Deep Vein Thrombosis and Air Travel Group Litigation, Re

This appeal arises out of group litigation between the appellants ("the passengers") and the respondents ("the carriers"). Most of the passengers allege that they have suffered deep vein thrombosis ("DVT"), leading to serious injury, as a result of travelling in the carriers' aircraft. In some particularly tragic cases the DVT proved fatal and the claim is brought by the personal representatives of the passenger.

The passengers number 24. Their claims include an allegation that the carriers are liable to them under the provisions of Article 17 of the 1929 Warsaw Convention as amended and incorporated into English law by the Carriage by Air Act 1961. The carriers number 18. They include 15 out of the leading 25 international airlines. They deny that the passengers suffered DVT as a result of travelling in their aircraft, but contend that, even if they did, this does not give rise to valid claims under Article 17.

On 20 December 2002 Nelson J gave judgment in respect of four preliminary issues. This appeal relates to only one. At the outset of his judgment Nelson J described this issue as "whether, on the basis of the agreed specimen matrix, the Claimants have a claim under Article 17 of the Warsaw Convention for the DVT allegedly suffered by them, and in particular whether the specimen matrix discloses an 'accident' for the purposes of Article 17". The judge held that the specimen matrix did not disclose an "accident" and the passengers appeal against that decision with the permission of the judge.

The following articles of the Warsaw Convention are relevant to the issue before us:

Article 22 imposes limits of the liability that is imposed by Articles 17, 18 and 19.

The specimen matrix is in the following terms:

It is thus agreed, for the purpose of the issue before us, that the specimen claimant suffered DVT which was "caused by the flight". For the passengers, Mr Cakebread contends that this statement in the specimen matrix falls to be supplemented by the following pleaded "Further Information", supplied under Part 18 of the Civil Procedure Rules:

Mr Robert Lawson, who appears for all the carriers except Qantas, has not objected to this course, subject to one reservation. He submits that for the purpose of the preliminary issue these particulars must be treated as typical and usual features of travel by air at the material time. Neither Mr Cakebread nor Mr Bankim Thanki QC, who appears for Qantas, has demurred to that submission, and I shall proceed on the basis suggested by Mr Lawson.

The issue before us depends, essentially, on the meaning of "accident" in Article 17. The parties start from common ground. In the context of Article 17 "accident" refers to the cause of the injury and not to the injury itself. That proposition and the relevant test of an "accident" were accurately identified by the United States Supreme Court in Air France v Saks (1985) 470 US 399 at p 405:

Mr Cakebread has made submissions both as to the meaning that should be accorded to "unexpected or unusual" in the Saks definition and as to the meaning to be accorded to "event or happening".

He submits that the manner of operating an aircraft will be "unexpected or unusual" in so far as it departs from the manner in which the airline could "reasonably be expected" to operate its aircraft, applying an objective test and having regard to all the material circumstances.

As to "event or happening", Mr Cakebread submits that this covers both acts and omissions of those operating the aircraft. Thus if those operating the aircraft act in a way which one would not reasonably expect, this will constitute an "accident". Equally, if they fail to take some action which one would reasonably expect them take, this also will constitute an "accident".

Mr Cakebread applies these principles to the facts in the specimen matrix and the Further Information as follows. The carriers knew that the usual features of travel by air, as detailed in the Further Information, created a risk of causing DVT. Having regard to that knowledge they could reasonably be expected to avoid or mitigate those features. Subjecting passengers to those features constituted "unexpected or unusual" conduct and, in consequence, an "accident" or "accidents" within Article 17.

Mr Cakebread further submits that the facts in paragraph 7 of the specified matrix disclose further accidents, as follows. The operators of the aircraft were aware of the risk of DVT. They could, without difficulty or expense, have warned passengers of that risk and given them advice as to how to reduce it. In these circumstances they could reasonably have been expected to give the warning and advice in question. Their omissions to do so constituted "accidents" within Article 17.

Mr Lawson accepts that the definition of "accident" advanced by the Supreme Court in Saks has to be applied flexibly and in a purposive manner. Thus "accident" has been held to cover a deliberate attack by terrorists or molestation by a fellow passenger. There is, however, no escape from the requirement that there should be an accident causing the injury. It is thus necessary to identify some event which can properly be described as an accident.

