Family Practice
© Oxford University Press 1996
VoL13, No 3
Printed in Great Britain
Sore throat management in general practice
Paul Little and Ian Williamson
Little P and Williamson I. Sore throat management in general practice. Family Practice 1996;
13:317-321.
This paper discusses primary care management of sore throat in the context of recent
national 'consensus' guidelines from the Drugs and Therapeutics Bulletin. The guidelines
advise taking a throat swab, using typical clinical features where swabs are not available,
and suggest that antibiotics shorten the duration of symptoms and prevent complications.
Systematic reviews and individual studies indicate that the evidence for prescribing antibiotics for most presentations of sore throat in general practice is marginal, and the benefits
are probably outweighed by the likely costs of antibiotics. Using clinical scorecards or symptom clusters to identify individuals who would benefit from treatment is insensitive with
low predictive value, although inexpensive. Using throat swabs as a gold standard for
diagnosis is inappropriate since they are neither very specific nor sensitive, and will greatly
increase costs of management. The relative lack of evidence for the efficacy of antibiotics
and for the use of throat swabs from primary care research, and also an unbalanced perspective of dangers and complications related predominantly to a secondary care setting,
underlines the problem of achieving valid consensus guidelines. Guidelines not firmly based
on evidence appropriate to the intended setting are more likely to be received sceptically
and hinder getting research into practice.
Keywords. Sore throat, clinical guidelines, tonsilitis, pharyngitis, throat swab.
Introduction
What is the evidence in each of these areas? This
deserves critical discussion not only because they are
important scientific and clinical issues for a very common clinical problem, but also in the context of
guidelines and applying evidence in practice. This paper
first discusses the burden of disease in primary care.
Then we will deal with the evidence for the efficacy
of antibiotics, of the use of throat swabs and of targeting
treatment to particular groups according to clinical
signs.
Although sore throat is one of the commonest conditions managed in primary care,1 management is still
controversial, and guidelines which suggest changing
management are likely to have significant implications
for both patients and the National Health Service." The
principle of 'getting evidence into practice'4 and the
use of guidelines is also particularly topical: guidelines
should be based on good evidence, be feasible and be
appropriate for the setting in which they will be used.3
What national guidelines are available for the management of sore throat in primary care? The respected and
widely read Drugs and therapeutics bulletin (DTB),
which develops consensus guidelines through an expert
panel system, recently reviewed the diagnosis and treatment of sore throat.6 All UK general practitioners
(GPs) receive the DTB and thus it is a very influential
source of advice and guidance for primary care. Their
guidelines advised (i) taking a throat swab since "the
precision in diagnosis is greatly increased"; (ii) where
throat swabs are not possible, targeting treatment according to "typical clinical features", especially in
children; (iii) that antibiotics shorten the duration of
symptoms and prevent complications.
Burden of disease for primary care
Estimates of attendance rates (per capita per annum)
for sore throat vary: 0.08-0.20 for pharyngitis and tonsillitis in single practices,17* 0.2 for sore throats in a
region,9 and for the—possibly atypical—practices in
the national morbidity survey10 approximately 0.1
(assuming 1 in 4 'respiratory' attendances are for sore
throat8 and allowing for re-attendance1112). Using the
latter conservative figure along with UK populations
estimates,13 and assuming that a consultation costs
£10,M then the cost of the GP consultations for sore
throat is approximately £60 million per annum even if
no drugs are prescribed—let alone other costs (see
below). The health service burden is increasing—a 14%
increase for respiratory attendances since the last
Received 28 December 1995; Accepted 28 January 1996.
Primary Medical Care, Aldermoor Health Centre, Aldermoor
Close, Southampton SO 16 5ST, UK.
317
318
Family Practice—an international journal
10
morbidity survey —despite no obvious change in the
pathological basis for sore throat during the last 30
years.13
Extra cost implications of throat swabbing
The cost of swabbing all acute sore throats routinely
in general practice would be approximately £40 million
annually, assuming £4 per swab* and probably at least
£3 for the extra time (2-3 min organizing the swab,
feeding back results to patients, pulling notes, documenting results, writing notes, writing scripts, etc.).
Do antibiotics prevent complications?
