RADIAL KERATOTOMY

RADIAL KERATOTOMY

501 prediction rate for euthyroid subjects for total T3 and total and free T4. On the other hand, 27% (total T3), 33% (free T4), and 53% (total T4) o...

161KB Sizes 1 Downloads 122 Views

501

prediction rate for euthyroid subjects for total T3 and total and free T4. On the other hand, 27% (total T3), 33% (free T4), and 53% (total T4) of patients with hyperthyroidism were misclassified. These misclassification-rates fell to 20-27%, if three experienced doctors used the results of total T3, total T4, and free T4 simultaneously. However, a substantial reduction in the rate of false negatives, down to

7%,

was

possible only by incorporating

recent

advances

in

mathematical decision theory.3,4 The results of Caldwell et al should be amended to take into account the principle of external

validity. Medizinisches Strahleninstitut,

University of Tübingen, D-7400 Tubingen, West Germany

C. F. HESS

CF, Brodda K, von Schulthess GK, Thürlimann B, Wurstbauer K. On the significance of serum thyroid hormones for the diagnosis of hyperthyroidism: Considerable improvement of the physicians’ classifications by means of mathematical decision theory. Nucl Med (in press). 2. Lincoln EM. Statistical concepts fundamental to investigations. N Engl J Med 1985; 312: 890-97 3. Hess CF, Brodda K. On the optimum choice of categories for the classification of biomedical data patterns. Meth Inform Med 1979; 18: 222-28 4. Hess CF, Brodda K. On the optimum selection of diagnostic tests in computer-aided differential diagnosis. Comp Biomed Res 1984; 17: 389-98. 1 Hess

DISTINGUISHING BETWEEN RENAL ALLOGRAFT

REJECTION AND CYCLOSPORIN NEPHROTOXICITY SIR,-Dr Taube et al (July 27, p 171) describe the distinct histological pictures of acute transplant rejection and cyclosporininduced damage which are associated with similar impairments of renal allograft function. They conclude that making the distinction between these two conditions is important and difficult but that a kidney biopsy is often helpful in doing so. Although a biopsy is quite safe, it is nonetheless invasive and expensive. Staining the urinary sediment with methyl-green/pyronin and quantifying the number of pyronin positive ("turned on") mononuclear cells permits the early diagnosis of acute allograft rejection. 14 of 16 episodes of biopsy-proven acute rejection were diagnosed by the use of this method. This prospective study was done before the availability of cyclosporin so that it is not known whether examination of the urinary sediment is useful for differentiating cyclosporin damage from acute rejection episodes. However, the histopathological pictures which Taube et al describe lead me to believe that different urinary sediment findings might be expected. The large amount of mononuclear cell infiltration that is seen in renal biopsy specimens obtained during acute rejection episodes is consistent with the lymphocyturia previously described. The absence of this process with cyclosporin-induced renal damage might be expected to be reflected by a much less active urinary sediment. I would suggest that the diagnostic value of this simple test be compared with renal biopsy. easy and

Section of Medical Oncology, School of Medicine,

University of Minnesota, Box 286 Mayo Memorial Building, 420 Delaware Street SE, Minneapolis, Minnesota 55455, USA 1

WILLIAM J. M. HRUSHESKY

Hrushesky WJM, et al Lymphocyturia in human renal allograft rejection. Arch Surg 1972, 105: 424-26

POSSIBLE ROLE OF INSULIN DEFICIENCY AND GROWTH HORMONE IN SEVERE RETINOPATHY

SIR,-We read with interest the study of severe retinopathy in diabetics with apparently mild glucose intolerance by Dr Barnes et al (June 29, p 1465). This study highlights the difficulty in accurately matching non-insulin-dependent diabetics. Failure to match the patient groups invalidates comparisons. Retinopathy is commonly accepted as being associated with poor

long-term control-the only metabolic information provided in this study relates to the period of hospital follow-up, although the patients may have had hyperglycaemia for a long time before presentation with visual impairment. The group without retinopathy was matched for age, weight, and duration of treatment. We are

not

told how they presented and there is no reason to suspect

that the duration of glucose intolerance was similar to that of those with retinopathy. Since the glycaemic profiles for the two groups of diabetics were significantly different the actual duration of diabetes may also be different, which reinforces the suggestion that the groups studied were dissimilar. In view of the possibility, therefore, that those with retinopathy had had diabetes for longer than those without, it cannot be concluded that hypoinsulinaemia and growth hormone excess are important in the aetiology and progression of small-vessel disease in patients with non-insulin-dependent diabetes. When studying non-insulin-dependent diabetics, we agree that comparative data on duration of known diabetes should be given. For this metabolic study, however, diabetics without retinopathy should also have been matched for mean diurnal glucose concentration. Matching patients accurately takes time; biochemical variables have to be measured by different assays and interassay variability has to be asssessed. We wondered whether the controls were investigated at the same time as the patients with retinopathy. Baines et al’s study should be regarded only as an observation of metabolic derangement in a small group of patients with proliferative retinopathy which may be related to diabetes. Department of Medicine, East Birmingham Hospital, Bordesley Green East, Birmingham B9 5ST

P. E. JENNINGS O. ODUGBESAN J. FETCHER A. H. BARNETT

RADIAL KERATOTOMY

SIR,-As a corneal surgeon, I would add my whole-hearted agreement to the July 27 editorial by my senior colleague (instantly recognised by his literary style). Although results of radial keratotomy after a year may appear favourable (Ophthalmology 1985; 92: 177-98), the results after four years are disconcertingly bad (Arch Ophthalmol 1985; 103: 782-84). Surgeons and patients alike would be wise to await for the longer term results of prospective studies that are underway before embarking upon this largely cosmetic procedure. St. Thomas’ Hospital, London SE1 7EH

M. G. FALCON

DISAPPEARING COLLAGEN ANTIBODIES IN RHEUMATOID ARTHRITIS

SiR,-Serum antibodies to collagen, particularly collagen type II, have been described in rheumatoid arthritis. However, the frequency has varied between 3% and 71%, and depends on the assay used. In rodent models of arthritis, immunisation against type II collagen induces specific serum antibodies to this collagen2,3 and the arthritis can be transferred passively to other animals by such serum.In these models, antibody appears to be responsible for the primary arthritic lesion. However, in human rheumatoid arthritis the time and nature of the initial pathogenic insult is unknown and may predate the onset of symptoms by months or years. Using a conventional ELISA assay on serial samples from patients very early in their arthritic disease, we have found significant levels of collagen antibody at first clinical presentation in some cases, but antibodies are often no longer detectable by the time a firm diagnosis of rheumatoid arthritis is made and are thereafter persistently

negative. obtained from patients who first presented to a rheumatological clinic with undiagnosed arthritic symptoms. Up to five serial samples were obtained at follow-up of up to 2 years. Sera from 24 cases in whom definite or classical rheumatoid arthritis was proven, from 24 cases of inflammatory (non-rheumatoid) arthritis, and from 18 normal age and sex-matched controls were studied. The clinical status of most of these patients has already been described.5 Sera were stored at -20°C. Sera were tested at 1:500 dilution for IgG antibodies to native and denatured collagen types I-V by an ELISA method." Samples which gave a reading higher than 2 SD above the arithmetic mean of 18 normal control sera were considered to have raised antibody levels to that collagen type.

Blood

was