As to the requirement that the event should be "unexpected or unusual", Mr Lawson submits that the touchstone of what is expected or usual is the practice generally adopted by airlines. An act which complies with such practice cannot properly be described as an "accident" within Article 17. Idiosyncratic, though regular, behaviour by an individual airline is capable, by reason of being unexpected or unusual, of qualifying as an accident. The preliminary issue is proceeding, however, on the premise that the manner of operation of the specimen aircraft reflects the general practice of all the airlines who are party to this litigation. No aspect of that practice can have the quality of an "accident".

It is apparent from the judgment of Nelson J that Mr Cakebread argued before him that the construction of "accident" should be informed by two themes or principles which underlay the Convention -- the "fault based" theory of liability and the "apportionment of risk" theory of liability. The judge did not accept that the Convention reflected either theory, as advanced by Mr Cakebread, or that either theory was of relevance to the task of construing Article 17.

Mr Cakebread has not sought to build on those arguments before us. I consider that Nelson J was correct to conclude that these theories did not assist the task of construction. He cited a passage from the speech of Lord Hope in Morris v KLM Royal Dutch Airlines [2002] UKHL 7; [2002] 2 WLR 578 at p 601 as setting out the general character and purpose of the Convention. I shall not repeat that exercise, but simply make two points. (1) The liability imposed by Article 17 is not fault based. (2) The apportionment of risk in relation to death or bodily injury in the course of carriage by air depends upon whether or not it has been caused by an accident. If it has been caused by an accident the carrier will be liable, subject to the other relevant Articles. If it has not been caused by an accident, the passenger or the passenger's personal representatives will have no claim against the carrier.

I propose to consider this question in the first instance having regard to the natural meaning of the words used in Article 17, referring to authority only by way of illustration. I shall consider the effect of my conclusions on the issue under consideration, thereby reaching a provisional view as the outcome of this appeal. Finally I shall turn to the authorities to see whether they lead me to a result that differs from my provisional view.

The parties are agreed in taking as their starting point the passage in the Opinion of the Supreme Court in Saks that I have already quoted. In Morris v KLM [2001] EWCA Civ 790; [2002] QB 100 at p 110 I remarked that this definition gave the word "accident" a natural and sensible meaning in the context of Article 17. In the House of Lords, at paragraph 71, Lord Hope agreed. I remain of that view.

Article 17 speaks of the accident causing damage sustained in the event of death or wounding or bodily injury. This is generally treated as requiring that the accident should be the cause of the death, wounding or bodily injury -- see, for instance, Shawcross & Beaumont on Air Law 4th Ed at Chapter VII, paragraph 693. It seems to me that the concept of an accident taking place on board an aircraft which causes death, wounding or bodily injury is a simple one. The words naturally suggest an untoward event which impacts on the body in a manner which causes death, wounding or injury. The impact of the event may be physical or it may be an impact on the senses which results in bodily injury, but one would normally expect the untoward event to cause the death or injury directly.

I do not suggest that these reflections provide an exclusive description of the "accident" in Article 17. I note that in Saks the Supreme Court stated at page 406:

One can break down the definition of an accident into two elements. (1) There must be an event; (2) the event must be unusual, unexpected or untoward. I would endorse the approval by the United States Court of Appeals, Third Circuit, of the following charge to a jury as to the correct legal standard for determining the occurrence of an accident under Article 17:

I turn to consider in a little more detail the first element of an accident. It must be an event. The first meaning of accident that is given by the Oxford English dictionary, 2nd Ed, is: "Anything that happens. An occurrence, incident, event." That meaning is however obsolete, in that it has become replaced by the second meaning that the Dictionary gives: "Anything that happens without foresight or expectation; an unusual event, which proceeds from some unknown cause, or is an unusual effect of a known cause; a casualty, a contingency." The first element of an accident is to be found in all of these.

A critical issue in this appeal is whether a failure to act, or an omission, can constitute an accident for the purposes of Article 17. Often a failure to act results in an accident, or forms part of a series of acts and omissions which together constitute an accident. In such circumstances it may not be easy to distinguish between acts and omissions. I cannot see, however, how inaction itself can ever properly be described as an accident. It is not an event; it is a non-event. Inaction is the antithesis of an accident.