Rheumatic fever (RF) and glomerulonephritis (GN)
The studies on the prevention of RF were carried out
using penicillin injections on post-war barracked
military personnel, where attack rates were high
(0.3-5 %). 16 This may not be generalizable to the use
of oral antibiotics in a modern community setting with
lower attack rates, and where only 50% of children
complete a 10-day course.17 Individuals who have had
oral antibiotics in general practice may be as likely to
develop RF or GN as those not taking antibiotics.1819
The incidence of RF probably declined well before
the advent of antibiotics.20 In the USA, but not the UK
or Europe, the incidence temporarily increased in the
1980s21 and the steady decline since 1984 is difficult
to attribute to antibiotics since most subjects do not present to health professionals.11J2 The average UK GP
during a lifetime of practice has an approximately one
in five chance of seeing a case of either poststreptococcal GN or RF following a sore throat.1819
The principal problem of prescribing to prevent RF
and GN is that subjects may not have a significant sore
throat, and those with sore throats mostly do not attend
their GP.131819-22-24 Even if oral antibiotics in the community were of proven benefit, either GP surgeries
would need to be overwhelmed with patients or antibiotics would need to be freely available in the community to effectively prevent such complications. This
would result in greater likelihood of penicillin resistance
and greater use of more expensive drugs.^^
Septicaemia
The DTB underlined this complication, and quoted a
case review where the mortality from streptococcal
bacteraemia was 50%." Although there have been
more case reports of serious streptococcal infections
in Europe and the USA in the past 5 years,28-29 this
may not reflect a true increase. Furthermore, the majoTity are associated with an initial soft tissue focus (e.g.
skin, muscle, fascia) or deeper focus (e.g. empyema,
osteomyelitis) rather than pharyngitis,13'30 and with
strains not typical of pharyngeal types.13J0 This is also
borne out in the above UK case series:27 in only one
of the cases was the throat described as the possible
initial source, and the patient would not have been a
candidate for a trial of no antibiotics in any case (the
initial provisional diagnosis was meningitis). Thus it
is misleading to imply that streptococcal septicaemia
could be prevented by antibiotics prescription to patients
attending GPs with the normal range of presentation
of sore throat.
Suppurative complications
There is some evidence for a small protective effect of
antibiotics on the likelihood of developing either otitis
media or sinusitis.16 However, these studies are old
and took place when the complications were more common, involved small numbers of complications in any
case and were mainly on institutionalized servicemen,
so it is again very difficult to extrapolate the results to
modern general practice. The two studies addressing
the issue in general practice31-32 had wide confidence
intervals for the odds ratio for developing complications (greatly overlapping 1 for the prevention of otitis
media).
There have been few studies addressing the issue of
the prevention of quinsy using oral antibiotics: a postwar study in a military setting used intramuscular
penicillin.16 Quinsy did apparently decline significantly after the war,33 but how much of this is due to
improved socio-economic conditions and how much due
to antibiotics is not clear. John Fry, in a review of complications of sore throat in 30 years of practice (3000
cases of tonsillitis and presumably an equivalent number
of pharyngitis), documented three cases of serious cervical adenitis and 20 cases of quinsy34 all of which
settled with parenteral antibiotics (hence were possibly
peri-tonsillar cellulitis and not 'true' quinsy) and no
other complications; his prescribing for sore throats is
25% (i.e. a low prescriber). A search of computerized
notes (1990-1995) of 8000 current patients in our practice serving a deprived area of Southampton revealed
14 diagnoses of 'quinsy' or 'early quinsy': only three
attended with a prior sore throat; two of these had oral
antibiotics which did not prevent quinsy developing.
All cases settled with parenteral penicillin (hence were
possibly peri-tonsillar cellulitis). Thus quinsy or peritonsillar cellulitis is rare and most cases do not present
to the GP with prior sore throat.
Hence there is considerable doubt from current
evidence that oral antibiotics will effectively prevent
complications of sore throat, which are rare in any case.
Do antibiotics reduce symptom burden?
