Disease-specific risk of venous thromboembolic events is increased in idiopathic glomerulonephritis
2011; Elsevier BV; Volume: 81; Issue: 2 Linguagem: Inglês
10.1038/ki.2011.312
ISSN1523-1755
AutoresSean J. Barbour, Allen Greenwald, Ognjenka Djurdjev, Adeera Levin, Michelle Hladunewich, Patrick H. Nachman, Susan L. Hogan, Daniel C. Cattran, Heather N. Reich,
Tópico(s)Autoimmune Bullous Skin Diseases
ResumoThe risk of venous thromboembolic events is thought to be highest in patients with membranous nephropathy. This association has been recently questioned, and it is not known whether this simply reflects the severity of proteinuria. To better understand the relationship between histologic diagnosis and the risk of venous thromboembolic events we evaluated patients in the Toronto Glomerulonephritis Registry. Of 1313 patients with idiopathic glomerulonephritis, 395 were diagnosed with membranous nephropathy, 370 with focal segmental glomerulosclerosis (FSGS), and 548 with immunoglobulin-A nephropathy (IgAN). Risk factors were evaluated by Cox proportional hazards for 53 image-confirmed venous thromboembolic events in 44 patients during a median follow-up of 63 months. The risk was highest in patients with membranous nephropathy and FSGS (hazard ratios of 22 and 7.8, respectively) referenced to patients with IgAN. Following adjustment for gender, cancer history, proteinuria, and serum albumin by multivariable analysis, the histologic subtype remained an independent risk for venous thromboembolic events. This risk was still highest in patients with membranous nephropathy followed by FSGS with adjusted hazard ratios of 10.8 and 5.9, respectively. Thus, in this large cohort, histologic diagnosis was an independent risk factor for venous thromboembolic events. Further studies are needed to discover mechanisms responsible for this high risk in patients with membranous nephropathy. The risk of venous thromboembolic events is thought to be highest in patients with membranous nephropathy. This association has been recently questioned, and it is not known whether this simply reflects the severity of proteinuria. To better understand the relationship between histologic diagnosis and the risk of venous thromboembolic events we evaluated patients in the Toronto Glomerulonephritis Registry. Of 1313 patients with idiopathic glomerulonephritis, 395 were diagnosed with membranous nephropathy, 370 with focal segmental glomerulosclerosis (FSGS), and 548 with immunoglobulin-A nephropathy (IgAN). Risk factors were evaluated by Cox proportional hazards for 53 image-confirmed venous thromboembolic events in 44 patients during a median follow-up of 63 months. The risk was highest in patients with membranous nephropathy and FSGS (hazard ratios of 22 and 7.8, respectively) referenced to patients with IgAN. Following adjustment for gender, cancer history, proteinuria, and serum albumin by multivariable analysis, the histologic subtype remained an independent risk for venous thromboembolic events. This risk was still highest in patients with membranous nephropathy followed by FSGS with adjusted hazard ratios of 10.8 and 5.9, respectively. Thus, in this large cohort, histologic diagnosis was an independent risk factor for venous thromboembolic events. Further studies are needed to discover mechanisms responsible for this high risk in patients with membranous nephropathy. Complications of the nephrotic syndrome are an important cause of morbidity in patients with idiopathic glomerulonephritis (GN).1.Canadian Institute for Health Information. 