Wang v. Metropolitan Life Insurance Co. [Indexed as: Wang v. Metropolitan Life Insurance Co.]

72 O.R. (3d) 161

[2004] O.J. No. 3525

Docket No. C40967

Court of Appeal for Ontario,

Laskin, Charron and Armstrong JJ.A.

August 31, 2004

*Application for leave to appeal to the Supreme Court of Canada was filed December 30, 2004, and submitted to court March 31, 2005.

Insurance -- Life insurance -- Interpretation and construction -- "Accident" -- Death from amniotic fluid embolism during childbirth constituting "illness" and not "accident" for purposes of accidental death benefit rider in life insurance policy.

The insured died as a result of an amniotic fluid embolism that occurred during childbirth. An amniotic fluid embolism is the obstruction and constriction of pulmonary blood vessels by amniotic fluid entering the maternal circulation, causing obstetric shock. The insurer took the position that an accidental death benefit was not payable under the terms of the accidental death benefit rider in the insured's life insurance policy because the insured had died, not from an accident, but from a natural cause, and because the rider specifically excluded death caused or contributed to by an illness. The insured's husband commenced an action. On a motion for a determination of a question of law before trial, the motions judge held that the insured's death had resulted from an "accident" and not from an "illness". The motions judge relied on expert evidence that amniotic fluid embolism is a rare and devastating complication of pregnancy, which occurs at anywhere from 1/14,000 to 1/60,000 live births, and concluded that "the insured would not have expected to die. Nor would any reasonable person in her position. Her death was, therefore, accidental." In determining that the insured's death did not result from an illness, the motions judge noted that the death was the unexpected result of a normal life event, pregnancy, and that pregnancy is not an illness. The insurer appealed.

Held, the appeal should be allowed.

Per Charron J.A. (Armstrong J.A. concurring): The motions judge erred in her interpretation of what constitutes an "accident". The word "accident" is to be given its ordinary and popular meaning, since there is no definition of "accident" in the policy. Relying on the decision of the Supreme Court of Canada in Martin v. American International Assurance Life Co., she stated that the pivotal question was whether the insured expected to die. However, the expectation of death test is not appropriate in a case where death results solely from a natural cause. A person's unexpected death during his sleep, caused by aneurysm or other condition with fatal consequences, would not be described as an accident in ordinary and popular language. Nor would such an interpretation accord with the reasonable expectations of the parties. The insured's death was not accidental.

The motions judge erred in her determination of whether the death resulted from a physical illness by failing to consider that amniotic fluid embolism, not pregnancy, was the cause of death. The question was not whether pregnancy is an illness; it was whether amniotic fluid embolism is an illness. There was uncontradicted medical evidence before the motions judge that amniotic fluid embolism is a physical illness. The rarity of its occurrence during the course of pregnancy and childbirth does not mean that it is not a physical illness. Since amniotic fluid embolism is a physical illness and undoubtedly "caused or contributed to" the insured's death, the occurrence fell within the scope of the exclusion.

Per Laskin J.A. (dissenting): The motions judge correctly decided that the insured died as the result of an accident and that no physical illness directly or indirectly caused her death. The decision of the Supreme Court of Canada in Martin v. American International Assurance Life Co. established a general and [page162] simple principle for determining when a death is "accidental", which turns on the question whether the insured expected to die. If the answer is unclear, then the court may consider whether a reasonable person in the position of the insured expected to die. The motions judge asked that question, and determined that the insured did not expect to die, so that her death was accidental. That was a finding of mixed fact and law, and as the motions judge asked herself the right question, her finding was entitled to deference on appeal. Even if the finding was not entitled to any deference, the motions judge was correct. There is no suggestion in Martin that the expectation test is inapplicable where an insured dies from natural causes. On the contrary, the court stressed that the expectation test was a test that could be generally applied to apparent accidental deaths. The expectation test in Martin concerns the consequences flowing from the action or conduct of an insured. When that statement was applied to this case, the insured's death could only be characterized as accidental. One consequence of her action in becoming pregnant and giving birth was her death -- a consequence so unusual and so unexpected that it had to be accidental. The distinction between accidental death and death from natural causes is a false dichotomy. In all cases, the only question to be asked is whether the death was unexpected or whether the insured expected to die.

Charron J.A. adopted an overly narrow interpretation of "accident" in circumstances where a generous interpretation was called for. A generous interpretation was warranted because the insurer would always have limited the breadth of "accidental death" by expanding the scope of the exclusions. It did not do so. Moreover, the insurer's best position was that the meaning of "accident" -- and whether its meaning extended to the insured's death -- is ambiguous. The law unequivocally states that ambiguities in an insurance policy must be resolved against the insurer and in favour of the insured.

Amniotic fluid embolism is not a physical illness. Every pregnant woman has amniotic fluid emboli, which move into her uterine veins and travel to her heart and lungs, yet only a very few will suffer an embolism. To suggest that the ordinary person would think that any time a woman becomes pregnant she has a physical illness because she has a cluster of emboli and open uterine veins, which in .00007 per cent or fewer live births combine to produce a fatal fluid embolism, is nonsense. The existence of amniotic fluid emboli in pregnant women is a condition so remote in its potential mischief that common speech would call it not a physical illness, but at most a predisposing tendency. An alternative way of thinking of the insured's condition was to consider it a "temporary disorder" uniquely associated with pregnancy and childbirth. Many cases have held that "temporary disorders" do not preclude recovery under an accidental benefit insurance policy even where the policy denies benefits for loss caused or contributed to by disease or physical illness.

APPEAL from an order of Sachs J. of the Superior Court of Justice, dated October 16, 2003, determining a question of law.

