Nephrotic syndrome is a collection of symptoms and signs resulting from damage to the glomeruli (the filtering units of the kidney), which causes severe proteinuria (leakage of protein from the blood into the urine), low blood levels of protein, and swelling.
Nephrotic syndrome may occur as a result of diabetes mellitus, amyloidosis (accumulation of amyloid, an abnormal protein, in the tissues), or any type of glomerulonephritis (inflammation of the glomeruli). The conditions may also be due to severe hypertension (high blood pressure), reactions to poisons (such as mercury or cadmium), or adverse reactions to drugs, such as gold or penicillamine. Nephrotic syndrome may also be a complication of an infection elsewhere in the body, such as hepatitis B or malaria. In children, the syndrome is most commonly due to a condition known as minimal change glomerulonephritis.
The symptoms of nephrotic syndrome appear gradually over days or weeks and worsen as more and more protein is lost in the urine. Early signs include frothy urine and decreased urine production. The main symptom is swelling of the legs and face as a result of oedema (accu-mulation of fluid in the tissues). Fluid may also collect in the chest cavity, resulting in pleural effusion and shortness of breath, or in the abdomen, causing ascites. Lethargy and loss of appetite (leading to weight loss) may also occur.
Diagnosis and Treatment
Diagnosis involves blood tests; the urine must also be examined under a microscope and the amount and type of protein lost assessed over a 24-hour period (see urinalysis). In most cases, a kidney biopsy (removal of a small sample of tissue for microscopic analysis) is also required to establish the exact cause and help with treatment decisions. Treatment is aimed at the underlying condition. A low-sodium diet may be recommended and diuretic drugs may be given to reduce oedema. Protein may need to be given intravenously, but the effects are short lived. Corticosteroid drugs may also be prescribed and are often effective in cases of nephrotic syndrome in childhood.
The outlook for someone with nephrotic syndrome depends on the extent of the kidney damage. Recurrent episodes may occur even after treatment. Problems with excessive blood clotting are common and may require treatment with anticoagulant drugs. Infections may occur as a result of immunoglobulin loss into the urine. Nephrotic syndrome is usually associated with raised cholesterol, which increases the risk of myocardial infarction (heart attack) due to atherosclerosis (fatty deposits on the artery walls). In the most severe cases, chronic kidney failure, and eventually an irreversible loss of kidney function, may develop.
The Nephrotic Syndrome management, complications and pathophysiology in detail
The nephrotic syndrome (NS), recognized as an entity for over half a century is defined by massive continued urinary protein losses resulting in hypoalbuminaemia and oedema formation. These are associated with modifications in kidney function, and with complications such as increased susceptibility to infections, thromboembolism, altered lipid and carbohydrate metabolism, and losses of binding proteins in the urine.
Oedema, the usual presenting complaint, is the consequence of abnormal accumulation of interstitial fluid, and becomes obvious if in excess of 10 per cent of body weight. This is usually associated with proteinuria greater than 40 mg/m2 BSA/h (>3.5 g/24 h in a 70 kg adult) leading to hypoalbuminaemia of less than 25 g/l. However in slowly developing nephrotic syndromes significant proteinuria and hypoalbuminaemia may persist for prolonged periods without evident oedema, while in children with minimal changes and NS in relapse, oedema may appear even before a low serum albumin appears.
Thus, there is only an approximate boundary between the NS and persisting proteinuria, with borderline patients. The older the patient, the lesser the degree of proteinuria and hypoalbuminaemia at which oedema may be observed.
Clinical features and investigation
Nephrotic oedema is noticeable first only around the eyes in the morning, and the ankles in the evening, but with increasing retention there is permanent swelling of ankles and face, with a sacral pad and oedematous elbows. The effects of gravity are less evident in children than adults, in whom oedema is generally dependent: young children with NS may suffer considerable ascites and facial oedema without ankle oedema. In adults retention of up to 4 l of salt and water remains undetectable, revealed only by weighing. With increasing oedema, ascites may appear followed by pleural effusions, which are usually bilateral, occasionally unilateral, and usually limpid, but sometimes opaque and chylous. Genital oedema may be distressing, especially in males.
The oedema remains soft and pitting even when profound, but if it remains untreated for long periods it may become indurated and pit only with difficulty especially around the ankles. Ankle swelling may be asymmetrical if deep venous thrombosis supervenes. Striae may appear even if no corticosteroids are being given, and the skin may actually split and weep spontaneously.
Needlestick punctures may also weep profusely. The liver is often painlessly enlarged, especially in children. The jugular venous pressure is usually normal or low, but if raised in association with a low or normal blood pressure in an older adult with NS this raises suspicion of cardiac amyloidosis as a cause. The nails may show white bands corresponding to periods of hypoalbuminaemia. If extreme hyperlipidaemia is present (see below) xanthomas may form periorbitally and elsewhere. In patients with progressive forms of nephritis underlying the syndrome, symptoms and signs of uraemia and hypertension may appear. The NS is often chronic, and the psychological effects of a condition which (together with its treatment) distorts the patient's body image must not be forgotten.
Prolonged hospital admissions and fear of renal failure take their toll of morale. In the medical history, the ingestion of medicines (prescribed or bought over the counte), prior acute or chronic infections, allergies, or any features suggestive of a systemic disorder such as lupus erythematosus should be noted. A history of macroscopic haematuria may be obtained, but this is unusual except in postinfectious or mesangiocapillary glomerulonephritis. The possibility of an associated tumour, usually in the lung or large bowel should be kept in mind in older patients. Finally, the family history may be revealing on occasion, as in Alport's syndrome or the Finnish form of congenital nephrotic syndrome. Investigation of patients with the nephrotic syndrome follows closely on that of those with persistent proteinuria and haematuria.
The majority opinion today is that all nephrotic adults and all nephrotic children less than 1 and over 10 years of age should have a renal biopsy, even if their urine shows no casts or red cells. If no unusual features are present and persistent microscopic haematuria and red cell casts are absent, treatment can be started with corticosteroids. The underlying renal biopsy appearances differ according to the age of the patient.
Renal imaging usually by ultrasound examination will be needed. A glomerular filtration rate (GFR) measurement using a single injection of radioactive tracer or contrast is desirable, in addition to the obligatory measurement of plasma creatinine; in some very oedematous patients it may be better to wait until oedema has been minimized because of problems with equilibration of the isotope.
Almost all patients show some reduction in GFR when the data are corrected for ideal weight for height, for reasons discussed below, and an initially reduced GFR is not necessarily an indicator of structural renal damage. The proteinuria must be quantitated on several 24 h urines as a baseline. Serological measurements should include a slide test for antinuclear factors, together with a specific assay for ds-DNA antibody, hepatitis B and C serology, complement concentrations; in patients over 45 years of age a chest radiograph and protein electrophoresis searching for paraproteins should be carried out.
A VDRL or equivalent investigation may be needed to exclude syphilis, remembering that low titres of antibody against treponemes are induced by previous yaws. ANCA tests will rarely be needed since patients with vasculitis very rarely develop a nephrotic state, unless they have recovered function after severe renal damage. A routine anti-GBM antibody is likewise unnecessary because of the rarity of this type of presentation, but should of course be done if the immunohistology suggests linear capillary wall IgG deposits.
Proteinuria and urinary findings
The only symptom of profuse proteinuria itself is urinary frothing, which some patients may notice, and which may provide a valuable clue as to when major proteinuria began. Findings on urine microscopy or testing for haematuria depend upon the underlying cause of the proteinuria, and range from a bland sediment and no or only intermittent red cells and only fatty casts in the case of minimal changes, to an angry sediment with abundant red cells, red cell and granular casts or even visible blood in the urine, in cases of severe mesangiocapillary or crescentic nephritis.
Urinary losses of albumin are the major factor in the development of hypoalbuminaemia, but the possible role of catabolism of reabsorbed albumin in the tubule remains controversial. Thus, severe hypoalbuminaemia may be seen in the presence of only moderate albuminuria. Turnover studies have shown that the fractional catabolic rate of albumin is increased, but the absolute catabolic rate is decreased (Kaysen and Al Bander 1990). Losses of protein into the intestine as part of a generalized increase in capillary permeability often have been proposed (Schulze et al. 1980) but not confirmed.
Pathogenesis of nephrotic oedema
The pathogenesis of oedema formation in nephrotics is still not entirely understood (Schrier and Fassett 1998; Palmer and Alpern 1997; Vande Walle and Donckerwolcke 2001). The two major systems involved in oedema formation are the capillaries and the kidneys.
The extracellular compartment of body water is made up of plasma water and interstitial fluid. These are in dynamic equilibrium and fluid movement across the capillaries is the result of the balance between filtration and reabsorption due to changes in capillary and tissue hydraulic and oncotic pressure, and changes in capillary permeability. The Starling equation describes the contribution of these factors to fluid movement:
(Jv) = Kf Ã— S Ã— EFP,
where EFP (effective filtration pressure) is calculated as (Pc - Pi) - (Ï- Ï ); Kf is the ultrafiltration coefficient, and S is capillary surface area.
Fluid movement is different along the length of the capillary vessel, related in turn to changes in pressure. Along the capillary the hydrostatic pressure (Pc) decreases, whilst the colloid osmotic pressure increases slightly as water is filtered into the interstitium. The positive EFP at the arterial side (Â±4 mmHg) results in a high filtration rate (= 60 l/24 h), while the negative EFP value (-2.8 mmHg) promotes fluid reabsorption at the venous side. Net fluid filtered (filtered minus reabsorbed fluid = 4-6 l/24 h) is removed by lymph flow back into the circulation. This is associated with a slightly negative hydrostatic pressure (Pi) in the interstitium (-2 mmHg) (Taylor 1981.
Although the capillary permeability to albumin is low, the interstitial concentration is about 40 per cent of that of plasma, resulting in a Ï€i of 12 mmHg (Fig. 4a). In (nephrotic patients, thi(s normal fluid balance is disrupted. When Ï€c decreases because of urinary losses of albumin, there will be increased loss of fluid into the interstitium from the capillaries, resulting in oedema. However, important changes in the interstitium will limit oedema formation. The accumulation of interstitial fluid increases Pi and when Pi reaches zero, interstitial tissue resistance increases and limits filtration.
The filtered fluid from the capillaries has a lower protein content, leading to dilution of the interstitial fluid and this will reduce interstitial oncotic pressure. Thereby the interstitial fluid oncotic pressure will decrease in parallel with plasma colloid osmotic pressure (COP). Also, increased interstitial hydrostatic pressure accelerates lymph flow. The return of interstitial fluid to the vascular space is enhanced and the washout of the interstitium leads to further reduction of the concentration of albumin in the interstitial fluid. Koomans et al. (1985) have shown that in nephrotics a decrease of plasma COP from 23 to 10 mmHg will decrease the transcapillary COP gradient by only 2â€“3 mmHg. Thus, a new balance is achieved consisting of a higher tissue pressure, increased lymph flow and a decrease in both interstitial fluid and plasma albumin concentration (Fig. 4b).
However, in some situations this equilibrium will not be achieved. This is the case in patients with extremely low plasma albumin concentrations, such as in children with a congenital nephrotic syndrome. In these patients, plasma COP is so low that even a decrease of interstitial COP to values near zero will not allow achievement of a new equilibrium. Also in minimal change nephrotic syndrome during a rapidly developing relapse, the decrease in plasma COP may be too fast to be matched by an appropriate decrease of interstitial COP, increased lymph flow and vascular refill. During this stage this disequilibrium will be associated with insufficient vascular refill and decreased circulating blood volume (Vande Walle and Donckerwolcke 2001).