2010 Annual Report—Treatment of End-Stage Organ Failure in Canada, 1999 to 2008. http://secure.cihi.ca/cihiweb/products/corr_annual_report_2010_e.pdf.Google Scholar One such complication is venous thromboembolic events (VTEs) including deep and renal vein thromboses (RVTs), as well as pulmonary emboli. The reported risk of VTEs in patients with nephrotic syndrome is highly variable, due to diverse patient populations and variable screening practices. The risk of VTE is not well defined, and is reported to range from 3 to 48%, depending upon the site of the VTE and the intensity of diagnostic screening.2.Kayali F. Najjar R. Aswad F. et al.Venous thromboembolism in patients hospitalized with nephrotic syndrome.Am J Med. 2008; 121: 226-230Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar,3.Glassock R.J. Prophylactic anticoagulation in nephrotic syndrome: a clinical onundrum.J Am Soc Nephrol. 2007; 18: 2221-2225Crossref PubMed Scopus (161) Google Scholar Given the potentially significant morbidity and mortality associated with these events,4.Bellomo R. Atkins R.C. Membranous nephropathy and thromboembolism: is prophylactic anticoagulation warranted?.Nephron. 1993; 63: 249-254Crossref PubMed Scopus (96) Google Scholar a better understanding of the frequency of clinically detected VTEs would help to guide therapy. The nephrotic syndrome is considered to be a thrombophilic milieu due to a variety of elements including measurable changes in coagulation factors due to protein loss in the urine, altered platelet activity, intravascular volume contraction, venous stasis accompanying edema, and chronic inflammation.5.Llach F. Hypercoagulability, renal vein thrombosis, and other thrombotic complications of nephrotic syndrome.Kidney Int. 1985; 28: 429-439Abstract Full Text PDF PubMed Scopus (323) Google Scholar Historically, the cause of the underlying glomerular disease was proposed to be an important risk for VTE, with membranous nephropathy (MN) implicated as a risk factor for these events.3.Glassock R.J. Prophylactic anticoagulation in nephrotic syndrome: a clinical onundrum.J Am Soc Nephrol. 2007; 18: 2221-2225Crossref PubMed Scopus (161) Google Scholar However, studies that implicated MN as a risk factor for VTE primarily diagnosed asymptomatic thrombotic events on routine screening of uncertain clinical impact; did not adjust for the level of proteinuria or hypoalbuminemia, population age, or cancer; and did not compare with disease-specific controls.5.Llach F. Hypercoagulability, renal vein thrombosis, and other thrombotic complications of nephrotic syndrome.Kidney Int. 1985; 28: 429-439Abstract Full Text PDF PubMed Scopus (323) Google Scholar, 6.Llach F. Koffler A. Finck E. et al.On the incidence of renal vein thrombosis in the nephrotic syndrome.Arch Intern Med. 1977; 137: 333-336Crossref PubMed Scopus (62) Google Scholar, 7.Llach F. Papper S. Massry S.G. The clinical spectrum of renal vein thrombosis: acute and chronic.Am J Med. 1980; 69: 819-827Abstract Full Text PDF PubMed Scopus (182) Google Scholar, 8.Velasquez Forero F. Garcia Prugue N. Ruiz Morales N. Idiopathic nephrotic syndrome of the adult with asymptomatic thrombosis of the renal vein.Am J Nephrol. 1988; 8: 457-462Crossref PubMed Scopus (57) Google Scholar, 9.Chugh K.S. Malik N. Uberoi H.S. et al.Renal vein thrombosis in nephrotic syndrome—a prospective study and review.Postgrad Med J. 1981; 57: 566-570Crossref PubMed Scopus (51) Google Scholar In contrast to historical tenets, the largest study to address this topic showed an annual incidence rate of VTE in MN similar to that seen in other glomerular diseases.10.Mahmoodi B.K. ten Kate M.K. Waanders F. et al.High absolute risks and predictors of venous and arterial thromboembolic events in patients with nephrotic syndrome: results from a large retrospective cohort study.Circulation. 