Martin v. American International Assurance Life Co., [2003] 1 S.C.R. 158, 223 D.L.R. (4th) 1, 301 N.R. 127, [2003] 6 W.W.R. 1, [2003] I.L.R. ÂI-4171, 2003 SCC 16, 12 B.C.L.R. (4th) 201, [2003] S.C.J. No. 14; Smith v. British Pacific Life Insurance Co., 1965 91 (SCC), [1965] S.C.R. 434, 51 D.L.R. (2d) 1, 51 W.W.R. 417; Stats v. Mutual of Omaha Insurance Co., 1978 38 (SCC), [1978] 2 S.C.R. 1153, 87 D.L.R. (3d) 169, 22 N.R. 91, consd Other cases referred to Amend v. Equitable Life Assurance Co. of the United States, 342 N.Y.S.2d 284 (N.Y. Co.Ct. 1972); Berkowitz v. New York Life Insurance Co., 256 A.D. 324 (N.Y. 1939); Bobier v. Beneficial Standard Life Insurance Co., 570 P.2d 1094 (Colo. App. 1977); [page163] Bornstein v. J.C. Penney Life Insurance Co., 946 F.Supp. 814 (Colo. Dist. Ct. 1996); Carroll v. CUNA Mutual Insurance Society, 894 P.2d 746 (Colo. 1995); Chale v. AllState Life Insurance Co., 353 F.3d 742 (9th Cir. 2003); Claxton v. Travellers Insurance Co. of Hartford (1917), 1917 605 (QC CA), 36 D.L.R. 481 (Que. C.A.); Denton v. Travelers Insurance Co., 25 F.Supp. 556 (Md. Dist. Ct. 1938); Fiffick v. Econ-o-Check Corp., 85 Fed. Appx. 16 (5th Cir. 2004); Gaskins v. New York Life Insurance Co., 104 So.2d 171 (La. 1958); Gottfried v. Prudential Insurance Co. of America, 414 A.2d 544 (N.J.S. Ct. 1980); Hamlyn v. Crown Accidental Insurance Co., [1893] 1 Q @@.B. 750 (C.A.); Housen v. Nikolaisen, [2002] 2 S.C.R. 235, 219 Sask. R. 1, 211 D.L.R. (4th) 577, 286 N.R. 1, 272 W.A.C. 1, [2002] 7 W.W.R. 1, 30 M.P.L.R. (3d) 1, 2002 SCC 33, 10 C.C.L.T. (3d) 157; Howard v. National Education Assn. of New York, 984 F.Supp. 103 (N.Y. Dist. Ct. 1997); In a Life Insurance Co. v. Brundin, 533 P.2d 236 (Alaska 1975); J.C. Murphy v. Continental Casualty Co., 296 So.2d 507 (La. 1972); Jones v. Aetna Life Insurance Co., 439 S.W.2d 721 (Colo. App. 1969); Katskee v. Blue Cross/Blue Shield of Nebraska, 515 N.W.2d 645 (Neb. 1994); Landress v. Phoenix Mutual Life Insurance Co., 291 U.S. 491 (1934); Paulissen v. United States Life Insurance Co., 205 F.Supp.2d 1120 (Colo. Dist. Ct. 2000); Reed v. United States Fidelity & Guaranty Co., 491 P.2d 1377 (Colo. 1971); Silverstein v. Metropolitan Life Co., 254 N.Y. 81 (N.Y. 1930); Somersall v. Friedman, 2002 SCC 59, [2002] 3 S.C.R. 109, 215 D.L.R. (4th @@.) 577, 292 N.R. 1, [2002] I.L.R. Â1-4114, 2002 SCC 59, 25 M.V.R. (4th) 1, [2002] S.J. No. 60; Voison v. Royal Insurance Co. of Canada (1988), 1988 4736 (ON CA), 66 O.R. (2d) 45, 29 O.A.C. 227, 53 D.L.R. (4th) 299, [1988] I.L.R. Â1-2358 (C.A.); Waxman v. Waxman (2004), 2004 39040 (ON CA), 186 O.A.C. 201, 44 B.L.R. (3d) 165, [2004] O.J. No. 1765 (C.A.); Winchester v. Prudential Life Insurance Co., 975 F.2d 1479 (10th Cir. 1992) Rules and regulations referred to Rules of Civil Procedure, R.R.O. 1990, Reg. 194, Rule 21 Authorities referred to Bryant Jr., J.A., "What Conditions Constitute Disease Within Terms of Life, Accident, Disability or Hospitalization Insurance Policy", 461 A.L.R. 3d 822

Sergio Grillone, for respondent. Elisa A. Scali, for appellant.

CHARRON J.A. (ARMSTRONG J.A. concurring): --

Nature of the Appeal

[1] The question on this appeal is whether Stacey Chiu Lin's death from an amniotic fluid embolism that occurred during childbirth resulted from an "accident". Mrs. Lin's husband, the respondent Kristoffer Wang, and her daughter received the $480,000 Death Benefit under a policy of life insurance issued by the appellant Metropolitan Life Insurance Company ("Metlife"). However, Metlife took the position that the additional $200,000 Accidental Death Benefit was not payable under the terms of the accidental death benefit rider in the policy because Mrs. Lin had [page164] died, not from an accident, but from a natural cause, and further because the rider specifically excluded death caused or contributed to by an illness. Mr. Wang commenced this action. On a motion for the determination of a question of law before trial, the motions judge rejected Metlife's position and held that Mrs. Lin's death resulted from an "accident" and not from an "illness". Metlife appeals from this finding.

The Facts

[2] The facts are not in dispute and are contained in an Agreed Statement of Facts. Stacey Chiu Lin died on September 2, 1994 shortly after undergoing an elective caesarean section. The caesarean was performed in the standard fashion and was uneventful. At the time of delivery, partial placenta accreta was encountered requiring manual removal of the placenta. Subsequent to the closure of Mrs. Lin's abdomen, she encountered difficulty with her breathing, became cyanosed (blue in colour), and subsequently suffered a cardiac arrest. Despite efforts to resuscitate Mrs. Lin, she died a short while after. A post-mortem examination established the cause of death as "amniotic fluid embolism". The coroner's investigation summary stated the same cause of death and characterized the means as "natural". Each party filed an expert medical report describing the nature of an amniotic fluid embolism. I will refer to these reports later in these reasons.

[3] The life insurance policy in question was issued on April 26, 1994. It provided for base coverage in the amount of $480,000 for a yearly premium of $863.84 and an additional death benefit in the amount of $200,000 for a premium of $13.68. As noted earlier, there is no dispute over payment of the basic death benefit. The accidental death benefit rider, among other things, provides the following:

This rider provides additional insurance if the insured dies from an accident.

If we receive proof that the insured died, directly and independently of all other causes, as a result of an accident, we will pay under this rider:

Risks not Covered -- No payment will be made if the death:

2. Is caused or contributed to, directly or indirectly, by physical or mental illness or treatment for the illness.

[4] "Accident" is not otherwise defined in the policy. [page165]

The Motions Judge's Findings

[5] The question put to the motions judge was the following:

Whether the death of the Plaintiff's wife, Stacey Chiu Lin, was an "accident" pursuant to the accidental death benefit rider in the policy of insurance issued by the defendant.

[6] In light of the wording of the policy, the motions judge considered the following two questions: 1. Was the insured death's caused directly and independently of all other causes as a result of an accident? 2. Was the insured's death caused or contributed directly or indirectly by physical illness or treatment for the illness?

[7] On the first question, the parties agreed that the word "accident" or "accidental" "is an ordinary word to be interpreted in the ordinary language of the people": Stats v. Mutual of Omaha Insurance Co., 1978 38 (SCC), [1978] 2 S.C.R. 1153, 87 D.L.R. (3d) 169, at p. 1164 S.C.R. Mr. Wang took the position that, viewed from this perspective, the central question was whether a reasonable person in the position of Mrs. Lin would have expected to die. Metlife argued that a death from "natural" causes cannot be "accidental". In support of its position, Metlife relied on the decision in Smith v. British Pacific Life Insurance Co., 1965 91 (SCC), [1965] S.C.R. 434, 51 D.L.R. (2d) 1, at p. 438 S.C.R., where the court stated:

An "accident" is defined in Welford on Accident Insurance, 2nd ed., p. 268, as:

The word "accident" involves the idea of something fortuitous and unexpected, as opposed to something proceeding from natural causes; an injury caused by accident is to be regarded as the antithesis to bodily infirmity caused by disease in the ordinary course of events.

[8] The motions judge held that this "ratio" in Smith was later rejected by the Supreme Court of Canada in Martin v. American International Assurance Life Co., 2003 SCC 16, [2003] 1 S.C.R. 158, 223 D.L.R. (4th) 1 and that the pivotal question was, as submitted by Mr. Wang, whether the insured expected to die. The motions judge then relied on Mr. Wang's expert, who, in his report, stated the following about amniotic fluid embolism:

Amniotic fluid embolism is a rare and devastating complication of pregnancy. It occurs at anywhere from 1/ 14,000 to 1/60,000 live births . . .

The salient factors with respect to Amniotic Fluid Embolism are that it is: -- Rare -- Unpredictable -- Non-preventable . . . [page166]

[9] After noting that Metlife's expert did not disagree with Dr. Steinberg's description of amniotic fluid embolism, the motions judge concluded:

Given this, it is my view that the insured, Mrs. Lin, would not have expected to die. Nor would any reasonable person in her position. Her death was, therefore, accidental.

[10] The motions judge then determined the second question as follows:

The insured's death was the unexpected result of a normal life event -- pregnancy. Pregnancy is not an illness. Further, there is no evidence that the insured had any illness that caused or contributed to her death. Therefore, the insured's death was not caused or contributed to, directly or indirectly, by physical illness or treatment for the illness.