The presence and contribution of changes in general capillary permeability coefficient (Kf) is still a matter of debate. However, under conditions of hypoalbuminaemia the permeability of the capillary to protein seems to decrease and this will also contribute to maintenance of the oncotic gradient between capillaries and interstitium (Fadnes 1975; Golden et al. 1990). Based on all these considerations, oedema formation in the nephrotic syndrome cannot be related only to reduced plasma oncotic pressure inducing hypovolaemia with secondary renal sodium retention, but a primary renal defect in sodium and water excretion must be involved.
All patients with NS show renal sodium and water retention, but this may show important differences according to the stage of the NS. At onset of the NS and during the phase of oedema formation a positive sodium balance is found, but after an equilibrium is reached renal sodium excretion again matches intake. However, this is not a stable situation, and sodium retention may occur occasionally in longstanding NS (Vande Walle et al. 1996).
The classical hypothesis related renal sodium and water retention to hypovolaemia (underfill theory). Loss of albumin by the kidneys is not fully compensated by increased hepatic synthesis, the plasma albumin concentration decreases, and the resulting reduction of COP increases fluid movements out of the vessels into the interstitium and thereby decreases plasma volume. The reduction of plasma volume will activate the sympathetic nervous system, and the reninâ€“angiotensin axis, promote AVP secretion and suppress atrial natriuretic peptide (ANP). Interaction of these systems stimulates renal sodium and water retention. Arguments have been provided that at least in some patients and during limited phases in the development of the nephrotic syndrome, renal sodium retention is stimulated by hypovolaemia (Usberti et al. 1984; Kumugai et al. 1985). In children with NS, clinical symptoms of hypovolaemia are occasionally noticed but in these patients, even when blood volume is assessed by plasma and red cell volume measurements, a decrease was seldom found (Vande Walle et al. 1995).
In adult patients with minimal change NS, an exaggerated fall in plasma volume has been observed when patients go from recumbent to standing position (Joles et al. 1993) suggesting the presence of hypovolaemia. However, several other studies failed to show a consistent reduction in blood and plasma volume in nephrotic patients (Dorhout Mees et al. 1979).
Elevated concentrations of plasma renin and aldosterone have been found in nephrotics, and in some patients sodium and water excretion increases by manoeuvres known to expand blood volume (such as intravenous albumin infusion and head-out water immersion), and is associated with a decrease in concentrations of vasoactive hormones (Rasher et al. 1986). Blockage of endogenous aldosterone by large doses of spironolactone resulted in enhanced natriuresis only in nephrotics and not in controls while on a high sodium intake (Shapiro et al. 1990). Both are arguments for involvement of increased RAAS activity, probably related to hypovolaemia. Other findings pointing to hypovolaemia are the non-osmotic related increased production of antidiuretic hormone in patients with the NS and increased plasma catecholamine concentrations and urinary clearances found in patients with NS (Usberti et al. 1984).
In nephrotic patients, increased proximal as well as distal tubular sodium reabsorption are present, and this also points to extrarenal stimulation of sodium retention (Grausz et al. 1972; Vande Walle et al. 1996). All these data are arguments in favour of water and sodium retention mediated by hypovolaemia. However, such abnormalities were only found in patients with rapid progression to the nephrotic syndrome and in others with extremely low serum albumin values (Vande Walle et al. 1996).
While all these observations are arguments in favour of a contracted plasma volume as the initiating factor for water and sodium retention, there is ample evidence that this is not the case in the majority of patients with the nephrotic syndrome. One study has shown that renal sodium retention already starts in the incipient nephrotic syndrome when plasma albumin is only slightly decreased (Vande Walle et al. 1995). In these patients, no difference in blood volume was found when measurements made during this phase and when in remission were compared. In other studies of patients with incipient remission of oedema, natriuresis was shown to precede an increase in plasma albumin (Brown et al. 1982; Koomans et al. 1987). Both observations are arguments against reduced plasma oncotic pressure being the primary mechanism for sodium retention. Several reports have indicated the lack of increased renin and aldosterone concentrations during the phase of oedema formation in NS, or found only modest increases in renal sodium excretion in response to blood volume expansion by head out water immersion or albumin infusion (Geers et al. 1984; Rabelink et al. 1993). While manoeuvres to achieve blood volume expansion in nephrotic patients resulted in appropriate elevations of plasma ANP concentrations, such increase in endogenous production of ANP was associated with a blunted natriuresis (Perico and Remuzzi 1993).
The most convincing evidence against hypovolaemia as the primary cause of sodium retention in the nephrotic syndrome are studies in experimental animals. In a rat model of unilateral proteinuric renal disease induced by puromycin amoninucleoside infusion, impaired urinary sodium excretion was limited to the proteinuric kidney despite the fact that both kidneys were affected by similar changes in blood volume and plasma COP (Chandra et al. 1981; Ichikawa et al. 1983).
Thus, these data from the literature and our own observations (Koomans et al. 1987; Palmer and Alpern 1997; Schrier and Fassett 1998) allow us to reformulate the sequence of events leading to renal water and sodium retention in NS. The basic abnormality seems to be a primary renal disturbance and exists in all cases of NS. Sodium retention occurs at an early stage when no significant decrease in serum albumin is noticed. Sodium retention continues through the stage of oedema formation. By sodium and water retention and redistribution of albumin stores most patients are able to maintain plasma COP above a critical minimum and conserve their blood volume. In these patients, sodium retention is not related to increased vasoactive hormones but the primary renal defect has yet to be identified.
Some patients, especially minimal change NS in early relapse, experience temporary hypovolaemia due to a disequilibrium in albumin redistribution. If progression to the nephrotic stage is too fast and compensatory mechanisms are insufficient, blood volume can only be maintained by secondary stimulation of vasoactive hormones maximizing renal sodium retention. The majority of these patients eventually achieve a new equilibrium, characterized by oedema, low plasma albumin concentrations and stable blood volume. A persistently unstable circulation is found only if the plasma COP decrease below a critical level (Â±8 mmHg). The relative roles of sodium retention due to hypovolaemia, and primary sodium retention by intrarenal defects, probably vary during the different stages in the development of the NS (Vande Walle and Donckerwolcke 2001).
Changes in kidney function
In nephrotics, conflicting observations on GFR have been made (Berg and Bohlin 1982; LÃwenborg and Berg 1999), although in the majority a decreased GFR was found (Geers et al. 1984; Shapiro et al. 1986). If GFR and renal plasma flow (RPF) were only volume dependent, they would be increased in hypervolaemic patients and during hypovolaemia the decrease of RPF would be more important than the decrease in GFR, resulting in increased filtration fraction (FF). In addition, a decrease in plasma oncotic pressure would also result in an increased GFR and RPF, but changes in GFR will be larger than changes in RPF and result also in increased FF. However an inverse correlation between FF and oncotic pressure was found in patients with the NS. Therefore, the decrease in FF commonly found in nephrotic patients implicates a decrease of the permeability of the glomerular basement membrane (Kf) (Berg and Bohlin 1982; Myers and Guasch 1993).
Different nephron segments are involved in sodium retention in the NS. One study attributed impaired sodium excretion to decreased GFR and showed correction of both sodium excretion and GFR by volume expansion (Shapiro et al. 1986).
Several tubular segments may be involved in renal sodium retention. In incipient nephrotic syndrome and during hypovolaemia increased tubular sodium retention was found at both proximal and distal tubules. In the distal tubule, at least part of increased sodium reabsorption is related to increased sodium/potassium exchange and is stimulated by aldosterone (Vande Walle and Donckerwolcke 2001). In patients with stable circulatory volume, sodium retention is not related to increased vasoactive hormones and seems to be localized in the most distal nephron segments. The molecular mechanism of renal sodium avidity has been attributed in PAN nephrotic rats to stimulation of Na, K-ATPase and increased expression of epithelial sodium channels of cortical collecting duct cells (Deschenes et al. 2002). Increased distal tubular sodium reabsorption has been attributed to resistance to ANP (Perico and Remuzzi 1993) and experience in nephrotic animals and patients are in accordance with this assumption. The abnormality was related to a specific cellular alteration in the ANP signalling pathway. Evidence has been provided that increased cGMP phosphodiesterase activity blunts the cellular actions of cyclic guanosine monophosphate, the second messenger of ANP normally produced in response to ANP's interaction with its biologically active receptor (Valentin et al. 1992).
A defect in urinary concentration capacity was found in humans with NS and similarly in adriamycin-induced NS in rats (Fernandez-Llama et al. 1998). The abnormality was associated with a decreased expression of Na+ transporters (Na“K2Cl, Na+/H+ exchanger NHE-3 and Na, K-ATPase, Î±1) in the thick ascending limb, and of renal medullary aquaporin-1 water channels in the thin descending limb of Henle's loop and of aquaporin 2 and 3 in the collecting duct. The concentrating defect was related to the failure to generate a high osmolality in the renal medulla and osmotic equilibrium in the collecting duct. The defect found in experimental studies, however, may follow a toxic effect of adriamycin (Apostol et al. 1997; Fernandez-Llama 1998). We and others observed a defect in urinary diluting capacity following a water load in patients with the NS (Vande Walle et al. 1996). This abnormality maybe related to decreased Na“K2Cl transport and thereby failure to generate a hypotonic luminal fluid by the cortical thick ascending limb of Henle's loop. However, we found the abnormality only in patients with hypovolaemia and suggest that decreased Na+ absorption in Henle's loop maybe related to low distal nephron sodium delivery. It is not clear how the reduced Na+ reabsorption in Henle's loop fits with the decreased sodium excretion in patients with NS.
Symptomatic treatment of nephrotic oedema
When oedema formation is severe, symptomatic treatment is required, independent of any specific measures that may be available to treat the underlying condition. The first step is to ensure a reasonably low intake of sodium compatible with a relatively normal diet, but in patients with hypovolaemia who require maximal sodium retention to maintain circulatory volume, particularly children, this policy can be dangerous. In patients with hypervolaemia, and in those requiring diuretics a low sodium intake (50-70 mmol/24 h in adults) should be given. An additional goal of sodium restriction is potentiation of the antiproteinuric effects of ACE inhibitors, discussed below.
Patients with the NS often show relative resistance to diuretic (Kirschner et al. 1992), which has been attributed to a multifactorial decreased delivery of the drug to the active sites in the tubular brush border of the kidney (Wilcox 2002). Diuretics are protein-bound, which limits the diuretic to the vascular space and maximizes its delivery to the kidney. Hypoalbuminaemia will result in a larger extravascular distribution and reduce its availability (Keller et al. 1982), as well as increasing its inactivation to the glucuronide within the kidney (Pichette et al. 1996). In addition, it has been shown that binding of free filtered frusemide to urinary proteins decreases the inhibitory effect on loop Cl- reabsorption (Kirschner et al. 1990); although Agarwal et al. (2000) found that displacement of frusemide from urinary albumin using sulfasoxazole made little difference to its diuretic activity. Experimental studies have shown that the active drug is less effective in inhibiting Cl reabsorption in the loop of Henle of nephrotic rats than in control rats (Keller et al. 1982). The tubuli of analbuminaemic rats are resistant to frusemide, but of practical importance this can be overcome by intravenous administration of the frusemide bound to albumin (Inoue et al. 1987). While the extent of absorption of frusemide in healthy individuals and nephrotics may be identical, a delayed delivery to the active sites may diminish the response to oral therapy. Other drugs may have a better bioavailability such as bumetanide (1-10 mg) or torsemide (10-50 mg), because of their hepatic rather than renal metabolism (Brater et al. 1984; KrÃ¤mer et al. 1999).
