Evaluation of Clinical Outcomes and Renal Vascular Pathology among Patients with Lupus : Clinical Journal of the American Society of Nephrology

Journal Logo

Original Articles

Evaluation of Clinical Outcomes and Renal Vascular Pathology among Patients with Lupus

Barber, Claire*; Herzenberg, Andrew; Aghdassi, Ellie; Su, Jiandong; Lou, Wendy§; Qian, Gan§; Yip, Jonathan; Nasr, Samih H.; Thomas, David; Scholey, James W.**; Wither, Joan*,‡,††; Urowitz, Murray*,‡; Gladman, Dafna*,‡; Reich, Heather**; Fortin, Paul R.*,‡,‡‡

Author Information
Clinical Journal of the American Society of Nephrology 7(5):p 757-764, May 2012. | DOI: 10.2215/CJN.02870311
  • Free

Abstract

Introduction

Renal pathology in SLE is characterized by mesangial cell proliferation, inflammation, necrosis, basement membrane abnormalities, immune complex deposition, and vascular abnormalities (1,2). Renal vascular lesions (RVLs) have been classified in a number of studies (35); however, they are not included in the International Society of Nephrology/Renal Pathology Society (ISN/RPS) revised 2004 criteria for lupus nephritis (LN) (6). Although associations between RVLs and renal outcomes have been proposed (3,4,7), the literature is hampered by its largely retrospective nature, and it has been argued that, because no clear definitions exist for many of the described lesions, their prognostic significance is not well understood (3,5).

Our study describes the clinical characteristics at time of renal biopsy of patients with RVLs and elucidates the following information: are vascular lesions found on renal biopsy in patients with SLE associated with (1) increased mortality, (2) increased risk of extrarenal vascular events, or (3) poor renal outcomes?

Materials and Methods

Patients with renal biopsy data were selected from the database registry of the University of Toronto Lupus Clinic (UTLC). Since 1970, the UTLC has prospectively followed patients who fulfilled at least four of the 1971 or 1982 American College of Rheumatology classification criteria (8) or three criteria and had positive histology. All patients gave their informed consent and were followed according to a standard protocol, which has had continuous approval from the University Health Network Research Ethics Board.

Renal biopsies (207), performed as clinically indicated, from 164 patients with SLE were identified from 1970 to 2007, reviewed by two independent pathologists, and scored based on the ISN/RPS revised 2004 criteria for LN (6); a third pathologist was used for consensus scoring if disagreement occurred. Only one biopsy per patient was included in the study. Three patients with overlapping lupus vasculopathy (LV) and thrombotic microangiopathy (TMA) were excluded to avoid confounding, leaving 161 patients for analysis. RVLs were defined according to the classification described in the work by Appel et al. (1,5). (1) LV: necrotizing changes in the vessel wall associated with abundant immune deposits causing luminal narrowing or occlusion that are often positive for fibrin, immunoglobulin, and complement with absence of inflammatory cells. (2) TMA: a luminal narrowing and occlusion by accumulation of eosinophilic and fuchsinophilic material with staining for fibrin with absence of discrete immune, histologically identical to hemolytic uremic syndrome–thrombotic thrombocytopenic purpura (TTP/HUS) lesions. (3) Arterial Fibrinoid Necrosis (lupus vasculitis): the small- and medium-sized arteries are affected; there is a prominent inflammatory cell infiltrate with mural inflammation and fibrinoid necrosis resembling microscopic polyangiitis. (4) Uncomplicated vascular immune deposits (UVIDs): lesions with vascular immune deposits that, when visualized by light microscopy, reveal that, despite the vessels appearing normal, immune complex deposits are present in the walls of arterioles and to a lesser extent, in the veins; no thrombosis or inflammatory infiltrate is present, and immunofluorescence is positive for immunoglobulins and complement. (5) Arterial sclerosis (AS) and arteriolar hyalinosis: control patients were selected who had renal biopsies without evidence of RVL.

SLE Disease Activity and Damage

Disease activity was determined using the SLE disease activity index 2000 (SLEDAI-2K) within 3 months of renal biopsy (9). The Systemic Lupus International Collaborating Clinics/American College of Rheumatology damage index (10) was used to assess cumulative damage at the time of the relevant biopsy or within 6 months.