[11] In determining the second question, the motions judge made no reference to the opinion letter filed by Metlife's expert, Dr. William H. Noble. I will set out the contents of that report later in these reasons.

Issues

[12] Each party has framed the issues on appeal somewhat differently. In light of the argument before this court, I would state the issues as follows: 1. Did the motions judge err in her interpretation of the case law on what constitutes "an accident"? 2. Was Mrs. Lin's death as a result of amniotic fluid embolism an "accident" according to the terms of the policy? 3. Did the motions judge err in restricting her analysis on whether the death resulted from a physical illness to the question whether pregnancy was an illness? 4. Based on the totality of the evidence, was the death caused or contributed to by a physical illness within the exclusion contained in the policy?

[13] For reasons I will set out, it is my view that the motions judge erred in her interpretation of what constitutes an accident. It is therefore incumbent upon this court to determine whether the death was accidental based on the applicable principles of law. In my view, it was not. I am also of the view that the motions judge erred in her determination of whether the death resulted from a physical illness by failing to consider that amniotic fluid embolism, not pregnancy, was the cause of death. Based on this [page167] record, I conclude that the death was caused or contributed to by a physical illness within the meaning of the exclusion.

Analysis

1. Did the motions judge err in her interpretation of the case law on what constitutes "an accident"?

[14] There is no dispute that the word "accident" is to be given its ordinary and popular meaning since there is no definition of "accident" found in the policy. I agree with the motions judge that the decision in Stats provides a useful point of reference in the application of this general test.

[15] In Stats, the payment of a death benefit turned on whether the injuries were "accidental". The insured died as a result of injuries suffered when she drove her car, erratically and at increasing speed, past a stop sign, across a street, over a sidewalk and into a brick building, killing both occupants of the car. An autopsy revealed that she had very high blood and urine alcohol levels at the time of the accident. Upon this evidence, the insurance company refused to pay the claim of the named beneficiary under the life insurance policy.

[16] The Supreme Court of Canada stated the following at pp. 1162-64 S.C.R.:

The word "accident" found in an insurance policy is to be given its ordinary and popular meaning. There is no technical definition of "accident" to be applied.

A variety of dictionary definitions have been attempted and text writers have used very astute and logical analyses of what would constitute an accident, but remembering that it is an ordinary word to be interpreted in the ordinary language of the people, I ask myself what word would any one of the witnesses of this occurrence use in describing the occurrence. Inevitably, they would have used the word "accident". I am ready to agree that one has to have a knowledge of all the circumstances before one's use of the ordinary language can have a determinative effect but even with all the knowledge of the circumstances which I have outlined in such detail, the ordinary person would still use the word "accident". Pigeon J., in Canadian Indemnity Company v. Walkem Machinery & Equipment, supra, adopted Halsbury's words, "any unlooked for mishap or occurrence", and in Fenton v. Thorley & Co. Limited [[1903] A.C. 443.], Lord Macnaghten said at p. 448:

. . . the expression "accident" is used in the popular and ordinary sense of the word as denoting an unlooked-for mishap or an untoward event which is not expected or designed.

[17] The Supreme Court of Canada therefore concluded in Stats that the injuries that resulted in death were accidental, despite the fact that they were occasioned by the negligent conduct [page168] of the insured. Although not referred to by the court, the decision in Stats was certainly a move away from the court's earlier decision in Smith. In Smith, the insured, who suffered from a heart condition, died from a coronary thrombosis and occlusion at the steering wheel of his car. At the time, he was rocking the car back and forth in an attempt to assist his friend who was pushing the vehicle out of a snow bank. The court, basing its decision on earlier precedent, concluded that the insured's death was not caused by accident because the loss resulted from the deliberate conduct of the insured.

[18] The view that the category of deaths caused by accidental means excludes accidental deaths that are the natural effects of deliberate actions was expressly put to rest in Martin. The court referred to Smith and other authorities that reached a similar conclusion and rejected that view (at para. 12):

This view seems to me, however, to be problematic. Almost all accidents have some deliberate actions among their immediate causes. To insist that these actions, too, must be accidental would result in the insured rarely, if ever, obtaining coverage. Consequently, this cannot be the meaning of the phrase "accidental means" in the policy. Insurance policies must be interpreted in a way that gives effect to the reasonable expectations of the parties: Reid Crowther & Partners Ltd. v. Simcoe & Erie General Insurance Co., 1993 150 (SCC), [1993] 1 S.C.R. 252, at p. 269. A policy that seldom applied to what reasonable people would consider an accidental death would violate this principle.

[19] Hence the motions judge was correct in her conclusion that the ratio in Smith was overturned by the Supreme Court of Canada in Martin. However, she erred in concluding that, in rejecting the principle in Smith, the court expressly rejected the distinction between an accidental death and a death from natural causes upon which Metlife relies in this case. The Supreme Court of Canada in Martin never opined on this issue. I will say more about this later.

[20] The motions judge then referred to the following passage in Martin, paras. 20-21:

It follows that death is not non-accidental merely because the insured could have prevented death by taking greater care, or that a mishap was reasonably foreseeable in the sense used in tort law. Nor does a death that is unintended become "non-accidental" merely because that person was engaged in a dangerous or risky activity. As this Court emphasized in Canadian Indemnity, supra, at p. 316, the jurisprudence assigns a generous meaning to "accidental", in the absence of language to the contrary in the policy.

The pivotal question is whether the insured expected to die. The circumstances of the death -- what the insured said, or did, or did not do -- may point to the answer. However, to the extent that the answer is unclear when [page169] the matter is viewed solely from the perspective of the insured, the court may consider whether a reasonable person in the position of the insured would have expected to die . . . .

[21] The motions judge concluded that the determining question was whether Mrs. Lin or any reasonable person in her position would have expected to die. Given the rarity of occurrence of amniotic fluid embolism, she concluded that there would be no expectation of death, and that consequently, the death was accidental.

[22] Mr. Wang supports the motions judge's analysis and conclusion. He submits that it does not matter whether the death is precipitated by a natural cause or a deliberate action. So long as it is unintended and unexpected, it is accidental. Counsel for Mr. Wang fairly conceded in argument that this test would conceivably be wide enough to include, in the category of accidental deaths, an unexpected and sudden death during sleep from any number of natural causes such as heart failure, aneurysm, thrombosis or embolism.

[23] In my view, the motions judge's interpretation of the test adopted in Martin ignores both the factual context of that decision and the ordinary person test upon which it is premised. In Martin, the insured was a physician who had developed an addiction to opiate medications for which he was being treated. He had been placed on a program of gradual withdrawal from morphine and Demerol. He was found dead in his office, death having been caused by an intravenous injection of Demerol. The level of Demerol was found to be at the low end of the range for lethal doses. The court found, in looking at the full chain of events, that the insured neither intended nor expected to die and that death was caused by an accidental overdose.

[24] In discussing what constitutes death by "accidental means", the court reiterated the principle in Stats that the word "accident" is "an ordinary word to be interpreted in the ordinary language of the people". The court continued as follows [at paras. 19-20]:

Hence, . . . we must focus on the ordinary person's understanding of the phrase, and on "whether in ordinary and popular language the event as it happened would be described as an accident" (para. 26). Only in this way can the reasonable expectations of both the insured and insurer be protected. We must therefore inquire how the phrase "death by accidental means" is used in ordinary language.

As a starting point, we note that the accidental nature of a particular means of death depends, in ordinary parlance, on the consequences that the insured had or did not have in mind. When we speak of an "accidental" means of death, we normally have in mind a situation in which someone's action has had results that this person did not intend or expect. Unintentional or [page170] unexpected death is seen as accidental; intentional or expected death as non-accidental. [Emphasis in original omitted]

[25] The court referred to other similar authorities and to a number of cases where the general rule that unexpected death is accidental was applied. The court concluded at para. 23 that "[t]he expectation test can be applied generally to all cases in which death appears to be accidental".

