In the absence of a chest x-ray, the British Thoracic Society defines pneumonia as symptoms of an acute lower respiratory tract infection, including a cough and at least one other lower respiratory tract symptom, together with at least one systemic symptom, and new focal signs on chest examination.⁵² In most reported series of patients with community acquired pneumonia, no pathogen is identified in 50% or more of cases. The role of viruses will become clear with increased use of modern molecular diagnostic techniques (see Annex 2 to this guideline on the SIGN website). A wide array of organisms may cause acute pneumonia and published reports vary in the organisms isolated due to differences in patient groups, presence of epidemic organisms and diligence of the investigation:^53,54

• Streptococcus pneumoniae is the most frequently identified pathogen in community GP samples. Other organisms commonly reported include Mycoplasma pneumoniae, Staphylococcus aureus, Haemophilus influenzae and influenza viruses. A raised incidence of Staphylococcus aureus is found during influenza epidemics.¹⁴
• Organisms such as Mycoplasma pneumoniae show seasonal variation with peak incidence at four yearly intervals.
• Legionella spp. were initially reported in large scale outbreaks but this organism is now known to occur sporadically particularly in patients with comorbidity such as congestive cardiac
  failure, diabetes and chronic obstructive pulmonary disease.^54,55 Approximately three-quarters of Legionnaires' disease cases have a history of recent travel abroad.⁵⁶
• Chlamydia pneumoniae has recently been identified as a human pathogen, the extent to which this organism causes respiratory disease is as yet undetermined. Infection by Chlamydia psittaci should be considered when there is exposure to birds.
• Patients with a history of recent foreign travel may have pneumonia caused by a wide variety of organisms rarely found in Scotland. Additionally, common pathogens such as Streptococcus pneumoniae which are acquired abroad may exhibit more resistance to common antibiotics than is currently seen in Scotland.^57,58,59,60
• Anaerobic infection due to aspiration into the lower respiratory tract is found in patients with an alcohol problem, and in other conditions predisposing to aspiration into the respiratory tract.

5.1 Investigation

5.1.1 SPUTUM CULTURE

If sputum is available and the patient has not had prior antibiotic treatment then a Gram stain is a good indicator of the causative organisms.^54,61 Overnight culture will provide confirmation and the chance to perform susceptibility studies, allowing modification of empirical therapy. Culture is also helpful in establishing the pathogenicity of any isolates.^{62,63,64,65,66,67} Evidence level 2+

5.1.2 BLOOD TESTS FOR C-REACTIVE PROTEIN (CRP)

In a study of adults with respiratory tract infection in general practice,⁶⁸ CRP was the best test discriminating between pneumonia and non-pneumonic LRTI. In a further larger study of 402 adults, this finding was confirmed and it was found that, in the first week of the illness, viral LRTI could also produce high CRP values; with the likelihood ratio for pneumonia in the presence of a high CRP increasing after the first week of illness.⁶⁹ In this study, although those with radiological evidence of pneumonia had a higher mean white blood cell count (WBC), a WBC of >=10.4 was not helpful in predicting radiologically defined pneumonia unless symptoms had been present for seven days or more. Another study showed that in a hospital population with community acquired pneumonia, failure of the CRP to fall was a useful indicator of treatment failure.⁷⁰ Evidence level 2+,3

An assessment of whether the availability of CRP tests to GPs can reduce antibiotic prescribing for respiratory tract infections in the community looked at prescribing for both upper and lower respiratory tract infections in two groups of patients: those in whom the CRP value was available and those in whom the decision was based on clinical grounds alone.⁷¹ This RCT found no difference in prescribing between the two groups. In previously well patients, presenting with LRTI, a CRP level >50 mg/dl was seen more frequently in patients with indirect indications and microbiological evidence of infection but the sensitivity and specificity of the test were insufficient for it to be of value for routine management in primary care.⁷² Evidence level 1+,3

CRP levels are of limited use as a diagnostic tool for community acquired pneumonia and should not be performed routinely.

5.1.3 PULMONARY FUNCTION TESTS

One study has followed up a group of 95 patients presenting to their GP with an episode of cough associated with diffuse wheeze or crackles.⁷³ Three years after their initial presentation, the patients completed a questionnaire and performed spirometry and methacholine challenge testing. A total of 34% of this group had findings consistent with a diagnosis of asthma or COPD. Thus a presentation with a cough associated with diffuse wheeze or crackles may raise the suspicion of an underlying airway problem. Evidence level 3

Consider spirometry in the convalescent period to diagnose asthma or COPD in patients with community acquired pneumonia presenting with a cough associated with diffuse wheeze or crackles.

5.1.4 CHEST X-RAY

Chest x-ray evidence of pneumonia is reported in around 40% of patients thought by their GPs to have an acute lower respiratory tract infection associated with new focal chest signs.^14,19 The absence of any signs of abnormality (i.e. pulse, respiratory rate, temperature, and chest examination) makes the diagnosis of radiologically-defined pneumonia unlikely. Evidence level 2+,3

In a study of 402 consecutive adults presenting to general practice in Sweden with symptoms of respiratory tract infection, 5% were shown to have pneumonia on chest x-ray.¹⁵ However, in this study, lung crackles and other abnormal chest findings were interpreted too frequently as features of pneumonia. Similarly in a study of 153 adult patients with LRTI, only one of nine with pneumococcal pneumonia, and two of seven with mycoplasma infection, had radiographic evidence of pneumonia.⁷⁴ Evidence level 2+,3

There has also been debate regarding the value of follow-up chest x-rays in those found to have pneumonia. A retrospective review of case notes of 1,011 patients admitted to hospital with pneumonia found 13 patients with bronchial carcinoma.⁷⁵ In eight cases this diagnosis was apparent on the initial chest x-ray. Bronchial carcinoma was thus found on convalescent chest x-ray in just 0.58% of patients. The authors therefore recommended a clinical review one-two months after diagnosis, and x-raying only those with ongoing symptoms. In a separate prospective study, a convalescent chest x-ray was recommended in those patients who make a good recovery because they found that six out of 36 smokers over the age of 60 with pneumonia, had an underlying bronchial carcinoma.¹⁴ Evidence level 2+,3

Chest x-ray should not be used routinely for patients with acute symptoms of community acquired pneumonia.

Consider chest x-ray in the convalescent period in community acquired pneumonia patients who smoke, or if patients do not make satisfactory progress.

5.2 Treatment

Although there is no direct evidence due to trials not having been conducted and due to the fact that it is no longer ethical to conduct such trials, the longstanding consensus is that antibiotic treatment is essential for pneumonia.⁵² Evidence level 4

Early administration of antibiotics in patients with pneumonia is essential.

The antibiotic chosen should be effective against Streptococcus pneumoniae. Treatment with an aminopenicillin or a macrolide is appropriate.

In younger patients (aged <50 years) Mycoplasma pneumoniae should be considered, particularly if it is an epidemic year and any of the following clinical features are present:

• upper respiratory tract symptoms
• headache
• symptom duration >1 week.

In these cases, and in those with a diagnosis of chlamydial pneumonia, treatment with a macrolide or tetracycline is appropriate since aminopenicillins are ineffective.

For patients with indices of severity who might normally be referred to hospital, but for various reasons are managed in the community, aminopenicillin and macrolide combination treatment and close follow-up is recommended.

Patients with features of pneumonia should be reviewed after 48 hours, or earlier if clinically indicated, when severity should be reassessed.