1040 and decreased menstrual loss is
negligible. It appears, therethat to fore, contrary previous suggestions, tubal occlusion does not influence subsequent menstrual loss. Comparing patients with themselves before and after operation introduces the variable of time, but this seems preferable to accepting the many variables inherent in selecting an alternative control group. Counting vulval pads is a very crude method of measuring menstrual loss, but at least it is objective and it may be presumed that the standard of hygiene of the individual woman remains the same before and after operation. Department of Obstetrics and
Gynæcology, University of Liverpool.
BRIAN ALDERMAN
Secondly, cost is mentioned in relation to the use of the most efficient systems. My first action on being appointed a consultant anaesthetist was to ask for the use of piped-gas systems in our group of hospitals beginning in the areas of highest usage of medical gases. This "low priority" capital request was suddenly implemented after the incredibly short time of two years, and, since that date, the district has saved 3000 per annum on ansesthetic gas supplies, increasing, obviously, as inflation has taken its hold. Apart from giving me the moral satisfaction of knowing that I am saving the group my own salary, I feel that the 25-30% saving associated with the use of piped-gas systems gives leeway for specifying the best equipment in piped-gas systems, even if this results in a slight increase in capital cost. Lewisham Hospital, London SE13 6LH.
ARE BENZODIAZEPINES ANTIPSYCHOTIC AGENTS?
SIR,-Dr Trabucchi and
Dr Ba (Nov. 1, p. 868) claim that in the treatment of schizobe useful may
benzodiazepines phrenia. Their use of these substances is based on the fact that benzodiazepines mimic the effects ofG.A.B.A. and that baclofen (’Lioresal’), a G.A.B.A. derivative, has been found of some value in the therapy of schizophrenia. It is worth noting that benzodiazepines interact with a variety of other putative transmitter systems. For example, they are able to change the activity of brain catecholaminergicl2 and cholinergic3 neurones and to bind glycine receptors of synaptic membranes.4 On the other hand evidence based on in-vivo and in-vitro experiments denies that baclofen can act through G.A.B.A. mechanisms.5-9 In their open trial Dr Trabucchi and Dr Ba administered 10 mg of diazepam three times a day, even though diazepam is a
long-acting drug with a half-life of 20 (or more) hours.10 Frequent administration might produce accumulation of diazepam and its metabolites and is therefore seldom necessary unless considerable sedative effects are required. It would also be important to know which rating system these authors employed to evaluate the effect of their treatment and what maintenance dose of neuroleptics their patients received before the start of this trial. Thus the answer to the question "Are benzodiazepines antipsychotic agents?" might be "Benzodiazepines are of no particular value in the treatment of schizoSome schizophrenic patients, in fact, appear to dephrenia teriorate when chronically treated with chlordiazepoxide or
M. CUNDY
RADIOIMMUNOASSAY OF PLASMA—PROLACTIN
SIR,-We have been using the radioimmunoassay kit (V-L-S 2) for measuring plasma-prolactin in man, generously supplied for research purposes by the National Institute of Arthritis, Metabolism and Digestive Diseases (N.LA.M.D.D.), National Institutes of Health (N.I.H.), Bethesda, successfully for the past six months. It has come to our attention on the bush telegraph that several of our colleagues elsewhere have had difficulty setting up the assay using these reagents. We feel that modifications to the suggested N.I.H. assay protocol, which both increase its sensitivity and reliability, will be of interest to them, and will prevent needless frustration. Chloramine-T iodination of human prolactin (V-L-S 2) by the method outlined in the N.I.H. protocol produces a label which, in our experience, fails to give a usable standard curve at any antiserum dilution. By using the lactoperoxidase iodination procedure shown in the flow chart (see accompanying FLOW CHART FOR IODINATION AND PUR]F;CAT!ON OF H-PRL
To
20 :J of 0.01 M ammonium bicarbonate, pH 8.2, containing 2 ug h-PRL add
senally: 10 ul of sodium acetate buffer (0.4 M; pH 5.6) 5 ul of carner free Il2’ (Amersham IMS 30) containing 500 0 ofI"’. of 0003‘ v/v H,0, m acetate buffer (04 M; pH 10 ’Jl of acetate buffer (0.4 M ; pH 5.6) containing 1 ug lactoperoxidase
5.6)’
10 :JI
...
BSlgma). Mix.
diazepam".lo University Institute of Pharmacology, Viale G. B. Morgagni 65, 50134 Florence, Italy.
