INDEXED
WORLD HEALTH 0 R GAN l Z AT l 0 !\1
ORGANISATION MONDIALE DE LA SANT~ WHO/TB/Techn.Guide/3* 22 February 1963
THE wHO STANDARD 'IUBERCULIN TEST CONTENTS
INTRODUCTION • • • • • • • • • • • • • • • • • • • • • • TUBERCULIN PRODUCT AND DOSAGE EQUIPMENT AND STERILIZATION TECHNIQUE OF INJECTION MEASUREMENT OF REACTIONS~
.. . ...... • • • • • • • • •
2
• • • • • • , • • • • •
~
2
. . . . . . . .. . . . . . . . . ......... . ... . ..
·'-1-
5
AND RECORDING
• • • • • • • • • • • • • • • • •
7 8·
REPORTING AND EVALUATION OF RESULTS DISCUSSION • • • • · • • • • • • • • •
,
.... ............. . .
.. . . .... .. . . . .. . REFERENCES • . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . APPENDIX. Directions for preparing dilutions of purified tuberculin RT 23 • .. . • . • • . • • . . • • . • , •. • • • . • ... • • • ,
13
17
19
* This guide supersedes document WHO/I'BC/Int.;46 first issued on 21 Ai.igust i959 and re-issued with some revision on 1 March 1960 (WHO/TBC/Int./46 Rev,l).
HHO,/r.S/'Vecttn..:GJ:U.d.e.,IY-
page 2
INTRODUCTION
The tuberculin test forms an important part of several of the measures in the tuberculosis programme: surveys of the prevalence and incidence of tuberculous '-
infection, case-finding and BCG vaccination. . . · .. ' Much valuable information may be obtained by comparing data on tuberculin testing from different countries, or parts of a country, and for different periods. Such comparisons are, however, very difficult unless the data to be compared have ...
been obtained in a uniform manner. tuberculin'given'in technique.
a
The testing must have been done with a s~andard ...... standard dose and according to a scrupulously standardized followec.~
Similarly, standards must have been
in the reading, recording
and reporting of'the results.
WHO has developed, in recent
year~,
a.si;-al'!-dard tuberculin test based orr.detailed
studies of the various aspects of tuberculin testing and extensive practical experience gained in all parts of the world. This test -.a low-dose, intradermal· It
procedure - is used in all tuberculosis activities with ltlhich \rJHO is associated. has been adopted also by a number of national health authorities. The standard test is based on the use of a single batch of PPD-tuberculin, RT 23. a A s.tandard .low dose of ·this t·uberculin dissolved in 0.1 ml of a special
stabilizing diluent is injected intradermally.
The test is read three to four days
later by measur.in.g .the transv.erse diameter ·of pal·pable ·induration. · · ·
TUBERCULIN PRODUCT AND DOSAGE Batch RT 23 was prepared with UNICEF and WHO. 11
by the Statens Seruminstitut, Copenhagen, by agreement
The method of preparation was the same as for the two preceding
batches from this Institute, RT 19-20-21 and RT 22, both of which had been widely used in internationally-assisted tuberculosis control projects. The batch was assayed,
6
~ Stock. solution of RT 2~5 containing one milligram (50 000 "TU") per ml may be obtained from the Statens Seruminstitut, Copenhagen (see Appendix).
WHO/IB/Pechn...Guhle/3' page 3
in several human populations, against the International Standard for Purified Protein Derivative of Mammalian Tuberculin (PPD-S) and against the batches of PPD previously used in international work. The total weight of the,batch in. dry form was 670 g of
which today, after four years of use, 538 g (or about 27 000 000 000 doses, not cmmting losses and wastage) remain. One of the main reasons for preparing such a
large batch was to make unnecessary the standardization of frequent new batches of tuberculin: such standardizations, if they are to be satisfactory, require very Another important consideration was
extensive and costly assays on human subjects.
that the general and continued use of tuberculin fr::)m a single batch could be expected to increase the comparability of data on tuberculin testing from different areas and for different periods. It is expected that this batch will suffice until eventually
a much better kind of tuberculin than PPD has been developed (e.g. a more specific preparation) and even after that time will be available in cases in which such a new kind of tube~~ulin
is not adopted for some particular reason.
Thus, RT 23 ma3r well
be the last batch of PPD (for intradermal testing) ever to be prepared. The tuberculin is diluted with a special stabilizing diluent - isotonic phosphate..
buffered saline, pH 7 .38, to which 0 .0,)5 per cent. of the non-ionic detergent Tween 80 has been added. Purified tuberculin in the high dilutions used for intradermal
testing has been found to be subject to
~~rke~ unpredictable variations in potency5 • 9
and a series of studies has indicated that this instability is largely due to a rapid adsorption of the tuberculin to the glassware and containers in which the dilution is 8 15 kept. • Addition of Tween 80 to the diluent prevents this adsorption and thus the 10 variable attenuation of the dilutions. Provided that they are kept at a temperature not exceeding 20 C, except for short 19 periods, and are not exposed unduly to direct sunlight or strong daylight, the dilu t ions may be used for six ~onths.
