Diagnostic evaluation of dogs was performed at the various coauthors' institutions; analysis of data and manuscript preparation was performed at Purdue University, West Lafayette, IN. Preliminary results of this study were presented as a research abstract at the 28th annual ACVIM Forum, Anaheim, CA, 2010.
Corresponding author: Barrak M. Pressler, Department of Veterinary Clinical Sciences, School of Veterinary Medicine, Purdue University, West Lafayette, IN 47907; e-mail: email@example.com.
Background: Nephrotic syndrome (NS) develops most commonly in people with glomerular diseases associated with marked albuminuria. Hypernatremia, hypertension, and progressive renal failure are more prevalent in nephrotic than nonnephrotic human patients.
Hypothesis/Objectives: Dogs with NS have higher serum cholesterol, triglyceride, and sodium concentrations, higher urine protein:creatinine ratios (UPC) and systolic blood pressure, and lower serum albumin concentrations than dogs with nonnephrotic glomerular disease (NNGD). NS is associated with membranous glomerulopathy and amyloidosis. Affected dogs are more likely to be azotemic and have shorter survival times.
Animals: Two hundred and thirty-four pet dogs (78 NS dogs, 156 NNGD dogs).
Methods: Multicenter retrospective case-control study comparing time-matched NS and NNGD dogs. NS was defined as the concurrent presence of hypoalbuminemia, hypercholesterolemia, proteinuria, and extravascular fluid accumulation. Signalment, clinicopathologic variables, histopathologic diagnoses, and survival time were compared between groups.
Results: Age, serum albumin, chloride, calcium, phosphate, creatinine, and cholesterol concentrations, and UPC differed significantly between NS and NNGD dogs. Both groups were equally likely to be azotemic at time of diagnosis, and NS was not associated with histologic diagnosis. Median survival was significantly shorter for NS (12.5 days) versus NNGD dogs (104.5 days). When subgrouped based on serum creatinine (< or ≥1.5 mg/dL), survival of NS versus NNGD dogs was only significantly different in nonazotemic dogs (51 versus 605 days, respectively).
Conclusions and Clinical Importance: Presence of NS is associated with poorer prognosis in dogs with nonazotemic glomerular disease. Preventing development of NS is warranted; however, specific interventions were not evaluated in this study.
Nephrotic syndrome (NS) is defined as the concurrent presence of hypoalbuminemia, proteinuria, hypercholesterolemia, and extravascular fluid accumulation.1,2 In people, urine protein:creatinine ratios (UPC) of >2.0–3.5 or 300–350 mg/mmol are referred to as nephrotic-range proteinuria.3,4 Serum albumin concentration in these patients is often <2.5 g/dL and total cholesterol concentration is usually >240 mg/dL.3,5 People with NS are at higher risk than those with nonnephrotic glomerular disease (NNGD) for development of complications, including hypernatremia, thromboembolic disease, hypertension, and progressive renal failure.1,2,6–8 Minimal change disease and membranous nephropathy are the primary glomerular diseases most commonly diagnosed in children and adults with NS, respectively, whereas diabetic nephropathy and amyloidosis are the most commonly diagnosed secondary glomerular diseases in adults with NS.2,9 Prognosis in people with NS is influenced primarily by the underlying glomerular disease, but more severe urinary protein loss increases the risk of progressive renal failure.2,10
NS is well-recognized in dogs.11–22 However, the relationship between NS and signalment, type of glomerular disease, severity of proteinuria, hypoalbuminemia, and development of complications is unknown. Amyloidosis, membranous nephropathy, and minimal change disease may be more likely to result in NS because of greater urinary albumin loss than other glomerulopathies.12,13,20–23 Although the presence of proteinuria decreases time until first uremic crisis and renal-related death in dogs with chronic kidney disease,24 whether or not prevalence of azotemia is greater in NS than NNGD dogs and whether or not presence of NS is associated with decreased survival time have not been investigated.
