//
© 2026 Zoe. All rights reserved.
Zoe is a legal information platform. Always consult the Federal Register of Legislation for authoritative text.
[2018] FCAFC 7
respondent. The appellants pay 90% of the respondents' costs of and incidental to the appeal including in relation to the notice of contention.
1 cited instrument linked from this judgment.
After losing the main patent appeal, Idenix argued it should only pay 70% of Gilead's costs because Gilead had also lost on its notice of contention arguments. The Full Court accepted that the notice of contention took some time but decided the bulk of the appeal's enormous complexity and cost had been created by Idenix's attempt to re-run the entire trial on factual findings. The notice of contention was essentially a back-up position. A 30% cut was too large; instead the Court ordered Idenix to pay 90% of Gilead's costs.
AI-generated legal information, not legal advice. Zoe can make mistakes. Check the cited source, and for advice about your situation consult a qualified Australian lawyer.
2,196 words · generated 24/04/2026
This costs judgment is the sequel to the substantive decision in Idenix Pharmaceuticals LLC v Gilead Sciences Pty Ltd [2017] FCAFC 196. In the primary proceeding Gilead had succeeded in obtaining orders that Idenix's Australian patent relating to nucleoside analogues for the treatment of hepatitis C was invalid. Idenix appealed to the Full Court on multiple grounds, mounting what the Court later described as an "overly elaborate" challenge directed at disturbing numerous factual findings made by the primary judge and in substance seeking a re-trial of complex factual questions.
Gilead filed a notice of contention seeking to uphold the invalidity findings on additional or alternative bases. Those grounds included contentions that claims 7 to 41 of the patent (a) did not derive priority from the 949 application, (b) were not fairly based on the 949 application or its incorporation by reference, (c) were not fairly based on the specification itself because the embodiments on page 100 did not provide a real and reasonably clear disclosure of the compounds of claim 7, and (d) did not derive priority from the 350 application for reasons additional to those given by the primary judge.
The Full Court (Nicholas, Beach and Burley JJ) heard the appeal over three days. It dismissed Idenix's appeal in its entirety and also rejected each ground in Gilead's notice of contention. At [327] of the principal reasons the Court noted that, prima facie, Gilead would be entitled to costs but that its lack of success on the notice of contention might affect that position. Directions were then made for the parties to file short written submissions on costs. Those submissions were filed by 21 December 2017 and the present judgment was delivered on the papers on 2 February 2018.
Idenix contended that a 30% reduction in the costs otherwise payable to Gilead was appropriate. It pointed to the discrete nature of the notice of contention grounds, the 25 pages of separate written submissions devoted to them (approximately 38% of total submissions), the 65 paragraphs of the appeal reasons that addressed those issues (approximately 33% after excluding background material), and the two hours of hearing time occupied by the contention. Gilead responded that the notice of contention had occupied only a limited proportion of the three-day hearing, had not occasioned significant additional costs, was not raised unreasonably, and was in any event a legitimate response to Idenix's own expansive approach. It relied on the principle that a successful party with a single cause of action should not be penalised for advancing alternative arguments that might have succeeded.
The Court accepted that there was "some force" in Idenix's submissions that time had been spent on the notice of contention points, but concluded that the length and complexity of the appeal had been generated principally by Idenix. The notice of contention was, in reality, reactive. A 30% reduction was not warranted. Instead a "small reduction of 10%" was ordered, so that Idenix must pay 90% of Gilead's costs of and incidental to the appeal, including costs relating to the notice of contention.
The Court began from the orthodox position that s 43 of the Federal Court of Australia Act 1976 (Cth) confers a wide discretion that must be exercised judicially. The ordinary rule is that costs follow the event, but that rule is not absolute. Where a successful party has enjoyed only mixed success on discrete issues, the Court may award less than all of that party's costs or may apportion costs by reference to the time taken, the degrees of difficulty, and overall fairness. The judgment cites Firebird Global Master Fund II Ltd v Republic of Nauru (No 2) (2015) 327 ALR 192 at [6] for the proposition that, if success on the event is contestable when separate issues are examined, an issue-by-issue analysis may be required.
