E cdip/15/inf/2 original: english date: january 8, 2015 Committee on Development and Intellectual Property (cdip) Fifteenth Session Geneva, April 20 to 24, 2015

VI. Results

Figure 1 shows the time trends for the unique product-active ingredient combinations. There is a marked increase in the share using patents during the mid-1990s. Also the share relying only on trademarks increases substantially beginning the second half of the 1990s.

Figure 1: Time-trend patents and trademarks for unique product-active ingredient combination

When we examine the ownership of this IP, we see striking differences in the regional patterns. Figure 2 shows the share of trademark and patent filings coming from domestic and foreign entities in Chile, by date of the corresponding ISP registration. Almost all the patent filings are by entities based in Europe and the U.S., with the exception of a small increase in Chilean-origin filings during the most recent period. The total share of Chilean-origin filings is less than two per cent of total pharmaceutical patent filings, and none of these filings match to active ingredients in the ISP registration data. In contrast, over half the trademark filings are by Chilean entities, with the other half largely from Europe and the U.S.

Figure 2: Time-trend patents and trademarks by foreign and domestic entities

An examination of the ISP registrants makes the reasons behind this phenomenon clear. Table  2 shows the number of ISP registrants from each region of the world that are listed as performing each type of function when they register at the ISP (a registrant may list several functions for each registration). Clearly Chilean firms do almost all of the importing, distributing, quality control, and local packaging of the drugs, whereas foreign firms are the source, licensor, or foreign packagers. Manufacturing can be done in either location. However, it is fairly clear that foreign firms are the source of originator drugs and therefore hold almost all of the patents, whereas marketing and distribution as well as the production of generics is the province of the Chilean registrants, who therefore hold a large number of Chilean trademarks.

Table 2: Functions performed by ISP registrants by country of origin

However, note that Latin American countries other than Chile are frequently the source of the drugs (probably at the expense of U.S. firms, judging from the patenting numbers). It appears that some multinationals use one or two Latin American countries (usually Argentina, Uruguay, or Panama, sometimes Brazil or Mexico) as a base for production and distribution throughout Latin America.

Figure 3 looks at the distribution of the status of the patent and trademark applications as a function of whether they have matched to our ISP product list. This figure also shows some interesting differences. Broadly speaking, trademark and patent applications associated with products and active ingredients that have been registered at the ISP are more likely to be either pending or granted, and much less likely to have been rejected, abandoned, or withdrawn. This is not unexpected, but it does suggest that these applications are of more economic relevance than the unmatched applications.³¹

Figure 3: Legal status of patents and trademarks


Primary versus secondary patents
Our main interest is in the use of secondary patents by foreign originator companies in Chile. Collecting the relevant data for investigating this question is challenging. In this paper we rely on the identification of our patents as primary or secondary that was done by internal and external patent examiners at INAPI. Of the 504 Chilean pharmaceutical patents that match to our list of active ingredients, 113 (22%) were identified as primary patents, with the remaining 78% being secondary. This ratio of 1:4 is significantly lower than the ratio of 1:7 found by the pharmaceutical sector inquiry of the European Commission. If we look at all granted patents regardless of whether they have matched to a product registered at the ISP, we find that there are more primary than secondary patents. Of course, this may simply reflect the fact that secondary patents are more often rejected by the patent office precisely because they do not cover a new active ingredient.
The 504 matched patents are associated with 322 of the 2,630 active ingredients. Of these active ingredients, slightly more than one third (185) have at least one primary patent. In about half the cases (89 or 48%), the associated primary patent is the first patent on that ingredient, but in the remaining cases, there is a secondary patent preceding the primary patent.

Figure 4 shows the trends in ISP-matched pharmaceutical patent applications for the two types of patents separately. During the 1990s after the introduction of pharmaceutical product patents, both types of applications increase but after 2005 there is substantial decline for reasons unknown to us, although this may reflect the worldwide slowdown in the introduction of new pharmaceuticals.

Figure 4: Pharmaceutical patent applications by type

Figure 5 shows the trend in the share of matched patents that are identified as being primary patents in two ways: by the year of the patent application, and by the year of first ISP registration that includes the covered active ingredient. The trends show contrasting results suggesting highly variable lags between a patent application and the associated ISP registration. Although the share of patent applications that are primary declines very slowly over time, the share of patent-related new ISP registrations increases steadily over the period.

Figure 5: Primary patents by patent application year and product registration year

To investigate the timing between a Chilean patent application and the first associated ISP registration further, we computed the lag between the two and plotted the distributions for primary and secondary patents in Figure 6.³² This figure clearly shows that the great majority (86%) of the primary patents are applied for before the first time the associated ingredient is registered at the ISP. In contrast, only 56% of the secondary patents are applied for before the initial ISP registration. A nonparametric test of the difference between the two lag distributions yields a χ²(1) of 37.5 and is highly significant. The median lag for primary patents is 6 years and for secondary patents it is 2 years. In a number of cases, the lags are over 5 years, which suggests delayed entry into the market.

Figure 6: Lag between patent application and product registration

Table 3 looks at the number of patents that protect a given active ingredient. About 55 per cent of the active ingredients are protected by a single patent and 34 per cent of active ingredients that are patent protected are protected by 2 or 3 patents. Very few active ingredients are associated with a larger number of patents. When we look at the breakdown into primary and secondary patents 72% of active ingredients that are protected by a single patent are in fact protected by a secondary patent. Among drugs that are protected by several patents, in most cases they are protected by only secondary patents or a combination of primary and secondary patents.
Table 3

To gauge the effect of secondary patents on potential patent terms extensions, Figure 7 looks at the lag between the application date of the first primary patent and that of the latest secondary patent by active ingredient. The figure shows that in most cases the lag is positive, meaning the application for secondary patent was filed after the primary patent, and in many cases this lag amounts to several years. If the secondary patent offered exclusivity to some degree, Figure 7 would suggest that in some cases, companies could gain a number of additional years of patent exclusivity through the filing of secondary patents.

Figure 7: Lag between earliest primary patent and latest secondary patent by active ingredient (active ingredients protected by both primary and secondary patents)