2008; 117: 224-230Crossref PubMed Scopus (254) Google Scholar Therefore, the disease-specific risk of VTE in patients with idiopathic GN remains an area of significant uncertainty. Accordingly, we sought to define the disease-specific risk of clinically evident VTEs in patients with idiopathic GN after adjusting for thrombophilic risk factors common in glomerular diseases such as proteinuria, hypoalbuminemia, and malignancy. We studied a large cohort of patients with idiopathic immunoglobulin-A nephropathy (IgAN), focal and segmental glomerulosclerosis (FSGS), and MN, who were followed up longitudinally in the Toronto Glomerulonephritis Registry in order to better understand the disease-specific risk of VTE in patients with idiopathic GN. The study cohort comprised 1313 patients, including 370 subjects with FSGS, 548 with IgAN, and 395 with MN. Reasons for exclusion from the study are outlined in Supplementary Figure S1 online. Baseline characteristics are described in Table 1. The cohort comprised 63% male patients, followed up for a median of 63 months with baseline proteinuria of 3.1g/day and albumin of 32.9g/l. Download .doc (.05 MB) Help with doc files Supplementary Figure S1Table 1Baseline (within 6 months of presentation) and longitudinal characteristicsOverallFSGSIgANMNP-valueNumber of patients1313370548395Length of follow-up (months; median (LQ, UQ))63.0 (33.3, 108.0)67.8 (35.9, 110.1)62.6 (33.4, 105.5)59.6 (30.3, 104.9)0.3Age at presentation (years; mean (s.d.))42.2 (15.2)42.5 (15.8)38.1 (13.1)45.6 (15.6)<0.0001Sex (% male)63.162.461.366.10.3Race (%)<0.0001 Caucasian67.669.058.678.5 African American5.69.03.25.4 Asian15.311.026.64.0 Other11.611.011.612.0Estimated CrCl at presentation (ml/min per 1.73m2; mean (s.d.))74.4 (30.0)71.6 (30.9)73.5 (29.2)78.4 (30.0)0.0007Proteinuria at presentation (g/day; median (LQ, UQ))3.1 (1.5, 6.5)3.7 (1.8, 6.8)1.6 (0.8, 2.9)5.6 (3.3, 9.9)<0.0001TA proteinuria during follow-up (g/day; median (LQ, UQ))2.6 (1.4, 4.6)3.3 (1.8, 4.9)1.8 (0.8, 2.9)3.9 (2.0, 6.5)<0.0001Albumin at presentation (g/l; mean (s.d.))32.9 (9.1)33.1 (9.7)38.2 (6.6)26.8 (7.1)<0.0001TA albumin during follow-up (g/l; mean (s.d.))36.5 (6.2)36.6 (6.8)39.1 (4.5)33.0 (6.0) 8g/day) at presentation, and sustained over time, compared with other GN subtypes (P<0.0001). There were 53 VTEs in 44 patients, with the types of events described in Table 2. A greater proportion of patients with MN developed VTEs, compared with FSGS or IgAN (7.85% vs. 2.97% vs. 0.36%, respectively, P<0.0001). Even when isolated RVT was excluded from the analysis, a larger proportion of patients with MN experienced a VTE compared with patients with FSGS or IgAN (4.8% vs. 2.7% vs. 0.36%, P<0.0001).Table 2The number and nature of VTEsOverall, N=1313FSGS, N=370IgAN, N=548MN, N=395P-valuePatients with a VTE (number (%))44 (3.4)11 (3.0)2 (0.4)31 (7.9)<0.0001Number of VTEs of each type DVT10415 PE208111 RVT192017 Other4004Days to first VTE (median (LQ, UQ))272 (0, 1080)1094 (401, 1604)453 (363, 542)151 (0, 447)0.07Abbreviations: DVT, deep vein thrombosis; FSGS, focal segmental glomerulosclerosis; IgAN, immunoglobulin-A nephropathy; LQ, lower quartile; MN, membranous nephropathy; PE, pulmonary embolism; RVT, renal vein thrombosis; UQ, upper quartile; VTE, venous thromboembolic event. Open table in a new tab Abbreviations: DVT, deep vein thrombosis; FSGS, focal segmental glomerulosclerosis; IgAN, immunoglobulin-A nephropathy; LQ, lower quartile; MN, membranous nephropathy; PE, pulmonary embolism; RVT, renal vein thrombosis; UQ, upper quartile; VTE, venous thromboembolic event. Venous thromboses occurred at a median of 272 days after first presentation. Overall, 57% of VTEs occurred within 1 year and 70% within 2 years of presentation. The VTEs tended to occur closer to the time of GN diagnosis in patients with MN; however, there was a wide range in the timing of the event, and this did not reach statistical significance. The risk of VTE over time is illustrated in the Kaplan–Meier curve in Figure 2. The risk of thrombosis during the follow-up period differed across all three groups, with the lowest risk in IgAN patients and the highest in MN patients (P<0.0001 across groups). The characteristics of patients at the time of VTE according to underlying GN diagnosis are summarized in Table 3; at the time of the VTE, patients with MN and FSGS were hypoalbuminemic and had nephrotic-range proteinuria.Table 3Description of patients with (VTE+) and without (VTE-) VTEsFSGS, N=11 