The Court then evaluated the particular circumstances. It acknowledged that the notice of contention had raised discrete matters—priority and fair basis questions not raised by the notice of appeal—and that those matters had consumed measurable resources: two hours of oral argument, 25 pages of dedicated submissions, and 65 paragraphs of reasons. Nevertheless three countervailing considerations persuaded the Court that only a modest 10% reduction was fair. First, the notice of contention was reactive; it responded to the breadth of Idenix's attack rather than introducing extraneous controversy. Second, the overwhelming bulk of the appeal's length and complexity was generated by Idenix's "overly elaborate" forensic choices, including its attempt to re-agitate a multitude of factual findings in a manner that effectively invited the Full Court to conduct a re-trial. Third, the Court was not persuaded that the notice of contention had occasioned costs that were disproportionate to the overall scale of the three-day patent appeal.
The Court expressly distinguished Sandvik Intellectual Property AB v Quarry Mining & Construction Equipment Pty Ltd (No 2) [2017] FCAFC 158, in which no costs adjustment had been made because the notice of contention had occupied only limited hearing time and was inferred not to have generated significant costs. While the present case involved more substantial argument, the reactive character of the contention and Idenix's responsibility for the overall scale of the dispute meant that a 30% reduction would over-compensate. The 10% figure was therefore selected as better reflecting "the relative significance of the notice of contention points in relation to the hearing time and the like". The result was an order that Idenix pay 90% of Gilead's costs, including those referable to the notice of contention.
Prior to this judgment the applicable principles were already well settled. Ruddock v Vardarlis (No 2) (2001) 115 FCR 229 had emphasised that the costs discretion, while broad, is not unfettered and must be exercised judicially; departure from the usual rule that costs follow the event is justified only where fairness or justice requires it. Firebird Global Master Fund II Ltd v Republic of Nauru (No 2) (2015) 327 ALR 192 at [6] had clarified that where success on the overall event is contestable once separate issues are isolated, courts may undertake an issue-by-issue analysis, taking account of the degrees of difficulty of issues, time taken to argue them, and like matters. Les Laboratoires Servier v Apotex Pty Ltd (2016) 247 FCR 61 at [297]-[303] had reiterated that a successful party may recover less than its costs or suffer apportionment where it has failed on discrete issues. Hockey v Fairfax Media Publications Pty Ltd (No 2) (2015) 237 FCR 127 at [88] had noted a modern judicial willingness to apportion costs in recognition of the increasing factual and legal complexity of contemporary litigation and the tendency of parties to run multiple claims involving different factual enquiries in the one proceeding.
This judgment did not change those principles. It applied them to the specific context of a patent appeal in which a notice of contention had been advanced and rejected. The decision confirms that, even in intellectually dense patent litigation, the Court will not engage in minute "undue analysis" of every unsuccessful argument but will instead take a holistic view that weighs the reactive nature of the contention, the party primarily responsible for the scale of the dispute, and the proportionate significance of the unsuccessful points. The 10% reduction, rather than the 30% sought, illustrates that the discretion is exercised with an eye to practicality and to avoiding disincentives for parties to canvas all legitimate alternative bases that might uphold a judgment.
Paragraph 3 states: "The ordinary rule is that costs follow the event, although a successful party may be awarded less than its costs, or costs may be apportioned, based upon success on the issues." In plain English, winning overall does not automatically guarantee recovery of every dollar spent; if you lose on distinct, severable points that took real time and effort, the Court can trim the costs award or divide the costs issue by issue.
At paragraph 8 the Court records Idenix's submission, drawn from Ruddock v Vardarlis (No 2) and Firebird, that departure from the general rule is justified where circumstances make it reasonable or fair, or where justice requires it, and that an issue-by-issue approach may be warranted when success on the event is contestable. This passage simply reminds practitioners that the "costs follow the event" starting point is not a straitjacket.
The dispositive reasoning appears at paragraphs 15 and 16. Paragraph 15 acknowledges "some force in Idenix's submissions that some time was spent dealing with the notice of contention points". Paragraph 16 then concludes: "In all the circumstances, we consider that there should be a small reduction of 10% in the costs we would otherwise order in Gilead's favour. We do not consider that a 30% reduction is warranted, notwithstanding Idenix's arithmetic concerning written submissions. The length and complexity of the appeal was generated principally by Idenix, with the notice of contention in reality reactive thereto. We consider that a 10% reduction more fairly represents the relative significance of the notice of contention points in relation to the hearing time and the like."
Translated, the Court is saying: yes, the notice of contention took some ink and some court time, but the real reason this appeal was so long and expensive was Idenix's own forensic choices. Gilead's extra arguments were a sensible back-up position, not the main driver. Therefore a modest 10% discount is fair; a 30% cut would give Idenix an unjustified windfall.
