Smithkline Beecham Plc Glaxosmithkline UK Ltd & Ors v Apotex Europe Ltd & Ors

This is an appeal and cross-appeal from a judgment of Lewison J of 26^th July 2005, [2005] EWHC 1655 (Ch); [2005] FSR 44; [2006] 2 All ER 53; [2006] IP & T 307. Mr Nicholas Strauss QC advanced the argument for the appellants (including the Canadian companies mentioned below) and Mr Andrew Waugh QC that of the respondents (collectively “GSK”).

GSK were between them the owners and exclusive licensees of Patent No. 2,297,550. This covered a process for making a particular form of paroxetine hydrochloride. GSK sold paroxetine hydrochloride as a pharmaceutical under the trade mark Seroxat on a very large scale. Its principal ultimate customer was the Department of Health.

At various times from November 2001 to the autumn of 2002 the “Defendants” as I shall collectively call Apotex Europe Ltd, Neolab Ltd. and Waymade Healthcare Ltd. each evinced an intention to sell paroxetine hydrochloride in the UK, an intention of which GSK became aware. Everyone knew that the particular product envisaged was to be made in Canada and imported. In correspondence with solicitors for the Defendants GSK contended that the manufacture would involve the use of the patented process and so the importation into and sale in the UK of the product would infringe the patent pursuant to s.60(1)(b) of the Patents Act 1977.

The Defendants first began litigation by starting, on October 9^th 2002, an action for revocation of the patent. GSK responded immediately with an infringement action within which they applied for an interim injunction. In due course the actions proceeded together and were heard together, essentially as an infringement action with defences of non-infringement and invalidity. The action for revocation in effect became a counterclaim.

The only parties to the two actions were GSK and the Defendants. The latter are all English companies. The companies who were to manufacture in Canada, namely Apotex Pharmachem Inc. (formerly Brantford Chemicals Inc. (who made the bulk chemical) and Apotex Inc Inc. (the formulators), were not parties. I call them collectively “the Canadian companies”.

Until determination of the application for an interim injunction, from October 29^th 2002 until November 28^th 2002, the Defendants undertook to the court not in the UK to deal in any pharmaceutical preparation containing paroxetine hydrochloride. There was no prohibition on mere importation or keeping. In return GSK gave a cross-undertaking, the precise form of which was in substance the same as that which was to replace it save that it referred to the undertaking rather than the injunction

On 28^th November, when a first instance judge, I granted an interim injunction substantially in the same terms as the Defendants’ prior undertaking. GSK gave the following cross-undertaking:

On 8^th December 2003 Pumfrey J held the patent not infringed and invalid. The interim injunction was continued briefly pending a hearing on 18^th December to determine the final form of order. There was a cross-undertaking in damages given by GSK. Although differently worded yet again, it was in substance the same as that which had run until judgment.

Two days before that hearing, GSK told the Defendants they would not ask for an interim injunction pending appeal. The day before the hearing the Canadian companies applied to be joined as defendants in the action. Although the witness statement in support did not say so, the real objective behind this was to obtain a ruling that such joinder would have retrospective effect so that they would be entitled to claim on the cross-undertaking.

The hearing was on 18^th December 2003. Pumfrey J ordered an inquiry as to damages on the cross-undertaking, ordered points of claim by the “Apotex parties” (i.e. the Defendants) and subsequent pleadings. However he also stayed the inquiry pending determination of the application to join the Canadian companies.

On appeal the patent was held valid but not infringed [2004] EWCA Civ 1568; [2005] FSR 524. So the Court of Appeal left the relevant part of his order intact. The cross-undertaking remained enforceable.

Procedurally things then rather went awry, neither side complying with the order of Pumfrey J or the Rules of Court. Notwithstanding the stay of the inquiry, draft points of claim were served on GSK. These named both the Defendants and the Canadian companies as “claimants on the cross-undertaking”. The draft included claims by the Canadian companies in their own right as well as claims by the Defendants not only for themselves but for the Canadian companies. When GSK’s solicitors pointed out the draft included claims by the Canadian companies, the solicitors (by then) both for the Defendants and the Canadian companies, said in a letter of 6^th April:

GSK never formally responded to the invitation made in the second paragraph. What it did was to apply for an order for determination of preliminary points, namely:

whether [the Canadian companies or either or them] can recover any sums on the enquiry on the cross-undertaking ordered by Pumfrey J on 18^th December 2003 (including any sum by way of restitution) (the basis upon which recovery is claimed having been identified in paragraphs 11-13, 16-17, 32-36, 39-40, 46-49 and 50-56 of the Points of Claim dated 11^th March 2005), and

whether [Apotex Europe] can recover any sums in respect of any loss suffered by [the Canadian companies or either of them] by reason of the interim injunctions granted in these proceedings (as referred to in paragraph 38 of the Points of Claim dated 11^th March 2005), and

whether Neolab, Apotex Europe or Apotex Inc can recover any sums in respect of any payments which would have been payable under clause 2.6(1) of the Neolab Supply Agreement (as Pleaded in paragraphs 41 to 45 of the Points of Claim dated 11^th March 2005)”

Quite why GSK went about it this way was never explained. After all neither the Defendants nor the Canadian companies even had permission to serve a pleading. No doubt the general idea was to get the other claims determined at the same time as the Canadian companies’ application to be joined under the slip rule. But, for that, joinder of the Canadian companies was necessary for reasons other than the slip rule – they were advancing a positive case as claimants. What should have happened is that the stay be lifted, the draft pleading be converted to a served pleading with a concurrent application to join the Canadian companies both on the basis of the slip rule application and on the basis that they were claimants in their own right. Then GSK’s application would have been on a proper procedural basis.

What actually happened was that Pumfrey J refused to order the preliminary points. Instead he ordered that:

So when the matter came before Lewison J it was in an odd state. He described it this way:

There were five points:

Whether the cross-undertakings could or should be amended under the "slip" rule. The point here was the form of cross-undertakings was not in the form set out in the Practice Direction to Part 25, PDA para.5.1 which says that:

Whether the claim that GSK was estopped from denying that the Canadian companies were entitled to claim their own losses under the cross-undertakings should be struck out;

Whether the court should strike out the Canadian companies’ claim that they are entitled to recover in restitution the losses they have sustained.

