On repeat biopsy, it is not uncommon to find cases of early FSGS in patients who were initially diagnosed with MCD [19]. Thus, this is one of the confusing clinical scenarios in nephrotic syndrome, when LM reveals no abnormalities in the glomerulus and the patient’s response to steroid therapy is weak. EM is not always available either due to higher costs or to the longer time needed to carry out the examination and to get the results. Accordingly, any significant finding in the LM picture should never be neglected, as it may help predict the prognosis and explain the disease pathogenesis, ultimately affecting the treatment decisions. Moreover, with the recent discoveries on the role of podocytes in the pathogenesis of this group of glomerulonephritis, it has become essential to fully describe podocyte abnormalities in both LM and EM reports. One of the limitations of the available podocytopathy studies is the limited number of patients who had EM examinations [20, 21]. If we could assess their podocytes by LM, this might open the door for clinical studies with a larger sample size. In our examination of kidney biopsies of patients with NS without abnormalities in the LM picture, we observed few cells in the urinary space without attachment to the BM, as shown in Fig. 2. We hypothesized that these cells were detached podocytes and that this could direct the diagnosis toward early FSGS instead of MCD. This differentiation works in harmony with the Barisoni classification of primary podocytopathy which considers lost podocytes as the differentiating factor between FSGS and MCD [4].

Thus, we aimed to assess the role of LM in detecting detached podocytes. To the best of our knowledge, no study has discussed the ability of LM to detect podocyte detachment by routinely used stains. All the available studies discussed podocyte detachment as described with  EM or by special stains or by counting all the currently available podocytes within the glomerulus [9, 10, 16, 22]. We made the hypothesis that if we succeeded in confirming the ability of LM to detect detached podocytes, this would add a new, inexpensive, and easy method for in-depth evaluation of LM sections in patients with apparent MCD.

Several previous studies that tackled MCD and FSGS had a major risk of bias, which is related to the absence of EM examination. Accordingly, a large percentage of the included patients might have been misdiagnosed as MCD [20, 21, 23, 24]. EM is essential for unmasking cases of early FSGS that may present as MCD. MCD is not well studied in Egyptian adults. We retrieved all cases with the previously mentioned dilemma “apparent MCD by LM” in the period from 2015 to 2020. There were 118 patients; 50 adults and 68 children. In our study, of the 83 EM biopsies, 74 (89%) showed findings other than MCD, 69 (83%) showed podocyte detachment, 48 (57.8) were diagnosed as early focal segmental glomerulosclerosis that was not revealed until EM was done, only 9 (10.8%) were diagnosed as MCD, 25 (28%) had BM abnormalities, 3 (3.6%) had an early membranous pattern of GN that was not revealed until EM was done and 22 (24%) had EM characteristics of Alport syndrome.

To assess the accuracy of LM in detecting these detached cells, we had to compare the results to a gold standard. Thus, we compared the LM findings to EM and we also correlated them to the response to steroid treatment, as steroid response is considered a better predictor of early FSGS than EM [25].

In our study, the presence of detached cells in the LM picture was significantly correlated with the degree of proteinuria; cases with detached podocytes had a higher incidence of nephrotic range proteinuria. This agrees with the results obtained by Hanamura and colleagues who reported that detached podocytes detected in urine samples were associated with higher proteinuria [10]. Hanamura and colleagues also failed to find any association between urinary detached podocytes and BP, hematuria, or eGFR; these findings were similar to our study.

Response to steroid treatment is expected to be better in cases without podocyte detachment, as they encompass cases of MCD. It is reported that response to steroids is a valuable tool to differentiate between MCD and FSGS [25]. In the present study, multivariate logistic regression showed a similar finding; only podocyte detachment by LM was a statistically significant independent predictor of steroid resistance. Cases with LM detachment had 4.5 times higher odds to exhibit steroid resistance. Taneda et al. and Abumregha et al. also reported a higher incidence of steroid resistance among FSGS cases in comparison to MCD [26, 27].

Detached podocytes in LM were confirmed by EM in 93% of cases. When there was detachment in ≥ 25% of biopsied glomeruli or mean detached cells per affected glomeruli > 1, EM confirmed podocyte detachment in 100% of the cases. This indicates a high specificity and high positive predictive value of LM, which is expected, since LM was used to count the total number of podocytes per glomerulus, and podocytes were identified as cells that reside within the tuft of the glomerulus but outside the glomerular basement membrane [10, 11]. Podocyte detachment by LM was not confirmed in two cases by EM. This can be explained by the focal distribution of the lesions. Routine EM preparations choose only one or two glomeruli for examination; if the examined glomerulus was normal, the results could be non-representative, while LM is more representative because it has a larger sample size. Although the positive predictive value was good, there were several false-negative cases, i.e., 34 out of 63 cases with no LM detachment, and this led to low sensitivity (46%). Thus, Cohen’s Kappa analysis showed slight agreement between LM and EM in assessing detached podocytes, even with a significant p value of 0.028.

