judge Henley went beyond the concept of mere manifestation of symptoms. He noted the significant differences between the lead poisoning at issue in Field, a "well-known recognized industrial disease[ ]," 291 F. Supp. at 374, and "an insidious and slowly developing side effect of a medicinal drug ingested pursuant to a physician's prescription." Id. He suggested the moment of accrual should be established by reference to a plaintiff's awareness of his or her condition, including both the fact of the injury and the probable causal connection between the injury and a product's use; in essence, the court or jury must determine the time at which an individual would have sufficient knowledge and understanding of an alleged injury to initiate a lawsuit.
In its discussion of Schenebeck, Lederle stresses the fact that in affirming the District Court, this Court commented that "[s]light damage initiates the accrual of the cause of action." 423 F.2d at 925 (emphasis added). That emphasis, however, disregards this Court's actual analysis of Mrs. Schenebeck's situation, which dealt less with the physical appearance of her symptoms than with the nature of the disease and the difficulty in detecting the "slow and progressive" onset of its permanent damage. Id. This would indicate that "slight damage" refers to the quality and not the amount of harm; the limitations period can begin running if the manifested damage, even if slight, reveals the nature of the injury. For example, Mrs. Schenebeck's knowledge that her vision was blurry could not be considered adequate for limitations purposes until she knew of its probable permanence. Mr. Spickes, on the other hand, understood the nature of his injury at the time his defective pacemaker was removed; his allegations dealt only with his discovery of the scope of the defendant's conduct regarding the product which caused the injury. To know the nature of an injury, therefore, is not necessarily to know its "full extent," that is, precisely how much damage has been inflicted. This concept of the nature of a plaintiff's condition could be described either as his or her awareness of the condition, as outlined in the District Court opinion in Schenebeck, or by the notion of "informed diagnosis" as presented to the jury in Mrs. Mulligan's case.
Here, the jury had sufficient evidence to conclude either that the true nature of Mrs. Mulligan's illness did not manifest itself until 1976 or that it took until 1976 for Mrs. Mulligan to obtain a diagnosis that informed her of the nature and cause of her condition. The plaintiff, a trained medical technician, did find miscroscopic red blood cells in her urine in 1967, but was unsuccessful in her diligent efforts to locate a physician who could diagnose her condition. Only in 1976 did two events lead her to the diagnosis that is the basis of this lawsuit: in March or April, she first discovered cast in her urine, which she knew to be a sign of glomerulonephritis, and in June, she came across an article on drug-related nephritis.
III.
In its second ground on appeal, the defendant argues that the District Court erred in submitting the question of punitive damages to the jury and in not granting a judgment notwithstanding the $100,000 punitive-damage verdict returned for plaintiff. Under Arkansas law, punitive damages are appropriate when a defendant knew or should have known that its course of conduct was about to inflict injury and yet continued its activities with conscious indifference to the consequences. Brown v. Missouri Pacific R.R., 703 F.2d 1050, 1052 (8th Cir.1983). The punitive-damages award must be upheld if there exists "any substantial evidence to support the jury's finding." Wallace v. Dustin, 284 Ark. 318, 320, 681 S.W.2d 375, 377 (1984). Although the question is close, we hold that Mrs. Mulligan presented sufficient evidence.
The jury had before it evidence that Varidase was linked to a variety of side effects, some similar to those claimed by Mrs. Mulligan, in particular, 1955 and 1956 reports of nephritis and renal problems. Both internal office correspondence and physician complaints indicated that Lederle knew of these reactions and that it experienced troublesome difficulties with the drug's pyrogenicity and purity tests. In the 1950's, company officials acknowledged in response to complaints that studies showed a significant number of those given Varidase suffered a low-grade fever; yet the 1960 PDR said nothing whatsoever about the pyrogenic potential of Varidase.10 Plaintiff's expert witnesses described this warning as unquestionably inadequate. The medical literature introduced by Mrs. Mulligan included a 1953 account in the Journal of the American Medical Association of possible anaphylactoid reactions to Varidase; a 1952 case of circulatory collapse, reported in the same publication, and referring to a 50 per cent. rate of febrile reactions among those given the drug; a 1955 article from the Journal of Clinical Investigation dealing with the use of streptokinase, which dealt largely with pyrogenic reactions and depressions in blood pressure, but also mentioned temporary abnormalities noted in urinalysis of those given Varidase. Given the possible problems with the drug, Dr. Steele went so far as to testify that "Varidase should never have been on the market because it was of no efficacy for anything.... I think that the risk/benefit ratio was such that the drug should never have been marketed." Tr. at 468. Also introduced was the 1976 letter sent by Mrs. Mulligan's physician to Lederle, fully detailing her symptoms and asking for "any information you could give me concerning adverse reactions to this product [and] ... any help your immunology department could offer as to the future management of this patient." Plaintiff's Exhibit No. 10. The defendant's response, said the physician, was a phone call from a Lederle doctor who said "he did not believe that Varidase could cause the symptoms that had been listed in the letter." Tr. at 360.