eventsMN, N=31 eventsIgAN, N=2 eventsVTE+VTE-VTE+VTE-VTE+VTE-Albumin at presentation (g/l)26.1 (8.7)33.2 (9.7)21.9 (6.1)27.2 (7.1)40, 46aValues presented are first available.38.2 (6.6)TA albumin over entire follow-up (g/l)33.3 (5.4)36.7 (6.8)32.7 (6.1)33.0 (5.9)44.0, 46.039.1 (4.5)Proteinuria at presentation (g/day)8.2 (3.0, 13.7)3.6 (1.8, 6.7)6.5 (4.4, 11.2)5.6 (3.2, 9.7)3.5, 0.5aValues presented are first available.1.6 (0.8, 2.9)TA proteinuria over entire follow-up (g/day)5.9 (4.1, 7.0)3.1 (1.8, 4.9)4.9 (2.7, 9.3)3.7 (1.9, 6.5)0.5, 2.81.8 (0.8, 2.9)Albumin at time of VTE (g/l)32.8 (8.2)NA25.5 (7.4)NA40.0bAlbumin at time of VTE was not available for one patient.NAProteinuria at time of VTE (g/day)6.2 (3.9, 10.7)NA9.8 (3.7, 11.7)NA3.4, 0.5NADays to first VTE1094 (401, 1604)NA151 (0, 447)NA363, 542NAAbbreviations: FSGS, focal segmental glomerulosclerosis; IgAN, immunoglobulin-A nephropathy; MN, membranous nephropathy; NA, not applicable; TA, time averaged; VTE, venous thromboembolic event.Data are provided as median (lower quartile, upper quartile) or mean (standard deviation), except for IgAN where individual data are provided, as there were only 2 VTEs.a Values presented are first available.b Albumin at time of VTE was not available for one patient. Open table in a new tab Abbreviations: FSGS, focal segmental glomerulosclerosis; IgAN, immunoglobulin-A nephropathy; MN, membranous nephropathy; NA, not applicable; TA, time averaged; VTE, venous thromboembolic event. Data are provided as median (lower quartile, upper quartile) or mean (standard deviation), except for IgAN where individual data are provided, as there were only 2 VTEs. The results of univariable analyses to identify clinical features associated with the development of VTE are shown in Table 4. The underlying histological diagnosis was closely associated with VTE risk; the risk was highest in patients with MN (unadjusted hazard ratio 22.0, 95% confidence interval 5.3–92.1, P<0.01) and FSGS (hazard ratio 7.8, 95% confidence interval 1.7–35.2, P<0.01) compared with IgAN (reference group). Male sex, proteinuria at presentation, TA proteinuria, albumin at presentation, and TA albumin were all associated with VTEs (see Table 4).Table 4Univariable analyses: clinical variables associated with risk of venous thromboembolic eventsHR95% CIP-valueUnderlying disease0.0001 IgANReference FSGS7.81.7–35.20.008 MN22.05.3–92.1<0.0001Male sex2.61.2–5.60.01Proteinuria at presentation by categories0.0003 8g/day17.02.3–127.60.006TA proteinuria by categories0.0002 8g/day25.33.2–198.70.002Albumin at presentation by categories 38g/lReference 29–38g/l4.30.5–37.10.2 <29g/l27.72.8–203.9 39g/lReference 35–39g/l3.41.3–8.90.01 <35g/l5.52.2–13.5 38g/lReference 29–38g/l2.70.3–23.90.4 <29g/l9.61.2–76.40.03Proteinuria at presentation0.7 8g/day2.60.3–20.20.4Underlying disease0.006 IgANReference FSGS5.91.3–27.90.02 MN10.82.4–49.40.002Abbreviations: CI, confidence interval; FSGS, focal segmental glomerulosclerosis; HR, hazard ratio; IgAN, immunoglobulin-A nephropathy; MN, membranous nephropathy. Open table in a new tab Abbreviations: CI, confidence interval; FSGS, focal segmental glomerulosclerosis; HR, hazard ratio; IgAN, immunoglobulin-A nephropathy; MN, membranous nephropathy. VTEs are a potentially life-threatening complication of the nephrotic syndrome. The risk of VTE has traditionally been thought to be highest in patients with MN; however, it is not clear whether this disease-specific risk is independent of clinical variables such as patient age or degree of proteinuria. In this study, we sought to determine whether the underlying histological diagnosis is an independent risk factor for the development of clinically evident VTE. We have shown that the risk of clinically evident VTE in patients with idiopathic GN is closely related to the underlying histological diagnosis, with the highest risk in MN, the lowest in IgAN, and an intermediate risk in FSGS. The higher frequency of VTEs was not