J.
After
exactly 10 seconds
at room
temperature, stop the
reaction
by addition
of
A. NISTRI
UNSAFE ANÆSTHETIC SUPPLY SYSTEMS
SIR,-An informative letter from Dr Dinnick and others (Oct. 11, p. 712) does dispel any feeling that the problems of installation might not be receiving adequate attention. However, there are still two factors which are of concern
piped-gas
500 vIof phosphate buffer (0.05 M; pH 7.4) The lOdlnated h-PRL is purified by barbitone buffer (0 06M; pH 861 elution from a ’Sephadex G-75’ column (0.9 x 10 cm) which has been saturated with 2 ml of 10‘o w/v bovine serum albumin (B.S.A.) in the same buffer. Fracuons of 500 yl volume are collected mto tubes containing 100 ul of 10’li w/v B.S.A. IS assessed bv The of the fractions covering the first peak of chromatoelectrophoresis and radioactivity scanning. Aliquots of the purest fraction are stored at -20°C and diluted, when required, to give 10 000 c.p m. (approxtmately 50 pg h-PRL) per 100 ul of reagent. The label remams usable for 3--4 weeks.
punty
radioactivity
to me.
Firstly, a system of banding can never be as effective as manufacturing the pipe in a clearly distinguishable colour, particularly when banding has been applied to existing pipes; one cannot guarantee that enough bands will be fitted to the pipe to make it clearly identifiable. 1.
K. M , Laverty, R. in The Benzodiazepines (edited by S. Garattini, Mussini, and L. O. Randall); p. 191. New York, 1973. Lidbrink, P., Corrodi, H., Fuxe, K., Olson, L. ibid. p. 203. Ladinsky, H., Consolo, S., Peri, G., Garattini, S. ibid. p. 241. Young, A. B., Zukin, S. R., Snyder, S. H. Proc. natn. Acad. Sci. U.S.A. 1974, 71, 2246. Curtis, D. R., Game, C. J. A., Johnston, G. A. R., McCulloch, R. M. Brain
Taylor, E.
2. 3. 4. 5.
Res.
1974, 70, 493.
Davies, J., Watkins, J. C., ibid. 1974, 70, 501. Davidoff, R. A., Sears, E. S. Neurology, Minneap. 1974, 24, 957. Nistri, A., Constanti, A. Experientia, 1975, 31, 64. Nistri, A. ibid. p. 1066. 10. Greenblatt, D. J., Shader, R. I. New Engl. J. Med. 1974, 291, 1011, 1239. 6. 7. 8. 9.
a sensitive and reproducible standard curve is obtained with a working antiserum dilution of 1 in 60 000. This is four times the suggested working dilution, and enables up to four times as many assays to be performed with the same amount of antiserum. The assay conditions adopted follow those set out in the method sheet supplied with the kit, with the following exceptions :
table),
(1) The rabbit anti-human prolactin antiserum is used at a working dilution of 1 in 60 000 (final dilution 1 in 300 000). (2) F.D.T.A. is omitted from all assay buffers except the second antibody diluent buffer. Heat-inactivated normal rabbit serum (1; v/v is included in the first antibody diluent buffer, and bound counts are separated by the addition of donkey anti-rabbit antiserum (Welcome Reagents Ltd.). Tubes are incubated at 4°C for three days and one day (second antibody incubation).
(main incubation)
1041 The detection limit for the method is approximately 0.1 ng/tube (0’5 ng/ml plasma) and, in our assay, one milliunit of M.R.C. human prolactin (h-PRL) Standard A (t): 71-222) produces equivalent displacement of h-PRL label as 34 ng of ’V-L-S 2’ h-PRL. We hope our colleagues will find these suggestions of practical use in setting up the method. Department of Biochemistry, University of Surrey, Guildford, Surrey, and St. Helier Hospital,
P. J. WOOD M. M. SHAHWAN V. MARKS.
Carshalton,
Surrey.
malignant disease if virus infection can be excluded or if the raised values persist" is open to serious question on two grounds. First, these assays have at best a considerable variability and, at least in our laboratory, they are not always precise. Secondly, it is our experience that a great number of patients without obvious viral infections or malignancies have very high C4 levels. In no way are we suggesting that Dr Bach-Mortensen’s data or our own are incorrect. On the contrary, they deserve more study, but from a critical and cautious angle. cate a
Department of Pediatrics, Upstate Medical Center, State University of New York, Syracuse, New York 13210, U.S.A.
