Drug use and prisoners SIR,-Prevalence data show that in Belgium the proportion of drug users in prisons increased from 5 6% in 1982 to 9-7% in 1985 and to 11-9% in 1986.' From a prevalence study in 1982 it seems that 5 5% of prisoners had used drugs before being sentenced (298/5454) and that 0 1% started to use drugs while in prison (6/5454). Dr Anthony Maden and colleagues found that two thirds of the women in prison who were dependent on drugs reported having attended treatment services before entering prison.' Similarly, they found that of 189 men identified as drug dependent, 85 had had treatment before entering prison.' These data do not contradict our findings among a population of drug users who were interviewed while attending help agencies in Antwerp, Belgium.' Among the drug users surveyed over three months in 1989 (n= 126) a high proportion had a history of imprisonment (45%). Of those who had consulted help agencies more than once, half seemed to have been in prison. Surprisingly, of those who consulted only once for treatment, only a quarter had been in prison. Among the drug users interviewed during three months in 1990 (n=234) 30% (59/194) admitted to having been in prison, 55% once and 45% more than once. As with the data for 1989, 35% of those who consulted a help agency more than once had been in prison, compared with only 15% of those who consulted only once for treatment. As these differences could not be explained by age or sex we assume that the problems related to drug use and addiction can be considered as the main responsible factor. The role of the judicial services in dealing with the drug problem is therefore important. Our data confirm that even if a large proportion of the people intend to seek treatment when released this will be no guarantee of their staying out of prison. As Dr Maden and colleagues suggest,4 prisons have an important role in preventing and treating drug addiction problems. On the other hand, good cooperation among prison, social services, and treatment agencies is essential to help people after they have been released from prison. Drug treatment agencies should be able to start their programme in prison in order to continue afterwards when people have been released. PIERRE VAN DAMME GUIDO VAN HAL Department of Epidemiology and
MATTI CLAUWERS RUDI PEETERS
Community Medicine, University of Antwerp, 2610 Antwerp, Belgium 1 Peeters R. Prevalentie van druggebruik in Vlaanderen. [Prevalence of drug use in the Flanders.] Epideiologtsch Bulletin
1990;2:8-13. (In Flemish.) 2 Casselman J. Aantal druggebruikers in de gevangenis. [Number of drug users in prison.] Panopticon 1988;9:63-74. (In
Flemish.) 3 Maden A, Swinton M, Gunn J. Women in prison and use of illicit drugs before arrest. BMJ 1990;301:1133. 4 Maden A, Swinton M, Gunn J. Drug dependence in prisoners.
BMJ 1991;302:880. (13 April.) 5 Van Hal G, Van Damme P, Clauwers M, Eylenbosch W. Illicit drug use in the city of Antwerp, 1990. Scientiflc report of the National Fund for Scientific Research. Antwerp: Department of Epidemiology, University of Antwerp, 1990. (NFWO
3.0019.90.)
General practitioners' attitudes to drug users SIR,-I look after a young woman who has symptomatic HIV disease and is a drug user. I treat her HIV disease jointly with the local hospital. Her drug use is stable, and she has not used illicit drugs for almost a year. Because of her illness she was able to be rehoused in more appropriate accommodation. This meant leaving my list and having to find a new general
1464
practitioner, and I decided that I would try to transfer her care directly. I was shocked by the response I got from the general practitioners I spoke to. In the end I had to speak to 11 general practitioners before one would finally take her on, and even then the doctor agreed to take her on only in terms of her HIV disease and not her drug prescription. The responses of these general practitioners varied from a polite "No thank you" to "People like this ought to be shot." Not one was willing to listen to the advances that she had made; all seemed to be completely blinded by prejudice. If drug users are going to be the next wave of people affected by HIV what hope do they have of good care in general practice?
without unacceptable toxicity emerges, will then commence the parallel group studies required in the United States. A possible solution to Professor Levy's dilemma of recruiting patients to a parallel group study, in which half the patients will not receive the drug, is the Amery-Dony design.' This design, like all designs, has its own problems and it may be less suitable for Alzheimer's disease than for epilepsy. However, all the patients are exposed to the drug and the responders who have not also experienced unacceptable side effects later enter a double blind, placebo controlled, parallel group phase, which I have been led to believe is acceptable to the Food and Drug Administration.