The defendant argues that, at best, this evidence shows it inadequately warned of the potential side effects from Varidase and does not represent the kind of reckless disregard required to support punitive damages, comparing its situation to that in Forrest City Machine Works v. Aderhold, 273 Ark. 33, 616 S.W.2d 720 (1981). In that case, involving a child who was injured climbing off a tractor, the Arkansas Supreme Court could not find sufficient evidence to sustain a punitive-damage award, even though there had been at least two lawsuits filed in connection with the same tractor model. Plaintiffs in Aderhold, however, apparently did not offer the type or quantity of evidence suggesting defendant's knowledge presented here. More important, the Supreme Court has since elaborated on the level of conduct necessary for punitive damages. In Airco, Inc. v. Simmons Nat'l Bank, 276 Ark. 486, 638 S.W.2d 660 (1982), a surgical patient suffered irreversible brain damage when, because of human error, the hoses on an artificial breathing machine were improperly connected. Evidence showed that the manufacturer was aware through complaints and technical articles of the possibility of such a misfunction, yet failed to change the design of the apparatus. The court rejected defendant's argument that its conduct fell in the same category as the tractor distributor's in Aderhold. Two factors were used to distinguish Airco from Aderhold. First, the court said, "the patient's very life always depended upon the artificial breathing ... [while] the sale of farm machinery ... did not necessarily involve a similar continuous possibility of death or serious injury." 276 Ark. at 492, 638 S.W.2d at 663. The second differentiating factor was the nature of the proof against the defendant in Airco, "that the manufacturer knew from the outset, by its own testing, that an unnecessary component of the product was so deadly" that it should not have been marketed. Id.
We believe that the proof submitted by Mrs. Mulligan is closer to that involved in Airco than that in Aderhold. It is true that the side effects exhibited in Varidase users were not necessarily of the same immediate life-threatening potential as that in Airco.11 Nevertheless, we consider prescription drugs to be more akin to an artificial breathing apparatus than to a tractor. Adverse reactions to prescription drugs are more likely to be insidious; moreover, the drug consumer uses a product chosen by a third party--the physician--and is particularly dependent on the quality of a manufacturer's warning. Here, Lederle was aware that Varidase, which was described by plaintiff's expert as "worthless," Tr. at 402, had a propensity to cause fevers and other reactions, and that there was a likelihood it would be given to patients, such as Mrs. Mulligan, with minor clotting problems, whose physicians were unaware of the possible risks. The defendant emphasizes the fact that Mrs. Mulligan's condition is not representative of the side effects reported to Lederle in the 1950s and was unique both in the time it took to manifest itself and in its duration. Lederle argues it should not be made to pay punitive damages for a condition it could not foresee. But the fact that Mrs. Mulligan did not suffer the exact adverse effects of which Lederle was aware does not shield the defendant from the threat of punitive damages. The deterrence goal of such damages, Aderhold, 273 Ark. at 44, 616 S.W.2d at 726, would be subverted if they were awarded only when a defendant was aware that the precise injury suffered by a plaintiff was possible.
Lederle also argues that Mrs. Mulligan's punitive-damages proof was largely based on post-1960 evidence and therefore immaterial to the question of defendant's alleged reckless disregard of what would happen to the plaintiff when she received Varidase a quarter century ago. As support, defendant cites Freeman v. Anderson, 279 Ark. 282, 651 S.W.2d 450 (1983), which upheld a refusal to admit as evidence of punitive damages in an auto-accident case the fact that the defendant-driver failed to stop at the scene. Only its knowledge and attitude when the injury was caused is therefore at issue here, Lederle says. We agree that had plaintiff's pre-1960 evidence been insufficient, later actions of defendant would have been irrelevant. Nevertheless, we conclude that Lederle's reading of Freeman is too cramped. There, the automobiles apparently suffered only minor damage, and plaintiff did not seek medical attention until eight months after the accident. This is not a situation where a defendant, after causing an accident, fled the scene where bodies were thrown from burning automobiles. Given the minor character of the accident in Freeman, the Arkansas Supreme Court could fairly say that defendant's conduct (however unsavory) did not exhibit "such a conscious indifference to the consequences that malice might be inferred." 279 Ark. at 286, 651 S.W.2d at 452. When considering punitive damages in the present case, the jury might properly have considered the company's response to the physician's 1976 inquiry about any connection between Mrs. Mulligan's symptoms and Varidase, and his request for advice in managing her condition. The jury could have concluded that Lederle's denial of such a link represented "conscious indifference to the consequences" of its previous actions, particularly in light of the 1971 book, Hazards of Medication, written by the company's director of medical communications, mentioning a link between abnormal renal function and streptokinase-streptodornase.