The purpose of this study was to characterize signalment, selected clinicopathologic findings, extravascular fluid accumulation, glomerular disease histologic type, and survival time in dogs with NS. We hypothesized that NS dogs would have significantly greater UPCs, serum cholesterol, triglyceride, and sodium concentrations, and systolic blood pressure, and significantly lower serum albumin concentrations and antithrombin (AT) activity than NNGD dogs. We also hypothesized that NS would be associated with membranous glomerulopathy and amyloidosis, but not with membranoproliferative glomerulonephritis (MPGN). Finally, we hypothesized that NS dogs would have higher serum creatinine and blood urea nitrogen (BUN) concentrations and decreased survival time as compared with NNGD dogs.
Materials and Methods
Dogs were identified using medical record database searches at 8 university veterinary teaching hospitals (Purdue University, January 1, 1987–August 1, 2008; The Ohio State University, January 1, 2000–May 11, 2009; University of Pennsylvania, January 1, 1990–March 1, 2009; University of Tennessee, January 1, 1977–January 1, 2009; North Carolina State University, January 1, 2005–April 1, 2009; University of Illinois, January 1, 1992–December 31, 2006; Virginia Tech, January 1, 1984–August 31, 2008; Kansas State University, May 1, 2004–April 1, 2009). Identified records were first reviewed by an ACVIM board-certified small animal internal medicine diplomate with an interest in nephrology/urology from that institution. Data from dogs from all institutions were then reviewed again by a single co-author (B.M.P.) before final inclusion.
Dogs in the NS were identified using the database search terms “nephrotic syndrome” and “dog.” Inclusion in the NS group required the concurrent presence of all 4 abnormalities used to define NS in people. Hypoalbuminemia and hypercholesterolemia were based on each institution's reference ranges. Proteinuria of glomerular origin was diagnosed either by a UPC > 1.0 or by >“1+” urine protein concentration by dipstick examination with concurrent absence of other urine dipstick or sediment examination findings that suggested a possible postglomerular source of increased urine protein, any additional diagnostic test results (ie, more specific evaluation of liver and gastrointestinal tract function, if performed), and the record reviewers' clinical judgment that alternative causes of hypoalbuminemia or proteinuria were unlikely. Extravascular fluid accumulation was diagnosed by physical examination or diagnostic imaging. Dogs that had all 4 criteria noted during the course of evaluation, but not simultaneously, were excluded from the NS group.
Dogs in the NNGD group were identified by the database search terms “dog” and either “protein-losing nephropathy,”“glomerular disease,” or “glomerulopathy.” Two dogs that presented within 6 months of each NS dog were then selected to control for variations over time in clinical pathology assays and institutional reference ranges, diagnostic imaging modalities used, and recommended treatments for NNGD and NS. Specific clinicopathologic findings required for a diagnosis of NNGD were not established a priori; however, any dog with insufficient diagnostic testing to satisfy both reviewing authors' clinical judgment that this diagnosis was appropriate was excluded and replaced. Dogs were not excluded from the NNGD group if glomerular disease was suspected to be because of a nonrenal inflammatory disease, and no specific diagnostic testing for diseases commonly associated with increased urine protein loss were required for inclusion. Dogs with glomerular disease and any 3 or fewer criteria used to define NS were eligible for inclusion in the NNGD group, whereas dogs with insufficient diagnostic testing to exclude presence of NS were excluded. All dogs ultimately included in the NNGD group had either (1) a UPC > 1.0 and the record reviewers agreement that pre- or postglomerular causes of proteinuria had been appropriately excluded based on diagnostic test results and clinical judgment, (2) concurrent hypoalbuminemia and >“1 +” urine protein as determined by urine dipstick examination, absence of other urine dipstick or sediment examination findings that suggested a possible postglomerular source of increased urine protein, and the record reviewers' agreement that alternative causes of hypoalbuminemia or proteinuria had been appropriately excluded based on diagnostic test results and clinical judgment, or (3) historical clinicopathologic findings consistent with glomerular disease, >“1 +” urine protein as determined by urine dipstick examination, absence of other urine dipstick or sediment examination findings that suggested a possible postglomerular source of increased urine protein, and confirmation of glomerular disease on necropsy examination.