Whether the claim that Apotex Europe was entitled to claim the losses of the Canadian companies as "third party" losses under the cross-undertaking.

Whether the Canadian companies could or should be joined as defendants with retrospective effect.

Lewison J ruled against the Defendants and the Canadian companies on each of these points save for (ii), estoppel. That he held was arguable and had to go to trial. It was accepted before him that if the Canadian parties could not be joined as defendants with retrospective effect, the application for their joinder would not be pressed. When he so ruled the Canadian companies ceased to be parties in the proceedings, either as defendants or as counterclaimants or as respondents to GSK’s application. Lewison J however did give them permission to appeal the joinder and slip rule applications along with the “declaration” which he granted, namely:

Lewison J went on to order that fresh Points of Claim be served. These were to be confined to the only issue which he found arguable, namely the claim based on estoppel. Such a document was served containing much the same material as had been in the draft but with an allegation of estoppel by representation. Its viability is the subject of GSK’s cross appeal.

The Canadian companies decided not to appeal Lewison J’s slip rule decision nor his decision that they could not be joined with retrospective effect. So when the appeal came on before us, the Canadian companies were no longer parties at all and were not seeking so to be. We asked Mr Strauss whether nonetheless he was appearing for the Canadian companies. He said he was. We pointed out that he had a problem but nonetheless at that stage allowed the argument to proceed as though they were still in the case.

In due course Mr Strauss offered an undertaking by the Canadian companies to issue such proceedings as were necessary to cover the claims made by them in their own right. I am content to consider the case on the basis that such an application has been made and is opposed on the grounds that it is doomed to fail. I do so not least because it was GSK who embarked on this procedural chaos and in effect brought the Canadian parties in. However I wish to make two observations:

The procedural irregularity was not merely technical – it leads and has led to conceptual muddles, particularly in relation to the claim based on an estoppel. For once matters are seen in a proper procedural state it is possible more clearly to discern that the Canadian companies have to base their claim on the estoppel relied upon.

It is in any event wholly undesirable that in matters as weighty as this, things should not be on a proper footing. Treating or deeming a series of important procedural steps (including one as basic as making a company a party to the proceedings) to have happened is, in principle, unsatisfactory. A.P. Herbert’s Lord Mildew comes to mind: “there is too much of this damned deeming.”

Before us the appellants raise only points (iii) and (iv) namely restitution and third party loss. GSK cross appeal the estoppel point, (ii); the slip rule and joinder decisions are unchallenged

Before turning to the points in issue it is necessary to explore the nature of a cross-undertaking in damages. This is largely uncontroversial and so can be taken shortly. The practice of requiring a cross-undertaking from a plaintiff who sought an interlocutory (now called “interim”) injunction developed in the 19^th century. The reason was that the court at the interlocutory stage did not know who the ultimate winner would be. So if an injunction was granted but the case ultimately failed, the person enjoined would have a remedy.

The legal justification of the practice stems from the discretion of the court to grant or withhold an injunction given by what is now s.37 of the Supreme Court Act 1981 (repeating earlier enactments to the same effect):

A party who is granted an interim injunction but who ultimately loses the full trial is not regarded as a wrongdoer because he got an interim injunction. Sometimes, for convenience and want of a better term, the expression “wrongful injunction” is used, but in truth there is nothing wrongful about it. The decision whether or not to grant it is made on the basis of a necessarily incomplete picture. The decision depends on all the circumstances of the case, generally whether or not damages to an ultimately victorious claimant would be an adequate remedy, whether the claimant can show a serious issue to be tried and so on. There is no need here to recite all the various considerations which may arise – American Cyanamid v Ethicon[1975] AC 296 and the many other cases about grant or refusal of an interim injunction are discussed in all the books. Lord Diplock put it broadly in NWL v Woods[1979] 3 All ER 614 at 624:

But a number of particular considerations do call for mention in the context of this case. First is that the existence of the cross-undertaking in damages is a very material consideration in whether or not an injunction should be granted. The fact that an ultimately unsuccessful claimant will have to compensate the defendant for having “wrongly” stopped his proposed activity is a major factor in assessing the balance of risk.

Second is that the position of third parties or the public who may be affected by the injunction is a matter which the court can take into account in exercising its discretion. This is so in limited circumstances even for a final injunction, see e.g. Miller v Jackson [1977] 1 QB 966 and Chiron v Organon (No.10) [1995] FSR 235.

Next is that third parties unconnected with the dispute (“neutrals”) who may incur expense in complying with the order may specifically be covered by a cross-undertaking. This was first required by Robert Goff J in a freezing order case, Searose v Seatrain [1981] 1 WLR 894. The practice in freezing order cases of requiring from the plaintiff an express undertaking to indemnify any third party affected by the order against all expenses reasonably incurred in complying with the order and all liabilities flowing from such compliance was endorsed by this court in Z v A-Z [1982] 1 QB 558.

In this connection it may be noted that the form of order in the current Practice Direction is not without its difficulties – the trouble is that it uses the words “other party” which could mean “party to the action” or “other person” more generally. The point is discussed by Mr Steven Gee QC in his forthcoming article in Lloyds Maritime and Commercial Law Quarterly[2006] LMCLQ 181 – he was kind enough to send me an advance copy when he learned that I was a member of the court in this case. The Rules Committee would do well to look at this again. The difficulties do not arise in this case because the standard form of order was not used.

But what is clear is that until this case, it has always been assumed that an outsider (as I will call a legal entity which is not a party to the proceedings) does not have a claim as of right against a party who obtained a “wrongful” interim injunction which caused him expense or damage. The only possible exception to that is Lord Denning MR’s statement that a party who gives notice of a freezing injunction to a bank or innocent third party enters into an implied contract whereby he will pay for the costs of that party complying with the order – see Z v A-Z at p.575. The other members of the Court (Eveleigh and Kerr LJJ) do not comment on this point – it is unnecessary to explore it further here for no contract claim is made.

Next the Court has jurisdiction to require a party seeking an interim injunction to give a cross-undertaking for the benefit of third parties affected by the injunction, see Allied Irish Bank v Ashford Hotels[1997] 3 All ER 309. This stems from the discretionary nature of the jurisdiction conferred by s.37 of the 1981 Act.