Frusemide is most often used in doses of 40 mg up to 200 mg/24 h, doses higher than used in other diseases. When the efficiency of oral diuretics is impaired, a continuous intravenous administration should be considered. Constant infusion is preferred to bolus administration, because it prevents a rebound in postdiuretic sodium reabsorption. This effect has been documented by several studies (KrÃ¤mer et al. 1999). Addition of other distal-acting diuretics may be necessary to obtain an effect, but in patients with low plasma aldosterone concentrations, the effect of added spironolactone is often disappointing unless given in very high dosages (600 mg/24 h), which often induce nausea. Additional metolazone 5-20 mg on alternate days has been used also, with success reported by some authors.
Intravenous albumin administration would seem at first sight to be the treatment of choice in patients with hypoalbuminaemia and hypovolaemia, but is rapidly excreted into the urine, and may lead to circulatory overload and hypertension in patients with hypervolaemia. Therefore the detection of hypovolaemia is of particular importance. Clinical symptoms alone are often unreliable for evaluation of blood volume and plasma albumin and oncotic pressure do not allow assessment of the blood volume status. Intrinsic renal diseases may be more often associated with sodium retention independent of volume status.
Thus it would be useful to have clinical measures of blood volume status to guide treatment. Concentrations of vasoactive hormones are not available for a quick diagnosis. We have shown that the association of low FENa+ and increased [UK+/(UK+ + UNa+)] ratio correlated best with clinical symptoms of hypovolaemia and significantly elevated levels of plasma renin, aldosterone, noradrenaline, and vasopressin. If [U(K+)/(UK+ + UNa+)] Ã— 100 is higher than 60 per cent, salt-free albumin infusion (1.5 ml/kg BW of a 20 per cent solution) can safely be administered (Donckerwolcke et al. 2003). The diameter and collapsibility of the inferior vena cava and the left atrial volume measured by ultrasonography have been advocated as an index of central filling in nephrotic children (DÃnmez et al. 2001) and deserve more exploration.
Intravenous salt-free albumin has been used also rather sparingly and controversially (Dorhout Mees 1996) by some authors in adults, but more commonly in children with severe hypovolaemia, which is uncommon in adult nephrotics. A dose of 100 ml of 20 per cent solution, combined with a high dose (usually 250-500 mg) of oral or 250 mg per 1.72 m2BSA intravenous frusemide was recommended in adults (Davison et al. 1974) and in children (Haws and Baum 1993). However, intravenous albumin alone has little effect in inducing a diuresis in patients with NS (Rabelink et al. 1993) or in hypoalbuminaemic cirrhotic patients (Chalasani et al. 2001). Direct comparison of diuretic alone versus a diuretic plus intravenous albumin showed no advantage for the latter regime in one study (Akcicek et al. 1995), but a modest benefit in another (Fliser et al. 1999). In cirrhotics (in whom the confounding variable of proteinuric binding of frusemide was absent) albumin alone or in combination did not enhance the diuresis (Chalasani et al. 2001). Each of these manoeuvres can be dangerous if employed alone, since IV albumin may induce severe or even fatal pulmonary oedema if a diuresis is not induced, and high-dose frusemide can induce severe volume depletion and hypotension. If employed, it is best to use the double regime on alternate days, to allow re-equilibration of interstitial and plasma albumin and saline. Despite the fact that the albumin passes quite quickly into the urine and the lack of evidence of an effect of albumin alone, this regime does seem to permit rapid removal of up to 35-40 l over 2 weeks in severely oedematous nephrotic adults.
Finally, gentle ultrafiltration using either a haemodialysis machine or (less commonly) peritoneal dialysis has been used in such resistant patients ever since the introduction of dialysis (Fauchald et al. 1985). This may lead to acute renal failure if pursued too vigorously, however, and perhaps albumin could be used in combination with ultrafiltration, but this has not been adequately tested.
Reduction of proteinuria
Apart from symptomatic relief of oedema, much evidence has accumulated (Burton and Harris 1996; Abbate et al. 1999; Jafar et al. 2001) that suggests proteinuria is directly toxic to the renal tubules and leads eventually to an interstitial infiltrate and renal fibrosis. Additionally, reduction in proteinuria leads to improvement in plasma albumin concentrations and consequent improvement in lipid and coagulation parameters (see below). Thus, reduction of proteinuria has become a major goal in the management of patients with persisting proteinuria from chronic conditions leading to renal failure.
Since the first descriptions (Michielsen et al. 1973; Arisz et al. 1976), a number of manoeuvres (Table 1) have been outlined that can reduce nephrotic-range proteinuria arising from any type of histological change, when specific treatment to reduce or eliminate proteinuria is not available. All of these seem to act through reversible haemodynamic changes, and are accompanied by a greater or smaller reduction in the GFR, important in patients with either structural glomerular damage and/or reversible reduction in GFR. This therapeutic ratio is more favourable using ACE inhibitors than with any other agent except perhaps dipyridamole. Whilst any reduction in the mean arterial blood pressure will reduce proteinuria, the effect of ACE inhibitors is present even when the reduction in proteinuria is corrected for that achieved by lowering the blood pressure to an equivalent extent using other agents (Gansevoort et al. 1995a).
The effect of these manoeuvres in reducing proteinuria is additive (Table 1) and all may have a final common action in reducing intraglomerular capillary hydraulic pressure, Pc (de Jong et al. 1992). On the one hand, non-steroidal anti-inflammatory drugs (NSAIDs), protein restriction, and cyclosporin most likely exert their antiproteinuric effect by increasing afferent glomerular arteriolar tone; whilst ACE inhibitors, AT1 antagonists (Kurokawa 1999) and possibly dipyridamole, act through decreasing efferent glomerular arteriolar tone (de Jong et al. 1992). The combination of ACE inhibitor and AT1 antagonist is additive (Russo et al. 2001). Improvements in glomerular permeability seiving profiles during treatment with indomethacin (Golbetz et al. 1989) or ACE inhibitors (de Zeeuw et al. 1990) are almost certainly the result of these haemodynamic changes.
Table: Inhibition of proteinuria in (non-diabetic) nephrotic patients by different agents
||Arisz et al. (1976)
|Low protein diet
||Kaysen et al. (1991)
||Tejani et al. (1987)
||Gansevoort et al. (1995a)
|AT1 receptor antagonists
||Kan et al. (1974), de Jong et al. (1988)
|Indomethcin + lisinopril
||Heeg et al. (1991)
|Enalapril + low protein diet
||Ruilope et al. (1992), Gansevoort et al. (1995b)
|Captopril + ibuprofen
||Allon et al. (1990)
|ACE inhibitor + AT1 antagonist
||Russo et al. (2001)b, Laverman et al. 2002
|a Mainly lisinopril and elalapril have been studied, but captopril, fosinopril (Keilani et al. 1993), ramipril, and benazapril have similar effects.
|b Note that patients in this study all had IgA nephropathy, with a mean proteinuria of only 1.9 g/24 h.
A number of controlled trials of the renoprotective effects of ACE inhibitors in proteinuric patients with various underlying renal conditions have been performed (see Gansevoort et al. 1995a and Chapters 11.1 and 11.2), which suggest that the strategy of using a long-acting ACE inhibitor such as enalapril, ramipril, or lisinopril to reduce protein excretion in long-term nephrotic patients should form a part of the long-term management of all such patients today. AT1 receptor antagonists also are under trial, as are combinations of both. It is important to restrict sodium at the same time to 50-60 mmol/24 h, since the inhibitory effect on proteinuria is much greater in sodium-depleted than sodium-replete patients (Heeg et al. 1989), and the effect can be completely abrogated by a sodium intake of 200 mmol/24 h of sodium. Also, the maximum effect will not be achieved for several weeks with ACE inhibitors, in contrast to NSAIDs, with which a reduction in proteinuria is evident within a week or two. Lisinopril 5 mg is usually effective in reducing proteinuria by one third, and today is probably the recommended treatment; 10 mg/24 h will halve the proteinuria, but with a 25 per cent reduction in GFR. Addition of a small dose and an NSAID into the regime can be considered in both adults (Allon et al. 1990; Heeg et al. 1991) and children (Trachtman and Gauthier 1988), but the further reduction in GFR will carry with it a greater risk of acute renal failure.
Nephrectomy to eliminate proteinuria
In a few unfortunate individuals, proteinuria remains torrential despite aggressive treatment, postural hypotension is prominent, oedema persists, the plasma creatinine increases, and protein malnutrition becomes increasingly severe. These are usually young and show severe ("malignant") focal and segmental glomerulosclerosis; if elderly, amyloidosis. In this situation, it may be useful to contemplate nephrectomy, dialysis, and intensive nutrition, an approach now standard in the Finnish congenital NS before the infants have become too malnourished (Kim et al. 1992). Unilateral nephrectomy has been used also in neonatal nephrotics (Mattoo et al. 1992) and also in focal segmental glomerulosclerosis, with decrease in proteinuria and an increase in albumin in some cases, but no change in others.
Bilateral surgical nephrectomy through a midline incision is the standard procedure (Bienz et al. 1994), but other means have been employed to achieve a "medical" nephrectomy. These include the administration of mercurial diuretics or NSAIDs in very high doses (Bamelou and Legrain 1982), cyclosporin and angiotensin (Rieu et al. 1994), or the injection of polymers, autologous thrombus (Olivero et al. 1993), coils, gelfoam, or fat into the renal artery. Both approaches have advantages and disadvantages; we have used only surgery and NSAIDs under these circumstances.
Complications of the nephrotic syndrome
The NS, mainly through alterations in concentration of plasma proteins, affects every cell and every tissue in the body. There is selective loss of low molecular weight protein in the urine, together with a general overproduction of all hepatically synthesized proteins. Thus, low molecular weight proteins tend to be depleted in the plasma, with accumulation of the higher molecular weight species. Proteins smaller than 180-200 kDa show lower plasma concentrations, and above this size increased levels. Proteins with fast turnovers in health are more resistant to depletion than those with slower normal synthetic rates.
The abnormalities found in NS result from the effects of these alterations in the protein environment, directly as a result of altered concentrations (e.g. in coagulation) or as a secondary result of induced alterations in cellular function (e.g. binding of increased LDL to platelets or endothelium).
Infections in nephrotic patients
Incidence and clinical features
Although no longer a major problem in developed countries, infections remain a formidable challenge in nephrotics throughout the developing world (Elidrissy 1982; Gulati et al. 1997). Early papers from developed countries described a similar situation (Arneil 1961) but such reports are rare today (Rusthoven and Kabins 1978). However, sepsis remains an important cause of deaths in children with NS even today: 6/10 deaths in 389 children with minimal change NS were from sepsis (International Study of Kidney Disease in Children 1984) and Gorensek et al. (1988) also showed that peritonitis remains a threat.
Although minimal changes underlies the NS in most cases, this may reflect only the fact that it is the most common form in childhood; we have seen pneumococcal peritonitis also in children with mesangio-capillary glomerulonephritis and in Henoch-Schonlein purpura. Most organisms implicated are encapsulated, and the predominance of pneumococcal infections is striking; others, such as staphylococcus, are notable for their almost complete absence. The clinical infections rate is higher in children treated with cytotoxic drugs than in those treated with prednisone alone, being highest in those treated with chlorambucil (6.5 per cent) compared with cyclophosphamide (1.5 per cent) (Latta et al. 2001).
This is a particular feature of children with NS. the oldest case reported by Professor J. Cameron at Guy's Hospital, London was aged 21, but even older patients have been noted (Chuang et al. 1999).
