Respiratory infections are caused by viruses, or bacteria, or both. If the illness is entirely viral in origin, an antibiotic will not help. If there is a bacterial component, antibiotic treatment will sometimes help, and may be vital. It is often difficult to recognize when bacteria may be involved in respiratory infection as secondary bacterial infection may complicate viral respiratory infections. However, the key question in the decision about treatment with antimicrobials is: do the benefits to the patient outweigh the risks? It is not necessary to prescribe antimicrobials for all bacterial respiratory infections.
When considering the role of antimicrobial chemotherapy it is important to reflect on the epidemiology of infection in the twentieth century (Fig. 19.1). There are two striking features: the doubling of mortality in 1918, caused by the influenza pandemic; and the steep decline in mortality through the first half of the century, before the arrival of antimicrobial chemotherapy or vaccines. The introduction of antimicrobial chemotherapy did accelerate the decline in mortality from some respiratory infections (e.g. otitis media, pneumonia, and tuberculosis) but had no impact on others (e.g. bronchitis). Two conclusions can be drawn from Fig. 19.1first, it is clear why there is so much concern about the possibility of an influenza pandemic given the massive impact on mortality of the 1918 pandemic. Second, antimicrobial chemotherapy has not had the dramatic effect on mortality from infections that is popularly attributed to it. The major impact in the first half of the twentieth century came from improvements in public health, which is why death from all infections still increases with increasing socio-economic deprivation in the twenty-first century. Antimicrobial chemotherapy is just one component of an overall strategy to prevent and treat infections.
Fig. 19.1 Infectious diseases mortality in the USA during the twentieth century. Reproduced from Armstrong GL, Conn LA, Pinner RW. Trends in infectious disease mortality in the United States during the 20th century. Journal of the American Medical Association 1999; 28: 61-66 with permission from American Medical Association.
Upper respiratory tract infection
This is one of the commonest acute problems seen in general medical practice, with an incidence of 100 cases per 1000 inhabitants per year, although only a minority of these will present to a doctor. It is commoner in females than men. Symptoms include: sore throat with anorexia, lethargy, and systemic illness. On examination there may be inflamed tonsils or pharynx, a purulent exudate on tonsils, fever, and anterior cervical lymphadenopathy.
Sore throat may be part of the early symptom complex of many upper respiratory viral infections, in which case cough is a common additional feature. Occasionally, it may be a presenting symptom of acute epiglottitis or other serious upper airway disease.
There is no evidence that bacterial sore throats are more severe or long-lasting than viral ones. The most commonly identified organism isStreptococcus pyogenes, the group A β-haemolytic streptococcus. Most other cases are caused by adenoviruses. There is no reliable way to distinguish between bacterial and viral causes based on symptoms and signs.
The gold standard for diagnosis of streptococcal infection in the throat includes a positive anti-streptolysin O (ASO) titre in addition to culture of Str. pyogenes from the throat. There is a high asymptomatic carrier rate for the organism (up to 40%) and it is common to culture it from sore throats when there is no serological evidence of infection. Moreover a negative culture does not rule out Str. pyogenesas a cause of sore throat. Neither culture of throat swabs nor rapid tests based on detection of streptococcal antigen are helpful in most cases.
Most people with sore throat manage the condition successfully without seeing a doctor. Paracetamol is an effective analgesic, with less risk of adverse effects than non-steroidal anti-inflammatory drugs. Aspirin should be avoided in children because of the risk of Reye's syndrome. The immediate benefits from antimicrobial chemotherapy are actually very meagre. Symptoms usually persist for 5-7 days with or without antibiotics, which only shorten illness by 24 h. The same control of symptoms can probably be achieved with paracetamol.
Streptococcal sore throat is important because it may lead to serious complications, particularly rheumatic fever, which is still prevalent in many countries. Evidence about the effectiveness of antibiotics for preventing nonsuppurative and suppurative complications comes from studies on military personnel living in overcrowded barracks in the late 1940s and early 1950s. This evidence has little relevance to management of sore throat in modern communities, at least in the developed world, where rheumatic fever is now very uncommon. Similarly experience with the use of antibiotics to prevent cross-infection in sore throat comes mainly from army barracks and other closed institutions. It is very unlikely (and unproven) that trying to eradicate Str. pyogenes with routine antibiotic therapy for sore throat will produce any measurable health gain in the general public in Western countries, whereas it is likely that this would increase the prevalence of antimicrobial resistance.
