Do increased risks of cancer mean we need to change our approach to treatment?

As we learn more about inflammatory diseases, we find they are associated with other risks, including major, serious issues like cardiovascular events and cancer. Do these risks imply that we need to treat or screen patients differently than we normally would? When we interpret data from studies that show increased risk, we need to assess the absolute risk of the events. Often, studies conclude there are significant risks and report the relative risks. To put those risks into clinical perspective, we need to know the absolute magnitude of the risk.

For example, in JAMA Dermatology, Jung et al. (1) reported the results of a large, population-based study of patients with hidradenitis suppurativa (HS) and concluded there is a ‘significantly increased risk of overall cancer as well as several specific cancers’ in HS patients. This may be misleading. The word ‘significantly’ is a bit ambiguous; it can mean ‘not likely due to chance” or it could mean “clinically meaningful.’ While the observed associations in Jung et al.’s study were not likely due to chance, they do not appear clinically meaningful.

The size of the database used allowed for the detection of minute differences. Of the 29 cancer subtypes listed, HS was statistically significantly associated with 6 per the adjusted hazard ratios. An adjusted hazard ratio of about 3 was observed for oropharyngeal cancers, which means there was a 3-fold increased risk of such cancers in HS patients. This sounds scary and ‘significant.’ However, this does not fully convey the magnitude of the risk. Of HS patients, 12 in 105,992 patient-years had oropharyngeal cancer; in the control group, 33 patients in 853,330 patient-years had oropharyngeal cancer. Thus in HS patients there were 0.00011 cases of oropharyngeal cancer per patient-year or 1 case per 10,000 patient-years, which means if we observed 10,000 HS patients for a year we would expect 1 to develop oropharyngeal cancer. In the control group, there were 0.00004 cases per patient-year or 4 cases per 100,000 patient-years, which means if we observed 100,000 patients in the general population for 1 year we would expect 4 to develop oropharyngeal cancer. The excess risk in HS patients compared to controls is around 7 per 100,000 patient-years. Even if we assume the observed relationship was not due to some other unmeasured variable that biased the results (which is hard to exclude), it would still take 13,415 patients followed for 1 year to observe 1 additional case of oropharyngeal cancer due to HS. It is doubtful that we would need to change how we treat or screen HS for such a small risk. It is far more likely that the costs (financial and other risks) of performing screening (with the likelihood of identifying many more false positives than true positives) would far outweigh the benefits.

Physicians and researchers tend to overestimate the importance of effects expressed in relative terms; using absolute terms allows for improved risk assessment (2). Using data from Jung et al.’s published paper, we calculated the number needed to harm (NNH) for each of the reported ‘significant’ associations (Table 1). Seeing that a very large number of patients-years is required to expect one more event makes us reticent about drawing conclusions about what cancer screening should be done differently for these patients. Recommendations for testing should weigh the very low increased risk of cancer with the costs of testing including the costs associated with false positive results and of inducing increased worry in patients (3). Given the low rates of cancer reported in this study and lack of clinically relevant increased cancer risk in HS patients, the costs of more aggressive cancer surveillance might well outweigh any potential benefit.

Table 1. Number needed to harm (NNH) calculations for statistically ‘significant’ cancer associations.

Cancer type	Number needed to harm (patient-years)
Oral cavity and pharyngeal	13,415
Colorectal	7006
Non-melanoma skin	16,204
Prostate	5947
Central nervous system	16,878
Hodgkin lymphoma	44,554
Overall	1486

Whenever we’re told ‘there’s a significantly increased risk” and “we need to do something about it,’ we should scrutinize the data to assess the absolute magnitude of the risk and consider both the benefits and costs of doing something about it.

Disclosure statement

Dr. Feldman has received research, speaking and/or consulting support from a variety of companies including Galderma, GSK/Stiefel, Almirall, Leo Pharma, Boehringer Ingelheim, Mylan, Celgene, Pfizer, Valeant, Abbvie, Samsung, Janssen, Lilly, Menlo, Merck, Novartis, Regeneron, Sanofi, Novan, Qurient, National Biological Corporation, Caremark, Advance Medical, Sun Pharma, Suncare Research, Informa, UpToDate and National Psoriasis Foundation. He is the founder and majority owner of www.DrScore.com and founder and part owner of Causa Research, a company dedicated to enhancing patients’ adherence to treatment. Aditi Senthilnathan has no conflicts to disclose.