Standardization of the One-stage Prothrombin Time for the Control of Anticoagulant Therapy

Transcription

1 8 1 Janary 19 Mental Illness in Adolescence-Henderson et al. MEDICAL BRIMISH JOURNAL this way. The ability of parents and friends to perceive distress in adolescents appears to vary with social class. At a time when there has been mch revised legislation dealing with the management of yong offenders it is srprising that only si teenagers were referred by the cort for psychiatric assessment. The setting p of family corts as recommended in the Kilbrandon Report (19) and the establishment in Edinbrgh of a nit specifically for adolescent psychiatric illness shold improve the sitation. Smmary A srvey is reported of referrals of adolescents to psychiatrists in Edinbrgh. The annal referral rate was 5. per thosand tat risk. For one-third of the series the consltation took place after an act of deliberate self-poisoning or self-injry; almost cne-half were referred by their family doctor and one-tenth were self-referrals. The social and emotional difficlties arising from the natral hesitance of adolescents to discss their problems with adlts are discssed. We are gratefl to Professor G. M. Carstairs for his help and advice ; to Drs. Norman Kreitman and John Evans for their criticism and advice ; and to the physician-sperintendents and consltant psychiatrists in the region. REFERENCES Backett, E. M., Shaw, L. A., and Evans, J. C. G. (1953). Proc. roy. Soc. Med.,,. Hare, E. H., and Shaw, G. K. (195). Mental Health on a New Hosing Estate. London. Kessel, W. I. N. (19). Brit.. prev. soc. Med., 1, 1. Kilbrandon Committee (19). Children and Yong Persons, Scotland. Scottish Home and Health Department, Scottish Edcation Department. H.M.S.O., London. Philip, A. E., and McClloch, J. W. (19). Brit.. prev. soc. Med.,, 1. Siegel, S. (195). Nonparametric Statistics for the Behavioral Sciences. New York. Srole, L., Langner, T. S., Michael, S. T., Opler, M. K., and Rennie, T. A. C. (19). Mental Health in the Metropolis: The Mid-town Manhattan Stdy, vol. 1. New York. Taylor, Lord, and Chave, S. (19). Mental Health and Environment. London. Toolan, J. M. (19). Amer.. Psychia'., 118, 19. Warren, W. (195).. Child Psychol.,, 1. Standardiation of the One-stage Prothrombin Time for the Control of Anticoaglant Therapy Brit. med. J., 19, 1, 8-88 ROSEMARY BIGGS,* M.D.; K. W. E. DENSON,* D.PHIL. The one-stage prothrombin time test is widely sed for the control of anticoaglant therapy and has proved on the whole satisfactory for this prpose. One difficlty is that the reslts are epressed in many different ways, and even if this is done apparently in the same way the reslts of one centre are seldom directly comparable to those of another. Ths if a patient moves from one centre to another there may be a change in the degree of anticoaglation. Another difficlty is that clinical trials carried ot at more than one hospital may be difficlt nless special steps are taken to standardie the method. In recent years efforts have been made to solve this problem, and co-ordination shold greatly improve the sitation. In the United States the College of American Pathologists has reviewed the techniqes in se and made some recommendations abot acceptable techniqes (Miale and Lafond, 193). The Committee for Haemostasis and Thrombosis is devising a method of comparing reslts based on a standard thromboplastin (Biggs, 195 ; Biggs and Denson, 19). In England Poller (19) has been stdying the preparation of a reagent to be sed in a wide area, so that all the hospitals shall have a nified method. These stdies have been made from rather different points of view and are complementary to each other rather than competing, and conseqently a mch more constrctive approach shold be possible. Coaglation Defect cased by Comarin Drgs The comarin drgs case variable redction in the blood concentrations of clotting factors VII, IX, X, and prothrombin (II). The one-stage prothrombin time records abnormality in factors II, VII, and X, and ths in most patients gives a fair * Medical Research Concil, Blood Coaglation Research Unit, Chrchill Hospital, Oford. record of the abnormality. Since the defect is composite and it is impracticable to measre all of the factors separately, an overall assessment of the abnormality is made from the reslts of the test. Possibility of Standardiation The details of techniqe will not be discssed, bt in all of the methods the abnormality is recorded as a lengthening of clotting-time of the patient's plasma in comparison with normal. Table I gives eamples of reslts obtained with different methods. One difficlty in comparing one method with another is clearly the different clotting-times of normal plasma -the reslts in Table I show a range from 13. to 5 seconds. This difficlty is avoided if the clotting-time ratios are calclated (Table I), becase all of the normal clotting-times are epressed as a clotting-time ratio of 1. It will be seen from Table I that if the ratios are taken as an epression of the degree of abnormality, then the methods vary a good deal in this TABLE I.-One-stage Prothrombin Time was Carried Ot on a Normal Plasma Sample and on Three Samples from s Receiving Anticoaglant Therapy. Si Different One-stage Techniqes were Used. The Clotting-times Obtained are Epressed in Seconds: the Clotting-timne ratio is calclated by Dividing the Clotting-time for the Abnormal Sample by that of the Normal (e.g., -5 = 1.5) Method 1 Method 1 Method 3 Method Method 5 Method Samnple C. - R. C. R. C. R. C. R. C. R. C. R. Normal C. = Clotting-time. R. Ratio. Br Med J: first pblished as 1.113/bmj on 1 Janary 19. Downloaded from on October 18 by gest. Protected by copyright.

