Respiratory tract infection is infection of the breathing passages, which extend from the nose to the alveoli (the tiny, balloon-like sacs in the lung). This type of infection is divided into upper and lower respiratory tract infections. Upper respiratory tract infections affect the nose, the pharynx (throat), the sinuses, and the larynx (voice-box). Examples are the common cold, and inflammatory conditions such as pharyngitis, tonsillitis, sinusitis, laryngitis, and croup. Lower respiratory tract infections, which involve inflammation of the trachea (windpipe), bronchi, and lungs, include acute bronchitis, acute bronchiolitis, and pneumonia.

Upper respiratory tract infection (URTI) in detail - technical

Essentials

Acute upper respiratory tract infections (URTIs) include acute pharyngitis/tonsillitis and acute rhinitis. Acute sinusitis, acute otitis media, and influenza also come under the umbrella of infections of the upper respiratory tract. Otitis media and influenza will be discussed elsewhere: this chapter concentrates on acute pharyngitis/tonsillitis, acute rhinitis, and acute sinusitis.

Acute URTIs are the commonest reason for patients to seek medical advice in the United Kingdom, and nearly all cases are managed in primary care. Respiratory tract infections are also the commonest reason for antibiotics to be prescribed, leading to serious concern that the inappropriate use of antibiotics for predominantly self-limiting conditions will foster the development of antibiotic resistance, with the danger that serious infections will become untreatable. Thus it is currently an international priority to discourage the use of antibiotics where there is poor evidence of their efficacy. The evidence for the effectiveness of treatments for URTI in this chapter comes from a search of the Cochrane Library databases of systematic reviews and randomized controlled trials.

Pharyngitis/tonsillitis

Clinical presentation

Pharyngitis is caused by both bacterial and viral organisms, and has been somewhat arbitrarily divided into nasopharyngitis (with nasal symptoms, i.e. rhinitis), and pharyngitis or tonsillopharyngitis (without nasal symptoms). Causal organisms include: group A β-haemolytic streptococcus; adenoviruses; influenza A and B; parainfluenza 1, 2, 3; Epstein–Barr virus (EBV); enteroviruses; Mycoplasma pneumoniae; and Chlamydia pneumoniae.

In addition to a sore throat, pharyngitis is often accompanied by fever, headache, nausea, vomiting, anorexia, and sometimes abdominal pain, with or without enlarged and tender cervical lymph nodes, tonsillar erythema and exudate. Scarlet fever has a characteristic ‘scarlatiform’ rash caused by group A β-haemolytic streptococcal exotoxins. Infectious mononucleosis due to EBV may present with or without exudative tonsillitis, cervical or general lymphadenopathy, palatal petechiae, splenomegaly, rhinitis, and cough.

Throat swabs, rapid tests, and clinical algorithms

Antibiotics can be targeted to those patients who have positive throat swabs for group A streptococcus, a positive rapid streptococcal test, or clinical characteristics associated with a positive throat swab (e.g. the Centor criteria of fever, tonsillar exudate, anterior cervical adenopathy, and absence of cough). However, the throat swab has limitations: in both unselected and clinically selected populations in primary care practice it is neither particularly sensitive nor specific when compared to a rise in antistreptolysin O titres (ASOT) or anti-DNAse B titres. A rise in ASOT or anti-DNAase B is probably a better indicator of serious infection and predicts complications, but these are not suitable for clinical diagnosis. The results of throat swabs take days to return to the clinic, and they greatly increase the costs of managing what is mostly a self-limiting condition. Furthermore, evidence suggests that in practice clinicians do not use the results, even of rapid tests, and that the overall accuracy of decision-making is little changed when they are used.

Attempts to derive algorithms or clinical decision rules based on the throat swab have the same limitations of validity as the throat swab itself. Although clinical scoring methods may provide a crude method of identifying patients at a higher risk of complications (see below), better evidence is needed of the effects of using such scoring methods in practice.

Treatment

Antibiotics for symptoms

The Cochrane review of the efficacy of antibiotics for the treatment of sore throat indicates that antibiotics have modest benefit in reducing the duration of symptoms—by approximately 16 h for an illness lasting on average 8 days in total—and even less for trials which did not restrict selection according to the results of throat swabs. This marginal benefit of antibiotics in resolving symptoms suggests that, for patients who are not unwell systemically, the physician should either not prescribe, or use a delayed prescribing approach, advising the patient to wait for several days before collecting or using their prescription. Both these approaches have been shown in a large randomized controlled trial to be acceptable, to change attitudes and behaviour, and not to delay symptom resolution appreciably.