This study focuses on determining the diagnostic value of detecting those detached cells which are assumed to be podocytes detaching from the basement membrane. To be able to differentiate early FSGS from MCD based on this LM finding, there are two steps for verification. The first step is to ensure that these cells actually are podocytes (which has been done successfully in this study) by reviewing the EM and searching for detached podocytes. Standardized reporting of these cells is mandatory as they have a high positive predictive value for the EM picture of detached podocytes, and this was the main aim of our study. Such finding is essential not only in the pathology of FSGS but also in multiple other glomerular disorders including diabetic nephropathy, membranous nephropathy, lupus nephritis, and others.

The second step should be to ensure that patients with detached cells have FSGS and not MCD, and this is the debatable point. Detachment is the primary pathogenic step that leads to the formation and progression of sclerosis. Animal models of podocyte injury using diphtheria toxins or Puromycin confirm this pathogenetic pathway; animals with no, or minimal podocyte detachment, resemble MCD cases. Detachment—if present—should be below the compensation limit of the glomerulus. Increasing the dose and duration of the toxic substance will increase the podocyte injury beyond the compensation limits leading to FSGS [28]. In humans, Royal et al., classified renal biopsies from > 200 patients with minimal change disease or FSGS according to ultrastructural pathological changes into six clusters. Clusters 4, 5, and 6, which included the majority of FSGS cases, had the highest probability of including detached podocytes and vice versa [29]. Additionally, in a recent study, Zee et al. described the detachment of podocytes in LM images as a halo around the podocytes. This halo was present in 19% of FSGS cases and absent in MCD and MCD-like cases, with clinical correlations similar to those found in our study [30]. Furthermore, Ahn and Bomback stated that in their approach to diagnosing primary podocytopathy: “There are hallmark differences in the histologic appearances of MCD and FSGS, which in turn represent distinct pathogenic models after initial podocyte injury (e.g., no change in podocyte number in MCD vs podocyte detachment and death in FSGS)” [5]. A similar concept was mentioned in Wiggins’ spectrum of podocytopathy: “Podocytes are not lost into the urine in MCD with the maintenance of the normal podocytes numbers compatible with the excellent long-term prognosis for renal function”.

To confirm the diagnostic and prognostic value of podocyte detachment we need a study that has follow-up biopsies for patients who were initially diagnosed as MCD, after which the second biopsy showed FSGS. Thus the first biopsies have to be reviewed for the presence or absence of detached cells. In our cohort, we found only one patient with repeat biopsy; the first biopsy showed MCD with detached cells, and his second biopsy showed FSGS. Of course, one patient is not enough. Thus, we can not be sure that if detached cells are detected, the diagnosis is FSGS. A helpful approach for such patients is to confirm the state of podocytopenia by a quantitative assessment of the remaining podocytes using special podocyte stains that were not feasible in our study. Even though we cannot confirm the diagnosis of FSGS, these cases would be better classified as podocytopathy with detached podocytes rather than MCD, since podocyte loss is a change that should not be reported as “Minimal”.

One of the unexpected findings in the present study is the high incidence of Alport syndrome features in EM results. The majority of similar studies reported a lower incidence than in our study where approximately 25% of cases showed features of Alport syndrome [20, 24, 31]. This unusual finding may be explained by the presence of a different genetic background in our cohort with different inherited mutations and consanguinity patterns, in addition to differnce in inclusion and exclusion criteria between studies, and selective availability of EM. In our community, previous studies reported Alport syndrome to be the most common cause of hematuria in Egyptian children and Alport syndrome represents > 16% of the original kidney disease in transplanted children [13, 32]. Alport syndrome mostly presents with hematuria, but it may present with nephrotic syndrome even in the absence of hematuria [33, 34].