This decision is likely to be invoked in any appeal (particularly but not exclusively in intellectual property) where the respondent has succeeded on the ultimate disposition but has failed on one or more grounds contained in a notice of contention. Trigger facts include: (1) the notice of contention raises discrete legal or factual questions that are not co-extensive with the appellant's grounds; (2) measurable but not dominant resources are devoted to those grounds (for example, two hours out of three hearing days, roughly one-third of written submissions, and roughly one-third of the reasons); (3) the notice of contention can fairly be characterised as reactive to the breadth or intensity of the appellant's attack; and (4) the overall scale and cost of the appeal have been driven primarily by the appellant's litigation choices, such as an attempt to re-litigate numerous factual findings at appellate level.
Conversely, the decision suggests that a larger reduction (approaching the 30% sought by Idenix) may be appropriate where the notice of contention introduces substantial new controversy, occupies the majority of hearing time, or cannot be characterised as merely defensive. The judgment also implies that in patent cases the public interest in removing invalid patents from the Register may tilt the discretion against overly punitive costs orders on unsuccessful but non-frivolous alternative arguments.
The judgment itself carefully applies and distinguishes earlier Full Court and High Court authority. It cites Firebird and Les Laboratoires Servier for the proposition that apportionment is available on mixed success, yet emphasises that such apportionment must remain proportionate and not descend into "undue analysis". It distinguishes Sandvik on the basis that, unlike the limited contention in that case, the present notice of contention had occupied measurable time, yet the reactive character of the contention and Idenix's responsibility for complexity prevented the larger reduction that might otherwise have been indicated. It also draws on Hockey v Fairfax to acknowledge the modern readiness of courts to apportion in complex litigation, while declining to translate that readiness into a 30% discount on the facts before it.
The decision therefore stands as an illustration of calibrated, holistic costs reasoning rather than a mechanical percentage exercise based on page counts or paragraph counts alone. It reinforces that the party who generates the bulk of the complexity will ordinarily bear the greater costs burden, even if the successful party has failed on some alternative grounds.
Several questions are left for future cases. First, what degree of hearing time or written submission length is required before a notice of contention will attract a reduction materially greater than 10%? The Court rejected Idenix's "arithmetic" (25 pages of submissions, 65 paragraphs, two hearing hours) as insufficient to justify 30%, but did not articulate a bright-line threshold. Second, how heavily will courts weigh the "reactive" character of a contention? If a respondent advances an ambitious notice of contention that is not strictly necessary to answer the appeal, will a larger discount follow even if the appellant's own conduct inflated the overall hearing length?
Third, the judgment notes that in patent cases the public interest in removing invalid patents may reinforce the principle that parties should not be discouraged from raising all legitimate arguments. The precise weight to be given to that public interest in costs decisions remains unsettled. Finally, the decision was made on the papers; it is unclear whether an oral hearing on costs would have altered the percentage arrived at or the level of detail in the reasons. These nuances will continue to require careful factual assessment in each subsequent costs dispute.
THE COURT: 1 On 7 December 2017, we published our reasons for judgment (the appeal reasons). In relation to costs, we ordered that each party file and serve a short submission within 14 days. The parties have now filed written submissions and these reasons deal with costs. These reasons should be read together with (and adopt the abbreviations used in) the appeal reasons (see Idenix Pharmaceuticals LLC v Gilead Sciences Pty Ltd [2017] FCAFC 196). 2 On the issue of costs, we relevantly stated at [327] of the appeal reasons that: Idenix's appeal must be dismissed. Gilead's grounds of contention are also rejected. Prima facie, Gilead would ordinarily be entitled to costs, but that position may be affected in part by its lack of success on the grounds of contention. 3 The power of the Court in relation to costs is well established. Section 43 of the Federal Court of Australia Act 1976 (Cth) gives the Court a wide discretion in awarding costs. The exercise of the Court's discretion is not without principles or practices; it must be exercised judicially (Les Laboratoires Servier v Apotex Pty Ltd (2016) 247 FCR 61 at [305] per Bennett, Besanko and Beach JJ). The ordinary rule is that costs follow the event, although a successful party may be awarded less than its costs, or costs may be apportioned, based upon success on the issues (Firebird Global Master Fund II Ltd v Republic of Nauru (No 2) (2015) 327 ALR 192; [2015] HCA 53 at [6] per French CJ, Kiefel, Nettle and Gordon JJ; Les Laboratoires Servier at [297] to [298] and [303]). 