Mr Thanki drew our attention to a description of the natural meaning of an event given by Lord Mustill, in a quite different context, in Axa Reinsurance (UK) Plc v Field [1996] 2 Lloyd's Rep 233 at p 239:

Mr Cakebread has submitted that it is, or may be, possible to identify two "accidents" in the specimen matrix, as augmented by the Further Information. The first can, I believe, be fairly described as subjecting the passengers to carriage by air that has the features itemised in the Further Information. These features have been described by the judge, and by Mr Cakebread, by the generic description "providing cramped seating". The Further Information also makes allegations about air pressure and the atmosphere and temperature in the cabin. The act of altering the pressurisation, the supply of oxygen or the temperature in the cabin would all, in my opinion, be capable of constituting an "event" that satisfied the first element of an "accident". As I understand it, however, the allegations in respect of those matters are intended to treat them, like cramped seats, as integral features of carriage pertaining throughout the flight rather than the effects of acts of the crew in the course of the flight.

I do not consider that existence of these permanent features of the aircraft, or the subjecting of the passengers to carriage in aircraft with these features, is capable of amounting to an event that satisfies the first limb of the definition of an accident which "took place on board the aircraft or in the course of any of the operations of embarking or disembarking". For this reason alone I would hold that the flight itself, albeit that under the matrix it is assumed to have caused the DVT, was not capable of amounting to an "accident" for the purposes of Article 17.

I turn to the failure to warn of the risk of DVT, or to advise on precautions which would avoid or minimise that risk. I cannot see how this failure can be categorised as an accident. It was simply something that did not happen -- a non-event. For this reason alone I would hold that the failure to warn of or to advise about the dangers of DVT is not capable of amounting to an accident under Article 17.

Nelson J reached the same conclusion that I have in relation to the question of whether the matrix disclosed an event. He observed in paragraph 118: "Here, the agreed matrix of fact reveals no event or happening which occurred on the flight." At paragraph 121 he repeated: "Nor on the factual matrix is there any event which is external to the passenger save for the happening of illness itself which does not qualify under article 17 . . . " These terse findings were fatal to the passengers' case. For the reasons that I have given, and subject to consideration of the case law, I consider that they were correct.

The question of whether there took place on the flight events which were "unexpected or unusual" was and is unreal, having regard to the judge's conclusion and my provisional conclusion that there was no relevant event. Nonetheless much time was devoted before us to discussion of whether the features relied upon by Mr Cakebread as constituting accidents had the necessary quality of being unexpected or unusual.

The judge remarked of the matrix, at paragraph 118, "the flight can only be described as normal and unremarkable", and in the following paragraph: "If however nothing happened during an ordinary and unremarkable flight that involved the actions of anyone except for the passenger himself or herself and his or her atypical reaction to a normal and unremarkable flight, there has been no unexpected or unusual event or happening." It is plain the judge thought that this fairly described the position.

Mr Cakebread criticised these conclusions on the part of the judge. He suggested that he "failed to apply the Saks test objectively and flexibly but instead applied a subjective test, using the carrier's own standard procedure as a benchmark." This criticism was misconceived. As I have already observed, the specimen matrix must be considered as setting out, not the idiosyncratic procedure of a single airline, but the typical procedures of the industry as a whole at the material time.

Mr Lawson submitted that this fact precluded any suggestion that the features relied upon by Mr Cakebread, including the failures to warn and advise, were "unexpected or unusual". In answer to this Mr Cakebread advanced an argument which, so it seems to me, was a refinement on the arguments advanced before the judge. He contended that the test of "unusual or unexpected" did not involve a comparison between the normal incidents of a flight at the time and what in fact occurred. The test involved asking "would it have been reasonable to expect the carriers to act as they did, having regard to the knowledge that they had about the risk of causing DVT?"

Mr Cakebread submitted that a carrier could reasonably be expected to take reasonable steps to warn of, minimise or exclude known dangers to the health, safety and lives of its passengers. Its failure to do so was "unusual and unexpected" within the Saks test.

Mr Lawson took a stand of principle against this approach. He submitted that it was, in effect, a test of whether the carrier was at fault. Because Article 17 did not involve a fault based test, it was not legitimate to have regard to questions of fault when considering whether an event was unusual or unexpected.