Finally there are a limited amount of authorities on the principles of assessment of quantum on a cross-undertaking. Rather than consider them at this point it is more convenient to deal with them when dealing with the third party loss claim.

This is founded on the basis that GSK made more money than they would have done if there had been no interim injunction. That must be so – GSK neither lost sales nor were forced to reduce their price by reason of competition. The Department of Health paid them more money, probably a lot more money, than they would have paid if there had been no injunction.

It is at GSK’s “extra” money that the restitution claim is aimed. The Canadian companies say that it was made at their expense and so they should be entitled to a share of it. That share, they say, is to be calculated by reference to the profits they would have made but for the injunction. They submit that the law of restitution will give it to them. They say that the claim for “lost” profits put this way is not a claim in damages, which they accept they cannot make.

It is a far-reaching and novel contention. If right it would follow, for instance, that a very big trick indeed, amounting to a fundamental assumption, was missed in Hoffmann-La Roche. There the Crown sought an interlocutory injunction to restrain the defendants from charging more for their drugs than permitted by a statutory order which the defendants were contending was invalid. The question, debated all the way up to the House of Lords, was whether the Crown should have to give a cross-undertaking in damages as a condition of grant of the injunction. It was held not: that the Crown was essentially seeking to enforce what was prima facie the law and not protecting its proprietary interests. It was accepted, however that the Crown would be richer as a result of the injunction because the NHS would pay less for the drugs. If a claim in restitution because of the “wrongful” injunction lay, then the whole debate about whether the Crown should have to give a cross-undertaking was pointless – the Crown could be compelled by this alternative route to disgorge the benefit it had received by reason of the injunction.

Mr Strauss did not shrink from that. He said that no-one had considered restitution. Lord Diplock said at p.361:

Likewise what I said in R v Medicines Control Agency ex parte Smith and Nephew (Primecrown Intervening) [1999] RPC 705 does not explicitly exclude such a claim:

With that I turn to Mr Strauss’ more detailed argument. He began by reminding us that the law of restitution is still under development, pointing to a number of statements to that effect in Westdeutsche Bank v Islington [1996] AC 669 (per Lord Goff at p.691 and Lord Woolf at p.722), Woolwich Equitable v IRC [1993] AC 70 (per Lord Goff at p. 172 and Lord Browne-Wilkinson at p.197and Kleinwort Benson v Lincoln CC [1999] 1 AC 221 (per Lord Goff at 372 and Lord Lloyd at 393). It is not necessary to set them all out here. Perhaps the best-known of them is that of Lord Goff in Westdeutsche Bank:

Mr Strauss accepted that the facts of these cases are far from the present case – he relied upon them to say that the general approach to restitution is not limited to particular issues about interest, mistake of law and the like. The key, he submitted, is the general approach to restitution which is a response to a “call to justice.”

More specifically Mr Strauss then turned to the principles applicable when a court of appeal reverses a prior judgment, saying that they applied by analogy to the case of a “wrongful” interim injunction. His general assertion was that the court has power to ensure that a litigant who has used its process to gain an advantage and then lost on appeal will be obliged by the court to disgorge that benefit to a party adversely affected. Moreover, he asserted, that power extends to making the litigant disgorge that benefit to a third party who was adversely affected by the “wrongful” order.

I start with that last assertion. Mr Strauss could find no authority to support it. Not even his best case, the Israeli Supreme Court decision in Palimport v Ciba-Geigy (23^rd Dec 1974), goes that far. I do not see how any rational rule of restitution could. After all when an injunction is granted, there are often non-parties who are affected. Some will benefit, others will lose. Here, for instance, others besides GSK will have benefited. They will include those of GSK’s customers who have gained from the higher price which resulted from the injunction, GSK’s chemical suppliers and so on. It would be absurd to say that a restitutionary claim would lie against them. Yet if Mr Strauss were right, I can see no reason why it would not. In particular they cannot be distinguished on the basis that they are not wrongdoers whereas the “wrongful” injunctor was. For the injunctor is not a wrongdoer.

Likewise there are likely to be “losers”. Here, as it happens, the intention of Apotex Europe was to sell the Canadian companies’ product. But the injunction prevented sale by the Defendants whoever made the product. In another case where the injunctee proposed to buy from one or more of a number of manufacturers outside the UK would they all have a right to “restitution”? And would the intended carriers of the foreign manufacturers be able to say they lost trade and so have a claim? Putting it generally, the court cannot put the clock back so as to undo all “winnings” and restore all “losses” caused to all who were affected by the “wrong” injunction.

There is another aspect of the suggested right of non-parties to a claim in restitution. It could only, even on Mr Strauss’ case, apply to the “unjust benefit” gained by the injunctor. That would be a pot limited in quantity. But the losses of non-parties might exceed that pot. What is to be done then? Some sort of apportionment between different claimants? On what principles would that work? And, given that a claim in restitution would have at least a six year limitation period, suppose an early claimant had used up the pot, what would be the position of a late one? All this shows that what is really being claimed under the guise of restitution is damages.

The conclusion must be that there is no possible rational basis for a third party to have a claim in restitution in respect of benefits which accrue to a “wrongful” injunctor.

So it is not necessary to consider whether a party the subject of a “wrongful” injunction or other order has an independent claim in restitution. In practice that is almost invariably an academic question because there will nearly always be a cross-undertaking in place. This is so not only for interim injunctions but in the case of final injunctions. In respect of these the court will normally only refuse a stay pending appeal if the successful injunctor is willing to give a cross-undertaking in damages should the appeal be successful, see the 3M case, Minnesota Mining & Manufacturing Co v Johnson & Johnson [1976] RPC 671

However, because the point has been fully argued, and counsel’s researches have extended far and wide, I think it worth exploring the question of whether there is a general restitutionary power which was simply overlooked in the past.

