• Community-acquired pneumonia
• severe pneumonia
• complicated pneumonia
• childhood
• paediatric lung disease
• respiratory infection
• tuberculosis
• viral infection

In their letter, Thomas and Spencer1 claim that our assertion that Staphylococcus aureus is the most important cause of necrotising pneumonia2 is wrong on the basis of the most recent evidence and that Streptococcus pneumoniae is currently the major cause.1 However, although the studies cited by them document an increase in the incidence of necrotising pneumococcal pneumonia,1 we still believe that S aureus is the most important pathogen clinically and therapeutically.

It has recently been found that the incidence of complicated pneumonia in children is increasing and that there is a concurrent increase in the incidence of community-associated methicillin-resistant S aureus (CA-MRSA) infections.3–5 The emergence of CA-MRSA was initially reported in the USA and in subjects with skin infections, but paediatric necrotising pneumonia due to S aureus has been reported in healthy subjects of different ages and patients with underlying diseases such as cystic fibrosis.4 ,5 A high level of suspicion is required in patients with severe community-acquired pneumonia, and CA-MRSA needs to be considered early in its differential diagnosis because it can lead to rapid deterioration and death unless it is appropriately and immediately treated.

Moreover, the specific antimicrobial treatment of CA-MRSA is different from that using the traditional antimicrobial agents currently prescribed for community-acquired pneumonia. Antimicrobial agents that specifically inhibit exotoxin production, such as clindamycin and linezolid, should be preferred.6

Finally, no prophylaxis against CA-MRSA is yet available. On the contrary, necrotising pneumonia due to S pneumoniae is mainly due to serotype 3 and, in a minority of cases, serotypes 1 and 19A.2 All of these are included in the 13-valent pneumococcal conjugate vaccine that has been recently licensed and included in the immunisation schedules approved for children in most industrialised countries.7 This means that the incidence of infections due to S pneumoniae may be significantly reduced in the future, whereas those associated with CA-MRSA can only remain stable or proportionally increase.