[26] In my view, it is noteworthy that the entire analysis of the court relates to the consequences flowing from the actions or conduct of an insured. Indeed, a good part of the analysis concerns cases where persons engage in risky activities. No reference is made to death resulting from natural causes. The focus of the analysis, as is usually the case, is obviously driven by the factual context of the case, and in that context, it makes eminent sense that "the pivotal question is whether the insured expected to die".

[27] In my view, the expectation test is not appropriate in a case where death results solely from a natural cause. Such a death is not one "in which death appears to be accidental". If we come back to the examples referred to earlier, a person's unexpected death during his or [her] sleep, caused by aneurysm or other condition with fatal consequences, would not be described as an accident in ordinary and popular language. Nor would such an interpretation accord with the reasonable expectations of the parties.

[28] I therefore conclude that the motions judge erred in her legal analysis on this question.

2. Was Mrs. Lin's death as a result of amniotic fluid embolism an "accident" according to the terms of the policy?

[29] There is no dispute that Mrs. Lin died as a result of an amniotic fluid embolism. There is also no dispute that this condition, although rare, happens naturally during childbirth. Would an ordinary person describe what happened as an accident? Perhaps only in the figurative sense as a tragic "accident of nature", but in my view, not in the ordinary sense of the word. In respect of this kind of occurrence, I find that the Welford definition of "accident" approved by the Supreme Court of Canada in Smith [at p. 438 S.C.R.] is still informative. For convenience, I reproduce it again:

The word "accident" involves the idea of something fortuitous and unexpected, as opposed to something proceeding from natural causes; an injury caused by accident is to be regarded as the antithesis to bodily infirmity caused by disease in the ordinary course of events." [page171]

[30] I therefore conclude that Mrs. Lin's death was not the result of an accident.

[31] Even if I am wrong on this point and the death can properly be characterized as accidental, it is still necessary to consider whether the occurrence is excluded from coverage as a death resulting from a physical illness.

3. Did the motions judge err in restricting her analysis on whether the death resulted from a physical illness to the question whether pregnancy was an illness?

[32] As noted earlier, the accidental death benefit is not payable under the terms of the policy if it was "caused or contributed to, directly or indirectly, by physical or mental illness or treatment for the illness". The motions judge held that pregnancy is not an illness and that, consequently, this occurrence did not fall within the exclusion provision of the policy.

[33] With respect, I think that the motions judge's analysis on this question is also flawed. Although pregnancy was part of the chain of events that resulted in the ultimate death, pregnancy was not a cause of death. The cause of death was indisputably amniotic fluid embolism. The question therefore is not whether pregnancy is an illness; obviously, it is not. Rather, the question is whether amniotic fluid embolism is an illness. It does not appear that the motions judge considered this question other than perhaps where she added: "Further, there is no evidence that the insured had any illness that caused or contributed to her death." This leads us to the final issue.

4. Based on the totality of the evidence was the death caused or contributed to by a physical illness within the exclusion contained in the policy?

[34] Mr. Wang filed a medical opinion report from Dr. Steinberg, an expert in obstetrics and gynecology, who described amniotic fluid embolism as "a rare and devastating complication of pregnancy" which occurs "at anywhere from 1/ 14,000 to 1/60,000 live births". Dr. Steinberg attached to his report some of the relevant literature on amniotic fluid embolism. Each of these articles points out that, although rare, this condition is one of the most, if not the most, common causes of maternal death in western industrialized societies. Dr. Steinberg also included and commented upon recent studies by another expert, Dr. Clarke, who believes that amniotic fluid embolism is a misnomer and suggests that it should be replaced by "anaphylactoid syndrome of pregnancy". Dr. Steinberg explains that: [page172]

The term anaphylactoid in medical practice denotes a sudden change in hemodynamics commonly found in severe allergic reactions which leads to morbidity and/or death. This new concept in the understanding of amniotic fluid embolism arises out of studies by Clarke and others which indicate that there is a sudden maternal cardiovascular collapse related to immunological changes which are unpredictable.

[35] After summing up Dr. Clarke's findings, Dr. Steinberg concludes his report by identifying the salient factors with respect to amniotic fluid embolism:

The salient factors with respect to amniotic fluid embolism are that it is: -- rare; -- unpredictable; -- non-preventable; -- associated with an anaphylactic-like response; -- associated with a high neurological impairment in survivors; -- likely associated with changes in the immunological status; -- resistant to specific modalities of treatment; -- associated with a high mortality rate.

[36] He then concludes that:

It would be reasonable to conclude that such a rare and catastrophic complication of an elective cesarean section could be classified as: -- unforeseen; -- unexpected; -- unusual.

[37] Metlife filed a medical opinion report from Dr. Noble who has practised anaesthesiology and intensive care medicine since 1970. As noted by the motions judge, Dr. Noble did not disagree with anything that Dr. Steinberg stated in his report. In fact, he expressed his agreement as follows:

I had already reviewed the world literature on Amniotic Fluid Embolism and can tell you the articles Dr. Steinberg selected are current and appropriate and represent the medical thinking on Amniotic Fluid Embolism. As a result of reading these and other materials I will respond to the questions you have posed.

Dr. Steinberg's report quotes Dr. Clark and correctly states his position. Dr. Clark believes there is an allergic component to the amniotic fluid embolus syndrome but also states there is an obstructive component. I agree entirely with Dr. Steinberg's opinion which succinctly describes amniotic fluid embolism that rapidly causes the death of rare pregnant women. [page173]

[38] Dr. Noble went on to respond to questions that were specifically put to him. In answer to the question "What is an amniotic fluid embolus?" he opined:

It is a disease that affects pregnant women resulting from the passage of amniotic fluid from within her uterus to the mother's venous circulation. This material moves to the mother's heart and lungs causing death in rare cases (1/ 40,000 cases).

[39] In answer to the question "Is an amniotic fluid embolus a physical illness?", he answered:

Yes, Amniotic Fluid Embolus is a physical illness. It results when the separation of amniotic fluid from the mother's circulation is crossed as uterine veins are opened by the birthing process (regardless of method). The Amniotic Fluid Emboli move into the mother's veins and go to her heart and lungs. This may begin the classic disease process described in Stacey Chui Lin's case. Since every mother, while giving birth by any method, may suffer an amniotic fluid embolus, this is a physical illness. A comparison may be made between Amniotic Fluid Embolism and a stroke or a heart attack. All North Americans over 55 years have atherosclerosis (hardening of the arteries) but only a few will suffer the stroke or heart attack that may result depending on severity. It cannot be considered an accident -- it is a rare event created by a process within our body. The same applies to Amniotic Fluid Embolism. All of the conditions required for a death from amniotic fluid embolism (pregnancy and open maternal uterine veins) exist in all pregnant women but the load of emboli is small or the reaction is not major and so most women do survive pregnancy. Rarely, this physical illness claims a life as it unfortunately did in Stacey Chiu Lin's case.

[40] Mr. Wang argues before this court that Dr. Noble's opinion that amniotic fluid embolism is an illness is not supported in the literature. It is not apparent that this argument was ever made to the motions judge. She certainly did not allude to it in her reasons and, further, the Rule 21 procedure [Rules of Civil Procedure, R.R.O. 1990, Reg. 194] agreed upon by the parties belies any suggestion that there is a material factual dispute. In any event, I do not accept Mr. Wang's argument on this point. There is no conflict between the two medical opinions. Dr. Steinberg described amniotic fluid embolism as "a rare and devastating complication of pregnancy" and "an often devastating obstetrical syndrome" but did not provide an opinion on whether it was a physical illness. His opinion appeared to be confined to the question whether the catastrophic complication could reasonably be foreseen. Further, I see nothing in the literature that detracts from Dr. Noble's opinion. Hence, his medical opinion that amniotic fluid embolism is a physical illness stands uncontradicted.