A patient information leaflet may be of value in the management of acute sore throat and may assist in managing future episodes at home without general practitioner involvement. Patients who are sceptical about withholding antibiotics can be given a prescription with the suggestion that they do not use it unless their symptoms persist for more than 3 days. Only about 30% of patients who are given delayed prescriptions go to the pharmacy to get their antibiotics.
If antibiotics are to be prescribed the drugs of choice are penicillin V or a macrolide, and these should be given for at least 10 days to eradicate the organism and prevent recurrence. Glandular fever commonly causes symptoms and signs that are indistinguishable from streptococcal throat infection (including a very impressive purulent exudate on the tonsils). Ampicillin, amoxicillin, and co-amoxiclav should not be used, as they will cause a rash if the sore throat is the herald of glandular fever. Tetracyclines are also inappropriate because of the high incidence of resistance among streptococci.
Other infections that may present with sore throat
Noisy difficult breathing, hoarseness, and stridor are common signs of croup, a distressing condition that is usually viral in origin. Treatment is supportive; the condition is usually self-limiting and resolves in 2-4 days if uncomplicated, but severe cases may require endotracheal intubation or tracheostomy. Acute epiglottitis is a much less common, but much more dangerous, cause of croup caused by infection withHaemophilus influenzae type b; it can occur in adults as well as in children. There is systemic illness as well as local respiratory difficulty, and the swollen oedematous epiglottis can cause complete airways obstruction with dramatic suddenness. It is this complication that makes acute epiglottitis such a life-threatening condition. Treatment is as much concerned with maintaining the airways as with controlling the infection.
If breathing difficulty is present in a patient with croup, urgent referral to hospital is mandatory and attempts to examine the throat should be avoided.
Although rare in countries with effective vaccination policies, diphtheria is still prevalent in many parts of the world. Diagnosis is made on clinical grounds, notably the presence of a characteristic membranous exudate on the tonsils and pharynx. Treatment with antitoxin should be given immediately without waiting for laboratory confirmation. Antibiotics have no part to play in treating the infection, but penicillin or erythromycin is effective in eradicating the infection to prevent spread.
Oral thrush, infection of the mucous membrane with the yeast Candida albicans, is predominantly a neonatal infection. Candida is a common vaginal commensal, especially in pregnancy, and the infant acquires infection during passage through the birth canal. It presents in the first few days of life as white curdy patches on cheeks, lips, palate, and tongue. Treatment is with local nystatin.
In adults, oral thrush may follow treatment with antibacterials or corticosteroids. However, it may be indicative of serious underlying disease, such as diabetes or immunodeficiency (Chapter 28). In all these conditions one of the oral polyene or azole derivatives may be used to control the candida.
Acute otitis media
Three-quarters of cases of acute otitis media occur in children; one in four children will have an episode during their first 10 years of life. Acute otitis media should be distinguished from otitis media with effusion, commonly referred to as glue ear; as many as 80% of children suffer this infection at least once before the age of 4.
Acute otitis media is an inflammation of the middle ear of rapid onset presenting with local symptoms (earache, rubbing or tugging of the affected ear) and systemic signs (fever, irritability, disturbed sleeping). It is often preceded by other upper respiratory symptoms such as cough or rhinorrhoea. On examination a middle ear effusion may be present but in addition the drum looks opaque and may be bulging.
The condition is caused predominantly by H. influenzae and Streptococcus pneumoniae. Staphylococci, Str. pyogenes, and α-haemolytic streptococci are less often involved. However, acute otitis media should not be treated routinely with antibiotics. As with sore throat, antibiotics only have a small impact on the duration of acute symptoms, which can be controlled equally effective with paracetamol. If an antibiotic is to be prescribed a 5-day course is sufficient; the antibiotic of choice is amoxicillin; erythromycin and coamoxiclav are logical alternatives and may be necessary if β-lactamase-producing H. influenzae is involved. Decongestants, antihistamines, and mucolytics are not effective. As with sore throat, patient information leaflets and delayed antibiotic prescriptions are effective strategies for reducing the unnecessary use of antibiotics.