I start with Rodger v The Comptoir d’Escompte de Paris (1871) LR 3 PC 465. A judgment for a sum of money had been given but reversed on appeal. The money had been paid following judgment. The appeal court had ordered the return of the money and the question was whether there should also be interest on the money. Lord Cairns said, in reporting to Her Majesty:

This reasoning is not based on a general power to put any harm caused by a
“wrongful” order right – it is simply the application of the principal that if A has had the use of B’s property, he must account for the benefit he has had from that use – its “fruits.” The principle goes back at least to Sympson v Juxon (1624) Cro Jac 649. There a first instance judgment had wrongly given possession of land to the plaintiff. Upon successful appeal the defendant had his land restored along with profits made meanwhile:

The same principle was applied in Trustee of the property of F.C Jones (a firm) v Jones[1997] Ch 169. A bankruptcy order was made in 1984 (so under the Bankruptcy Act 1914). Under that Act the trustee in bankruptcy got title to all the assets of the bankrupt as of the date of the act of bankruptcy. So, the trustee owned the partnership assets. The wife drew £11,700 out of those assets and invested in potato futures. By November 1984, she had made £50,000 using that money. The trustee sought to recover that sum and succeeded. It was the fruit of his money. The case is not an example of the sort of wide, generalised, doctrine of restitution needed if the Canadian parties are to succeed here. (Actually, although Millett LJ said that equity had no role to play in that case, for myself I cannot see why – it looks like a plain case of a constructive trusteeship).

Mr Waugh sought to rely on Hillgate House v Expert Clothing [1987] EGLR 65. Some landlords took possession after a successful, at first instance, forfeiture claim. The tenant succeeded on appeal and then brought a claim for the wrong of breach of the covenant for quiet enjoyment, relying on Rodger. The claim failed. Browne-Wilkinson J first held that the lease had been in existence all the time, despite the “wrong” order for forfeiture. But he went on to hold that since the acts complained of were done pursuant to a court order they could not give rise to a cause of action. He said:

As far as other common law courts are concerned, it seems clear that both the US and Australian courts have clearly decided that there is no general principle of restitution available to a party harmed by a “wrong” court order. In the US, Greenwood County v Duke Power (1939) 107 F(2d) 484 clearly decides this. A “wrongful” injunction granted at the behest of a power company had stopped the county from receiving or using Federal funds to build a rival power station. Upon reversal of the decision and dissolution of the injunction the county claimed for the profits made by the power company from the injunction and for its losses. Judge Parker said:

Another US case is United Motor Service v Tropic-aire (1932) 57 F.2d 479 (Court of Appeals for the 8^th Circuit). The defendant said he had suffered damage in excess of that provided for by a bond put up by way security as a term of the grant of an interim injunction in an action which ultimately failed. Judge Kenyon said:

Mr Strauss sought to get some comfort from Arkadelphia Milling v St Louis Southwestern Rly (1918) 249 US 134 (Supreme Court). A “wrongful” injunction had restrained a State Railroad Commission from enforcing certain shipping tariffs which it had fixed. Two frequent shippers were also enjoined as representative defendants. The injunction was directed against “the Commissioners …, the individual shippers named as defendants, and all other patrons of the road in the shipment of freight.” It was dissolved on appeal. So the injunction had compelled the “patrons of the road” to overpay. The question was whether there could be a restitutionary claim on their behalf as a class. It was held there could. But this was because they were the class that were injuncted. The case does not establish that third parties not subject to the injunction can bring a claim. Nor that a general restitutionary claim by a defendant who is enjoined will lie. The class was allowed to recover that which by reason of the injunction it had overpaid – a true restitution.

Actually if one thinks about it, the reasoning in Arkadelphia Milling is rather against a general restitutionary claim. For if such a claim lay at the behest of anyone adversely affected by an injunction which had resulted in profits to the injunctor, the reasoning would have been quite different. It would have said just that, and had no need to go into the fact that the injunction was a class injunction.

Finally as regards the US position, if a restitutionary claim lay for every non-party who loses money or has to pay more as a result of a “wrongful” injunction, whether interim or final, I would expect there to be masses of claims by way of class action run on a contingency fee basis. But no-one has pointed to any such case.

The point has also received detailed consideration in Australia. The leading case is National Australia Bank v Bond Brewing Holdings [1991] 1 VR 386. Ex parte orders had been granted on the application of the defendant company’s bankers appointing receivers and managers. In due course it was held that was wrongful. No cross-undertaking was given on the making of the order (as it was held should have been). The company sought a remedy against the banks for the “wrongful” orders it had obtained. It was held that there was none.

The Supreme Court of Victoria analysed the problem in detail. Kaye J said at p.576

Brooking J began a particularly learned judgment thus at p.591:

Contrary to Mr Strauss’ submission that the case was concerned only with a claim in damages, Brooking J addressed a restitutionary claim directly at pp.593-4:

After review of further authorities he said (p.597):

As the argument developed before us, it became clear that for restitution to work Mr Strauss needed to show that something had been taken from the Canadian companies and given to GSK by the injunction. He submitted that “what had been taken from” the Canadian companies was “a business opportunity” and that the injunction had in substance transferred that to GSK. So, he submitted, restitution entitled his clients to the fruits of that transferred business opportunity, namely that part of GSK’s “extra” profits which would have been made by Canadian companies but for the injunction.

He only found one case which he could suggest supported that view. This was the Israeli Supreme Court case I mentioned earlier as his “best” case, Palimport. The plaintiff had succeeded at first instance in a patent infringement action and was granted a final injunction. There was no stay of injunction but it was discharged when an appeal was allowed. The defendant wanted recompense for the damage done whilst it was in force. It seems (for there is no reference to it) that the courts of Israel had not adopted the English practice settled in 3M.

The claim was made in restitution, not damages. as was pointed out by Sussman (Acting President):

Of particular importance is the fact that Israel has a special statutory provision about appeals. The Acting President mainly founded his judgment in favour of a restitutionary claim upon this statute. He developed the proposition that no claim in damages lay, referring, inter alia, to the article in the Law Quarterly Review by Gordon ((1958) LQR 517. And went on to say:

With great respect I cannot accept that Rodger justifies a general unjustified enrichment claim. The analysis in the National Australia Bank case shows why. The Acting President also relied upon Arkadelphia Milling. But he overlooked the class nature of the injunction which was what justified the class restitution claim. Nor does he deal with the Greenwood case, though (it appears from the judgment of Kahan J) it was cited.

The Acting President then allowed the restitution claim giving two reasons:

The second reason was:

Kahan J explicitly held that restitution was not limited to infringement of a property right:

The judgments of Cohen J and Kister J are largely based on Rabbinical law which is irrelevant to present considerations.