CHRIS FORD
E H REYNOLDS
London NW6 6RY
Department of Neurology, King's College Hospital, London SE5 9RS
Tribulations for clinical trials
1
SIR,-Though I agree with Mr I M C Macintyre's views on clinical trials,' an often overlooked issue in conducting such trials is the importance of being able to recruit adequate numbers of subjects in a given trial centre. One of the problems in recruiting subjects may be (and in my experience has been) the number of trials with the same target population being conducted in the same centre. A suggestion for circumventing this problem would be to charge local ethics committees with the responsibility for assessing the number of trials that are feasible at any one time. In my opinion this is an important ethical issue, for if patients are entered into trials that wilf fail to recruit adequate numbers to yield meaningful results they are being asked to participate in pointless research and the researchers are wasting efforts and resources that might be better used elsewhere. P POWER-SMITH
Royal Hallamshire Hospital, Sheffield S 10 2JF 1 Macintyre IMC. Tribulations for clinical trials. BMJ 1991;302: 1099-100. (11 May.)
SIR,-Professor Raymond Levy's concerns about the probable influence of the American Food and Drug Administration's new guidelines on the design of trials of potential new pharmacological agents for Alzheimer's disease' have already been experienced for some years in the field of epilepsy, for which the Food and Drug Administration also demands parallel group studies. The new drug vigabatrin (Sabril; Merrell Dow) illustrates the point. The drug was first licensed in this country on 20 November 1989, the evidence for efficacy having been accepted by the Department of Health on the basis of six European double blind crossover trials2 and one European trial of Amery-Dony design,34 which includes a double blind parallel group phase. Since then, several other European countries have licensed the drug on the same evidence. By a remarkable coincidence, the advisory committee of the Food and Drug Administration met on 20 November 1989 to consider the further evaluation of vigabatrin in the United States. I was present at that meeting and it was my clear impression that a somewhat reluctant committee, concerned mostly about toxicity, eventually agreed to advise resumption of clinical trials in the United States under the pressure of the European experience, including the decision to license the drug in the United Kingdom. Parallel group trials of vigabatrin are now under way in the United States and, as in the cases of carbamazepine and valproate, it is likely to be several years behind Europe in marketing the drug. If the epilepsy experience is replicated in Alzheimer's disease Professor Levy may find that American companies will first undertake crossover trials in Europe and, when evidence of efficacy
Less' R.'Fribulatiotastforclinicaltrials.BM1,7l991;302:1271. 25
May.i 2 Reynolds EH. Vigabatrin. B,J 1990;300:277-8. 3 Amerv W, Dony J. A clinical trial design avoiding unduc placebo treatment.]J Clin IPharmacol 1975;15:674-9. 4 Rinig HA, Heller AJ, Farr IN, Reynolds EH. Vigabatrin: rational treatment for chronic epilepsy. J Neurol Neurosurg Psvchiatry 1990;53: 105 1-5.
SIR,-I share Mr I M C Macintyre's concern that too few patients are being entered into clinical trials and welcome his suggestions for increasing recruitment. It is of interest that, in the Scottish breast conservation trial, only four of 42 eligible patients (10%) were deterred by the random element.2 Most non-entrants (8/12; 75%) refused to receive the elective adjuvant chemotherapy that was prescribed for all patients with negative or low oestrogen receptor levels, regardless of randomisa-
nion. This suggests that most patients, if properly counselled, are willing to accept randomisation. Low entry rates may therefore often be caused by the reluctance of clinicians to consider patients for trials. As patients may fare better if treated in trials' would it not be appropriate to require that all patients eligible for trials are given the opportunity? Surely it is just as unethical to deny a patient the option of entering a trial from which they may benefit as it is to include a patient without informed consent. M N GAZE
Departmetit of Radiation Oncology, University of Glasgow, Glasgow G61 1BD 1 Macintyre IMC. Tribulations for clinical trials. BMJ 1991;302: 1099-100. (11 May.) 2 Jack WJL, Chetty U, Rodger A. Recruitment to a prospective breast conservation trial: why are so few patients randomised?
BMJ 1990;301:83-5. 3 Karjalainen S, Palva I. Do treatment protocols improse end results? A study of survisal of patients with multiple mycloma in Finland. BMJ 1989;299:1069-72.
Appraising journals' reviewing procedures SIR,-Drs F G R Fowkes and P M Fulton highlighted the difficult task of a reviewer in undertaking a critical appraisal of published research.' In line with this, we suggest ways in which peer review and editorial standards of papers submitted to medical journals might be improved. We believe that the time has come for medical journals to operate a double blind review procedure -that is, authors' names and affiliations should be removed from manuscripts sent for review and referees should remain anonymous. Authors ought to receive copies of the referees' full reports with the editor's decision and be given the chance to appeal against an editorial decision.