0
9,10
The standard dose is 0.02 microgram ("1 Tun) of RT 23 in 0.1 ml of stabilizing diluent. This dose produced reactions of about the same size as 0.06 microgram
("3 TU" )a of the International Standard for Purified Protein Dc;rivati ve of f/Iammalian
.,.. The WHO Expert Committee on Biological Standardization, at its fourteenth meeting in 1960, decided that the existing definition of the International Unit of Purified Protein Derivative of Mammalian Tuberculin should no longer apply.21 In the present document, th:,;refN·o, the doses are defined by weight; for easy reference_, the equivalent in the previously used international tuberculin units is given in brackets.
a
WHO/IB/rech.n...Gui.de/3 page lJ. 6
Tuberculin (PPD-S) in 0.1 ml of diluent without Tween 80.
The reactions obtained softer~ however~
with 0.02 microgram of RT 23 with Tiveen tend to be somewhat those elicited with 0.06 microgram of PPD-S. Therefore~
than a
in some
programmes~
higher dose of RT 23 - usually 0.04 microgram - is preferred; that are not only ~arger~
this produces reactions such
but also more firm so that they are easier to read;
use of a higher dose Hill of course increase the proportion of inconveniently strong reactions.
EQUIP~lliNT
AND STERILIZATION
A special set of syringes and needles is used for the standard tuberculin test. This is kept entirely separate from the syringes and needles used for other is sterilized in a special container. In order to permit precise measurement ~f
tests~
and
the volume injected, the syringes
should be graduated in hundredths of a millilitre (or at least in fiftieths of a millilitre) and should be tight. injection~
Owing to the heavy pressure needed for intradermal
ordinary syringes - whether of the Record or all-glass type - will often If there is much leakage_, it is If
show some leakage between plunger and barrel.
impossible to measure accurately the volume (dose) of tuberculin injected. syringes of these types are used~
each syringe should be tested for leakage before A simple apparatus has
it is used for the first time, and periodically afterwards.
. 4 been constructed for thls purpose. discarded.
Syringes with significant leakage should be
In recent years, a new type of syringe has been introduced which is provided with a washer or piston ring - of rubber or plastic - fitted into a groove near the lower end of the piston. Such syringes when properly constructed are absolutely intraderrr.~l
tight and are therefore preferred for the
injection of tuberculin in
routine testing as well as in research work; Record syringes. The syringes are used with 25- or
they do not cost much more than ordinary
2,6-gauge~
10-mm-long, platinum or steel needles. their surface
A disadvantage of steel needles is that they are affected by flaming:
becomes roughened by oxydation which makes rapid and painless insertion in the skin difficult.
WHO /'TB/'I'echn...e/3-page 5
A small burner, usually with ethyl alcohol, is used for flaming the needle point before each injection. The reactions are measured by means of a small transparent ruler (suitable length: 10 em) calibrated in millimetres.
Syringes and needles are sterilized at least once on each day of use, preferably in the autoclave or, if that is impossible, by boiling in distilled water for at least 10 minutes. water may be used.) (If distilled water cannot be obtained, cleanly-collected rain-
A syringe and needle may be ltsed for many injections during a if it is necessary to change the needle, a freshly-sterilized Furthermore, before each
single session, provided the needle, once mounted, is not taken off the syringe after the first injection:
syringe should be taken at the same time, and vice versa.
injection, the tip (2-4 mm) of the needle is heated in a flame, in the case of steel needles to at least l00°C (when this temperature has been reached, one can usually hear the escape of steam from inside the needle); preferably to a dull-red heat. Irr~ediately
and, for platinum needles,
after the desired temperature has been no
attained, a small amount of diluent is squirted out in order to cool the needle;
attempt is made to adjust the position of the plunger relative to a particular point of the scale, but the position of the plunger between two points is noted an~
memorized, and a corresponding position between two points 0.1 ml further on the scale is aimed at when the injection is made. the whole needle is flamed. before injection. Before each filling of the syringe,
Normally, no attempt is made to sterilize the skin
TECHNIQUE OF INJECTION Usually the test is given in the dorsal aspect of the forearm although, when necessary, the volar aspect may be used without significantly affecting the size of the reactions. 2
However, it is important not to give the test in a site previously Reactions to tuberculin injected in the site of a they appear sooner, fade
used for tuberculin t'esting.
previous test differ from those obtained in a new site:
sooner {and a fading reaction may be more difficult to perceive), attain greater 20 . . " presen t a muc h h" " max1.mum s1.ze, ana 1.g1wr frequency o f b u ll ae. 3 • 7 • Therefore, wh enever retesting is carried out, the test sites should be systematically rotated.
WHO/TB/Techn-Ouide/) page 6
How large an area is affected is not.precisely known but it is generally assumed that the effect is limited to the site of the old reaction, possibly including surrounding oedema or double erythema. In practice, it would seem safest to use
no more than two, or at most three, well-defined test sites at the dorsal aspect of the forearm and the same at the volar aspect. The effect is still very
pronounced a year or more after the old test, and it is no doubt advisable not to use the site of a previous test for a number of years thereafter. The needle point is inserted in the superficial layer of the skin of the forearm while the skin is ligb.tly stretched in the direction of the needle and lengthwise of the arm. The syringe is held by the barrel only and the plunger is not touched The volume of 0.1 ml is
until the needle point has been satisfactorily inserted.
slowly injected, and the finger removed from the end of the plunger before the needle
is withdrawn.
