126 Chapter 6 Discussion This study showed that there is poor agreement between body composition measured with ADP and body composition measured with DXA in preterm born infants at term equivalent age. Compared to ADP, DXA showed higher fat mass percentages. Furthermore, estimations of fat mass based on the ∑SFT showed poor agreement with the actual fat mass measured with DXA. Various studies in term infants report high correlations between fat mass measured with ADP and fat mass measured with DXA (9, 10). Nevertheless, a high correlation does not imply both methods found the same value and doesn’t provide information about the test quality (24). Similar to studies performed in full term infants, DXA gave higher estimates of fat mass in our cohort compared to estimates by ADP (9, 10). In agreement with that, early animal studies showed that DXA seems to overestimate fat mass (25, 26). To our knowledge, no data has been published on the comparison of DXA and ADP in extremely preterm infants. Nevertheless, one recent study in South-African term infants also showed higher estimates of fat mass by ADP compared to DXA (8). Moreover, several reviews have highlighted that both DXA and ADP have reasonable reproducibility, but only modest accuracy. According to these reviews ADP actually seems to underestimate fat-free mass percentage or fat-free mass expressed in grams per liter (fat-free mass density). Especially, when the fat-free mass percentage or density gets higher, the underestimation becomes larger (15, 27). In actual fact the fat-free mass percentage or density may be a more relevant parameter to assess, as in practice it may be more insightful to properly predict fat mass and fat-free mass percentage than it is to predict absolute fat and fat-free mass. Nonetheless, in all these studies it is to be questioned whether an appropriate reference method for body composition has been used. In practice, both DXA and ADP, as well as deuterated water, have been deemed as reliable methods, however, there seems to be no universally accepted preferential reference method in living infants. In contrast to DXA, ADP takes into account that hydration status is different in infants as compared to adults. Particularly during the first week of life, infants’ fatfree mass hydration is higher, and therefore DXA estimations of fat and fat-free mass may not be as accurate as ADP estimations in this period (28). Moreover, the algorithms used in DXA software are not open for critical analysis. In addition, DXA quality is negatively influenced by movement, while moderate movement does not affect body composition measurements taken using ADP. Moreover, infants are exposed to a low dose of radiation. Therefore, in this study, and presumably in others as well, ADP measurements may have been more reliable. In line with others, we could not externally validate the model by Schmelze et al. for fat mass prediction (29). To date, predictive models for the estimation of fat mass using the SFT, have only been validated in term and late preterm infants and the
RkJQdWJsaXNoZXIy MTk4NDMw