Information collected included breed, sex, neuter status, date of birth, date of 1st evaluation for NS or NNGD, and date azotemia was first noted. Extravascular fluid accumulation in NS dogs was categorized as “ascites,”“subcutaneous edema,”“pleural effusion,” or “other,” with more than one accepted if applicable. Clinicopathologic data extracted from medical records included absolute platelet count, serum albumin, cholesterol, triglyceride, creatinine, BUN, sodium, chloride, and potassium concentrations, serum AT activity, prothrombin time (PT), activated partial thromboplastin time (APTT), urine specific gravity, UPC, and indirect systolic blood pressure. Extravascular fluid nucleated cell count and protein concentration were collected from medical records of dogs with NS. Renal biopsy and necropsy reports were reviewed when available; final glomerular disease diagnosis was based on light, electron, and immunofluorescence microscopy results as available, but examination of kidney tissue with all 3 modalities was not required for inclusion. Survival in days was calculated from the date of 1st evaluation for NS or NNGD until the date of death or euthanasia. Dates that dogs were last known to be alive before being lost to follow-up or date of death or euthanasia were determined from medical records at the participating institutions or by phone contact with referring veterinary practices.
The number of dogs included in statistical analyses was not uniform because some results were not available for all dogs. The date of 1st evaluation at each institution was considered to be the date of diagnosis for purposes of calculating survival time, even when dogs had evidence of glomerular disease before 1st presentation.
Statistical analyses were performed by commercially available software.a Normally distributed variables (age at diagnosis, serum potassium concentration, systolic blood pressure) were compared between groups by the 2-sample t-test with equal variances, and nonnormally distributed continuous variables (all other variables) were compared between groups by the 2-sample Wilcoxon rank-sum (Mann-Whitney) test. Categorical data were compared by the Pearson χ2 test. Correlation was determined by the Spearman rank correlation coefficient (rs). For all tests, P-values < .05 were considered significant. For survival analysis dogs were compared based solely on categorization within NS or NNGD groups and when further subgrouped based on presence or absence of azotemia, defined as serum creatinine concentration >1.5 mg/dL. Dogs were included in the survival analysis if they had lived beyond the date of diagnosis and date of death was known; dogs were censored from survival analysis if they were still alive at the time of data collection or if they were known to have survived beyond initial evaluation but were subsequently lost to follow-up. Kaplan-Meier survival curves were plotted and survival times were compared by the log-rank test.
Seventy-eight dogs with NS met all inclusion criteria. Seventy-six of 78 NS dogs were first evaluated after January 1, 1990. Two dogs with NNGD were selected for each NS dog based on the dates of 1st diagnosis, with 156 NNGD dogs successfully identified for this comparison population. The number of NS cases contributed from the 8 participating institutions ranged from 2 to 14 (median, 11). Median number of dogs newly diagnosed with NS per year ranged from 0.4 to 2.6, with an overall median of 0.5 newly diagnosed cases per year. Initial diagnosis of NS was made at the same time as 1st diagnosis of glomerular disease in 70 of 78 (90%) dogs. Of the 8 NS dogs that had been diagnosed with NNGD before diagnosis with NS, 7 developed NS within 1 month, and 1 dog developed NS after 128 days (median, 15.5 days).
There was no significant difference in sex or neuter status between NS and NNGD dogs (P= .383; Table 1). There were 37 different pure breed varieties represented in the NS dog group, and 47 pure breed varieties in the NNGD group. Distribution of breeds with >10 individuals present in the total population of 234 dogs (Cocker Spaniels, Dalmatians, Golden Retrievers, Labrador Retrievers, Shetland Sheepdogs, and mixed breed dogs) was not significantly different between NS and NNGD dogs (P= .125; Table 1). Dogs with NS were significantly younger (mean ± SD, 6.2 ± 2.9 years) at 1st evaluation than NNGD dogs (mean ± SD, 8.4 ± 3.2 years; P< .001).