Glue ear is an inflammation of the middle ear with accumulation of fluid in the middle ear but without symptoms or signs of acute inflammation. It is often asymptomatic and earache is uncommon. On examination a middle ear effusion is present but with a normal looking ear drum. Antibiotics should not be given.
Acute sinusitis presents with pain originating in the maxillary, frontal, ethmoid, or sphenoid sinuses, with the maxillary sinus being by far the commonest. Onset of facial pain is often preceded by non-specific symptoms of upper respiratory inflammation and there may be systemic signs of inflammation. The bacterial causes of acute sinusitis are the same as acute otitis media. If X-ray or culture confirms the clinical diagnosis then antibiotics can substantially reduce the duration of symptoms. However, neither of these investigations is routinely available in primary care.
Culture of the sinuses requires percutaneous sinus puncture and aspiration, which is not a procedure that most general practitioners are trained to do (or that many patients would consent to). Unfortunately antibiotic treatment of patients with symptoms suggestive of sinusitis but without confirmation by X-ray or culture is no more effective than symptomatic relief.
As with acute otitis media antibiotics for acute sinusitis should be reserved for the more severe cases. Penicillin V or amoxicillin are as effective as newer antibiotics. The recommended duration of treatment is 10 days in the absence of evidence that shorter courses are as effective.
Lower respiratory tract infections
Acute cough is the most common symptom of lower respiratory infection, whether as a new symptom or as an exacerbation of chronic symptoms. Cough is not a universal feature: some patients with pneumonia present with pleuritic chest pain or with symptoms of systemic inflammatory response (fever, malaise, headache, or myalgia) without cough. The most important diagnosis to make is pneumonia because it can be life threatening and its outcome can be improved with antimicrobial chemotherapy. However, it is not possible to distinguish reliably between pneumonia and other causes of lower respiratory tract infection from clinical history and signs. Consequently, in primary care management must be based on an assessment of severity of illness and need for referral to hospital.
The incidence of lower respiratory tract infection in the UK is between 40 and 90 cases per 1000 population per year, being commoner in the very young and old and in the winter months. In the UK there is about a four-fold higher incidence in the most deprived communities in comparison with the most affluent communities.
Mortality is highest in the elderly. The 30-day mortality associated with lower respiratory tract infection in people over 65 years old is 10%. However, many of these elderly people die ‘with’ rather than ‘of’ the infection. Bronchopneumonia is often recorded as the immediate cause of death in people with chronic, life-threatening diseases. Mortality from ‘pneumonia’ has actually increased in developed countries since the introduction of antibiotics, but more people are living longer and most of this mortality is from bronchopneumonia.
Most people with lower respiratory infections manage their own symptoms without seeking medical attention. Of 1 million people with lower respiratory tract infection only 300 000 will see a primary care physician. Of these 1 in 4 (70 000) will be treated with antibiotics, although only about 1 in 10 (7000 people) will have a diagnosis of pneumonia. From the original 300 000 people who presented to a primary care physician only about 200 (0.7%) will be admitted to hospital with pneumonia.
Management in primary care
The key to management of lower respiratory tract infection in primary care is to distinguish between patients who have severe infection that should be referred to hospital and the majority (99%) who can be managed safely at home. There are four questions to address:
The answers to these questions distinguish between four broad populations of people with lower respiratory tract infection. These will be discussed starting with the most severe (but least common).
Box 19.1 Features of severity of lower respiratory tract infection that can be easily assessed in primary care (items in bold are most important)
Patients with features indicating severe infection
Referral to hospital should be considered in patients who exhibit one or more of the features of severity (Box 19.1), especially if they are over the age of 50. This applies whether or not the patient has additional physical signs indicating pneumonia because the absence of these signs is not a reliable method for excluding pneumonia. The final decision should be based on clinical judgement that includes social factors. Even a relatively well patient who lives in poor social circumstances or in an isolated rural area with no home support may require referral to hospital. Conversely patients who are 65 years old and have signs of pneumonia can be managed safely at home if they have sufficient social support.
