groups in whom bed sharing is common before this practice is .... families had a greenhouse or an orchard. Of the .... hours work on call; the frequency of on call work is set by the ... understanding of how these doctors are required to work.
from the syndrome.5 Further work is needed on the physiological and behavioural aspects of the parent-infant interaction that occurs among groups in whom bed sharing is common before this practice is implicated in the pathogenesis of the sudden infant death syndrome. I SADAF FAROOQI GREGORY Y H LIP D GARETH BEEVERS
Department of Medicine, Dudley Road Hospital, Birmingham B18 7QH 1 Scragg R, Mitchell EA, Taylor BJ, Stewart AW, Ford RPK, Thompson JMD, et al. Bed sharing, smoking, and alcohol in the sudden infant death syndrome. BMJ 1993;307:1312-8. (20 November.) 2 Mitchell EA, Stewart AW, Scragg R, Ford RPK, Taylor BJ, Becroft DMO, et al. Ethnic differences in mortality from sudden infant death syndrome in New Zealand. BMJ 1993; 306:13-6. 3 Balarajan R, Raleigh VS, Botting B. Sudden infant death syndrome and postneonatal mortality in immigrants in England and Wales. BMJ 1989;298:716-20. 4 Farooqui S, Perry U, Beevers DG. Ethnic differences in infantrearing and their possible relationship to the incidence of sudden infant death syndrome (SIDS). Paediatric and Perinatal Epidemiology 1993;7:245-52. 5 Gantley M, Davies DP, Murcott A. Sudden infant death syndrome: links with infant care practices. BMY 1993;306: 16-20.
cohort were known to be chronic carriers of hepatitis B virus: all were positive for antibodies to hepatitis C virus but none was positive for hepatitis C virus RNA. Only four of the 37 patients who were not carriers of hepatitis B virus were negative for hepatitis C virus RNA. This difference was significant (P < 0-001, Fisher's exact test). Our data suggest that replication of hepatitis B virus inhibits replication of hepatitis C virus in coinfected chronic carriers of hepatitis B virus. This raises the possibility that inhibiting hepatitis B virus replication by treating such patients with interferon may result in reactivation of hepatitis C virus infection once the treatment is stopped. In conclusion, the virological outcome of coinfection with hepatitis B virus and hepatitis C virus may not be as straightforward as Mimms and colleagues suggest. It may depend on which virus infects a patient first as well as the relative doses of infection. WILIIAM L IRVING G DOLAN
T H LEIGH
1 O'Brien PMS. Helping women with premenstrual syndrome. BMJ 1993;307:1471-5. (4 December.) 2 American Psychiatric Association. The diagnostic and statistical manual, third edition, revised. Washington, DC: APA, 1987: 367-9.
Concurrent infection with hepatitis B and C viruses ED1TOR,-Larry Mimms and colleagues report an interaction between hepatitis B and C virus in patients who received both viruses through contaminated blood transfusions: coinfection with hepatitis C virus reduced the multiplication of hepatitis B virus.' Our experience with patients coinfected with these viruses is different.2 We tested serum samples from a cohort of 60 haemophilic subjects for markers of hepatitis C virus infection by second generation enzyme linked immunosorbent assay (ELISA) and recombinant immunoblot assay and for presence of hepatitis C virus RNA by reverse transcriptase polymerase chain reaction. Forty two of the serum samples tested were positive for antibodies to hepatitis C virus by ELISA, of which 35 were positive and seven were indeterminate (c22 band only) by recombinant immunoblot assay. RNA from hepatitis C virus was detected by the polymerase chain reaction in 33 of these 42 samples, including four of those that were indeterminate by recombinant immunoblot assay. Five of the
BMJ VOLUME 308
15JANuARY 1994
ELISA CALZOLARI
Istituto Genetica Medica, Universita, Ferrara, Italy PIER PAOLO MASTROIACOVO
Clinica Pediatrica, Universita Cattolica, Rome, Italy GRAZIA PETRELLI AMEDEO SPAGNOLO
Istituto Superiore di Sanita, Laboratorio Epidemiologia e Biostatistica, Rome, Italy ROMANO TENCONI
1 Mimms LT, Mosley JW, Hollinger FB, Aach RD, Stevens CE, Cunningham M, et al. Effect of concurrent acute infection with hepatitis C virus on acute hepatitis B virus infection. BMJ
1 Paduano M, Mc Ghie J, Boulton A. Mystery of babies with no eyes. Observer 1933 January 17:5. 2 Handysides S. Report of anophthalmia under scrutiny: BMY 1933;306:416. 3 Russel EBAW. Microphthalmos and environmental pollutants. BMJ 1993;306:790. 4 Dolk H, Elliot P. Evidence for "clusters of anophthalmia" is thin. BMY 1993;307:203. 5 Gilbert R "Clusters" of anophthalmia in Britain. BMJ 1993;307:340-1.
haemophilia. BryHaematol. 1993;85:611-2.