I ask myself two questions: first is there a real prospect that the Israeli case would be followed here? And second, if so, would it be extended to non-parties to the original action? In asking the first question I am conscious that I may be too favourable to Mr Strauss, for on a strike-out claim it is the duty of the court to apply English law “as it now stands” (per my Lord the Chancellor when he was Vice-Chancellor at [39] in Baird Textiles v Marks & Spencer [2001] EWCA Civ 274, [2002] 1All ER (Comm) 734). Judge LJ was to the same effect at [55]:

I have no doubt that it would not be so followed. Firstly it is clear that the analysis and citation of authority, though extensive, was not as detailed as that in the Australian case. Second for the reasons I have given, I do not think that either the Rodger or Arkadelphia Mining case were correctly analysed. Third the case in part depends on the Israeli Execution Law which is irrelevant here, Fourth I cannot accept that a mere freedom to trade can be fairly be regarded as property or anything in the nature of property and restitution remains based on restoring property or the fruits of property.

In so holding I have taken into account the articles by Professor Friedmann of Tel Aviv University, Columbia Law Review1980 p.504 and that in Essays in the Law of Restitution (1991) edited by Professor Andrew Burrows. Professor Friedmann is enthusiastic about the case (describing it as a “landmark decision” in his contribution to the latter). But he seems to be alone – the principle the case is said to establish has not in practice been a seed which has taken root outside Israel.

Even if that were wrong, I cannot envisage any extension to non-parties for all the reasons I mention above about limitation, multiple claims by different persons and so on. Nor does Professor Friedmann, for all his enthusiasm for Palimport, go that far. If one were to go that far, then I can see no stopping point. Court orders, as I have already observed, can affect third parties beneficially or adversely. Litigation itself can have the same effect. For instance, it is the everyday experience of any litigation intellectual property lawyer that the very existence of litigation against a defendant can have a severe adverse affect on his sales – why buy or put in your catalogue a product which your supplier may be prevented from supplying? It is settled beyond doubt that no action lies, save for very limited exceptions in the nature of abuse of process, for recompense for damage caused by litigation itself (see Metall und Rohstoff v Donaldson Lufkin & Jenrette [1990] 1 QB 391. That is so even in respect of a claim by a party who claims he was harmed by the litigation – so it is inconceivable that a non-party could have such a claim.

What then of Mr Strauss’ submission that the Canadian companies had a “business opportunity” of which they were deprived? He says that, unlike other third parties they were in a special position, having their own process and plant, to exploit the market. I do not think that can be equated in any way with anything in the nature of property, what Professor Friedmann calls “quasi-property” (see p.512 of the Columbia Review article). Cases in which courts have spoken of a “business opportunity” have been quite different – the opportunity has been something in the nature of confidential information, see e.g. CMS Dolphin v Simonet[2001] BCLC 704 at p. 733. Moreover the cases have been about the defendant taking and using the quasi-property for himself – thereby unjustly enriching himself at the plaintiff’s expense. The Canadian companies’ special process, assuming it was special, was in not itself something which was taken from them and given to GSK.

I conclude the Judge was right to strike out the claim. I reach this conclusion without the need to consider whether, as Mr Waugh contended, the Canadian companies, if they had applied to the court, could have had the benefit of the undertaking extended to them. I say this because I am not entirely sure that is so and this is not the place to debate in detail what would have happened if they had so applied.

I elaborate a little. In Primecrown at p.708I said:

The principles about who can successfully intervene and ask for such an undertaking and whether such an intervener can be put on terms as to costs or the like are far from worked out. They may need development on a case-by-case basis. I understand that the current practice of the Patents Court when an application for an interim injunction in respect of a pharmaceutical is sought is to require the patentee to give notice to the Department of Health of the application in case it wishes to seek such an undertaking. To date there has been no ruling on the point.

One consideration which may matter is whether the person applying is himself a potential infringer. In many intellectual property cases there is a chain of potential infringers starting with the manufacturer (or importer) running through middlemen to the ultimate consumer. That is so in patent cases unless the ultimate consumer can avail himself of the defence given by s.60(5)(a) of the Act – that his use is for a private and non-commercial purpose. Even keeping a patented product can infringe (s.60(1)(a)). So if the claimant picks on only one member of the chain as a target for an interim injunction, it might be said that he was artificially limiting his liability on the cross-undertaking and the others in the chain should have the benefit of one too, perhaps on terms as to costs if the action succeeds. On the other hand the position of non-infringers, such as third party foreign manufacturers may be different. I say no more for the present, for it was not argued by Mr Strauss that the Canadian companies would not have been granted a cross-undertaking if they had intervened.

I should also mention some other arguments advanced by Mr Strauss. One was by way of appeal to policy considerations. He suggested that a claim in restitution had an advantage over a cross-undertaking regime. This was that a claimant would know where he stands – he cannot be ordered to pay more than the pot of “extra” money he gains by the “wrongful” injunction. By way of contrast, suggested Mr Strauss, if cross-undertakings had to be given not only to actual injunctees but third party interveners who might be adversely affected, liability could be extensive, going well beyond the value of the injunction. That, he suggested, could operate as an undesirable deterrent to a party in need of an interim injunction.

I am not impressed by this. First there are all the difficulties which I have mentioned about multiple claims to a single pot. The so-called “advantage” is fraught with practical difficulties on top of the intellectual difficulties. Second I cannot see that the “deterrent” is necessarily a bad thing. A party who may have to recompense third parties significantly damaged by an interim injunction, is likely to weigh carefully whether he really has a case which he is likely to win at trial, as opposed to one which raises no more than a “serious issue to be tried.”

Another argument was based on the fact that a successful claimant in a patent action is entitled to elect the remedy of an account of profits made by the infringer rather than damages, see s.61(1)(d) and s.61(2) of the Act. Mr Strauss suggested that there ought to be some sort of symmetry – a “wrongly” injuncted defendant ought similarly to be able to claim the profits made by the patentee if his action fails. I do not see why. That might even put the defendant in a better position than if he had never been sued. Moreover the appeal to symmetry works the other way as regards restitutionary claims by third parties. Here, for instance, if there had been no interim injunction but GSK had succeeded at trial, there is no way that the Canadian companies, as non-defendants, could have been required to disgorge their profits.