uniquely attributable to a higher frequency of RVT events; a higher frequency of VTEs was observed in patients with MN even when RVTs were excluded. In our cohort, when considered in isolation, both proteinuria and hypoalbuminemia increased the risk of VTE; however, the disease-specific risk of VTE was independent of the degree of proteinuria, serum albumin levels, and cancer history. Although patients with MN were more likely to have higher levels of proteinuria and lower albumin levels, as well as high rates of cancer, these differences did not fully account for differences in VTE risk. Data regarding the disease-specific risk of VTE in patients with the nephrotic syndrome are divergent. Historically, MN has been associated with a particularly high risk of VTE. This was supported by studies that used screening renal venograms, and a substantial proportion of VTEs captured in the patient population were asymptomatic RVTs; in these studies, the risk of VTE was reported to be as high as 86%.5.Llach F. Hypercoagulability, renal vein thrombosis, and other thrombotic complications of nephrotic syndrome.Kidney Int. 1985; 28: 429-439Abstract Full Text PDF PubMed Scopus (323) Google Scholar, 6.Llach F. Koffler A. Finck E. et al.On the incidence of renal vein thrombosis in the nephrotic syndrome.Arch Intern Med. 1977; 137: 333-336Crossref PubMed Scopus (62) Google Scholar, 7.Llach F. Papper S. Massry S.G. The clinical spectrum of renal vein thrombosis: acute and chronic.Am J Med. 1980; 69: 819-827Abstract Full Text PDF PubMed Scopus (182) Google Scholar, 9.Chugh K.S. Malik N. Uberoi H.S. et al.Renal vein thrombosis in nephrotic syndrome—a prospective study and review.Postgrad Med J. 1981; 57: 566-570Crossref PubMed Scopus (51) Google Scholar, 11.Wagoner R.D. Stanson A.W. Holley K.E. et al.Renal vein thrombosis in idiopathic membranous glomerulopathy and nephrotic syndrome: incidence and significance.Kidney Int. 1983; 23: 368-374Abstract Full Text PDF PubMed Scopus (118) Google Scholar However, the clinical impact of these asymptomatic VTEs is unclear. Furthermore, not all studies uniformly found a similarly high frequency of VTE in MN.3.Glassock R.J. Prophylactic anticoagulation in nephrotic syndrome: a clinical onundrum.J Am Soc Nephrol. 2007; 18: 2221-2225Crossref PubMed Scopus (161) Google Scholar In addition, previous studies did not adjust for proteinuria or albumin level; therefore, the question of the independent disease-specific risk of VTE remained unanswered.5.Llach F. Hypercoagulability, renal vein thrombosis, and other thrombotic complications of nephrotic syndrome.Kidney Int. 1985; 28: 429-439Abstract Full Text PDF PubMed Scopus (323) Google Scholar, 11.Wagoner R.D. Stanson A.W. Holley K.E. et al.Renal vein thrombosis in idiopathic membranous glomerulopathy and nephrotic syndrome: incidence and significance.Kidney Int. 1983; 23: 368-374Abstract Full Text PDF PubMed Scopus (118) Google Scholar, 12.Kendall A.G. Lohmann R.C. Dossetor J.B. Nephrotic syndrome. A hypercoagulable state.Arch Intern Med. 1971; 127: 1021-1027Crossref PubMed Scopus (194) Google Scholar The largest study to date, to address disease-specific risk included 298 patients with nephrotic-range proteinuria, of whom 157 had idiopathic primary GN.10.Mahmoodi B.K. ten Kate M.K. Waanders F. et al.High absolute risks and predictors of venous and arterial thromboembolic events in patients with nephrotic syndrome: results from a large retrospective cohort study.Circulation. 2008; 117: 224-230Crossref PubMed Scopus (254) Google Scholar In this study, there was no observed difference in the incidence of VTE between MN, FSGS, and minimal change disease, and only proteinuria above 8.2g/day was associated with increased VTE risk.10.Mahmoodi B.K. ten Kate M.K. Waanders F. et al.High absolute risks and predictors of venous and arterial thromboembolic events in patients with nephrotic syndrome: results from a large retrospective cohort study.Circulation. 