In the context of a likely streptococcal infection, trial evidence suggests that delaying the prescription results in 20% fewer recurrences than the immediate prescriptions of antibiotics, presumably because antibiotics modify local or systemic immune mechanisms. Thus, any marginal symptomatic benefit from an immediate prescription of antibiotics for the current illness must be weighed against the disadvantage that the patient is more likely to suffer symptoms from a recurrence.

Antibiotics to prevent complications

The Cochrane review of antibiotics for treating a sore throat supports the use of antibiotics to prevent complications, but the evidence is limited by both clinical importance and generalizability. For the commoner complications, e.g. otitis media, 200 people would have to be treated to prevent one case of a self-limiting illness: in other words, it is not important clinically. For the rarer complications—rheumatic fever and glomerulonephritis—the evidence is not generalizable; for instance, evidence of efficacy in rheumatic fever is based largely on trials where intramuscular penicillin was used in barracked military personnel after the Second World War. This evidence cannot be sensibly applied to modern settings where the attack rate is much lower and oral antibiotics are used. However, the benefits of antibiotics are likely to be greater in settings where complications are much more common.

The commonest complication of practical importance to health services is quinsy (peritonsillar abcess), but this is still relatively uncommon—about 1 in 400 following presentation in primary care with sore throat. The Cochrane systematic review, which demonstrates that antibiotics prevent quinsy, relies on data from patients with tonsillitis who were systemically unwell enough to be admitted to hospital shortly after the Second World War, when the prevalence of quinsy in untreated patients was very high (1 in 18). Clearly, this data cannot be extrapolated to patients presenting from modern populations who are not systemically unwell, treated with oral antibiotics, and where the prevalence of quinsy is much lower.

Quinsy following sore throat is possibly slightly more common (1 in 60) in those who are unwell, with three out of four Centor criteria, most of whom have fever. Rigorously conducted placebo-controlled trials in patients with these criteria suggest quinsy may be prevented by oral penicillin, but in routine clinical practice, where compliance is not assessed, the preventive benefit of penicillin is not likely to be 100%, as reported in the trials where compliance was assured. There are limited routine data which suggest that many patients who develop quinsy after being seen in primary care do this despite being given penicillin. Whether using the clinical Centor criteria is better than the primary care physician’s assessment of how unwell patients are is unclear: 20% of those considered ‘not to be very unwell systemically’ by the physician will still have three out of four of the Centor criteria, and we have limited data to assess whether the criteria predict the very few individuals who will develop quinsy.

Thus, where the primary care physician judges the patient to be both systemically unwell and/or have three out of four of the Centor criteria, it would be reasonable to treat with penicillin, or at least discuss with patients the likely risks of nontreatment. There is some observational evidence to suggest that there has been no increase in admissions with quinsy since the uptake of the delayed-prescribing strategy in the United Kingdom.

Lemierre syndrome—a rare complication

This syndrome, caused by fusobacterium—an anaerobe that is part of normal throat flora—has been highlighted recently following a rise in reports between 1990 and 2000 to about 20 per year in the United Kingdom. A patient with pharyngitis does not improve, remains pyrexial, and develops pharyngeal swelling due to a local abscess. Internal jugular thrombosis or embolism to the lungs commonly occurs or is suspected, and in such cases prompt referral to hospital is needed. The condition responds to metronidazole, but—since the differential diagnosis is incipient quinsy—high-dose penicillin should also probably be given. All the isolates in recent case series have been sensitive to metronidazole, 2% resistant to penicillin, and 15% resistant to erythromycin. However, to encourage increased prescribing of antibiotics on the basis of an increase in Lemierre’s syndrome is unwarranted: it would increase the dangers of both resistance and anaphylaxis—and anaphylaxis, although rare, is still commoner than Lemierre’s syndrome.

Which antibiotic and for how long?

A systematic review has concluded that cephalosporins may provide some additional benefit to penicillin in streptococcal tonsillitis. However, there are several problems with advocating cephalosporin use as first-line agents: the extra benefit is small (85.8% vs 93.6% clinical cure); the trial data assume that the subgroup of patients with streptococcal infections has been identified, so the effect will be less in unselected patients; and whether cephalosporins provide additional benefit to a high dose of penicillin V is unclear. Thus penicillin is probably still the best first-choice antibiotic.