Table 1. Selected characteristics of dogs with glomerular disease with or without nephrotic syndrome.a
Nonnephrotic Glomerular Disease
All data presented as number of dogs (%).
b Includes 32 additional breeds in the nephrotic syndrome group and 42 additional breeds in the nonnephrotic glomerular disease group.
Extravascular fluid in NS dogs was most commonly noted in the peritoneal cavity (n = 58; 74%), followed by the subcutaneous tissues (n = 48; 62%) and pleural space (n = 16; 21%). Additional sites of fluid accumulation were the intramural intestinal tract (2 dogs), retroperitoneal space (1 dog), and perinephric region (1 dog). Protein concentration in extravascular fluid was <2.5 g/dL in all 23 dogs for which fluid analysis was performed. Total nucleated cell count was <1,000 cells/μL in 19 of 22 (86%) samples and <1,500 cells/μL in all (100%) samples.
Selected Clinical and Clinicopathologic Findings
Systolic blood pressure did not significantly differ (P= .089) between NS dogs (N = 57; mean ± SD, 177 ± 35 mmHg) and NNGD dogs (N = 97; mean ± SD, 167 ± 33 mmHg). Method of blood pressure measurement (ie, oscillometric versus Doppler ultrasonography; anatomic site; whether the recorded value was the result of a single or multiple averaged measurements) was not compared because of absent or incomplete information available in the medical records of most dogs.
Clinicopathologic findings are summarized in Table 2. Median serum albumin concentration was significantly lower and median serum cholesterol concentration and UPC significantly higher in NS dogs than in NNGD dogs. Median serum triglyceride concentration was similar in NS (N = 3; median, 79 mg/dL; range, 64–94 mg/dL) and NNGD (N = 9; median, 85 mg/dL; range 44–325 mg/dL) dogs, but differences were not statistically analyzed for significance because of the small number of dogs for which results were available. Serum albumin concentration and UPC ranges overlapped for NS and NNGD dogs (Fig 1).
Table 2. Selected clinicopathologic findings in dogs with glomerular disease with or without nephrotic syndrome.a
Nephrotic Syndrome-Associated Findings
Nonnephrotic Glomerular Disease
Additional clinicopathologic findings reported in text.
bFor clarity, despite normal distribution, median and range displayed for serum potassium concentration rather than mean and standard deviation.
Median serum sodium and potassium concentrations were not significantly different between NS and NNGD dogs, whereas median serum chloride concentration was significantly higher in NS dogs, although the difference (4 mEq/L) was small. Median serum total calcium concentration was significantly lower in NS dogs than in NNGD dogs. There was positive but only moderate correlation between serum total calcium and albumin concentrations in the total population of dogs, in NS dogs, and in NNGD dogs (rs= 0.473, 0.659, and 0.341, respectively; P < .001 for all correlations). Median serum phosphate concentration was significantly higher in NS dogs than in NNGD dogs. There was positive, good correlation between serum phosphorus and creatinine concentrations in the total population of dogs, in NS dogs, and in NNGD dogs (rs= 0.753, 0.743, and 0.781, respectively; P < .001 for all correlations).
There were no significant differences between the 2 groups in any measured coagulation parameters (absolute platelet count, PT, PTT, serum AT activity). Serum AT activity was <80% in 10 of 17 (59%) NS dogs and 12 of 17 (71%) NNGD dogs. There was no significant correlation between AT activity and serum albumin concentration in either NS (rs= 0.228; P= .378) or NNGD (rs= 0.386; P= .114) dogs (Fig 2). Of the 34 dogs with known serum AT activity, 7 of 12 (58%) dogs with serum albumin concentration < 1.8 g/dL and 14 of 22 (64%) dogs with serum albumin concentration ≥1.8 g/dL had serum AT activity <80%.