The evidence for symptomatic relief in sore throat is
also marginal. Of the five placebo controlled trials
done in general practice, there may be a small improvement in the number of patients well after 3 days
Sore throat management
FIGURE 1 Percentage of patients without sore throat following
presentation to a general practitioner; patients randomized either
to penicillin (n = 256) or placebo (n = 272)
Reproduced by kind permission of the authors and the practitioner
from Whitfield and Hughes, 1981.33
comparing penicillin to control,16 the figures vary
from 12 % (the largest trial) to 38 %. However the largest
trial (n = 528) so far33 showed this benefit only for a
small subgroup, and furthermore demonstrated that
illness was not shortened at all irrespective of initial
presentation with fever, purulent tonsils or lymphadenitis (see Figure 1).
Another major problem in prescribing antibiotics to
ease symptoms is the evidence from placebo controlled
trials (in Group A Beta Haemolytic Streptococcus
(GABHS)-swab-positive patients) that the recurrence
rate is significantly higher in subjects treated with
penicillin immediately compared to treatment delayed
for 48-56 hours (37% and 16% respectively relapsed
within 4 months). This may be due to preventing immunity to streptococcus developing.12 The effect of
delaying antibiotics on symptom resolution is similar
to not giving antibiotics at all since most of the symptom reduction is in the first few days of presentation.12
Thus in disease that has a time course of approximately
7-10 days either in treated or in untreated individuals,16 and identifying patients with GABHSpositive throat swabs, patients treated early with antibiotics will be symptomatically better off by 12% for
the 3-5 days following taking antibiotics,12 but 20%
more patients will be exposed to the whole range (i.e.
100%) of symptoms due to relapse that they would not
have been exposed to if they were not given antibiotics,
i.e. probably cancelling any net benefit. Thus antibiotics
probably do not reduce the symptom 'burden'.
Could a subgroup be targeted to improve
outcome, and would this affect prescribing?
Could particular subgroups be identified? Unfortunately,
symptom clusters do not seem to be a good indicator
of streptococcal infection or antibiotic response.33 The
319
'typical' features of sudden pain, fever, exudate and
tender adenopathy 'especially' in children* are only
present in approximately 15% of cases of streptococcal pharyngitis.3* In US emergency centres, the Breese
symptom score although specific (94%) was only
modestly predictive (positive predictive value 59%) and
not sensitive (26%) compared to throat swabs.37
This leads on to the guidelines using the throat swab
as a diagnostic 'gold standard'.6 The sensitivity and
specificity of the throat swab are low—26-30% and
73-80% respectively—when compared to a rise in the
anti-streptolysin O titre (ASOT) titre.24-33 The ASOT
titre is likely to be more important than throat swab in
predicting outcome in view of the 100% association between definite cases of RF and either high initial level
of ASOT or a rise in ASOT.24 The low predictive value
of throat swabs relates to the high symptomless carrier
rates—6-^0 %24-SM1 especially in children.6^ The DTB
conclusion that streptococcal "infection" is "more common" in children may just reflect higher carriage rates.
The high carrier rate would explain why ASOT rises,
in only 20-50% of individuals with sore throat and
GABHS-positive swab,394243 and may contribute to the
apparent 'failure' rate of approximately 20% with
penicillin therapy." Although ASOT titres are liable
to be a much better indication of immunologically
significant infection rather than the throat swab (which
cannot distinguish carriage of streptococci from infection), the delay, cost and inconvenience of serial ASOT
titres rule out routine clinical use.
Even if throat swabs were taken there is the difficulty of using the results in changing treatment decisions: even with a rapid antigen test, where the results
were available comparatively quickly (within hours,
unlike the 48-56 hours for the throat swab), the results
made a difference to prescribing decisions in only 13 %
of cases.44
Costs of prescribing
The marginal benefit of prescribing must be weighed
against the possible costs. Routine prescribing encourages dependence and re-attendance:11 40% re-attend
during the following 6 months if not prescribed an
antibiotic for sore throat versus 60% if an antibiotic is
prescribed, thus taking up surgery time with a largely
self-limiting condition and resulting in considerable
opportunity cost for both the GP and the patient. The
effect of consistent advice over a long time is likely to
further reduce re-attendance. Taking the 20% figure
as a conservative estimate of the reduction in attendance,
this would save an estimated £12 million annually.