INCOMPLETE C4 CONSUMPTION IN LEUKÆMIA
SIR,-The letter by Dr Bach-Mortensen and his associates (Sept. 13, p. 499) is of interest to us. For two years we have been studying the role of complement in childhood acute lymphoblastic leukaemia (A.L.L.). Our findings are similar in several respects to those presented in the letter, but there are important differences. We have not measured C1-esterase inhibitor in our patients but we have found that, at diagnosis or in relapse, addition of immune precipitates to the patient’s serum results in incomplete consumption of C4.’ Thus, in 14 of 20 children with A.L.L., incubation
of serum with
a
B.S.A.-goat-IgG/anti-B.S.A.
immune precipitate made at equivalence results in hmmolytic-C4 consumption which is less than 3 standard deviations below the mean for consumption in our age-matched control sera. This incomplete C4 consumption slowly disappears during induction therapy with vincristine and steroid, so that by remission nearly 75% of these children have a normal immuneprecipitate assay. From Dr Bach-Mortensen’s data, we would interpret this failure to completely consume C4 at diagnosis or relapse as being secondary to elevated serum-levels of Clesterase
inhibitor.
It is of interest that
ha:molytic-C4 titres were not high in our patients diagnosis relapse, whereas levels of C3 and properdin factor B, measured by radial immunodiffusion,2 were clearly elevated at these times. In addition, properdin and properdin-convertase values were also normal at diagnosis or relapse. The mean C4 values were equal to the mean for our control group and in only 2 patients were the levels of C4 greater than 3 standard deviations above the mean. By contrast, 60% of all sera examined at diagnosis or relapse have C3 at
or
and factor-B levels above 3 S.D. above the mean. What is most provocative in our data, however, is that during induction therapy there is a striking fall in serum C4, C3, and factor-B concentrations, so that 60% of these levels are below 3 S.D. from the normal mean. These levels return to normal in most patients in remission. Whether or not this precipitous drop in C4, C3, and factor-B concentrations represents invivo consumption of complement during clearance of tumour cells is not clear. Stastistical analysis of the paired values for C4 and C3, as opposed to factor B and C3, indicates that the C3 concentration is clearly related to the level of factor-B, rather than to that of C4 (r
negligible. It appears, therethat to fore, contrary previous suggestions, tubal occlusion does not influence subsequent menstrual loss. Comparing patients with themselves before and after operation introduces the variable of time, but this seems preferable to accepting the many variables inherent in selecting an alternative control group. Counting vulval pads is a very crude method of measuring menstrual loss, but at least it is objective and it may be presumed that the standard of hygiene of the individual woman remains the same before and after operation. Department of Obstetrics and
Gynæcology, University of Liverpool.
BRIAN ALDERMAN
Secondly, cost is mentioned in relation to the use of the most efficient systems. My first action on being appointed a consultant anaesthetist was to ask for the use of piped-gas systems in our group of hospitals beginning in the areas of highest usage of medical gases. This "low priority" capital request was suddenly implemented after the incredibly short time of two years, and, since that date, the district has saved 3000 per annum on ansesthetic gas supplies, increasing, obviously, as inflation has taken its hold. Apart from giving me the moral satisfaction of knowing that I am saving the group my own salary, I feel that the 25-30% saving associated with the use of piped-gas systems gives leeway for specifying the best equipment in piped-gas systems, even if this results in a slight increase in capital cost. Lewisham Hospital, London SE13 6LH.
ARE BENZODIAZEPINES ANTIPSYCHOTIC AGENTS?