Renal Variables

Estimated GFR (eGFR) was calculated using the Modification of Diet in Renal Disease Study Group equation (11). ESRD was defined as serum creatinine of ≥200 μmol/L and/or eGFR ≤15 ml/min per 1.73 m2 and/or dialysis for >6 months or having a kidney transplant (12). A secondary renal outcome was CKD as defined according to Kidney Disease: Improving Global Outcomes guidelines (12,13) as a sustained eGFR <60 ml/min per 1.73 m2.

Vascular Risk Factors

Age, mean arterial pressure (MAP), smoking history, body mass index, blood lipids, history of diabetes, and use and cumulative dose of steroids were recorded at the time of biopsy. Arterial vascular events (composite variable) included myocardial infarction, angina, cerebrovascular accident, and transient ischemic attacks.

Outcomes Analyses

To explore associations between renal vascular findings and clinical manifestations of disease, a subset of the cohort of the 161 patients was selected if they had a period of observation in the UTLC database covering 8 years prebiopsy and 7 years post-biopsy. This 15-year time window was chosen to maximize our use of the registry data, and it allowed a sufficient pre- and post-biopsy time period for events to occur and have biologic plausibility. Methodologically, it anchored the period of observation to a comparable period within an individual patient’s time frame and between patients. This subset included 133 patients with a renal biopsy before December 31, 2003.

Demographic information, including ethnicity, and medication profile were also retrieved. Presence of antiphospholipid antibody syndrome was recorded according to Sydney criteria but without the test for the anti–β2-glycoprotein 1 antibody assay, because it is not available (14).

Statistical analysis was performed using SAS (version 9.2; SAS institute, Cary, NC) statistical software. When comparing two groups for demographics and baseline characteristics, the t and the Wilcoxon rank sum tests were used for continuous variables, and the chi-squared and Fisher exact tests were used for categorical variables. Four series of univariate and multivariate models based on logistic regressions were constructed to evaluate associations between the variables assessed and clinical outcomes. Specifically, for each clinical outcome, its association with each of the variables, such as RVL, patient demographic features, and baseline renal characteristics, was first examined, and variables that were significant and clinically relevant were considered for inclusion in the multivariate models, which were then assessed according to goodness of fit criteria (15). Given the small number of events, the exact method was used (16), and continuous variables in the model were dichotomized using clinically relevant cutoffs. Both the exact and mid-P values were reported as recommended in the literature for analyzing data with small sample sizes; the mid-P value was provided as a sensible way to overcome the conservatism in the exact method because of discreteness (17,18). A sensitivity analysis was conducted for each model by considering four scenarios for those patients who died or were lost to follow-up: (1) both cases had no event; (2) the former (died) had no event, and the latter (lost to follow-up) had an event; (3) the former had an event, and the latter had no event; and (4) both cases had an event. In cases of death, sensitivity analyses were conducted only for scenarios 3 and 4.

Results

In this study, 207 biopsies from 164 patients were examined. Three patients who had both TMA and LV on renal biopsy were excluded to avoid confusion. The present analysis is based on 161 patients: 13 (8.1%) patients had TMA, 5 (3.1%) patients had LV, 10 (6.2%) patients had UVID, and 0 patients had lupus vasculitis. AS alone was frequent, occurring in 93 patients (57.8%). Lesions of mild to moderate AS were seen in 27 (96.4%) patients with TMA, LV, or UVID (only one case of TMA had concomitant severe AS); 40 patients (24.8%) had no RVL and therefore, were used as controls (Figure 1).

F1-11
Figure 1:
Proportion of cases of renal vascular lesions distributed among the renal biopsies based on the class of lupus nephritis. There was no significant difference between classes of nephritis with respect to the relative distribution of renal vascular lesions. AS, arterial sclerosis; LV, lupus vasculopathy; TMA, thrombotic microangiopathy; UVID, uncomplicated vascular immune deposits.

Demographic Characteristics

The characteristics of subjects at the time of renal biopsy are reported in Table 1. There was no difference between TMA and control subjects with respect to gender distribution, but AS and LV patients were older. Lupus disease activity at time of biopsy was higher in the LV group compared with controls. All subgroups of patients with RVLs had higher percentages of people with damage (Systemic Lupus International Collaborating Clinics/American College of Rheumatology damage index score ≥1) than controls, except the UVID group (Table 1); 0 of 13 patients with TMA had a diagnosis of TTP/HUS at the time of renal biopsy or during the course of follow-up.

