Photo of Bexis

This post discusses litigation holds.  Litigation holds aren’t sexy.  They aren’t going to take counsel to the Supreme Court.  They aren’t going to make the nightly news – at least we hope.  But you know and we know that in a mass tort between 90 and 99% of the cases generated by plaintiffs’ solicitation machines aren’t worth very much.  So how do plaintiffs engage in the litigation equivalent of turning dross into gold?  Pretty much anything they can to divert attention from the lousy merits of most of their claims to something, anything else.

Speaking of something else, here’s another notorious characteristic of mass torts.  The great bulk of the expense, particularly to defendants, is in discovery – mostly responding to discovery.  Even though we advocate ediscovery for defendants, the inescapable fact remains that our side has most of the documents, and even a greater percentage of the electronically stored information (“ESI”), that are going to be at issue in any given litigation.  So discovery is the other side’s number one target for diverting attention from the lousy merits of most of their claims. Plaintiffs would much rather bring motions alleging spoliation of ESI than defend motions concerning Daubert, preemption, or heaven help us all, the merits.

Plaintiffs’ counsel, particularly in mass torts, are very good at what they do.  A defendant’s litigation holds, therefore, need to be carefully prepared not to provide the other side with fodder for discovery disputes.  The contretemps we’ve seen in three recent prescription medical product MDLs illustrates some of the problems – defendants end up getting sanctioned over discovery that, if done correctly, probably wouldn’t have amounted to very much.  See In re Actos (Pioglitazone) Products Liability Litigation, 2014 WL 355995 (W.D. La. Jan. 30, 2014) (“Actos”); In re Ethicon, Inc. Pelvic Repair Systems Products Liability Litigation, ___ F.R.D. ___, 2014 WL 439785 (S.D. W.Va. Feb. 4, 2014) (“PRS”), and In re Pradaxa (Dabigatran Etexilate) Products Liability Litigation, 2013 WL 6486921 (S.D. Ill. Dec. 9, 2013) (“Pradaxa”).

In each of these MDLs, plaintiffs have blown discovery disputes that, in individual cases might well not have even been litigated, out of proportion to shift the focus from the merits of their cases to defendants’ supposed “bad faith” in the preservation of relevant evidence.  Plaintiffs predictably sought sanctions that went to the merits of the litigation – to bolster their silent majority of weak cases – rather than just discovery-related issues – demanding default judgments, adverse inference jury instructions, and of course even more cost-shifting, when defendants are already saddled with the great majority of the cost of producing far more information and data than is ever used in any trial.  That’s in addition to the predictable pursuit of attorneys’ fees and costs.

Unfortunately, in each of these three recent MDLs, plaintiffs were at least partially successful – and litigation holds were featured prominently every time.


Proceeding alphabetically, we’ll start our brief and unsettling survey with Actos.  A couple of months ago, the Actos MDL court ordered a number of sanctions for what it considered spoliation of ESI.

The Actos MDL was created in late 2011.  See In re  Actos Products Liability Litigation, 840 F. Supp. 2d 1356 (J.P.M.D.L. Dec. 29, 2011).  The litigation hold that the defendant in Actos was sanctioned for allegedly violating, however, dates from well before this litigation began – and that’s the first problem.  It’s dangerous to draft any litigation hold so broadly that it can be flipped  from one litigation to a completely different matter.

The primary issue was sales rep files – notoriously difficult to keep current because of staff turnover.  and because sales reps are a mobile workforce without benefit of offices within corporate headquarters.  The sanctions in Actos involved (in part) the efiles of four sales representatives who had left the defendant’s employ “between 2001 and 2006,” 2014 WL 355995, at *2.  That’s right, the Actos plaintiffs sought – and obtained – sanctions involving the destruction of emails involving personnel that left the company between five and ten years before any of the litigation in the MDL was ever filed.  See Actos, 840 F. Supp.2d at 1357 (docket numbers reflecting 2011 filing dates).  We’re talking about records of sales reps who left the defendant well beyond the expiration of all but the most extreme statutes of limitations.

There were also some missing electronic files for substitute sales reps (for whom dates of departure are not clear) and non-American employees (Japanese and European) who also left the company “between 2001 and 2011.”  2014 WL 355995, at *2.  Not as egregious (at least some of the requests) as to time, but it looks like any overseas defendant now must also import American discovery standards anytime they export goods.  Some of the stuff being discussed goes back to 1993 – and it’s from Japan. Id. at *4.

And the defendant gets sanctioned for this?

You bet.  In this crazy, ridiculously expensive post-Zubulake era of ediscovery, defendants can be sanctioned for routine recycling of electronic equipment, even though literally thousands of relevant emails were available from other sources and were produced. Id. at *4-5 nn. 13-18 (numbers of emails actually produced).