Tavistock Clinic, Tavistock Centre, London NW3 5BA
ANNA CALABRO
Clinica Pediatrica, Universita, Perugia, Italy
Dipartimento Pediatria, Universita, Padua, Italy
1993;307:1095-7. (30 October.)
EDITOR,-P M S O'Brien's review of the premenstrual syndrome is misleading in its discussion of late luteal phase dysphoric disorder.' O'Brien is incorrect in stating that this diagnosis "refers only to the mood disturbance of premenstrual syndrome"; the diagnostic criteria for the disorder listed in the Diagnostic and Statistical Manual of Mental Disorders, third edition, revised include physical and behavioural symptoms as well,2 giving a category essentially identical with what O'Brien terms "primary premenstmal syndrome."
FABRIZIO BLANCHI
Istituto Fisiologia Clinica del CNR, Dipartimento Epidemiologia e Biostatistica, 56124 Pisa, Italy
University Hospital, Queen's Medical Centre, Nottingham NG7 2UH
2 Hanley JP, Dolan G, Day S, Skidmore S, Iring WL. Interaction of hepatitis B and hepatitis B and hepatitis C virus infections in
Helping women with premenstrual syndrome
(48/1032) of the control parents (odds ratio-0 9 (95% confidence interval 0-23 to 2 55)). On the basis of these results, the association between use of benomyl and anophthalmia and microphthalmia seems unlikely.
Clusters of anophthalmia No link with benomyl in Italy ... EDIToR,-Since the Observer reported a possible link between apparent clusters of anophthalmia and microphthalmia in England and Wales and the pesticide benomyl,' several articles have commented on possible strategies to manage public and media concern, the methodology to identify clusters, the difficulty in assessing disease clusters with unreliable data, and the weaskness of retrospective studies.2-5 Analysis of cases routinely reported to the national congenital malformation surveillance scheme (CMSS) in England and Wales and to the European congenital malformations registries (EUROCAT) suggested that pesticides based on benomyl are unlikely to be major causes of anophthalmia and microphthalmia.45 In Italy the National Committee of the Congenital Malformations Registries also made a descriptive analysis of cases of anophthalmia and microphthalmia collected during 1986-90 and presented the results to the Ministry of Health. After exclusion of malformations associated with chromosomal and other genetic syndromes, 33 cases of anophthalmia and 78 cases of microphthalmia were reported in 940 615 births, giving prevalences of 0 35 and 0-83 per 10000 births respectively; results similar to those of the European registries.4 The prevalance of the two malformations, separately and together, was evaluated for 18 regions: interregional distribution was not significantly heterogeneous (x2- 18-61, df- 17, P-03), and in no region was the prevalence significantly different from the mean prevalence. A correlation analysis by region was then made between the prevalence of the malformations and use of benomyl based pesticides (derived from reports of the National Statistical Institute (ISTAT)). There was no significant correlation by region, nor when the regions were grouped by four levels of pesticide consumption. An analysis by parental occupation was performed on a subset of the data (768 005 births) with babies with preauricular tags acting as controls. Of the 95 parents of babies with anophthalmia and microphthalmia, 4-2% (two fathers and two mothers) had agricultural occupations compared with 4-7%
... or in Norway EDrrOR,-The relation between parental exposure to the fungicide benomyl and anophthalmia in children has received attention.' We have investigated the hypothesis that a maternal occupational activity that could result in exposure to benomyl during the first trimester is a risk factor for anophthalmia or microphthalmia. A Norwegian multiregister study offered the opportunity to explore this association. A national cohort of farming families has been established from all five agricultural or horticultural censuses during 1969-89. All 157 360 personal farm holders bom later than 1924 have been identified, and