692
ported by a grant from McGaw Laboratories. Financial support for
cates that the abnormalities are particularly common
more than a week after major surgery, at a time when
many of the patients had sepsis. These changes were
almost entirely unrecognised and untreated in our pa-
tients before the survey. Indeed, few of the patients had
even had their weight measured.
It is not known if disorders of nutrition such as we
have detected have any influence on the outcome after
surgery. But it seems to be generally agreed that treat-
ment to improve nutrition is indicated in patients who
are "malnourished" after major surgery, particularly
when the patient has sepsis as well. Our findings suggest
that little attention is being given to the nutritional state
of these very ill patients. Whatever else is done to draw
attention to this situation, we, like others,’ 15 suggest
that greater emphasis on nutrition is required in the
medical curriculum.
We should like to thank the surgeons and the nursing and technical
staff of all the departments involved in the survey. R. L. B. was sup-
Screening for Disease
A STRATEGY TO DETECT &bgr;-THALASSÆMIA
MINOR
IAN SHINE S. LAL
Thomas Hunt Morgan Institute of Genetics, Inc., 628 North
Broadway, Lexington, Ky. 40508, U.S.A.
Summary Over the past three years 25 302 adults
in Kentucky have been tested for
hæmoglobinopathies, and of these, hæmoglobin A2 was
measured on 3734, 1973 with microcytosis and 1761
within the normal range. The best methods of detecting
&bgr;-thalassæmia minor using red-blood-cell indices were
compared. No method detected all heterozygotes. A new
method was devised consisting of three parts: (1)
hæmoglobin electrophoresis, (2) calculation of the pro-
duct of the square of the mean corpuscular volume
(M.C.V.) multiplied by the mean corpuscular hæmoglobin
(M.C.H.) measured in units of one hundred, (3) A2 deter-
mination on all AA samples with (M.C.V.)2 × M.C.H.
<1530 and on those with variant genotypes consistent
with thalass&aelig;mia. In this series this new method
detected 137 out of 138 heterozygotes with 4&middot;4% false-
positives.
INTRODUCTION
THE thalasstmias are a group of clinically important
disorders caused by inherited defects in the rate of syn-
thesis of normal haemoglobin. This paper is concerned
only with the carriers of the most common of the tha-
lassaemias, p-thalassaemia, which in the homozygous
state causes the death of about 100 000 children in the
world per annum.’ As for any autosomal recessive dis-
order, theoretically the incidence of the homozygote
could be reduced almost to zero if the heterozygote
could be detected. An ideal test for the detection of
heterozygotes should have no false-negatives, few false-
positives, and should be simple and reliable. Several
tests have been used including measurement of osmotic
fragility,2 fetal haemoglobin,3 free red-cell porphyrin,4’
blood-smears,5 and haemoglobin A2,6yet none is readily
the vitamin assays was provided by Roche Products Ltd.
Requests for reprints should be addressed to G.L.H., University
Department of Surgery, The General Infirmary, Leeds 1.
REFERENCES
1. Randall, H. T. in Manual of Surgical Nutrition (edited by W. F. Ballinger,
J. A. Collins, W. R. Drucker, S. J. Dudrick, and R. Zeppa). Philadelphia,
1975.
2. Bistrian, R. B., Blackburn, G. L., Hollwell, H., Heddle, R. J. Am. med. Ass.
1974, 230, 858.
3 Leevy, C. M., Cardi, L., Frank, O., Gellene, R., Baker, H. Am. J. clin. Nutr
1965, 17, 259.
4. Bollet, A. J., Owens, S. ibid. 1973, 26, 931.
5. Butterworth, C. E. Nutrition Today, 1974, 9, 4.
6. Bartholomew, R. L., Delaney, A. Proc. Aust. Ass. clin. Biochem. 1964, 1,
64.
7. Mancini, G., Carbonara, A. O., Heremans, J. F. Immunochemistry, 1965,
2, 235.
8 Smithells, R. W., Sheppard, S., Schorah, C. J. Archs Dis. Childh. 1976, 51,
944.
9. Stanolovic, M., Miletic, D., Stock, A. Clinica. chim. Acta, 1967, 17, 353.
10. Brocklehurst, J. C., Griffiths, L. L., Taylor, G. F., Marks, J., Scott, D. L.,
Blackley, J. Geront. clin. 1968, 10, 309.