[41] Subsequent to drafting these reasons, I have read the draft reasons of my colleague, Laskin J.A. I wish to briefly [page174] address his point that the medical evidence is irrelevant in determining whether amniotic fluid embolism is a physical illness. Neither party took this position; however, I will accept his proposition that what counts is the ordinary person's understanding of the term.

[42] First, my colleague expresses the view that no ordinary person would think that any time a woman becomes pregnant she has a physical illness because of the remote risk of amniotic fluid embolism. He states that "none of the usual indicia that ordinary people think of when they think of a physical illness are present -- no symptoms, no outward manifestations of illness, no progressively worsening condition".

[43] With respect, I believe my colleague falls into the same error as the motions judge when he focuses on "pregnancy" as the illness as opposed to amniotic fluid embolism. Amniotic fluid embolism is consistently described in the literature appended to Dr. Steinberg's opinion letter as a "complex condition", a "syndrome" which is "classically characterized by the abrupt onset of hypotension, hypoxia, and consumptive coagulopathya". Although rare, it is usually devastating, "with mortality as high as 80 percent". I think an ordinary person would characterize this condition as a physical illness.

[44] Secondly, my colleague relies on Cardozo J.'s common sense distinction between a disease or infirmity, and a mere predisposing tendency in support of his conclusion that amniotic fluid embolism is not a physical illness. Following Cardozo J.'s analysis, I come to the opposite conclusion.

[45] As Cardozo J. put it, "a morbid or abnormal condition of such quality or degree that in its natural and probable development it may be expected to be a source of mischief . . . may fairly be described as a disease or an infirmity". Amniotic fluid embolism, as I have just described in the previous paragraph, certainly falls in that category of cases. It is a drastic departure from an ideal or perfect norm of health and, its potential mischief, in "its natural and probable development", is usually death within a short period of time. I respectfully disagree with Laskin J.A.'s conclusion that amniotic fluid embolism is not a physical illness because of the rarity of its occurrence during the course of pregnancy and childbirth.

[46] I would therefore conclude that amniotic fluid embolism is a physical illness. Since it undoubtedly "caused or contributed to" Mrs. Lin's death, the occurrence falls within the scope of the exclusion.

[47] I would allow the appeal, set aside the motions judge's order, and substitute an order in accordance with these reasons. [page175] The appellant is entitled to its costs fixed at $7,500, inclusive of disbursements and GST.

LASKIN J.A. (dissenting): --

A. Introduction

[48] I have read the draft reasons of my colleague Charron J.A. I do not agree with either her analysis or her conclusion. In my view, the motions judge, Sachs J., correctly decided that Mrs. Lin died "as the result of an accident" and that no "physical illness" directly or indirectly caused her death.

[49] My colleague has summarized the factual background. The key facts are these. The respondent, Kristoffer Wang, is the beneficiary of his late wife's life insurance policy with the appellant, Metropolitan Life. The policy contains a $200,000 accidental death benefit rider, under which the insurer agrees to pay if the insured died "directly and independently of all other causes, as the result of an accident". The rider lists a number of exclusions of which only one is relevant: "No payment will be made if the death is caused or contributed to, directly or indirectly, by physical or mental illness or treatment for the illness". The insurer chose not to define either "accident" or "physical illness".

[50] After giving birth to their child, the respondent's wife Mrs. Lin died of an amniotic fluid embolism, a rare and devastating complication of pregnancy, occurring in 1 in 14,000 to 1 in 60,000 live births. The record does not disclose any other cause of death; and does not disclose that the insured Mrs. Lin suffered from any pre-existing illness.

[51] Although Charron J.A. lists four issues, in my view there are essentially two. 1. Did the motions judge err in deciding that Mrs. Lin died as the result of an accident? 2. Is amniotic fluid embolism (or embolus) a "physical illness"?

B. Analysis

First issue: Did Mrs. Lin die as the result of an accident?

a. The general principles

[52] The main Canadian case on the meaning of "accidental death" in an insurance policy is the Supreme Court of Canada's recent judgment in Martin v. American International Assurance Life Co., 2003 SCC 16, [2003] 1 S.C.R. 158. [page176] Martin confirmed or established three propositions relevant to this appeal: 1. Where the word or phrase "accident", "accidental", "accidental means" or "accidental death" is not defined in the policy, it should be given its ordinary and popular meaning, not the meaning that might be given by a person with expert knowledge. Considering how the word or phrase is used in ordinary language will protect the reasonable expectations of both the insurer and insured. 2. The distinction drawn by some courts -- including the Supreme Court in its earlier decision in Smith v. British Pacific Life Insurance Co., 1965 91 (SCC), [1965] S.C.R. 434, 51 D.L.R. (2d) 1 -- between "accidental means" and "accidental death" (or "accidental results") should be eliminated. Those courts that adhered to this distinction held that a beneficiary could recover only if both the insured's death and the means by which it was brought about were accidental. By contrast, in Martin, McLachlin C.J.C. held that the two phrases "accidental means" and "accidental death" have essentially the same meaning. In so holding, she relied on Cardozo J.'s dissenting judgment in Landress v. Phoenix Mutual Life Insurance Co., 291 U.S. 491 (1934) where he termed the distinction a "Serbonian Bog" at pp. 499-500:

The attempted distinction between accidental results and accidental means will plunge this branch of the law into a Serbonian Bog. "Probably it is true to say that in the strictest sense and dealing with the region of physical nature there is no such thing as an accident." . . . On the other hand, the average man is convinced that there is, and so certainly is the man who takes out a policy of accident insurance. It is his reading of the policy that is to be accepted as our guide, with the help of the established rule that ambiguities and uncertainties are to be resolved against the company . . . . The proposed distinction will not survive the application of that test.

When a man has died in such a way that his death is spoken of as an accident, he has died because of an accident, and hence by accidental means. (Citations omitted)

The Supreme Court of Canada had to address this distinction because the policy provision in question in Martin provided coverage only if the insured's death "resulted directly, and independently of all other causes, from bodily injury effected solely through external, violent and accidental means . . .". The policy provision in issue here does not use [page177] the phrase "accidental means". Nonetheless, the elimination of the distinction is important for the present appeal because it obliges the court to focus on the result, on the consequence of the actions or events producing the death, and to decide whether that result or consequence was "unexpected" and thus "accidental". In short, death is accidental if the result is accidental, whether or not due to accidental means. An insured's voluntary act, even intentional or deliberate act, may nonetheless result in an "accidental death" if the death was not expected but was instead "an unlooked for mishap". 3. Most important, Martin established a general and simple principle for determining when a death is "accidental". According to McLachlin C.J.C., "[t]he pivotal question is whether the insured expected to die" (para. 21). If the answer is unclear then the court may consider whether a reasonable person in the position of the insured expected to die. Thus, the Chief Justice distinguished between unintentional or unexpected death, which is seen as accidental, and intentional or expected death, which is seen as non-accidental. As she observed, "[t]he general rule that unexpected death is accidental has been repeatedly applied" (para. 22), and therefore "[t]he expectation test can be applied generally to all cases in which death appears to be accidental" (para. 23).

