Responses
Other responses
Jump to comment:
- Published on: 22 March 2019
- Published on: 22 March 2019A more careful interpretation of meta-regression is required
I read with interest Dr Rao's commentary on Nath et al.,'s meta-analysis of trials of atraumatic and traditional lumbar puncture needles. This is a high quality paper which complies with PRISMA guidelines for the reporting of systematic reviews and meta-analyses and provides strong evidence for the use of atraumatic needles to reduce the incidence of postdural post puncture headache (PDPH).
However, Dr Rao is incorrect to state that the subgroup analysis of patients <18 years showed a significant difference in PDPH in this population. In fact the opposite is true as the p-value is >0.05 and the confidence interval for the RR spans 1. Instead Nath et al., show that having pre-specified age as a potential interactor/confounder there is no significant difference in the risk of PDPH for <18yr vs >18yr.
This is a subtle, but important distinction. First because it is possible the meta-regression was not adequately powered to detect a difference if one is present (a false negative). Second because age is a continous variable and so dichotomising in this way reduces statistical power to detect differences at different ages (e.g. there is a benefit in older children but not in younger children).
A more accurate interpretation of the study is that it shows that overall atraumatic lumbar puncture needles have lower risk of PDPH and that there is no evidence that this is not the case for patients under 18.
Conflict of Interest:
None declared.