The injection should raise a flat, anaemic weal with pronounced pits If the injection is made into the deeper layers of the
and a steep border line.
skin {as shown by a dome-shaped and less anaemic weal), this will scarcely affect the size of the resultant tuberculin reaction but may tend to make it more difficult to 2 read. The volume injected should not, as is frequently done, be gauged by the size of the anaemic weal raised by the injection. As shown by several studies - with . 2 an d Wlvn ·~· BCG vacc1ne . 14 ' 18 - ~ 'h e wea 1 s1ze . . an inaccurate gauge. t u b ercu1 1n 1s
F or
one thing it is impossible, without frequent measuring, to .. decide whether the weal raised is 6, 7, or 8 mm in diameter; and such relatively small differences in weal An increase in weal
size correspond to great variations in the volume injected.
size of about 1.5 mm, for instance, may correspond to a doubling of the volume injected (and thus of the dose). affects the size of the weal: Moreover, the depth of the injection markedly
0.05 ml given superficially or 0.2 ml given deeply
produce weals of approxoimately the same size.
Also, th2 size of the weal raised by 2
a given volume varies considerably with the sex and age of the person tested, being largest in adult males, intermediate in adult females, and smallest in children.
WHO/ffi~
page 7
.f\'JEASUREMENT OF REACTIONS, AND RECORDING
The test is read (examined) three or four days after it has been given. reading is limited to a single aspect o:::' the reaction, viz. the induration (infiltration).~
The
a
The test site is carefully palpated and if an induration is
present its limits are determined and its transverse diameter (transverse relative to the arm) is measured.in millimetres. The induration may be more or less easily
recognizable, varying from a firm, well-circumscribed density in the skin to a soft, ill-defined swelling. The latter type of induration may easily escape notice
unless the test site is palpated with a light touch. The widest transverse diameter of the induration is recorded in millimetres. If there is no palpable induration,
"o"
is recorded.
The presence of additional The recordings should be
features such as bullae and lymphangitis may be noted. 22,23 made preferably on an individual card • The readings should be made by trained observers. 11
The abili t3r to make
accurate and consistent readings is acquired only after long training in careful quantitative 11 reading. 1
An inexperienced reader, for instanc2, may tend to overConversely, he may
read a strong reactinn by including the surroundinr; oedema.
under-read a weak reaction by failing to include its softer, flatter, marginal part. Small systematic differences are observe<i even between the readings of experienced readers. Therefr"~re,
when two or 1:1ore readers work in the same proThis comparison is made by a
gramme, their readings are compared periodically. so-called blind double reading:
the reactions of a relatively large number of The
subjects are read by both readers, each working independently of the other.
results of these readings are entered in a correlation table which permits closer . 1,22 analysis of the between-reader dlfferences. ·
~ The measurement of several aspects of the reaction at the same time (e.g. erythema and density, besides induration) is only of interest in special research, and in particular in assays· of tuberculin products •
......
~THO j4'B/J'ecl'1Ih Gui {le/p
page 8
Hhen reading a reaction, the reader should not know whether the person has been BCG-vaccinated, so as to avoid psychological bias in the readings; similarly, he
should be unaware of the results of any previous tuberculin test, or radiological or bacteriological examination. Thus he should have no access to the records of the Furthermore, when post-vaccination testing is
person whose reaction he is reading.
done, it may be of advantage to test also the unvaccinated subjects belonging to the same group.
REPORTING AND EVALUATION OF RESULTS I'Jhen a population group has been tested, the tuberculin test results are sorted according to the recorded size of the reaction to determine how many persons have reactions of 0 mm, l mm, 2 mm, etc. The resultant frequency distribution of The use to be made of whether it is used
reactions represents the basic data and is always reported. the frequency distribution depends on th~
purpose of the test:
in (a) unvaccinated groups either for epidemiological purposes or in order to separate the infected from the uninfected, or (b) vaccinated groupR to assess the allergy induced by BCG vaccination. (a) Testing of unvaccinated groups Not only Hhen the tuberculin test is used to select subjects for BCG vaccination or chest X-ray, but also in most epidemiological studies, the purpose of the test is the same: to distinguish between those who are infected with tubercle bacilli and How effectively this can be done under differing circumstances
those t'Vho are not.
will be discussed belovJ. Fig. l shows the frequency distribution of reactions to 0.1 microgram ("5 TU") of RT 19-20-2la observed among tuberculosis patients in Libya and in the Sudan. To
permit an analysis of the distribution, it has been presented in the form of a histogram in which the horizontal scale shows the sizes of the reactions and the heights of the columns show the percentage of persons with reactions of specified sizes. is seen that the reactions of the patients, i.e. persons known to be infected, It
a
This dose, dissolved in ordinary diluent, corresponds approximately to 0.02 microgram of RT 23 in stabilizing diluent.