11. Densen, K. W., Bowers, E. F. Clin. Sci. 1961, 21, 157.
adaptable to mass screening.7 With the advent of elec-
tronic particle counters, estimates of red-cell indices
became more attractive and were combined or used
singly to identify heterozygotes.8 In the present study
the three best methods were tested on 25 302 samples
and all were found to miss known heterozygotes. To
overcome this insensitivity a new method was developed
which used the information already obtained from
haemoglobin electrophoresis and extracted the maximum
information from the red-blood-cell indices. It comes
close to an ideal method.
SUBJECTS AND METHODS
A consecutive series of 2302 healthy ambulant adults were
tested for inherited ansemias. 70.3% were Black and 29’7%
were White of mixed North European ancestry with a small
Mediterranean admixture.
After informed consent, a 5 ml sample of venous blood was
drawn into a vacuum tube containing 7 mg E.D.T.A. The
haemoglobin type was identified by cellulose-acetate electro-
phoresis with "tris" E.D.T.A.-borate buffer at pH 8.4 and
citrate agar electrophoresis in citrate buffer at pH 6.2 by Sch-
neider’s method.9 Whenever a band migrating as F was pres-
ent, it was quantified by the Betke technique.1o As samples
were often 1-2 days old, red cells were not stained for fetal
haemoglobin. Haemoglobin Az was measured in duplicate for all
possibly thalassaemic samples, that is all samples with a mean
corpuscular volume (M.c.v.) of less than 80 m3, and onallwith
banding genotype AF, FA, or SA; it was also measured on as
many samples within the normal range as the daily work-load
would allow. Haemoglobin Az was separated from the whole
blood on a D.E.A.E. cellulose minicolumn in glycine buffer and
measured on a spectrophotometer at 415nm." The measure-
ment of A2 was reproducible, except when the sample con-
tained haemoglobin S. The extent of inaccuracy was measured
on a pooled sample from 50 people with sickle-cell trait; the
pooled "haemoglobin Az" fraction was concentrated by dialysis
and separated on a long column into 83-7% A2 and 14.3% S,
A Coulter counter model S, calibrated daily against normal
and abnormal controls, was used to generate red-cell indices.
The indices were used to calculate the England and Fraser dis-
criminant (M.c.v. minus the red-blood-cell count [x.s,c.]
minus five times the haemoglobin, minus 34 <1)"
(M.c.v.-R.B.c.-5Hb-3.4 <1) to detect Pearson’s threshold of
M.c.v. of less than 79 1,’3 and to calculate Mentzer’s tentau-
vely proposed ratio, M.C.v./R.B.C. < 13.14

COMPARISON OF FOUR METHODS OF DETECTING THALASSAEMIA MINOR
A new method was developed heuristically to reduce the
number of false-positives and false-negatives that occurred
using other methods. The product of (M.c.v.)2 x mean corpus-
cular haemoglobin (M.C.H.), measured in units of 100, was cal-
culated on each sample. Due to the known incompatibility of
genotype AS with the diagnosis of &bgr;-thalass&aelig;mia, and the non-
random distribution of genotypes among red-cell indices," it
was only necessary to use the product for samples with geno-
type AA. Because males had 6.2% higher scores than females,
all male scores were reduced by 6-2% to provide a uniform cut-
off point.
RESULTS
-
The flow chart (fig. 1) indicates the numbers gener-
ated by each procedure. Out of a total of 25 302 there
were 138 people identified with haemoglobin A2 >4.5%.
Fig. 1-Resutts generated by each step in detection of &bgr;-thalass&aelig;-
mia.
Fig. 2-Distribution
of (M.C.V.)’M.C.H. showing thatassaemia
minor compared to normal and anaemic subjects.
693
I I i
Four methods of detecting these 138 people were com-
pared (see accompanying table).
The (M.C.V.)2M.C.H. values showed a wide separation
between the two groups (fig. 2). The mean score for the
thalassaemics was 1094&plusmn;239; the mean score for the AA
non-thalasssemics was 2218&plusmn;386. This difference was
highly significant, t 20-5, p<0-0001. The single per-
=
son who was missed had haematological values within
normal limits, and would elude detection by any method
based on red-cell indices. Similar carriers with normal
haematological values are probably quite rare, since none
were found among 1761 people with microcytosis for
whom A2 was measured and only one example was
found among 1973 people without microcytosis for
whom A2 was measured.