BMJ
VOLUME 302
15 JUNE 1991
We agree with Drs Fowkes and Fulton that most submitted articles are neither exceptionally good nor exceptionally poor as different aspects of a paper may vary in quality. As both authors and referees we have found rating scales on specific aspects of a paper (for example, theoretical contribution, methodological accuracy, recommendation for publication in the journal) together with free comments helpful and informative. As referees we have found the circulation of completed rating scales and comments among reviewers after the editorial decision informative and polite. If revision of the manuscript is recommended it is pleasing to be notified of the final outcome. This simple way of providing feedback to referees makes an often difficult and time consuming task worth while, increases motivation, and encourages conscientious appraisal. Publishing the names of reviewers in the journal is also a way of acknowledging their efforts. Others may be able to contribute further suggestions that provide an incentive for referees and a fair and instructive appraisal procedure for authors. E I. C ONG
D)epartnient of InfectioIs t)iseases,
JABBOTT Adult Cystic Fibrosis Unit, AMonsall Hospital, 10 Manchester M I5OWR I Fowkes FGR, Fulton I'M. Critical appraisal ot ptiblished
research: introductorv gtidlelines. B.I{J 1991;302:1136-40. (1I May.)
-*-We are well aware of the evidence that double blind reviewing of papers may result in a better outcome, but our judgment-and that of most other editors-is that the evidence is not yet solid enough to justify changing to such a system. The results of further studies will be published soon. We do feedback our final decisions to reviewers, but it might well be that we should feedback more information about how we reached our decisions. Finally, we pay our reviewers, but we are exploring further ways of rewarding them for what is a difficult, time consuming, and vitally important task. -ED, BMJ.
Benzodiazepines and tinnitus SIR,-In his discussion of the causes and management of tinnitus Mr M Hawthorne points out that benzodiazepines should not be used to treat tinnitus because their long term use is detrimental.' I should like to add to this that withdrawal of benzodiazepines after long term use is a well documented but often unrecognised cause of tinnitus.` It may be related to the hyperacusis characteristic of benzodiazepine withdrawal, and it usually disappears once withdrawal is completebut I have seen patients in whom tinnitus persisted for many months after withdrawal from benzodiazepines was complete, and similar cases have been reported.4 In one of the reported cases the tinnitus was relieved by diazepam. Tinnitus is common and has many causes, but questioning patients about their use of benzodiazepines may elicit a cause that usually has a good prognosis. LINDA BEELEY Queen Elizabcth Hospital, Birmingham B 15 2TH
SIR,-Dr D D Sandeman and colleagues' paper on microvascular dilatation in diabetic patients raises some important points.' We agree that an impairment of the hyperaemic response to injury may be important in the pathogenesis of foot ulceration. It is interesting that they found such an impairment in the feet of people with newly diagnosed non-insulin dependent diabetes. We are surprised, however, they did not find such an impairment in the group of patients with insulin dependent diabetes matched for age and sex, especially when they quote a previous study that showed such an impairment.2 This may be due to type II error resulting from the small numbers in the groups. There seems to be a trend for lower hyperaemic responses in this group of patients than in normal controls (median in normal controls= 1 65, median in insulin dependent diabetic patients= 145). Another explanation may lie in the selection of patients as patients with type I diabetes with complications (such as retinopathy and neuropathy) have impairment of the reactive hyperaemic response to injury while those without complications have not.' We presume that none of the patients had a history of foot ulceration, and we are not told the extent of neuropathy in the two groups of diabetic patients, so the patients with insulin dependent diabetes selected (median duration of diabetes of 19 years) may represent a group of diabetic patients who do not have neuropathy and have been spared microvascular and macrovascular complications. The same cannot be said of the patients with newly diagnosed non-insulin dependent diabetes. We have a further reservation concerning the stimulus given. We are not convinced that testing the hyperaemic response to heating assesses the hyperaemic response to injury. This is because this test does not distinguish between the neurogenic response to injury and the direct effect of the stimulus on the vessels. This has been investigated,' and diabetic patients with foot ulceration have been shown to have an impairment of the neurogenic response due to C fibre neuropathy but not of the direct response to injury. Possibly impairment of the neurogenic response is the important factor in the pathogenesis of such ulcers. We do not think that the criteria suggested by Dr Sandeman and colleagues can eliminate the presence of macrovascular disease. Ankle brachial indices may be abnormally high because of incompressibility of arteries in diabetic patients due to sclerosis and calcification, which are common in diabetes, and without further investigations such as wave form analysis, oscillotonometry,' and arteriography macrovascular disease in this group cannot be excluded. The presence of foot pulses also does not necessarily denote a disease free proximal arterial tree, especially as diabetic patients seem to be spared atherosclerotic changes in the foot vessels. Finally, we emphasise that foot ulceration in diabetic patients has a multifactorial pathogenesis. In addition to large vessel disease and microvascular impairment, abnormal areas of pressure on the feet, foot deformities, limited joint mobility, sensory neuropathy, minor trauma, and infections may play an important part in this debilitating
condition.