SIR,-Dr Trabucchi and
Dr Ba (Nov. 1, p. 868) claim that in the treatment of schizobe useful may
benzodiazepines phrenia. Their use of these substances is based on the fact that benzodiazepines mimic the effects ofG.A.B.A. and that baclofen (’Lioresal’), a G.A.B.A. derivative, has been found of some value in the therapy of schizophrenia. It is worth noting that benzodiazepines interact with a variety of other putative transmitter systems. For example, they are able to change the activity of brain catecholaminergicl2 and cholinergic3 neurones and to bind glycine receptors of synaptic membranes.4 On the other hand evidence based on in-vivo and in-vitro experiments denies that baclofen can act through G.A.B.A. mechanisms.5-9 In their open trial Dr Trabucchi and Dr Ba administered 10 mg of diazepam three times a day, even though diazepam is a
long-acting drug with a half-life of 20 (or more) hours.10 Frequent administration might produce accumulation of diazepam and its metabolites and is therefore seldom necessary unless considerable sedative effects are required. It would also be important to know which rating system these authors employed to evaluate the effect of their treatment and what maintenance dose of neuroleptics their patients received before the start of this trial. Thus the answer to the question "Are benzodiazepines antipsychotic agents?" might be "Benzodiazepines are of no particular value in the treatment of schizoSome schizophrenic patients, in fact, appear to dephrenia teriorate when chronically treated with chlordiazepoxide or
M. CUNDY
RADIOIMMUNOASSAY OF PLASMA—PROLACTIN
SIR,-We have been using the radioimmunoassay kit (V-L-S 2) for measuring plasma-prolactin in man, generously supplied for research purposes by the National Institute of Arthritis, Metabolism and Digestive Diseases (N.LA.M.D.D.), National Institutes of Health (N.I.H.), Bethesda, successfully for the past six months. It has come to our attention on the bush telegraph that several of our colleagues elsewhere have had difficulty setting up the assay using these reagents. We feel that modifications to the suggested N.I.H. assay protocol, which both increase its sensitivity and reliability, will be of interest to them, and will prevent needless frustration. Chloramine-T iodination of human prolactin (V-L-S 2) by the method outlined in the N.I.H. protocol produces a label which, in our experience, fails to give a usable standard curve at any antiserum dilution. By using the lactoperoxidase iodination procedure shown in the flow chart (see accompanying FLOW CHART FOR IODINATION AND PUR]F;CAT!ON OF H-PRL
To
20 :J of 0.01 M ammonium bicarbonate, pH 8.2, containing 2 ug h-PRL add
senally: 10 ul of sodium acetate buffer (0.4 M; pH 5.6) 5 ul of carner free Il2’ (Amersham IMS 30) containing 500 0 ofI"’. of 0003‘ v/v H,0, m acetate buffer (04 M; pH 10 ’Jl of acetate buffer (0.4 M ; pH 5.6) containing 1 ug lactoperoxidase
5.6)’
10 :JI
...
BSlgma). Mix.
diazepam".lo University Institute of Pharmacology, Viale G. B. Morgagni 65, 50134 Florence, Italy.
J.
After
exactly 10 seconds
at room
temperature, stop the
reaction
by addition
of
A. NISTRI
UNSAFE ANÆSTHETIC SUPPLY SYSTEMS
SIR,-An informative letter from Dr Dinnick and others (Oct. 11, p. 712) does dispel any feeling that the problems of installation might not be receiving adequate attention. However, there are still two factors which are of concern
piped-gas
500 vIof phosphate buffer (0.05 M; pH 7.4) The lOdlnated h-PRL is purified by barbitone buffer (0 06M; pH 861 elution from a ’Sephadex G-75’ column (0.9 x 10 cm) which has been saturated with 2 ml of 10‘o w/v bovine serum albumin (B.S.A.) in the same buffer. Fracuons of 500 yl volume are collected mto tubes containing 100 ul of 10’li w/v B.S.A. IS assessed bv The of the fractions covering the first peak of chromatoelectrophoresis and radioactivity scanning. Aliquots of the purest fraction are stored at -20°C and diluted, when required, to give 10 000 c.p m. (approxtmately 50 pg h-PRL) per 100 ul of reagent. The label remams usable for 3--4 weeks.
punty
radioactivity
to me.
Firstly, a system of banding can never be as effective as manufacturing the pipe in a clearly distinguishable colour, particularly when banding has been applied to existing pipes; one cannot guarantee that enough bands will be fitted to the pipe to make it clearly identifiable. 1.
K. M , Laverty, R. in The Benzodiazepines (edited by S. Garattini, Mussini, and L. O. Randall); p. 191. New York, 1973. Lidbrink, P., Corrodi, H., Fuxe, K., Olson, L. ibid. p. 203. Ladinsky, H., Consolo, S., Peri, G., Garattini, S. ibid. p. 241. Young, A. B., Zukin, S. R., Snyder, S. H. Proc. natn. Acad. Sci. U.S.A. 1974, 71, 2246. Curtis, D. R., Game, C. J. A., Johnston, G. A. R., McCulloch, R. M. Brain
Taylor, E.
2. 3. 4. 5.
Res.
1974, 70, 493.
