Editor,—Following their studies of whole gut lavage fluid, Smyth et al have suggested that a non-idiopathic intestinal inflammation occurs constituitively in patients with cystic fibrosis (CF), as a consequence of a proinflammatory effect of the patient's CFTR mutations.1They reported marginally elevated excretion of IgG, IgM, interleukin 1 (IL-1), neutrophil elastase, and eosinophil cationic protein, and much more significant increase in excretion of IL-8 and albumin, but no increase in excretion of α₁ antitrypsin or IgA. In this study where lavage fluid was administered continuously, and intestinal effluent was collected in discrete samples, pooling of the effluent before analysis would have allowed small differences in calculated inflammatory marker outputs to be interpreted as representative of gastroentestinal output. Of all the inflammatory markers presented, only IL-8 shows a range of cytokine outputs in CF patients with or without fibrosing colonopathy that did not extend into the range seen in controls, in these non-parametric datasets. The author's evidence for intestinal inflammation therefore relies heavily on the validity of their IL-8 Quantikine assay (R&D Minneapolis) protocol.

I have explored the validity of this assay for use in supernatants of faecal homogenates in children with cystic fibrosis and found it wanting. Recovery of a 500 pg/ml spike of IL-8 progressively increased from 41% in samples which were a 12-fold dilution of faeces to 189% in samples which were a 120 000-fold dilution of faeces, when used according to manufacturer's instructions. Prediluting the samples 50/50 in newborn calf serum, and using calf serum for further dilutions gave this assay (R&D catalogue no D8000) mean (SD) spike recovery of 92.1 (12.5)% and coefficients of variation of 3.46% (intra-assay) and 6.85% (interassay).2

Without knowledge of the IL-8 ELISA validation data of Smythet al, I assume that this assay returns similarly spuriously high IL-8 concentrations in polyethylene glycol based wholegut lavage fluid to my 120 000-fold dilution faecal supernatant. The absence of a significant difference between CF patients and controls in their α₁ antitrypsin outputs suggests that intestinal inflammation was not present in the CF patients. Overestimation of the WGLF IL-8 concentration would explain the apparently implausibly large volumes of swallowed sputum that the authors estimate would be required to account for their results. In this study which could not turn off the mucociliary escalator, but did dramatically increase the rate of intestinal transit and exclude exogenous pancreatic enzymes, swallowed sputum is the most likely explanation for the results.