I don’t know about you, but I am getting really irritated with people searching desperately to find problems with the long term outcomes of preterm infants.

A new publication exemplifies this. A linked database study from Sweden has compared the frequency of hospitalisation for serious psychiatric disorders according to the gestational age at birth. (Nosarti C, Reichenberg A, Murray RM, Cnattingius S, Lambe MP, Yin L, MacCabe J, Rifkin L, Hultman CM: Preterm Birth and Psychiatric Disorders in Young Adult LifePreterm Birth and Psychiatric Disorders. Archives of General Psychiatry 2012, 69(6):610-617.) http://archpsyc.jamanetwork.com/article.aspx?doi=10.1001/archgenpsychiatry.2011.1374

The first question to ask is whether the data appear likely to be reliable. The subjects were born between 1973 and 1985, at a time when few women were getting antenatal ultrasounds, so individual errors are likely to be frequent, but the overall distribution of gestational age should be OK. But, the distribution of gestational ages is highly suspect: 17% of the deliveries were at 42 weeks or more, whereas 4.1% were less than 37 weeks and 0.4% less than 32 weeks, this should immediately ring alarm bells about the reliability of these data. These distributions are so far from what we would expect to see; either there has been a huge change in human biology over the last 30 years, or these gestational ages are systematically wrong.

But if we were to assume for a moment that the gestational age data were reliable, what do they actually show?

Well let’s focus on what is supposed to be the most striking finding, an enormous increase in bipolar disorder, relative risk among infants less than 32 weeks of 7.2, ‘fully adjusted’ relative risk of 7.4.

If we examine these numbers a little closer, rather than looking at the ‘relative risk’ let us examine the actual frequencies. Among full term infants there were about 150 hospitalisations among just over 1 million individuals. Among infants of less than 32 weeks gestation, there were ….. wait for it… 4 hospitalisations among 5125 individuals. That’s right, this paper, considered worthy of a news item in the BMJ, this paper, which makes broad general comments about brain development in the preterm and how surveillance of preterm infants should continue for life, this paper is based on 4 cases of hospitalisation among over 5000 preterm patients.

In addition, apart from the unreliability of the gestational age data, the authors had little information to adjust the frequencies for other risk factors, many of which could easily differ between premature and full term individuals.

So even if it were all true, the actual attributable risk of being preterm is 1 admission for bipolar disorder per 10,000 patient years. Now as an exercise in epidemiology, perhaps there is some value here, if there is some tiny increase in risk of serious psychiatric disorder among preterm infants that could have some importance (though I have difficulty imagining what it is), but to state, as these authors do, that these data show some general abnormalities in brain development among preterm infants is ridiculous.

If we add together all of the disorders that the authors claim to be significantly related with prematurity, we find that the term individuals had 3,114 hospitalizations for depression, nonaffective psychosis, bipolar disorder and eating disorders, out of the same million people. The premature infants had 41 out of 5,125. Sure the relative risk is increased, but the actual attributable risk is about 10 admissions more for every 10,000 patient years.

Rather than the ludicrous generalizations that the authors make about premature infants brain development, and need for surveillance, these data (even if they were reliable, which I doubt) say exactly the opposite. as far as serious psychiatric illness is concerned there is almost no difference between premature infants and the full term.