So how did sanctions in Actos come about?

The primary culprit was overly broad and vague litigation holds from prior litigation. The defendant issued a litigation hold for Actos/bladder cancer (we assume that’s the major injury at issue) on February 15, 2011.  2014 WL 355995, at *6.  This timing compares to the first bladder cancer “claim [being] made known to the company” in September, 2010; and “the first bladder lawsuit was filed during the summer of July 2011.”  2014 WL 355995, at *16.  There was apparently one screw up with respect to that hold, 2014 WL 355995, at *6, *16, in a company with employees on at least three continents.

It turns out there was a prior litigation hold from 2002 – having nothing to do with the current litigation or injuries – involving a variety of drugs “including Actos.”  Id. at *7.  The hold was drafted very broadly, “preserve any and all documents and electronic data which discuss, mention, or relate to Actos,” and instead of being allowed to expire when whatever prior litigation was concluded, it was simply “refreshed” several times (in 2003, 2006, 2007, and 2008) prior to the current litigation.  Id. at *8.  Sure, in some ways it’s simpler, easier to administer, and more conservative to draft very general litigation holds, rather than to tailor them to particular litigation. The same concerns might also lead to leaving a hold in place, instead of rethinking matters, when different litigation comes along.

We fully appreciate the heads-you-win, tails-I-lose position in which attorneys drafting litigation holds find themselves:  Define their preservation terms too narrowly and our side risks being second-guessed down the road for “under-preserving” or, worse yet, spoliation.  If we go too broad, our side faces create substantial preservation burdens, perhaps unnecessarily and perhaps for a very long time.  All things considered, we can sympathize with the cautious attorneys in the trenches who prefer erring on the side of preserving too much, rather than too little.  After all, it’s ordinarily better to have and not need than to need and not have.

But when a litigant is ultra-conservative with litigation holds, then it may well be obligating itself to save everything, essentially for forever.

[Defendant’s] argument is undercut by the fact that [its American subsidiaries], nonetheless, chose to institute a litigation hold sweeping in its scope and breadth, and containing no . . . limitation to or identification of a particular malady, in 2002.  The 2002 Litigation Hold is, by its own language, sweeping in nature and instructed to be broadly interpreted to encompass “any and all documents and electronic data which discuss, mention, or relate to Actos.”  [Defendant’s] own language in no way limits the hold to, or references in any fashion, one specific malady above any other; rather, the scope is defined as that “which discuss, mention, or relate to Actos” and specifically instructs the hold is to be interpreted “in its broadest sense.”

Actos, 2014 WL 355995, at *17.  With this kind of litigation hold language, a litigant runs the risk (as here) of obligating itself to produce the same stuff repeatedly – along with whatever’s necessary to keep information compiled on obsolete technology accessible.  As in Actos, these obligations can include electronic data for people who leave the litigant’s employ years, even a decade, before the litigation in which production is demanded.

Defendants, who issue broad litigation holds and conservatively keep them in place for years without regard to the litigation that originally spawned them, better follow them to the letter or risk being declared “spoliators” years later in unrelated litigation that happens to fall within the broad, vague terms of the original hold.  If a party places a litigation hold this way, the lesson of Actos is it’s going to have to live with it.  Id. at *21 (“[defendant’s] attempt to alter the clear and unambiguous language and instruction of the hold in the face of its clear language of the hold, itself, and the evolution of the holds, is wholly unpersuasive”).

Pelvic Repair Systems

The first ESI litigation hold relevant to PRS was issued in 2003, for a single case in Oregon.  The MDL was not created until 2012, almost a decade later.  2014 WL 439785, at *1.  The initial hold “was somewhat ambiguous as to whether documents prepared after [the date of the hold] were subject to a litigation hold and did not specify the nature of the claims or the injuries alleged in the civil action.”  Id. After the Oregon case was settled, the hold “was not explicitly withdrawn.”  Id.

A litigation hold in a second case, three years later, “essentially covered the same type and category of documents outlined in the 2003 notice,” but added an instruction to employees to create a litigation-specific computer file folder for “all relevant e-mails and attachments.”  Id. at *2.  Once again, when that case was resolved, “[t]here is no indication that this document hold was ever withdrawn.”  Id.

Additional cases involving different products filed in 2008 resulted in another hold “specific to the complaint and applicable to all documents involving [that product].”  Id. at *2. After the number of cases expanded to the point where plaintiffs sought “centralized management,” the defendant issued a “consolidated” hold that incorporated “the original dates on which ‘company-wide’ hold notices had been issued for various products involved in litigation.”  Id.