Two decisions of this Court are cited by defendant to support its contention that the punitive-damage evidence was insufficient. We consider neither to be controlling. In DeLuryea v. Winthrop Laboratories, 697 F.2d 222 (8th Cir.1983), we sustained the District Court's decision not to submit the issue of punitive damages in a product-liability case involving a prescription drug. While the defendant's warnings in DeLuryea were inadequate, they did refer to the reactions suffered by the plaintiff. The Court also distinguished the DeLuryea punitive-damages proof from that in a similar case, Hoffman v. Sterling Drugs, Inc., 485 F.2d 132 (3d Cir.1973), on the ground that DeLuryea "relied almost exclusively on expert evidence and did not present the devastating documentary evidence presented in Hoffman." 697 F.2d at 231. We conclude that Mrs. Mulligan's documentary proof, if not devastating, was sufficiently damning to support a punitive-damage award. Lederle also cites Ferren v. Richards Mfg. Co., 733 F.2d 526 (8th Cir.1984), in which we reversed a punitive-damage award in favor of a plaintiff whose artificial hip prosthesis cracked, necessitating replacement. The heart of plaintiff's proof was the fact that defendant had discovered a surface crack in one piece of an eight- to ten-piece sample of the material used to make prostheses and two articles published after plaintiff's prosthesis was sold dealing with potential problems. There is a difference between the knowledge imputed to defendant in this case and that in Ferren, where "the one small crack that was discovered during quality control testing was ... superficial and easily eliminated." Id. at 530. The jury here had before it evidence of persistent and troublesome difficulties with the product, from the company's testing of Varidase, physician complaints, and scientific journals.
Accordingly, the judgment of the District Court is
Affirmed.
The Hon. George Howard, Jr., United States District Judge for the Eastern and Western Districts of Arkansas
In the 1960's, an oral-tablet version of the drug was also developed
Mrs. Mulligan initially alleged she also developed "floaters" in her eyes as a result of receiving Varidase but abandoned the claim and presented no evidence on this question at trial
Plaintiff initially filed this lawsuit in June 1979. The action was non-suited in November 1982, re-filed a year later, and tried to a jury in October 1984. There is no question that the re-filing was timely. Originally, Mrs. Mulligan was co-plaintiff with her daughter, Mary, who claims also to show microscopic hematuria in her urine, allegedly from her exposure in utero to her mother's condition. With the approval of plaintiff, the District Court dismissed Mary Mulligan's claim with prejudice after opening statements
Defendant argued that complaints of mouth sores developing after Varidase use dealt with the buccal, and not the intramuscular product
Defendant cites several cases based on Ark.Stat. Sec. 37-205 for the proposition that Arkansas law rejects the notion that a plaintiff must be aware of her injury and its cause before the limitations period begins. See, e.g., Steele v. Gann, 197 Ark. 480, 123 S.W.2d 520 (1939); Williams v. Edmondson, 257 Ark. 837, 520 S.W.2d 260 (1975). Because of the contrast between Sec. 34-2803 and the much more rigid provisions of Sec. 37-205, we do not consider those cases to be appropriate support here
This argument appears to have been a last-gasp effort for the plaintiff, whose lawyer thought the limitations period for the 1977 pacemaker failure was five years. Spickes v. Medtronic, Inc., 275 Ark. 421, 422, 631 S.W.2d 5, 6 (1982)
Field involved a newspaper pressman who developed lead poisoning; the disease first appeared as a lesion on plaintiff's toe, but his lawsuit was not filed until more than three years later, after the disease had spread to plaintiff's legs, requiring amputation. The action was dismissed
Lederle argues that the value of Schenebeck has been diminished by the passage of the Arkansas Product Liability Act of 1979, Ark.Stat. Sec. 34-2801, et seq. However, as we have noted, ante at p. 863, the Act apparently was designed to include, and not wipe away, prior precedent. Moreover, we take into consideration the fact that the Arkansas Legislature and Supreme Court have never disavowed Schenebeck. It is also of interest that Judge Howard served on both the Arkansas Supreme Court and the state Court of Appeals
The 1960 Physician's Desk Reference dealing with intramuscular and topical Varidase stated:
PRECAUTIONS: Do NOT administer intravenously. Although true allergic reactions are rare, these enzymes are antigenic. In administering them, take every possible precaution to prevent or arrest allergic reaction. Do not employ VARIDASE Jelly in the presence of active hemorrhage.
Plaintiff's Exhibit No. 2.
There was a 1959 report to the drug company that a five-year-old child with a broken arm was given Varidase to break up a large hematoma and died three days later. However, as Dr. Steele acknowledged, Tr. at 426, and defendant stresses, "the autopsy showed no cause of death."