I Hawthornc Al. Ann quiestions. B.IJ 1991;302:1266. (25 May.) 2 Edwards (G, (Cantopher 1', Olivieri S. Benzodiazepine dependencc
and the problems of withdrawal. Postgrad Med j 1990; 66,suppl 2::S27-35. 3 Ashton H. Bcnzodiazcplnc withdrawal: an unftitished story.
BM7 1984;288: 1135-40. 4 Schweizer E, Rtickels K, Case WG, Greerblatt DJ. Longterm theraputic use of bcnzodiazepincs. II. Effccts of gradtual taper. Arch Gbeen Psvshtatrs 1990;47:908-15. 5 Busto U, Fornazzart L, Naranjo CA. Protracted tinnittis after discoiltinuationi of longterm therapeutic uise of benzo-
diazepines. 7 Cl/n Psst/copharntacol 1988;8:359-62.
BMJ VOLUME 302
Microvascular vasodilatation in non-insulin dependent diabetes
15
JUNE
1991
S K SHAMI S J CHITTENDEN S SARIN
L)eparimenit of Surgery, Utniversity College and Middlesex School of M\edicine, L onadon Wi1N 8AA 1 Sandeman DD, Pyrm CA, Green EM, Seamark C, Shore AC, rooke JE. Microvascular vasodilatation in fteet of newly diagnosed non-insulin dependent diabetic paticnts. BMJ 1991;302:1122-3. ('11 May.)
2 Rasman G, Villiams SA, Speicer PD, Smaje LH, Wise PH, Tooke JE. Impaired microsascular hyperaemic response to minor skin trauma in type I diabetes. BAIJ 1986;292:1295-8. 3 Walmslev D, Wales JK, W'iles PG. Reduced hvperaemia following skin trautna: evidence for an impaired microvascular response to injurv in the diabetic foot. Diabetologta 1989;32: 736-9. 4 Parkhouse N, Le Quesne PlM. Impaired neurogcnic vascuLlar response in patients with diabetes and neuropathic foot lesions. N Englj7 Med 1988;318:1306-9. 5 Tovey Fl, Karanjia N, Plant GR, Rees M. Oscillotonometry: a simple non-invasive method of differentiating proximal artcrial obstruction from distal microangiopathy in the vascuLlar assessmcnt of diabetics. Practical Dliabetes 1990l;7:203-5.
Perinatal bereavement SIR, - No one would disagree with the considerable advances that have been made by the profession in the management of perinatal death over the past 25 years described by Drs Stanford Bourne and Emanuel Lewis.' I would, however, like to draw attention to a major omission in their commentary -that of the role of the primary care team in this situation. For most women who experience either a stillbirth or a miscarriage (spontaneous or therapeutic, for whatever reason), the greater part of their continued care will be provided by their general practitioner and their health visitor-particularly the health visitor if there are already children in the family. Obstetricians, paediatricians, midwives, and staff on special care baby units can to a large extent only provide care for "the immediate painful reactions." Although I would fully agree that the immediate reaction and behaviour of the attending hospital staff are of paramount importance, the ensuing support of those professionals who have responsibility for the continued care of the whole family needs to be acknowledged. General practitioners need full and immediate communication from their hospital colleagues as to the circumstances of the loss and some details of what has already been said to the parents. As time passes it is members of the primary care team, rather than the obstetric team, who need to be attuned to the "danger signals" of unresolved mourning. To reinforce their role primary carers need to receive sufficient training in this area, for although tragedies such as perinatal death occur infrequently in any one practice, miscarriage is extremely common and general practitioners have been found to be lacking in their handling of parents at this time.2 The mourning process should be carefully monitored and the use of booklets such as that produced by the Stillbirth and Neonatal Death Society (SANDS) is a useful adjunct for both parents and carers. Finally, of course, it is essential that prompt referral to an expert is available for those in whom the grieving goes wrong. DEBORAH J SIHARP Departmcnt of General Ilractice, United Medical and Dental Schools of Guv's and St 'I'homas's Hospitals, Londott SE I1 6SP I
Bourtte S, Lewis 1167-8.
