You are on page 1of 5

Rheumatology 2007;46:14871491 Advance Access publication 5 August 2007

doi:10.1093/rheumatology/kem180

Causes and predictors of death in South Africans with systemic lupus erythematosus
S. Wadee1, M. Tikly1 and M. Hopley2
Objectives. Little is known about the long-term outcome and mortality patterns in systemic lupus erythematosus (SLE) in sub-Saharan Africa. We undertook a retrospective study of SLE in mainly black, unemployed patients, seen at a tertiary institution in Soweto, South Africa, to determine the causes and predictors of death. Methods. Demographic, clinical and laboratory data and outcome were extracted from the case records of patients attending the Lupus Clinic at Chris Hani Baragwanath Hospital. Results. Of the 270 case records with a diagnosis of SLE, 226 met the American College of Rheumatology classification criteria for SLE. The female to male ratio was 18 : 1. The mean (S.D.) age at presentation was 34 (12.5) yrs. Arthritis, nephritis and neuropsychiatric disease had a cumulative frequency of 70.4, 43.8 and 15.9% of patients, respectively. During the course of a mean follow-up period of 54.9 months, 193 (85.3%) and 89 (39.3%) patients were treated with oral corticosteroids and immunosuppressive agents, respectively. There were 55 (24.5%) known deaths and 64 (28.6%) patients were lost to follow-up. The estimated 5 yr survival rates were between 57 and 72%, depending on whether the group of patients lost to follow-up was classified in the analysis as either alive or dead. Infection (32.7%) was the commonest cause of death followed by renal failure (16.4%). Univariate analysis revealed that nephritis, neuropsychiatric disease and hypocomplementaemia were associated with an increased mortality, but multivariate analysis showed nephritis as the only significant predictor of mortality. Conclusion. Our findings suggest that SLE in indigent South Africans not only carries a poorer prognosis but also the main cause of death, infection and renal failure differ from those reported recently in industrialized Western countries. Nephritis is common in our patients and is the only independent predictor of poor outcome.
KEY
WORDS:

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on October 31, 2012

Lupus, Mortality, Africa, Blacks.

Introduction
Systemic lupus erythematosus (SLE) is a chronic multisystem autoimmune disorder that has a predilection for young women. Advances in the early serological detection of antinuclear antibodies coupled with discovery of a variety of immunosuppressive agents, including corticosteroids, have had a major positive impact on the outcome of the disease over the last 50 years [1]. In addition, progress in several other areas of clinical medicine, including diagnostic imaging, intensive care services, dialysis and transplantation and antimicrobial agents, have contributed to the reduction and mortality associated with SLE [2, 3]. Since the mid-1990s, several centres from the industrialized world have reported 5-yr survival rates in excess of 90% [47], compared with an appalling 40% in 1956 [8]. Notwithstanding these achievements, standardized mortality rates for SLE exceed the general population 2.4-fold in the industrialized world [9]. The outlook for SLE patients in the developing world is less optimistic. Studies from India, Tunisia, Senegal, Thailand and Curacao have found that, even in the 1990s and early 21st century, 5-yr survival rates are well below 90% [6, 1014]. Moreover, there are regional differences in the patterns and causes of death in SLE. In industrialized countries, causes of death are to some extent a function of disease duration. Early mortality, defined as a death occurring within 5 yrs of disease onset, is mainly related to disease activity or infections. Late mortality, occurring beyond 5 yrs disease duration, is frequently due to malignancy or cardiovascular disease [6, 15]. In contrast, studies from the developing world indicate that infections and active disease, particularly renal
1 Division of Rheumatology, Department of Medicine, Chris Hani Baragwanath Hospital and University of the Witwatersrand, Johannesburg, South Africa and 2 Boehringer-Ingelheim, South Africa.

disease, are the major causes of death and that the early mortality is higher than that in the industrialized world [10, 13]. One of the enigmas of SLE in sub-Saharan Africa is that while SLE is extremely rare in tropical Africa, it is, paradoxically, more common and carries a worse prognosis in Africans and people of African extraction living outside of the tropics [16]. This has led to the tropical gradient hypothesis [17] which postulates that certain tropical infections, especially malaria, alter the immune response and protect against auto-immune disease [18]. In the absence of any recent large studies on outcome and causes of death in SLE in sub-Saharan Africa, we undertook a retrospective study to investigate causes, patterns and predictors of mortality in a mainly urbanized unemployed cohort of SLE patients attending a tertiary care facility in South Africa. The study was approved by the Human Ethics Committee of the Faculty of Health Sciences, University of the Witwatersrand.

