COMMENTARY The glomerulus is a remarkable structure. The glomerular tuft is a meshwork of specialized capillary vessels that operates at high hydrostatic pressures (approximating 55 mm Hg) supported by specialized type IV collagen with high tensile strength, to produce glomerular ultrafiltration into Bowman’s space as the first stage of urine formation. Intraglomerular pressures are controlled by differential afferent and efferent resistance vessels, which are controlled by tubular-glomerular feedback, the renin–angiotensin–aldosterone system and prostaglandins. The glomerulus is composed of 600–700 cells (mostly mesangial and endothelial cells, fewer visceral epithelial cells). Their health within a transplanted kidney can affect function, proteinuria, and outcomes. Measuring glomeruli would seem like a good idea, and multiple studies in transplanted and native kidneys have found relationships with glomerular size with outcomes, although not always in a common direction.1,2 In current study published in this edition of Transplantation,3 Grande et al obtained 1027 indication biopsy samples for transplant dysfunction and measured glomerular planar surface area and diameter by image analysis on all glomeruli containing a vascular pole (excluding sclerosed glomeruli) from the 2 cohorts of the Deterioration of Kidney Allograft Function study. BKV infection and recurrent glomerulonephritis were excluded. The study reported glomerular size increased with time (r = 0.317) and was greater in the later cross-sectional cohort (mean 7.7 ± 5.6 y posttransplant) compared with early prospective cohort which used the first biopsy-for-cause samples (at 1.7±1.4 y), congruent with the findings from studies. An adaptive response to hyperfiltration (a weak association with recipient BMI was noted), or more likely, a greater prevalence of chronic-active antibody-mediated rejection may be the reason. Strong correlations of glomerular size with histological markers chronic-active antibody-mediated rejection included “active” scores of glomerulitis and peritubular capillaritis (“microvascular inflammation”), and chronic structural markers of transplant glomerulopathy (Banff cg scores) and associated mesangial response of increased matrix formation (Banff mm scores) and segmental glomerulosclerosis.4 Glomerular area results was independent of tubulointerstitial damage, and unrelated with C4d (some late AMR can be “C4d negative”) and donor specific antibodies (DSA) detection. AMR rates have fallen with better HLA matching and solid-phase antibody detection; however, late rejection involving DSA remains a threat associated with underimmunosuppression and nonadherence. These histological relationships were observed in both cohorts, although subanalysis by cohort found trend differences for clinical factors such as retransplant, male recipient sex, and DSA for early cohort, versus prior acute rejection, posttransplant diabetes mellitus, and Black race for late biopsies. These bigger glomeruli could reflect a glomerular response to local inflammation and chronic structural change related to antibody effects, which has biological plausibility and constitutes an adverse risk factor as such. Against that, glomerular size weakly correlated with estimated glomerular filtration rate (a possible beneficial adaptive response). The unadjusted hazard ratio for death-censored graft failure was unaffected by glomerular size. Only after extensive statistical adjusted of covariates by Cox methodology, produced a significantly reduced hazard ratio for graft failure of 0.967 (95% confidence interval [CI], 0.948-0.986) to 0.973 (95% CI, 0.955-0.992), for the best 2 models (per 1000 µm2 glomerular area), implying that larger glomeruli may be beneficial. What do these results mean? WHAT DETERMINES GLOMERULAR SIZE? Glomerular size varies widely and depends on the number, mean dimension, and structural integrity of each individual glomeruli, which are affected by nephron endowment of the transplanted kidney and pathological factors that alter morphological structure. Glomerular size and its variability are markers of glomerular stress, and have been associated with birth weight, (donor kidney) age, BMI, hypertension, and glomerular number: risk factors for future kidney disease.2 Small sclerotic glomeruli can follow inflow vascular changes of the afferent arterioles from calcineurin inhibitor nephrotoxicity or hypertension, observed as shrunken ischemic glomeruli, arterial wall thickening of interlobular arteries, and arteriolar hyalinosis. Atubular glomeruli are disconnected from downstream damaged tubules, still maintain limited ultrafiltration, are also reduced in size and common in damaged allografts. Enlarged (or nonshrunken) glomeruli are seen in recurrent diabetic nephropathy, infiltrative diseases, glomerulonephritis, and with transplant glomerulopathy, the pathological hallmark of chronic AMR. DOES GLOMERULAR SIZE MATTER IN KIDNEY TRANSPLANTATION? The results of glomerular size and transplant outcomes are inconsistent. Other transplant studies have shown that larger glomerular size, measured by point-counting, predicted late allograft dysfunction from implantation,5 and in 4-mo protocol biopsies,6 especially if associated with functional impairment. Those results suggest that smaller glomeruli do better in the long-term, independent of other variables, and contradict the current results. Reduction in functioning nephrons has been postulated to lead to glomerular hypertrophy, hypertension, and increased single-nephron hyperfiltration, resulting in glomerulosclerosis and additional nephron loss. This common pathway of hyperfiltration leading to end-stage renal disease remains debated. Glomerulosclerosis and glomerular size are correlated in the rat remnant kidney model, some human glomerular diseases, and in rare congenital diseases such as oligomeganephronia, where the few, very large glomeruli progressively undergo glomerulosclerosis and renal insufficiency. IS GLOMERULAR SIZE A USEFUL PREDICTIVE TOOL FOR TRANSPLANT CLINICIANS AND PATIENTS? There are several reasons why the answer would be a simple no. First, the extensive biological scatter (against time and estimated glomerular filtration rate, see figures) is consistent with the known variability of glomerular area, but increased standard deviations would reduce any test predictability. Second, measuring glomeruli and deriving reproducible, quantitative information from biopsy sections is far more difficult than expected at the first sight.1,2,7 Glomerular size is that most popular estimate of inferred total glomerular mass, but as a single estimate of glomerular volume, is acknowledged to provide limited information of true glomerular dimensions. Technical issues with dimensional glomerular profile measurement (including the diameter or area) include section thickness, optically lost caps, fixation effects (lower size with immersion-fixed tissue versus perfusion fixation), and the cortical zone sampled (juxta-medullary glomeruli are larger than superficial glomeruli). Cut section planes do not always correspond to the central “equatorial” glomerular section, and inferences assume spherical glomeruli (most are actually ellipsoid shape), and generally underestimate real glomerular size. Stereological evaluation of glomeruli (as in this article) involves definition of a region of interest by manual outline tracing are time consuming, and entail specialist training, additional image analysis equipment, and reproducibility validation testing before clinical implementation. Even a mathematical average of glomerular area within the section is a poor representation of the biological populations of glomeruli within any scarred kidney, which comprises variable numbers of small pathological glomeruli admixed with enlarged hyperfiltering normal glomeruli. Deep-learning algorithms of transplant biopsy tissue digital image analysis (including glomerular profiles, which can be recognized and sized using convolutional neural networks to produce multiclass segmentation) may be the future8; however, even these still need conceptual support from consistent morphometry results. And finally, the lack of significant association of area with death-censored graft survival in the raw actuarial survival curves by quartiles of glomerular area will be the ultimate arbiter that defines the clinical utility of this pathological marker, despite a modest effect produced after extensive statistical adjustment of covariates. Hence, it seems that a simple averaged glomerular size in isolation does not seem to matter much, however, but may become important if assessed by alternative or additional morphological techniques (including glomerular volume, glomerular number, or areal density), or sophisticated machine learning techniques, and always within the context of whole transplanted kidney cortical volume and donor organ quality.