Onset may be insidious but is usually sudden, and should be suspected in any nephrotic child who develops abdominal pain. Unfortunately, this is a common symptom associated with hypovolaemia, and the diagnosis must be confirmed by direct microscopic examination of a Gram stain or an immunochemical search for bacterial antigens in ascitic fluid removed by needle. Blood cultures are usually positive also, but of course take much longer to perform. Hypotension, shock, and even acute renal failure (Cavagnaro and Lagomarsino 2000) may follow rapidly, sometimes with disseminated intravascular coagulation (see below).
In the past, the organism was almost always Streptococcus pneumoniae, but other organisms have become relatively more common; these include -haemolytic streptococci, Haemophilus and Gram-negative bacteria (Speck et al. 1974; Tain et al. 1999). Even so, S. pneumoniae remains the most important organism: half in Krensky et al.'s (1982) series, with E. coli in 25 per cent and Gorensek et al. (1988) noted 11/37 cases. S. pneumoniae peritonitis also may be more common in Black than White children (O'Regan et al. 1980). Although now treatable, primary pneumococcal peritonitis is still a major illness, and remains an important cause of death in nephrotics in both the developing (Elidrissy 1982) and developed worlds (Gorensek et al. 1988). Penicillin-resistant strains of S. pneumoniae are now emerging (Giebink 2001).
This may arise from skin lesions, either spontaneous or as a result of venepuncture, and are usually the result of Î²-haemolytic streptococci or a variety of Gram-negative bacteria. Usually the clinical diagnosis is clear with obvious demarcation of the infected area; the patient may be toxic, febrile, or even become hypotensive. Other patients run a more indolent course with an area of infection, which remains localized. It is difficult to stain or culture organisms from fluid aspirated from the area, but as in primary peritonitis, blood cultures are usually positive.
Viral infections Relapses of the minimal change NS often follow not only viral upper respiratory infections, but also viral infections, especially measles, which may induce remission of the minimal change NS, and the disease was even used as a treatment for the condition before the advent of corticosteroids (Janeway et al. 1948). The major threat is either varicella (see below) or measles, especially to children receiving either corticosteroid or cytotoxic agents.
Other organisms Particularly in the developing world, infections (including peritonitis) may arise from a variety of other organisms, such as Bacterium alkaligenes, Bacteroides, Aerobacter, and Streptococcus viridans (Choudhry and Gai 1977). Tuberculosis, common in developing areas, affected almost 10 per cent of all nephrotic children in one large series from India (Gulati et al. 1997).
Pathogenesis of infections in nephrotics
Although physical factors such as fluid collections in cavities, ruptured fragile skin, and dilution of local humoral defences by oedema are important, it seems likely that immunological factors play the major role in the susceptibility of nephrotic patients to infection.
Low serum IgG concentrations are characteristic of nephrotic patients (Giangiacomo et al. 1975) for unclear reasons. However, low IgG does not seem to play an important role in infections in NS, since in patients with inherited common hypogammaglobulinaemia concentrations less than 2 g/l are needed to provoke serious infections, a level rarely seen in nephrotics. Also in deficiencies of IgG, infections tend to be staphylococcal, chronic, and sinopulmonary in site.
Factor B is crucial in the primitive, antigen non-specific alternative pathway of the complement system. Its molecular weight is only 55 kDa and thus is lost in the urine in the NS with lowered plasma concentrations (McLean et al. 1977; Anderson et al. 1979). The alternative pathway is of particular significance to the opsonization of encapsulated organisms, such as S. pneumoniae, and whilst the addition of IgG had no effect on opsonization by nephrotic sera, addition of purified factor B resulted in improvement. Matsell and Wyatt (1993) have pointed also to the importance of urinary losses of another alternative pathway component, factor I. By 20 years of age, most individuals have acquired specific antibodies against a variety of pneumococcal capsular antigens. Thus, it is only during childhood that there is this peculiar vulnerability to S. pneumoniae.
Transferrin is essential for normal lymphocyte function, and acts as a carrier for a number of metals, including zinc. Warshaw et al. (1984) suggested that, in vitro, added transferrin could restore defective function in lymphocytes from nephrotics to normal, and Bensman et al. (1984) suggest that urinary losses of zinc result in diminished production of the zinc-dependent thymic hormone thymulin.
There is abundantâ€”but confusingâ€”evidence that lymphocyte function is depressed in nephrotic patients, especially those with minimal change disease. This depression of T-cell function appears to arise from serum factor(s) which may persist in remission and which may or may not be specific for minimal change disease, independent of treatment. There are fewer data on polymorph function, but Yetgkin et al. (1980) and Sillix et al. (1983) demonstrated impaired phagocytosis of Staphylococcus aureus and E. coli by polymorphonuclear leucocytes from nephrotics, even when the assays were performed in normal serum.
Treatment and prophylaxis
Obviously, prompt induction of remission of oedema or proteinuria are the most important goals and the decline in death rate from infection in nephrotic children is probably the result of this. Meticulous care must be taken over asepsis concerning breaches in skin.
In children who have already experienced pneumococcal infection, there is a good case for using prophylactic penicillin at least whilst the child is oedematous since recurrences are well known (Anderson et al. 1979; Moore et al. 1980). For some time pneumococcal vaccines against capsular antigens have been available, and their use in nephrotic children seems logical (Tejani et al. 1985) as the response when given in remission is sufficient. During relapse, however, when treated with high doses of prednisone (~20 mg/kg/day) or with cytotoxic therapy, vaccination may not give protection. We recommend delaying vaccination up to 14 days after discontinuation of these treatments. However, recurrent pneumococcal sepsis has been reported in children immunized after a first attack (Primack et al. 1979; Moore et al. 1980). Persistence of the antibody is low in relapsing cases (Spika et al. 1986), and protection should not be trusted too far (GÃ¼ven et al. 2004). The emergence of non-vaccine serotypes in immunized individuals, and antibiotic resistance, are increasing problems (Giebink 2001).
The most important feature of treatment for established sepsis in a nephrotic patient is that it should begin quickly. This, in turn, rests with anticipation, suspicion, and rapid diagnosis. The ESR is useless in nephrotic children, the white cell count may be misleading in those taking corticosteroids or immunosuppressants and the C-reactive protein is not always available, or too late to be of use.
Parenteral antibiotics should always be used, even when the infection appears to be localized, because septicaemia is almost always present. Antibiotics should be begun as soon as cultures have been taken; results of sensitivity tests should not be awaited. In children, benzylpenicillin should always be a component of the antibiotic therapy; conversely when pneumococcus is observed, backup broad-spectrum therapy is needed, because it may not be the only organism present. A broad-spectrum cephalosporin, perhaps with an aminoglycoside, may be used as initial blind treatment in adults. Blood pressure, peripheral temperature, and venous pressure should be watched carefully, colloid given as necessary and heparin given for anticoagulation because of the danger of secondary thrombosis (Futrakul 19785). Finally, the need for supplementary corticosteroids in those taking these drugs, or in those who have just stopped them, should be remembered. Any severe infection should prompt discontinuation of cytotoxic therapy.
In nephrotic children with active disease, varicella/chicken pox is a threat. Those taking high-dose corticosteroids or other immunosuppressive agents within the previous 3 months are at risk of severe progressive disseminated disease. Following contact with infected persons, zoster immune globulin should be administered within 72 h. Active varicella zoster infection should be treated with intravenous acyclovir, and cyclophosphamide or chlorambucil should be discontinued. In several countries, an effective varicella vaccine is available. Low to moderate doses of prednisone do not contraindicate administration of the live virus vaccine. In these patients, seroconversion is similar to healthy children, but relapses of the NS following vaccination have been noticed (Alpay et al. 2002).
Thrombosis in both the arterial and the venous circulations is a relatively frequent and serious complication of NS (Cameron 1984; Cameron et al. 1988) being associated with clinical symptoms in about 10 per cent of nephrotic adults and 2-3 per cent of children (Andrew and Brooker 1996; Lilova et al. 2000). Subclinical thrombotic complications hve been reported to occur in up to 50 per cent of adults and 28 per cent of children with NS (Hoyer et al. 1986).
Abnormalities of coagulation
It remains unclear what causes the increased tendency to thromboembolism in nephrotic patients. No single protein factor appears overwhelmingly important, although most changes in NS might be expected to lead to enhanced coagulability (Table 2) (see Cameron 1984; Cameron et al. 1988; Kanfer 1990; Rabelink et al. 1994 for details).
Table: Concentrations of proteins important in coagulation in the nephrotic syndrome in relation to molecular weight
||Concentration in nephrotic plasma
|Zymogens and cofactors
|Von Willebrand factor
|Factor I (fibrinogen)
|Factor II (prothrombin)
Nephrotic patients are relatively immobile and may have haemoconcentration from hypovolaemia, especially in childhood. Whole blood viscosity is increased by both the increase in haematocrit with changes in red cell deformability, and increased plasma viscosity arising from high fibrinogen concentrations (Zwaginga et al. 1994). These conditions also favour margination of platelets and increased platelet aggregability (Rabelink et al. 1994).
Alterations in zymogens and cofactors
Numerous studies of adults (Cameron et al. 1988) and children (Kanfer 1990; Elidrissy et al. 1991; Andrew and Brooker 1996) survey concentrations and/or activities of procoagulant zymogens and their cofactors in NS. In general, plasma concentrations of factors of the intrinsic pathway of coagulation are reduced, and occasional patients with a bleeding tendency from this have been reported (Handley and Lawrence 1967); but all other zymogens or cofactors are either present in normal or more usually raised concentrations, particularly von Willebrand factor, fibrinogen, and factors V, X, and VII. Concentrations of prothrombin (factor II) are in general normal. Whether increases in zymogen concentrations, over and above the great excess in which all are present in normal circumstances will induce a "prethrombotic" state is not clear; none is rate-limiting at the concentrations found in normal plasma. However, it is worth noting that in normal individuals, fibrinogen and factor VIIc concentrations are independent variables predicting vascular disease. Zwaginga et al. (1994) and Rabelink et al. (1994) point to a crucial role of raised fibrinogen concentrations, particularly under flow conditions.
Alterations in fibrinolytic and regulator proteins
Some of this group of proteins are concerned both with inhibiting fibrinolysis (which limits thrombosis) but also thrombin generation (which promotes thrombosis), complicating prediction of likely effects. Also, activity of the several modulators of plasmin and thrombin summate; concentrations of some may go down and some up. Thus, even though concentrations of antithrombin III are very low in NS (10-20 per cent normal) (Kauffmann et al. 1978), total functional antithrombin activity is normal (Kanfer 1990) because of the increase in Î±2-macroglobulin concentration (Boneu et al. 1981). Rydzewski et al. (1986) also noted low concentrations of Î±1-antitrypsin, but felt that the increases in Î±2-macroglobulin did not compensate for losses of antithrombin III. Consistently, plasminogen concentrations havebeen found to be low in nephrotics, but the role of plasmin inhibitors (Du et al. 1985) is not yet clear.
The vitamin K-dependent natural anticoagulant protein C, and its cofactor protein S (Clouse and Comp 1986; Preissner 1990), are activated by the thrombin-thrombomodulin complex and inherited deficiencies are associated with venous thromboses. They antagonize the action of activated factors V and VIII and stimulate release of plasminogen activator from endothelial cells. Protein S is in addition responsible for binding of protein C to platelets and endothelial cells, where anticoagulant activity is expressed. Protein S exists in the plasma in a free active form, and also in an inactive form bound to C4b-binding protein. Despite a molecular weight of 62 kDa, which leads to the expected urinary losses (Cosio et al. 1985), concentrations of protein C remain normal or even increased in nephrotic plasma (Rostoker et al. 1987; Gouault-Heilmann et al. 1988); while there are conflicting reports concerning the functional activity of the molecule (Vaziri et al. 1988). Only the free protein S is active, but there is disagreement about even total concentrations of protein S in nephrotic plasma (Gouault-Heilman et al. 1988; Vaziri et al. 1988; Citak et al. 2000), although functional levels are usually low, which might play a part in nephrotic hypercoagulability (Vigano-Angelo et al. 1987).