I am not persuaded that, when considering whether an action has the unusual, unexpected or untoward quality necessary to enable one to qualify it as an accident under the Saks test, the questions of whether the actor has knowledge which makes it culpable to perform the action can never be relevant. I am, however, persuaded that it is simply not possible to apply to a state of affairs, or an omission to act, the test that is relevant to deciding whether an event is an accident.

Looking at the position overall, which is the only realistic way of looking at it, the matrix, as augmented by the Further Information, discloses nothing which is capable of constituting an "accident" within the meaning of Article 17.

There is no English case law that bears directly on the issue before us. We are, however, dealing with an international convention and decisions of courts of foreign jurisdictions in relation to that convention can be of great assistance and deserve careful consideration. Thus the general approach to the meaning of accident of the Supreme Court in Saks has been widely accepted and has been adopted by the parties as the starting point for considering the issue on this appeal.

Many of the cases are decisions of courts in the United States. There, until a recent decision of the Supreme Court, most trial courts entertained claims against carriers in circumstances that fell outside Articles 17, 18 and 19, based on causes of action other than those afforded by the Convention. Such claims were not subject to the Convention limits of liability. In consequence it was commonplace to find the carriers contending that a passenger's injury had been caused by an accident and that their liability was limited, while passengers were contending that they had a good claim that did not involve an injury caused by an accident. The position changed when, in 1999 in El Al Israel Airlines Ltd v Tseng (1999) 525 US 155, the Supreme Court held that Article 24 of the Convention precluded claims other than those brought under the Convention. Thereafter it was the passengers who strove to establish that they had sustained injuries caused by accidents and the carriers who denied this.

Where passengers fall ill in the course of carriage by air they reasonably expect that they will receive any appropriate medical care that the cabin staff are in a position to provide and where the illness is life-threatening, aircraft will often make an emergency diversion. Some of the cases involve situations in which the carrier is alleged to have failed to make the positive response to the medical emergency that was reasonably to be expected. In such circumstances the comment Lord Hope of Craighead made in Sidhu v British Airways plc [1997] AC 430 at pp 453-4 is particularly apposite:

Rather than refer to a mass of authority, I shall quote the useful general summary of the position in Shawcross & Beaumont at Chapter VII, paragraph 701:

That decision was approved in Saks at pp 404 and 405. It accords with my provisional view that there is nothing in the specimen matrix that is capable of constituting an accident. I must now turn, however, to a number of cases, some of which are not easy to reconcile with Abramson, or indeed with my provisional view.

Fulop v Malev Hungarian Airlines (2001) 175 F Supp 2d 651 is a decision of District Judge Marrero, sitting in the United States District Court, SD New York. Mr Fulop, the passenger, had a heart attack in the air and claimed that the consequences were aggravated because the airline refused to deviate so that he could receive urgent attention. The airline sought summary judgment. The judge held that there was a case to go for trial. The essence of his reasoning appears in the following passage of a lengthy judgment:

The judge considered a number of authorities, including Abramson. He distinguished these on the basis that Mr Fulop alleged that, in refusing to deviate, the airline had departed from its own internal procedures and that this was capable of amounting to an unusual or unexpected event which constituted an accident.

I do not find it impossible to reconcile the result in this case with my provisional conclusion that failure to warn or to advise in respect of the risk of DVT cannot constitute an accident. The failure to deviate was the corollary of the decision to fly on to the plane's destination notwithstanding the fact that a passenger had had a heart attack. If this was contrary to standard practice, or even the airline's internal practice, then it is perhaps arguable that there occurred an unusual or unexpected "event". In these circumstances and having regard to the conflicting authority, to which the judge had regard, it is understandable that he was not prepared to give summary judgment in favour of the airline. The reasoning in this case does not, and cannot, lead to the conclusion that mere inertia, such as the failure to warn or advise, is capable of constituting an accident.

One decision which influenced Judge Marrero was the decision of the United States District Court N.D. California in Husain v Olympic Airways (2000) 116 F Supp 2d 1121. That decision was subsequently appealed to the Ninth Circuit, Court of Appeals. It was a claim in respect of the death in flight of Dr Hanson. Dr Hanson suffered from asthma. He was particularly sensitive to second-hand smoke. He, his wife and family requested and obtained seats in the non-smoking section of the plane, but only a few rows in front of the smoking section. On three occasions his wife asked if he could be moved to a seat further from the smoking section. The flight attendant refused this request, untruthfully asserting that there were no spare seats. Smoke from the smoking section percolated to him and made him ill. He collapsed and died.