Suspected community-acquired pneumonia without features of severity
These patients have new focal signs in the chest (crackles or altered breath sounds), but are not severely ill. In the absence of chest X-ray (not available in many primary care settings) pneumonia can be diagnosed from symptoms of an acute lower respiratory infection (cough or dyspnoea or pleuritic chest pain) with at least one systemic symptom of infection (fever or tachycardia) and new focal signs on chest examination. However, only 50% of those with all of these features will actually have an abnormal chest X-ray.
Below the age of 45 very few patients with pneumonia also have chronic obstructive pulmonary disease. Between the ages of 45 and 64 the proportion is up to 10% and rises to 20% between the ages of 75 and 84. Pneumonia in these patients is more likely to be associated with severity criteria (Box 19.1).
A wide variety of organisms can cause pneumonia, including viruses. The commonest bacterial cause is Str. pneumoniae, which accounts for about 70-80% of cases in which a bacterial pathogen is identified. Atypical bacteria (Mycoplasma pneumoniae, Chlamydophila (Chlamydia)pneumoniae, Chlamydophila psittaci, Legionella pneumophila, and Coxiella burnetii) collectively account for 10-20% of cases and the remainder are caused by H. influenzae or Staphylococcus aureus. The latter is particularly associated with secondary bacterial infection following influenza.
With current technology neither sputum culture nor blood tests such as C-reactive protein or white cell count provide sufficient added value to the diagnosis to justify routine use. Sputum culture may be recommended in areas with a high prevalence of penicillin-resistant pneumococci.
Pneumonia is a life-threatening illness. None the less, patients with no features of severity (Box 19.1) can be managed safely at home with oral amoxicillin, a macrolide, or a tetracycline. There is no need to give combination therapy. A macrolide or tetracycline may be preferred if there are clinical features suggesting infection with one of the atypical bacteria (e.g. prominent upper respiratory symptoms, headache, or symptom duration for >1 week) particularly in younger patients or during an epidemic year for M. pneumoniae.
Patients with underlying chronic respiratory disease
These patients often have no new signs in the chest other than dyspnoea and sputum purulence. In the absence of signs of severity (Box 19.1) or of pneumonia the diagnosis is an acute exacerbation of the underlying condition. The likely bacterial pathogens are H. influenzae, Str. pneumoniae, and Moraxella catarrhalis. The development of green (purulent) sputum is a good indicator of a high bacterial load in the sputum. However, even in these patients antibacterial treatment has only a slight impact on the course of an acute exacerbation, shortening an illness of 5-7 days by no more than 1 day. Antibacterial treatment does not benefit patients with acute exacerbations of chronic obstructive pulmonary disease who do not have purulent sputum. The prevalence of resistance to aminopenicillins in H. influenzae is 10-30% and is much higher in Mor. catarrhalis. Despite this the clinical effectiveness of amoxicillin is just as good as co-amoxiclav or fluoroquinolones, probably because of the modest benefit from any antibacterial treatment. For the same reason routine sputum culture is not recommended and should be reserved for patients with symptoms that persist despite treatment with amoxicillin. A macrolide or tetracycline is appropriate for patients who are allergic to penicillin, or who have not responded to amoxicillin treatment. Fluoroquinolones should not be used empirically in the management of exacerbations of respiratory disease in primary care.
Non-pneumonic lower respiratory infection (acute bronchitis)
Most patients with no signs in the chest, who have been previously well and do not have other features of severity, have non-pneumonic infection, most of which are caused by viruses. A few cases are caused by M. pneumoniae, Bordetella pertussis, C. pneumoniae, Str. pneumoniae, or H. influenzae. Patients will have an illness lasting several days with or without antibiotics, which should not be prescribed unless patients have signs in the chest or features of severity (Box 19.1). Sputum purulence alone is not an indication for antibiotics in a previously well patient with no chest signs. As with sore throat and acute otitis media, patient information leaflets and delayed prescriptions are effective strategies for reducing unnecessary antibiotic treatment.