T1-11
Table 1:
Demographic features of SLE patients with renal vascular lesions compared with controls (n=161)

Renal Variables

Renal variables are reported in Table 2. At the time of renal biopsy, the eGFR was lower in patients with AS and LV compared with controls. The majority of the patients had proliferative LN in proportions that were similar among all groups; 10 patients had no evidence of LN on biopsy (5 patients had only AS lesions, and 2 patients had only TMA) (Table 2). The chronicity score was higher in patients with all RVLs groups compared with controls, except the UVID group.

T2-11
Table 2:
Baseline renal and cardiovascular characteristics in patients with and without renal vascular lesions on renal biopsy

Cardiovascular Risk Factors

The presence of cardiovascular disease risk factors is reported in Table 2. The MAP was significantly higher in all patient groups with RVLs. The remainder of the risk factors was not different between groups. In total, 140 patients (86.9%) had at least one available anticardiolipin antibody or lupus anticoagulant test on record ever. There were no significant differences in the number of patients who met criteria for the antiphospholipid antibody syndrome (2 control [5.0%], 2 TMA [15.4%], 0 LV, 0 UVID, and 4 AS [4.3%] patients).

Outcomes: Univariate and Multivariate Analyses

Four models were generated to determine if RVLs were predictive of death, arterial vascular events, or renal outcomes on a subset of 133 patients with available follow-up. For the models studying arterial vascular events, the association was tested over a period of 15 years (8 years before and 7 years after the biopsy). A sensitivity analysis was conducted for each model by considering four scenarios, as described above, for those patients who died or were lost to follow-up for each outcome.

A total of 20 patients died within 7 years of follow-up: 2 patients in the control group (2/34, 5.9%), 11 patients in the AS group (11/76, 14.5%), 3 patients with TMA (3/10, 30.0%), 2 patients with LV (2/4, 50.0%), and 2 patients with UVID (2/9, 22.2%). A total of 16 patients were lost to follow-up during this period of time. On univariate regression analysis, RVLs were not associated with increased mortality, and this finding did not change with sensitivity analysis. On multivariate analysis, higher SLEDAI-2K at the time of renal biopsy was associated with mortality, and the use of angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers was protective (Table 3).

T3-11
Table 3:
Associations between clinical features and death, CKD, and arterial vascular events

A total of 14 patients had systemic arterial vascular events within this 15-year period: 2 patients in our control group (2/34, 5.9%), 4 patients in the AS group (4/76, 5.3%), 3 patients in the TMA group (3/10, 30.0%), 2 patients in the LV group (2/4, 50.0%), and 3 patients in the UVID group (3/9, 33.3%). On sensitivity analysis, considering scenario 1, we assumed that, if a patient died or was lost to follow-up within 7 years of renal biopsy, then the patient had no thromboembolic event. On univariate analysis, the association between RVL groups (for RVL types and control) and thrombotic events was significant (Fisher exact test, P=0.002; exact conditional scores test, P=0.006), with the greatest group difference occurring between AS and UVID (odds ratio [OR; 95% confidence interval (CI)=8.57 [1.02, 65.71], P=0.047 and 8.57 [1.33, 51.87], mid-P=0.03); none of the individual lesions (RVL types) were marked differently from the controls for thrombosis. On sensitivity analysis, the results from scenario 2 were similar to the results from scenario 1; the results from scenario 3 were similar to the results from scenario 4. When patients who died during follow-up were all assumed to have had an arterial vascular event (scenario 3), the presence of TMA was associated with arterial vascular events (OR [95% CI]=5.5 [1.0, 34.9]). On multivariate analysis under scenario 1, the association between RVL groups and arterial vascular events was significant (exact conditional scores test, P=0.04), and the difference between AS and UVID was most noticeable (OR [95% CI]=8.35 [0.98, 83.12], P=0.05); compared with the control group, there is no significant association for individual lesion type (Table 3). Additionally, patients with ≥4.4 g/d proteinuria at the time of renal biopsy, which represents the upper third quartile in our cohort, were more likely to have a systemic arterial vascular event (Table 3).