(18l5Mav.)E.
Perinatal
bereavement.
B3AIJ 1991;302:
2 Friedman T. Wotrncn's experience of gencral practitioner management of miscarriage. ,7 R Coil Gcit IPract 1989;39:456-8.
SIR,-The recent editorial on perinatal bereavement highlights the improvements in the care of parents who experience late loss of pregnancy and mentions the value of counselling after termination for fetal abnormality in the middle of pregnancy.' Regrettably, the authors also make a series of statements about early loss of pregnancy that are unsubstantiated and cannot be allowed to pass
unchallenged. Although spontaneous miscarriage is about 10 times commoner than perinatal death, the psychological impact of early pregnancy loss has received scant attention in clinical practice and 1465
MATTI CLAUWERS RUDI PEETERS
Community Medicine, University of Antwerp, 2610 Antwerp, Belgium 1 Peeters R. Prevalentie van druggebruik in Vlaanderen. [Prevalence of drug use in the Flanders.] Epideiologtsch Bulletin
1990;2:8-13. (In Flemish.) 2 Casselman J. Aantal druggebruikers in de gevangenis. [Number of drug users in prison.] Panopticon 1988;9:63-74. (In
Flemish.) 3 Maden A, Swinton M, Gunn J. Women in prison and use of illicit drugs before arrest. BMJ 1990;301:1133. 4 Maden A, Swinton M, Gunn J. Drug dependence in prisoners.
BMJ 1991;302:880. (13 April.) 5 Van Hal G, Van Damme P, Clauwers M, Eylenbosch W. Illicit drug use in the city of Antwerp, 1990. Scientiflc report of the National Fund for Scientific Research. Antwerp: Department of Epidemiology, University of Antwerp, 1990. (NFWO
3.0019.90.)
General practitioners' attitudes to drug users SIR,-I look after a young woman who has symptomatic HIV disease and is a drug user. I treat her HIV disease jointly with the local hospital. Her drug use is stable, and she has not used illicit drugs for almost a year. Because of her illness she was able to be rehoused in more appropriate accommodation. This meant leaving my list and having to find a new general
1464
practitioner, and I decided that I would try to transfer her care directly. I was shocked by the response I got from the general practitioners I spoke to. In the end I had to speak to 11 general practitioners before one would finally take her on, and even then the doctor agreed to take her on only in terms of her HIV disease and not her drug prescription. The responses of these general practitioners varied from a polite "No thank you" to "People like this ought to be shot." Not one was willing to listen to the advances that she had made; all seemed to be completely blinded by prejudice. If drug users are going to be the next wave of people affected by HIV what hope do they have of good care in general practice?
without unacceptable toxicity emerges, will then commence the parallel group studies required in the United States. A possible solution to Professor Levy's dilemma of recruiting patients to a parallel group study, in which half the patients will not receive the drug, is the Amery-Dony design.' This design, like all designs, has its own problems and it may be less suitable for Alzheimer's disease than for epilepsy. However, all the patients are exposed to the drug and the responders who have not also experienced unacceptable side effects later enter a double blind, placebo controlled, parallel group phase, which I have been led to believe is acceptable to the Food and Drug Administration.
CHRIS FORD
E H REYNOLDS
London NW6 6RY
Department of Neurology, King's College Hospital, London SE5 9RS
Tribulations for clinical trials
1
SIR,-Though I agree with Mr I M C Macintyre's views on clinical trials,' an often overlooked issue in conducting such trials is the importance of being able to recruit adequate numbers of subjects in a given trial centre. One of the problems in recruiting subjects may be (and in my experience has been) the number of trials with the same target population being conducted in the same centre. A suggestion for circumventing this problem would be to charge local ethics committees with the responsibility for assessing the number of trials that are feasible at any one time. In my opinion this is an important ethical issue, for if patients are entered into trials that wilf fail to recruit adequate numbers to yield meaningful results they are being asked to participate in pointless research and the researchers are wasting efforts and resources that might be better used elsewhere. P POWER-SMITH
Royal Hallamshire Hospital, Sheffield S 10 2JF 1 Macintyre IMC. Tribulations for clinical trials. BMJ 1991;302: 1099-100. (11 May.)