Mr Strauss also made some appeal to the inherent jurisdiction of the court. But important and powerful though this is (see the article by my late father [1970] CLP 73) it cannot seriously be suggested it goes so far as to create positive legal rights for non-parties. The inherent jurisdiction is essentially procedural – a tool in the administration of justice - not part of substantive law creating free-standing rights for third parties of the kind claimed here.

Lord Diplock, in Hoffmann-La Roche at 361, expressed the view, obiter, that quantification of compensation under a cross-undertaking should be on the basis of a breach of a notional contract between the parties to the effect that the injunctor would not prevent the injunctee from doing the injuncted acts. That brings into play the rule in Hadley v Baxendale (1854) 9 Exch 34: briefly, the “wrongful” injunctor will pay only for such damage as he could reasonably foresee would be caused to the injunctee by the injunction, either from his general knowledge of the circumstances or from his knowledge of the particular circumstances of the injunctee.

In Primecrown I reviewed the rather exiguous authorities on the principle by which compensation is to be assessed and concluded that the notional breach of a notional contract basis may be too narrow in some cases (see p.714). For present purposes that does not matter. What matters is that the entire discussion in the cases is about how the compensation which the injunctor has undertaken to pay the injunctee (or other party identified in the cross-undertaking) is to be assessed. Whether it is the contractual or wider basis, one must not lose sight of what it is that is being quantified.

Founding himself on the basis of Lord Diplock’s notional contract, Mr Strauss argued that a party identified in the cross-undertaking as one whom the injunctor will compensate (here particularly Apotex Europe) can claim compensation not only for having been himself prevented from doing the restrained acts but also for others who were adversely affected by the injunction. He relied particularly on Linden Gardens v Lenesta Sludge [1994] 1 AC 85 and Alfred MacAlpine v Panatown [2001] 1 AC 518.

I think the argument fails for two reasons. First there is in fact no contract between the parties. The undertaking is given to the court – it is to compensate the parties identified in the undertaking. That is all. Second, even if there really had been a contract between the parties, the principles expounded in the cases do not and could not be expanded to cover this kind of third party loss.

The first reason is incapable of much elaboration. Lewison J put it elegantly as follows:

Putting it another way, the argument converts what is only a principle of assessment of compensation for breach of contract into both a notional breach of a notional contract and a method of assessment of breach. The cases relied upon are only about what damage can be claimed following a breach of contract.

Turning to the second point, even if there had been a real breach of a real contract not to prevent the Defendants from selling paroxetine, I do not see how the cases could establish a viable cause of action. Mr Strauss’ argument runs this way. Suppose that GSK and the Defendants had entered into such contract. GSK would have known that if they were to breach it, the Canadian manufacturing companies would be adversely affected. So Apotex Europe can claim for the losses of the Canadian companies in addition to (not by way of substitution for) its own losses by virtue of an exception to privity of contract.

Linden Gardens and Panatown were cases where one party (B) had contracted with (A) to perform certain works to land and was in breach of that contract. In Linden Gardens by the time of breach A had transferred the land to C. In Panatown the land was already owned by C at the time of contract and B had also entered into a direct deed with C whereby it agreed to pay C in respect of [in effect] any breach of its contract with A. In each case B contended that A had suffered no damage (the land being no longer, or never having been, his).

In Linden Gardens the benefit of the contract between A and B was, by its terms, non-assignable. Lord Browne-Wilkinson (with whom Lords Keith, Bridge and Ackner agreed) said:

Lord Griffiths, whilst agreeing with Lord Browne-Wilkinson, favoured a broader basis of liability. This was based on A’s interest in the performance of the contract – he has not received what B contracted to provide and B must pay for that. The liability did not depend on any proprietary right ever having been vested in A. He put it this way at p.97:

The principles were further considered in Panatown. A majority of the House held that A had no claim for substantial damages. Their Lordships did not all reason the same way. But, whether one goes to the majority or minority speeches, I cannot find any basis which would justify a claim by Apotex Europe for the Canadian companies’ losses.

Specifically the broadest approach, based simply on a “performance interest” which gives a contracting party an entitlement to sue the other contracting party for substantial damages suffered by a contemplated third party who is out of pocket by reason of the breach, could not apply here. Apotex Europe’s performance interest was in itself not being restrained from selling paroxetine hydrochloride. It had no other “performance interest.” There is no legal “black hole” of the sort contemplated in Linden Gardens.

Putting it another way, none of the speeches in either case begin to regard, as within the concept of “performance interest,” anyone who might be adversely affected by a breach of contract. Here for example besides the Canadian companies, others have been affected by the injunction – employees or even potential employees of the Canadian companies, shareholders of the Canadian companies, suppliers of chemicals which would have been used to make the product and carriers.

Mr Strauss seeks to put those parties on one side, arguing that the notional contract with Apotex Europe was not for their benefit whereas it was for the benefit of the Canadian companies. He says this is because the Canadian companies “are most closely associated with Apotex Europe”. He frankly adds:

I do not accept that the principles of law on either topic are artificial. What is artificial is this elaborate combination and extrapolation of those principles. I do not think the third party loss claim is viable. The Judge was right to so hold.

I turn to GSK’s cross appeal from the Judge’s refusal to strike out the claim by the Canadian companies based on estoppel, although he had “considerable sympathy” with Mr Waugh’s submissions. His principal reason for taking this course was that he considered that going into the detailed documents relied upon would amount to conducting a “mini-trial” which is not permitted by current practice under CPR Part 24.

Following this decision the Judge required the Canadian parties to serve a pleading setting out their estoppel case. It was essentially the same as the relevant parts of the earlier draft Amended Points of Claim. However I shall refer to the document as actually served – and consider this appeal as an appeal from a refusal to strike the document out. This has the merit of being procedurally orthodox. More than that it most clearly brings out that the estoppel relied upon is essentially being used to justify the claim – as a sword and not a shield to use the classic phrase.

There is no dispute as to the approach of the court at this stage. The judge summarised it thus:

The impugned pleading has 5 claimants, the Defendants and the two Canadian companies. They are collectively described as “Claimants on the Cross-Undertaking.” The key parts of the impugned pleading read:

In support of its claim to compensation, Apotex Inc, will contend that SKB and GSK UK are estopped from denying that it is entitled to claim the benefit of the cross-undertaking as to damages contained in the Injunctions, and will further contend that the inquiry should proceed on the assumption that it is so entitled. In support of this contention, Apotex Inc. will rely on the following facts and matters.