2008; 117: 224-230Crossref PubMed Scopus (254) Google Scholar Our study findings differ likely because of the fact that our large sample size of over 1300 patients may be better powered to detect disease-specific differences in risk. Our study results prompt two important questions for future consideration. First, why is MN independently associated with a particularly high VTE risk? Next, how could these data inform decisions regarding prophylactic anticoagulation? The mechanisms responsible for the increased VTE risk require further study and elucidation. The nephrotic syndrome itself is purported to cause changes in hemostatic factors that favor thrombosis, including increased fibrinogen and coagulation factor levels and decreased antithrombin III, protein C and S, and plasminogen levels.4.Bellomo R. Atkins R.C. Membranous nephropathy and thromboembolism: is prophylactic anticoagulation warranted?.Nephron. 1993; 63: 249-254Crossref PubMed Scopus (96) Google Scholar However, these findings are inconsistent, and less is known about mechanisms by which MN may specifically place patients with nephrotic syndrome at higher risk of VTE. Given that the absolute quantity of proteinuria may not independently explain the VTE risk, the nature of the proteinuria may differ between types of GN. The particular molecular weight of proteins lost in patients with MN may result in disease-specific alterations in proteins that affect susceptibility to thrombosis.13.Bazzi C. Petrini C. Rizza V. et al.Characterization of proteinuria in primary glomerulonephritides: urinary polymers of albumin.Am J Kidney Dis. 1997; 30: 404-412Abstract Full Text PDF PubMed Scopus (14) Google Scholar Antibodies to α-enolase have been observed in patients with MN, and these may have antifibrinolytic activity.14.Wakui H. Imai H. Komatsuda A. et al.Circulating antibodies against alpha-enolase in patients with primary membranous nephropathy (MN).Clin Exp Immunol. 1999; 118: 445-450Crossref PubMed Scopus (64) Google Scholar,15.Lopez-Alemany R. Longstaff C. Hawley S. et al.Inhibition of cell surface mediated plasminogen activation by a monoclonal antibody against alpha-Enolase.Am J Hematol. 2003; 72: 234-242Crossref PubMed Scopus (89) Google Scholar There is also a suggestion that factor V Leiden mutation may be associated with MN, providing a second hypercoagulable risk factor in addition to the nephrotic syndrome.16.Elinav E. Rubinger D. Hiller N. et al.Renal vein thrombosis and membranous glomerulopathy in a patient homozygote for factor V Leiden mutation: a mere coincidence?.Thromb Haemost. 2006; 95: 740-743PubMed Google Scholar A relationship between antibodies to the M-type phospholipase receptor (PLA2R) and VTE risk also requires further exploration.17.Beck Jr, L.H. Bonegio R.G. Lambeau G. et al.M-type phospholipase A2 receptor as target antigen in idiopathic membranous nephropathy.N Engl J Med. 2009; 361: 11-21Crossref PubMed Scopus (1503) Google Scholar The next logical stage in analysis is an evaluation of the risks and benefits of prophylactic anticoagulation in patients with GN. Our data suggest several characteristics that may identify patients with proteinuric GN who may be at highest risk of thrombosis, including male gender, significant hypoalbuminemia, and an underlying histological diagnosis of MN. The timing of VTE in MN appeared to be earlier than in those with FSGS (see Table 2), supporting the need for additional analyses to identify disease-specific time periods of highest thrombosis risk. These could then be used in a formal decision analysis to help guide the use of prophylactic anticoagulation. Several considerations regarding how our study design may have affected our findings merit comment. Patients did not undergo predefined screening tests to diagnose venous thrombotic events. It is therefore possible that patients with more severe nephrotic syndrome or those with a diagnosis of MN may have been more aggressively investigated for VTEs