By Kevin E. Noonan --
Lex Machina, a commercial venture spinning out of the "quantitative statistics" trend in patent scholarship popularized by Mark Lemley, Kimberly Moore, David Schwartz, and others, has released a Report on ANDA litigation that provides a slew of statistics ripe for interpretation (or at least speculation).
Written by Brian C. Howard (Legal Data Scientist and Director of Analytics Services) and Jason Maples (Legal Data Analyst), the Report covers the years 2009 to date and begins with a brief summary defining what an ANDA case entails. Then it dives right into the statistics, some of which are not surprising. For example, the Report shows (Figure 1) that most ANDA litigation occurs in the District of Delaware (678 cases), the District of New Jersey (481 cases) and the Southern District of New York (148 cases); the Northern District of Illinois, with 39 cases, is a distant fourth in ANDA case frequency. Within those judicial districts, the judges hearing ANDA cases are easily recognizable to those representing ANDA clients (Figure 2): Judges Sleet (195 cases), Robinson (187 cases), Stark (139 cases) and Andrews (122 cases) in Delaware; Judges Pisano (64 cases), Cooper (49 cases), Chesler (38 cases) and Sheridan (34 cases) in New Jersey; and Judge Stein (33 cases) in New York, as well as Judge Keeley (27 cases) in the Northern District of West Virginia. Similarly, that 98% of drugs involved in ANDA litigation are prescription drugs will come as no surprise and, indeed, is almost a requirement for ANDA litigation (albeit about 2.5% of drugs involved in the cases reported are "over the counter" drugs; Figure 15).
Filing trends have been relatively consistent over this period (Figure 3), during which they have averaged between 239 and 293 cases per year between 2009-2013; this year has seen 323 cases filed by the end of June with the biggest uptick having occurred in the first two quarters of the year (Figure 4). Curiously, most of these cases seem to have been filed in Delaware (Figure 6). The Report provides (Figure 7) some perspective, however, noting that while there have been 323 ANDA cases filed so far this year, that number pales in comparison to the 3,564 "non-ANDA" patent cases filed so far in 2014 (and the 5,167 and 5,793 such cases filed in 2013 and 2012, respectively).
The Report also notes (Figure 8) that in 70.3 percent of ANDA cases an injunction is obtained by the prevailing patentee, compared with a 57.6% rate in non-ANDA cases; generally, a prevailing patentee is rewarded an injunction to prevent "at risk" launch and the FDA does not approve the generic drug until the Orange Book listed patents have expired. Thus, this statistic may merely reflect the percentage of the time an NDA holder prevails in ANDA litigation where the Court grants relief only under 35 U.S.C. § 271(e)(4)(a).
Using "word cloud" graphics (Figures 12 and 13) the Report identified oxycodone and metformin as the drugs involved in the greatest number of ANDA litigations, which is not surprising in view of the patient populations existing (and growing) for these drugs. These drugs are also involved in ANDA litigation having some of the largest number of Orange Book listed patents at issue (metformin, 20 patents, exceeded only by testosterone having 23 asserted patents) and the greatest number of cases (oxycodone, more than 50 cases; metformin, 38 cases) as shown in Figure 14.
The surprising statistics are set forth at the end of the Report, having to do with the time in an Orange Book-listed patent's term that ANDA litigation ensues. For example, Figure 20 shows that the median "patent age" at ANDA filing has decreased, from 10 years in 2009 to 5.1 years in 2014 (with the trend beginning in earnest last year). Conversely, the time between the median date of FDA approval and ANDA litigation filing has increased, from 3.9 years in 2009 to 4.9 years in 2014. One explanation for the former trend is a statistical one: looking at the "raw" data in Figure 22, the decreased age may have something to do with a sharp skew in the data for 100-200 patents challenged within the first three years of their term. Alternatively, this trend may reflect slower innovation rates in the industry, so that truly "new" drugs are fewer than the combinations or follow-on drugs. This has a lot to do with economics, and with the "low hanging" fruit of these drugs having been harvested, as well as competition by biologic drugs which are more expensive (and may be more profitable) but also more specific for their biological targets. The second trend (Figure 21) is more difficult to sort out, but perhaps it just reflects generic challengers waiting (or having to wait) for the innovator to grow the market to a size worth pursuing. The raw data supporting these numbers (Figure 23) shows "negative" times wherein an ANDA case is filed before approval, which suggests that the statistics include declaratory judgment actions filed by generic challengers trying to invalidate the patents before they are even listed in the Orange Book.
Finally, the time before expiration at approval (based on expected expiration dates) has risen, from 8.1 years in 2009 to 15.4 years in 2014 (Figure 25). This may merely reflect faster approval times by the FDA, perhaps spurred by the increase in combination therapies using individual drugs already approved, or the frequency of filings for approval of related," evergreening" forms of a drug such as purified enantiomers, or a greater reliance on patent term extensions to increase patent life, or some combination of these factors.
While these statistics are fascinating, it is well to remember Twain's imprecation against "lies, damned lies, and statistics" when considering them (as well as the proverbial "grain of salt"). Perhaps more important than any of these statistics are recognized trends in the industry, which include at least the following:
• There are more ANDA filers for each branded drug than in the past (when, if you were not the first filer you might not file at all)
• There are more follow-on or derivative branded drugs, or combinations of drugs which are based on older, well-established drugs
• The Supreme Court's recent Actavis decision, and FTC actions (and private party actions supported by the FTC), make it harder to settle
And, of course, the ANDA statute requires a branded drug maker to file suit (or be subject to a DJ action that does not postpose generic approval), so that the actions of the generic filers and trends on that side of the fence influence what branded, NDA holders do (and must do). While the type of statistics provided by Lex Machina are interesting and may even be informative, they are just a reflection of these, and other, larger trends in the pharmaceutical industry. These include, of course, the 800 pound gorilla of biologic drugs which most industry observers have noted are the fastest growing sector of the pharmaceutical industry and are completely outside the scope of ANDA litigation or the Hatch-Waxman regime.
For additional information regarding this topic, please see:
• Silverman, "Sue Me, Sue You Blues: More Generic Patent Litigation is Being Filed," WSJ Pharmalot, November 5, 2014