At the time the Injunctions were ordered and throughout the period during which they were in force, there was a common assumption on the part of the Claimants on the Cross-Undertaking (including Apotex Inc.) and the Respondents on the Cross-Undertaking (i) that the Injunctions and the cross-undertakings as to damages contained in them applied to the manufacturer and/or intended supplier of the paroxetine hydrocholoride anhydrate proposed to be sold by Neolab and Waymade in the UK and/or (ii) that Apotex Inc. was to be treated as if it were a defendant to the action, without formal joinder.

This common assumption is evidenced by or to be inferred from (amongst other things) (i) the correspondence between the parties’ solicitors both before and during the proceedings, (ii) the parties’ applications, disclosure, witness statements and skeleton arguments, and (iii) the judgments and orders in the Court in the proceedings, as further described in Schedule 9 to these Points of Claim.

The making of that assumption on SKB and GSK UK’s part is further evidenced by or to be inferred from a Glaxo-SmithKline press release dated 25 June 2003 in which it was stated that:

Further or alternatively, by the correspondence, applications, witness statements and skeleton arguments written or produced on SKB and GSK UK’s behalf as particularised in Schedule 9 to these Points of Claim, SKB and GSK UK represented to Apotex Inc., and thereby led Apotex Inc. to believe, that it was being treated as a defendant to the action.

The parties dealt with each other (and, for the avoidance of doubt, communications passed between them) during the action on the basis of that assumption, and Apotex Inc. conducted itself during the action in reliance upon the said representation, as is apparent from the materials identified in sub-paragraph (b) above.

Further, because of the said assumption and/or representation, Apotex Inc. did not apply to be joined as a defendant to the action and did not apply for an express cross-undertaking as to damages in respect of any loss which it might suffer as a result of the Injunctions.

In the premises, it will inequitable to allow SKB and GSKUK to renege on the said common assumption and/or representation, and to deny that their cross-undertaking as to damages covers the loss caused by the Injunctions to Apotex Inc. as the manufacturer and/or supplier of the product and/or as a defendant to the action.

In the premises, SKB and GSK UK are estopped from denying that which was at all material times commonly assumed and/or represented to Apotex Inc. whilst the Injunctions were in force, and the Court should therefore proceed with this inquiry as if the assumption and/or representation were correct, that is, as if Apotex Inc. had the benefit of SKB and GSK UK’s cross-undertaking as to damages and/or were a defendant to the action.”

Two alternative kinds of estoppel are raised, estoppel by convention (where the parties act on a common, erroneous, assumption) and estoppel by representation (where one party has represented, erroneously, a state of affairs to be so and the other has relied upon that representation to his detriment). It was not disputed that for either kind of estoppel, the representation or common assumption as the case may be, must be unambiguous and unequivocal. That is inherent in the very nature of an estoppel.

Before us the argument covered more points than were raised or dealt with by the Judge. Mr Waugh contends that the pleading is not sustainable, even to the Part 24 standard, for three distinct reasons:

An inter-partes estoppel cannot operate so as to expand or contract the effect of a court order.

An estoppel cannot be used as a key element of a claim (sword not shield) and particularly it cannot operate to create a legal relationship when there was none at the outset.

In any event the material pleaded is an inadequate basis to justify an estoppel.

The interim injunction was directed at the Defendants. The undertaking was given to the court, not the Defendants. The undertaking was, if the court so ordered, to pay compensation to the Defendants, no-one else. I think it inexorably follows that whatever passed between the parties cannot affect the scope of what was actually undertaken, or, for that matter, the scope of the injunction. After all the court is not alleged to be party to the estoppel – to be estopped from reading its order to mean what it says.

Suppose for instance, that one or other of the Canadian companies had themselves sold paroxetine hydrochloride in this country without any involvement of any of the Defendants. Could SKB have invoked contempt proceedings on the basis that by estoppel, the Canadian parties were, or were “to be treated as” actual defendants? The answer is obviously no – the court order did not forbid the Canadian companies from selling direct in the UK. No doubt if they had done so or threatened so to do, GSK would have applied for an interim injunction against them. But the basis of the application would have been actual or threatened infringement, not any estoppel.

Mr Strauss recognised the force of that. He accepted that the Canadian companies were not in fact parties to the order and an estoppel could not make them so automatically. There could be no direct enforcement of the order by contempt proceedings against the Canadian companies. He took a more elusive position: that the Canadian companies “were to be treated as though they were parties to the order.” There are echoes of Lord Mildew’s aphorism here. Mr Strauss submitted that if the Canadian companies had themselves infringed, GSK could apply to the Court to have them added as defendants in the existing action on the basis of an estoppel – as if there were a binding agreement that they should be treated as parties. The court would then formally make them so. Similarly, submitted Mr Strauss, by reason of the estoppel, the Canadian companies were entitled to apply to the court to be made defendants entitled to the benefit of the cross-undertaking.

I do not accept that analysis. An estoppel cannot create an agreement. No doubt if there were an actual binding agreement by which the parties agreed that the Canadian companies should be defendants for all purposes, not only for the future but retrospectively, the Court would give effect to it. But that would be by reason of the express agreement – an agreement that the existing order should be varied. And even then there is no way, supposing there had been past infringement by the newly added party, that that party could be a contemnor by reason of those past acts. Even an express agreement cannot change the meaning of the order. All an express agreement can do is lead the court to varying its order with effect for the future.

So I just cannot see how the meaning of the court’s order can be varied by an estoppel.

GSK’s duty was to the court: to obey any order of the court that it should pay compensation to the Defendants. There was no pre-existing rule of law such as tort or contract by which GSK came under a direct duty to the defendants or had a direct liability to the Defendants to pay compensation. And there was never any legal relationship between GSK and the Canadian companies. So the proposed estoppel, said to operate between GSK and the Canadian companies would have the effect of creating a legal relationship between parties who had never, on the matter pleaded, even communicated with each other. That would be an astonishing result – and not one which I think the law could countenance.