SIR,-Professor Raymond Levy's concerns about the probable influence of the American Food and Drug Administration's new guidelines on the design of trials of potential new pharmacological agents for Alzheimer's disease' have already been experienced for some years in the field of epilepsy, for which the Food and Drug Administration also demands parallel group studies. The new drug vigabatrin (Sabril; Merrell Dow) illustrates the point. The drug was first licensed in this country on 20 November 1989, the evidence for efficacy having been accepted by the Department of Health on the basis of six European double blind crossover trials2 and one European trial of Amery-Dony design,34 which includes a double blind parallel group phase. Since then, several other European countries have licensed the drug on the same evidence. By a remarkable coincidence, the advisory committee of the Food and Drug Administration met on 20 November 1989 to consider the further evaluation of vigabatrin in the United States. I was present at that meeting and it was my clear impression that a somewhat reluctant committee, concerned mostly about toxicity, eventually agreed to advise resumption of clinical trials in the United States under the pressure of the European experience, including the decision to license the drug in the United Kingdom. Parallel group trials of vigabatrin are now under way in the United States and, as in the cases of carbamazepine and valproate, it is likely to be several years behind Europe in marketing the drug. If the epilepsy experience is replicated in Alzheimer's disease Professor Levy may find that American companies will first undertake crossover trials in Europe and, when evidence of efficacy
Less' R.'Fribulatiotastforclinicaltrials.BM1,7l991;302:1271. 25
May.i 2 Reynolds EH. Vigabatrin. B,J 1990;300:277-8. 3 Amerv W, Dony J. A clinical trial design avoiding unduc placebo treatment.]J Clin IPharmacol 1975;15:674-9. 4 Rinig HA, Heller AJ, Farr IN, Reynolds EH. Vigabatrin: rational treatment for chronic epilepsy. J Neurol Neurosurg Psvchiatry 1990;53: 105 1-5.
SIR,-I share Mr I M C Macintyre's concern that too few patients are being entered into clinical trials and welcome his suggestions for increasing recruitment. It is of interest that, in the Scottish breast conservation trial, only four of 42 eligible patients (10%) were deterred by the random element.2 Most non-entrants (8/12; 75%) refused to receive the elective adjuvant chemotherapy that was prescribed for all patients with negative or low oestrogen receptor levels, regardless of randomisa-
nion. This suggests that most patients, if properly counselled, are willing to accept randomisation. Low entry rates may therefore often be caused by the reluctance of clinicians to consider patients for trials. As patients may fare better if treated in trials' would it not be appropriate to require that all patients eligible for trials are given the opportunity? Surely it is just as unethical to deny a patient the option of entering a trial from which they may benefit as it is to include a patient without informed consent. M N GAZE
Departmetit of Radiation Oncology, University of Glasgow, Glasgow G61 1BD 1 Macintyre IMC. Tribulations for clinical trials. BMJ 1991;302: 1099-100. (11 May.) 2 Jack WJL, Chetty U, Rodger A. Recruitment to a prospective breast conservation trial: why are so few patients randomised?
BMJ 1990;301:83-5. 3 Karjalainen S, Palva I. Do treatment protocols improse end results? A study of survisal of patients with multiple mycloma in Finland. BMJ 1989;299:1069-72.
Appraising journals' reviewing procedures SIR,-Drs F G R Fowkes and P M Fulton highlighted the difficult task of a reviewer in undertaking a critical appraisal of published research.' In line with this, we suggest ways in which peer review and editorial standards of papers submitted to medical journals might be improved. We believe that the time has come for medical journals to operate a double blind review procedure -that is, authors' names and affiliations should be removed from manuscripts sent for review and referees should remain anonymous. Authors ought to receive copies of the referees' full reports with the editor's decision and be given the chance to appeal against an editorial decision.
BMJ
VOLUME 302
15 JUNE 1991
We agree with Drs Fowkes and Fulton that most submitted articles are neither exceptionally good nor exceptionally poor as different aspects of a paper may vary in quality. As both authors and referees we have found rating scales on specific aspects of a paper (for example, theoretical contribution, methodological accuracy, recommendation for publication in the journal) together with free comments helpful and informative. As referees we have found the circulation of completed rating scales and comments among reviewers after the editorial decision informative and polite. If revision of the manuscript is recommended it is pleasing to be notified of the final outcome. This simple way of providing feedback to referees makes an often difficult and time consuming task worth while, increases motivation, and encourages conscientious appraisal. Publishing the names of reviewers in the journal is also a way of acknowledging their efforts. Others may be able to contribute further suggestions that provide an incentive for referees and a fair and instructive appraisal procedure for authors. E I. C ONG
D)epartnient of InfectioIs t)iseases,
JABBOTT Adult Cystic Fibrosis Unit, AMonsall Hospital, 10 Manchester M I5OWR I Fowkes FGR, Fulton I'M. Critical appraisal ot ptiblished
research: introductorv gtidlelines. B.I{J 1991;302:1136-40. (1I May.)