Patients and methods


Chris Hani Baragwanath hospital (CHBH) is a tertiary referral centre servicing mainly the indigent, predominantly black population of Soweto and surrounding peri-urban and rural areas of South Africa. Case records of patients fulfilling the 1997 updated American College of Rheumatology (ACR) classification criteria [19] and attending the Lupus Clinic at CHBH between January 1986 and July 2003 were reviewed. Patients who met only 3 of the 11 classification criteria were classified as having lupus-like disease. Demographic data, clinical and laboratory features (as defined in the ACR criteria), drug therapy and outcome were abstracted from the case records. Antinuclear antibodies (ANA) and anti-dsDNA antibodies were tested by indirect immunoflourescence testing using HEp2 cell line and Crithidia luciliae as substrates, respectively. Causes of death were ascertained by review of case records, by discussion with the attending physician and post-mortem reports, where available. Outcome was categorized as known deaths, known alive and unknown outcome for patients lost to follow-up.
1487

Submitted 31 October 2006; revised version accepted 7 June 2007. Correspondence to: M. Tikly, FRCP, PhD, Rheumatology Unit, Chris Hani Baragwanath Hospital, PO Bertsham 2013, South Africa. E-mail: tiklym@medicine.wits.ac.za

The Author 2007. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

1488 The main clinical or pathological process resulting in recorded as the primary cause of death. Survival was from the date of first presentation to the Lupus admission to hospital with clinical features of lupus to or until death or loss to follow-up.

S. Wadee et al. death was calculated Clinic or July 2003, significantly higher death rate in the SLE group compared with the lupus-like group (24.3 vs 11.4%, respectively, P 0.09). In the SLE group, the commonest clinical features at presentation were arthritis (50.9%) and malar rash (46.5%); nephritis was present in 34.1% of patients (data not shown). Cumulative frequencies of the ACR clinical and serological criteria are summarized in Table 2. Of the 99 patients who met the ACR criteria for nephritis, 82 patients had renal biopsies. The initial histological classes were as follows: class I in one, class II in 12, class III in 30, class IV in 14, class V in 23 and class VI (endstage renal disease) in two patients. In 16 of the patients with class II, III or IV disease, additional class V changes were observed. Three patients underwent repeat biopsies all of whom showed class switches, one from class II to class III, one from class V to class III and one from class III to class V. Corticosteroids (CS) and immunosuppressive agents were prescribed in 193 (85.3%) and 89 (39.3%) patients, respectively (Table 2). More specifically, higher dose CS (HD-CS), defined as a maximum daily dose >15 mg, were prescribed in 141 (62.4%), cyclophosphamide (intravenous only in all, except three) in 64 (28.3%), azathioprine in 49 (21.8%), methotrexate in 13 (5.8%), cyclosporine in six (2.7%) and mycophenelate mofetil in one (0.4%) of the patients. The mean follow-up period was significantly longer for patients known to be alive compared with those known to be dead (82.2 months vs 46.7 months, respectively, P < 0.0001). Univariate analysis showed that nephritis, neuropsychiatric disease and hypocomplementaemia were significantly more common and haematological abnormalities less frequent in the patients known to have died compared with those known to be alive and those lost to follow-up. No significant differences were observed between the subgroups with respect to thrombocytopaenia. Multivariate analysis demonstrated that only nephritis was associated with death [OR 2.1 (95% CI 1.13.8), P < 0.001]. No significant association was observed with specific WHO histological classes of nephritis.

Statistical methods
Univariate analysis was performed applying either the one-way ANOVA test in the case of continuous variables or chi-square test (with Yates correction) for categorical variables. KaplanMeier statistic was applied to construct the survival curves. Multivariate analysis was done using the Cox proportional hazard regression model to test for independent predictors of death. All statistical analyses were performed using Statistica v6 software (StatSoft, Tulsa, USA). A P-value <0.05 was considered to be statistically significant.

