David Ogilvie takes you through this paper

What causes different diseases? Since the late nineteenth century, the causes of many infectious diseases have been identified. This means, for example, that we can now say that the disease we call measles is caused by a particular virus, and, as a result, we have been able to develop a vaccine to protect the population against the disease. But things are not so straightforward in the twenty first century, when we are increasingly preoccupied with chronic diseases for which there is no obvious villain, such as a single virus or toxic substance, to point the finger at.

In order to get some idea of what might cause a disease, epidemiologists like to begin by looking for patterns in the occurrence of the disease. Describing the distribution of a disease in the population like this is often referred to as descriptive epidemiology, to distinguish it from analytical epidemiology, which means testing hypotheses about specific possible causes of the disease. In practice, a particular study may include elements of both.

The classical approach of descriptive epidemiology is to look at how the occurrence of a condition varies according to time, place, and person. For example, epidemiologists have observed that colorectal cancer is commoner in western countries than in Africa (place), and that lung cancer is commoner in people who smoke (person). In this paper, the authors have looked at time patterns in the occurrence of Crohn's disease.

What did they do?

The authors identified people in Denmark who were diagnosed with Crohn's disease before the age of 20 during a 15 year period. It is worth noting how they found their cases: they used a registry of hospital discharges. Some countries collect very good information about diagnoses made on hospital inpatients, and this can be a useful source of cases for an epidemiological study. But it can, obviously, only identify people who were admitted to hospital. This is fine for many serious illnesses, but you probably would not choose this approach to study the epidemiology of a condition which is usually treated in primary care or in outpatients—or one which people do not tell doctors about. For the purposes of this study, it seems fair to assume that even if everyone with Crohn's disease does not get admitted to hospital, the proportion who do is not likely to vary according to which month they were born in.

The authors then analysed their cases by the month of their birth—making allowances for the total numbers of births in each month of the year. They used a statistical technique called regression to produce a curve (the solid line on the graph) which best fits the data. They tell us that the incidence of Crohn's at the peak of the curve (August births) is significantly higher than the incidence at the trough (March births). So, in this study, summer babies were more likely to go on to develop Crohn's disease than winter babies.

The pattern they found is an example of cyclical variation, meaning that the incidence of the disease varies with time on a repeating cycle—in this case, every year. Many infectious diseases have seasonal patterns like this—think of flu, for example. But the time pattern of a disease can also vary in other ways. Some show a secular trend—a trend over a longer period of time, such as the decline in tuberculosis over the twentieth century. Others show sudden rises—epidemics—and then fall back again to “normal” levels.

So what?

The cyclical variation shown in this study is consistent with the idea that Crohn's could be related to exposure to some infectious agent around the time of birth. But no one would claim that it proves it. This type of study provides just one piece of evidence which might help to illuminate the cause of Crohn's.

As usual, the paper raises some difficult questions. Firstly, why did a similar British study find that the risk of Crohn's was higher in the first half of the year and not the second? And secondly, why is the incidence of Crohn's so much higher in Britain than in Denmark? As the authors point out, this makes you wonder whether the disease we call Crohn's might be caused by different factors in the two countries. It sounds as if we should expect to hear more on this subject.