4 Gilead contends that Idenix should pay its costs of the appeal without reduction. It says that it succeeded, the appeal was dismissed and costs should follow that event. 5 Gilead submits that although the Court can apportion costs, it should not "cast undue analysis upon whether a successful party has been unsuccessful on 'particular issues' in the context of its predominant success in the controversy overall" (NV Sumatra Tobacco Trading Company v British American Tobacco Services Ltd [2011] FCA 1295 at [17] per Greenwood J). It says that a party with a single cause of action should generally be entitled to bring forward all alternative arguments which may be successful. If successful, generally speaking, that party should be regarded as having succeeded in the event and should be entitled to costs (GG v Australian Crime Commission (No 2) [2010] FCAFC 63 at [5] per Downes, Jessup and Tracey JJ). It says that justice may not be served if parties are dissuaded from canvassing all issues which might be material by the risk of adverse costs orders (Dodds Family Investment Pty Ltd (formerly Solar Tint Pty Ltd) v Lane Industries Pty Ltd (1993) 26 IPR 261 at 271; [1993] FCA 346 per Gummow, French and Hill JJ), and that the public interest in ensuring that invalid patents are removed from the Register reinforces this principle in patent cases. 6 Gilead contends that where a notice of contention occupies only a limited amount of time at the appeal hearing, it may be inferred that it did not occasion significant costs and its rejection need not lead to a varied costs order (Sandvik Intellectual Property AB v Quarry Mining & Construction Equipment Pty Ltd (No 2) [2017] FCAFC 158 at [12] per Greenwood, Rares and Moshinsky JJ). Noting that it is almost inevitable in complex litigation (including patent cases) that the successful party may not win every argument, Gilead submits that where a litigant has succeeded only upon a portion of its claim, provided that the unsuccessful points were not frivolous and did not unduly waste time and costs, full recovery of costs remains appropriate (Bradken Resources v Lynx Engineering Consultants (No 2) [2012] FCA 1478 at [5] per McKerracher J; Damorgold Pty Ltd v JAI Products Pty Ltd (No 2) [2014] FCA 377 at [32] to [44] per Middleton J). Accordingly, the fact that we rejected Gilead's notice of contention should not, so Gilead contends, lead to a variation of the usual order. 7 Gilead notes that its contentions occupied two hours of the three-day appeal and submits that we should infer that it did not occasion significant costs, relatively speaking. Further, it says that there was no suggestion that its contentions were raised unreasonably. Lastly, quoting the appeal reasons, Gilead says that Idenix's approach to the appeal was "simply to 'put the dice into the box for another throw'" (at [146]) and to challenge factual findings in an "overly elaborate way and with a view … to having [the Full Court] conduct a re-trial of the complex factual questions litigated before [the primary judge]" (at [141]). Consequently, it says that there is no reason to relieve Idenix of the costs burden of the appeal. 8 Contrastingly, Idenix submits that while a successful party is ordinarily generally entitled to an award of costs, this is not an automatic rule; the Court may depart from the general position where circumstances make it reasonable or fair to do so, or where justice requires it (Ruddock v Vardarlis (No 2) (2001) 115 FCR 229 at [11] per Black CJ and French J; Bowen Investments Pty Ltd v Tabcorp Holdings Ltd (No 2) [2008] FCAFC 107 at [5] per Finkelstein and Gordon JJ; Dodds Family Investments Pty Ltd at 271-272). It submits that such departure may be justified where there has been mixed success on the issues; a successful party may then be awarded less than its costs or there may be an order apportioning costs (Les Laboratoires Servier at [297]; H Lundbeck A/S v Alphapharm Pty Ltd (No 2) [2009] FCAFC 118 at [8] per Bennett and Middleton JJ; Dodds Family Investment Pty Ltd at 271-272). It also refers to Firebird Global Master Fund II Ltd where the High Court observed that if it can be argued that the event of success is contestable, by reference to how separate issues have been determined, circumstances may warrant departure from the general rule and costs may be determined "on an issue-by-issue basis, involving apportionments based on degrees of difficulty of issues, time taken to argue them and the like" (at [6]). 9 Idenix further refers to White J in Hockey v Fairfax Media Publications Pty Ltd (No 2) (2015) 237 FCR 127 at [88], who stated: ... courts are now more ready to apportion the costs awarded to a party who succeeds in only some of the claims he or she brings. This may reflect the increasing factual and legal complexity of modern litigation and the multiplicity of factual and legal issues it entails, and the tendency of applicants to pursue multiple claims involving different factual enquiries in the one proceeding. 