Abstract
With great interest we read the report by Berden et al that tackles the question of necrotizing lung infections in patients requiring ECMO 1 and would like to provide the following comments:
Firstly, we are surprised by the high number of necrotizing lung infections of 30% in their cohort of patients on VV-ECMO (38 of 125 patients) and thus would like to comment on their definition of necrotizing lung infection. While diagnosis of cavitation and abscess is usually straightforward, necrotizing pneumonia is less well defined. The authors use the presence of low post-contrast attenuation, or cavitation within consolidated areas, as radiologic criteria for necrotizing pneumonia and necrotizing cavitation, respectively. Based on their definition, 36 of 38 patients were diagnosed with necrotizing pneumonia. 1 The authors assessed the ratio of the extent of necrotic tissue in relation to the volume of non-affected lung parenchyma. However, in they measure extent of consolidation to normal lung parenchyma which would not reflect degree of necrosis. Further, a focal abscess in a consolidated area does not necessarily qualify as necrotizing pneumonia. Other studies define necrotizing pneumonia as an area of non-contrast enhancement within a consolidated lobe of >50% which seems to be a reasonable cutoff 2 and may lead to a lower number if applied to this cohort.
The second but related issue is how a focal necrotic lung process could lead to recalcitrant respiratory failure. Majority of patients with lobar pneumonia do not develop severe hypoxia due to hypoxic vasoconstriction directing blood flow to ventilated areas. Positive pressure ventilation may compromise this redistribution by re-direction of blood flow to poorly compliant areas of the lungs. If necrotising pneumonia is defined as large areas of non-perfused lung tissue, lack of blood flow to affected areas should lead to less hypoxia, unless collateral damage affects wider areas of the lungs, such as endobronchial spread of material, extensive hemorrhage, or development of ARDS. As this paper shows, patients requiring ECMO spanned the entire spectrum of necrotizing infections, from single lung abscess to diffuse necrotizing pneumonia with extensive cavitation. In a recent series of 50 patients with necrotizing pneumonia, only 50% of patients with extensive cavitation required mechanical ventilation. 2 So, what in fact is it that makes these patients sick? Is it the extent of parenchymal destruction and cavitation, or is it ensuing acute lung injury and extent of parenchymal infiltrates?
It is noteworthy that several patients developed necrotizing infections during their ECMO run, and it is fair to ask whether systemic inflammation and impairment in pulmonary blood flow may have contributed to development of necrosis. On the other hand, the use of ECMO may have facilitated more aggressive endobronchial toileting to gain control of the infection. Mortality was 100% for all five patients who underwent anatomic lung resection. Presumably, the goal of surgery in such cases is source control, based on the assumption that antibiotics do not penetrate necrotic tissue, and that cavitation represents a bacterial reservoir responsible for continued soiling of the airway. To our knowledge these claims have been neither explicitly stated nor supported by any substantive theoretical or empirical argument.2–4 In addition, surgical resection may not be the only means of achieving source control. 5 It remains incumbent upon those who advocate resection to clearly define and support surgical indications, if any such exist at all.
