Cancer Risk

A study published online today found that CFS patients have a higher risk of developing certain kinds of cancer. I cannot get full access to the study (curse you, paywall!), so this post is based only on the abstract.

The study used the Medicare database to identify 1.2 million cancer cases and 100,000 controls between the ages of 66 and 99 years old. CFS was identified using linked Medicare claims at a prevalence in both groups of 0.5%. The association of CFS with cancer was calculated using odds ratio. An odds ratio is the ratio of the odds of an event occurring in the first group to the odds of it occurring in the second group. If the odds ratio is greater than one, then the condition or event is more likely to occur in the first group; an odds ratio of less than one, then the condition is less likely to occur in the first group.

The study reported the following odds ratios for cancer: non-Hodgkin’s lymphoma 1.29; pancreas 1.25; kidney 1.27; breast 0.85; and oral cavity and pharynx 0.70.  It is very important to note that only the non-Hodgkin’s lymphoma odds ratio withstood multiple comparison adjustment, which is why the study only draws a conclusion about the association of non-Hodgkin’s lymphoma with CFS. The study concludes that, “Chronic immune activation or an infection associated with CFS may play a role in explaining the increased risk of NHL.”

This is scary stuff. An increased association with or higher risk of cancer in people with CFS has long been reported by Dr. Dan Peterson and others. But based on reading the abstract, I have some serious questions about this study.

First, CFS was identified in these cohorts based on linked Medicare claims. That means that a doctor had to code the patient as having CFS. We know that CFS is under-diagnosed, with only 20% or fewer cases actually receiving a diagnosis. In this study, 0.5% of the cohorts had been coded as having CFS. But that prevalence of 0.5% is higher than the prevalence rate found by Leonard Jason’s community-based study. Jason’s study found a prevalence of 0.42%. That is not consistent with the maxim that CFS is under-diagnosed. Furthermore, there is no information provided about how CFS was diagnosed in these patients. It seems highly likely to me that an unknown number of these cases are actually cases of chronic fatigue, and not CFS.

Second, odds ratio is not the same thing as relative risk. By way of example, let’s assume that 90 out of 100 sixth grade boys laugh at fart jokes, but only 20 of 100 sixth grade girls laugh at fart jokes. The odds of a boy laughing at a fart joke are 9:1. The odds of a girl laughing at a fart joke are 1:4, or 0.25:1. The odds ratio is calculated 9/0.25, or an odds ratio of 36. But this does not mean that boys are 36 times more likely to laugh at fart jokes than girls. The relative position is calculated 90/20, or boys being 4.5 times more likely to laugh at fart jokes than girls. That’s still a huge disparity, but it’s nowhere near the odds ratio of 36.

In this study, the odds ratios are very small: 1.29 in non-Hodgkin’s lymphoma cases. And this number tells us nothing about the relative risk of a CFS patient developing non-Hodgkin’s lymphoma. Based on these numbers alone, and the uncertainty about how accurately CFS was diagnosed in any of these cases, I’m not sure this paper should be sending any of us to push the panic button.