Presence of Azotemia and Altered Urine Concentrating Ability
Median serum creatinine concentration, but not median BUN concentration, was significantly higher in NS dogs than NNGD dogs (P= .013 and .057, respectively; Table 2). When azotemia was defined as having a serum creatinine concentration ≥1.5 mg/dL, NS and NNGD dogs were equally likely to be azotemic at 1st diagnosis (NS dogs, 68% [53/78]; NNGD dogs, 52% [84/154]; P= .060). Urine specific gravity was not significantly different between the 2 groups (P= .216). Urine specific gravity was > 1.012 in 60 of 75 (80%) NS dogs and 113 of 152 (74.3%) NNGD dogs, including 42 of 51 (82%) azotemic NS dogs and 60 of 80 (75%) azotemic NNGD dogs in which serum creatinine and specific gravity had also been recorded.
Glomerular Disease Subtype
A histologic diagnosis was determined in 35 (45%) NS dogs and 50 (32.1%) NNGD dogs (Table 3). Histologic diagnosis was determined solely from renal biopsy in 37 dogs, including 18 NS dogs (23% of all NS dogs, 51% of NS dogs with a histologic diagnosis) and 19 NNGD dogs (12.2% of all NNGD dogs, 38% of NNGD with a histologic diagnosis), and solely from kidney tissue collected at time of necropsy in 46 dogs, including 16 NS dogs (21% of all NS dogs, 32% of NS dogs with a histologic diagnosis) and 30 NNGD dogs (19.2% of all NNGD dogs, 60% of NNGD with a histologic diagnosis); 2 dogs (1 NS dog, 1 NNGD dog) had both renal biopsy and necropsies. Six antemortem renal biopsies (3 NS dogs, 3 NNGD dogs) were considered nondiagnostic.
Table 3. Renal disease histologic subtypes diagnosed in >5 dogs with glomerular disease.a
Nonnephrotic Glomerular Disease
There was no significant difference in the frequency of these diagnoses between dogs with nephrotic syndrome or nonnephrotic glomerular disease (P= .364).
Data presented as number (%) of samples.
b Includes 1 dog with nonnephrotic glomerular disease concurrently diagnosed with membranous glomerulopathy and amyloidosis.
Specific histologic diagnoses made in >5 dogs in the total population included membranous glomerulopathy (11 NS dogs; 18 NNGD dogs), MPGN (8 NS dogs; 14 NNGD dogs), and amyloidosis (9 NS dogs; 7 NNGD dogs); 1 dog with NNGD was concurrently diagnosed with membranous glomerulopathy and amyloidosis (Table 3). Additional diagnoses included proliferative glomerulonephritis (1 NS dog, 4 NNGD dogs), hereditary glomerulopathy (1 NNGD dog), focal segmental glomerulosclerosis (1 NNGD dog), and severe tubular disease without glomerular lesions (1 NNGD dog); this final dog was accepted into the glomerular disease group because of a markedly increased UPC (13.8), concurrent hypoalbuminemia (2.2 g/dL), and lack of azotemia (serum creatinine 0.8 mg/dL), which persisted at subsequent evaluations. Renal biopsies with no lesions noted on light microscopy (2 NS dogs, 1 NNGD dogs) or glomeruli with severe global “end-stage” glomerulosclerosis (4 NS dogs, 2 NNGD dogs) that precluded identification of the original glomerular histologic abnormality by light microscopy were also recorded. There were no significant differences between NS and NNGD dogs (P= .364) in the frequency of diagnoses made in at least 5 dogs (membranous glomerulopathy, MPGN, amyloidosis, or “end-stage kidneys”; Table 3).
Date of death or euthanasia was known for 62 (80%) NS dogs and 96 (61.5%) NNGD dogs. Dogs not included in survival analysis included 64 (27.4%) dogs (16 NS dogs, 48 NNGD dogs) that were lost to follow-up immediately after initial evaluation, and 10 dogs (1 NS dog; 9 NNGD dogs) euthanized on the day of diagnosis. Dogs that were censored from survival analysis included 12 NNGD that were known to be alive at discharge but were lost to follow-up at a later date; follow-up for these dogs ranged from 6 to 1,856 days (median, 1,055 days). Length of follow-up for 4 NNGD dogs known to still be alive at the time of data collection ranged from 264 to 1,147 days. Insufficient data were available in most records to definitively determine cause of death for most dogs, and therefore comparison of causes of death between treatment groups or with length of survival was not attempted.