Fig. 1 Distributions of tuberculin sensitivity in tuberculosis patients in Libya and the Sudan (size of reactions to "5 TU .. of RT 19· 20· 21) NUMBER: 274 MEAN: 16.0
30
Data from WHO-assisted surveys in Libya and the Sudan
Fig. 2 Distribution of tuberculin sensitivity in an unvaccinated village population in Ben Ulid district, UJ>ya {size of reactions to "5 TU" of RT 19-20-21)
30 <.!>
NUMBER: 1283
w
~20
~ 10 w a.. 8 12 16 20 24 28 INDURATION {mm) Data from WHO- assisted survey in Libya WHO 3120
z w
~~0/TE/Techn.Guide/) page S
constitute a "beil-shaped" or
11
normal" distribution, being grouped fairly symmetriTne reactions are characteristically
cally and compactly around acentral value. large, few being smaller than 8 mm.
Essentially similar distributions have been 16 obtained among tuberculosis patients in all parts of the world. A further distribution of reactions to the same dose of tuberculin, also from Libya, is shown in Fig. 2. population. It has been obtained among groups of the general
The reactions are seen to fall into two distinct groups, one (to the
left) of zero or small reactions, and the other (to the right) of l~rge-sized ~eactions
distributed almost symmetrically arounQ a maximum at 19 wn.
The right-
hand group closely corresponds in shape and position on the size scale to the distribution for the tuberculosis patients, anG. is presumed to represent persons infected with tubercle bacilli. Conversely, the left-hand group woulc:J. represent 13 those uninfected. The small reactions shown by many of the uninfected may be due to weak~
non-specific allergy (more about which below);
or they may be
tralli~atic
in
origin, that is, due to the injection as such and not to any sensitivity to tuberculin. In fact, distributions of reactions similar to that of the left-hand group in Fig. 2 . d by t es t" lng h ave b een obt alne "th Wl d"l ~ a l one. - l3 l uenu A small in which the
It will be noted that the two groups are not completely separated. proportion of reactions falls into largest of the 11
a~
intermediate zone (from 6-12 11
~n)
negati ve" reactions and the smallest of the
positive" reactions
seem to overlap. ~ ..
However, there are relatively few-. of these intermediate,
ndoubtful" reactions, and if the limit between positive and negative reactions is set at 8 or 10 will be small • The technical error of the tuberculin test is partly of complete separation. r~sponsible
for instance, the proportion mis-classified in either direction
for the lack
Owing to random. variatinns in injecting the tuberculin and
in reading the reaction, the observed reaction size may deviate more or less from 12 the "tru.e 11 size • Some (presumably half) o::" the deviations in infected persons will be towards the left and some of the deviations in non-infected persons will be to the right. Thus, any tendency for overlapping of sensitivity will become
accentuated in the empirical distribution, through the random errors.
WHO/ffi/Techn~Guioe/)f
page 10
The bipartite distribution of reactio:J.s described above has been found in practically all countries in the temperate zone and in many countries in the subtropics. It may :;e concluded that, whenever the reactions obtained in
unvaccinated groups show this type of distribution, the tuberculin test may be expected to provide a reasonably efficient distinction between infected and uninfected. A certain proportion of those tested will be wrong~y
classified
~ut
this proportion will be relatively small.
?or the purposes of rov.tine testing,
a sui table size-limit will be established on the basis of a careful tube.rc'Ulin survey. An analysis of the distribution of reactions obtaineQ in the survey will The choice of tl1is ;, routine" limit may also For instance, when the test
indicate where the limit should be set.
be influenced by the particular purpose of the test.
is used as a screening procedure to select children for X-ray examination of the chest, it may be considered desirable to exclude from examination only those who are definitely uninfected. size. In this case, the limit will be set at a lower reaction11
This "iill increase the probability that those classified as
negative" are
in fact uninfected;
at the same time, of course, a higher number of uninfected As another
children will be erroneously classified as "positive" and exar.nned.
example, when the test is used to select children for chemoprophylaxis, it may be considered desirable to restrict the selection to those definitely infected. this case, the limit "lill be set at a higher reaction size. In
This "'ill improve the
probability that those classified as Hpositive" are in fact infected with tubercle bacilli; at the same time, of c~urse,
an increased number of infected children will •
be missed by beinr; classified as "negative 0
vfuen the test is used for epidemiological purposes - for instance, to estimate the prevalence of tuberculous infection - it is generally preferable to base the estimate not on a routine criterion but ·-on a direct analysis of the way in which the reactions are distributed according to size. In most tropical countries and in some parts of the sub-tropics, the reactions observed in the general population are not distributed according to the clear bipartite pattern described above, but the pattern is more or less obscured by a varying proportion of intermediate-sized reactions • Sudan, are given in Fig. 3.