District Judge Breyer held that his death had been caused by an accident, rejecting the airline's contention that no unusual event had occurred. His reasoning appears in the following passage at pp 1131-1132:

The Court of Appeals upheld this decision - (2002) 316 F 3d 829. Their reasoning appears at p 837:

I have no difficulty with the result in this case but, with respect, I question the reasoning of the judge in both events. The refusal of the flight attendant to move Dr Hanson cannot properly be considered as mere inertia, or a non-event. It was a refusal to provide an alternative seat which formed part of a more complex incident, whereby Dr Hanson was exposed to smoke in circumstances that can properly be described as unusual and unexpected. The existence of the non-smoking zone provided the opportunity for Dr Hanson, if suitably placed within it, to avoid exposure to the smoke that threatened his health and, as it proved, his life. The direct cause of his death was the unnecessary exposure to the smoke. The refusal of the attendant to move him could be described as insistence that he remain seated in the area exposed to smoke. The exposure to smoke in these circumstances could, in my view, properly be described as an unusual or unexpected event. While smoke in that part of the cabin was not itself unusual or unexpected, the same cannot be said of Dr Hanson's enforced exposure to that smoke.

Once again the inaction, if so it can properly be termed, of the flight attendant is not comparable to the failure to warn or to advise, that Mr Cakebread contends constitutes an accident in the present case. This decision does not cause me to reconsider my provisional view that the matrix discloses no accident.

In August 2001 District Judge Kent, sitting in the United States District Court, SD Texas, considered the effect of Abramson and similar decisions in McCaskey v Continental Airlines (2001) 159 F Supp 2d 562. The relevant allegation in that case was that the airline had failed to divert after a passenger had suffered a stroke and that this failure was an accident under Article 17, which caused his death. The airline's motion for summary judgment was rejected. After referring to Abramson and other cases, the judge said at p 574:

My comments in respect of Fulop apply equally in respect of this decision. It does not put in question my provisional view as to the result in the present case.

Much more pertinent is the decision of the same judge in Blansett v Continental Airlines (unreported, 11 December 2002). Mr Blansett suffered a cerebral stroke on a flight to London, allegedly as a consequence of developing a DVT. His claim against the airline alleged that they failed to warn, advise and/or take adequate precautions to inform about and/or prevent DVT on their international flight. The airline's claim for summary judgment was rejected. After considering the authorities, including those to which I have referred, the judge held at pp 6-7:

I can readily appreciate the policy considerations that influenced this decision. I do not, however, consider that the reasoning adequately addresses the need to identify an event as an essential element of any accident. A failure to institute routine safety procedures, to reprint safety placards and to make announcements once or twice in the course of flight cannot be said to constitute an event or an "accident", without abusing the flexibility that is proper when defining that word beyond breaking point. In short, I do not agree with this decision.

Finally I come to the decision upon which Mr Cakebread's submissions to us were largely founded. On the day that Nelson J delivered his judgment, Bongiorno J, sitting in the Supreme Court of Victoria, gave judgment in Povey v Civil Aviation Safety Authority and Others [2002] VSC 580. Because of the time difference between Melbourne and London Nelson J was able to consider this decision before he handed down his judgment. It did not cause him to change his mind, or to call for further argument. We are told that the Melbourne case was a test case upon which hundreds of similar cases turned.

In issue was an application by Qantas and British Airways for summary judgment in respect of a claim brought by a passenger against them under Article 17. The bodily injury alleged to have been caused was DVT, which led to a pulmonary embolism, a stroke and permanent disability. The passenger's pleaded case gave particulars of the alleged accident which were restricted to the conditions within the cabin. The accident was said to have occurred throughout the duration of two flights -- one from Sydney to London and the return flight from London to Sydney a few days later.

The judge held that the pleaded particulars, whether individually or collectively, were incapable of falling within the most liberal definition of "accident". He went on to hold, however, that the matter did not end there. Counsel for the claimant had orally advanced two additional allegations which, if pleaded, would disclose a viable claim. Accordingly summary judgment should not be entered.

