Comment on ‘Allograft function and muscle mass evolution after kidney transplantation’ by Gaillard et al.

Abstract

We read with interest the article by Gaillard et al., entitled ‘Allograft function and muscle mass evolution after kidney transplantation’.¹ This article highlights the association between measured glomerular filtration rate (mGFR) and urinary creatinine excretion rate (CER) at 3 months and 1 year after kidney transplantation (KT) and the association between CER and mortality in kidney transplant recipients (KTr). We previously investigated, in a population of 200 KTr, the prevalence and consequences of disturbances of muscle parameters assessed on an unenhanced cross-sectional CT-scan taken at the level of the third lumbar vertebra.² We had determined age-specific and sex-specific normality thresholds on 130 healthy subjects. Twenty-five per cent of KTr had a muscle density below the 2.5th percentile of the reference population. Myosteatosis thus defined was independently associated with mortality. Conversely, in our study, only 5% of KTr had a skeletal muscle mass index (SMI) below the 2.5th percentile of the reference population, and these few patients had no apparent increased risk of mortality.

The discrepancy between the results of these two studies may be explained by differences in the methods used to estimate skeletal muscle mass. Although SMI is adjusted for height squared, the authors did not adjust CER for body size. This may account for the unbalanced distribution of women in the lowest tertile of CER, with patients in this group being also older, and smaller. Although the authors subsequently incorporated morphometric parameters into the multivariate analyses, their methodological choice to use a muscle mass-derived variable, namely, the CER, without adjustment for height may have impacted their results.

Comparison of muscle mass assessed by CER or by methods based on CT-scan segmentation, with or without adjustment for body size, is therefore of interest for the aim of standardizing clinical practice.

In 127 of the 130 healthy subjects in our study, CER had been measured from four timed periods of 40 min. We investigated in these patients both the correlation and agreement between CER and SMI, CER and total lumbar muscle cross-sectional area at the third vertebra (MCSA) and then between CER and the product of MCSA by height. Because the latter provides a muscle volume rather than a muscle surface area, we hypothesized that MCSA × height could be a better surrogate marker of total muscle mass. We used the Pearson test to assess the correlation between the variables. Agreement was assessed both by Fleiss' kappa coefficient, and by calculating the proportion of patients classified in the same tertiles between the variables of interest.

The Pearson correlation coefficients were 0.723, 0.890 and 0.910 in the correlation analyses between CER and SMI, MCSA and MCSA × height, respectively (Figure 1). 64%, 81% and 93% of patients were classified in the same tertiles of CER and SMI, CER and MCSA and CER and MCSA × height, respectively. Fleiss' kappa coefficients were 0.47 (0.35–0.56), 0.71 (0.59–0.83) and 0.90 (0.78–1), indicating moderate, substantial and almost perfect agreement between CER and SMI, CER and MCSA and CER and MCSA × height, respectively.

These results confirm, as expected, that CER, measured in a standardized method using timed urine collections, is a surrogate marker of total muscle mass, but not of muscle mass adjusted to the body size of subjects.

Besides the issue of adjustment to body size, the discrepancies between the results of the two studies can be explained by the fact that Gaillard et al. did not assess muscle mass in the year before or at the time of kidney engraftment, as we did in our study. Early complications specific to KT may have influenced both mortality and early muscle mass loss and thus be confounding factors. The fact that CER depends not only on muscle mass but also on protein intake could also be a cause of discrepancy.³ The authors made adjustments based on urinary urea excretion. However, the assessment of protein intake from 24-h urinary urea excretion may be inaccurate because of urine collection errors. But overall, the excellent correlations and agreement between CER and the product of CSMA by height that we found suggest that the impact of protein intakes on CER is not highly significant, at least in healthy subjects.

We would also emphasize that CT-scan provides data on muscle quality (myosteatosis) and on other body composition data (visceral and subcutaneous fat masses). CT scan can be performed in patients on dialysis, whether or not they are registered on a transplant waiting list. Moreover, segmentation of CT scans for body composition analysis has become easier, more accurate and more reproducible with current and future software solutions.^{4, 5} Conversely, CER measured by repeated timed urine collections is a time and human resources-consuming procedure, and to our knowledge, there is no data assessing its inter-observer and intra-observer reproducibility. The main disadvantage of CT scan is the exposure to radiation, limiting its use mostly to patients having CT scan in the care setting. On the contrary, CER measurement has the advantage of being repeatable without worrying about radiation.

In conclusion, CER measurement and the CT-scan segmentation-based method for estimating skeletal muscle mass each have advantages and disadvantages. Both methods seem to be efficient for estimating total muscle mass, as we found an almost perfect agreement between CER and the product of MCSA by height. However, the concordance and agreement between SMI (adjusted to body size) and CER (not adjusted) were poorer. We fully endorse the European consensus guidelines on the definition and diagnosis of sarcopenia, which advocate adjusting both whole-body skeletal muscle mass and appendicular skeletal muscle mass to body size.⁶ These guidelines covered muscle mass assessment methods using bioelectrical impedance analysis, dual-energy X-ray absorptiometry, CT-scan and MRI. Similarly, we advocate that adjustment of CER to height squared or body surface area be required in clinical practice and for clinical research when the goal is to identify patients with low muscle mass.

All of the authors of this manuscript have no conflicts of interest to disclose.