13
Tv1o
examples, both from the·
In the upper histogram, two g.coups of reactions are
Fig. 3 Distribution of tuberculin sensitivity in unvaccinated children, 8-12 years old, in two provinces of the Sudan (size of reactions to TU" of RT 19·20-21}
,,5'
30 BLUE NILE PROVINCE 20 10
NUMBER: 2819
NUMBER:3306
KORDOFAN PROVINCE 20 10 0~~~~~~~~~~~~--~
0
4
8
12
16
20
24
28
INDURATION ( mm} Data from WHO- assisted survey in Libya WHO 3121
WH0/"'3/Techn....GU:ide/3 page 11
still discernible, but a large proportion of intermediate-sized reactions has appeared. In the lower histogram, separate groups can no longer be distinguished: These "non-bipartite" patterns do not reflect a infectec~).
the groups seem to have merged.
change in the right-hand group (the
This is evicl.enced, among other
things, by the absence of change in the distributions for tuberculosis patients (those known to be infected): the form and position of these distributions on the
size scale is essentially the ::>ame in the tropics as in the temperate zone {see Fig. 1). The observed patterns rather correspond to a change in the composition
of the left-hand group (the uninfected), with fewer zero or small reactions and more intermediate-sized reactions. The interpretation of this change is that there are two kinds of sensitivity in the population: as everywhere, there is a specific high-grade sensi"tivity reflectlow-grc-.~.de
ing infection with tubercle bacilli, and, in addition, there is a appearing as intermediate-sized reactions. termed 11
sensi ti vi ty
T'he latter type of sensitivity has been
non-specific", as it seems not to be related to tuberculous infection.
13 16 '
Its cause is as yet unknown, but there is increasing evidence suggesting that it may be due to infection by acid-fast, essentially non-pathogenic organisms antigenically related to Mycobacterium tuberculosis. The term nlow-grade" should not be taken to mean that non-specific sensitivity is invariably weak; to another. it may, in fact, vars· in intensity from one geographical region
The !.1igher its intensity, the more marked is the shift of the left-hand
group towards higher reaction sizes, and the more overlapping th8re is of the largest non-specific reactions and the smallest specific ones. It is evident that the 11
overlapping 11 seriously reduces the efficacy of the
tuberculin test as a means of separatin::; those infected ''Ti th tubercle bacilli fror11 those who are not. vJherever the border-line bet\-veen positive and negative is drawn, If the size-
a substantial proportion of the re<J.ctions will be wrongly classified.
limit is set low, for instance at 5 rnm, few cf the infected will be classified as negative, but a large proportion of the uninfected (those with low-grade sensitivity) will be considered positiv~.
On the other hc:nd, if the size-limit 1s set high, say
at 12 mm, the classification of uninfectcd persons as positive will be largely avoided, but many of the infected ''rill be labelled as negati ·ve.
vJHO/TB/rechn.Guicle/3 page 12
In practice, the decision on where to set the routine size-limit will be based on: (i) an analysis of distributions of reactions obtained by testing of sample this will give
groups of the general population and tuberculosis patients:
some indication of how much, and where on the size scale, the reactions of the uninfected overlap those of the infected; (ii) the purpose of the testing: fo~
if the test is used to select subjects for
BCG vaccination,
example, the limit may be set high, the inclusion of most
of the uninfected being considered more important than the exclusion of all the infected. The non-bipartite pattern 0f the distributions is somewhat less of a drawback when the test is used as an epidemiological tool to study the frequency of tuberculous infection. The problem here is not to separate the infected from the uninfected but This may be done by attempting to
rather to estimate the proportion of infected.
estimate the size of the (more or less obscured) right-hand group in the distribution. As shown previously, it may be assumed that, whatever the pattern of the distribution, the reactions of the infected will follow a 11
no:rmal" distribution, and it may be
possible if the overlapping is only slight to reconstruct the separate distributions of 22 "positive" and "negative", by means of appropriate statistical tools. However, if the overlapping is pronounced, such statistical manipulations may not be permissible. Some, very crude, relative estimates may still be obtained by considerin:g only reactions above, say, 17 or 19 mm; the number of such strong reactions may be assumed
to constitute an approximately constant proportion of all infected persons, since the strength of tuberculin allergy tends to be fairly independent of age (at least for children and young adults) and of the epidemiolosical situation. It should be
emphasized once more that this latter method usually will be veDr. inaccurate. (b) Testing of ~CG-vaccinated
groups
The purpose of the tuberculin test when used in vaccinated groups is to assess the tuberculin sensitivity resulting from the vaccination.
Fig. 4
Distributions of tuberculin sensitivity in schoolchildren vaccinated with one of two BCG vaccines of differing strengths (size of reactions to 115 TU 11 of RT 19·20-21) I
30 20 10 w 0
VACCINE LOT 1124
NUMBER: 281 MEAN: 14.9
(!) <(
1-
z
u
w 0::
0..
w
30 20 10
VACCINE LOT 230
NUMBER: 476 MEAN: 10.1
o~~~~~~~~~J_~~~
0
8 12 16 20 24 28 INDURATION (mm) Data from WHO-assisted survey in the Sudan WHO 3122
4
\IJHO/TB/:'echm CiuJ..de/3"" page 13
In Fig. 4 two examples are given of the distribution of reactions (to 0.1 micrograrr~,
"5 TU11 , of batct RT 19-20-21) obtained in groups vaccinated intradermally with The two examples refer to two groups of Sudanese children vaccinated with one It is seen that in each case the distribution of the
BCG.