-*-We are well aware of the evidence that double blind reviewing of papers may result in a better outcome, but our judgment-and that of most other editors-is that the evidence is not yet solid enough to justify changing to such a system. The results of further studies will be published soon. We do feedback our final decisions to reviewers, but it might well be that we should feedback more information about how we reached our decisions. Finally, we pay our reviewers, but we are exploring further ways of rewarding them for what is a difficult, time consuming, and vitally important task. -ED, BMJ.
Benzodiazepines and tinnitus SIR,-In his discussion of the causes and management of tinnitus Mr M Hawthorne points out that benzodiazepines should not be used to treat tinnitus because their long term use is detrimental.' I should like to add to this that withdrawal of benzodiazepines after long term use is a well documented but often unrecognised cause of tinnitus.` It may be related to the hyperacusis characteristic of benzodiazepine withdrawal, and it usually disappears once withdrawal is completebut I have seen patients in whom tinnitus persisted for many months after withdrawal from benzodiazepines was complete, and similar cases have been reported.4 In one of the reported cases the tinnitus was relieved by diazepam. Tinnitus is common and has many causes, but questioning patients about their use of benzodiazepines may elicit a cause that usually has a good prognosis. LINDA BEELEY Queen Elizabcth Hospital, Birmingham B 15 2TH
SIR,-Dr D D Sandeman and colleagues' paper on microvascular dilatation in diabetic patients raises some important points.' We agree that an impairment of the hyperaemic response to injury may be important in the pathogenesis of foot ulceration. It is interesting that they found such an impairment in the feet of people with newly diagnosed non-insulin dependent diabetes. We are surprised, however, they did not find such an impairment in the group of patients with insulin dependent diabetes matched for age and sex, especially when they quote a previous study that showed such an impairment.2 This may be due to type II error resulting from the small numbers in the groups. There seems to be a trend for lower hyperaemic responses in this group of patients than in normal controls (median in normal controls= 1 65, median in insulin dependent diabetic patients= 145). Another explanation may lie in the selection of patients as patients with type I diabetes with complications (such as retinopathy and neuropathy) have impairment of the reactive hyperaemic response to injury while those without complications have not.' We presume that none of the patients had a history of foot ulceration, and we are not told the extent of neuropathy in the two groups of diabetic patients, so the patients with insulin dependent diabetes selected (median duration of diabetes of 19 years) may represent a group of diabetic patients who do not have neuropathy and have been spared microvascular and macrovascular complications. The same cannot be said of the patients with newly diagnosed non-insulin dependent diabetes. We have a further reservation concerning the stimulus given. We are not convinced that testing the hyperaemic response to heating assesses the hyperaemic response to injury. This is because this test does not distinguish between the neurogenic response to injury and the direct effect of the stimulus on the vessels. This has been investigated,' and diabetic patients with foot ulceration have been shown to have an impairment of the neurogenic response due to C fibre neuropathy but not of the direct response to injury. Possibly impairment of the neurogenic response is the important factor in the pathogenesis of such ulcers. We do not think that the criteria suggested by Dr Sandeman and colleagues can eliminate the presence of macrovascular disease. Ankle brachial indices may be abnormally high because of incompressibility of arteries in diabetic patients due to sclerosis and calcification, which are common in diabetes, and without further investigations such as wave form analysis, oscillotonometry,' and arteriography macrovascular disease in this group cannot be excluded. The presence of foot pulses also does not necessarily denote a disease free proximal arterial tree, especially as diabetic patients seem to be spared atherosclerotic changes in the foot vessels. Finally, we emphasise that foot ulceration in diabetic patients has a multifactorial pathogenesis. In addition to large vessel disease and microvascular impairment, abnormal areas of pressure on the feet, foot deformities, limited joint mobility, sensory neuropathy, minor trauma, and infections may play an important part in this debilitating
condition.
I Hawthornc Al. Ann quiestions. B.IJ 1991;302:1266. (25 May.) 2 Edwards (G, (Cantopher 1', Olivieri S. Benzodiazepine dependencc
and the problems of withdrawal. Postgrad Med j 1990; 66,suppl 2::S27-35. 3 Ashton H. Bcnzodiazcplnc withdrawal: an unftitished story.