Factor V, previously thought of as a purely procoagulant zymogen, is also a cofactor of protein C (Svensson and Dahlbock 1994). Mutations of factor V that permit its procoagulant action but eliminate its anticoagulant cofactor function have been described in as many as 30-40 per cent of otherwise normal patients who suffer deep vein thromboses (Factor VLeiden) (Voorberg et al. 1994). However, study of small number of nephrotic patients with thrombosis did not suggest an increased frequency of potentially thrombogenic factor V mutations (Irish 1997).
Thus, the immensely complex balance of coagulant, fibrinolytic, and regulator proteins is much disturbed by the dysproteinaemia of the nephrotic state. It seems unlikely that the increases in zymogens and other promoters of coagulation will enhance coagulation in view of the excess normally present in plasma, except perhaps fibrinogen. Urinary losses of regulator proteins may be important; there is no convincing evidence that total antithrombin activity is reduced in nephrotics when all the changes are summed up. Fibrinolysis has been less studied in nephrotics, and it may be here that the effects of dysproteinaemia, if any, are to be found. Conventional tests of coagulation such as the prothrombin time are usually normal in NS (Andrassy et al. 1980). The partial thromboplastin time may be prolonged in the few patients with low intrinsic pathway activities (factors XII, XI, and IX). However, it is safe to biopsy these patients without fresh frozen plasma, as it is in those with a lupus "anticoagulant" (antiphospholipid antibody) (Kant et al. 1981).
Alterations in platelet function
Three possible mechanisms by which platelet hyperaggregability may arise in NS have been explored (Cameron 1991): effects of lowered serum albumin on platelet prostaglandin metabolism, effects of hyperlipidaemia on platelet membrane lipids, and effects of increased von Willebrand factor concentrations.
The platelet count in nephrotic patients has been reported as either normal or mildly raised (Cameron 1984). Abnormalities of platelet aggregation and thromboxane production in platelets from nephrotic patients studied ex vivo are well recognized. To what extent these data reflect events in vivo is not clear, but the data of Zwaginga et al. (1994), who studied whole nephrotic blood under flow conditions, throw doubt on their relevance. In vitro, hyperaggregability of stirred nephrotic platelet-rich plasma to ADP stimulation, with occasional spontaneous aggregation, is seen. This can be reversed by addition of concentrated urine protein (Bang et al. 1973) or of purified albumin both in vitro and in vivo by infusion (Remuzzi et al. 1979). This is unlikely to depend upon increased availability of arachidonic acid for thromboxane synthesis as proposed by Jackson et al. (1982) and Schiepatti et al. (1984), at least in vivo.
Hyperlipidaemia is a prominent feature of the NS. Lipoproteins bind to platelets, LDL-enhancing and HDL-inhibiting aggregation; these effects may be mediated through effects on adenylate cyclase. Hyperlipidaemia is associated in NS with spontaneous aggregation (Jackson et al. 1982) and dietary lipids will alter the composition of platelet membrane lipid. There are also opposing effects of LDL and HDL on endothelial cell production of prostacyclin (see below), which is a powerful modulator of platelet activity. LDL from nephrotic plasma is enriched in lysolecithin (Joles et al. 1994), and lysolecithin-containing LDL are toxic to endothelial cells, diminishing nitric oxide production (Kugiyama et al. 1990), which is an inhibitor of platelet adhesion under flow conditions.
Hyperaggregability to ristocetin (a cationic compound) is present in in vitro inplatelets from nephrotic adults, which appeared to be independent of the depletion of albumin, or the availability of arachi-donate but is probably related to the greatly increased concentrations of von Willebrand factor in nephrotic plasma (Bennett and Cameron 1987), to which ristocetin binds in a triple complex with the platelet. Suggestions of altered charge on platelets from nephrotic patients based on indirect methodology (Levin et al. 1985) have not been sustained by direct measurement (Cohen et al. 1988; BÃhler et al. 1992).
Alterations in endothelial cell function
Endothelial cells are known to play a very active role in maintaining the free flow of blood in vivo. Their function has been studied little in nephrotic patients, but Stroes et al. (1995) showed impaired response to endothelium-dependent vasodilatation, possibly as the result of the alterations in circulating lipids (Kugiyama et al. 1990; Joles et al. 1994). Alterations in endothelial function could be important in nephrotic hypercoagulability.
Role of drugs
Corticosteroids The description of thromboses in nephrotic patients coincided with the introduction of corticosteroids in their treatment (Calcagno and Rubin 1961); in Egli's (1974) paediatric survey, 26 of 59 patients who developed thrombosis were taking steroids at the time. Corticosteroids increase the concentration of some zymogens (Ozsosylu et al. 1962), and prothrombin times and APTT may be shortened under steroid treatment (Ueda et al. 1987). The action of warfarin is also antagonized by corticosteroids (Menczel and Dreyfuss 1960). On the positive side, steroids tend to increase concentrations of antithrombin III (Thaler and Lechner 1978) and inhibit platelet aggregation (Glass et al. 1981), at least in large doses.
Most patients who develop thrombosis are receiving diuretics, since almost all are oedematous; some develop thrombosis during diuresis (Lilova et al. 2000) possibly from the increased haematocrit induced by diuretics. The already increased blood viscosity in nephrotics as a result of high fibrinogen concentrations increases steeply with increasing haematocrit. It may be that the more judicious use of diuretics during the 1970s and 1980s has led to a decrease in the frequency of thromboses, for example, of the pulmonary artery in children (Egli et al. 1973; Egli 1974), which we no longer observe. Another factor may be the common practice of giving albumin infusions at the same time as powerful diuretics, especially when the latter are given intravenously.
Peripheral venous thrombosis and pulmonary embolism
Deep vein thrombosis
Thrombosis of the deep calf veins is common in NS, overt in 3-12 per cent of adults (Llach 1982; Cameron 1984), whilst occult thrombi may be detectable in as many as 25 per cent by Doppler ultrasonography (Andrassy et al. 1980). It is not clear how common deep vein thrombosis is in childhood; less than 1 per cent of the 4158 paediatric patients identified through a MEDLINE search by Andrew and Brooker (1996) showed clinically-evident deep vein thrombosis, and Mehls et al. (1987) reported only two cases amongst 204 nephrotic children. However, the study of Hoyer et al. (1986) discussed below on pulmonary emboli in children with the nephrotic syndrome, suggests that thrombi must be more common than this at a subclinical level, and 8/16 thrombotic episodes in the series of NS patients in Bulgaria (Lilova et al. 2000) were in the leg.
Not surprisingly in view of these data, pulmonary emboli are also common: at a clinical level about 0-15 (median 7) per cent of adult nephrotics have embolism (Cameron et al. 1988), and in our series of those with minimal changes, 8 per cent. However, if ventilation-perfusion (V/Q) isotope scanning is done routinely, 9-26 (median 12) per cent of adult nephrotics show evidence of pulmonary emboli, often symptomless (Cameron 1984). Nevertheless only a single patient in our own adult series up to 1982, followed for a total of 2100 patient-years, had actually died from pulmonary embolism. As for DVT, children seem to be less affected: Egli's data (1974) discovered only a single child of 3377 with clinical embolism. But, in contrast, 2 of 204 cases of Mehls et al. (1987) had pulmonary emboli, and Jones and HÃebert (1991) reported a major pulmonary vein embolus in a child, and Lilova et al. (2000) another. As in adults, a systematic study of pulmonary V/Q scans in 26 nephrotic children (Hoyer et al. 1986) showed evidence of emboli in a much higher proportion (35 per cent).
Other venous thrombi
Other venous thrombi are much less common, with the exception of renal vein thrombosis, discussed below. However subclavian or axillary, jugular, iliac, portal, splenic, hepatic, and mesenteric vein thrombosis have all been described (see Cameron 1984 and Cameron et al. 1988 for references). Sagittal sinus thrombosis has been described both in children and adults, and we have seen a fatal case in which the thrombus extended to include superficial cortical veins. We have also seen a nephrotic child with priapism.
In all series reported in adults, arterial thrombosis is much less common than venous thrombosis (Cameron 1984). Also, in children the proportion of thromboses at a venous site (73-81 per cent) was greater than in arteries (19-27 per cent) (Andrew and Brooker 1996; Lilova et al. 2000). Thrombosis of almost every artery has been recorded, and is summarized in detail in Cameron (1984) and Cameron et al. (1988). Multiple arterial thrombi in the same patient have been recorded also (Lye and Tan 1991) as well as both venous and arterial thrombi in the renal vessels (Kennedy et al. 1991).
One of the most common sites is the femoral artery, found usually in children (10 cases in Egli's survey of 1973) often in association with attempts at femoral vein puncture in hypovolaemic subjects, and also in adults (Kanfer et al. 1970). A femoral artery thrombosis was the only instance amongst our 90 nephrotic adults with minimal change. Pulmonary artery thrombosis was common in the past, being the single most common site in the paediatric study of Egli (1974), and has been in an adult (Kanfer et al. 1970). Now, it seems almost to have disappeared. It is not clear how many of these cases may have in fact have originated as pulmonary emboli (Jones and HÃ©bert 1991). Haemoconcentration seems the likely factor, since the only other condition in which pulmonary artery thrombosis may be seen is polycythaemic congenital cyanotic heart disease.
Renal venous thrombosis in nephrotic patients
This controversial topic has been reviewed extensively in the past (Llach 1982, 1985). Controversy exists not only about its incidence, but also how energetically it should be sought, and how it should best be managed especially when symptomless. All agree that it is most commonly found in association with membranous nephropathy, usually idiopathic but also in secondary forms such as in lupus (Appel et al. 1976), after gold therapy (Nelson and Birchmore 1979), and after transplantation (LiaÃ±o et al. 1988), as well as in the Heymann model of membranous nephropathy in rats. The reason for this predilection for membranous nephropathy has so far eluded explanation.
The apparent prevalence of renal vein thrombosis varies according to whether a clinical or a venographic diagnosis in symptomless patients is considered. Also, since renal vein thrombosis and membranous nephropathy are associated, those series with a high proportion of patients with membranous nephropathy (as in several American series) will show a higher incidence of renal vein thrombosis. Rather consistent clinical prevalences from 4 to 8 per cent have been reported in membranous nephropathy (Trew et al. 1978; Noel et al. 1979; Cameron et al. 1988), whilst in several American series using venography, up to 50 per cent have been reported (Llach 1982; Wagoner et al. 1983). In contrast, in Europe the prevalence is much lower, from 13 to 18 per cent (Andrassy et al. 1980; Cameron et al. 1988; Rostoker et al. 1992). In another American series, it was only 9 per cent (Pohl et al. 1984). In other forms of NS, the incidence of renal venous thrombosis is much lower, clinically 0-5 per cent (Andrassy et al. 1980; Llach 1982); of 322 such nephrotics in our own series only five (1.5 per cent) were affected (Cameron et al. 1988). Venography demonstrated subclinical thrombi, but only in 0â€“16 per cent of cases (Llach 1982).