of two different vaccines.
reactions approximates a 1'normaln distribution.
rhus, the reactions are distributed
similarly as for tuberculosis patients (see Fig. l), although they differ in being situated lower on the size scale: the sensitivity induced by BCG vaccination is usually
I'Jeaker than that resulting from natural virulent infection. From the pattern of the distributions, it is clear that the vaccinated children form a homogeneous group with respect to tuberculin sensitivity, differing from one another only in degree of sensitivity; those l'li th 11
zero'· reactions may be assumed to
have a very weak degree of sensitivity demonstrable only by testing with a higher dose of tuberculin. Thus it would be biologically meaningless to divide the reactions 11
invaccinated persons into two categories arbitrarily called
positive" and "negative".
17
Instead, the mean size of the reactions (including the nzero 11 reactions) is used to describe the level of allergy; as long as the distribution of reactions retains a
fairly symmetrical unimodal form, the mean size may be considered the rnost accurate and suitable single expression of the allergy in the group tested. In the two examples given, the vacc."ces used were apparently of different allergenic potency: the reactions of the children vaccinated with lot 1124 average
14.9 mm, while those of the children given lot 2)0 average only 10.1 r.rr.n.
DISCUSSION The introduction of Tween 30 as a stabilizing additive is the result of a long series of investigations?, ,~, ? designed to define the importance of the loss of potency observed when purified tuberculins are highly diluted.. These investigations
' 8
n
l'
did not c·upport the commonly-:-held belief that the ac"ci vi ty of the dilution decreases gradually witn the duration of storage, and that this source of error could be eliminated. by using freshly-prepared dilutions. In fact, it was found that the
potency of the dilution decreases very rapidly - in a matter of hours - and quite unpredictably. Large va~iations
in potency could be demonstrated not only between
dilutions prepared at different times but also between ampoules filled with the same
v!FO/""Hf!'ecJ~" J' riLJE'/~
Pd.68
14 the dilu-l,ion f:L"om 011<2
dilution: another.-' 0
ampoule might b2 only half s.s potent as that from experlments~ -
As shovm by further
8~15
these variations in potency could
be largely explained by variable acsorption oi tuberculin to the inner surface of the a . con t a1ner.·lJi th a suj_ table eluent, for insta:::1ce~ variable amounts of tuberculin could be recovered from gJ_ass ampoules that had previously contained tuberculin c.~.ilution~
the largest BX10'--l-Ylts being recoverable from those that previously contained dilut~_o::..
a particularly Heak
Another
finding~
also of practical
importance~
was
that a further loss of activ-ity takes place in ampoules filled only partially with r; dilutioYl.Experim2ntc:.l evidence suggests that this loss is a.lso due to adsorption: apparently the decreased ratio of volume of dilution to area of liquid-solid interphase in the almost empty c:.mpoule favours further aC::.sorption. These different forms
of reduction in potency may be prevented by adding a stabilizing c:.gent to the diluent. The most effective agent was found to be the non-ionic detergent Tween 80~ in a concentrc-.tion of 5 parts per 100 000. Batch RT standard~/J'1 23~
which was issued at the time when TvJeen diluent was introduced, was
ac1.inst PPD-S s_nd some pre-,riously-used purified products_, both in 6 ord_:_nary clil1.l ::-it,a c:.nJ in dilution stabilized vd th Tween. Tl1.e standardization was done by dupllcate intrade:cmal testing) chiefly in human subjects. One important aim
was to defi"le "equipo"c-cnt 11 doses of RT 23 with and without 'TWeen~ i.e. doses evoking the same ::tver:.:-.12,-J r..:-~ction E:
lze L1 p2rscns infected with tuberculosis, as the standard The results showedJ among other things, that RT 23
doses of the other products. prepared with Tween without Tween. ~-'~
abo'.lt three times more potent, in these terms, than RT 23
'Ihe usc of 'Tween ciluent was found not only to increase the potency
of the tuberculin dilution in te::·ms of reaction size b'Jt also to result in reactions of a different character~
as has been confirmed by subsequent field experience.
In
persons vvi th a modera. te-1:.o-high deGree o:L allerg:c, a dose of 0 .02 microgram of R'l' 23 1'-Ti th TvJee'J. elicits reactions of about the same size as the usually-employed dose of RT 19-20-21 or P.T 22. inte:'l.se~
C' 5
TU")
Howev•3r, the reactions to tuberculin vlith Tween are less The relative softness of
i.e. they are softer and less frequently bullous.
the reactions ma1ws them more difficult to 'jread11 ~ especially to an inexperienced observer~
their size may be under-estimated or they may be missed altogether.
Ano t;her "quali ta ti v2 11 difference is t[1a t the potency ratio of tuberculin with ween to tuberculin vJithout '11Hcen is not the same in persons i<Tith weal-: allergy as in those with
~ Most of thest: experimen~s were carried out with the PPD products used in vmo/ UNICEF-assisted programmes, bu-;; the phenor:lenop occurs also with other tuberculins, a;Lthough for some it ~-s m~-..:r, less pr~nounced. b
WHO/"D/Thchn.Guic.ie/3. page 15
strong allergy.