2 Anticoaglant respect. This is a phenomenon well known to coaglation workers and is one of the main reasons why it is difficlt to compare one method with another. If comparative tests are carried ot on a large nmber of samples from patients on anticoaglant therapy by different methods a certain reglarity is disclosed. If the clotting-time ratios obtained by two methods are plotted one against the other a straight line can be drawn throgh the points (Fig. 1). From the reslts shown in Fig. 1 it wold be reasonable to say that on average a clotting-time ratio of by method 1 was eqivalent to one of. by the second method and that a ratio of 3 was eqivalent to one of 3.5, and so on. This process of comparison has been made on a very large nmber of samples (Biggs, 195 ; Biggs and Denson, 19) and the relation has been fond to hold good by varios methods and in many different laboratories. This reglarity observed in comparing techniqes means that it shold be possible to calibrate all methods in terms of a standard. The problem is to establish and define the standard. Epression of Reslts Most of the commercial methods provide a calibration crve which converts the reslts to "percentage of normality." These diltion crves do not ensre any niformity between methods of assessing the degree of abnormality, becase the sensitivity of the methods to the comarin defect is not trly represented by these diltion crves. The crves are made by testing the clotting-times by the one-stage techniqe of diltions of normal plasma, and they vary very considerably according to the dilent sed, to the natre of the thromboplastin, and to other details of techniqe. To many the diltion crve has seemed 1 _- _ 1_- A - -I_ Or -- - s 1. s It * ~ sch, Therapy-Biggs and Denson clottirag-time ratios have been preferred. Ths it may be the half times the normal clotting-time this will represent qite object ive to maintain a therapetic range of two to two and a different activities if the diltion crves are taken to represent half tiimcs the normal clotting-time, the tre abnormality. The difference is delineated by the horiratio method has the advantage that it does not carry ontal lines in Fig., which incldes the range two to two and The any Einjstified implication of scientific validity. Bt the a half times the clotting-time. If, on the other hand, the 1 Janary rati BRITISH MEDICAL JOURNAL 85 diltion crve and ratio methods of calclating the reslts are nlikely to give the same information. This can be visalied by reference to Fig., which presents hypothetical diltion crves for two modifications of techniqe. If the clinician oso FIG. 1.- receivin g anticoaglant therapy with si different test methods and a standard method. The diltion crve is appropriate to a particlar clottingthe - - I 8 1 PERCENTAGE ACTIVITY FIG..-Two hypothetical diltion crves for the one-stage prothrombin-time method. The horiontal lines show the limits of the " therapetic range " if the criterion sed is two to two and a half times the clotting-time. The vertical lines show the range if the criterion is 1 to % of activity Unless the diltion crves for the methods are identical the two criteria mst give a different degree of abnormality. an arbitrary scale ot measrement that the se ot simple decides to maintain a therapetic range of two to two and a clinician decides to se a percentage range, say 1 to % of normality, then clearly the two methods will differ in the ratios -- / ~ 5 eqivalent to this range. -'/ / These dilemmas abot the epression of the reslts of the one-stage prothrombin time XX 5-s are well known, and it is towards a satis / factory soltion to them that the efforts of I-,X standardiation are directed. The aim shold ] X ttx be to make a realistic assessment of the 3 diltion crve for a particlar method, - X becase if the clotting-time means anything / X (as it certainly does) then it is the amont X,,, residal activity which is important The diltion crve is ths the central point 5, in this stdy. In 199 Biggs and Macfarlane - / / - Y. pointed ot that the diltion crve cold be epressed linearly by plotting the clotting- / time against the reciprocal of the concentra- j /, 5 -oe 5 - m/ 5 :tion of wvhole plasma in the dilted mitres. /N/ / /More recently it has also been shown (Biggs, fi ~ X 'X sx diltion crve can be epressed linearly. XXX X /X X 195) that the clotting-time ratios for the / 31 / 3 XXt 3 X This point is illstrated by making this I I mathematical nmber cf conversion for a 9- ft - - -diltion crves (Fig. 3). This method of epressing the reslts has the advantage the normal is epressed as nity by that all FOR VARIOUS TECHNIQUES methods. This mathematical conversion does -One-stage prothrombin-time investigation was carried ot on samples from patients not of corse solve the problem of which p-time ratios were calclated, and for each of the methods the ratio was plotted against method for the control of anticoaglant o obtained with the standard method. The solid circles represent the ratios likely to give the best average reslts in comparing the two methods. therapy. Br Med J: first pblished as 1.113/bmj on 1 Janary 19. Downloaded from on October 18 by gest. Protected by copyright.