Median (12.5 days) and range of survival (0–2,783 days) for deceased NS dogs included in the survival analysis were significantly shorter than median (104.5 days) and range of survival (0–3,124 days) for deceased NNGD dogs (P < .001; Fig 3). When azotemia was defined as serum creatinine concentration ≥1.5 mg/dL, survival of deceased nonazotemic dogs (median, 497 days) in the total study population was significantly longer than survival of deceased azotemic dogs (median, 13 days; P < .001). When dogs in the NS versus NNGD groups were analyzed separately based on presence or absence of azotemia (Fig 4), significantly shorter survival of nonazotemic NS dogs (median, 51 days; mean, 390 days) was noted compared with deceased nonazotemic NNGD dogs (median, 605 days; mean, 800 days; P= .023). Survival time of deceased azotemic NS dogs (median, 10 days; mean, 71 days) was not significantly different from that of deceased azotemic NNGD dogs (median, 45 days; mean, 316 days; P= .099).
Age at time of initial diagnosis was the only signalment-associated factor significantly different between NS and NNGD dogs. Dogs with NS had significantly lower total serum calcium concentrations and significantly higher serum phosphate and creatinine concentrations and UPCs. Dogs with NS had significantly greater serum cholesterol and decreased serum albumin concentrations as compared with NNGD dogs; however, these differences are likely partially attributable to selection bias, as inclusion in the NS group required the presence of hypoalbuminemia and hypercholesterolemia. Unlike in people with NS, associations with hypernatremia, hypertension, or specific types of glomerular disease were not noted. Finally, presence of NS was associated with decreased survival in nonazotemic dogs, whereas azotemia was associated with decreased survival regardless of concurrent presence of NS.
People with UPC > 2.0–3.5 (300–350 mg/mmol) are said to have nephrotic-range proteinuria.3,4 Although quantification of urine protein excretion allows some prediction of risk, NS may occur in glomerular disease patients with serum albumin concentrations >2.5 g/dL, and conversely some patients may have marked hypoalbuminemia without concurrent edema.1,25 Albumin hypercatabolism, down-regulated synthesis, and compartmental redistribution contribute to glomerular disease-associated hypoalbuminemia, and marked hypoalbuminemia may occur despite sub-nephrotic–range proteinuria.1,25 Additionally, NS may develop in some animals with significant hypoalbuminemia, but with progressive glomerular damage urinary protein loss may decline in spite of persistent fluid accumulation. The large overlap in UPC ranges we observed in NS versus NNGD dogs is presumptively also because of the variable contributions of these other edema-promoting factors; fluid extravasation is likely multifactorial in origin, and thus serum albumin concentration is correlated with, but is not the only determinant of, NS. Regardless of the cause, this overlap precluded establishment of a cut-off value for nephrotic-range proteinuria that would aid in identifying dogs at risk of NS.
The cause of hypercholesterolemia in human patients with NS is unclear, but it may reflect a nonspecific increase in hepatic protein synthesis in response to hypoalbuminemia, low plasma oncotic pressure, or altered plasma viscosity.5 Unlike hypercholesterolemia, hypertriglyceridemia is not a defining feature of NS in people, and increases in serum cholesterol concentration are not always accompanied by increases in triglycerides.5,26 Hyperlipidemia is associated with increased prevalence of atherosclerosis and progressive azotemia, and human patients with persistent hypercholesterolemia commonly are treated with the “statin” class of 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors.5,25 Lipid-lowering therapy is not recommended in dogs with NS because atherosclerosis is uncommon in this species.27 Further studies are needed to determine whether prognosis in dogs with NS can be improved by interventions which decrease serum cholesterol concentration.