information concerning the activity at the farm at the time of the censuses has been gathered for the individual holding. By means of their national personal identification number, the spouse and children of the holders were identified in the central person register. The linking procedures have identified 248 696 farmers and spouses born during 1925-74 and 323359 children bom during 1952-91. Linkage with the medical birth registry of Norway gave information on the perinatal outcome of 192416 pregnancies during 1967-91. Benomyl was introduced in Norway in 1971. It is mainly used as a fungicide in greenhouses and orchards; it is also used to a limited extent during the spring in fields of vegetables and, since the mid-1970s, in the late autumn in grain fields. Its use has been limited in Norway: the maximum sale was 1682 kg active compound in 1980. We considered farm activities after 1970 that might result in exposure to benomyl at a relevant time: work in greenhouses throughout the year; in orchards for children conceived in April-June; with vegetables grown in fields for children conceived in JanuaryApril; or cultivating grain for children conceived in August-November. Eleven per cent (n=21 843) of the entire cohort of infants born in 1967-91 had potentially been exposed to benomyl according to these criteria. Four of the 192 416 children had been diagnosed as having anophthalmia or microphthalmia (International Classification of Diseases, eighth revision, codes 744.0-744.1), yielding a crude prevalence of 0 21/10 000 new-
205
born infants. In comparison, the crude prevalence in the total population born in 1981-91, according to the medical birth registry, was 0-43/10000 newborn infants (26 cases, population 602 301). In all four cases the infants had been singletons born alive at full term with birth weights over 3500g. One had additional birth defects. The four families lived at farms in different parts of Norway, with no temporal clustering of the pregnancies (conceptions between 1971 and 1988). None of the families had a greenhouse or an orchard. Of the 21843 infants potentially exposed to benomyl, one was diagnosed as having microphthalmia (0 46/ 10 000). This child came from a farm that grew field vegetables. Our data do not indicate that parental agricultural activity is a specific risk for anophthalmia in children. Even with this large cohort, comprising roughly a seventh of all births in Norway during the period studied, no firm conclusions concerning this rare birth defect can be drawn. The use of proxies as exposure variables adds to the uncertainty. Nevertheless, our data do not support the suggestion that benomyl is a risk factor. PETTER KRISTENSEN
National Institute of Occupational Health, PO Box 8149 Dep, N-0033 Oslo, Norway
programmes. We do not, however, know the proportion of pregnant women exposed to benzimidazole; it is likely to be low. Thus changes in rates of anophthalmia and microphthalmia over time would not be expected to be seen, even if benzimidazole was teratogenic. The large variations in the prevalences of the condition and in the proportions of morphological subtypes between the programmes suggest a non-uniform definition of anophthalmia and microphthalmia despite the specificity given in the protocol. Since England and Wales are at the extreme of this broad variation, this must be taken into consideration in further investigations, including case-control studies. EDUARDO E CASTILTA
Intemational Centre for Birth Defects, 00195 Rome, Italy 1 Handysides S. Reports of anophthalnia under scrutiny. BMJ 1993;306:416. 2 Gilbert R "Clusters" of anophthalmnia in Britain. BMJ 1993;307:340-1. 3 Intemational Clearinghouse for Birth Defects Monitoring Systems. Congenital malformations world-wide. Amsterdam: Elsevier, 1991. 4 International Clearinghouse for Birth Defects Monitoring Systems. 1984 annual report. Copenhagen: ICBDMS, 1986.
LORENTZ M IRGENS
Medical Birth Registry of Norway, University of Bergen,
Bergen, Norway 1 Gilbert R. "Clusters" of anophthalmia in Britain. BMY 1993; 307:340-1. (7 August.)