The first such allegation was that the airlines had certain knowledge about DVT, its causes, its relationship to the cramped conditions in economy class and precautions which could be taken to minimise the risk of it occurring. The second allegation was that it was commonplace for airlines to issue warnings and advice to passengers in respect of risks of which they had notice.

The judge cited Husain as demonstrating that "an accident, as Saks uses the word, may include action or inaction by airline staff". He continued at paragraph 39:

The most critical part of the judgment is that in which the judge considered whether a failure to warn or advise could constitute an event on board an aircraft. He said at paragraph 44:

There can be no quarrel with the last sentence of this paragraph but, with respect, I cannot accept that the first two sentences deal adequately with the difficulty of bringing a failure to warn or advise within the definition of an accident. It is true that the distinction between acts and omissions can, in some circumstances, be dubious. Husain, upon which the judge relied so heavily, is an example of such circumstances, as I have already explained. But there is no dubious distinction in the case of a failure to warn or give advice. Such a failure is mere inaction or inertia. There is no way in which it can be suggested that such inaction is tantamount to action. As for Article 25, it cannot properly be inferred that because this Article refers to an "act or omission", Article 17 makes a carrier liable for inactivity. "Act or omission" is a general phrase that is apt to cover situations where the distinction between the two is indeed dubious. Furthermore, the ambit of Article 25 is wider than that of Article 17. If, for instance, an employee recklessly fails to carry out maintenance on a plane, with the result that it suffers an engine failure and crashes, passengers or their personal representatives will have claims under Article 17 by reason of the fact that the crash was an accident. No independent claim will lie under Article 17 for the employee's failure to carry out the maintenance, which plainly falls outside the Article. But, so it seems to me, Article 25 will apply so as to preclude the right to limit

For these reasons I am unable to concur with the conclusion of Judge Bongiorno that a failure to warn or to advise in relation to DVT is capable of constituting an accident under Article 17. Nelson J clearly had the same reaction as mine.

The decision in Povey is in conflict with an earlier decision of Knight J, sitting in the District Court of New South Wales -- Van Luin v KLM (unreported, 11 October 2002). The plaintiff claimed in respect of a DVT, allegedly sustained in flight. The defendant airline sought to have the statement of claim struck out as disclosing no cause of action. The plaintiff sought to save the pleading by applying to amend to allege a failure on the part of the air crew to advise her that she ought to move around the cabin and drink extra fluid. The judge refused permission to amend and dismissed the claim on the ground that it disclosed no reasonable cause of action. He held at paragraph 50:

A month earlier, in the Superior Court of Justice in Ontario, Hermiston J in McDonald v Korean Air (unreported, 18 September 2002) had dismissed a similar claim for the same reason. The Ontario Court of Appeal affirmed this decision in a judgment consisting of a single sentence on 18 February 2003.

On 7 February 2003 Boulton DCJ, sitting in the District Court of Queensland, in Rynne v Lauda-Air Luftfahrt AG [2003] QDC 004 struck out the plaintiff's claim on the grounds that it had no real chance of succeeding. The plaintiff claimed to have suffered DVT in the course of a flight from Vienna to Sydney. She relied by way of an accident under Article 17 essentially on failures to warn and advise about DVT. The judge referred to Povey, to Van Luin, to McDonald and to the decision of Nelson J in the present case. He observed at paragraph 30 that he found the analysis of Nelson J to be absolutely compelling.

I share the view of Boulton DCJ, and had been tempted to follow the example of the Ontario Court of Appeal by giving judgment in a single sentence. I have not done so because I feel sympathy for the claimants in this group litigation and feel that they deserve a full explanation as to why their appeal has not succeeded. I would add that, if their claim goes no further, this may prove a blessing in disguise. It seems to me that if they had got over the hurdle of establishing an arguable case under Article 17, they would have faced litigation involving difficult issues of causation which would have been very costly to resolve and the outcome of which would have been questionable.

For the reasons which I have given, I would dismiss this appeal.

LORD JUSTICE JUDGE: I agree with the Master of the Rolls. I shall add a few words of my own by way of emphasis only. In this appeal we are not considering a claim for damages for negligence, but rather whether, on the basis of simple facts, undisputed for present purposes, a number of international air carriers would be liable under Article 17 of the Warsaw Convention to passengers who developed deep vein thrombosis while the aircraft in which they were travelling were in flight. Liability will arise only " .... if the accident which caused the damage so sustained took place on board the aircraft .... "; otherwise it will not.