If an oral antibiotic is to be prescribed, then it is probably preferable to give a narrow-spectrum antibiotic (penicillin V) to minimize side effects and the risk of resistance developing. If penicillin V is used, there are arguments for using a large dose given the variable absorption (e.g. 2 g/day adults and 1 g/day children). A 10-day course will better eradicate streptococcus, but the clinical significance of this is unclear. Longer courses have the disadvantage of poorer compliance, and greater likelihood of antibiotic resistance developing in the long term. Twice-daily dosing of penicillin V compared with the same total dose over four doses per day may result in better compliance and better clinical/microbiological outcomes. Intramuscular injection of penicillin can be used, although in practice is rarely employed in the United Kingdom. Ampicillin will cause a rash in patients with infectious mononucleosis, so erythromycin is a suitable second-line agent among patients with penicillin allergy.

Treatment of patients with rheumatic fever

Patients who have had one attack of rheumatic fever are at a higher risk from new infections since they are likely to develop recurrent attacks of rheumatic fever and complications. Although most of the evidence for the prevention of rheumatic fever comes from old trials in unusual settings, it seems reasonable to treat patients with a past history who are at a high risk of recurrence and secondary complications, since what evidence there is suggests penicillin prevents rheumatic fever. 

Other medical treatments

Treatment with aspirin in children is contraindicated because of the small but avoidable risk of Reye’s syndrome. There are several trials of the use of nonsteroidal antiinflammatory drugs (NSAIDs) in providing effective relief of pain and fever in tonsillitis and pharyngitis. However, the limited trial evidence comparing them with standard treatment (paracetamol) does not clearly demonstrate their superiority. Limited trial data suggest that other useful analgesic adjuncts may include caffeine, and benzydamine hydrochloride gargle. There is also evidence from a systematic review that patients with more severe presentations who are receiving antibiotics may benefit from steroids. However, there is probably less benefit from oral steroids compared with parenteral preparations, and the effectiveness of steroids when no antibiotics are prescribed and in less severe presentations is unknown.

Recurrent attacks

Surgery

A Cochrane review has assessed the role of surgery for recurrent sore throat. The much-quoted Paradise trial, documented in the Cochrane review, assessed tonsillectomy for selected children with severe symptoms defined by the ‘Paradise’ criteria as meaning seven or more episodes of well-documented, clinically important and adequately treated throat infection in the preceding year, or five or more times per year for each of the preceding 2 years, or three or more times per year for each of the preceding 3 years. The study found approximately one less episode of throat infection rated as moderate or severe per child in the surgical group, but the trial was small and was criticized by the Cochrane review for imbalances of important baseline characteristics, although these are perhaps unlikely to affect the inferences. Other trial evidence has shown that for children with moderately frequent throat infections (on average three in the previous year) a wait-and-see approach results in acceptable symptom control, although those with more than three infections per year had some benefit from immediate operation—one fewer episode of sore throat. When discussing options with parents the modest benefits of surgery must be weighed against its disadvantages: tonsillectomy will result in considerable postoperative pain and some complications (4–7% requiring operative surgery for haemorrhage, or other significant symptoms such as severe nausea and dehydration).

Other treatments

There is preliminary trial evidence for the use of α-streptococci spray, immune stimulants, and pneumococcal vaccination, but further confirmation is required.

Nasal congestion and rhinorrhoea

Nasal symptoms are a common reason for attending the doctor. They may be due to a variety of causes—commonly acute viral infection (common cold), allergic rhinitis and sinusitis, vasomotor rhinitis and rhinitis medicamentosa, and less commonly atrophic rhinitis, hormonal rhinitis, and mechanical/obstructive rhinitis. Colds are responsible for significant morbidity: on average there are 0.4 episodes and 1.2 days of restricted activity per person per year for the common cold.

Acute rhinitis

Symptoms are acute nasal congestion and rhinorrhoea, mild malaise, sneezing, sore throat, variable loss of taste and smell, and usually last from 1 to 2 weeks unless sinusitis is present. Examination reveals a hyperaemic and oedematous mucosa, with or without purulent secretions.

Treatment

Symptomatic

Trial evidence supports the use of both oral and topical decongestants for the symptoms of rhinitis. Intranasal ipratropium bromide is also effective symptomatic treatment, but is only available (in the United Kingdom) on prescription. However, topical decongestants should probably not be used for more than a maximum of 7 days: rhinitis medicamentosa starts to develop at 10 days. Because of their moderate systemic effects, care should be taken with oral decongestants in patients with heart disease and hypertension. Saline drops are commonly advocated, but saline or medicated nose drops have been shown to be ineffective in trials in both children and adults. A Cochrane review suggests that steam may provide some relief of symptoms.

Antibiotics

The use of antibiotics for the common cold has been assessed in a Cochrane systematic review and shown to provide modest benefit.

Other

Reviews of trials indicate little benefit from antihistamines or zinc lozenges. A Cochrane review of the herb echinacea demonstrated positive results in most studies, but there was not enough evidence to recommend the use of a specific product.