[53] The Supreme Court of Canada's decision in Martin is entirely consistent with the predominant American jurisprudence. American state courts have had a long history of interpreting accidental death provisions in insurance policies. As has our Supreme Court, most have adopted Cardozo J.'s dissent in Landress, and have eliminated the distinction between accidental means and accidental results. And most American courts have also employed the expectation test in determining whether a death is "accidental". The decision of the Supreme Court of Colorado in Carroll v. CUNA Mutual Insurance Society, 894 P.2d 746 (Colo. 1995) at p. 752 is typical:

An effect which is the natural and probable consequence of an act or course of action is not an accident, nor is it produced by accidental means. It is either the result of actual design, or it falls under the maxim that every man must be held to intend the natural and probable consequence of his deeds. On the other hand, an effect which is not the natural or probable consequence of the means which produced it, an effect which does not ordinarily follow and cannot be reasonably anticipated from the use of those means, an effect which the actor did not intend to produce and which he cannot be [page178] charged with the design of producing under the maxim to which we have adverted, is produced by accidental means . . . . Landress, 291 U.S. at 501 n. 2 (Cardozo, J., dissenting) (quoting Western Commercial Travelers' Ass'n. v. Smith, 85 F. 401, 405 (8th Cir. 1898)). Thus, under this definition, as long as the resulting injury or death is unexpected, unintended, and unforeseeable then the injury or death is accidental. (Emphasis added) See also Fiffick v. Econ-o-Check Corp., 85 Fed. Appx. 16 (5th Cir. 2004).

[54] In addition, Martin is also consistent with this court's main decision on the meaning of accidental injury or death in an insurance policy, Voison v. Royal Insurance Co. of Canada (1988), 1988 4736 (ON CA), 66 O.R. (2d) 45, 53 D.L.R. (4th) 299 (C.A.). In Voison, the insured became totally disabled with a spinal cord injury after assuming an awkward position while renovating his kitchen. Our court held that his injury was "accidental" and he was thus entitled to recover under his personal disability policy, which insured against loss resulting from "accidental bodily injuries". As McLachlin C.J.C. did in Martin, Robins J.A. stressed at p. 49 O.R. that an accidental injury (or death) is one that is "unexpected, unusual and unforeseen":

The word "accidental", like "accident", is, as the cases demonstrate, not susceptible of precise definition. In the context of an accident insurance policy which contains no express definition, it is well established that these words are to be given their ordinary, usual and popular meaning as indicating an unlooked for mishap or an untoward event which is not expected or designed; or as an event which takes place out of the usual course of events without the foresight or expectation of the person injured; or as an injury happening by chance unexpectedly, or not as expected. In determining whether a certain result is accidental, the occurrence is to be viewed from the standpoint of an ordinary reasonable person to see whether or not, from his or her standpoint, it was unexpected, unusual and unforeseen. It is irrelevant that a person with expert knowledge would have expected the occurrence or regarded it as inevitable.

[55] Robins J.A. also emphasized at pp. 49-50 O.R. that even where the insured engages in a voluntary act, the resulting injury will be accidental if the harm could not reasonably have been foreseen or expected.

Recovery under a policy insuring against accidental injury is not necessarily confined to cases where there is "an accident" in the sense of an antecedent mishap from which injury results or where injury results from circumstances which can be separately visualized and described as "an accident". An injury may be regarded as accidental where an insured engages in a voluntary act not intending to cause himself harm and the consequent harm could not reasonably have been foreseen or expected. Where, for instance, an insured voluntarily subjects his body to stress, strain or exertion and suffers injury as a result, it is difficult to visualize a separate or external event preceding the injury which could be classified as "an accident". The fact that the [page179] injury happened through the insured's own act does not, however, necessarily prevent it from being an accidental injury. Where the injury is unforeseen, unexpected and without design, and not likely to result naturally or ordinarily from the voluntary or intentional act, but rather constitutes an unusual result, it may be said that it is an accidental injury.

[56] Thus, a common theme runs through the Supreme Court's decision in Martin, the predominant American case law and this court's decision in Voison. Whether a death is accidental turns on the answer to a simple question, which may be framed in several ways: did the insured or a reasonable person in the insured's position expect to die?; was the death unexpected, unintended and unforeseen?; was the death an unlooked for mishap or an unusual result?

b. This case

[57] The motions judge asked the very question that the Supreme Court in Martin said was pivotal: did Mrs. Lin expect to die from her pregnancy and the delivery of her child? The motions judge's answer to that question was "no", and so she found that Mrs. Lin's death was accidental. That finding is a finding of mixed fact and law. As the motions judge asked herself the right question -- that is, she applied the correct legal principle -- her finding is entitled to deference on appeal. Absent palpable and overriding error -- and I see none here -- her finding should not be disturbed. See Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235, 211 D.L.R. (4th) 577; Waxman v. Waxman, 2004 39040 (ON CA), [2004] O.J. No. 1765, 44 B.L.R. (3d) 165 (C.A.). On this basis alone, I would not give effect to the insurer's position on this first issue.

[58] But even if the motions judge's finding is not entitled to any deference, in my view, she is correct. Mrs. Lin became pregnant and gave birth to a child. She hardly expected to die from doing so. Nor would any reasonable person in her position. The risk of death from pregnancy and childbirth was so minute -- no higher than 1 in 14,000 live births -- that Mrs. Lin's tragic demise was entirely an unexpected and unusual result, an unlooked for mishap.

[59] Charron J.A., however, argues that the motion judge's analysis is flawed. My colleague contends that the expectation test in Martin is inapplicable where an insured dies from natural causes. She uses as an example, death from an aneurysm while sleeping. In my colleague's opinion, the expectation test applies only to "consequences flowing from the action or conduct of an insured". She concludes that an amniotic fluid embolism happens "naturally" during childbirth and therefore Mrs. Lin's death was [page180] not accidental. I answer Charron J.A.'s argument with the following five points.

[60] First, in my view, Charron J.A.'s argument finds no support in Martin itself. Quite the contrary. McLachlin C.J.C. stressed that the expectation test was a test that could be generally applied to apparent accidental deaths. She did not carve out any exception for death from natural causes. Indeed, one of the examples the Chief Justice used to illustrate an accidental death -- "the insured may simply have miscalculated the effects of his actions on his body" -- applies to Mrs. Lin's death, since Mrs. Lin had undoubtedly miscalculated the effects of childbirth on her body.

[61] I agree with Charron J.A.'s statement that the expectation test in Martin concerns "the consequences flowing from the action or conduct of an insured". However, when that statement is applied to this case, Mrs. Lin's death can only be characterized as accidental. One consequence of Mrs. Lin's action in becoming pregnant and giving birth was her death -- a consequence so unusual and so unexpected that it must be accidental.

[62] Second, I consider the distinction urged by Metropolitan Life and accepted by my colleague between accidental death and death from natural causes to be a false dichotomy. I assume that Charron J.A. uses the phrase "death from natural causes" in the same sense as the insurer's expert, Dr. Noble, to mean death from processes within one's own body, not death triggered by an external event. I accept that some deaths occurring naturally from processes within one's body may not be accidental: for example, serious heart failure resulting in a fatal heart attack, or severe hardening of the arteries resulting in a fatal stroke, or simply death from old age. But I consider the distinction to be false because in all cases, the only question to be asked is whether the death was unexpected or whether the insured expected to die. See Bornstein v. J.C. Penney Life Insurance Co., 946 F.Supp. 814 (Colo. Dist. Ct. 1996). In my examples, the death is not accidental, not because it occurred naturally, but because it was not unexpected. However, other deaths from "natural causes" may be entirely unexpected and, therefore, accidental.

[63] As Robins J.A. pointed out in Voison, even where an insured voluntarily subjects her body to stress, strain or the like -- as Mrs. Lin did in becoming pregnant and giving birth -- any resulting injury or death will nonetheless be classified as accidental if it is unexpected or unusual. Moreover, the case law is filled with examples where the court has characterized as accidental the unexpected injury or death of an insured, from natural [page181] processes within the insured's own body. Indeed, in several of the cases the injury or death occurred as a result of a normal daily activity.