Numerous authorities from around the world, designed to assist our understanding of the meaning of these brief words, have been drawn to our attention. Notwithstanding the examination and re-examination of Article 17 in different factual situations, as and when they have arisen, in the present litigation the question which requires an answer is whether, in the deemed but so far unproved circumstances accepted as accurate for present purposes by all the parties to the litigation, the passengers can establish that an accident within the meaning of Article 17 took place during their respective flights.

The starting point for the proper understanding of Article 17 is found in the decision of the Supreme Court of the United States in Air France v Saks (1985) 470 US 392. Valerie Saks alleged that she had become permanently deaf as the result of an injury caused during flight which resulted from negligent maintenance and operation of the aircraft's pressurisation system. By the time the case came before the Supreme Court, the argument proceeded on the basis that the evidence showed that the pressurisation system had worked normally throughout the flight. Therefore, it was contended that the injury did not qualify as an "accident" within Article 17. The claimant had successfully argued, before the Court of Appeal for the Ninth Circuit, that normal changes in cabin pressure did so qualify.

In a close analysis of the text of the Warsaw Convention in both French and English, and after a study of the negotiation which led to the Convention, Justice O'Connor delivered the Opinion of the Supreme Court. She noted that the word "accident" was not defined in the Convention, and observed that the Convention itself did not help to illuminate its meaning. She pointed out that "accident" was a word which could be used "in many senses", and referred to an observation by Lord Lindley in Fenton v J Thorley and Co (1903) AC 443, at p 453, that "the word 'accident' is not a technical legal term with a clearly defined meaning".

Justice O'Connor described "two clues" to the meaning of "accident" in Article 17. First, as different words were used in Article 17 (accident) and Article 18 (occurrence), it was unlikely that the two words had the same meaning. In short, they could not simply be transposed. Second, Article 17 was concerned with "an accident which caused the passenger's injury, and not an accident which is the passenger's injury".

In the result, Justice O'Connor came to the conclusion that for liability under Article 17 to arise, the injury must be caused by an "unexpected or unusual event or happening that is external to the passenger", later adding " .... but when the injury indisputably results from the passenger's own internal reaction to the usual, normal and expected operation of the aircraft, it has not been caused by an accident, and Article 17 of the Warsaw Convention cannot apply".

Significantly, in the present context, the judgment referred to three examples where the claimant failed to establish an accident. These were Abramson v Japan Airlines 739 F 2d 130 [CA3 1984] (sitting in an airline seat during normal flight which aggravated hernia, not an "accident"; McDonald v Air Canada 439 F 2d 1402 [CA5 1971] (fainting while waiting in the terminal for one's baggage, not shown to be caused by an "accident"; Sherer v PanAmerican World Airways Inc 54 App Div 2d 636, 387 NYS 2d 580 (sitting in an airline seat during normal flight which aggravated thrombophlebitis, not an "accident"). None of these decisions was disapproved or doubted. At the same time, examples were provided of a number of cases in which, although the arrangements for the flights themselves proceeded normally, and no fault could be found with the aircraft, an accident within Article 17 had occurred. These included terrorist attacks (Evangelinos v TransWorld Airlines Inc 550 F 2d 152 [CA3 1977], and Krystal v British Overseas Airways Corporation 403S Supp 1322 [CD Cal 1975]), and the activities of a drunken passenger who fell and injured another passenger (Oliver v Scandinavian Airline System 17 CCH Av Cas 18, 283 [Md 1983]). Again, no disapproval was expressed of these decisions. And, in a clear distinction involving Saks itself, attention was drawn to Weintraub v Capital International Airways Inc 16 CCH 8v Cas 18, 058 [NY Supp Ct 1st Dept, 1981] where it was contended that a "sudden dive" had caused pressure changes which resulted in hearing loss

It is not my intention to suggest or imply that it is normally helpful to embark on a lengthy recital of different decisions. My purpose is to demonstrate that the Supreme Court in Saks did not simply identify a principled approach to the meaning of "accident" within Article 17 but that the Opinion itself was illuminated by practical examples of individual cases which did, and did not, fall within the principle.

