Article Text

Bladder cancer risk in painters
  1. Paolo Vineis
  1. Correspondence to Professor Paolo Vineis, MRC/HPA Centre for Environment and Health, Imperial College, London W2 1PG, UK; p.vineis{at}

Statistics from

The paper by Guha et al (see page 568) on the risk of bladder cancer in painters, which appears in this issue of OEM,1 is important on several grounds. Starting with the more general reasons, it is a clear example of how reporting scientific information in detail, with extensive analyses of the available data, is an exercise in ‘deliberative democracy’. A certain simplification of scientific messages, such as the use of the words ‘carcinogenic’ or ‘non-carcinogenic’, does not help in decision-making if it is divorced from a complete and intelligent description of the evidence at hand. Simplification of messages, as frequently done by the press or, worse, by actors with a vested interest, implies little consideration of the ability of the lay public to understand scientific facts and their interpretation. The evidence in this case is clearly laid down and is strikingly consistent. Out of 41 studies reviewed, 37 show a RR greater than one. Excess risks are found in all continents, in both genders, with different definitions of ‘painter’ and after adjustment by smoking. There is no evidence of publication bias, and there is little heterogeneity across studies. The RR for those exposed for more than 10 years is 1.8, a non-negligible excess.

Why is such detail important? First, because it makes the causal inference (‘sufficient evidence’) very persuasive, since it shows the reasoning behind the final judgement. Second, a rigorous quantitative approach (a meta-analysis) reinforces the traditional more qualitative approach used by the IARC Monographs Working Groups until now.2 The concepts are the same (amount of information, strength of association, consistency, lack of bias and confounding), but they are potentiated by the application of formal meta-analytical tools. Third, the overview of the case of painters has a historical interest, because from the first imperfect studies evidence from a number of sound studies has accumulated. In other words, it is a situation in which the early reports definitely turned out not to be ‘false positives’.

Another meta-analysis has been recently published, that covers most (but not all) of the studies included in the report under discussion.3 The meta-analysis by Bachand et al3 is, however, much less clear. Although it essentially comes to very similar conclusions (an overall risk estimate of 1.23 for bladder cancer), the paper is unfortunately rather confusing. The authors used what they call external adjustment for smoking, which they correctly identify as a potential confounder. Their external adjustment consists in applying to the risk estimates for painters the following assumptions: (a) an estimate of strength of association between smoking and bladder cancer; (b) an estimate of the proportion of smokers (current and ex-) and never smokers among the general population; and (c) estimates of the proportion of smokers among painters. While for points (a) and (b) I do not see particular objections, for point (c) they use an old estimate from Theodor Sterling, who in the 1970s claimed that occupational exposures explained the association between lung cancer and smoking (see also In any case, they do not seem to use Sterling's data when they come to their results section. Overall, their calculations are difficult to follow, and external adjustment does not seem to substantially falsify the positive association with painters. Also adjustment by socio-economic status is applied to the estimates (in a rather questionable way) but does not dramatically change the overall conclusion.

Going back to ‘deliberative democracy’, the IARC Monographs have for a long time championed such an approach to the translation of scientific evidence into prevention. To cite Robert Frost, the American poet, “the only way forward is the way through”: there is no alternative to facing the entire body of evidence. Applying rigour to causality criteria,2 together with a full and detailed account of the data at hand, is the only way to help society address risks for health from occupational and environmental exposures. This paper is a further step in the same direction. However, I have heard criticisms of the fact that broad categories or industries (such as the rubber industry, or hairdressers, or painters) are addressed in IARC evaluations. This attitude would be against the scientific principle of an analytical approach that calls for specificity. For example, one cannot test the hypothesis of risks in painters or the rubber industry in rats, or judge the biological plausibility of such associations. First, this is not completely true, because of course one can conduct studies on the specific agents used by painters. Second, again, it is a question of ‘deliberative democracy’: if the best available evidence is on painters (and not on lower-level categories), and such evidence is sufficient, it is a duty of the researcher to evaluate it as it is. Like the ‘duty of care’, there is also a ‘duty of prevention’, and the right question is not whether one should evaluate the evidence for painters as such, but what happens next. The authors of this paper correctly suggest that one should look into specific exposures such as metal coatings, wood varnishes or stains. I would add that the existence of a risk is itself a reason to clean up the working environments for painters while studies are conducted to disentangle the possible causal agents. Otherwise we will find ourselves repeating again and again the 1966 statement by Case, referring to dye manufacturers and rubber workers: “In retrospect, it now seems astonishing that the epidemiological evidence that had accrued, although it was fragmentary, should have been so lightly discounted”.4

View Abstract


  • Linked articles 051565.

  • Competing interests None.

  • Provenance and peer review Commissioned; not externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Review
    Neela Guha Nelson Kyle Steenland Franco Merletti Andrea Altieri Vincent Cogliano Kurt Straif