CA125: Holy Grail or a Poisoned Chalice

The history of (and perhaps the lessons to be learned from) peritoneal CA125 investigation presents no better example of the diffi culties in assigning the status of a ‘marker’ predictive of membrane dysfunction to factors measured in dialysis effl uent. In 1995, Visser et al. [3] were the fi rst to suggest that CA125 could be used as ‘a bulk marker of mesothelial cell mass in stable PD patients’. This was suggested based largely on in vitro and ex vivo approaches that suggested a mesothelial-cell-specifi c location (i.e. it is not expressed by leucocytes). Their data also suggested passive secretion of the antigen and a relationship to the number of mesothelial cells in culture. These data on the ‘regulation’ of secretion, however, were at odds with both pre-existing and subsequently produced data suggestive of induction of CA125 by pro-infl ammatory cytokines [4, 5] . In the current issue, Br e borowicz et al. fan the fl ames of this debate further by suggesting, based exclusively on in vitro data, that CA125 production by peritoneal mesothelial cells neither correlates with, nor is regulated by pro-infl ammatory cytokines [6] . The evidence ascribing CA125 as a clinical marker of mesothelial cell or membrane integrity was apparently strengthened by the studies of Ho-dac Panekeet et al. [7] who presented cross-sectional data suggesting a relationIdentifying easily measurable markers that are of prognostic value in predicting membrane failure in peritoneal dialysis (PD) patients remains one of the major challenges. Facilitating this endeavour has always been the ease of sampling of drained peritoneal effl uent within which, with the development of ever more sensitive measurement techniques, any number of mediators can be quantitated. Local production of these ‘factors’ (i.e. of peritoneal membrane or infi ltrating leucocyte origin) would then theoretically, at least, allow changes to be correlated with clinical outcomes and thereby their usefulness as indicators of altered membrane status. So far then there is an attractiveness and inescapable logic to this type of approach that would predict that after almost 20 years of investigation (since the fi rst measurements of prostaglandins in peritoneal effl uent [1, 2] ) there would now be suffi cient data on a number of markers predictive of clinical and membrane-associated changes in these patients. Unfortunately, however, despite signifi cant scientifi c endeavour (and as we will discuss a number of long-term clinical studies) we are really little nearer to identifying our holy grail. Published online: April 7, 2005