In this jurisdiction we accept an obligation to seek a uniform or consistent construction of the Warsaw Convention with that adopted by courts in the countries of other signatories to the Convention (Sidhu v British Airways [1997] AC 430, and Morris v KLM [2002] 2 AC 628). As it seems to me, however, we need to be a little cautious not to assume that every decision from every other jurisdiction, at whatever judicial level in the country concerned, carries identical weight. In the context of Article 17, we must respectfully recognise that the pre-eminence of the judgment of the Supreme Court of the United States in Saks has been universally acknowledged, and we should apply it.

The problem that has arisen is that the explanation by the Supreme Court of the meaning of "accident" has sometimes directed attention away from the single word actually used in Article 17, "accident". The Article has not been re-written, at greater length, incorporating the explanation provided by the Supreme Court. There is no new text. So while we must bear in mind that the decision in Saks explains, and limits, the meaning of the single word "accident", and the circumstances in which liability can arise under Article 17, neither the word nor the Article are amended. Inevitably, the forensic endeavour to bring an individual case within the ambit of Article 17 has led to increasingly refined and subtle arguments which, speaking generally, focus on the broader words of explanation used in the judgment in Saks rather than the words of the Article themselves, and their inapplicability to an injury resulting from the passenger's own "internal reaction" to what is otherwise a normal, untroubled flight.

In the final analysis, in this jurisdiction at any rate, this court is bound by the decision and reasoning in Saks (see, for example, the summary by Lord Hope of Craighead of the of the domestic authorities in Morris v KLM, at paragraph 71). It will also be bound by the Master of the Rolls' analysis of the remaining authorities, with which I expressly agree. Unless cases emerge from other jurisdictions at the equivalent level of the Supreme Court in the United States of America, or the House of Lords here, it is difficult to envisage very many cases involving this aspect of Article 17 in which an expansive disquisition about the authorities would hereafter either be necessary or appropriate, whether from counsel in argument, or for that matter in the judgment of the court.

In my judgment, in agreement with the Master of the Rolls, and Nelson J below, on the basis of the agreed specimen matrix of facts, an "accident" for the purposes of Article 17 is not disclosed.

Accordingly, I agree that the appeal should be dismissed.

LORD JUSTICE KAY: I have little doubt that when the Warsaw Convention was framed, those responsible for the language of Article 17 thought that "accident" was a relatively straightforward concept that would be readily understood. The extensive authorities to which we have been referred have served to demonstrate that such thinking was misplaced. However, in my judgment, the court cannot lose sight of the fact that the foundation for liability for personal injury that was chosen and agreed was that there had been an "accident".

Inevitably, for the reasons explained by Justice O'Connor in Air France v Saks 470 US 392, circumstances would arise which lay close to the borderline between what was an accident and what was not an accident. To resolve such issues, the Supreme Court gave the helpful guidance to which reference has already been made by my Lords. However, I have no doubt that it was never the intention underlying that guidance to substitute some other form of words for those used in Article 17 of the Convention so as effectively to bring within the Article circumstances that could not on any sensible view be seen as involving an accident on any recognised meaning of that word. Otherwise there would be a failure to give effect to the agreement that resulted in the Convention.

If one asks oneself the simple question was there, on any recognised meaning of the word, an accident in circumstances where a person suffered a deep vein thrombosis merely because of the effect of a flight on an aeroplane without there being any triggering event, the answer in my judgment is quite simply no.

If the arguments that have been advanced by the appellants in this case were correct and the intention was to provide for liability for the consequences of air travel of the kind with which this litigation is concerned, it seems to me inevitable that the language of Article 17 would have been different and the word "accident" would never have been used or certainly would never have been used without qualification or without an alternative being expressed.

Some of the decisions to which we have been referred suggest that courts have taken as their starting point other decisions on the application of the Article rather than, as I believe is inevitably correct, an initial consideration of the language of the Article itself. The result is a gradual journey so far away from the source that the origins can no longer be clearly seen. Those failings are, I believe, not to be found in the judgments of either the Master of the Rolls or Judge LJ, or that of Nelson J, and I entirely endorse and agree with both the conclusion that my Lords have reached and the reasons which they have given.

I too would dismiss the appeal.