A total of 71 patients had CKD within 7 years postrenal biopsy, including 17 patients in the control group (17/34, 50.0%), 42 patients with AS (42/76, 55.3%), 5 patients with TMA (5/10, 50.0%), 2 patients with LV (2/4, 50.0%), and 5 patients with UVID (5/9, 55.6%). On univariate analysis, none of the RVLs were associated with CKD on sensitivity analysis (data not shown). In multivariate analysis, at the time of renal biopsy, an MAP >100 was associated with CKD, and a higher eGFR was protective. None of the RVLs were independently associated with CKD. Our sensitivity analysis did not alter these findings.

A total of five patients had ESRD within 7 years after renal biopsy, including one patient in the AS group (1/76, 1.3%), three patients in the TMA group (3/10, 30.0%), one patient in the LV group (1/4, 25.0%), and zero patients in the UVID or control groups. We could not perform additional analysis, because the number of outcomes was too small.

Discussion

To our knowledge, this study is the first to link RVLs to prospectively followed and recorded clinical outcomes in a cohort of patients with SLE. This study shows that RVLs, especially AS, are prevalent in patients with nephritis. Although RVLs may not be independently associated with renal outcomes or mortality, they are associated with an increased risk of systemic arterial vascular events. Interestingly, two variables unrelated to RVLs were significant in our multivariate model on mortality: use of an ACE inhibitor within 1 year of biopsy was associated with a strong protective effect, whereas a higher disease activity measured by SLEDAI at the time of biopsy was a predictor of increased mortality at 7 years.

Longitudinal observational studies such as our study are limited by attrition, and data in the study are dependent on capture, recording, and reviewer’s biases if not blinded to the hypothesis. The work by Gladman et al. (19) has previously reported that 11% of the UTLC is lost to follow-up, which is similar to our findings (12%). Because renal biopsies were conducted as clinically necessary and not part of a standard protocol, the analyses may be biased by indication, and the renal outcome correlations may have been obscured by lesions other than vascular lesions. However, we believe that these limitations reflect real-life practice, and contrasting the presence or lack of presence of vascular lesions in nephritis remains an interesting comparison. Our observations are also limited by large CIs because of a small number of observations (20), and in general, our study has limited power to make conclusions about subjects with LV. Because the numbers of vascular events were small, we were not able to comment on any association between type of vascular event (e.g., arterial or venous) and type of lesion, and we were unable to examine the association between RVLs and ISN/RPS subclasses of nephritis. AS lesions occurred on 96% of the biopsies that had other RVL lesions, and this finding may have confounded our results. The AS observed in all but one case with LV or TMA was of mild or moderate severity and could be compared with a group of patients with AS without TMA or LV. Lastly, other vascular and thrombophilic risk factors, such as clotting factor deficiencies, are not captured.

RVLs in SLE were first described in autopsy series (21,22). However, clinical outcomes were first described in the work by Banfi et al. (4), which retrospectively analyzed 285 renal biopsies; 27.7% had RVLs (9.5% with LV, 8.4% with TTP/HUS-like changes, 7.0% with AS, and 2.8% with vasculitis). The prevalence of RVLs and more specifically, AS in our study was much higher. However, in line with the findings of our present study, patients with RVLs had higher serum creatinine and were more likely to be hypertensive compared with controls, whereas degree of proteinuria was similar between the various groups.

In contrast to previous reports (3,4), we found no association between the proliferative subtypes of lupus nephritis and the type of RVLs. This finding may be partially explained by the introduction of the ISN/RPS (6) classification criteria for lupus nephritis and may have resulted in reclassification of some subjects. The chronicity and activity indices in our patients were also instructive. Unlike previously published data (3,4), we found that renal biopsies with RVLs did not differ from controls in their activity scores, although they did have higher chronicity scores in keeping with previous reports (3,4).

In our study, five patients in the RVL group and zero patients in the control or UVID groups developed ESRD over 7 years of follow-up. The work by Descombes et al. (3) reported that patients with arteriosclerotic changes or LV on biopsy had similar renal outcomes compared with those patients without RVLs; however, the latter group received more immunosuppressive medications, suggesting more aggressive disease. This finding was not felt to be the case in our cohort. In contrast, the work by Banfi et al. (4) reported that the probability of renal survival was decreased in patients with RVLs (hazard ratio=5.49 [95% CI=2.6, 11.3]); however, the cohort had more significant renal failure at the time of biopsy. All studies concluded that survival is not predicted by the presence of RVLs.