10 Idenix says that an adjustment in the costs awarded to the successful party in an appeal may take into account lack of success on some bases including on notices of contention (AstraZeneca AB v Apotex Pty Ltd (2014) 226 FCR 324 at [446] per Besanko, Foster, Nicholas and Yates JJ; Garford Pty Ltd v Dywidag Systems International Pty Ltd (2015) 110 IPR 30; [2015] FCAFC 6 at [144] per Dowsett, McKerracher and Nicholas JJ; Regency Media Pty Ltd v MPEG LA, LLC (2014) 231 FCR 588 at [38] per Bennett and Pagone JJ). 11 Consequently, Idenix contends that while Gilead should be awarded costs, the amount should be reduced in the order of 30% to reflect the time and costs related to the notice of contention. It says that this approach is supported by the following. First, to the extent that the notice of contention alleged that the primary judge should have found that claims 7 to 41 of the Patent: (a) did not derive priority from the 949 application (ground 1(b)); (b) were not fairly based on the matter described in the specification based on the disclosure of the 949 application and its incorporation by reference into the specification of the Patent (grounds 2(b) and (c)); and (c) were not fairly based on the matter described in the specification as, contrary to the primary judge's finding, "the embodiments on p 100 of the Patent did not, when read in the context of the specification as a whole, provide a real and reasonably clear disclosure of the compounds of claim 7 as the invention" (ground 2(a)), those were discrete matters that were not raised by the notice of appeal and were rejected. 12 Second, Idenix says that Gilead was unsuccessful in its contention that the claims 7 to 41 did not derive priority from the 350 application for reasons additional to those of the primary judge (ground 1(a)). It says that although ground 9 of the notice of appeal challenged the primary judge's conclusion that claim 7 (and dependent claims) of the Patent was not fairly based on the 350 application, we noted at [240] that this was done to protect its position in the event that Gilead's contention regarding the findings of fair basis on the 949 application was successful. Had Gilead not sought to overturn the primary judge's finding on fair basis in relation to the 949 application, Idenix submits that no challenge as to fair basis in relation to the 350 application would have been necessary. In any event, says Idenix, to the extent that there is considered to be any overlap between the notice of appeal and the notice of contention in relation to the 350 application, this can be addressed by a reduction in global costs, rather than separate costs orders in respect of each. 13 Third, Idenix submits that in fact significant time and effort was involved in dealing with Gilead's notice of contention. It says that it was the subject of 25 pages of separate written submissions in chief, in answer and in reply (as opposed to 40.5 pages of submissions in relation to the appeal, being approximately 38%). Further, it says that a substantial 65 paragraphs of the appeal reasons were devoted to addressing the issues raised in the notice of contention (approximately 33% after excluding paragraphs regarding background science, the Patent specification and earlier applications). 14 Idenix submits that the appeal could have been conducted more economically if the grounds in the notice of contention had not been raised. It says that this is very different from Sandvik Intellectual Property AB, where no adjustment was made for costs of a rejected notice of contention when it "occupied only a limited amount of time at the hearing of the appeal, and, it may be inferred, did not occasion significant costs" (at [12]). 15 We are of the view that Gilead should substantially have its costs, but there is some force in Idenix's submissions that some time was spent dealing with the notice of contention points. 16 In all the circumstances, we consider that there should be a small reduction of 10% in the costs we would otherwise order in Gilead's favour. We do not consider that a 30% reduction is warranted, notwithstanding Idenix's arithmetic concerning written submissions. The length and complexity of the appeal was generated principally by Idenix, with the notice of contention in reality reactive thereto. We consider that a 10% reduction more fairly represents the relative significance of the notice of contention points in relation to the hearing time and the like. I certify that the preceding sixteen (16) numbered paragraphs are a true copy of the Reasons for Judgment herein of the Honourable Justices Nicholas, Beach and Burley.
SCHEDULE OF PARTIES NSD 640 of 2016 Appellants Fourth Appellant: UNIVERSITE DE MONTPELLIER Fifth Appellant: IDENIX (CAYMAN) LIMITED
Idenix Pharmaceuticals LLC
Gilead Sciences Pty Ltd
[1993] FCA 346
(2014) 226 FCR 324
(1993) 26 IPR 261
(2015) 327 ALR 192
(2015) 110 IPR 30
(2015) 237 FCR 127
(2016) 247 FCR 61
(2014) 231 FCR 588
(2001) 115 FCR 229
The appellants pay 90% of the respondents' costs of and incidental to the appeal including in relation to the notice of contention.