Fractional clearance studies have determined how size affects the ability of molecules to cross the glomerular filter. When tritiated neutral dextrans were used to test glomerular permeability in the rat, their fractional clearance was 1 (no measurable restriction of filtration) when the effective hydrodynamic radius, a_e, was ≤20 Å; but with increasing a_e values, the fractional clearance decreased progressively until it approached 0 with radii greater than approximately 40 Å (19,20). Numerous studies over the years have confirmed that for neutral solutes, the sieving coefficient decreases with increasing molecular size (21). Charge also plays a role in controlling the transglomerular passage of macromolecules as suggested by the observation of Michael et al. (22) that there was a reduction in negative charge in the glomerular capillaries associated with proteinuria in rats with aminonucleoside nephrosis and in humans with MCD. Since these initial observations were made, fractional clearance studies using different sizes of neutral dextran and comparing their fractional clearances with those of negatively charged dextran sulfate of similar size and cationic dextran molecules supported the importance of electrostatic charge (23,24). However, others have argued that there is no charge selectivity and that negatively charged molecules of small size such as albumin may pass freely through the glomerulus to be reabsorbed by tubules (25). A raging controversy developed that is discussed below in the context of the properties of the capillary wall. Shape of the molecule undergoing filtration is another important determinant of filtration (26,27).

The various elements of the glomerular capillary wall act together in series to control the filtration of macromolecules forming both size and charge barriers. The barrier is dynamic with cross talk between the epithelial and endothelial cells (28,29). The size-selective barrier resides in all layers of the glomerular filtration barrier. Originally, the GBM was thought to be the principal size-selective barrier (7), but this is no longer the prevailing opinion (29). Nonetheless, the occurrence of proteinuria associated with mutations in laminin (a protein found only in GBM) supports a contribution of the GBM to the size selectivity of the barrier (29). The role of the filtration slit diaphragm, first described by Rodewald and Karnovsky (17), was confirmed by Kestila et al. (14) following their discovery of nephrin mutations in congenital nephrotic syndrome (CNS) of the Finnish type in association with a loss of the slit diaphragms. More recent work shows a heteroporous structure in the slit diaphragms with variation in the size of the pores under pathologic conditions (18). Edwards et al. found that the sieving curve was highly dependent on changes in filtration slit width and not in changes in the GBM (30). They concluded that size selectivity is most sensitive to the size of the slit diaphragm and the hindrance coefficient of the GBM (30). Most recently, a role for the endothelial surface layer has also been described for size restriction (31). This layer is composed of glycocalyx composed of proteoglycans bound to the cell membrane and an endothelial cell coat attached to the glycocalyx. This cell coat consists of proteoglycans, glycosaminoglycans, glycoproteins, and various plasma proteins such as
albumin and orosomucoid (31). Deen (32) using mathematical modeling suggested that the endothelial and epithelial cell layers are most important for size selectivity.

Originally, it was believed that the charge barrier resided chiefly in the anionic sites of the GBM (33). It is now generally accepted that the charge barrier also resides in the endothelial and/or epithelial cells (7,21,28,29). Jeansson and Haraldsson (34) gave hyaluronidase and heparitinase III to mice and measured glomerular selectivity in both in vivo and isolated perfused kidneys. They found evidence for both size and charge selectivity using a heterogeneous charged fiber model. Treatment with hyaluronidase increased albumin permeability fourfold without changing the selectivity of similar sized neutral molecules. Additional evidence for the presence of both size and charge barriers was obtained from a study of adriamycin-induced NS in the mouse. The investigators demonstrated increased clearance of larger Ficolls in treated mice as compared to control mice (35). Furthermore, loss of charge selectivity was also established by comparison of the clearance of neutral as compared to anionic albumin (35). These changes were associated with a thinning of the glycocalyx of the endothelial cells and down-regulation of synthesis of the heparan sulfate proteoglycans (35). However, it has also been suggested that no charge barrier exists. Some investigators believe that albumin passes through the capillary wall but is then returned to the plasma by a proximal tubular albumin retrieval pathway in which megalin, a low-affinity albumin-binding receptor, directs albumin to lysosomes or via transcytosis to the basolateral membrane and cubilin, a higher-affinity albumin receptor, directs albumin to lysosomes (25). These authors suggest that proteinuria is due to inhibition of this retrieval pathway rather than increased glomerular permeability. This group using in vivo two-photon microscopy demonstrated that the glomerular sieving coefficient for albumin was higher than expected to a degree that supported lack of a charge barrier (36). This raised a storm of controversy with contentious articles and editorials. Several other investigators using the same two-photon technique demonstrated a glomerular sieving coefficient for albumin that was close to or at the expected value (37,38) supporting the presence of a charge-selective barrier. Most recently, Sandoval et al. (39) have shown that the glomerular sieving coefficient may vary in different strains of rats and under different conditions. Thus, the controversy is not yet entirely resolved and requires further consideration as our understanding evolves (40). Nonetheless, most investigators agree that a charge-selective barrier is present in the glomerulus. Additional evidence for the presence of a glomerular charge barrier is supported by the association of proteinuria with glomerular injury particularly the accumulating evidence of mutations to various elements of the glomerular filter and the occurrence of proteinuria (28).