Results
Of the 270 patient records with a diagnosis of SLE, 226 fulfilled the ACR classification criteria for SLE and the remaining 44 had lupus-like disease. Their demographic and outcome data are summarized in Table 1. In the SLE group, the mean age of 38.9 yrs at presentation in males was higher than the 33.7 yrs in females, but this difference failed to reach statistical significance. Compared with the SLE group, the lupus-like group had a significantly shorter follow-up period (54.9 months vs 27.3 months, respectively, P < 0.0001). Half the patients in the lupus-like group were lost to follow-up, compared with 28.3% in the SLE group (P < 0.01). There was also a trend towards a

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on October 31, 2012

TABLE 1. Demographic data of SLE and lupus-like groups of South African patients SLE n 226 Male, Female Ethnic groups Black Mixed race Asian Caucasian Mean (S.D.) age (yrs) Mean (S.D.) follow-up (months) Known deaths (%) Lost to follow-up (%)

lupus-like n 44 1.15 37 1 5 1 35.9 (13.4) 27.3 (31.3) 5 (11.4) 22 (50)

Total n 270 1.16 247 8 12 3 34.3 (12.7) 54.2 (48) 60 (22.2) 86 (31.9)

1.18 210 7 7 2 34 (12.5) 59.4 (49) 55 (24.3) 64 (28.3)

Causes of death
In the SLE group, cause of death was ascertainable in 40 of the 55 known deaths. In the balance of cases, a cause of death could not be determined either because the patients died at home or because the death certificate did not specify a cause of death. Infections were the primary cause of death in 18 (45%) and a contributory cause in an additional six (15%) of the remaining 40 deaths (Table 3), with acute infections accounting for the vast majority of the cases (Table 4). In two of the three tuberculosis (TB) patients,

P < 0.0001 vs lupus-like group;

P < 0.001 vs SLE group.

TABLE 2. Cumulative frequency of clinical and laboratory features and drug therapy in 226 South African SLE patients ACR criteria/feature Mean (S.D.) age (yrs) Mean (S.D.) follow-up (months) Malar rash (%) Discoid lupus (%) Oral ulcers (%) Photosensitivity (%) Serositis (%) Arthritis (%) Neuropsychiatric disease (%) Nephritis (%) Haematological abnormalities (%) Thrombocytopaenia (%) Positive ANA (%) Immunological criteria (%) Anti-dsDNA antibodies (%) Anti-Sm antibodies (%) Antiphospholipid antibodies (%) Hypocomplementaemia (%) Oral corticosteroids Immunosuppressive agents

Total n 226 33.9 12.5 59.4 49 132 (58.4) 94 (41.5) 87 (38.5) 88 (38.9) 41 (18.1) 159 (70.4) 36 (15.9) 99 (43.8) 118 (52.2) 29 (12.8) 224 (99.1) 179 (79.2) 125 (55.3) 92 (40.7) 61 (27) 147 (65) 193 (85.3) 89 (39.3)

Known dead n 55 34 14.3 46.7 43.3 34 (61.8) 19 (34.5) 28 (50.9) 20 (36.3) 12 (21.8) 35 (63.6) 14 (25.4) 35 (63.6) 21 (38.2) 6 (10.9) 55 (100) 45 (81.8) 33 (60) 23 (41.8) 16 (29.1) 44 (80) 52 (94.6) 21 (38.2)

Known alive n 107 33.7 11.1 82.2 46.3 60 (56.1) 47 (43.9) 39 (36.4) 37 (34.6) 16 (14.9) 83 (77.5) 12 (11.2) 39 (36.4) 65 (60.7) 12 (11.2) 105 (98.1) 83 (77.5) 61 (57) 41 (38.3) 32 (29.9) 67 (62.6) 91 (85.1) 51(57.3)

Lost to follow-up n 64 34.4 13.3 32.3 39.5 38 (59.3) 28 (43.8) 20 (31.3) 31 (48.4) 13 (20.3) 41 (64.1) 10 (15.6) 25 (39.1) 32 (50) 11 (17.2) 64 (100) 51 (79.6) 31 (48.4) 28 (43.8) 13 (20.3) 36 (56.3) 50 (78.1) 17 (26.6%)

P-value NS 0.0001 NS NS NS NS NS NS 0.03 0.002 0.01 NS NS NS NS NS NS 0.03 NS NS

Known alive vs known dead.