3 8 1 Janary 19 It is. however. posible to pt together the reslts of correlation stflies, sing different methods, with this linear method of epressing the diltion crve. Let s take, for eample, the reslts in Fig. 1, which shows correlation graphs of si different methods and a standard method. From these graphs an average eqivalence between the methods can be worked ot. FXor methods 1 and 3 of Fig. 1 these eqivalent ratios are given in Table II. If a diltion crve is drawn for the standard method, then, on the basis of the figres in Table II, diltion crves cr -. ~~~~~~~~~~~~.o FIG. 3.-Diltion crves for a variety of methods are presented by plotting the reciprocal of normal plasma concentration against the clotting-time ratio. The lowest crve was obtained with hman brain etract by dilting normal plasma with plasma from a patient with very severe vitamin-k deficiency. The other crves are derived from crves issed with commercial scale reagents. It shold be noted that these calibration graphs bear no relation to the sensitivity of the methods to the defect cased by comarin drgs. can be drawn for methods 1 and 3 (Fig. ). It will be seen that if the standard method is epressed as a straight line then se of the correlation points from Fig. 1 prodces straight-line graphs for methods 1 and 3. If the diltion crve appropriate to each method is sed the clotting-times for the individal samples sed for Fig. 1 may be converted to percentage of normality. The reslts of this TABLB II.-Ratios Eqivalent to the " Standard " Method of Fig. 1 are Taken from the Solid Circles of Fig. 1 Ratio by Standard Method Ratio by Method 1 Ratio by Method 3 a- C: _ S ~~- Anticoaglant Thcrapy-Biggs alnd Denson,',, ' METHOD a - "" STANDARD A" ok- X a-,/ ----omethod 3 '.' - J Ot-- I 8 1 I 1 lb 18 FaG. -Calibration graphs for the standard method and methods 1 and 3 from Fig A BRITISI MEDICAL JOURNI conversion are given in Table III. The correspondence between the reslts of the methods sing appropriate diltion crves is obvios. This particlar eample incldes a nmber of very long clotting-times and was chosen for this reason to illstrate the correlation graphs ; ths many of the " percentage " reslts are low. This sort of correlation has in fact been made with very many samples over a wide range of clotting-times and seems to hold generally (Biggs, 195). TABLE III.-Percentage Reslts for Two Different One-stage Methods Compared with Those of an Arbitrarily Selected Standard Method Using the Diltion Crves of Fig. to Ascertain the Percentages Sample Standard <5 19 <5 <5 Method 1 Method ) * < -5 <5 < 5 < 5 Since the standard is to be sed merely as a basis for nified comparison, this method of rendering the reslts of one techniqe on a scale adjstable to all can form the basis of the standardiing procedre ; it remains to select a standard. Fig. shows that the methods differ from each other in the steepness of the line epressing the diltion crve. It is easy to see that the steeper the line the more " sensitive " the method to the defect cased by the comarin drgs. The redctio ad absrdmn wold be a line along the base showing no alteration in clotting-time with treatment. Ths a satisfactory method to se as a standard is one with a good sensitivity and a steep line. For general prposes a method with a reasonably short clotting-time for normal plasma wold probably be most generally acceptable. Standard Preparation A standard preparation based on a good deal of preliminary work can now be proposed. This is a reagent containing absorbed o plasma and sing hman brain as a sorce of thromboplastin. This method is provided with an arbitrary diltion crve relating clotting-time ratios with percentage of normality, which has been fond in practice to cover the most sally employed degrees of anticoaglation. Sch a crve based on ratios is shown in Fig. 5. This can probably more seflly be presented as a Table (Table IV). j CI 3, X X I - 8 \ \-/ 1_-X- - / CONCENTRATION-PER CENT FIG. 5. Diltion crve for the standard method. Br Med J: first pblished as 1.113/bmj on 1 Janary 19. Downloaded from on October 18 by gest. Protected by copyright.