There is no evidence about the further medicalization or anxiety-inducing effects of using throat swabs
and a 2-3 day delay in 'diagnosis', but this is likely
to increase the medicalizing effect (by creating the
320
Family Practice—an international journal
perception that it is important to do a throat swab with
a sore throat to allow proper management and hence
that it is important to attend the GP to have a swab
taken). Like all screening it may exacerbate anxiety
unless handled carefully.45 In the context of greater attendance for respiratory and trivial problems, we need
to be sure that policies will not encourage this process
further by unnecessarily medicalizing self-limiting
conditions.
In addition to the re-attendance costs for patient and
surgery, there are other costs: side-effects of antibiotic
use such as the 3.8% allergy rate,4* between 10 and
60% of children experiencing diarrhoea on treatment,474* oral and genital Candida albicans particularly
if broad spectrum antibiotics are used, and unwanted
pregnancy following reduced efficacy of oral contraceptives.
Considering dangerous side-effects, following
penicillin treatment the estimated incidence of
anaphylaxis is between 1.5 and 4 cases per 10 000 with
two patient deaths per 100 000.^ If every case of
acute pharyngitis and acute exudative tonsillitis were
treated with penicillin, i.e. approximately 100-500 cases
per GP per year, then in the average working lifetime
(35 years) a GP would expect to have an approximately
one in three to one in fourteen chance of seeing death
by anaphylaxis from treating sore throat, i.e. of the same
order of magnitude as seeing a case of nephritis or RF
following a sore throat, neither of which have a high
fatality rate.
Behaviour and prescribing
Discussing prescribing as if patients were just symptom and disease clusters is highly artificial. Psychosocial
factors for both the doctor and the patient are important determinants of prescribing,49-30 and good practice
requires their acknowledgement and exploration before
prescribing. GPs probably perceive more pressure to
prescribe than exists, given that 41 % of patients entering consultations expect a prescription but 67 % leave
with one.31 Patients probably find explanation, symptomatic treatment, and no antibiotics an acceptable
option even in painful upper respiratory conditions as
evidenced by the uncontrolled report of the acceptability
of no antibiotic treatment in otitis media,52 and a controlled trial of no antibiotic prescription with advice in
sore throat.11
probably outweighed by the likely costs of antibiotics:
the complications, including death (rare); diarrhoea
and allergy (common); relapse (common); medical
dependence and associated health service costs (common); and possible unknown psychological costs. The
exception would be the rare individuals who are septicaemic or with quinsy where the risks of not treating
are likely to be much greater.
Using clinical scorecards or symptom clusters to identify individuals who would benefit from treatment is
insensitive with low predictive value, although inexpensive. Using throat swabs as a gold standard for
diagnosis is inappropriate since, although they are
reasonably specific, they are insensitive and not very
predictive and will greatly increase costs of
management.
GPs should share with patients the evidence about
the modest efficacy, likely complications and relapse
rate with antibiotics for most presentations of sore
throat, and negotiate to improve the symptomatic
management of sore throat without relying routinely on
antibiotics.
The relative lack of evidence for the efficacy of antibiotics and for the use of throat swabs from primary
care research, and also an unbalanced perspective of
dangers and complications related predominantly to a
secondary care setting, underlines the problem of
achieving valid consensus guidelines. Guidelines which
are not adequately based on evidence in the setting in
which they will be used, and do not sufficiently take
account of the practical implications of implementation,
are likely to be received sceptically and adhered to
poorly and will hinder getting research into practice.4-5
References
1
2
3
4
3
6
7
Summary
We argue that the evidence for prescribing antibiotics
for sore throat is marginal. Complications are rare, and
on current evidence may not be preventable by prescribing for sore throat in primary care. The benefits are
8
9
10
Hodgkin K. Towards earlier diagnosis in primary care. Edinburgh: Churchill Livingstone, 1978.
Little PS, Williamson I. Controversies in management Are antibiotics appropriate for sore throats? Costs outweigh the
benefits. BrUedJ 1994; 309: 1010-1011.
Schwartzman P. Controversies in management. Are antibiotics
appropriate for sore throats? Careful prescribing is beneficial.
BrUedJ 1994; 309: 1011-1012.
