IP Cases & Articles

Numerical limits revisited. Smith & Nephew v Convatec

Numerical limits in patent claims are, it seems, always open to interpretation, despite numerous cases that have considered the issue. A recent case in the English Court of Appeal, Smith & Nephew v Convactec [see note 1 below for case details], serves to illustrate why.

The case concerned a process for the silverisation of gel-forming fibres used in wound dressings. A key step in the process involved subjecting the gel-forming fibres to an agent which facilitates the binding of the silver to the fibres, wherein the according to the claim the agent must be present in a concentration between 1% and 25% of the total volume of treatment. The defendant wished to use a process that involved a concentration of binding agent which did not exceed 0.77%, and sought a declaration of non-infringement to that effect.

First instance decision

At first instance, the judge held that the proposed process did not infringe. He did so applying "significant figures" interpretation to the numerical limits in the claim. On that basis, the lower end limit was expressed to one significant figure and accordingly "1%" meant 0.95% - 1.4%. Thus 0.77% was below the claimed range. Interestingly, the upper end was expressed to two significant figures, and therefore 25% here meant 24.5%-25.4%. It is worth noting that the accuracy of deterination at the different ends of the claimed range would therefore have to be different, under this interpretation.

The Court of Appeal's approach

Convatec, having lost at first instance, appealed on the basis that the claim limits should be construed as being to the nearest whole number – thus 1% encompassed 0.5% and above, and therefore there was infringement at 0.77%. By contrast, Smith & Nephew contended that the claim limits should be interpreted precisely as 1% and 25%, with no rounding, an approach that has been adopted in a number of European Patent Office (EPO) decisions [see note 2 below].

The judge reviewed numerous UK and EPO decisions involving numerical limits.

Kirin Amgen Inc v Hoechst Marion Roussel

The judge stated that, under UK law, the principles of purposive construction in the leading UK decision on interpretation of patent claims generally, Kirin Amgen Inc v Hoechst Marion Roussel [see note 3 below for case details], applied equally to numerical limits as they do to any other claim integers. Two principles were particularly relevant to this issue:

  1. The reader of the claim brings common general knowledge to that reading and understands that the purpose of the limit is to demarcate an invention.
  2. The patentee will have chosen the words of the claim on the basis of skilled advice and therefore in so far as the patentee has cast their claim in specific rather than general terms, is likely to have done so deliberately.

PLG v Ardon 

[See note 4 for case details]. A claim limit of "not less than 75%" was infringed where the relevant aspect of the accused product varied from 60% to above 75%, depending on where the measurement was taken. In the judge's view, this was a question of an immaterial variant rather than an expansive interpretation of a numerical limit, and it would be very rare for a numerical limit to be given such an expansive interpretation.

Auchincloss v Agricultural & Veterinary Supplies

[See note 5 below for case details]. The judge agreed that as a general matter numerical limits were quite different to descriptive words when considering "variants". Consideration of variants sits more easily with a word such as "vertical" than it does with an expression such as "between 87.0° and 93.0°".

Lubrizol v Esso & TH Goldschmidt v EOC

[See notes 6 & 7 below for case details]. The Court of Appeal had interpreted "at least 1.3" as including "1.25 and above" on the basis that 1.3 was expressed to two significant figures, and would have been interpreted as such by the skilled addressee using the conventions adopted by scientists. Similarly, in TH Goldschmidt v EOC a range of "pH 5 to 8" was interpreted as encompassing values from 4.6 to 4.9. Had the patentee wished to exclude such values, the claim could have been expressed as "5.0 to 8.0", and the skilled addressee would have known that measurement to this level of accuracy was conventional. The other UK authorities come to very similar conclusions.

In this case, the claimed range was expressed to different significant figures at each end. How therefore should it be interpreted? The judge set out five principles that he felt were of particular relevance to this kind of claim. Perhaps the most important are that the meaning and scope of a numerical range must be ascertained in the light of the common general knowledge and the context of the patent as a whole; and that the skilled person may understand from all of these circumstances that the degree of precision set out in the claim is intended to include all values within that range, stated to the same degree of precision.

Having reviewed the patent as a whole, the judge rejected the "exact values" approach contended for by the defendant. This was inconsistent with the description which set out a number of values stated to several decimal places – had "exactly 1 to 25" been intended, this would have been inconsistent with that disclosure. He also rejected the first instance judge's approach of applying significant figures to the ends of the claimed range. He noted the anomaly that there would be different degrees of precision at each end and hence asymmetry in the claim. On the evidence and having reviewed the patent, he considered the "nearest whole number" approach to be the right one. The numbers in the claim conveyed a degree of accuracy with which the skilled addressee needs to make a determination of the relevant concentrations, and this should logically be same at both ends of the claim. . Accordingly, the limits were interpreted as being to the nearest whole number, which therefore meant the range extended between 0.5% and 25.4%. There was therefore infringement.

The clear message from this decision is that there is no single applicable approach to interpreting numerical limits. Significant figures, exact values or whole numbers each have their place. Numerical limits will be construed in context, taking into account what the patentee will be understood to have been trying to convey by choosing to express the limits in the way they were. This involves considering such limits with care, both when claiming and considering the validity or infringement of such a claim.


  1. Smith & Nephew v Convactec [2015] EWCA Civ 607
  2. It should be noted that this is far from being the only approach adopted in the EPO.
  3. Kirin Amgen Inc v Hoechst Marion Roussel Ltd [2004] UKHL 46
  4. PLG v Ardon [1993] FSR 197
  5. Auchincloss v Agricultural & Veterinary Supplies Ltd [1997] RPC 649
  6. Lubrizol v Esso [1997] RPC 195
  7. TH Goldschmidt v EOC Belgium (25 January 2000)