For example, when 0.02 microgram of RT 23 with Tween is used instead
of a dose of, say, n5 TU" of RT 19-20-21, persons with moderate-to-strong sensitivity (such as follows infection, or vaccination V<Tith a potent BCG vaccine) will show reactions of the usual size, whereas persons with weak sensitivity will have smaller reactions than usual. The latter phenomenon must be taken into consideration when
evaluating the results of retesting after vaccination vlith a relatively weak BCG vaccine. The post-vaccinati::m tuberculin reactions elicited with 0.02 microgram of
RT 23 with Tween will show in this case a wider distribution by size (larger variance) and a lower average size, very small reactions being more frequent than with corresponding doses of tuberct<lin Ni thout Tween 80. If reactions below a
specified size are classified as "negative", a marked increase in the proportion of "negatives" may be observed, following the introduction L1f Tween, in groups of persons vaccinated with weak vaccine, while groups of persons vaccinated with uniformly strong vaccine will continue to b<.: "almost 100 per cent. positive" with either kind of tuberculin dilution. (Whether it is a disadvantage or an advantage
to use a tuberculin preparation that will emphasize so strongly the difference between more and less potent vaccines is perhaps a somewhat delicate question.) More generally speaking, comparison of the test res1.-~lts
obtained before and after
the introduction of Tween is possible only in the case of persons with infection allergy or vaccinated with relatively strong vaccinEos. Test results obtained in
persons reacting weakly, because of non-specific sensitivity or vaccination with a weak vaccine, cannot be expected to be comparable. Two possible explanations for the qualitative differences between purified 6 tuberculin diluted with and without Tweeu have been suggested. One is that the addition of Tween to the diluent makes the tuberculin more specific, i.e. that it prevents the loss - by adsorption - of a highly specific fraction of the tuberculin. Another is that Tween has an in vivc effect on the tuberculin reaction, which is dependent on the size of the reaction, Accumulating evidence seems to suppcrt the
second hypothesis, so that·, tu the extent that the speclal character of the reactions to Tween-tuberculin is considered a disadvantage, it would t)e
desirable to have an Studies of this
alternative stabilizing agent with no effect on the reaction. problem are at present under way.
\tJHOjT'B/P.echn..:tGui.-de/3' page 16
However, it must be emphasized that the special character referred to is at most a very minor disadvantage relative to the importance of having tuberculin dilutions of accurate potency and high stability. Therefore, until another
stabilizing agent has been demonstrated to be decidedly superior, Tv1een
80 will
continue to be used in \I]HO tuberculosis control programmes as an additive to the tuberculin diluent.
\iJHOji:'B/Techn .. Guide/3 page 17
REFERRNCES 1.
Edwards, L. B., Palmer, C. E. 8: I•1agnus, K. (1S:~3). BCG vaccination (World Health . t. "" . .:.er~es, .~ ~ N J.on: Honograpn ·o. 1" c:., p. 34) 0 rg&nJ.za Guld, J. (1954) ~uantitative aspects of the intradermal tuberculin test in humans. II. rl'hs relative importance of accurate injection technique, Acta tube~c. scand. 30, 16 Gclld_, J. (1957) Interpretation of tuberculin reactions in populations with a high proportion cf BCq-vaccinated p~rsons, Bull. 1\ild Hlth Org. 17, 225 0 1 ' J • e.. -t.LCL, G
2.
3. lj. •
Rua, ' 0. ~ (l';.r_ :';) ,· ·· ~ 1 ea.Kac:e '· f t u b ercu l" · ::; ..•. .. 1easurement or oJ.n syrJ.nges, Brit. med. J . l_, 368
5. 6.
Guld, J ., f/lagnus, K. & Magnusson, riJ. (1;::)5) Instability of the potency of tuberculin dilutions, Amer. Ii.ev. Tuberc. '72; 126 Guld, J. et al. (1958) Standardization of a new batch of purified tuberculin (PPD) intended for international use, Buli.·v!ld Hlth Org. 19, 81+5 Ikegami, ~1. (1956) Studies on the standardization of the evaluati.on method of tuberculin test. Report I. Errors in the result of tuoel"'CUlin test due to the accelerated reaction occurring at its customary application site, Kekkaku, 31, 383 _; Report II. Comp.ariso~ 6f the i~tensi ties of tuberculin tests at v;:-rious parts of the flexor sic..~e of the forearm, Kekkaku, 31, 44.); Report III. On the size of the double erythema of tuberculin test, Kel\:kaku, 31, 510; Report IV. Selection of the sites for the tuberculin test, showing equal intensities of reactions, ICekkak:u., 31, S30 !1agnus, K. et al. (1956) Instability of the potenc;,- of tuberculin dilutions. A second report, Amer·. Htv. Tuberc. 71J. , 297 r1at71us, K. et al. (1<;'58) Stability of purified tuberculin in hic;h dilution, Bull. \Jld Hl th Org. 19, 765 J.'.1agnusson, r.'l. et al. (1953) Diluents for stabilization of tuberculin, Bull. Hld Hl th Org. 19, 799 Magnusson, M. &. Bentzon, M. v,f. (1958) Preparation of purified tuberculin RT 23, Bull • vJld Hl th Org • 19, 829 Nissen Meyer, S., Hougen, il._. & Edwards, P. (1951) Experimental error in the determination of tuberculin sensi ti vi t~r, Publ. Hl th Rep. (!.lash.) 66, 561 Nyboe, J. (1960) The efficacy of the tuberculin test. An analysis based on results from 33 countries, Bull. Hld Hlt1 Org. 22, 5
7.