4 1 Janary 19 Anticoaglant Therapy-Biggs and Denson TABLE IV.-Clotttng Time Ratios Eqivalent to Varios Arbitrary Percentages for the Standard Method Clotting- % Clotting- Clotting- % time Ratio time Ratio time Ratio Use of the Standard Preparation in Practice The standard preparation shold be sed in parallel with the method preferred in a particlar laboratory on a nmber of normal samples and samples from patients treated with the comarin anticoaglant in se. An eample of sch a comparative stdy is presented in Table V. The figres for the TABLE V.-Twenty-one Plasma Samples from s Receiving Anticoaglant Therapy were Tested by Two Methods, One of Which was Taken as the Standard. The Clotting-time Ratios and Percentages from Appropriate Calibration Tables are Given. These Figres are Presented in Figs. and Sample Standard Method Ratio I ' r- _E Method to be Tested Ratio 1t i O ig two methods are plotted against each other and the best line to fit the points is drawn (Fig. ). The ratios of the two methods are then compared. If the ratio of by the standard method is taken as a fied point, then by the second method it can be seen that a ratio of 1.9 wold indicate the same degree of abnormality. The appropriate diltion crve for the method can ths be fied, or perhaps more conveniently a table of ratio eqivalents to percentage may be sed. Table VI presents a composite set of figres which cold be sed for all methods. At the top of this table are given the figres for the ratio eqivalent to that of the standard ratio of. In the eample BRITISH MEDICAL JOURNAI 8 given this was 1.9. Ths for this method the colmn headed 1.9 wold be sed. The figres for conversion to percentages appear in Table V. In Fig. the percentage reslts for the two methods have been plotted ; clearly, if the therapetic range is taken to be 1 to % both methods give a satisfactory assessment of the samples. Or P- (_ or 5I i/ // A o 3 5 CLOTTING-TIME RATIO FIG..-The one-stage method was carried ot on a nmber of samples from patients having anticoaglant therapy, a "standard" method and a test method being sed, and the clotting-time ratios for the two methods are plotted against each other. Commercial Scale / Preparations The principle of the method of standardiation depends qite simply on the comparison of the method in qestion with a reference preparation of known sensitivity. The two methods are sed on a nmber of samples from patients receiving anticoaglant therapy and the reslts compared as in Table V and Fig.. The method cold very well be sed to standardie commercial thromboplastins, which wold then be issed with an approved scale. Minimm Standard of Sensitivity A preparation and method which is insensitive to the defect cased by comarin drgs can never give satisfactory reslts. At present the majority of preparations are calibrated more by their reaction to saline diltions of normal plasma than by their sensitivity to the defect cased by comarin drgs. Ths some TABLE VI-.Calibration Table for Use with the One-stage Prothrombin-time Methods of Varying Sensitivity to the Comarin Defect Ratio Eqivalent to Standard Ratio of ! j ! * I I I Br Med J: first pblished as 1.113/bmj on 1 Janary 19. Downloaded from on October 18 by gest. Protected by copyright.