[64] Here is a sampling of the case law where the injury or death was considered "accidental": a wrenched knee and dislocated cartilage suffered while picking up a marble dropped by a child, Hamlyn v. Crown Accidental Insurance Co., [1893] 1 Q.B. 750 (C.A.) (expressly approved of by Robins J.A. in Voison at p. 51 O.R.); a hernia suffered from swinging a golf club, Claxton v. Travellers Insurance Co. of Hartford (1917), 1917 605 (QC CA), 36 D.L.R. 481 (Que. C.A.) (also expressly approved of by Robins J.A. in Voison at p. 52 O.R.); death from a ruptured aneurysm suffered during sexual intercourse, Carroll v. CUNA Mutual, supra (although recovery denied because the insured's death was predominantly caused by her pre-existing aneurysm and hypertension, not by the accident); death from aspiration or asphyxiation brought on by vomiting after eating, Jones v. Aetna Life Insurance Co., 439 S.W.2d 721 (Colo. App. 1969) and Bobier v. Beneficial Standard Life Insurance Co., 570 P.2d 1094 (Colo. App. 1977); death from a ruptured esophagus caused by severe vomiting after eating, J.C. Murphy v. Continental Casualty Co., 296 So.2d 507 (La. 1972); death from high altitude pulmonary edema during a trek through the Himalayas, Paulissen v. United States Life Insurance Co., 205 F.Supp.2d 1120 (Colo. Dist. Ct. 2002); and Chale v. AllState Life Insurance Co., 353 F.3d 742 (9th Cir. 2003); a fatal heart attack after a pick-up basketball game, Gottfried v. Prudential Insurance Company of America, 414 A.2d 544 (N.J. S. Ct. 1980); a fatal heart attack after a firefighting training exercise at a factory, Winchester v. Prudential Life Insurance Co., 975 F.2d 1479 (10th Cir. 1992) (although recovery was denied because there was no separate bodily injury as required by the policy); a fatal heart attack or stroke after surgery, In a Life Insurance Co. v. Brundin, 533 P.2d 236 (Alaska 1975) and Bornstein v. J.C. Penney Life Insurance Co., supra.

[65] In her reasons, Charron J.A. supports her conclusion with the example of the person who dies of an aneurysm while sleeping. A similar example might be the outwardly healthy 40- year-old who dies suddenly of a heart attack while out for an afternoon walk. According to my colleague, these deaths are not accidental. I agree that in some of these cases the insurer may legitimately refuse to pay accidental death benefits. The insured may have had a pre-existing illness or disease entitling the insurer to rely on the typical exclusion clause in the insurance policy, or other specific language in the policy may preclude coverage. [page182]

[66] But in other cases, why shouldn't the death be characterized as accidental? Some courts have so held. For example, in Gottfried v. The Prudential Insurance Company of America, supra, a five-judge panel of the New Jersey Supreme Court concluded that under a policy insuring for "accidental bodily injury", a 44-year-old insured's death from a heart attack following a pick-up basketball game was "accidental". In the court's opinion an insured would reasonably expect coverage for this unexpected death resulting from vigorous exercise. A contrary conclusion would reinstate the requirement done away with by Cardozo J. in Landress: the need to show accidental means as well as an accidental result. Under the language of the policy in issue in Gottfried -- which is similar to the language of Mrs. Lin's policy -- the insurer was liable because the heart attack was an accidental result. See also Amend v. Equitable Life Assurance Co. of the United States, 342 N.

Y.S.2d 284 (N.Y. Co. Ct. 1972). In addition, the case law listed above argues for coverage. The debate over coverage for these cases leads to my third and fourth points.

[67] My third point is that I do not consider the comparison between Mrs. Lin's death and the insured who dies of an aneurysm while sleeping to be an apt one. Pregnancy and childbirth bear no similarity to routine or normal daily activities. Women may eat, rest, sleep, work and exercise daily; it is not everyday that they become pregnant and give birth to a child. If some separate or triggering event is required for an unexpected death to be deemed "accidental" then pregnancy and childbirth is that event. See Howard v. National Education Assn. of New York, 984 F.Supp. 103 (N.Y. Dist. Ct. 1997) on the need for a triggering event.

[68] Fourth, I consider my colleague to have adopted an overly narrow interpretation of "accident" in circumstances where a generous interpretation is called for. A generous interpretation is warranted because the insurer could always have limited the breadth of "accidental death" by expanding the scope of the exclusions. After all, the argument that death from natural causes should not be considered accidental is not novel. If Metropolitan Life wished to preclude recovery for death from a "natural cause" it could easily have said so in the policy. It did not do so.

[69] Fifth, it seems to me the insurer's best position is that the meaning of "accident" -- and whether its meaning extends to Mrs. Lin's death -- is ambiguous. The law unequivocally states that ambiguities in an insurance policy must be resolved against the insurer and in favour of the insured. See for example, Somersall v. Friedman, 2002 SCC 59, [2002] 3 S.C.R. 109, 215 D.L.R. (4th) 577. [page183]

[70] For all these reasons, I would not give effect to the insurer's position that Mrs. Lin's death was not the result of an accident.

Second issue: Is amniotic fluid embolism (or embolus) a "physical illness"?

[71] Metropolitan Life's obligation to pay under the accidental death benefit rider is qualified by a list of exclusions. The one exclusion relevant to this appeal provides that the insurer need not pay if the death "is caused or contributed to, directly or indirectly, by physical or mental illness or treatment for an illness". Mrs. Lin died of an amniotic fluid embolism. Therefore, causation is not in question; she had no mental illness; and she was not treated for illness. The resolution of the second issue turns on whether amniotic fluid embolism is a "physical illness", or perhaps, more precisely, whether amniotic fluid embolus is a "physical illness".

[72] I frame the question in the alternative to reflect the distinction properly made by the insurer's expert, Dr. Noble, though not by counsel, between an embolus and an embolism. An embolus (the plural is emboli) is a detached blood clot or a bit of tissue. An embolism is an obstruction or occlusion of a vessel by an embolus. Typically the embolus travels through the blood stream and becomes lodged in a blood vessel, usually in the heart, lungs or brain. An amniotic fluid embolism is the obstruction and constriction of pulmonary blood vessels by amniotic fluid entering the maternal circulation, causing obstetric shock. See the Signet/Mosby Medical Encyclopedia (New York: New American Library, 1985); and Stedman's Medical Dictionary, 27th ed. (Baltimore: Lippincott Williams and Wilkins, 2000). I turn now to the issue. I do so bearing in mind that because the term "physical illness" is contained in an exclusion clause it must be interpreted narrowly.

[73] My colleague contends that the motions judge's analysis on this issue is flawed because she held only that pregnancy is not an illness and she did not expressly say that amniotic fluid embolism is not an illness. However, in holding that the insurer could not rely on the exclusion clause, it seems to me the motions judge implicitly concluded that amniotic fluid embolism or embolus is not a physical illness. Nonetheless, I will accept Charron J.A.'s point and tackle her argument head on.