Acute sinusitis

Diagnosis

Acute sinusitis, usually defined as an infection that lasts for less than 3 weeks, is an uncommon complication of coryzal illness and pharyngitis. There is no absolute standard against which symptoms and signs can be compared for accuracy of diagnosis: aspiration by sinus puncture is probably the definitive investigation, since it indicates the presence of infecting organisms, but for obvious reasons this is rarely performed, and contamination by commensal organisms can occur.

Four-view radiographs show acceptable agreement with aspiration and culture, although only moderate interobserver agreement. The United States Agency for Health Care and Policy Research has reviewed the diagnosis and treatment of sinusitis: combining all studies comparing sinus radiographs with sinus puncture demonstrated a sensitivity of 73% and specificity of 80%. A history of purulent nasal discharge, maxillary toothache, purulent secretions on examination, poor response to decongestants, and abnormal illumination of the sinuses, are all predictive of sinusitis defined using four-view radiographs as the standard: four or more symptoms or signs giving a likelihood ratio of a positive test of 6. A problem with sinus illumination as a diagnostic tool in primary care is that it performs differently in different settings, probably due to operator sensitivity. There is preliminary evidence comparing symptoms with CT as the ‘standard’, which is justified since the presence of fluid and total opacification of the sinuses on CT predicts antibiotic response. Purulent rhinorrhoea, purulent secretion in the cavum nasae, a history of ‘double sickening’ (getting better, then getting worse again), and an erythrocyte sedimentation rate higher than 10 are predictive of a CT diagnosis of sinusitis—three of these features giving a likelihood ratio of a positive test of 1.8.

However, using a four-item clinical risk score—of purulent rhinorrhoea with unilateral predominance, local pain with unilateral predominance, bilateral purulent rhinorrhoea, and presence of pus in the nasal cavity—is as sensitive and specific as any other method in predicting the results of sinus puncture. Thus, for acute sinusitis, diagnostic tests are not currently indicated, and until valid near-patient tests are available, clinical targeting probably performs as well as any other method.

Treatment

Antibiotics

A Cochrane review of all controlled trials suggests that the absolute benefit for symptom resolution is moderate, and must be balanced against the disadvantages of prescribing antibiotics. Furthermore, this review mostly includes trials where there was radiological confirmation—which is not appropriate for diagnosis in primary care—and does not include all the trials from primary care, which show moderate or no effect. An IPD (individual patient data) meta-analysis of trials using a clinical diagnosis documented a number needed to treat of 15 for all patients and 8 for those with purulence, and no greater benefit for those with symptoms for longer than a week. Thus both the effectiveness and cost-effectiveness of antibiotic treatment of acute sinusitis in primary care is questionable for most patients.

Other

Preliminary trial evidence shows that decongestants are unlikely to be helpful. There is limited evidence that antihistamines may be helpful for patients with a history of allergic rhinitis who develop sinusitis, and some evidence that proteolytics (e.g. bromelain) and mucolytics may help. There is mixed trial evidence for the benefit of topical steroids. Although trials of NSAIDs suggest they are helpful, they may not be significantly more effective than paracetamol.

Further reading  

Systematic review of diagnosis of sore throat

Del Mar C (1992). Managing sore throat: a literature review. I: Making the diagnosis. Med J Austral, 156, 572–5.

Antibiotics and recurrent sore throat, the ‘medicalizing’ effect of prescribing antibiotics, and the use of delayed prescriptions

Little PS, et al. (1997). An open randomised trial of prescribing strategies for sore throat. BMJ, 314, 722–7.

Little PS, et al. (1997). Reattendance and complications in a randomised trial of prescribing strategies for sore throat: the medicalising effect of prescribing antibiotics. BMJ, 315, 350–2.

Use of the ‘Centor’ criteria to target antibiotic prescribing for sore throat

Zwart S, et al. (2000). Penicillin for acute sore throat: randomised double blind trial of seven days versus three days treatment or placebo in adults. BMJ, 320, 150–4.

Diagnosis and treatment of sinusitis

US Department of Health and Human Services (1999). Evidence report/technology assessment number 9: diagnosis and treatment of acute bacterial rhinosinusitis. AHCPR, Rockville, MD.

Young J, et al. (2008). Antibiotics for adults with clinically diagnosed acute rhinosinusitis: a meta-analysis of individual patient data. Lancet, 371, 908–14.

Diagnosis and management of rhinitis

Joint Task Force on Practice Parameters. (2008). The diagnosis and management of rhinitis: An updated practice parameter. J Allergy Clin Immunol, 122, S1–S84.

Cochrane reviews