In PRS, the plaintiffs sought sanctions because, for “22 current or former employees,” the defendant’s “productions of custodial files contained few, if any, documents.”  Id. at *3.  These people were at all levels of corporate responsibility from the “worldwide president” down to “sales representatives.”  Id. at *4-5.  We don’t know how pervasive the problem was, because PRS doesn’t say how many defendant employees’ files were sought during discovery, so it could be 10% − it could be 0.1% − we can’t say.  But the court found these 22 missing or partially missing files sufficient to impose spoliation sanctions.

Precedent provided “minimal guidance . . . as to the effect that an isolated product liability action has on a manufacturer’s duty to preserve evidence that may be relevant to different plaintiffs in future litigation based on the same or similar product.”  PRS, 2014 WL 439785, at *8.  PRS cited Actos as the only example of a case holding “that the duty to preserve evidence for the benefit of subsequent plaintiffs was triggered by the broad and sweeping language of a litigation hold notice.”  Id.

PRS rejected the Actos ruling as overbroad as incompatible with the “litigation is reasonably anticipated” Fifth Circuit standard for when a duty to preserve arises.

Although [defendant’s] legal department did not expressly withdraw the [2003 litigation hold], key personnel in the company would have known that the lawsuit was resolved, which then eliminated the duty to segregate and preserve. . . . Nothing before this court suggests that [defendant] had any reason in 2003 to anticipate additional litigation relating to its . . . products.  Indeed, it was two years after the 2003 complaint was dismissed before another lawsuit was filed.  Moreover, an isolated lawsuit, or even two, would not reasonably lead [defendant] to believe that large scale nationwide products liability litigation was down the road.

2014 WL 439785, at *11.

Instead of earlier dates involving one-off cases, the court picked a later date as when large-scale litigation was “reasonably anticipated” under the Fifth Circuit standard. However, that date (2007/2008) was still dangerously early, significantly preceding the first FDA notice suggesting problems with these devices.  Id. at *12.  PRS chose this date in reliance on what the defendant had stated in a “consolidated” litigation hold issued in 2011 – three to four years (depending on product) after the fact:

The undersigned determines the . . . dates to be appropriate in this case given that those were the triggering dates identified in a “Consolidated Hold Notice . . . issued on February 18, 2011.  The consolidated notice was prepared by [defendant’s] in-house counsel, the corporate individual most knowledgeable about the corporation’s litigation exposures, and the one responsible for preparing and disseminating document preservation notices.  . . .[I]in the Consolidated Notice, counsel described litigation as “ongoing” and reminded employees that “[h]old notices relating to the individual products previously have been distributed company-wide on the following dates:” [citing to hold notices from 2007 and 2008].  Considering that this notice was issued at a time when sixty cases were being consolidated [in] New Jersey for centralized management, the only reasonable interpretation of the notice is that counsel for [defendant] believed that documents relating to [the products] were already being preserved company-wide for the purpose of large scale litigation. Accordingly, the February 2011 hold notice was intended to “refresh” the 2007 and 2008 document preservation notices.

Id. at *12.

PRS also relied on the defendant’s litigation hold language in finding a breach of duty, even though the defendant’s preservation efforts might otherwise have satisfied the federal rules’ “proportionality” standard:

[Defendant] contends that it had no reason to keep the e-mails of every employee, the drafts found on their hard drives, and other incidental records before the case consolidation in 2010.  While this position could potentially have merit, the undersigned need not address it because the scope of the preservation was established by [defendant’s] in-house counsel in the document preservation notices.  In each case, counsel instructed employees to segregate and hold “all documents, memoranda, notes, files, emails, etc.” relating to the device at issue and pertaining to labeling; pharmacovigilence; regulatory; discovery, research and development; product communications; marketing and sales; manufacturing; and distribution.

Id. at *13 (emphasis added).

With both the date and scope of the duty to defend in PRS established, even the sanctioning court conceded that the litigation holds at issue could well have exceeded any preservation obligation required by the relevant rules.  Thus, the missing files (nine of the 22 employees remained with the defendant for at least some period after the 2007/2008 litigation holds) resulted in spoliation sanctions.  The defendant was sanctioned even though “no specific loss ha[d] been identified,” and the missing information could well have been duplicated elsewhere, particularly in “common source” locations where “most employees stored or transferred such information.”  Id. at *19.

However, because neither scienter nor prejudice was established, sanctions in PRS were limited to monetary sanctions (currently being briefed), which could be substantial given the intensity with which issues, particularly discovery issues, are litigated (we would say over-litigated) in MDL proceedings.  Id. at *23; see id. at *20 (“[p]laintiffs have not provided the court with any concrete evidence of prejudice to their cases”).