Hemodynamic factors, including blood flow, convection, diffusion, transcapillary hydraulic pressure difference, and intra-glomerular pressure, have been shown to play an important role in the passage of macromolecules across the glomerulus (19). Fractional clearance of a solute depends on the relative transport of water and the molecule in question. The latter takes place by both convection and diffusion for intermediate-sized macromolecules (an a_e value between approximately 20 and 34 Å). For small and large macromolecules, convection is the predominant mode of solute transport. An increase in the glomerular filtration rate (GFR) results in comparable increments in water flux and solute transport by convection so that no change in fractional clearance occurs when diffusion is not a factor. However, in the case of intermediate-sized molecules, an increase in GFR produces a decrease in diffusion and, therefore, a reduction of fractional clearance. Afferent arteriolar plasma protein concentration may also affect these factors. These findings, predicted on theoretical grounds (20), were confirmed using dextran under conditions of different glomerular plasma flow rates, one of the determinants of the GFR. Other evidence that glomerular hemodynamics may affect glomerular permeability is provided by the observation that angiotensin-converting enzyme (ACE) inhibition reduces proteinuria. Another structural feature that may affect water or macromolecular flux through the glomerulus is the SPS first demonstrated in three-dimensional reconstruction of the glomerulus (41). This space is bounded by the podocyte cell body and the filtration barrier. Filtrate leaves the SPS and enters the urinary space through small exit pores between cell bodies, which produces a high-resistance space that may be controlled by the podocyte (3,41,42).

A recent contribution from Hausmann et al. (43) suggests the possibility that extracellular potential differences could be a determinant of glomerular filtration. They studied the Necturus maculosus (common mudpuppy), which has sufficiently large glomeruli to undertake such a study, and found a potential difference of 0.09 mV at a perfusion pressure of 20 cm H₂O that could be eliminated by perfusion with a cation such as protamine. They explained this phenomenon as a streaming potential. This concept was further elucidated in an accompanying editorial in which a dilute ionic solution passes through a negatively charged filter, the positive ions will distribute in a uniform fashion (44). However, application of flow under pressure will cause the positive ions to flow through the filter where some will stick to the nonplasma side. Some of the negative ions will be reflected resulting in a streaming potential difference. Another editorial in the same issue suggests caution in the interpretation of these results (45).

Thromboembolic episodes occur in 20% to 42% of adults and in 1% to 9.2% of children with NS (92) and constitute the most life threatening of the complications of the NS. Deep venous thrombosis, renal vein thrombosis (RVT), and pulmonary embolism are the most common forms, but there are reports of arterial thrombosis and coronary thrombosis. Arterial thrombosis accounts for a higher percentage of thrombotic complications in children than in adults and results in a high rate of limb loss/death (93). In one study, the incidence of RVT in patients with NS was 35%, while the overall incidence of thrombotic complications outside the kidneys was 20% (94).

The pathogenesis of hypercoagulability in NS is multifaceted, stemming from an imbalance between coagulation and fibrinolysis (92). First, one must consider the presence of a genetic predisposition such as antithrombin deficiency that would increase the likelihood of thrombus formation (92). Many procoagulant factors are increased secondary to the lower serum albumin to which they are usually bound. These include fibrinogen; factors V, VII, and VIII; and von Willebrand factor (7,92,95). Increased fibrinogen leads to increased platelet aggregation and β-thromboglobulin. On the other hand, some anticoagulant substances are lost in the urine, including factors XI and XII, plasminogen, and antithrombin III (ATIII) (7,92,95). The loss of ATIII is thought to be particularly important as patients with NS who have such a deficiency have a 50% to 70% risk for thromboembolism (93). The fibrinolytic system is tilted toward decreased fibrinolysis with decreased plasminogen and plasminogen activator (92). It has been shown that the fibrin clot structure in NS may be less porous so that the thrombi are more resistant to fibrinolysis (92). Furthermore, prothrombotic microparticles, formed from various blood cells, may be increased in NS (92).

