Indirect estimation of nephron number: a new tool to predict outcomes in renal transplantation? Free

Since the 1980s there has been unprecedented interest in the total number of nephrons in the kidney. This interest was initiated by the work of Barker et al. [1, 2], whose pioneering epidemiological studies led to the developmental origins of health and disease (DoHAD) hypothesis, and Hostetter et al. [3] and Brenner et al. [4], who postulated that an inherited or acquired deficit in nephron number may be associated with increased risk of adult hypertension and renal disease. By combining these two hypotheses, the nephrology community began to consider, for the first time, that premature birth or low birthweight might lead to kidneys with low nephron endowment, which increases the risk of adult renal and cardiovascular disease [5]. Low nephron endowment is especially problematic, because human nephrogenesis ends before term birth, so any deficit in nephron number is permanent.

Since the early 1990s, nephron number has been estimated in autopsy studies of Danish [6], French [7], German [8], white and African American [9–11] and white and Aboriginal Australian kidneys [10, 11]. The average nephron number in a single kidney reported in most of these studies was ∼900 000, but the surprising finding in each study was the large range in nephron number in kidneys without evidence of renal disease. For example, in the largest study to date of 176 African Americans, nephron number ranged from 210 332 to 2 702 079, a 12.8-fold range [10, 11].

Is this large variability in adult nephron number present at birth (i.e. variable nephron endowment) and/or is it due to different rates of nephron loss with advancing age? Experimental studies have shown conclusively that a range of genetic variants/mutations influence the processes of nephrogenesis and thereby nephron endowment [12]. Similarly, a range of perturbations to the feto-maternal environment can lead to low nephron endowment [5]. Importantly in this regard, Hughson et al. [13] showed that nephron number in adult white and African Americans was directly associated with birthweight, with each additional kilogram in birthweight being associated with ∼260 000 additional nephrons per kidney. Hoy et al. [9] also showed conclusively that nephrons are lost at a rate of ∼6700 per year in adulthood.

The estimates of human total nephron number cited above were obtained using either acid maceration or stereological analyses on autopsy kidneys [14]. For estimates of nephron number to be utilised in clinical nephrology, either a non-invasive approach possibly involving imaging is required, or an indirect approach is necessary. Unfortunately, a non-invasive method for estimating nephron number in human kidneys has yet to be described. More than 20 years ago, Basgen et al. [15] estimated the number of nephrons in dog kidneys by combining estimates of renal cortical volume (obtained with MRI) with estimates of glomerular volume density and mean glomerular volume (obtained using needle biopsies). On average there was very good agreement between the fractionator stereological method and the MRI/renal biopsy method for the 10 dog kidneys measured, although a 36% difference between estimates for an individual kidney was observed for the two methods. In 2014, Beeman et al. [16] published the first report in which MRI was used to estimate total nephron number in whole ex vivo human kidneys. While this study utilized an exogenous marker of glomeruli (cationic ferritin) and was ex vivo, it represents an important step towards estimation of nephron number in whole human kidneys.

Given the current lack of non-invasive methods for estimating nephron number, clinical surrogate markers of low nephron number [5] may serve as a bridge between scientific theoretical knowledge and potential clinical applications. These surrogate markers may determine (i) low nephron endowment (i.e. low birthweight), (ii) nephron loss (i.e. older age) or (iii) a combination of both (i.e. glomerular volume).

The concept of an indirect measure of nephron number based on clinical data is both intuitive and intriguing. In the present edition, Schachtner and Reinke [17] provide a new method to estimate nephron number using an extrapolation strategy. Hughson et al. [13] and Hoy et al. [9] had previously provided linear regression analyses linking nephron number, birth weight and age for a Caucasian population. In a rather logical exercise, Schachtner and Reinke [17] calculated adult nephron number based on birthweight and age, which allowed them to estimate nephron endowment, and subtract the expected degree of age-related nephron loss. The results were consistent with previous reports using methods of direct estimation [10, 11].

The nephron loss after nephrectomy induces compensatory changes in the remaining kidney. Despite the 50% decrease in nephron number, the estimated glomerular filtration rate (eGFR) in living donors increases to 60–70% of pre-donation levels, suggesting an adaptive glomerular hyperfiltration [18], which may potentially lead to different outcomes in living donors compared with healthy non-donors. Two recent studies showed that the incidence of end-stage renal disease (ESRD) was 8–11 times higher in living donors than in healthy non-donors [19, 20]. In addition, another study showed that the incidence of proteinuria in living donors is significantly higher than in control groups [21]. However, the 15-year cumulative incidence of ESRD remains <1% and many studies have reported that living donors have similar life expectancy as the general population [22]. Interestingly, the risks for hypertension and CKD are significantly increased in African American, Hispanic or Aboriginal Australian donors compared with the risks in white donors. In these groups, low birthweight and low nephron number might contribute to the increased risk after nephrectomy [23].

Given that transplant kidneys are precious, minimally invasive procedures are performed and access to direct measurements of renal morphology is limited. Therefore, there is a clear need to develop prognostic tools that will determine transplant as well as donor outcomes. Schachtner and Reinke [17] investigated associations between indirect estimated nephron number in 91 living kidney donors and their kidney function, hypertension and proteinuria at 12, 36 and 60 months (median follow-up was 78 months). Indirect nephron number was strongly and positively correlated with eGFR. Furthermore, donor birthweight was inversely associated with the development of hypertension and proteinuria.

Overall, these findings suggest that indirect estimates of nephron number based on donor birthweight and age may provide useful clinical information that can assist clinicians in the education and selection of living kidney donors. However, we should be cautious when interpreting this information. While birthweight is a well-established determinant of nephron endowment, its use in isolation is almost certainly an oversimplification of a far more complex system. An adverse feto-maternal environment, even in the context of normal birthweight, may have a direct effect on renal morphology and thereby nephron endowment [5]. The use of multiple variables to assess gestational adequacy may provide a more realistic perspective. Similarly, the use of age as a unique determinant of nephron loss is also problematic. Multiple studies have shown that other variables such as body size and hypertension may be direct contributors to nephron hypertrophy and loss [24–26]. Therefore, further studies are urgently needed in order to construct more complex and realistic mathematical models that take into account such variables as sex, race, body size, hypertension and prematurity on nephron number.

It is certainly important to develop indirect measures of renal morphology, especially if they can have potential clinical applications. As an example, Grams et al. [27] recently developed a new tool for estimating the long-term risk of ESRD according to 13 predominant demographic and health characteristics. However, neither birthweight nor nephron mass was analysed in their cohorts. The study by Schachtner and Reinke [17] provides an important first step towards the development of clinical algorithms for the indirect estimation of nephron number. Is this tool ready to be used in day-to-day practice? The answer is probably not. However, we should applaud the efforts of clinicians who try to use available clinical data to generate otherwise unobtainable information such as nephron number. Whether a clinical algorithm for indirect estimation or direct in vivo measurements, we believe that nephron number has a place in the assessment of renal health and disease.

(See related article by Schachtner and Reinke. Estimated nephron number of the remaining donor kidney: impact on living kidney donor outcomes. Nephrol Dial Transplant 2016; 31: 1523–1530)

REFERENCES