No further clues from global investigation EDrroR,-Newspaper reports of geographic clusters of anophthalmia and microphthalmia in England and Wales possibly related to maternal exposure to a benzimidazole pesticide (benomyl) prompted a news article' and an editorial.2 The International Clearinghouse for Birth Defects Monitoring Systems3 reacted by conducting a global investigation. Data were requested on cases of anophthalmia and microphthalmia in their isolated and associated (syndromic) forms, excluding confounding defects (blepharophimosis, microcornea, coloboma, cryptophthalmus, and "small eyes") and cases of chromosome trisomies. Reporting programmes also provided information on the availability of benzimidazoles and on any unusual frequency distribution of anophthalmia and microphthalmia. Further information on the 13 programmes contributing data has been published.3 Reported prevalences of anophthalmia and microphthalmia at birth varied widely among the programmes, with extreme values in England and Wales (93/4149283-022/10000) and Atlanta (178/696057-2-56/10000). The rates for the 11 other programmes varied around 1/10000 (1102/10666512=1-03). One quarter of the cases were reported as anophthalmia in every programme except that in England and Wales, where more than half of the cases were reported as anophthalmia. Multiple malformations occurred in 80% of the cases of anophthalmia and microphthalmia except in England and Wales, where the proportion was only 66%. A low prevalence of anophthalmia and microphthalmia in England and Wales compared with the other programmes was found previously by the clearinghouse and attributed to underascertainment.4 The present investigation failed to provide clues to either support or refute the hypothesis of an association between exposure to benzimidazole and anophthalmia and microphthalmia. Secular trends were stable, and no rumours or alarms were reported from any of the participating
206
Making the internal market work EDrroR,-I agree with one of Lucy Moore and Maureen Dalziel's conclusions with regard to making the internal market work-namely, that there is a need to create incentives in the market.' Many providers have risen to the challenges of the market and collectively effected great improvements in the delivery of health care, but there are, as the market is currently constituted, no incentives for maintaining success. This is an important flaw. For many clinicians and managers whose service or hospital has enjoyed success since the introduction of the market there are no rewards. If trusts were commercial organisations (and I am not suggesting that they should be) those that achieved success, measured in broad terms by treating more patients and achieving financial stability, would have little difficulty in attracting further investment to develop new services or expand existing ones. What is preventing successful trusts from building on their success is not only continuing tension over the question of whether the market should be managed or whether competition should be allowed to dominate but also a lack of acknowledgment that the principles underlining the introduction of the market have been only half developed. Until there are incentives and rewards for doing better and succeeding, I predict increasing resentment, cynicism, and lack of support for the original objectives associated with the introduction of the market. NEILGOODWIN
St Mary's Hospital,
LondonW2 INY 1 Moore L, Dalziel M. Making the intemal market work. a case for managed change. BMJ 1993;307:1270-2. (13 November.)
Working hours in the career grades EDrroR,-Linda Beecham notes that the NHS Management Executive does not believe that career grade doctors will be bound by the European Union's directive on the organisation of working time.' The management executive believes that consultants determine their own working patterns and would therefore be covered by one of the derogations written into the directive. We suggest that consultants' intensity of work is
determined by purchasers and the health of the local population. Also, for many consultants a large proportion of the hours worked is out of hours work on call; the frequency of on call work is set by the number of consultants participating in the rota. Therefore the working pattem is determined not by consultants but by their employers' trusts or directly managed units. The fact that the management executive considers that associate specialists determine their own working pattems shows a complete lack of understanding of how these doctors are required to work. We seriously doubt that the executive's statement would stand up to legal scrutiny. Article 17 derogation 2.1 (c) (i) concerns "services relating to the reception, treatment and/or care provided by hospitals or similar establishments." It allows derogation from the sections concerning daily rest, breaks, weekly rest periods, and length of night work. It does not allow derogation from a maximum working week of 48 hours. It also states that derogation can be given only "provided that the workers concerned are afforded equivalent periods of compensatory rest or that, in exceptional cases . . . workers concerned are afforded appropriate protection." The NHS Management Executive is correct to state that, under the final provisions of the directive (article 18 para 1(b)), any member state may allow any worker to exceed 48 hours (subject to health and safety considerations) so long as there has been prior agreement by the employee. This derogation applies only to the rule concerning the maximum weekly average of hours worked; sections concerning rest periods will still have to be adhered to. The directive also states that no employee can be "subjected to any detriment by his employer because he is not willing to . . . perform such work." In conclusion, this directive must affect career grade doctors. No doctor can be forced to work more than 48 hours a week against his or her wishes, and a doctor's willingness to do so could not legally be used to affect the decision to appoint him or her at interview. Also, due consideration must be given to rest periods during the working week. GRANTINGRAMS SHARON BINYON
Binringham B29 4LS 1 Beecham L Working time directive will not affect career
grades. BMJ 1993;307:1362. (20 November.)
Prescribing costs Fundholders had a head start EDITOR,-Jean Bradlow and Angela Coulter's findings on the effect of fundholding on general practitioners' prescribing costs' need to be interpreted with caution. Firstly, as the authors describe in their discussion, the indicative prescribing amounts of fundholding and nonfundholding practices were not derived in the same way. Fundholders had, and used, the opportunity to negotiate their indicative prescribing amounts upwards to "take full account of the needs and circumstances of each practice." This included additional allowances made for historically low prescribing costs. This opportunity was not available to non-funiholding practices. We need to know the extent of this upward revision before we can interpret the performance of practices relative to their indicative prescribing amount. Fundholding practice 4 and non-fundholding practice 10 (in tables V and VI) both showed 11% increases after inflation in net ingredient costs. The fundholding practice made a 10% saving on its prescribing amount while the non-fundholding practice overspent its prescribing amount by 10%.
BMJ VOLUME 308
15 jNuAY 1994