Another novel finding of our study was the demonstration of an increased risk of systemic arterial vascular events in patients with RVL. The precise magnitude of this effect is difficult to appreciate because of wide CIs. This is a finding that warrants exploration in future studies. The increased risk of vascular events may be partially explained by an increased prevalence of hypertension in patients with RVLs.

This study is the first of patients followed in a prospective cohort with RVLs on biopsy. The presence of any RVLs was not independently associated with renal outcomes or overall survival in our cohort, although we had limited power to examine LV outcomes. The mortality at 7 years post-biopsy was decreased in those patients prescribed an ACE inhibitor within 1 year of their biopsy and increased in those patients with a higher lupus disease activity at the time of biopsy. The presence of RVLs was found to be associated with hypertension and diminished renal function at time of renal biopsy, and these two variables were the only significant predictors of poor renal function in the multivariate analysis. We found an association between RVL and systemic arterial vascular events in the absence of other traditional risk factors for cardiovascular disease and independently of proteinuria.

Disclosures

None.

Acknowledgments

This study was funded by an operating grant from the Canadian Institutes of Health Research (QNT 78341) entitled “LuNNET—The Lupus Nephritis New Emerging Team.” Canadian Network for Improved Outcomes in SLE (CaNIOS) is supported in part by Lupus Canada, Lupus Ontario, the Lupus Foundation of Ontario, and BC Lupus, as well as the Arthritis and Autoimmune Research Centre Foundation. The Centre for Prognostic Studies in Rheumatic Disease—University of Toronto Lupus Clinic is supported in part by The Smythe Foundation, Lupus Ontario, the Dance for the Cure, the Flare for Fashion, and the Lupus Foundation of Ontario. J.W. is supported in part by the Arthritis Centre for Excellence. P.R.F. is a Distinguished Senior Investigator of the Arthritis Society with additional support from the Arthritis Centre of Excellence, University of Toronto, and Université Laval, Quebec City, Quebec, Canada. C.B. is supported by a the UCB-CRA-TAS Post-Graduate Rheumatology Fellowship Award (UCB, Canadian Rheumatology Association, and The Arthritis Society) and an Alberta Innovates Health Solutions Clinical Fellowship.

Present address: Dr. Claire Barber, Division of Rheumatology, Department of Medicine, University of Calgary, Calgary, Alberta, Canada.

Present address: Dr. Paul R. Fortin, Division of Rheumatology, Centre Hospitalier Universitaire de Québec et Université Laval, Quebec, Quebec, Canada.

Published online ahead of print. Publication date available at www.cjasn.org.