The discovery disputes in Pradaxa involved more than preservation, but with respect to that issue, they focused on ESI in the possession of certain employees (similarly to Actos and PRS) and failure to preserve business-related text messages from certain employees’ cell phones.  Some of the relief granted – requiring witnesses in a foreign country to come to the United States – was declared beyond the power of the MDL court in a mandamus decision.  In re Petition of Boehringer Ingelheim Pharmaceuticals, Inc., , ___ F.3d ___, 2014 WL 274084 (7th Cir. Jan. 24, 2014).  For purposes of this post, we’re interested in the effect of litigation holds on the finding of spoliation, rather than these other issues.

The litigation hold issue in Pradaxa was primarily related to the extent of the hold issued in response to the MDL rather than on the pernicious effect of previous hold orders.  “[Defendant’s] duty to preserve material relevant to this litigation arose in February 2012 when it received a lean letter regarding the first post-launch Pradaxa product liability suit.  [Defendant] has indicated that it issued a litigation hold shortly thereafter − in April 2012.”  2013 WL 6486921, at *6.  In June, a motion to create an MDL stated that there would be “at least 500” cases. Id. at *7.  This was enough for defendant to “understand the size and scope of this litigation.”  Id.

One aspect of the discovery problems in Pradaxa was the administration, rather than the scope, of the defendants’ overall litigation hold.  “The Court has examined the defendants’ ‘holds,’ . . . and does not take umbrage with the language or scope thereof.”  2013 WL 6486921, at *6.  Rather:

The defendants have taken a too narrow and an incremental approach to its “company-wide” litigation hold.  The Court has been relying on the common meaning of the words that that the defendants have a company-wide litigation hold on all persons who have custody of any documentation relevant to Pradaxa. The production requests of the plaintiffs are so broad as to cover any possible derivation of means to document someone’s thoughts, words and deeds. . . .  [A]ll of the materials that are discussed in this order were clearly covered by production requests and further anticipated by the Court as subject to the “company-wide” litigation hold.

Id.  Pradaxa thus rejected the defendants’ explanation that employees who had “not been identified [by plaintiffs] as custodians” when they left their employ were not subject to the hold.  Id. at *10 (discussing failure to preserve “custodial files”).  Rather, the duty to preserve was not “limited” to material “that has actually been requested by the opposing party.” Id. at *12.

The litigation hold specifically concerning sales representatives’ ESI, however, was held to be “grossly inadequate.”  Id. at *14.  The Pradaxa opinion identifies several deficiencies – all of which involved events occurring after the MDL’s creation:  (1) the hold was “initially applied” only to information generated by “the specific sales representatives who detailed specific plaintiff’s physicians,” and “nothing” was done to preserve information before these persons were identified; (2) the hold was initially expanded only to “sales representatives currently detailing” the drug; (3) after 13 months, the hold was expanded to “all sales representatives who had ever detailed” the drug; (4) detailing by persons not classified as sales representatives was not included in the hold until several more months had passed, and then only incrementally, as with the sales reps; and (5) “the defendants’ hold . . . did not expressly extend to text messages” for more than a year after such messages had been sought in discovery.  Id. at *14, 16.  The court concluded:

The litigation hold described by the defendants is wholly inadequate in light of the size and scope of this litigation.  The defendants were under a duty to preserve information that they knew or reasonably could foresee would be relevant to imminent or pending litigation. . . .  Once the duty to preserve was triggered, the defendants owed a duty to preserve evidence that may be sought during discovery and should have implemented an adequate plan to find and preserve relevant evidence.

Id. Given the size of the developing litigation, this court opined that anything short of a hold extending to all detailing personnel was inadequate.  Pradaxa was not a “regional case . . . with no indication of the litigation expanding into nationwide litigation,” and “nothing in any case management order . . . can be interpreted as suggesting such a tailored litigation hold was acceptable.”  Id. at 15.

Pradaxa further held that text messages should have been explicitly mentioned in the hold, to ensure that “employees understood that text messages were included.”  Id. at *17.  Further, defendant should have taken affirmative steps “to intervene in the automated deletion of employee text messages on company issued phones,” since “intervention in the routine operation of an information system is one aspect” of a litigation hold.  Id.

Pradaxa ultimately concluded that defendants’ actions and omissions with respect to both the scope and administration of litigation holds were in “bad faith.”  2013 WL 6486921, at *18-19 (identifying potential loss of some of a key witnesses’ files; the “inadequate” hold on detailing employees, and “failure to preserve text messages” as examples).  The extent of sanctions in Pradaxa has yet to be fully determined, but included almost $1 million in fines as well as a threat of “remedies yet to be addressed.” Id. at 20.

