Eurosurveillance - View Article

Differentiating inhalational anthrax from other influenza-like illnesses and acute respiratory infections in the setting of a deliberate release of Bacillus anthracis

Amanda Walsh (amanda.walsh@hpa.org.uk) and Dilys Morgan, Health Protection Agency Communicable Disease Surveillance Centre, London, UK

Many acute respiratory illnesses are viral in aetiology, and the infections are broadly grouped together as influenza-like illness (ILI) because of the similarity in presenting symptoms. Community acquired pneumonia (CAP) is often bacterial and frequently requires hospitalisation. Depending on the disease stage at presentation, inhalational anthrax (IA) may have features of either ILI or CAP. Recognition of inhalational anthrax, particularly amongst seasonal acute respiratory illnesses, is likely to be difficult if a deliberate release has not been recognised.

The classic clinical picture of IA is biphasic but non-specific, with an initial phase that is influenza-like with fever, malaise, myalgia and cough. The second phase begins abruptly with development of an acute dyspnoea and progresses rapidly to septic shock. The non-specificity of the early symptoms may make it unrealistic to diagnose in the absence of a known release, but prompt recognition is critical for appropriate management.

Two recent papers have attempted to define those clinical features that may assist diagnosis, and to develop a diagnostic algorithm [1,2]. Following recognition of a release or of other cases, a risk assessment is crucial with patients being assigned to a high risk or no risk group based on their exposure to anthrax spores. Not all potential cases will be encountered in such a setting, however, and certain groups, especially the elderly, are particularly vulnerable, since infection may result from exposure to a very low dose of spores [3]. Sporadic cases will require differentiation from other more common causes of acute respiratory illness, including CAP. Table 1 shows those signs and symptoms likely to be most useful for distinguishing clinical syndromes.

Table 1. Suggestive signs and symptoms for inhalation anthrax. (Adapted from references 1 and 2.)

Symptoms/test results favouring possible inhalation anthrax
	Widened mediastinum on CXR or CT scan
	Pleural effusions
	Haemoconcentration >50%
	Shortness of breath
	Chest discomfort or pleuritic chest pain
	Drenching sweats
	Tachycardia
	Nausea or vomiting
	Oxygen saturation <94% by pulse oximeter
	PaO2 <70mmHg on room air
	Elevated transaminases

Symptoms/test results favouring other influenza-like illness
	Coryza/rhinorrhoea,
	Sore throat
	Positive rapid tests for influenza or RSV

The following scoring systems have also been used to aid the differentiation of clinical presentations.

Table 2. Scoring system to distinguish between IA and ILI. (Adapted from reference 2.)

	Score
Low serum albumin	+2
Tachycardia	+2
No nasal symptoms	+2
No myalgia or arthralgia	+1
Low serum sodium	+1
No headache	+1
High haematocrit or haemoglobin	+1

A score of ≥ 4 was 100% sensitive and 96% specific for IA

Table 3. Scoring system to differentiate IA from CAP. (Adapted from reference 2.)

	Score
Elevated serum ALT and AST levels	+1
Low serum sodium level	+1
Normal WBC count	+1
Nausea or vomiting	+1
Tachycardia	+1

However this only poorly differentiates IA from CAP: Score ≥ 2, sensitivity100% but 48% specificity; Score ≥ 3, sensitivity 82%, specificity 81%.

The figure shows diagnostic investigations and outcomes for patients presenting with appropriate clinical signs/symptoms during an anthrax outbreak

Figure. Workup for patients with an ILI during an anthrax outbreak. (Reproduced with permission from reference 1.)

Key for algorithm:
ABGT - arterial blood gases CBCD – complete blood count and differential CR - chest radiograph CSF – cerebrospinal fluid CT - computed tomographic	GS - gram stain IA – inhalational anthrax IV – intravenous LFTs – liver function tests RSV respiratory syncytial virus

Conclusions
Following the recognition of a deliberate release of Bacillus anthracis, discrimination of IA from ILI could be readily achieved with high sensitivity and specificity. However differentiation from CAP, particularly that due to Streptococcus pneumoniae, is more problematic due to the fact that symptoms of each are generally indicative of sepsis. Recognition of the first cases of IA in a deliberate release is likely to remain problematic. An assessment of the known exposure or possible epidemiological risk may assist, but hospital admission and full diagnostic imaging, biochemical and microbiological investigations are required for definitive diagnosis.