The (M.C.V.)lM.C.H. values also effected a wide separa-
tion between the means of heterozygotes (1094&plusmn;239)
and anaemic people (1814_+601), defining anaemia as a
hsemoglobin concentration below 12 g/dl in males and
below 11 g/dl in females (fig. 2), (r=7.7, P<00001).
DISCUSSION
Currently in the United States many people are being
offered education, testing, and genetic counselling for
variant haemoglobins. The counsellor is frequently in the
awkward position of telling a person with genotype AA
that he or she has no risk of producing a severely
anaemic child when the presence or absence of thalassae-
mia minor is unknown. If the person has (s-thalassaemia
minor, the risk is 0.03% if he or she mates at random
among Whites; 3% if he or she mates at random among
Blacks; and 25% if the mate also has p-thalassaemia
minor or sickle-cell trait.
If, as Shaw suggests,16 it will soon be obligatory for
physicians to inform patients of all genetic risks, then
tests for variant hxmoglobins must be able to detect
p-thalassaemia minor. Perhaps it is naive to expect to
find a reliable single simple test to detect thalasssemia
minor which is, in fact, the heterozygous expression of
one of a set of genes, one of which produces no hmmo-
globin A, one of which interacts with &bgr;-chain variants
producing a considerable amount of haemoglobin A but
producing no increase in the R.B.C.,17 and one of which
has no haematological or clinical manifestations. In an
earlier attempt to find a single test,18 relying on the pro-
duct of M.c.v. and M.C.H., we missed about 20% of
heterozygotes. It was then realised that more informa-
tion was obtained by squaring the M.c.v., that heterozy-
gotes for haemoglobins C and S tend to have thalassoemic
red cells, and that the presence of SA or an increase of
F in association with thalassaemia minor shifted the
(M.C.v.)2M.C.H. value into the normal range. 19 The pres-
ent proposed method missed only one heterozygote, and
produced only 4-4% false-positives.
The measurement of haemoglobin A2 is considered the
most reliable indicator of -thalasssemia, although there
694
is much variation between methods and between labora-
tories.l&deg; Although 3.5% A2 is the conventional threshold
between thalassaemia minor and normal," in this inves-
tigation the threshold was taken as 4-5% because the in-
termediate range included several people for whom the
diagnosis could not be supported clinically or geneti-
cally. Since the purpose of this study was to compare the
value of different methods of detecting heterozygotes, we
decided to remove the intermediate range from considera-
tion until genetic studies had been completed on everyone.
The M.C.V.2M.C H. score also provides a measure of
the likelihood that an individual is a heterozygote since
the frequency of heterozygotes was inversely propor-
tional to the score. Until some perfectly reproducible
method of measuring A2% is available, the M.C. V. 2M.C.H.
is useful as a confirmatory test, particularly within the
range of overlap between the two distributions.
The difficulty of detecting thalassaemia minor in the
presence of iron deficiency was diminished but not over-
come as the 1116 false-positives show, although one or
two might have been people with recently treated poly-
cythaemia. While this frequency of false-positives is too
high, it was nonetheless lower than that produced by the
method of Pearson et al. which was the only other
method to approach an acceptable rate of false-nega-
tives.
We are extremely grateful to Dr R. G. Schneider, Dr F. Clarke
Making Do
Efficiency in the National Health Service
INFORMATION AND DECISION IN A
SURGICAL UNIT
EFFORTS to improve the efficiency of a surgical service
may be compared to the stimulation of industrial pro-
duction. Both activities, may be hampered by bottle-
necks, those components of the total system which limit
the level of output and may cause other parts of the pro-
cess to operate at low efficiency and with spare capacity.
The surgical components include doctors, nurses, tech-
nicians, beds, operating-theatres, drugs and supplies.
Surgical output is hard to measure, so the identification
of overload and underload will be less easy than in the
typical manufacturing process. Nevertheless, improve-
ments are possible if the staff in a surgical unit are will-
ing to take the four steps I outline below.
INFORMATION AND EXPERIENCE
Attempts to draw an accurate picture of what is hap-
pening in the unit often disclose gaps between what is
thought to happen and what actually does. Human
nature forgets selectively, remembers the times when the
theatre list overran, while forgetting the occasions when
the session started late or the expected patient failed to
turn up. Well-remembered incidents confuse discussion
when, for example, irritating difficulties in finding a bed
prompt sweeping conclusions about the shortage of beds.