because of their higher perceived risk. One might expect this to have the most impact with respect to detection of RVT; indeed, VTE frequency was disproportionately high even when RVTs were excluded from the analysis. Next, there was a trend suggesting earlier thrombotic events in patients with MN, which needs to be considered in the context of our inclusion of patients with at least 1 year of follow-up in our analysis. We chose 1 year of follow-up in order to assure adequate time to detect events across all GN subtypes, and to accrue sufficient clinical data to identify risk factors associated with VTE. This may have biased our results toward diagnosing fewer events, potentially underestimating the increased risk of VTE in MN. We verified that this was not likely to be a concern, given that only two VTEs were documented in patients with MN who had <1 year of follow-up (see Materials and Methods). In addition, an underestimation of the risk of events would not necessarily alter our findings of the risk factors associated with VTE. Finally, despite the large size of our cohort and the long follow-up period, there were relatively few events, which does limit the ability to identify predisposing risk factors. In conclusion, this underscores the rarity of clinically evident venous thrombotic events in this patient population and the need to identify those patients at highest risk to target clinical interventions. In summary, we have shown that patients with MN have a greater risk of VTE compared with those with FSGS or IgAN, even after adjustment for differences in degree of proteinuria, hypoalbuminemia, and malignancy. Further research is required to identify reasons for the disease-specific risk of VTE in patients with MN. Since 1974, all patients with biopsy-proven idiopathic GN in the greater Toronto area have been enrolled in the Toronto Glomerulonephritis Registry.18.Regional program for the study of glomerulonephritis. Central Committee of the Toronto Glomerulonephritis Registry.Can Med Assoc J. 1981; 124: 158-161PubMed Google Scholar Information is collected prospectively by registrars as of the time of first clinical presentation, and includes demographics, baseline and developing comorbidities, clinical and laboratory parameters, and medication use. We considered all patients in the Toronto GN Registry with idiopathic IgAN, FSGS, and MN for enrolment in our cohort. We excluded those patients with secondary causes of GN, who were younger than 16 years of age at presentation, had incomplete clinical data, or less than 12 months of follow-up. We included patients with a spectrum of disease severity, including those who had subnephrotic proteinuria, to more completely describe the risk of VTE associated with the underlying histological type of disease. The 12-month period was chosen to ensure that patients were followed up long enough to capture VTEs. There were 2790 patients considered for our cohort, of whom 1477 were excluded resulting in a cohort of 1313 patients, including 370 with FSGS, 548 with IgAN, and 395 with MN. Supplementary Figure S1 online shows the derivation of the cohort and reasons for exclusion. We determined that only two VTEs were detected in 94 MN patients with <1 year of follow-up, supporting the fact that restricting our analyses to patients with longer follow-up would not substantially underestimate the frequency of events. The start of the follow-up period was defined as the time of the first assessment with available clinical and laboratory data. This may have preceded the date of renal biopsy. Baseline parameters were taken as the first available within 6 months of the start of the follow-up period. Secondary causes of GN were defined as the presence of systemic lupus erythematosis, hepatitis B, hepatitis C, HIV, or other coexistent glomerular diseases on biopsy (such as diabetes). Proteinuria was measured in 24-h urine collections. The follow-up period was broken into 6-month blocks, the average proteinuria during each block was determined, and the mean of all such values was termed the TA proteinuria, and meant to reflect the burden of proteinuria during the follow-up period. Similarly, the TA albumin levels (TA albumin) were calculated. Creatinine clearance was estimated from the Cockroft-Gault formula and standardized to body surface area (ml/min per 1.73m2).19.Cockcroft D.W. Gault M.H. Prediction of creatinine clearance from serum creatinine.Nephron. 