Dunning M, McQuay H, Milne R. Getting a Grip. Health Serv
J 1994; Appendix H: 24-26.
University of Leeds, University of York, Royal College of
Physicians. Implementing Clinical Practice Guidelines.
Effective Health Care Bull 1994; 8: 1-12.
Diagnosis and treatment of streptococcal sore throat. Drugs
Ther BuU 1995; 33: 9-12.
Fry J. Medicine in three societies. Lancaster: Medical and
Technical Publishing Co. Ltd, 1969.
Howie J. A new look at respiratory illness in general practice.
J RCoUGen Pract 1973; 23: 895-904.
Howie J, Richardson I, Gill G, Dumo D. Respiratory illness
and antibiotic use in general practice. / R Coll Gen Pract
1971; 21: 657-663.
HMSO, OPCS. Morbidity statistics from general practice:
Fourth national study 1991. London: HMSO, 1994.
321
Sore throat management
11
12
13
14
15
16
17
18
19
20
21
22
23
Herz MJ. Antibiotics and adult sore throat—an unnecessary
ceremony. Fam Pract 1988; 5 : 196-199.
Pichichero M E , Disney F , Talpey W ex al. Adverse and
beneficial effects of immediate treatment of Group A
b-haemolytic streptococcal pharyngitis. Fed Infect Dis J 1987;
6: 635-643.
Central Statistical Office. Annual abstract of statistics. London:
HMSO, 1995.
Daly E, Vessey MP, Barlow D , Gray A , McPherson K, Roche
M. Hormone replacement therapy in a risk benefit perspective. In Berg G, Hammar M (eds). The modem management
of die menopause. London: Parthenon, 1993.
Shulman ST, Gerber M A , Tanz RR, Markowitz M D . Streptococcal pharyngitis: the case for penicillin-therapy. Fed
Infect Dis J 1994; 13: 1-7.
Del Mar C. Managing sore throat: a literature review n . Do
antibiotics confer benefit? MedJAust 1992; 156: 644-649.
Chamey E, Bynum R, Eldridge D et al. How well do patients
take oral penicillin? Pediatrics 1967; 4 0 : 188-195.
Howie J, Foggo B. Antibiotics, sore throat and rheumatic fever.
J R Coll Gen Pract 1985; 35: 223-224.
Taylor JL, Howie J. Antibiotics, sore throat and acute nephritis.
J R Coll Gen Pract 1983; 3 3 : 783-786.
White P. Changes in the prevalence of various types of heart
disease in New England. JAMA 1953; 33: 783-786.
H M S O . Hospital inpatient enquiry. 1979-1985.
London:
HMSO, 1986.
T a u b e r K A , Rowley A M , Shulman S T . A seven year
(1984-1990) U . S . nationwide hospital survey of acute
rheumatic fever (Abstract). Am Heart Assoc Sci Sessions
1992; 8 6 : I-36I.
Banks M , Beresford S, Morrell D , Waller J, Watkins C . F a c tors influencing the demand for primary care in women aged
2 0 - 6 4 years: a preliminary report. Int J Epidemiol 1975; 4 :
31
32
33
34
33
36
37
38
39
40
41
42
43
189-195.
24
23
26
27
28
29
30
Valkenburg H A , Haverkorn M J , Goslings W R O . Streptococcal
pharyngitis in the general population. II. T h e attack rate of
not treated with penicillin. J Infect Dis 1971; 124: 3 4 8 - 3 5 8 .
B a q u e r o F , Martinez-Beltran J, Loza E . A review of antibiotic
resistance patterns of Streptococcus pneumoniae in Europe.
J Antimicrob Chemother 1991; 2 8 (suppl C ) : 3 1 - 3 8 .
Tanz RR, Hofer C. A survey of U.S. pediatrician strategies for
managing Group A streptococcal pharyngitis. XII Lanoefield
Symposium, St Petersburg, Russia, 1993.
Francis J, Warren RE. Streptococcus pyogenes bacteraemia in
Cambridge: a review of 67 episodes. Q J Med 1988; 6 8 :
603-613.
Stevens D L , Tanner M H , Winship J. Severe Group A streptococcal infections associated with toxic shock like syndrome
and scarlet fever toxin. N Eng J Med 1989; 3 2 1 : 1-7.
