188 twenty seconds or so. So far I have not found the manoeuvre to fail or the hiccoughs to return. I imagine that the pressure on the meatus stimulates the auricular branch of the vagus, thus blocking momentarily the path of stimuli coming from the diaphragm or adjacent viscera. "Dacres", Alfnston, Sussex, UK PHILIP READING Chlamydia trachomatis detection and non-invasive sampling methods SIR,-Several points Dr Patel and colleagues (May 11, p 1169) make cannot pass without comment. Patel et al do not regard testing of early morning urine samples as feasible and test first-catch urine samples from men, irrespective of how long urination has been withheld. It is strange that they do not compare first-catch urine efficiency with that of testing early morning samples. If sensitive methods of detection are used, it is true that testing an early morning urine has no advantage.’ Unfortunately, it is difficult to determine the sensitivity of their methods. Although they confirm the positive results obtained by the NovoNordisk ELISA by Syva MicroTrak, they did not confirm the negative results and have probably been missing chlamydiae. One way to assess this would be to determine the prevalence of Chlamydia trachomatis in men with non- gonococcal urethritis (NGU). Examination of the Sheffield data does not allow this, but if all the men with symptoms had NGU then 30% would have been chlamydia-positive-which is less, by about 20%, than would be expected with a sensitive method .2 An alternative explanation for the low prevalence rate, apart from lack of test sensitivity, is that swabs were not taken properly. No doubt the Sheffield group would argue that this does not matter because urine samples provide results that are better than those obtained by urethral swabbing (94% versus 72%). This may be true, but to infer that the results are in any way a vindication of the non-invasive approach when there is so much at fault would be wrong. For others, including ourselves who have used a sensitive method of testing a urethral swab, swabbing is at least as sensitive as urine sampling. Finally, the boldness with which the Sheffield workers assert that urine sampling in women is of little value suggests that they have come to this conclusion on the basis of urine examinations. That they have done so is not mentioned. Urethral carriage of C trachomatis in only 1-3% of 455 new female attenders who they examined suggests an insensitive test, poor swabbing, or both. In a continuing study of women attending St Mary’s Hospital, as many urethral smears as cervical smears have been chlamydia-positive by MicroTrak; moreover, a centrifuged urine deposit has occasionally been the only specimen positive by MicroTrak. We are, therefore, unable to support Patel and colleagues’ contention that the detection of chlamydiae in urine probably represents contamination of the sample. Division of Sexually Transmitted Diseases, Clinical Research Centre, Harrow, and Jefferiss Wing, St Mary’s Hospital, London W2 1NY, UK P. E. HAY P. HORNER B. J. THOMAS D. TAYLOR-ROBINSON Department of Clinical Microbiology, University College Hospital, London WC1 G. L. RIDGWAY 1. Thomas BJ, Gilchrist C, Hay PE, Taylor-Robinson D. Simplification of procedures used to test urine samples for Chlamydia trachomatis. J Clin Pathol 1991; 44: 374-75 2. Hay PE, Thomas BJ, Gilchrist C, et al The value of unne samples from men with non-gonococccal urethritis for the detection of Chlamydia trachomatis. Genitourin Med 1991; 67: 124-28. *** This letter has been shown to Dr Patel and his colleagues whose reply follows.-ED. L. SIR,-We are surprised that Dr Hay and his colleagues criticise our work when so many of our findings are in keeping with those of other independent workers. 1,2 Only recently Crowley et alz reported a comparison of urethral swabs with first-catch urines and their detection rates for Chlamydia trachomatis with enzyme immunoassay (EIA) were 65-5% and 91%, respectively. The results we presented were preliminary findings of a much larger investigation looking at many aspects of urine versus urethral testing. Our study was set up not to compare early morning urines with first-catch urine but to compare first-catch urines with urethral swabs, which we still believe to be the most practical clinical use for the test. Although symptoms are a poor discriminator for the presence of non-gonococcal urethritis (NGU), Hay et al attempt to determine the sensitivity of our methods by assuming that all men with symptoms have NGU. Not surprisingly they underestimate the probable sensitivity of our methods. There were 65 men with NGU of whom 25 (38%) had evidence of chlamydial infection. Before and during the study we have monitored the quality of sample collection. Direct immunofluorescence (DIF) on all urethral samples from new attenders yields only an extra 2% than do EIA tests. The value of screening women with both urethral and cervical tests has been contested by many workers using sensitive testing.3 Our routine practice of cleaning the vulva before taking urethral swabs and the availability of tests to all patients rather than their restriction to a further selected high-risk group explains our low urethral detection rate. Hay et al seem to conclude that we are calling for the abandonment of sensitive testing (they themselves use a mixture of PCR and DIF). Unfortunately such techniques are not available to most clinics striving to offer chlamydia testing to all their patients; the reality remains a choice between various EIA methods.’ It is in the context of such a choice that we feel a first-catch urine test offers an excellent alternative to urethral swabbing. Department of Genitourinary Medicine, Royal Hallamshire Hospital, Sheffield S10 2JF, UK R. PATEL G. R. KINGHORN Public Health Laboratory Service, Sheffield G. KUDESIA R. VAN HEGAN 1. Caul EO, Paul ID, Milne D, et al. Non-invasive sampling method for detecting Chlamydia trachomatis. Lancet 1988, ii 1246-47. 2. Crowley T, Milne D, Paul ID, et al. Non-invasive diagnosis of Chlamydia trachomatis in males-time for a change? BSSI chlamydia symptoms, April 19, 1991: 79. 3. Bradley MG, Hobson D, Lee N, et al. Chlamydial infections of the urethra in women Genitourin Med 1985; 61: 371-75. 4. Radcliffe KW, Rowen D, Mercey DE, et al. Survey of the management of Chlamydia trachomatis infection of the cervix Genitourin Med 1991, 67: 41-43. Genomic imprinting SiR,—Dr Lubinski and Professor Hall (May 25, p 1288) suggest that there is a mechanistic and possibly temporal relation between genomic imprinting (differential parental gene expression), monozygous twinning, and X-inactivation. This hypothesis rests upon discordant phenotypic expression of genetic disorders in monozygous twins, being more common in females. Numbers of published cases are small, however. It is unlikely discordance in phenotypic expression of genetic disorders in monozygous twins would by itself warrant publication since this would often be dismissed as being due to environmental effects. We suspect many such cases may remain unknown to the scientific community. This notion is exemplified by two pairs of monozygous 46,XX twins who have the devastating dementia and behavioural disorders characteristic of Rett syndrome. One set shows phenotypic concordance; in the other set one twin is severely affected, is of short stature (126 cm), and is wheelchair bound while the other is still mobile at age 26, of normal height (162 cm), and generally abler than expected in Rett syndrome. We have communicated these two cases to Lubinski and Hall, and suggest that it might be a good idea to create a register to accumulate data that will substantiate or refute their observation that discordance in phenotypic expression of genetic disorders is more common in monozygous female twins than in males. If there is a firm association between sex-chromosome constitution and phenotypic discordance in monozygous twins, how could this come about? We suggest differential spatial