8. 9.
10. ll.
12.
1).
WHO/TB/TechrL GUide/3 page 18
11+.
Palmer, C. E. & Edwards, P. 0. (1953) Vario.tion in tecr..nique of intracutaneous BCG vaccination, Brit. med. J. l, 363 Waaler, H. et al. (1958) Adsorption of tuberculin to glass, Bull. Hld Hlth Org. 19, 783 WHO Tuberculosis Research Office (1955) Further studies of geographic variation in naturally acquired tuberculin sensitivity, Bull. Fld Hlth Org. 12, 63 \JJHO Tuberculosis Research Office (19,55) Certain characteristics of BOG-induced tuberculin sensitivity, Bull. Wld Hlth Org. 12, 123 ~VHO
15.
16.
17.
18.
Tuberculosis Research Office (1955) Suppurative lymphadenitis following intradermal BCG vaccination of pre-school children, Bull. \!ld Hlth Org. 12, llf3
19.
HHO Tuberculosis Research Office 8: Biophysics Laboratory of the Universit;)T of Copenhagen (1955) Effect of exposure of tuberculin to light, Bull. Wld Hlth Org. 12, 179 ...
20.
HHO Tuberculosis Research Office (1955) Repeated tuberculin tests in the same site, Bull. \tJld Hl th Org. 12, 197 \'Tld Hlth Org. techn. nep. Ser. 1961, 222, ll.J WHO (1~63)
21. 22.
Technical guide for assessment of BCG vaccination programmes (document HHO/TB/Techn.Guide/2 Rev.2)
23..
WHO (196-) Standard records for t!HO tuberculosis control programmes (document WHO/'IB/Techn.Guide/1+) (in preparation)
VJHO,/IB/'Vechn_.-Gui-de./3 page 19 APPEND DC DIRECTIONS FOR PREPARING DILUTIONS OF PURIFIED TUBERClJI,TI-J RT 23 DIRECTIONS FOR PREPARING DILUTIONS OF PURIFIED TU:3EHCULIN RT 23
A.
Stock solution of PPD Stock solutions containing 50 000 TU per ml are dispatched from the 3tatens
Serurninsti tut,. Copenhagen..
If- kept ~t.e:rile and at rJ.bout
.4°c_,
the st-ock solutions
can be used up to 12 months after preparation.
Fresh stock solutions·are prepared
at the Statens Seruminstitut on about l !•larch_. l June_, 1 September and l December. B.
Diluent
A stabilizing diluent consisting of (isotonic) phosphate-buffered saline_, pH 7.38~
with 0.01 per cent. chinosol and 0.005 per cent. Tween 80, should be used.
The constituents are prepared as follows: 1. Phosphate-buffered saline, pH 7.38 Potassium dihydrogen phosphate (Iili PO~)
2
l.~LS g
Sodium monohydrogen phosphate (Na HPo .2H 0) 4 2 2 Sodium chloride (NaCl) The above are distilled in 1.05 litres of distilled water. by autoclaving at 120°C for 30 minutes. 2. Chinosol: 10 per cent. chinosol. ll g
'7 .60 g
l.J-.8 g The solution is sterilized
of chinosol are dissolved
aseptically in 100 ml of sterile distilled water. 3. Tween
Bo
(Atlas Powder Company, Wilmington, Delaware, USA) is sterilized by
autoclavinc at l20°C for 20 minuces. 5 per cent. solution of Tween 80. I\
5 per cent. solution is prepared by
dissolving 1 ml of sterile Tween 80 in 19 ml of sterile, phosphate-buffered saline with 0.01 per cent. of chinosol. , Gea t• J.ng. 1
Dissolving of Tween 80 is facilitated by slight
Stabilizing diluent.
To 1 litre of sterile (isotonic), phosphate-buffered
saline, pH 7.38, 1 ml of 10 per cent. chinosol, and 1 ml of 5 per cent. of Tween 80 are added. 1 Sterile 5 per cent. solutions of Tween 80 may also be obtained from the Tuberculin Laboratory_, Statens Serum:.Ln.sti tut_, Copenhagen.
i-i'RO/TB/Tec.bn...Guide/3 page 20 Appendix
c.
Preparation of tuberculin dilutions 100 TU per 0.1 ml dilution (for further dilution only) To l litre of stabilizing diluent add 20 ml of PPD stock solution. l TU per 0.1 ml: to l litre of stabilizing diluent add 10 ml of 100 TU per
0.1 ml solution.
October
196~
Mogens Magnusson Tuberculin Laboratory Statens Seruminstitut Copenhagen