[74] My colleague relies on Dr. Noble's opinion to find that amniotic fluid embolism is a physical illness. For convenience, I reproduce in full his affirmative answer to the question whether amniotic fluid embolus is a physical illness. [page184]

Yes, Amniotic Fluid Embolus is a physical illness. It results when the separation of amniotic fluid from the mother's circulation is crossed as uterine veins are opened by the birthing process (regardless of method). The Amniotic Fluid Emboli move into the mother's veins and go to her heart and lungs. This may begin the classic disease process described in Stacey Chui [sic] Lin's case. Since every mother, while giving birth by any method, may suffer an amniotic fluid embolus, this is a physical illness. A comparison may be made between Amniotic Fluid Embolism and a stroke or a heart attack. All North Americans over 55 years have artherosclerosis (hardening of the arteries) but only a few will suffer the stroke or heart attack that may result depending on severity. It cannot be considered an accident -- it is a rare event created by a process within our body. The same applies to Amniotic Fluid Embolism. All of the conditions required for a death from amniotic fluid embolism (pregnancy and open maternal uterine veins) exist in all pregnant women but the load of emboli is small or the reaction is not major and so most women do survive pregnancy. Rarely, this physical illness claims a life as it unfortunately did in Stacey Chiu Lin's case.

[75] Although pregnancy itself is not an illness, according to Dr. Noble every pregnant woman has a physical illness. That is because every pregnant woman has amniotic fluid emboli, which move into her uterine veins and travel to her heart and lungs. In the same way every person on this continent over 55 has a physical illness, hardening of the arteries. But just as all North Americans over 55 have hardening of the arteries -- artherosclerosis -- yet only a few will suffer a stroke or heart attack, all pregnant women have amniotic fluid emboli and open uterine veins, yet only a very few will suffer an embolism.

[76] None of the medical and scientific literature appended to the opinion of the respondent's expert Dr. Steinberg supports Dr. Noble's opinion. That said, I disagree with Charron J.A. because in my view Dr. Noble's expert medical opinion is irrelevant to determining the meaning of physical illness, and thus is irrelevant to determining whether amniotic fluid embolus or embolism is a physical illness. See Voison, supra, at p. 49 O.R.

[77] The insurance policy does not contain a medical definition of physical illness. Indeed, Metropolitan Life chose not to define the term at all. Thus, as with the word "accident", the term "physical illness" must be given its ordinary and popular meaning. What matters is the ordinary person's understanding of the term. To suggest that the ordinary person would think that any time a woman becomes pregnant she has a physical illness because she has a cluster of emboli and open uterine veins, which in .00007 per cent or fewer live births combine to produce a fatal amniotic fluid embolism is, respectfully, nonsensical. None of the usual indicia that ordinary people think of when they think of a physical illness are present -- no symptoms, no outward manifestations of illness, no progressively worsening condition. [page185]

[78] To establish a sensible working definition of physical illness -- a definition that accords with insureds' reasonable expectations -- I again turn to Justice Cardozo. Exclusions for "physical illness", "disease" or "bodily infirmity" are typical in accidental death benefit policies and they essentially mean the same thing. In interpreting these terms, most American courts apply the reasons of Cardozo J. in Silverstein v. Metropolitan Life Co., 254 N.Y. 81 (N.Y. 1930). In that case, Cardozo J. held at p. 84, that the meaning of the term in common parlance should govern, and he distinguished between a "disease" or "infirmity" and a "predisposing tendency":

In a strict or literal sense, any departure from an ideal or perfect norm of health is a disease or an infirmity. Something more, however, must be shown to exclude the effects of accident from the coverage of a policy. The disease or the infirmity must be so considerable or significant that it would be characterized as disease or infirmity in the common speech of men . . . . A policy of insurance is not accepted with the thought that its coverage is to be restricted to an Apollo or a Hercules.

A distinction, then, is to be drawn between a morbid or abnormal condition of such quality or degree that in its natural and probable development it may be expected to be a source of mischief, in which event it may fairly be described as a disease or an infirmity, and a condition abnormal or unsound when tested by a standard of perfection, yet so remote in its potential mischief that common speech would call it not disease or infirmity, but at most a predisposing tendency. (Citations omitted)

See also J.A. Bryant Jr., "What Conditions Constitute Disease Within Terms of Life, Accident, Disability or Hospitalization Insurance Policy", 461 A.L.R. 3d 822; Katskee v. Blue Cross/ Blue Shield of Nebraska, 515 N.W.2d 645 (Neb. 1994).

[79] Cardozo J.'s common sense distinction shows that amniotic fluid embolus or embolism is not a physical illness. It is not an abnormal condition of such quality or degree that in its natural and probable development may be expected to be a source of mischief. Amniotic fluid emboli and open maternal uterine veins do not "naturally and probably" produce a fatal embolism. The opposite is the case. Their existence in pregnant women is a condition so remote in its potential mischief that common speech would call it not a physical illness, but at most a predisposing tendency. It could hardly be otherwise when at most .00007 per cent of pregnant women die from it.

[80] The flaw in relying on Dr. Noble's expert medical opinion to establish that amniotic fluid embolus is a physical illness may be demonstrated by considering the very comparison he uses: hardening of the arteries that affects all North Americans over age 55. Although medical opinion may regard artherosclerosis that affects all older people to be a physical illness, courts have [page186] taken exactly the opposite view. For good reason. Ordinary people who have normal hardening of the arteries for their age do not think of themselves as having a physical illness. The case law recognizes this. It recognizes that denying recovery under an accidental injury or death policy because of an "infirmity" caused by the normal aging process does not accord with an insured's reasonable expectations. Normal degeneration is inevitable. It does not amount to an illness or disease.

[81] An example from the case law is the judgment in Reed v. United States Fidelity & Guaranty Co., 491 P.2d 1377 (Colo. 1971). In that case, a 56-year-old firefighter died of smoke inhalation. His widow claimed under an accidental insurance policy. The insurer denied coverage, arguing that the deceased had a disease or infirmity: artherosclerosis. The medical evidence showed that his artherosclerosis was well within normal limits for a 56-year-old man, and the court therefore rejected the insurer's argument. In common parlance, the man was not "diseased" or "infirmed":

We note that both doctors testified that Mr. Reed's condition was well within the normal limits for a 56-year-old man, that had he not inhaled smoke he could have looked forward to a normal life expectancy, and that artherosclerosis was a part of the normal aging process. We also note that Mr. Reed's family doctor, his wife and his colleagues on the fire department agreed that prior to his accident, Mr. Reed was a man of apparently robust good health.

We do not believe that as a matter of law these facts describe a man who is "diseased" or "infirm" as those terms are used in common parlance . . . . (At p. 1380)

[82] In a similar way, Mrs. Lin had conditions found in all pregnant women. Sadly, she suffered an amniotic fluid embolism, or what Dr. Steinberg states should more properly be characterized as an anaphylactoid syndrome of pregnancy, akin to a severe allergic reaction. This severe allergic reaction -- a hypersensitivity to the conditions found in all pregnant women -- in ordinary language does not amount to a physical illness. See Denton v. Travelers Insurance Co., 25 F.Supp. 556 (Md. Dist. Ct. 1938); and Berkowitz v. New York Life Insurance Co., 256 A.D. 324 (N.Y. 1939); Fiffick v. Econ-o- Check Corp., supra; Gaskins v. New York Life Insurance Co., 104 So.2d 171 (La. 1958).

[83] An alternative way of thinking of Mrs. Lin's condition is to consider it a "temporary disorder" uniquely associated with pregnancy and childbirth. On this view, even Dr. Noble's opinion does not support characterizing amniotic fluid embolus as a "physical illness". Many cases have held that "temporary disorders" do not preclude recovery under an accidental benefit insurance policy even where the policy denies benefits for loss caused or contributed to by disease or physical illness. See J.M. Bryant, supra. [page187]

[84] Accordingly, I would not give effect to the insurer's position on the exclusion clause.

C. Conclusion

[85] In my opinion, the motions judge did not err in holding that Mrs. Lin's death was accidental, and she did not err in holding that the exclusion clause was inapplicable. I would therefore dismiss the appeal with costs to the respondent in the amount claimed, $8,300, inclusive of disbursements and GST.

Appeal allowed.