5 88 1 Janary 19 Anticoaglant Therapy-Biggs and Denson BMEDIsALHJOURNA commercial preparations are very insensitive. When the first reference preparation is available it wold be an advance for a minimm acceptable sensitivity to be established by standards instittions. Sch a concept wold encorage the preparation of satisfactory commercial preparations. LU 3 1 I a 1 3 TEST METHOD PERCENTAGE FIG..-Reslts presented in Fig. converted to percentage activity from appropriate calibration scales and the reslts plotted against each other. The vertical and horiontal lines give a reasonable " therapetic range." The limiting lines show the range of error to be epected from the method as a reslt of simple inaccracy in determining the clotting-time. Conclsion The work done in many centres on the standardiation of the one-stage prothrombin time for the control of anticoaglant therapy with comarin drgs is reaching a stage when it may be possible greatly to improve the general standard of control. The emphasis on technical care in carrying ot the test, always noted by Qick and more recently nderlined by Miale, is clearly important. The se of the same techniqe in a wide Brit. med.., 19, 1, 88-9 a Dependence on Detromoramide area also obviosly has great advantages (Poller, 19). The introdction of a reference preparation with an internationally recognied scale and a simple method of adjsting varios techniqes to the scale wold rond off this work. The calibration scale of commercial preparations cold be adjsted to the scale, preparations made individally in clinical laboratories cold be similarly calibrated. Moreover, attention to sensitivity to the comarin defect wold ndobtedly improve the reliability of control. A single scale properly applied at different centres wold ensre safety and niform dosage for a patient moving from one place to another and wold greatly improve the standard of clinical trials carried ot at more than one centre. Smmary The control of anticoaglant therapy by the one-stage prothrombin time sffers from the fact that the reslts of this test are reported differently at different centres. Ths clinical trials with a standard objective of treatment are difficlt to organie and a patient cannot travel from one centre to another with any assrance that the level of therapy will remain constant. A method of standardiing the test has been devised so that all may se the same scale of measrement adjsted in each case to fit the preferred thromboplastin. The method is based on the preparation of the standard thromboplastin with which other preparations can be compared. The work reported in this paper was ndertaken jointly on behalf of the M.R.C. Working Party on Anticoaglant Therapy and the International Committee on Haemostasis and Thrombosis. REFERENCES Biggs, R. (195). Transactions of the Conference Held nder the Aspices of the International Committee on Blood Clotting Factors, 19, pp and Denson, K. W. E. (19). Transactions of the Conference Held nder the Aspices of the International Committee on Haemostasis and Thrombosis, and Macfarlane, R. G. (199).. clin. Path.,, 33. Poller, L. (19). Brit. med..,, 55. Miale, J. B., and Lafond, D. J. (193). Prothrombin Srvey, College of American Pathologists. Standards Committee: Sbcommittee on Coaglation. R. SEYMOUR-SHOVE,* M.B., CH.B., D.P.M.; C. W. M. WILSON,t M.D., B.SC., PH.D. Detromoramide (Palfim) has been available as a possible sbstitte for morphine since 195. Several acconts presenting evidence of dependence on this drg have been pblished (Calesnick, 1959 ; Temple and Attisso, 191 ; Flavell Matts, 19 ; Thompson, 19 ; Johnson and Milner, 19). The Interdepartmental Committee on Drg Addiction (191) mentions for cases of addiction to the drg, thogh it makes no reference to addiction in its Second Report (195). The W.H.O. Epert Committee on Addiction-prodcing Drgs (19) sggested that the term " drg dependence," with a modifying phrase linking it to a particlar drg type, shold replace the term " drg addiction." * Clifton Hospital, York. f Department of Pharmacology, Trinity College, University of Dblin. According to the W.H.O. definition, detromoramide wold fall into the class of drgs prodcing dependence of the morphine type. The most characteristic featre of this type of dependence is prodction of the abstinence syndrome within a few hors of administration of the last dose of the drg ; the symptoms which occr in this syndrome are described in the W.H.O. report and by Eddy, Hlalbach, Isbell, and Seevers (195). They can be evalated by the Himmelsbach (1939) scale. The WX.H.O. report states that the abstinence syndrome may also be precipitated in a matter of mintes by the administration of a specific antagonist while contining to administer the agent responsible for the dependence. Wikler, Fraser, and Isbell (1953) have shown that the administration of nalorphine can act as a specific antagonist to morphine, methadone, and heroin in this way. Br Med J: first pblished as 1.113/bmj on 1 Janary 19. Downloaded from on October 18 by gest. Protected by copyright.