Clinical diagnosis and evaluation
Renal venous thrombosis may present acutely with loin pain, haematuria, renal enlargement, pain and swelling in the ispilateral testicle in males, and deterioration in renal function; or as a slower decline in renal function without dramatic signs or symptoms. Leg oedema increases if the vena cava is involved, although caval thrombosis can be surprisingly silent. Thrombosis of renal veins may be unilateral, and men are more commonly affected than women. About 35 per cent of patients with renal venous thrombosis will have pulmonary emboli which are clinically evident or show up on scanning (see Cameron et al. 1988 for detailed discussion).
It has yet to be established that seeking symptomless renal venous thrombi is useful (Rostoker et al. 1992), since their prognosis appearsto be benign (see below), or how frequently or at what intervals rescreening must be undertaken. Nor is it clear what diagnostic strategy should be adopted: until recently, renal venography was the standard investigation, but being expensive and invasive was little employed for screening symptomless patients with NS, even those with membranous nephropathy. However, colour duplex Doppler ultrasonography (Avasthi et al. 1983) is easy and safe to perform, and although it has never been evaluated against MRI scanning (Tempany et al. 1992) or spiral CT, both of which have advocates, it probably remains the best diagnostic approach today. Surprisingly in view of the divergence of opinion in the past, no recent studies using Doppler duplex ultrasound have beeen published in nephrotic patients.
The prognosis of untreated renal venous thrombosis, whether clinically evident or silent, is no longer clear since virtually all patients now receive heparin and/or warfarin as soon as a diagnosis is made. In particular, the prognosis of radiographically diagnosed, symptomless renal venous thrombosis in the absence of anticoagulation is not known. Functional renal impairment is an adverse prognostic sign (Laville et al. 1988) but in many instances recovery of renal function is possible and recanalization of the veins usually occurs.
Treatment and prophylaxis of thrombosis in nephrotics
Thromboses in children, being more frequently arterial, will on average be more serious clinically than in adults.
Treatment of evident thrombosis
Patients should be mobilized, sepsis avoided or treated promptly, dehydration from incidental causes (e.g. diarrhoea) treated, diuretics used with care, and haemoconcentration minimized. Anticoagulation carries the usual risks and presents additional difficulties in nephrotic patients. Heparin acts mainly through activation of antithrombin III, whose concentration may be diminished in nephrotics (Kauffman et al. 1978). Thus, higher doses of heparin are required in NS to achieve anticoagulation, although there are probably additional explanations (Vermylen et al. 1987b) and Sie et al. (1988) describe normal plasma concentrations of heparin cofactor II (molecular weight 66 kDa) in nephrotics. Heparin binds also to Î±2-macroglobulin and to endothelial cell surfaces, as well as promoting platelet aggregation in some patients. Warfarin similarly presents problems in nephrotics (Ganeval et al. 1986), mainly because it is bound to albumin whose concentration may change therapeutically or spontaneously; however, it increases anti-thrombin III concentrations (Andrassy et al. 1980).
Whether nephrotic patients with symptomless deep venous thrombosis should receive anticoagulation treatment has never been studied adequately; probably if they also have symptomless pulmonary emboli, most physicians would anticoagulate. If the physician does wish to anticoagulate every nephrotic with a symptomless deep vein thromboses, then he or she will have to anticoagulate one quarter.
Patients with symptomatic thromboses should receive anticoagulation using warfarin with an INR maintained at 2â€“4, with similar treatment for children (Andrew and Brooker 1996). Special consideration needs to be given to nephrotics with renal venous thrombosis (Kanfer 1994), irrespective of how the diagnosis was reached. In the face of an incidence of overt or occult pulmonary emboli of 35 per cent, it is difficult to argue against anticoagulating all of them, although evidence of benefit from doing this is not available in symptomless patients.
There appears to be no advantage either to thrombectomy or local fibrinolytic therapy (Laville et al. 1988). In most cases in which warfarin has been used and subsequent venography performed, the renal vein had recanalized (Kassirer 1979; Llach 1982), but in some it did not (Laville et al. 1988). Five of 27 patients in Laville's series had bleeding complications from anticoagulation (see above).
Whenâ€”and on what groundsâ€”anticoagulation can be stopped also remains unclear. Data from the general population with deep vein thrombosis (Petiti et al. 1986) suggests a minimum of 3 months, with some extra benefit from 6 months. However, stopping the warfarin at any time in the continuing presence of NS may lead to rethrombosis (Briefel et al. 1978), and a reasonable policy is to continue warfarin until the oedema remits, or at least until the serum albumin is greater than 25 g/l. This may result in anticoagulation being required for some years.
Patients with diagnosed thrombi and an antiphospholipid antibody should have a higher level of anticoagulation, with an INR of 3â€“4 (Piette 1994), which needs to be maintained as long as the antiphospholipid antibody persists. At the moment there are no clear indications for prophylactic anticoagulation in patients with antiphospholipid antibodies; high titres of IgG antibody carry the greatest risk of subsequent thrombosis, but at the moment data do not justify treatment although others have expressed the opposite point of view (see next paragraph).
Prophylactic anticoagulation in nephrotic patients?
Given the very high incidence of thrombotic complications (10 per cent of adults and 3 per cent of children), should all nephrotics or at least all those with membranous nephropathy receive prophylactic anticoagulation? Bellomo and Atkins (1993) and Sarasin and Schifferli (1994) conclude that warfarin anticoagulation is justified in all patients with membranous nephropathy, and should lead to lowered morbidity. Sarasin and Schifferli (1994) argue further that anticoagulation for every nephrotic may be justified. The problem lies with the quality of the data fed into the decision models used in these analyses, and it will be interesting to see if these reviews have altered clinical practice. After all, patients at risk of thrombosis from familial antithrombin III, protein C, or protein S deficiencies are not recommended for prophylactic anticoagulation (Pabinger et al. 1994). The benefit of prophylactic anticoagulation with warfarin in children with NS remains even more unclear, alhough some recommend treatment in patients with unstable steroid-resistant NS.
Other possible approaches to prophylaxis are subcutaneous low molecular weight or synthetic heparins but there are no reports of this in nephrotics. Also, there is a case for giving at-risk subgroups (e.g. those with membranous nephropathy or antiphospholipid antibody) low-dose aspirin (75 mg daily, which is safe with steroids) and dipyridamole (100â€“200 mg three times a day), as suggested by Andrassy (1980). This has the advantage of being safe, and might also be justified on the grounds of the extra risk of vascular disease.
Lipid abnormalities in the nephrotic syndrome
Hyperlipidaemia is the most common complication of the NS, to the point where some authors have considered it a part of the definition of the syndrome: hypercholesterolaemia is present in 90 per cent of patients with a urinary protein excretion of over 3 g/24 h (Kasiske 1998). The importance of hyperlipidaemia lies in its contribution to the development of atherosclerosis, and possibly also to the progression or induction of renal damage leading to or aggravating chronic renal failure (Samuelson et al. 1997). Abnormalities occur in all aspects of lipid metabolism in nephrotic patients.
Nephrotic hyperlipidaemia (Kasiske 1998; Wheeler 2001) is characterized by an increased plasma concentration of cholesterol, both free and cholesterol esters (total plasma cholesterol > 5.2 mmol/l). The hyperlipidaemia correlates inversely and strongly with serum albumin concentration [although dyslipidaemia has been reported in patients with nephrotic-range proteinuria without hypoalbuminaemia, and in remission (Zilleruelo et al. 1984)]. It is independent, however, of the underlying cause of the NS including lupus (JovÃ©n et al. 1993). Increases in fasting triglyceride (TG) concentration are less common, and more often found in patients with severe NSs. The profile of lipoproteins, mixtures of apolipoprotein and lipids, is characterized by increases in TG-rich very low-density lipoprotein (VLDL), and in parallel of intermediate density lipoprotein (IDL) and of low-density lipoprotein (LDL) particles. Kashyap et al. (1980) showed reduced apolipoprotein C-II in VLDL from nephrotics, which may be important since Apo C II is an activator of lipoprotein lipase (see below). Deighan et al. (1998) demonstrated that LDL from nephrotics are rich in the smaller, denser, more atherogenic LDL III than normal. In contrast, high density lipoprotein (HDL) concentrations are usually normal, although an important reduction of the HDL-2 fraction is found (Short et al. 1986), and some studies have reported low total HDL concentrations in very severe NSs (Appel et al. 1985). The composition of the HDL is also altered from normal (JovÃ©n et al. 1987).
Lipoprotein (a) [Lp(a)] concentrations are also increased (Stenvinkel et al. 1993; Doucet et al. 2000) and it has been demonstrated that increased plasma concentrations of apolipoprotein (a) (>300 mg/l) are a strong independent predictor of adverse vascular events in the general population. Plasma concentrations in the normal population vary 100-fold, on the basis of genetic heterogeneity and molecular polymorphism.
Lipiduria is well recognized in the NS, and fatty urinary casts are a characteristic feature (see Chapter 1.2). Cholesterol, phospholipid, free fatty acids, and TG are all present in nephrotic urine. HDL is also present (Short et al. 1986) and Gitlin et al. (1958) showed that labelled HDL appeared in the urine when injected into nephrotics, whereas VLDL did not, presumably on the basis of molecular size.
Pathogenesis of the hyperlipidaemia of the nephrotic syndrome
This is complex and understood only in part. Lipoproteins transport TGs in plasma. They may be categorized according to apolipoprotein composition and density into high density (which contain apolipoprotein AI) and VLDL, IDL, and LDL (containing apolipoprotein B) (Vaziri 2003). The plasma concentration of the various lipoproteins is dependent on the balance between synthesis and catabolism. The liver synthesizes and secretes VLDL into the circulation. The determinants of VLDL catabolism are endothelial-bound lipoprotein lipase (LpL) which requires the availability of apoprotein C-II (apo CII). Lipoprotein lipase tethers triglyceride-rich lipoprotein to the vascular endothelium, where it is hydrolyzed. Normal VLDL metabolism is also dependent on HDL by recirculation of apo CII from HDL-3 to VLDL. Nascent HDL is formed by surface constituents of VLDL, free cholesterol and phospholipids and matures in the circulation. During maturation HDL particles accumulate cholesterol, which is esterified by lecithin cholesterol acyltransferase (LCAT) to form cholesterol esters. These sink into the core as nascent HDL is converted to HDL-3 and subsequently to HDL-2. The TG-depleted VLDL is taken up by the liver or interacts with cholesterol ester-rich HDL-2. HDL-2 is processed by lipase to HDL-3. The end-product of hydrolysis of VLDL, low density lipoprotein (LDL), is also directly synthesized by the liver.
Hyperlipidaemia in the NS is the result of both increased synthesis and decreased catabolism of lipoproteins (Wheeler 2001). Major abnormalities are related to impaired conversion of VLDL to LDL, increased hepatic LDL secretion, increased Lp(a) production by the liver, and impaired HDL maturation. The following mechanisms have been reported to be involved:
Decreased receptor-mediated clearance of circulating VLDL (Warwick et al. 1990, 1992; Kaysen and de Sain-van der Velden 1999) now seems crucially responsible for the increase in VLDL concentrations, although initially this was thought to be the result of increased synthesis. Both increased synthesis and reduced degradation of TGs has been found in nephrotics. VLDL catabolism also is impaired by a reduced endothelial-bound LpL pool. While synthesis of LpL is apparently normal, its attachment to the vascular endothelium may be reduced by deficiency of apoprotein C-II. Although total apoprotein C-II is normal in nephrotics, its concentration is relatively reduced related to the increased VLDL (Kashyap et al. 1980). Patients with NS have VLDL particles deficient in apo C-II (Deighan et al. 2000). Thus, proteinuria, through loss of ApoC-II in the urine, contributes directly to the defect in LpL activity, independently of any effect on plasma albumin (Shearer et al. 2001). However, in addition, the concentration of free fatty acids (normally albumin-bound) is increased in nephrotic plasma, which inhibits LpL (NikkilÃ¤ and PykÃ¤pistÃ¶ 1968) and nephrotic sera inhibit LpL in vitro (Vermylen et al. 1987a). Detailed studies in analbuminaemic compared with nephrotic rats have shown that both hypoalbuminaemia and proteinuria affect lipoprotein catabolism in the NS in different ways (Shearer et al. 2001).