Pertussis (whooping cough)
Antibiotics are notoriously ineffective in controlling the distressing cough of pertussis; nevertheless, erythromycin has been shown to eradicate the organism from the respiratory tract and can also be used for the protection of susceptible close contacts. Vaccination is the only reliable way of preventing and controlling this early childhood infectious disease.
The susceptibility of patients with cystic fibrosis to pulmonary infection is well recognized and is often the cause of early death. Most lung infections in patients with cystic fibrosis are managed in the community, usually by outreach teams from secondary care. One of the striking features of chest infections in cystic fibrosis is that relatively few pathogens are involved. Early in the disease the organisms implicated are frequently Staph. aureus or H. influenzae, or both. As patients progress through adolescence to adulthood, these pathogens are replaced byPseudomonas aeruginosa. Major problems arise when Ps. aeruginosa is replaced by Stenotrophomonas maltophilia or Burkholderia cepacia; these organisms are often resistant to many antibiotics and treatment should be guided by laboratory findings. The selection of antibiotics in patients with cystic fibrosis should be determined by the specialist services that manage the patient.
Management in hospital
In hospital the clinical diagnosis can be confirmed with a chest X-ray, although it should be recognized that its sensitivity is not 100%. The gold standard for diagnosis of bacterial pneumonia is culture of bacteria from lung tissues or a needle aspirate from the lung but these tests are too dangerous to use in routine clinical practice. The point is that some patients with pneumonia can have a normal chest X-ray at presentation, so if the clinical features strongly suggest pneumonia it is reasonable to treat and repeat the chest X-ray after 24-48 h.
The severity criteria for community-acquired pneumonia are based on assessments of confusion, urea concentration, respiratory rate, and blood pressure for those 65 years of age and older (CURB-65 score; Table 19.1). It is similar to but importantly different from the classification of severity of sepsis (Chapter 13, p. 185). The CURB-65 score is specifically designed to be used in patients who present to hospital in order to identify low-risk patients who do not need to be admitted to hospital, whereas the classification of sepsis is intended to be used for any patient with infection (community or hospital acquired) to identify patients who are deteriorating and require more intensive therapy. The CURB-65 score identifies low-risk patients more accurately than the sepsis severity score. There are more complex pneumonia-specific scores (for example the pneumonia severity index used in North America) but these are no more accurate than CURB-65.
Table 19.1 The CURB-65 severity score for patients presenting to hospital with community acquired pneumonia and the mortality range measured in two prospective cohort studies. CURB-65: score one point for each of: Confusion; urea >7mmol/l; respiratory rate ≥30/min; low systolic (<90mm Hg) or diastolic (≤60mm Hg) blood pressure); age ≥65 years
Between 30 and 50% of patients who present to hospital with community-acquired pneumonia are found to be in the CURB-65 low-risk group. However, about half of these patients have other reasons for admission. Some will have co-morbidities that require inpatient management. In particular, patients with chronic obstructive pulmonary disease and pneumonia could be in respiratory failure and yet have a CURB-65 score of 0 (if they have a respiratory rate <30/min, which is likely if they have Type 2 respiratory failure). In addition to medical reasons for admission some patients will have poor social circumstances or insufficient support to be managed at home.
The management of patients admitted to hospital should be determined by their CURB-65 score. Low-risk patients who are admitted for other reasons can be managed in the same way as low-risk patients in the community, with either amoxicillin, a macrolide, or a tetracycline. Some guidelines do recommend that all patients admitted to hospital with pneumonia should receive antibiotics for pneumonia caused by atypical bacteria but several clinical trials shows that treatment with an aminopenicillin alone is just as effective for patients with low or intermediate risk pneumonia.
