Microalbuminuria in severe sepsis: timing is difficult, "normal" levels are uncertain

Excerpt

Sir: In their letter to the editor Gosling and colleagues [1] raise two questions regarding the interpretation of microalbuminuria in the critically ill. On one hand, they emphasize the importance of timing; on the other, they question the reference range used in our study [2]. Regarding their comments on timing, we fully agree that increased microalbuminuria occurs early after the initiating stimuli. However, based on previous and the current results we think that microalbuminuria is rather an acute marker of an inflammatory insult than the marker of generalized capillary leakage, but there is no firm evidence to support either theory [2, 3, 4]. In intensive care patients the time of onset of the acute insult is often unknown, and therefore we need reliable markers to reflect the actual situation at any given time. Gosling and colleagues assumed that we measured microalbuminuria far too late, and this is why increased extravascular lung water (EVLW) was not accompanied by an increased urinary microalbumin creatinine ratio (M:Cr). To date no studies have investigated capillary leakage in this context. In one of our current trial in patients following major abdominal surgery we found normal EVLW while there was a considerable increase in microalbuminuria as reported several times previously (unpublished data). In our recent study EVLW measurements were started within 24 h, when the need for invasive hemodynamic monitoring was indicated [2]. As our patients were severely ill, with at least two organ system failures, high EVLW and peripheral edema, one would assume that this is the result of systemic capillary permeability. However, we detected no significant changes in the urinary albumin excretion together with constantly increased EVLW. Therefore we feel that the M:Cr ratio cannot be relied on as a bedside test for ongoing capillary leakage in severe sepsis, but it is still a very useful marker of acute insults with some predictive value regarding outcome [3]. …