Details of disease and treatment in 10 children with Wilms' tumour diagnosed in 1980-2 who were overtreated Prognosis Dose of Chemotherapy based on radiotherapy Case Clinical histological postoperatively Duration No stage appearance (cGy) Drugs given (months)P>O-1).

ticularly when the preparation contains povidone.Unfortunately, our patient died from complications of unnecessary treatment given because the unfamiliar molar units used to report the phenobarbitone concentration were misinterpreted.ADDENDUM-Since writing this report we have en- countered a further case of serious but eventually reversible respiratory failure in a 46 year old man after aspiration of Medicoal.I Anonymous.Repeated oral activated charcoal in acute poistning.

SIR,-Dr G Rayman and colleagues made some interesting observations on the local hyperaemic response to minor skin trauma (needle insertion and warming) in insulin dependent diabetes (17 May, p 1295). However, their literature search may have been incomplete, as they do not mention our reports on the local hyperaemic responses to needle insertion and superficial injection of small volumes of bland liquids in diabetic and nondiabetic subjects. 14 Dr Rayman's protocol was similar to our own, except that we used photoelectric plethysmography, a non-invasive optical technique which yields similar results to the laser Doppler flowmetry method used by the authors'; the two methods have been compared (1 March, p 620). Injection depth was fixed at 7-8 mm under the centre of the photoelectric plethysmography probe by inserting the needle along an angled track drilled through a perspex block cemented to the side of the probe. Injection of 0-1-0 15 ml of saline or insulin diluent in normal subjects (n=6, mean age 24 years) produced a local hyperaemic "flare," which reached a peak representing a 1600/o increase over baseline flow at 2 minutes and had largely faded by 30 minutes. Simple needle insertion without injection produced a response of similar amplitude and duration, suggesting that this acute, transient hyperaemia was a non-specific response to injury. In contrast, there was prolonged and considerable hyperaemia after injection of local vasodilators such as prostaglandin El,l 2nicotinicacid,2aprotinin,3andinsulin.. In stable insulin dependent diabetics (n=7, mean age 22 years, mean duration of diabetes 9years) the response was slightly but not significantly less than in normal subjects, but in severely brittle diabetics with chronic metabolic instability (n= 8, mean age 18 years, mean duration of diabetes 10 years) the hyperaemic flare was significantly flatter and shorter than in both other groups (figure). These findings, unlike those of Dr Rayman and colleagues, therefore suggested that impaired local hyperaemic responses to needle insertion were related to long term metabolic control rather than the presence of diabetes itself; differences in patient selection may explain these divergences. Other points should be mentioned. Firstly, as discussed (1 March, p 620), neither photoelectric plethysmography nor laser Doppler flowmetry can measure absolute units of blood flow, although