The colon in medicine: nothing to do with the intestinal tract

BMJ 1997; 315 doi: (Published 20 December 1997) Cite this as: BMJ 1997;315:1657
  1. Hamish Cameron, head of medical affairsa,
  2. Annette Robertson, statisticiana
  1. a Medical Affairs Department, Zeneca Pharmaceuticals, Alderley Park, Macclesfield, Cheshire SK10 4TG
  1. Correspondence to: Dr Cameron


Recent work on systematic reviews led us to reflect on the titles of many medical papers and then wonder why there seemed to be more colons about than before. We report our efforts to explore this gut feeling.

Methods and results

We recorded the number of titles containing at least one colon in the first 100 original or short reports in the BMJ, Lancet, and New England Journal of Medicine every five years from 1970 to 1995. This produced a colon index by year and by editor. The colon indices were subjected to several exploratory tests—the {chi}2 test, the F test, and Student's t test.

All three journals started with a low index, and over time the Lancet and BMJ more often included colons in their titles (1). The {chi}2 test on the number of titles with a colon v the number without was not significantly different in 1970 but was significant in 1995 ({chi}2=26.34, df=2, P<0.0005). There were also significant differences between the journals— BMJ v Lancet ({chi}2 =5.10, df=1, P<0.025), BMJ v New England Journal of Medicine ({chi}2=25.99, df=1, P<0.005), and Lancet v New England Journal of Medicine ({chi}2=11.64, df=1, P<0.005).


Colon index in BMJ, Lancet, and New England Journal of Medicine during 1970-95

The BMJ maintained the highest index throughout the study, peaking at 26 in 1990, when one in four papers included a colon in their title, while the New England Journal of Medicine had only one colon. The New England Journal of Medicine and the Lancet used other methods to break up titles: leading (spacing) and secondary titles of different print sizes, an occasional full stop, and dashes. There was a significant difference between the slopes of colon index over time (F=13.37, df =1, 4; P<0.02), confirming different rates of increase between the journals. Comparisons between journals showed a significant difference in the rates of increase between the New England Journal of Medicine (slope 0.006 (95% confidence interval −0.036 to 0.048)) and both the Lancet (slope 0.35 (0.17 to 0.53); t=3.74, df=1, P<0.001) and the BMJ (slope 0.72 (0.35 to 1.09); t=3.72, df=1, P<0.001). The slope estimates for the BMJ and the Lancet were not significantly different.

We also examined the colon index with each editor. During the 25 years the New England Journal of Medicine and BMJ both had three editors while the Lancet had four. Dr Arnold Relman let in the one colon in the New England Journal of Medicine in 1985, while at the Lancet the last three editors consistently maintained an annual index of around 10. Successive editors at the BMJ have been associated with greater mean indices, peaking with the current editor, Dr Richard Smith, and an index of 23. As this index was based on one observation in 1995 we reviewed the first 100 papers of 1996; 42 titles contained a colon. The BMJ is thus nearing the point when every other paper will have a colon in the title. At this rate we predict that all BMJ titles will include colons by 1999.


We found a significant difference between the colon index of the three journals over time. With colonic absence in the New England Journal of Medicine, some presence in the Lancet and prominence (megacolon) in the BMJ, could this be a predominantly British disorder?

Is the journal totally to blame? Perhaps unscrupulous authors, aware of an editor's penchant, alter their titles in a bid to win acceptance (the perverse colon). We wonder whether British papers submitted to the New England Journal of Medicine undergo colonic resections? Is the phenomenon related to specialty—are there more colons in the titles of papers submitted by, say, gastroenterologists?

Perhaps this represents further evidence of the British obsession with matters intestinal, a variant of inflammatory vowel disease. With respect to the printed page, though, we must challenge this colonic surge and ask whether it is time for a washout.

View Abstract