CORRESPONDENCE
355
Correspondence Letters are published at the discretion of the Editor. Opinions e.rpressed by correspondents are not necessarily those of the Editor. Unduly long letters may be returned to the authors for shortening. Letters in response to a paper may be sent to the author of the paper so that the reply can be published in the same issue. Letters should be typed double spaced and should be signed by all authors personally. References should be given in the style specified in the Instruction to Authors at the front of the Journal.
MRI OF PARENCHYMAL LIVER DISEASE
SIR - We read with interest the recent paper 'Magnetic Resonance Imaging of Parenchymal Liver Disease: a Comparison with Ultrasound, Radionuclide Scintigraphy and X-ray Computed Tomography' (The Clinical NMR Group, 1987) but were surprised by the poor results using ultrasound in the diagnosis of cirrhosis and, in particular, fatty change. Several groups have now shown that ultrasound can detect fatty change in the liver with a sensitivity of up to 94% (Debongie et al., 1981; Saverymuttu et al., 1986). To achieve such results it is essential to use correct ultrasound criteria, in particular the presence of beam attenuation using the 5 MHz frequency (Saverymuttu et al., 1986; Taylor et al., 1986). It is unfortunate that the criteria for detection of fatty change used by the Aberdeen group with ultrasound were not stated. The use of the term hyper-reflectivity to describe an abnormal echo pattern is inadequate. Both fatty change and cirrhosis produce a hyper-reflective 'bright' echo pattern. However, fatty change can be identified by its characteristic tightly packed echoes compared with the coarse echoes of pure fibrosis. Fibrosis is more difficult to detect than fatty change. For example, in our recent series the sensitivity for fibrosis was only 57% although this increased to 71% in the subgroup with histologically established cirrhosis (Saverymuttu et al., 1986). In those instances where fibrosis leading to cirrhosis is present without fatty change an even higher detection rate is possible. It is therefore surprising that the Aberdeen group only found a sensitivity of 11% with ultrasound for the subgroup with established cirrhosis. Once again the ultrasound criteria for cirrhosis were not stated. In view of the atypical results for ultrasound in diagnosing parenchymal liver disease by the Aberdeen group it is misleading to include statements such as 'ultrasound is unreliable at detecting an increased in liver fat' without reference to other studies which have reached completely opposite conclusions regarding the accuracy of ultrasound in this area. S. SAVERYMUTI'U A. E. A. JOSEPH
Department of Radiology and Medicine St Georges Hospital Tooting London
References
Clinical NMR Group, Aberdeen (1987). Magnetic resonance imaging of parenchymal liver disease: a comparison with ultrasound, radionuclide scintigraphy and X-ray computed tomography. Clinical Radiology, 38, 495-502. Debongie, JC, Pauls, C, Fievez, M & Wibin, E (1981). Prospective evaluation of the diagnostic accuracy of liver ultrasonography. Gut, 22, 130-135. Saverymuttu, SH, Joseph, A E A & Maxwell, JD (1986). Ultrasound scanning in the detection of hepatic fibrosis and steatosis. British Medical Journal, 292, 13--15. Taylor, KJ, Ciely, CA, Hammers, L, Flax, Welting, G, Garcia-Tsao, G, Conn, H, Kuc, R & Barwick, KW (1986). Quantitative ultrasound attenuation in normal liver and in patients with diffuse liver disease. Importance of fat. Radiology, 160, 65--71.
SIR- In reply to the letter from the St George's Hospital Group concerning the ultrasound diagnosis of hepatic steatosis and hepatic fibrosis, I must point out that the ultrasound examinations of the patients included in our study were carried out prior to the publication of their own work in 1986. Since then, however, we have certainly looked at the differential beam attenuation using 3.5 and 5 MHz frequency transducers and we
agree that distinction between steatosis and hepatic fibrosis is possible in a number of patients. Attenuation of the 5 MHz frequency beam is also present in obese patients and this can cause significant difficulties in interpretation. A. P. BAYLISS
Department of Diagnostic Radiology Aberdeen Royal Infirmary Foresterhill Aberdeen
BARIUM ENEMAS ARE A HEADACHE SiR - We were surprised to read of the frequency of headache after
Picolax (sodium picosulphate) preparation for barium enema in the study by Kutt et al. (1988). We also require our outpatients to take two sachets of Picolax on the day prior to the barium enema examination in addition to a low residue diet. We specify that patients take at least halfa pint of clear fluid each hour from 9.00 to 20.00 and provide a section on the preparation instruction sheet to tick off these times as an aid to memory. The importance of adequate hydration has been previously emphasised (Miller, 1976). We questioned 41 consecutive patients attending for barium enema (29 female and 12 male, aged between 21 and 82 years with an average age of 57.4 years) about the symptoms they may have experienced following Picolax administration, specifically asking about headaches and thirst. Only three out of 41 patients (7.3%) experienced headache compared to 26.8% in the Bristol series and none complained of thirst. Our experience confirms that adequate rehydration can significantly reduce the incidence of headache. Our p~ients have found the specific guidelines in the preparation instruction sheet useful and simple to follow. J. A. SPENCER Department of Radiology D. J. NOLAN John Radcliffe Hospital Headington Oxford OX3 9DU REFERENCES Kutt, E, Hall, M J, Booth, A & Virjee, J (1988). Barium enemas are a headache. Clinical Radiology, 39, 9-10. Miller, RE (1976). The clean colon. Gastroenterology, 70, 289-290.
U L T R A S O U N D IN T H E FOLLOW-UP OF SPINALLY-INJURED
PATIENTS S m - We were very interested to read the article by Morcos and Thomas (1988) on the use of ultrasound in the follow-up of spinallyinjured patients. Over the past 3 years at Stoke Mandeville Hospital we have increasingly used ultrasound (US) and, to a decreasing extent isotope studies, for the evaluation of the renal tracts of our patients and we are in general agreement with the author's sentiments. The situation is, after all, rather analogous to the assessment of the patient with 'prostatic' symptoms, for whom it is now well established that a combination of plain abdominal radiographs (AXR) and US is as good as, if not actually superior to, intravenous urography (IVU). However, we feel that several comments are appropriate. In the spinally-injured patient, questions to be answered are: 1 Is there evidence of renal tract stones? 2 Is there upper tract and/or ureteric dilatation which may indicate either reflux or high bladder filling pressures wihout reflux? Urodynamics will not be able to demonstrate this.