It is as well to remember here the fundamental nature of an estoppel. An estopped party is precluded from asserting that a particular fact or set of facts or state of affairs is so. Thinking of it in procedural terms, an estoppel inherently must be raised by way of a riposte. The claimant pleads his case basing himself on alleged facts x, y, and z. The defendant then raises an estoppel saying: “you may not assert these facts by reason of an estoppel arising from representation, convention or whatever.”

The pleaded case here does not fit that essential nature, as can be tested by the procedural model. If the Canadian companies applied to the court to enforce the existing undertakings without asserting the estoppel in their claim, the court would simply dismiss the application as bad on its face. There would be no need for GSK to assert that which they are alleged to be estopped from contending, namely that the Canadian companies are not to be treated as parties to the order. Even if GSK did not appear to argue the point, the claim would fail without the estoppel.

Thus it is essential to the Canadian companies’ case that the estoppel be pleaded as part of the claim. The pleading is not a mere anticipation of a reply to a defence. This shows that what is relied upon is not an estoppel at all – it is a naked attempt to create a legally binding agreement when there never was one and never any intention to create one. That an estoppel cannot do, see for instance by way of a recent example, Baird Textiles.

It was on this ground that the Canadian companies succeeded. Mr Waugh contended that there was no question of a mini-trial: that all one had to do was to go through the documents pleaded: whether considered individually or collectively, none of them amount to an unequivocal representation, or demonstrate an unequivocal common assumption, that the parties treated the Canadian companies as though they were parties to the proceedings. This exercise did not involve any attempt to resolve any conflict of evidence – no “mini-trial”. So it was fit for determination under Part. 24.

I think Mr Waugh was right. Going through the documents to see whether they are capable of supporting an estoppel is not a mini-trial exercise – no resolution of disputed facts is called for. None of them in my view were so capable. It is only necessary to discuss two – those chosen by Mr Strauss as his best.

The first of these was an application notice issued by GSK’s solicitors on 11^th December 2002 – after the interim injunction had been granted. It sought procedural orders against what it called “Apotex,” for instance an order that “Apotex does … provide inspection of [the manufacturing process].” So, it is submitted, GSK were treating the Canadian companies (whom they knew were actually the manufacturers) as parties to the action. But the notice is directed at the existing Defendants only – for it is only they who are identified under the heading “Parties to be served.” At no point had the solicitors for the Defendants come on the record for the Canadian companies. In my opinion it is far too far-fetched to spell out of this an unequivocal mutual assumption, or representation by GSK, that the Canadian companies were to be treated as parties to the action as a whole, still less as parties to the cross-undertaking which had been given earlier. As a matter of convenience, and procedurally rather sloppily, the application was made against “Apotex”. In fact if the Canadian companies refused to abide by the order I do not think they would have broken it.

The other “best” document was the court order made on that application – an order made, after discussion about some terms immaterial to the present point, essentially by consent. The headings name the three Defendants and no others. The recital includes the words “Upon hearing counsel for the Claimants [GSK] and for the Defendants [“Apotex”]. So “Apotex” is defined to mean the actual defendants and none other. Nonetheless the order includes commands which could only be complied with if the Canadian companies co-operated. For instance paragraph 1 provides that “Apotex does …. provide inspection of the process.”

I cannot see how this document unequivocally represents or treats the Canadian companies as parties. If inspection was not provided (because the Canadian companies did not co-operate) that might have consequences for the actual defendants, but it was for them to comply with the order not the Canadian companies. The document is incapable of supporting the pleaded estoppels. So also for all the other documents pleaded in Schedule 9 to the Points of Claim and those set out in it.

I would add two further points. First it is not just that the pleaded documents do not support the estoppel. There is a letter emanating from the Defendants’ solicitors which positively contradicts it. The letter, of 4^th March 2003 was in response to a query about verification of the list of documents. It said:

The other point is in relation to sub-paragraph (f) of the pleading – which alleged that by reason of the common assumption and/or representation Apotex Inc did not apply for an express cross-undertaking. That is inconsistent with the statement on behalf of the Canadian companies that the first time they considered applying to become defendants in the action was following the handing down of judgment on 8^th December.

So I think the Judge was wrong to dismiss the strike out application on the basis that consideration of the point would involve a mini-trial. As the matter developed before us, Mr Strauss recognised that his case based on the pleaded documents might fail. So he retreated to a fall-back position. This was that even if all the specific documents pleaded did not support either kind of estoppel relied upon, nonetheless the pleading, stripped of all reference to those documents, still contained general averments of a common assumption or a representation by GSK to the effect that the Canadian companies were to be treated as parties and that was enough.

I do not accept this. The detailed particulars are put forward as justifying the general averments. If they do not justify the general averments (and I do not think they do) then the averments fall with them. It is no good saying: “Well there may be other material which we have or which may turn up on disclosure”. That does not demonstrate that a plea “has a real prospect of succeeding.”

Likewise I do not accept Mr Strauss’ reliance on the careful pleader’s precautionary “(amongst other things)” in para. 34(b). He submitted that could let in other things. I do not agree. One is entitled to assume with a modern pleading that a party is putting his best foot forward. If that best foot will not support the case, then this sort of reservation will not save it.

Mr Strauss suggested that Pumfrey J had restricted the nature of the Part 24 exercise when he refused to order a preliminary point and ordered instead that the matter be decided “according to the criteria under CPR Parts 3.4 and 24.” This is because Pumfrey J said in giving judgment: “I am quite satisfied that the purely legal argument that will have to be advanced on the basis that all the facts pleaded are true …...” The submission was that thereby the court was required to accept the general averments, stripped of their supporting particulars. I do not accept it. All that Pumfrey J was saying, and ordered, was that the usual criteria for Part 24 applied – there was to be no trial of a preliminary point and (because Part. 24 does not allow it) no mini-trial of any disputed point of fact. Reaching the conclusion I do involves neither of these.

I would add this. We had submissions on whether, as a matter of law, for estoppel by convention, a common understanding must actually be communicated by one party to the other. On one view Kerr LJ might be understood to have suggested this in The August Leonhardt [1985] 2 Lloyd’s Rep. 28 at p. 34:

In the result I would dismiss the appeals, allow the cross-appeal and refuse the application of the Canadian companies to advance claims on their own behalf.

I agree.

I also agree.