At the other end of the scale, patients at high risk should be treated with intravenous antibiotics that are effective against the full range of pathogens that may cause community-acquired pneumonia. Possible regimens include co-amoxiclav or cefuroxime plus a macrolide, or a fluoroquinolone with good activity against Str. pneumoniae (e.g. levofloxacin). The patient must receive the antibiotic(s) immediately and certainly within 4h of admission as later administration is associated with increased mortality. If patients are admitted through an Accident and Emergency Department they must receive their first dose of antibiotics there before transfer to the ward. If they are admitted direct to a ward the first dose must be clearly written for immediate administration, not left until the next drug round. In addition to intravenous antibiotics patients with severe pneumonia must have their oxygen requirements assessed by pulse oximetry or blood gas measurement within 4h of admission and receive high flow oxygen (5 litres per minute) if they are hypoxic. Adequate fluid replacement is also essential. Patients should be referred to a high dependency or intensive care unit if their vital signs do not improve rapidly. When young patients die from community-acquired pneumonia it is usually because of failure to recognize the need for intensive care.
The management of patients at intermediate risk falls between these two extremes and is a matter for clinical judgement. If in doubt it would be wise to treat as severe pneumonia while waiting for senior review.
Pneumonia is the leading cause of mortality resulting from infection acquired in hospital. The incidence of hospital-acquired pneumonia in intensive care units ranges from 10 to 65%, with case fatalities of 13-55%. It is often associated with mechanical ventilation. The risk of hospital-acquired pneumonia can be substantially reduced by using non-invasive methods for respiratory support instead of ventilation and by having clear care protocols for protecting host defences against respiratory infection during mechanical ventilation. Chemoprophylaxis plays a role through the use of selective decontamination of the digestive tract (see p. 241), which reduces the numbers of Gram-negative bacilli and hence the risk of infection.
The micro-organisms causing pneumonia within 5 days of admission are quite different from those seen in disease with a later onset. The bacteria responsible for early onset pneumonia are Str. pneumoniae, H. influenzae, Staph. aureus, and only rarely enteric Gram-negative bacilli. In contrast late onset infection is almost always caused by Gram-negative bacteria, mainly enterobacteria but also Ps. aeruginosaand Acinetobacter spp.
Methicillin-resistant Staph. aureus (MRSA) is becoming increasingly common in some units. Since tracheal aspirates are poor indicators of the cause of ventilator-associated pneumonia, bronchoalveolar lavage is recommended to confirm the diagnosis.
Empirical treatment for early onset pneumonia in patients who have not received antibiotics should be with co-amoxiclav or cefuroxime. Treatment of patients who have already received antibiotics or have late onset disease should be with a broad-spectrum cephalosporin such as cefotaxime, a fluoroquinolone or piperacillin plus tazobactam. Combination therapy is no more effective than monotherapy. Subsequent treatment should be directed by the results of broncho-alveolar lavage.
Other respiratory tract infections
Pneumonia developing in association with neutropenia following treatment with cytotoxic drugs, or in patients with immunosuppression, including those suffering from AIDS, may be due to Pneumocystis carinii, other fungi, or viruses. Appropriate treatment is discussed inChapters 27 and 28. The treatment of tuberculosis is considered in Chapter 25; influenza and other respiratory viral infections are dealt with in Chapter 27.
British Society for Antimicrobial Chemotherapy guidelines for hospital acquired pneumonia, Journal of Antimicrobial Chemotherapy in press 2006.
British Thoracic Society. British Thoracic Society guidelines for the management of community acquired pneumonia in childhood. Thorax2002; 57 Suppl 1: i1-i24. Available at: http://thorax.bmjjournals.com/cgi/content/full/57/90001/i1
British Thoracic Society. British Thoracic Society guidelines for the management of community acquired pneumonia in adults. Thorax2001; 56 Suppl 4: iv1-iv64. Available at: http://thorax.bmjjournals.com/cgi/content/full/56/suppl_4/iv1
British Thoracic Society. British Thoracic Society guidelines for the management of community acquired pneumonia in adults-2004 Update. Available at: http://www.brit-thoracic.org.uk/c2/uploads/MACAPrevisedApr04.pdf
Metlay JP, Fine MJ. Testing strategies in the initial management of patients with community-acquired pneumonia. Annals of Internal Medicine 2003; 138: 109-118.
Scottish Intercollegiate Guidelines Network (www.sign.ac.uk) Guideline 34: management of sore throat and indications for tonsillectomy Guideline 66: diagnosis and management of childhood otitis media in primary care Guideline 59: community management of lower respiratory tract infection in adults