There is a grey area between cases of FSGS and cases with a nephrotic presentation of Alport syndrome. Effacement of podocyte foot processes and podocytopenia were reported in cases of Alport syndrome [35, 36], but Tsuji et al. reported that no podocyte detachment was present [37]. Thus, these detached cells may indicate the concomitant presence of FSGS, a podocytopathy that is characterized by podocyte detachment and podocytopenia, according to the taxonomy of Barisoni and colleagues [4]. The presence of detachment in Alport syndrome may further increase the interest of the studied approach. Based on this fact, noticing a detached podocyte can distinguish MCD from early Alport syndrome, that might present with no abnormality by LM, in the absence of other differentiating parameters.

To the best of our knowledge, this is the first study that tries to validate the ability of LM to detect detached podocytes by routine stains. Moreover, this study discusses a common finding in nephropathology practice and tries to find an evidence-based answer to the following question: “should the nephropathologist ignore this detached cell or does it deserve to be reported?” Particularly, he/she does not have to use expensive special stains or sophisticated techniques to report such essential findings when it is present in the routine LM examination. Additionally, MCD has not been adequately studied among the Egyptian adult population. This is the first study to combine clinical, laboratory, and pathological (LM and EM) characterization in adult MCD patients in Egypt using EM in the majority of the studied population. Most studies that evaluated FSGS cases ignored the cases that were diagnosed as apparently MCD by LM. Our study is one of the few that focused on the hidden spectrum of FSGS cases and described its characteristic presentation, laboratory, and pathological features, and response to treatment.

Our study has limitations. First, it was retrospective. If we had the chance to carry out a prospective study, we would be able to add a few items to the studied parameters such as correlation with urinary podocytes. Second, limited resources led to the absence of genetic analysis for the studied population. Discovering the genetic background could help to better classify and understand the disease pathogenesis. Particularly, genetic podocytopathy does not always present with a positive family history or syndromic features, it has even improved with the administration of immunosuppressive agents in a few occasions [33, 38]. A higher number of MCD patients, particularly those who are steroid-responsive, would add more statistical power, but in routine practice, those cases usually do not proceed to kidney biopsy. This explains why we have a higher prevalence of steroid resistance among children in this study, as well as a high incidence of Alport syndrome in our series. Further in-depth studies of cases of Alport syndrome would solidify the diagnosis, not only by genetic analysis but also through audiograms, lens examination, special collagen stains for basement membrane, and examination of family members, all of which might help achieve a better understanding of such an under-reported disease. Podocytopenia may result from the actual loss of podocytes by detachment or from increased glomerular size with a fixed number of podocytes which is known as relative podocytopenia. Our study tackled only the first mechanism. Special podocyte stains and stereological methods for quantifying podocytopenia can indeed be more accurate in assessing and quantifying podocyte loss, but EM and response to steroids were the best available gold standards for comparison with our finding of LM detachment. Moreover, histological approaches might miss some cases due to sampling error. Response to steroids was reported by KDIGO GN 2021 guidelines as a better differentiating tool in such conditions. Moreover, there were concerns regarding the proper identification of podocytes by routine LM stains, as these free cells might be detached tubular, inflammatory, or even parietal epithelial cells. Morphological characteristics of podocytes are distinct from parietal cells, which are flat and not rounded, and from tubular cells that should normally be shed distally with the flow of the filtered urine away from the glomerulus. Moreover, the pathogenesis of this group of diseases includes only detachment of podocytes and not parietal cells without inflammatory cellular infiltration in the Bowman’s space.

We suggest that future studies should prospectively include staining for one or two special podocyte markers (such as ZO-1, WT-1, synaptopodin, CD2AP, podocin, podocalyxin, nephrin) to ensure better identification of these detached cells. The same approach can be followed in cases where LM detects detached cells and EM reports no detached podocytes. In the current study, we retrospectively tried to confirm our findings by using a special podocyte stain. Out of 29 patients who had detached cells on routine light and electron microscopy, 9 patients had sufficient glomerular tissue in their paraffin blocks. After staining by colloidal iron, 5 patients had detached cells. Considering the 3-dimensional configuration of the glomerulus, detached cells may be present in one section and not present in another within the same glomerulus. This explains why not all patients were positive by the special stain and adds to the low sensitivity of our finding. On the other hand, detached cells that were stained by colloidal iron in the five patients confirm the relatively high specificity of this test.

In our study, five cases showed binucleated podocytes indicating mitotic catastrophe. EM showed detached podocytes in all of them. Among those five patients, only one had positive LM detachment as well. Future studies may also include an assessment of the viability of podocytes while still anchored to the outer glomerular basement membrane; apoptosis, mitotic catastrophe, necroptosis, pyroptosis, or vacuolizations can be explored by LM among those patients with apparently normal glomerulI.