Fluid accumulation usually is limited to the subcutis in people with NS, with peripheral edema of the distal legs reported most commonly. Although subcutaneous fluid was diagnosed frequently in NS dogs in our study, peritoneal effusion was reported in an approximately equal number of dogs. Determining the precise site of subcutaneous fluid accumulation in all NS dogs was not possible, but the ventrum, jaw, and neck appeared to be most commonly affected. Regional variations in intravascular hydrostatic pressure (ie, higher in ventrally dependent regions in order to facilitate blood return to the heart against the force of gravity) or accumulation in gravity-dependent regions may be responsible for this difference between species.28
We did not find an association between NS and hypernatremia in dogs, although hypernatremia is a common finding in people with NS.2,7,29 The “overfill” hypothesis proposes that NS results from a primary renal tubular defect that induces excessive reabsorption of sodium, resulting in secondary water resorption and eventual fluid extravasation because of hypoalbuminemia-associated decreases in oncotic pressure.1,7,29 Although our results may provide indirect evidence that the pathogenesis of edema formation differs between people and dogs, the overfill hypothesis cannot be discounted solely because of this finding. Increases in sodium resorption should be accompanied by both osmotic drag of water and increased pituitary ADH release, thus blunting or completely abrogating any increase in intravascular sodium concentration. Plasma volume would be expected to be increased in these dogs, but this has not been studied in dogs with NS. The overfill hypothesis also predicts that hypertension should be more common in dogs with NS; we did not note a significant difference in systolic blood pressures in NS versus NNGD dogs. Alternatively, the “underfill” hypothesis29 suggests that decreased oncotic pressure alone is sufficient for development of NS, and the “vascular hyperpermeability” hypothesis30,31 suggests that presence of a primary vascular defect results in a generalized increase in vascular porosity. Whether different mechanisms of edema formation occur with the various glomerular diseases in dogs (as likely occurs in people) is unclear, and our findings do not preferentially support one of these theories.
Likelihood of azotemia at the time of diagnosis of NS in people is determined primarily by the underlying glomerular disease rather than NS itself. Nevertheless, human NS patients with uncontrolled proteinuria are at increased risk for acute renal failure and progressive renal damage.2 Although NS dogs had significantly higher serum creatinine concentrations than NNGD dogs, the small difference is of unknown clinical significance. Many NS and NNGD dogs in this study were azotemic at the time of 1st evaluation, similar to previous reports.12,13 This relatively high prevalence may be because of later diagnosis and referral of dogs with glomerular disease as compared with people; alternatively this may reflect a difference between species in disease pathogenesis or progression. Maintenance of partial to complete urinary concentrating ability in the face of renal azotemia was common in NS and NNGD dogs, consistent with previous studies.12,13 This finding, sometimes referred to as “glomerulotubular imbalance” in the veterinary literature,32 presumptively occurs because tubular damage has not sufficiently progressed to result in complete loss of urine concentrating ability despite glomerular injury having progressed sufficiently to prevent appropriate filtration of nitrogenous wastes. Minimal change disease is the most common cause of NS in human patients <6 years of age, proliferative glomerulonephritis and focal segmental glomerulosclerosis are the most common causes of NS in adult, nongeriatric human patients, and membranous glomerulopathy is most common in patients >45 years of age.2,3,9,28 In contrast, we found no association between any of the 3 most common histologic diagnoses and the development of NS in dogs. We hypothesized that membranous glomerulopathy and amyloidosis would be associated with NS because previous cases series have reported either a high prevalence of NS (membranous glomerulopathy)12,21 or a higher frequency of ascites or peripheral edema (amyloidosis).12,13 One other study reported no increase in risk of NS associated with these 3 histologic diagnoses, although the percent of NS dogs (15%) may have been insufficient to detect a difference.12 The relative prevalences of MPGN, membranous nephropathy, and amyloidosis were similar