Although total HDL concentrations may be normal in nephrotic patients, the distribution of the subclasses of this lipoprotein is abnormal, with a reduction of HDL-2 and an increase in HDL-3. This pattern has been attributed to the reduction in LCAT activity present in the NS (Cohen et al. 1980), which inhibits reverse cholesterol transport and thus impairs VLDL lipolysis by reducing the availability of apoprotein C-II. This loss of LCAT activity is probably multifactorial, since it is small enough to be lost in the urine, and also is potentiated by albumin which is reduced.
There is increased hepatic synthesis of all the lipid and protein moieties of VLDL and LDL which have been studied, in both human and in experimental NS (Kaysen 1991; Wheeler et al. 1991; Warwick and Packard 1993), an abnormality which in most cases returns to normal on remission. Probably an increase in apolipoprotein synthesis is the prime mover, and the accumulation of lipid a secondary event. Thus while conversion of VLDL to LDL is impaired, the principal pathway promoting the increase in LDL concentrations is through increased synthesis of LDL apoB100 by the liver (de Sain-van der Velden et al. 1998a); it is not clear yet what signal stimulates this (see below). The activity of cholesterol ester transfer protein (CETP), an enzyme that catalyzes the exchange of the cholesterol ester-rich core of HDL-2 for the TG-rich core of VLDL remnant particles, yielding LDL, is increased also in nephrotics, stimulating the normal pathway for LDL production (Kaysen 1992).
Elevated plasma Lp(a) concentrations result also from increased hepatic synthesis (De Sain-van der Velden et al. 1998a). Lp(a) consists of a molecule of LDL to which one molecule of the apolipoprotein apo(a) has been covalently attached to apoB100. Its molecular weight varies from 60 to 230 kDa according to the number of kringle repeats in the molecule, and both Lp(a) and its fragments are excreted in the urine in increased quantities in nephrotics (Doucet et al. 2000), the kidney also being an important site of metabolism of the molecule.
Chylomicra have been little studied in the NS (Levy et al. 1990; Warwick et al. 1992), but surprisingly in view of the data on impaired VLDL clearance, there seem to be no differences in the timing or height of chylomicron response to a fat meal in nephrotic subjects.
The primary stimulus responsible for triggering all these abnormalities of lipid metabolism in the NS has not been elucidated. The apparent coregulation of production of different proteins supports the hypothesis that reduced oncotic pressure or albumin concentration increases transcription of a group of liver-regulated proteins (Davis et al. 1980; Appel et al. 1985). Plasma viscosity has been suggested as the important factor, but this has also been denied (Appel et al. 1985). Infusion of albumin alone, or dextran, will acutely lower cholesterol and triglyceride in nephrotic patients or in nephrotic rats, but this effect may not operate in the long-term.
Consequences of hyperlipidaemia in nephrotic syndrome
Similar lipid abnormalities to those associated with NS have been shown to be major risk factors for atherosclerosis in the general population. However, the evidence to support the idea that this operates also within nephrotic patients, although plausible, remains scanty. The hyperlipidaemia of the NS was shown to increase the relative risk for myocardial infarction 5.5-fold and the relative risk of coronary death 2.8-fold compared to controls in a Californian population (OrdoÃ±ez et al. 1993). However, the only previous epidemiologic data, from the population in the South East of England (Wass et al. 1979) found no differences in incidence or outcomes compared to matched local controls. Numerous other small studies are inadequate and/or anecdotal, although the rapid appearance of aortic atheroma in some chronic nephrotics is striking, especially in children (Kallen et al. 1977; Hopp et al. 1994). The duration of the nephrotic hyperlipidaemia appears to be critical in inducing vascular damage, and it is patients with unremitting proteinuria and hypoalbuminaemia who are most at risk. Conversely, in remitting or minor NS vacular risk appears less. It must not be forgotten either that other risk factors, such as smoking, obesity, hypertension (Section 9), and hyperuricaemia must contribute strongly to vascular disease in nephrotic patients, as they do in the general population, and require management also.
The principal mechanisms of vascular damage appear to be associated with circulating LDL. Increased levels of LDL lead to increased scavenger-mediated LDL uptake by monocytes, inducing formation of foam cells within the vascular walls. Oxidized LDL is more rapidly taken up than normal LDL. Rupture of foam cells releases oxidized LDL and free radicals. Oxidized LDL causes apoptosis of endothelial cells, associated with endothelial dysfunction. The significance of the increased Lp(a) as a risk factor for atherosclerosis in the setting of NS is not yet known.
Elevated plasma TG concentrations are associated with a faster rate of deterioration of renal function in patients with glomerular disease (Hunsicker et al. 1997; Samuelson et al. 1997). Hyperlipidaemia may be associated also with the development of glomerular and interstitial renal disease in animal studies, but evidence remains persistently lacking of this in humans (Yang et al. 1999; Wheeler 2001). Endothelial damage may favour influx of lipoprotein into the mesangium, leading to proliferation and sclerosis. Also, filtered lipoproteins may accumulate in the tubular interstitium leading to a chronic inflammatory reaction. Interstitial foam cells are of macrophage origin, and accumulate in relation to duration but not intensity of the proteinuria (Neugarten and Schlondorff 1991).
Confounding factors such as proteinuria and hypertension made the assessment of any beneficial effect of treating hyperlipidaemia on progression of renal disease difficult in humans, although amply demonstrated in animal models (Harris et al. 1990). However, a recent meta-analysis of 13, generally small, controlled trials involving a total of 362 patients with diverse progressive renal diseases (but over two thirds diabetic), demonstrated that the rate of progression of renal disease to renal failure was slowed using various types of lipid-lowering therapy (statins, gemfibrozil, probucol). The rate of decline in GFR decreased by 0.16 (CI 0.03-0.27) ml/min/month [1.9 (CI 0.3-3.4) ml/min/year], comparable to the effect of reducing proteinuria using ACE inhibitors, and four times that achieved using protein restriction (Fried et al. 2001). The treatment of transient NS of less than some months' duration, however, does not at present seem justified (Keane et al. 1992; Olbricht and Koch 1992).
Treatment of hyperlipidaemia in the nephrotic syndrome
Although dietary fat restriction is usually recommended in hyperlipidaemic states, only severe fat restriction is associated with significant reduction in plasma lipids in NS. Dietary supplementation with fish oil has some lipid-lowering effect, but the effect on progression of renal disease has been found mainly in patients with IgA nephropathy (see Chapter 3.9). D'Amico and Gentile (1991) showed that a low-fat, low cholesterol diet in membranous nephropathy with NS, although it had no effect on proteinuria, reduced plasma total and LDL cholesterol concentrations by 24 and 27 per cent, respectively. Similar figures (28 and 32 per cent) were obtained using a vegetarian soy-protein, polyunsaturated fat diet; HDL cholesterol declined also, but by only 14 per cent.
Obviously, if major reduction of proteinuria or remission can be obtained, abnormalities of lipid metabolism will regress towards normal (Kaysen 1991). Gansevoort et al. (1994) reported that apolipoprotein(a) concentrations were reduced in nephrotic patients by giving lisinopril 10 mg/24 h together with indomethacin 150 mg/24 h, and Keilani et al. (1993) found that fosinopril 10â€“20 mg alone reduced proteinuria and also total cholesterol, LDL cholesterol, and Lp(a) concentrations. Several studies have documented the lipid-lowering effects of agents such as statins, bile acid sequestrants, fibric acids, fish oil, and probucol in patients with NS.
Statins (HMG CoA reductase inhibitors)
Lipid-lowering interventions, in order to decrease progression of renal disease should be directed towards lowering of TG-rich lipoproteins (Attman et al. 1999). While several drugs can achieve lowering of plasma lipids in nephrotics, HMG CoA reductase inhibitors (statins) have been shown to be the most effective agents (Massy et al. 1995). These drugs are inhibitors of the enzyme 3-hydroxy-3-methyl coenzyme A (HMG CoA), reducing cholesterol synthesis. They lead also to upregulation of hepatic LDL receptors, and thus to increased clearance of LDL from plasma in addition. They reduce both LDL and triglyceride levels, the latter through activation of lipoprotein lipase. Finally, these versatile drugs correct endothelial dysfunction through up-regulation of endothelial NO-synthase and decreasing endothelial superoxide production, as well as inhibiting proliferation of smooth muscle cells within the vascular wall (Dogra et al. 2002). They also have anti-inflammatory effects and inhibit cell proliferation, and thereby may decrease progression of renal disease (O'Donnell 2001).
Several statins have been shown to reduce LDL cholesterol in nephrotic patients by 35â€“40 per cent and TGs by 15â€“30 per cent in short-term studies. Simvastatin dosages ranging from 10 to 40 mg/24 h are required to lower cholesterol below 120 mg/l (3.1 mmol/l) (Rabelink et al. 1988; Warwick et al. 1992); there is no effect on Lp(a) concentrations. However, an effect on albuminuria and progression of renal disease has still to be shown in long-term studies (Samuelson 1998; Attman et al. 1999; Olbricht et al. 1999), although the recent meta-analysis of Fried et al. (2001) discussed above is encouraging. Other agents such as pravastatin (Toto et al. 1992) and lovastatin (Chan et al. 1992) have been used also in nephrotic patients with similar results, but simvastatin has been most studied in nephrotics. Olbricht et al. (1999) described 56 adults treated for 2 years, and noted a reduction in LDL cholesterol of 47 per cent with no change in HDL cholesterol. In general, statins are safe: however, both lovastatin and simvastatin rarely lead to rhabdomyolysis (Corpier et al. 1988) occasionally with acute renal failure, in the majority of cases the result of interactions with concomitant cyclosporin or gemfibrozil. Symptomless increases in muscle enzymes are often noted (Israeli et al. 1989; Olbricht et al. 1999), and the drug should be stopped if creatine kinase increases to greater than 500 i.u./l. Simvastatin has been reported to adversely affect growth in rats also, and has thus generally been avoided in children; pravastatin does not appear to have this disadvantage (Querfeld 1999).
Cholesterol is lowered less by fibrates than by statins, although they have a powerful effect on TGs in nephrotics as in other subjects (Groggel et al. 1989) since they stimulate lipoprotein lipase activity, and hence reduce VLDL cholesterol. However, since in nephrotics (unlike in renal failure) the main need is to reduce LDL cholesterol concentrations, it does not seem that fibrates are the best drugs to use - unless hypertriglyceridaemia is present as well. Fibrates induce muscle necrosis more frequently than statins, and fenofibrate produces a confusing increase in plasma creatinine which may be interpreted as a fall-off in renal function. Finally, the risk of gallstones is increased by both clofibrate and benzafibrate.
Probucol does not affect TG concentrations at all, but is effective in lowering cholesterol. Unfortunately, it may reduce HDL cholesterol concentrations as well as LDL cholesterol in many subjects. Its exact mode of action remains uncertain, and moreover no large population studies in cardiovascular disease are available. Its main attraction is that it reduces concentrations of the strongly atherogenic oxidized LDL by its antioxidant effect. Probucol has been used in nephrotic patients (Valeri et al. 1986; Iida et al. 1987), including children because of doubts about statins mentioned above (Querfeld et al. 1999), with a reduction in cholesterol concentrations; but it was taken off the market in 1999, and may not be reintroduced.