1976; 16: 31-41Crossref PubMed Scopus (13102) Google Scholar Race was self-reported. Comorbidities were prospectively collected by the registrars during routine chart reviews. We reviewed the longitudinal follow-up data of patients within the Toronto Glomerulonephritis Registry and extracted all venous thrombotic events. Routine screening for asymptomatic events was not performed, although some RVTs were detected on routine pre-biopsy ultrasound tests. Most events were clinically symptomatic, prompting further investigations, and were confirmed by imaging modalities. Pulmonary embolus was confirmed by pulmonary angiography, ventilation–perfusion scans, or computed tomography; deep vein thrombosis by ultrasound; and RVT by ultrasound, computed tomography, or percutaneous renal venograms. To ensure that cases were not missed, the records for all patients on warfarin were identified for repeat review to determine whether VTE may have been the indication for anticoagulation. Data were analyzed using Microsoft Excel and SAS software. Normally distributed variables were described as mean standard deviation and compared across groups using analysis of variance. Nonparametric variables were described as median (lower quartile and upper quartiles), and compared using the Kruskal–Wallis test. Categorical variables were compared using the χ2-test. All P-values are two tailed, with <0.05 considered statistically significant. The time to event was measured as the time from the start of the follow-up period to the documentation of a VTE. Differences in event-free survival were determined using the Cox proportional hazards test. The Cox proportional hazards test was used to identify clinical variables associated with the risk of VTE. Covariates considered included histological diagnosis, age, sex, self-reported race, cancer, proteinuria at presentation, TA proteinuria, albumin at presentation, and TA albumin. Because proteinuria is not normally distributed, it was analyzed as a log-transformed variable. All variables found to be significantly associated with VTE risk by univariable models were included in a multivariable model. However, proteinuria at presentation was found to be highly correlated with TA proteinuria; similarly, albumin at presentation was found to be highly correlated with TA albumin. Therefore, we chose to include proteinuria and albumin at presentation in our multivariable models. These were expressed as categorical variables to account for missing values, which were not associated with VTE risk (P=0.34 for missing albumin and P=0.31 for missing proteinuria). Multivariable analysis was repeated using TA albumin and TA proteinuria as covariates instead of values at presentation, with similar results (data not shown). We are grateful for the support of the nephrologists of the Greater Toronto Area for their ongoing participation in the Toronto Glomerulonephritis Registry. We thank registrars Naomi Ryan and Paul Ling for their diligent contributions to maintaining and updating the Registry data. HNR's work is supported by a KRESCENT New Investigator award, courtesy of the Kidney Foundation of Canada, the Canadian Society of Nephrology, and the Canadian Institute of Health Research. SJB is supported by the Clinician Investigator Program of the University of British Columbia, and the BC Renal Agency. Figure S1. Derivation of the cohort. Supplementary material is linked to the online version of the paper at http://www.nature.com/ki
Referência(s)