Experience differs from information, and leads to the
Fraser, Dr M. A. Spence, and Dr J. J. Hutton for their valuable cnu-
cisms of this manuscript.
Requests for reprints should be addressed to I.S.
REFERENCES
1. Lehmann, H., Huntsman, R. G. Man’s H&aelig;moglobins, p. 239. Amsterdam.
1974.
2 Malamos, B., Fessas, Ph., Stamatoyannopoulos, G. Br. J. H&oelig;mat. 1962, 8,
5.
3 Brook, S. R., Huntsman, R. G., Marshall, T. D., McClellan, D. S., Semple.
M. J., Crane, R. S. J. clin. Path. 1974, 27, 480.
4. Stockman, J. A., Weiner, L. S., Simon, G. E., Stuart, M. J., Oski, F. A. J. Lab
clin. Med. 1975, 85, 113.
5. Gouttas, A. in H&aelig;moglobin Colloquium (edited by H. Lehmann and K. Betke.,
p. 89. Stuttgart, 1962.
6. Kunkel, H. G., Ceppelini, R , M&uuml;ller-Eberhard, U., Wolf, J. J. clin. Invest.,
1957, 36, 1615.
7. Kabat, D., Koler, R. D. in Advances in Human Genetics (edited by H Harris
8.
and K. Hirschhorn); vol. 5, p. 168. New York, 1955.
Schmaier, A. H., Maurer, H. M., Johnston, C. L, Scott, R. B., Stewart,
L. M., J. Pediat. 1974, 84, 559.
9. Schneider, R G., Schmidt, R. M. in Abnormal H&aelig;moglobins and Thalass&aelig;-
mia (edited by R. M. Schmidt); p. 33. New York, 1975.
10. Betke, K., Marti, H. R., Schlicht, I. Nature, 1959, 184, 1877.
11. Huisman, T. H. J., Schroeder, W. A., Brodie, A. N., Mayson, S. M., Jakway,
J. J. Lab. clin. Med. 1975, 86, 700.
12. England, J. M., Fraser, P. M. Lancet, 1973, i, 449.
13. Pearson, H. A., O’Brien, R. T., McIntosh, S. New Engl. J. Med, 1973, 288,
351.
14. Mentzer, W. C. Lancet, 1973, i, 882.
15. Shine, I., Lal, S. Unpublished.
16. Shaw, M. in Proceedings of the Fifth International Congress of Human
Genetics 1976. Amsterdam, (in the press).
17. Pootrakul, S., Assawamunkongs, Na-Nakorn, S. Hemoglobin, 1976, 1, 75.
18. Lal, S., Shine, I. in Proceedings of the Fifth International Congress of
Human genetics 1976. Amsterdam, (in the press).
19. Shine, I. Unpublished.
20. White, J. M. in Abnormal H&aelig;moglobins and Thalass&aelig;mia (edited by R. M.
Schmidt); p. 19. New York, 1975.
21. Weatherall, D. J., Clegg, J. B. The Thalass&aelig;mia Syndromes, p. 113. Oxford,
1972.
often-expressed feeling that all parts of the service are
overloaded.
Spare capacity can be disguised by the Parkinsonian
expansion of work to fill the time and facilities available.
For example, extension of the patients’ stay in the unit
can mask the overprovision of beds, or excessive refine-
ments of surgical technique can hide a lack of routine
demand. Conversely, overloading can be partially dis-
guised : in a day-case theatre, the nursing staff were
employed on a 9-to-5 basis and yet the theatre schedule
almost guaranteed that patients were still awaiting dis-
charge at 7 P.M.
STANDARDS OF SERVICE
It is no part of the overall objective to increase the
number of patients treated at the expense of the quality
of service to the individual. The staff of the unit must
collaborate in agreeing on standards and on how per-
formances can be measured against them.
For example, in one hospital there was concern over
the effect of the number of beds and the amount of oper-
ating-theatre time on the throughput of patients. The
answer clearly depended on the standard set for the use
of theatre time, and particularly the inconvenience
caused by a session overrunning. A significant increase
was achieved in the number of patients who could be
treated if 90% of the theatre sessions finished within 30
minutes of the appointed time, rather than 20 minutes.
The effect of such decisions on performance standards is
so great that the efficiency exercise is pointless unless the
decisions are carefully debated in advance.
The administrative quality of the service must also be
considered. In the surgical unit, service standards will be
needed for (at least): the ability to accept emergency