in our study to those reported previously.12,13,22,33,34 The lack of association between NS and glomerular disease subtype may be because of differences in the pathogenesis of NS or specific glomerulopathies in dogs. Alternatively, these 3 subsets may have encompassed a more heterogenous group of glomerular diseases. Examination of renal tissues solely by light microscopy limited detection of ultrastructural or immunologic abnormalities. For example, a number of dogs had no histologic evidence of glomerular lesions despite the presence hypoalbuminemia and proteinuria. Some of these dogs may have had minimal change disease, which requires transmission electron microscopy to identify damaged podocytes. Minimal change disease is commonly diagnosed in people with NS and has been reported in dogs.23
Thromboembolic complications are associated with glomerular disease in both people and dogs, regardless of the presence of NS. Deep vein and renal thrombosis occurs more commonly in people with NS than those with NNGD.6,35 The reported prevalence of thromboembolism in dogs with glomerular disease has not been well-established, but up to 25.7% of all dogs with glomerular disease and 38.5% of dogs with amyloidosis have thromboemboli noted on necropsy examination.12,13,36 Decreased AT activity has been reported in dogs with glomerular disease,13 and case reports have noted circulating lupus-like anticoagulant,18 platelet hypersensitivity,15 and generalized evidence of dysregulated hemostasis.13,14,17 We did not find a significant difference in median platelet count, PT, or APTT between NS and NNGD dogs. Serum albumin concentration and AT activity are correlated in dogs and people,13,37,38 with 100% of dogs (n = 23) with serum albumin concentrations <1.8 g/dL having a serum AT activity <80%.13 However, normal or increased serum AT activity may be noted in some severely hypoalbuminemic human NS patients.39 We do not believe that our results affect the current recommendation that reduced serum AT activity should be assumed when serum albumin is <2.0 g/dL, and some client expense can be minimized by empirically treating with aspirin or equivalent agent.13,32 However, the lack of correlation between serum AT activity and serum albumin at concentrations >2.0 g/dL means that both NS and NNGD dogs with mild decreases in serum albumin concentration may also be at risk of thrombosis.40
We found that the presence of NS in nonazotemic dogs was significantly associated with shorter survival times. The reasons for this poorer outcome could not be adequately investigated in this retrospective study, as causes of death were oftentimes difficult to determine from record review, and some clinical signs noted immediately before death, such as central neurologic abnormalities, could potentially have been associated with glomerular disease in general but necropsies were not performed. In people with NS long-term prognosis is more heavily influenced by the underlying glomerular disease than the NS itself.4 However, several NS-associated complications, including cardiovascular disease and secondary infections, are increased in patients with uncontrolled edema, and do result in an overall worsened prognosis. Whether treatments to decrease edema and hypoalbuminemia in dogs with NS improve prognosis is unknown, but based on the poor survival identified here aggressive intervention is likely indicated. Slowing of disease progression in dogs with glomerular disease to delay or prevent development of azotemia appears to be critical, as survival of both NS and NNGD dogs was markedly shortened if serum creatinine concentration was ≥1.5 mg/dL. This finding appears to be a stronger influence on outcome than presence of absence of NS based on Kaplan-Meier curve analysis. Unfortunately, it is likely that nonazotemic dogs with glomerular disease are less likely to be referred for diagnostic evaluation and treatment than either azotemic NS or NNGD dogs. Despite increased awareness that diagnostic evaluation of dogs with asymptomatic proteinuria is indicated, and of the potential benefits of angiotensin-converting enzyme inhibitors,41 prognosis in dogs with NNGD in this study is the same as that reported almost 15 years ago.13
aStata version 10.2, StataCorp, College Station, TX
This study was not supported by any funding agencies or grants.
Conflict of interest disclosure: Dr Stephen P. DiBartola is Co-Editor-in-Chief of the Journal of Veterinary Internal Medicine (JVIM). Dr Barrak M. Pressler is Associate Editor of the JVIM. Dr George E. Moore is Consulting Editor for Experimental Design and Statistics of the JVIM.