Nicotinic acid (1â€“6 g/24 h) and its derivatives, and bile acid binding resins (cholestyramine and colestipol) have both been used wth successful reduction of cholesterol concentrations in nephrotic patients, but have severe and frequent side-effects which lead to poor compliance, and they are little used today.
In patients with FSGS, statin treatment associated with lipopheresis resulted in a prolonged reduction of LDL cholesterol. Reduction of LDL by 60-70 per cent was achieved which persisted following discontinuation of lipopheresis. The effect on protein excretion and serum albumin levels was not consistent in all treated patients, although some achieved remission of the NS (Brunton et al. 1999; Muso et al. 1999).
Protein wasting and protein intake
Many nephrotic patients become seriously wasted during the course of their illness, presumably the result not only of protein losses, but also catabolism of much larger amounts of protein in the renal tubules. The optimal dietary protein intake for patients with a persisting NS remains controversial. Although recommended in the past, it has long been known that a high protein intake (1.5 g/kg/24 h) leads to an increase in urinary protein excretion but without any increase in serum albumin or total plasma protein concentrations. In contrast, reduction in protein intake to 0.8 g/kg/24 h reduces proteinuria, although with controversial effects on serum albumin concentration (Mansy et al. 1989; Kaysen 1991) and a risk of protein malnutrition (Guarneiri et al. 1989). Therefore, a dietary protein intake of 0.8 g of protein of high biological value/kg/24 h, plus 1 g protein per gram of proteinuria, has been advocated widely in adults with the NS. In fact 0.8 g/kg/24 h represents an average protein intake in Europe, although less than American norms.
Alterations in carbohydrate metabolism in nephrotics
Unlike lipid metabolism, there appear to be no direct clinical consequences of the changes in carbohydrate metabolism in the NS (Bridgman et al. 1975; Loschiavo et al. 1983). Accelerated glycogenolysis, elevated resting insulin and impaired response to oral glucose loading with a diabetic-like plasma glucose curve have all been reported.
Loss of binding and transport proteins and hormones in the urine
A number of plasma proteins important in the transport of metals, hormones, and drugs are of relatively small molecular weight and thus are lost easily into the urine of nephrotic patients (Table below). Free protein hormones, especially of low molecular weight, are also lost. These findings, however, although interesting, are usually of little clinical significance.
Table: Loss of binding proteins in urine in the nephrotic syndrome
||Andrassay et al. (1980)
||Pohl et al. (1984)
||Cameron et al.1988)b
|a Kanfer et al. (1970); Kendall et al. (1971); Thompson et al. (1974); Bernard et al. (1977); Kauffman et al. (1978). All these papers reported concentrations of coagulation factors in nephrotics.
|b All minimum changes.
Low serum iron and losses of transferrin (molecular weight 87 kDa) and associated iron into the urine are features of NS (Brown et al. 1984). Transferrin turnovers show increased synthesis and degradation, presumably in renal tubules after reabsorption, and Prinsen et al. (2001) point to a failure of increased synthesis to replace the losses as the cause of the low plasma concentrations. True iron deficiency (Brown et al. 1984) is, however, rare since urinary losses of iron are at most 0.5-1.0 mg/24 h. In addition, the transferrin-iron complex binds to many proliferating cells, and the effects of transferrin depletion on immunity may not depend entirely upon zinc deficiency (Warshaw et al. 1984).
Caeruloplasmin has a molecular weight of 151 kDa is lost in the urine and its plasma concentration is low (Brown et al. 1984). Low red-cell and plasma copper concentrations heve been reported by some authors in nephrotic children (Stec et al. 1990) and rats (Pedraza-Chaverri et al. 1994), but no clinical consequences of these losses of copper have been found.
Zinc circulates bound mainly to albumin and also to transferrin, and zinc is reduced in plasma, hair, and white cells in NS (Reimold 1980; Stec et al. 1990) and in plasma in nephrotic rats (Pedraza-Chaverri et al. 1994). Increased synthesis of Î±2-macroglobulin and its tight binding of zinc lower the biological availability of the metal (de Sain-van der Velden et al. 1998b). Hypogeusia may be found in nephrotics (Mahajan et al. 1982), and the possible effects in reducing cell-mediated immunity have been discussed above (Bensman et al. 1984; Warshaw et al. 1984).
Loss of vitamins and hormones in the urine
A number of abnormalities of calcium and vitamin D metabolism have been described, in part the result of losses of vitamin D binding protein (molecular weight 59 kDa) and its associated vitamin in the urine (Vaziri 1993; Harris and Ismail 1994). Bone biopsies are usually normal, although some may show osteomalacia and/or hyper-parathyroidism (Mittal et al. 1999) but without symptoms, and vitamin D and/or calcium supplementation is not necessary (Mehls 1990). However, nephrotics with reduced renal function do more readily develop bone disease (Tessitore et al. 1984), and earlier treatment than usual with vitamin D may have a place in those evolving into uraemia.
Despite losses of thyroid-binding globulin (molecular weight 85 kDa) in the urine (Vaziri 1993) accompanied by bound T3 and T4, plasma concentrations of T3, T4, and TSH are usually normal in nephrotic patients, although the T3 may be rather low and the T4 raised with increased reverse T3. In general, this is of no clinical significance, although cases of hypothyroidism remitting once proteinuria ceased have been reported (Fonseca et al. 1991) and hypothyroidism is a regular feature of the Finnish congential NS (Mattoo 1994).
Cortisol-binding protein (molecular weight 52 kDa) is lost in nephrotic urine and in one study (Musa et al. 1967) plasma concentration was reduced, but not in another (Loschiavo et al. 1983). More important is binding of prednisolone to albumin (see below).
Erythropoietin has a molecular weight of 33 kDa and is thus readily lost in the urine, in both health and in proteinuric patients (Vaziri et al. 1992). A number of nephrotic patients, both children and adults, suffering anaemia disproportionate to their renal function have been described, whose anaemia responded to epoetin administration (Feinstein et al. 2001).
Despite the large number of drugs normally bound in part to albumin, the gross reductions in serum albumin in nephrotics give rise to remarkably few problems (Gugler and Arzanoff 1976). Those concerning warfarin (Ganeval et al. 1986) have been discussed above. Prednisolone is normally bound to albumin, but despite this even severe hypoalbuminaemia doses of prednisolone are not modified either in nephrotic adults (Frey and Frey 1984) or children (Miller et al. 1990; Rostin et al. 1990), since the levels of free drug equilibrate rapidly to nearly normal concentrations. However, Miller et al. (1990) suggest that differential pharmacokinetics may account for differing profiles of toxicity, as in the Boston drug surveillance programme.
Renal effects of the nephrotic syndrome
Acute renal failure in nephrotic patients (Smith and Hayslett 1992)
Apparently "idiopathic" acute renal failure is an occasional but important complication of NS, and can be distinguished from acute renal failure from identifiable causes such as interstitial nephritis, thrombosis, sepsis, or contrast media. About 100 such cases have been published (Koomans et al. 1992; Smith and Hayslett 1992) mostly in previous decades; it may be that this complication has become less common in recent years, but under-reporting is also possible.
This complication occurs mostly in older patients of either gender, overwhelmingly (81 per cent) in those with minimal change/FSGS histology, despite the overall preponderance of young children in the latter group. In such children, acute renal failure is, in contrast, very rare, and usually follows either sepsis or thrombosis (Cameron et al. 1988; Cavagnaro and Lagomarsino 2000). Most adult patients present already in acute renal failure; few develop it subsequent to diagnosis, or in relapse. One-sixth were judged to be seriously hypovolaemic or in shock, and all had a very low serum albumin. Urine volume is low, containing less than 5 mmol/l Na and unresponsive to diuretics and/or volume repletion, loaded with protein, and containing red cells and often red cell casts. Thus, renal biopsy is almost always necessary to establish a diagnosis, as this pattern of sediment suggests a proliferative nephritis rather than minimal changes.
The role of hypovolaemia and reduced renal perfusion is not clear; in many patients the circulation was full or overfull and hypertension was common. In six childhood patients during relapse of the NS, we observed a decrease in GFR below 20 ml/min/1.73 m2 associated with normal RPF and a significant decrease of FF, inversely correlated with vasoactive hormones (renin, noradrenaline) and positively correlated with plasma oncotic pressure. Following intravenous albumin infusion, in none of these patients did the FF increase. These data are in agreement with a severely reduced Kf as the cause of acute renal failure in patients with the NS (Vande Walle et al. 2004).
Moderate to severe tubular changes were present in the majority of reported cases, amounting to frank tubular necrosis in about one-fourth. Proteinaceous casts are an important feature in only a minority (Chamberlain et al. 1966; Venkataseshan et al. 1993). Interstitial oedema is usually present, perhaps indicating increased interstitial pressure (Lowenstein et al. 1970). Blood vessel changes are frequent, and have been thought by some to be important in the pathogenesis (Jennette and Falk 1990), but these may be the result of the age of the patients.
Management of these often elderly and severely ill patients is difficult. They continue to pass large amounts of protein in tiny amounts of urine, have very low serum albumin and sometimes unstable circulation, and are of course uraemic. If not already malnourished, they rapidly become so. Obviously, they require dialysis by some form of blood purification with gentle ultrafiltration and salt-free albumin infusions given cautiously to avoid pulmonary oedema. Supplementary intravenous nutrition may be required, with an obligation to filter large enough volumes to make this possible. Should they receive concomitant treatment with corticosteroids and/or cyclophosphamide? Clearly, the risks are high, but some authors have suggested that treatment in this fashion led to a diuresis, and it may help eliminate or reduce the proteinuria. Others have seen no benefit, and there is no clear answer to this question. Outcome is serious, with a mortality in published cases of 18 per cent, and another 20 per cent survive but never recover renal function. Recovery of renal function often takes several months rather than weeks. A few patients have gone on to have further relapses of their NS, but without renal failure in these episodes.
In nephrotic children, acute renal failure is very rare (Cameron et al. 1988; Cavagnaro and Lagomarsino 2000). In the majority of these few cases some other factor - major sepsis or thrombosis - was present. Recovery has been usual, although one of our patients lost both her legs as a result of associated aortic thrombosis, and was still relapsing 10 years later.
Renal tubular dysfunction in nephrotics
A Fanconi syndrome has been described in a small number of nephrotic patients. Some of these had reversible tubular defects dependent upon tubular damage from proteinuria (Shioji et al. 1974) but in most, the tubular defect arises from tubular damage as part of the underlying disease (Praga et al. 1991), and do worse than those without glycosuria/ aminoaciduria. In children, this finding (Boissou et al. 1980; McVicar et al. 1980) or Î²2-microglobulinuria (Portman et al. 1986) can help differentiate children with focal segmental glomerulosclerosis from those with only minimal change lesions.
Bernard, D. B. (1988). Extrarenal complications of the nephrotic syndrome. (Nephrology Forum). Kidney International 33, 1184-1202.
Cameron, J. S., Ogg, C. S., and Wass, V. J. Complications of the nephrotic syndrome. In The Nephrotic Syndrome (ed. J. S. Cameron and R. J. Glassock), pp. 849-920. New York: Marcel Dekker, 1988.
Kaysen, G. A., ed. (1993). The nephrotic syndrome: pathogenesis and consequences. American Journal of Nephrology 13, 309-428.
Orth, S. R. and Ritz, E. (1998). The nephrotic syndrome. New England Journal of Medicine 38, 1202-1211.