Shulman ST. Invasive and toxin-related diseases caused by
group A streptococci. Ped Infect Dis J 1991; 10: 5 2 8 - 5 3 1 .
Johnson D R , Stevens D L , Kaplan E L . Epidemilogical analysis
of G r o u p A streptococcal serotypes associated with severe
systemic infections, rheumatic fever OT uncomplicated
pharyngitis. / Infect Dis 1992; 166: 374-382.
44
43
46
47
48
Chappie PAL, Franklin LM, Paulett JD. Treatment of acute
sore throat in general practice. Therapeutic trial with observations on symptoms and bacteriology. Br Med J 1956; 1:
705-708.
Taylor B , Abbott G D , Kerr M , Fergusson D M . Amoxycillin
and co-trimoxazole in presumed viral infections of childhood:
placebo controlled trial. Br Med J 1977; 2 : 5 5 3 - 5 5 4 .
Peter G , Smith A . Streptococcal infections of the skin and
pharynx. N Eng J Med 1977; 2 9 7 : 3 1 1 - 3 1 7 , 3 6 5 - 3 7 0 .
F r y J. Acute throat infections. Update 1979; 1 8 : 1181-1183.
Whitfield M J , H u g h e s A O . Penicillin in sore throat. Practitioner 1981; 2 : 1061-1064.
Kiselica D . Group A beta haemolytic streptococcal pharyngitis:
current clinical concepts. Am Fam Phys 1994; 4 9 :
1147-1154.
Reed B D , Huck W , French T. Diagnosis of Group A b haemolytic strep using clinical scoring criteria, directigen
1-2-3 group A streptococcal test and culture. JAMA 1990;
150: 1727-1732.
Del Mar C. Managing sore throat: a literature review I. Making the diagnosis. Med J Austr 1992; 156: 572-575.
Caplan C . A case against the use of the throat culture in the
management of streptococcal pharyngitis. J Fam Pract 1979;
8: 485-190.
Chun L T , Reddy V. Occurrence and prevention of rheumatic
fever among ethnic groups of Hawaii. Am J Dis Child 1984;
138: 4 7 6 - ^ 7 8 .
Feery BJ, Forsell P, Gulasekharam J. Streptococcal sore throat
in general practice. MedJAust
1976; 1: 989-991.
Kaplan E L , Franklin HT, Dudding BA, Wannmaker L W .
Diagnosis of streptococcal pharyngitis: differentiation of active infection from the carrier state in the symptomatic child.
J Infect Dis 1971; 123: 4 9 0 - 5 0 1 .
Siegel A C , Johnson E E . Controlled studies of streptococcal
pharyngitis in a pediatric population II. Behaviour of type
specific immune response. N Eng J Med 1961; 2 6 5 : 5 6 6 - 5 7 1 .
Burke P , Bain J, Lowes A, Athersuch R. Rational decisions and
managing sore throat. Br Med J 1988; 296: 1646-1649.
Marteau T . Reducing the psychological costs. BrMedJ 1991;
301: 2 6 - 2 8 .
Madden TA. Adverse penicillin reactions in the records of a
general practice. / R Coll Gen Pract 1977; 2 7 : 73-77.
Bolme P , Eriksson M . Influence of diarrhoea o n the oral a b sorption of penicillin V and ampicillin in children. ScandJ
Infect Dis 1975; 7 : 141-145.
Hoigne R, Zoppi M , Sonntag R . Penicillins. In M N G Dukes
(ed.). Meyler's Side Effects of Drugs. Elsevier, 1992:
600-622.
49
Howie J. Clinical judgement and antibiotic use in general prac-
30
Howie J, Bigg A R . Family trends in psychotropic and antibiotic
prescribing in general practice. BrMedJ 1980; 1: 8 3 6 - 8 3 8 .
Cartwright A , Anderson R. General practice revisited: a second
study of general practice. London: Tavistock, 1981.
Pitts J. Shared decision making in the informed treatment of
acute otitis media. Practitioner 1987; 231: 1232-1233.
31
32
%tice.
BrMedJ 1976; 2: 1061-1064.