In children, the risk for thromboembolism is greater in those patients with CNS or with secondary forms of NS (92). In adults, the risk of RVT is higher in patients with membranous GN with 37% as compared to 24% for other forms of glomerular disease (92). Presentation of RVT varies with age. In young adults, it presents with flank pain and microscopic hematuria, whereas in older adults, presentation is characterized by a chronic syndrome. Furthermore, extrarenal thrombosis is also more common in older adults (92). Other clinical signs include acute renal enlargement, hematuria, oliguria, and acute renal failure (ARF) (96). Thromboembolic events occur early in the course of NS, usually within the first 6 months after diagnosis (97).

Infections were a major cause of mortality in patients with NS in the era before antibiotics (98), and bacterial infections remain a significant cause of morbidity. The prevalence of infections ranges from 8% to 38% in children (99), and one adult series reports a prevalence of 19%. Types of infections include sepsis, peritonitis, cellulitis, pneumonia, urinary tract infection (UTI), and meningitis (100,101,102). In 340 French children hospitalized with NS, severe bacterial infection was diagnosed 48 times (14%) in 32 patients. One half of the patients had spontaneous peritonitis, most often as the result of Streptococcus pneumoniae infection (99). The frequency of different types of infection varies by geographic location. A recent study from China found that 19% of admissions to a hospital for NS were also attended by infection (102). Pneumonia was most common followed by UTI, sepsis, peritonitis, and cellulitis. Age also affected the type of infection with children under 10 years suffering from pneumonia, while those older than 10 years had UTI most commonly (102). Infections with S. pneumoniae are sufficiently common that it is recommended that children with NS should be vaccinated against capsular antigens (100).

Comparisons of patients with NS with and without infections have cited risk factors for infection that include elevated levels of serum cholesterol (95), a serum IgG concentration of less than 600 mg/dL, a serum creatinine level of more than 2.0 (103), and increased fragility of the skin as a portal of entry (7). The heightened susceptibility to infection in patients with NS is usually attributed to hypogammaglobulinemia owing to urinary losses of immunoglobulin, increased catabolism, and a reduced rate of synthesis (7,95,100,101). Cellular immunity is also abnormal in the NS (95,100,101). These defects include a reduction in the numbers of circulating T lymphocytes and a depressed blastogenic response to mitogens (95). An increased incidence of infection results from these abnormalities in immunity and from inhibition of leukocyte chemotaxis (95). It has also been suggested that zinc deficiency resulting from urinary losses may contribute to an abnormal immune response in patients with NS (95). Loss of serum complement and opsonizing factors also contribute to the increased frequency of infection (95).

Hypercholesterolemia and hyperlipidemia are almost always present in patients with NS, and these biochemical abnormalities may contribute to the progression of glomerular disease (104). Hyperlipidemia in patients with NS may also cause systemic atherosclerosis. The lipid profile seen in patients with NS is characterized by hypercholesterolemia, elevated plasma LDLs and lipoprotein (a), hypertriglyceridemia, and alterations in HDL metabolism (81,105). The increases in cholesterol and LDL are due to increased synthesis of both moieties and decreased catabolism of cholesterol (83,105). Hypertriglyceridemia is due to decreased clearance of VLDLs and chylomicrons and due to the up-regulation of hepatic diacylglyceride acyltransferase (81,106). The changes in HDL metabolism are described in the section above on pathophysiology of hyperlipidemia. Increased lipoproteins and oxidized lipoproteins have also been described in children with NS (107). Recent evidence suggests that increased oxidative stress, a known risk factor for atherosclerosis, in NS may be due at least in part to decreased paraoxonase-1 activity in patients with NS (108). Paraoxonase-1 prevents oxidation of serum lipoproteins and reduces risk of atherosclerosis.

Ordonez et al. (109) compared the risk of coronary heart disease in 142 nondiabetic adults (older than 15 years old) with
NS (proteinuria more than 3.5 g per day) and nonproteinuric control subjects matched for age and sex. Neither the subjects nor the controls had a preexisting coronary heart disease. The authors estimated that the risk of myocardial infarction (MI) is between five- and sixfold higher for the patients with NS than for the controls and that all coronary heart disease events (MI, angina pectoris, and coronary insufficiency) and deaths from coronary heart disease are two- to threefold higher in these patients. Exclusion of 17 patients with MCD from the analysis did not substantially alter the results. The increase in MIs may be related to increased LDL, resulting in decreased nitric oxide-mediated vasodilation (110). Another possible mechanism of increased MIs is related to loss of lysophosphatidylcholine resulting in decreased deformability of erythrocytes and increased platelet adhesion (110). Given the apparent risk of catastrophic complications, patients with prolonged NS and hyperlipidemia should be considered for pharmacologic or dietary treatment aimed at lowering the plasma lipid levels (109).