BM7 1984;288: 1135-40. 4 Schweizer E, Rtickels K, Case WG, Greerblatt DJ. Longterm theraputic use of bcnzodiazepincs. II. Effccts of gradtual taper. Arch Gbeen Psvshtatrs 1990;47:908-15. 5 Busto U, Fornazzart L, Naranjo CA. Protracted tinnittis after discoiltinuationi of longterm therapeutic uise of benzo-
diazepines. 7 Cl/n Psst/copharntacol 1988;8:359-62.
BMJ VOLUME 302
Microvascular vasodilatation in non-insulin dependent diabetes
15
JUNE
1991
S K SHAMI S J CHITTENDEN S SARIN
L)eparimenit of Surgery, Utniversity College and Middlesex School of M\edicine, L onadon Wi1N 8AA 1 Sandeman DD, Pyrm CA, Green EM, Seamark C, Shore AC, rooke JE. Microvascular vasodilatation in fteet of newly diagnosed non-insulin dependent diabetic paticnts. BMJ 1991;302:1122-3. ('11 May.)
2 Rasman G, Villiams SA, Speicer PD, Smaje LH, Wise PH, Tooke JE. Impaired microsascular hyperaemic response to minor skin trauma in type I diabetes. BAIJ 1986;292:1295-8. 3 Walmslev D, Wales JK, W'iles PG. Reduced hvperaemia following skin trautna: evidence for an impaired microvascular response to injurv in the diabetic foot. Diabetologta 1989;32: 736-9. 4 Parkhouse N, Le Quesne PlM. Impaired neurogcnic vascuLlar response in patients with diabetes and neuropathic foot lesions. N Englj7 Med 1988;318:1306-9. 5 Tovey Fl, Karanjia N, Plant GR, Rees M. Oscillotonometry: a simple non-invasive method of differentiating proximal artcrial obstruction from distal microangiopathy in the vascuLlar assessmcnt of diabetics. Practical Dliabetes 1990l;7:203-5.
Perinatal bereavement SIR, - No one would disagree with the considerable advances that have been made by the profession in the management of perinatal death over the past 25 years described by Drs Stanford Bourne and Emanuel Lewis.' I would, however, like to draw attention to a major omission in their commentary -that of the role of the primary care team in this situation. For most women who experience either a stillbirth or a miscarriage (spontaneous or therapeutic, for whatever reason), the greater part of their continued care will be provided by their general practitioner and their health visitor-particularly the health visitor if there are already children in the family. Obstetricians, paediatricians, midwives, and staff on special care baby units can to a large extent only provide care for "the immediate painful reactions." Although I would fully agree that the immediate reaction and behaviour of the attending hospital staff are of paramount importance, the ensuing support of those professionals who have responsibility for the continued care of the whole family needs to be acknowledged. General practitioners need full and immediate communication from their hospital colleagues as to the circumstances of the loss and some details of what has already been said to the parents. As time passes it is members of the primary care team, rather than the obstetric team, who need to be attuned to the "danger signals" of unresolved mourning. To reinforce their role primary carers need to receive sufficient training in this area, for although tragedies such as perinatal death occur infrequently in any one practice, miscarriage is extremely common and general practitioners have been found to be lacking in their handling of parents at this time.2 The mourning process should be carefully monitored and the use of booklets such as that produced by the Stillbirth and Neonatal Death Society (SANDS) is a useful adjunct for both parents and carers. Finally, of course, it is essential that prompt referral to an expert is available for those in whom the grieving goes wrong. DEBORAH J SIHARP Departmcnt of General Ilractice, United Medical and Dental Schools of Guv's and St 'I'homas's Hospitals, Londott SE I1 6SP I
Bourtte S, Lewis 1167-8.
(18l5Mav.)E.
Perinatal
bereavement.
B3AIJ 1991;302:
2 Friedman T. Wotrncn's experience of gencral practitioner management of miscarriage. ,7 R Coil Gcit IPract 1989;39:456-8.
SIR,-The recent editorial on perinatal bereavement highlights the improvements in the care of parents who experience late loss of pregnancy and mentions the value of counselling after termination for fetal abnormality in the middle of pregnancy.' Regrettably, the authors also make a series of statements about early loss of pregnancy that are unsubstantiated and cannot be allowed to pass
unchallenged. Although spontaneous miscarriage is about 10 times commoner than perinatal death, the psychological impact of early pregnancy loss has received scant attention in clinical practice and 1465