Davies, J., Watkins, J. C., ibid. 1974, 70, 501. Davidoff, R. A., Sears, E. S. Neurology, Minneap. 1974, 24, 957. Nistri, A., Constanti, A. Experientia, 1975, 31, 64. Nistri, A. ibid. p. 1066. 10. Greenblatt, D. J., Shader, R. I. New Engl. J. Med. 1974, 291, 1011, 1239. 6. 7. 8. 9.
a sensitive and reproducible standard curve is obtained with a working antiserum dilution of 1 in 60 000. This is four times the suggested working dilution, and enables up to four times as many assays to be performed with the same amount of antiserum. The assay conditions adopted follow those set out in the method sheet supplied with the kit, with the following exceptions :
table),
(1) The rabbit anti-human prolactin antiserum is used at a working dilution of 1 in 60 000 (final dilution 1 in 300 000). (2) F.D.T.A. is omitted from all assay buffers except the second antibody diluent buffer. Heat-inactivated normal rabbit serum (1; v/v is included in the first antibody diluent buffer, and bound counts are separated by the addition of donkey anti-rabbit antiserum (Welcome Reagents Ltd.). Tubes are incubated at 4°C for three days and one day (second antibody incubation).
(main incubation)
1041 The detection limit for the method is approximately 0.1 ng/tube (0’5 ng/ml plasma) and, in our assay, one milliunit of M.R.C. human prolactin (h-PRL) Standard A (t): 71-222) produces equivalent displacement of h-PRL label as 34 ng of ’V-L-S 2’ h-PRL. We hope our colleagues will find these suggestions of practical use in setting up the method. Department of Biochemistry, University of Surrey, Guildford, Surrey, and St. Helier Hospital,
P. J. WOOD M. M. SHAHWAN V. MARKS.
Carshalton,
Surrey.
malignant disease if virus infection can be excluded or if the raised values persist" is open to serious question on two grounds. First, these assays have at best a considerable variability and, at least in our laboratory, they are not always precise. Secondly, it is our experience that a great number of patients without obvious viral infections or malignancies have very high C4 levels. In no way are we suggesting that Dr Bach-Mortensen’s data or our own are incorrect. On the contrary, they deserve more study, but from a critical and cautious angle. cate a
Department of Pediatrics, Upstate Medical Center, State University of New York, Syracuse, New York 13210, U.S.A.
INCOMPLETE C4 CONSUMPTION IN LEUKÆMIA
SIR,-The letter by Dr Bach-Mortensen and his associates (Sept. 13, p. 499) is of interest to us. For two years we have been studying the role of complement in childhood acute lymphoblastic leukaemia (A.L.L.). Our findings are similar in several respects to those presented in the letter, but there are important differences. We have not measured C1-esterase inhibitor in our patients but we have found that, at diagnosis or in relapse, addition of immune precipitates to the patient’s serum results in incomplete consumption of C4.’ Thus, in 14 of 20 children with A.L.L., incubation
of serum with
a
B.S.A.-goat-IgG/anti-B.S.A.
immune precipitate made at equivalence results in hmmolytic-C4 consumption which is less than 3 standard deviations below the mean for consumption in our age-matched control sera. This incomplete C4 consumption slowly disappears during induction therapy with vincristine and steroid, so that by remission nearly 75% of these children have a normal immuneprecipitate assay. From Dr Bach-Mortensen’s data, we would interpret this failure to completely consume C4 at diagnosis or relapse as being secondary to elevated serum-levels of Clesterase
inhibitor.
It is of interest that
ha:molytic-C4 titres were not high in our patients diagnosis relapse, whereas levels of C3 and properdin factor B, measured by radial immunodiffusion,2 were clearly elevated at these times. In addition, properdin and properdin-convertase values were also normal at diagnosis or relapse. The mean C4 values were equal to the mean for our control group and in only 2 patients were the levels of C4 greater than 3 standard deviations above the mean. By contrast, 60% of all sera examined at diagnosis or relapse have C3 at
or
and factor-B levels above 3 S.D. above the mean. What is most provocative in our data, however, is that during induction therapy there is a striking fall in serum C4, C3, and factor-B concentrations, so that 60% of these levels are below 3 S.D. from the normal mean. These levels return to normal in most patients in remission. Whether or not this precipitous drop in C4, C3, and factor-B concentrations represents invivo consumption of complement during clearance of tumour cells is not clear. Stastistical analysis of the paired values for C4 and C3, as opposed to factor B and C3, indicates that the C3 concentration is clearly related to the level of factor-B, rather than to that of C4 (r