References

1. D'Agati V: Renal Disease in Systemic Lupus Erythematosus, Mixed Connective Tissue Disease, Sjögren's Syndrome and Rheumatoid Arthritis. Heptinstall's Pathology of the Kidney, 6th Ed., edited by Jennette JC, Olson J, Schwartz MM, Silva F, Philadelphia, Lippincott Williams & Wilkins, 2007, pp 518–592
2. Seshan SV, Jennette JC: Renal disease in systemic lupus erythematosus with emphasis on classification of lupus glomerulonephritis: Advances and implications. Arch Pathol Lab Med 133: 233–248, 2009
3. Descombes E, Droz D, Drouet L, Grünfeld JP, Lesavre P: Renal vascular lesions in lupus nephritis. Medicine (Baltimore) 76: 355–368, 1997
4. Banfi G, Bertani T, Boeri V, Faraggiana T, Mazzucco G, Monga G, Sacchi G: Renal vascular lesions as a marker of poor prognosis in patients with lupus nephritis. Gruppo Italiano per lo Studio della Nefrite Lupica (GISNEL). Am J Kidney Dis 18: 240–248, 1991
5. Appel GB, Pirani CL, D’Agati V: Renal vascular complications of systemic lupus erythematosus. J Am Soc Nephrol 4: 1499–1515, 1994
6. Weening JJ, D’Agati VD, Schwartz MM, Seshan SV, Alpers CE, Appel GB, Balow JE, Bruijn JA, Cook T, Ferrario F, Fogo AB, Ginzler EM, Hebert L, Hill G, Hill P, Jennette JC, Kong NC, Lesavre P, Lockshin M, Looi LM, Makino H, Moura LA, Nagata MInternational Society of Nephrology Working Group on the Classification of Lupus NephritisRenal Pathology Society Working Group on the Classification of Lupus Nephritis: The classification of glomerulonephritis in systemic lupus erythematosus revisited. Kidney Int 65: 521–530, 2004
7. Yao GH, Liu ZH, Zhang X, Zheng CX, Chen HP, Zeng CH, Li LS: Circulating thrombomodulin and vascular cell adhesion molecule-1 and renal vascular lesion in patients with lupus nephritis. Lupus 17: 720–726, 2008
8. Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ, Rothfield NF, Schaller JG, Talal N, Winchester RJ: The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 25: 1271–1277, 1982
9. Gladman DD, Ibañez D, Urowitz MB: Systemic lupus erythematosus disease activity index 2000. J Rheumatol 29: 288–291, 2002
10. Gladman D, Ginzler E, Goldsmith C, Fortin P, Liang M, Urowitz M, Bacon P, Bombardieri S, Hanly J, Hay E, Isenberg D, Jones J, Kalunian K, Maddison P, Nived O, Petri M, Richter M, Sanchez-Guerrero J, Snaith M, Sturfelt G, Symmons D, Zoma A: The development and initial validation of the Systemic Lupus International Collaborating Clinics/American College of Rheumatology damage index for systemic lupus erythematosus. Arthritis Rheum 39: 363–369, 1996
11. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth DModification of Diet in Renal Disease Study Group: A more accurate method to estimate glomerular filtration rate from serum creatinine: A new prediction equation. Ann Intern Med 130: 461–470, 1999
12. Levey AS, Eckardt KU, Tsukamoto Y, Levin A, Coresh J, Rossert J, De Zeeuw D, Hostetter TH, Lameire N, Eknoyan G: Definition and classification of chronic kidney disease: A position statement from Kidney Disease: Improving Global Outcomes (KDIGO). Kidney Int 67: 2089–2100, 2005
13. Levey AS, Atkins R, Coresh J, Cohen EP, Collins AJ, Eckardt KU, Nahas ME, Jaber BL, Jadoul M, Levin A, Powe NR, Rossert J, Wheeler DC, Lameire N, Eknoyan G: Chronic kidney disease as a global public health problem: Approaches and initiatives—a position statement from Kidney Disease Improving Global Outcomes. Kidney Int 72: 247–259, 2007
14. Miyakis S, Lockshin MD, Atsumi T, Branch DW, Brey RL, Cervera R, Derksen RH, DE Groot PG, Koike T, Meroni PL, Reber G, Shoenfeld Y, Tincani A, Vlachoyiannopoulos PG, Krilis SA: International consensus statement on an update of the classification criteria for definite antiphospholipid syndrome (APS). J Thromb Haemost 4: 295–306, 2006
15. Hosmer DW, Lemeshow S: Applied Logistic Regression, 2nd Ed., New York, Wiley, 2000, p 373
16. Mehta CR, Patel NR: Exact logistic regression: Theory and examples. Stat Med 14: 2143–2160, 1995
17. Hirji KF: A comparison of exact, mid-P, and score tests for matched case-control studies. Biometrics 47: 487–496, 1991
18. Agresti A: Exact inference for categorical data: Recent advances and continuing controversies. Stat Med 20: 2709–2722, 2001
19. Gladman DD, Koh DR, Urowitz MB, Farewell VT: Lost-to-follow-up study in systemic lupus erythematosus (SLE). Lupus 9: 363–367, 2000
20. Greenland S, Schwartzbaum JA, Finkle WD: Problems due to small samples and sparse data in conditional logistic regression analysis. Am J Epidemiol 151: 531–539, 2000
21. Klemperer P, Pollack AD, Baehr G: Pathology of disseminated lupus erythematosus. Arch Pathol (Chic) 32: 569–631, 1941
22. Grishman E, Venkataseshan VS: Vascular lesions in lupus nephritis. Mod Pathol 1: 235–241, 1988
Copyright © 2012 The Authors. Published by Wolters Kluwer Health, Inc. All rights reserved.