*          *          *

So what take-aways can we offer from the adverse experiences of the defendants in Actos, PRS, and Pradaxa with respect to litigation holds?  We see several.

First, litigation holds are too important – their consequences too grave – to be left solely to in-house counsel, unless such counsel have significant mass-tort litigation experience under their belts.  Ideally (yeah, we’re litigators), litigation holds should be drafted by litigators, since litigators understand the consequences of getting things wrong, and will be in the line of fire if that happens.  For good reason, increasing numbers of litigators subspecialize in electronic discovery, including the formulation of preservation terms, custodian lists, and the like.  We consult with these specialists liberally when undertaking this sometimes-complicated task.

Second, litigation holds should always be proportionate to the potential scope of the litigation for which they are issued.  Opposing counsel relish playing Monday morning quarterback in litigation hold decisions and will seek to persuade a court to use hindsight to judge defendants.  What might be a proportionate scope at one point can become vulnerable as litigation evolves/expands into something (like Pradaxa) requiring a broader hold.  Salami slicing litigation holds when a mass tort is potentially (at least in hindsight) apparent – or worse already underway – can get a defendant sanctioned. On the other hand, imposing a mass-tort-sized litigation hold for one-off cases, without clearly thinking it through, is asking for trouble down the road.  Employees won’t take it seriously, and the effort to enforce that kind of hold according to its terms is difficult to maintain over time.  And believe us, the already substantial burdens of overly broad preservation only get worse as time passes

Third, litigation holds should conform to the products and injuries at issue and should be reviewed frequently so that their scope and implementation remain tailored to the needs of the litigation.  That sounds easier than it is – we’re taking for granted that the products and alleged injuries are readily apparent, which isn’t always so, even though one would think a preservation obligation would require it (just reread the cases we discussed).  Experience shows that this effort also requires some careful thought.  It can be expensive and asking for trouble to use conservative, catch-all language where only one of a number of related products or product risks is at issue.  A defendant issuing an overbroad litigation hold is likely to be held to its terms, even when they exceed the scope of any legal preservation duty, and as years pass those terms can be quite painful.

Fourth, we’d like to think that litigation holds can end.  Ideally they would, but perpetual preservation is one of the practical problems motivating the current effort to modify the discovery rules, including Rule 37 concerning sanctions.   If placing a litigation hold is difficult, the decision to lift a litigation hold is downright vexing, and requires fully considering the hypothetical landscape of “reasonably anticipated” disputes.  It’s no easy task to say when:  When is litigation truly over?  After a settlement is signed?  After the last dismissal is filed?  After the last objector’s appeal is dismissed?  After the statute of limitations expires?  Such decisions ought never be made alone.  In drafting litigation holds, consideration ought to be given to how, when, and under what circumstances it is proper to end them.  Similarly, when a case is over, litigation holds related to it should be revisited in light of the defendant’s litigation environment at that time.

Fifth, don’t expect employees to understand legal terminology or to make inferences from broad terms such as “documents” or “electronic information.”  Spell, it out. Follow up.  Tell them the full scope of what they are expected to preserve.  Offer technical support to make sure that it’s easy for employees to do the right thing.

Sixth, corporate culture needs to be taken into account in the issuance of litigation holds. Particularly if the hold extends to corporate entities outside the United States, it must be drafted – and internally enforced – to make sure that personnel unfamiliar with (and, indeed, resistant to) American mass tort litigation do what the letter of the hold requires them to do.  Whoever is responsible for issuing the hold must take ownership of it on an ongoing basis.

Seventh, expect every litigation hold to be enforced to the full extent of its terms, and plan for this. What a defendant writes in a litigation hold, it will have to live with it − even if, in the absence of the hold’s language the other side wouldn’t have a leg to stand on.  This is a classic description of a “gotcha game,” but we can make it harder for the other side to play.

Finally, plaintiffs should have to practice what they preach (see our ediscovery for defendants posts).  Plaintiffs have computers, too.  After the claims of each plaintiff in the MDL have accrued, are they deleting their Internet search history (relevant, at least, to the statute of limitations), and letting their own email autodelete functions continue unabated.  Have they “sanitized” their social media accounts?  It’s high time that plaintiffs are held to the same standards as defendants in the ediscovery arena.

We thank Dave Cohen, Pat Antezana, and Linda Werner, all of Reed Smith, whose recent client alert, “To Hold and To Have:  Issuing a Legal Hold is Not Enough,” inspired this post.