MCD is more common in boys than in girls; the ratio of boys to girls is approximately 2:1 (114,115,118). It is most commonly a condition of young children; with biopsy verification, the peak incidence was between 2 and 4 years (114,118). Almost 80% of histologically verified cases of MCD in the International Study of Kidney Diseases in Children (115) were found in children under the age of 6 years, with a median age of 3 years. However, MCD can occur at any age and is a common cause of NS in adults. It is more common among Caucasians, Asians, and Hispanics than among African Americans (137). The incidence is 2 to 16/100,000 per year in children under 16 years (137).

The clinical manifestations are similar to those seen in other forms of the idiopathic NS, and edema is the most common presenting sign (114). Microscopic hematuria was found in 36% of histologically confirmed cases in one study (114) and in 13% in another (118). Macroscopic hematuria was rare. Hypercholesterolemia is seen more frequently and is often elevated to very high levels. Blood pressure is usually normal at onset, and <20% of patients have hypertension (114,115,118). Transient depression of renal function has been recorded in 0.8% (138), 9.6% (114), and 19% (118) of children with MCD. In those cases in which the clearance is depressed initially, there is a return to normal with remission. The significance of decreased renal function in patients with MCD is discussed later (see “Acute Renal Failure in Minimal Change Disease”). Serum complement levels are not usually depressed (114,115,118). Proteinuria is almost always of the selective type (114,118). In fact, Adamson et al. (139) have suggested that the presence of higher levels of gamma globulins (greater
than 4.3%) on urine protein electrophoresis in conjunction with a normal GFR predicts an increased risk of FSGS.

MCD was defined in patients with NS, and there is a reluctance to make the diagnosis in patients with asymptomatic proteinuria. Hiraoka et al. (140) addressed the issue of “asymptomatic MCD” by reporting the cases of eight children with proteinuria discovered by chance on urinary screening. Two of the eight children underwent renal biopsy for coexistent hematuria, and the histopathologic diagnosis was MCD. None of the eight asymptomatic children relapsed for ≥1 year after completion of treatment, and only two experienced steroid-responsive relapses in greater than 6 years of follow-up. The authors concluded that among patients with steroid-responsive NS and a clinical diagnosis of MCD, those with mild manifestations without edema will have a favorable clinical course without relapse. Branten et al. (141) reported on a mother and two daughters with persistent proteinuria for 20 years in the mother but without any abnormality on biopsy including normal foot processes. These authors suggested the possibility that this represented a familial nephropathy different from MCD. On occasion, isolated FPE is found in patients with slightly elevated urine protein or no proteinuria. Because the relationship of isolated FPE to MCD is unknown and it is uncertain that the patients require treatment with steroids, I recommend that such patients be classified as having nondiagnostic glomerular changes. It is important to note that patients with MCD may have less than widespread FPE if the biopsy is performed when the patient is entering spontaneous remission or following treatment.

MCD is a major cause of the NS in adults (see Table 5.2), but in contrast to the situation with children, it is not the most common glomerular pathologic manifestation (111,142,143,144,145,146). The clinical presentation is usually similar to that seen in children (142,144), although some studies report a higher incidence of hypertension, ARF (111,142,143,144,145,146), and nonselectivity of proteinuria (111), particularly in the elderly (128). The condition is identical to MCD of childhood so far as the appearances by light, immunofluorescence, and electron microscopy are concerned, except that obsolete glomeruli, focal tubular loss and atrophy, and thickening of arterioles and arteries are seen more often in adults. Children have more frequent relapses (137). Secondary cases of MCD are more common in adults, and in some 10% of cases, it is associated with a drug reaction or a lymphoproliferative disorder (142,144). A familial form of steroid-responsive NS, likely representing MCD, has been reported in a Bedouin kindred (147).

De novo MCD has been described (148,149,150). In one series of 67 patients with posttransplant NS, five had MCD (150). Fifteen cases have now been reported, and eight were in living-related donor kidneys (148,150). Many of these cases occurred within a few months following transplantation. None of the recipients had FSGS as a primary disease. Most remitted with appropriate therapy. It is critical to exclude other possible causes of NS prior to making this diagnosis.