Lupus in South Africans a multidrug-resistant organism was cultured. Moreover, TB was a contributory factor in one death where the primary cause of death was a dilated cardiomyopathy. Renal failure in 16.4% of patients was the second commonest known cause of death. HIV infection contributed to only two infection-related deaths, from TB and acute pneumonia. Of the seven deaths in the miscellaneous group, one was due to a myocardial infarct, three were from dilated cardiomyopathy, two were from pulmonary hypertension and one was pregnancy related. No deaths due to malignancy were recorded. The causes of early death did not differ significantly from those that caused late death (Table 3). Five deaths occurred in the lupus-like group, three due to infection, one due to renal failure and in one case the cause was unknown. The KaplanMeier survival curves for the SLE group demonstrated an overall 5 yr survival probability of 57% and in the best case scenario, censored for patients lost to follow-up, the figure rose to 72% (Fig. 1). Survival was worse in patients

1489

with nephritis at presentation, but nephritis at any stage in the course of the disease was also associated with a poorer survival (P 0.002) (Fig. 2). The subgroup of infection-related deaths, when compared with the known alive group, were more likely to have had CNS disease (33.3 vs 11.2%, OR 4.0, P 0.03), less likely to be treated with azathioprine (5.9 vs 33.6%, OR 0.12, P 0.02.), and showed a trend towards being more likely to be treated with HD-CS (77.7 vs 56.7%, OR 2.7, P 0.12). No significant associations were observed with the use of other immunosuppressive drugs, in particular cyclophosphamide and infection-related deaths.

Discussion
Our findings suggest that the outcome of SLE in indigent South Africans continues to be considerably poorer than in Western industrialized countries. The best-case scenario 5-yr survival rate of 72% is not much different from the 68 and 78% reported in two smaller studies in 1973 and 1988, respectively [20, 21]. One of the likely biological reasons for this poorer outcome is that the disease is inherently more aggressive, and therefore carries a higher mortality in blacks, which has been demonstrated in AfricanAmericans [22, 23]. In particular, nephritis is more common in blacks and adversely affects outcome [24, 25]. In the present study, 43.8% of patients had nephritis, considerably higher than the 19.4% reported in Europeans [26]. More importantly, the 5-yr survival rate of 60% in patients with nephritis was significantly lower than the 84% 5-yr survival rate for patients without nephritis. Moreover, multivariate analysis showed that nephritis was the only independent predictor of poor outcome and renal failure was the second commonest cause of death. We were unable to investigate the role of socio-economic conditions and geographical place of residence (urban vs rural), but studies in the US have shown that these factors impact on outcome [27]. Inadequacies of state health services (a legacy of the previous apartheid system), cultural factors and the cost of travel across large distances are all factors that are likely to contribute to the delay in early detection of SLE and hence outcome. The problem of late presentation is particularly evident in the lupus nephritis patients, 77 of 99 (78%) of whom had clinically overt renal disease at first presentation. Furthermore,
Chi2 = 12.4, df = 3, p = 0.002 Known deaths Still attending clinic 1.0

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on October 31, 2012

TABLE 3. Causes of death in patients with SLE Cause of death Infection Renal failure Active disease Other Acute cardiovascular Cardiomyopathy Pulmonary circulatory Pregnancy related Unknown Total (%) <5 yrs 13 6 4 5 1 2 1 1 12 40 (72.7) >5 yrs 5 3 2 2 0 1 1 0 3 15 (27.3) Total (%) 18 9 6 7 (32.7) (16.4) (10.9) (12.7)

15 (27.3) 55

TABLE 4. Details of infective causes of death Type of infection Sepsis/Septicaemia unspecified Pneumonia unspecified Tuberculosis PCP pneumonia Meningitis Post-surgical sepsis Pyomyositis Total <5 yrs 4 4 1 1 1 1 1 13 >5 yrs 2 1 2 0 0 0 0 5 Total 6 5 3 1 1 1 1 18

Known deaths 1.0 0.9 Cumulative Proportion Surviving 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 24 48 72

Still attending clinic

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1


No nephritis Nephritis at presentation Nephritis during follow-up

Cumulative Proportion Surviving

96

120

144

168

192

0.0 0 24 48 72 96 120 144 168 192 Follow-up in months

Survival Time in months


FIG. 1. KaplanMeier survival curves for total SLE group, censored for patients lost to follow-up.

FIG. 2. KaplanMeier survival curves comparing SLE patients with and without nephritis, censored for patients lost to follow-up.

1490

S. Wadee et al. M.T. has received consultancy fees, honoraria and/or research support from Pfizer, Roche, Merck, Aspreva and Bristol Myers Squibb. M.H. is employed by Boehringer-Ingelheim, South Africa.

the under-resourced state sector has limited facilities for renal replacement therapy, including dialysis and transplantation. Infections, like in other emerging countries [10, 11], were the commonest cause of death, irrespective of disease duration. Studies by Ansell et al. [28] and more recently by Whitelaw et al. [29] have shown that even with specialized care in ICU, infections are associated with a high mortality in SLE in South African state hospitals. Of note in the present study is that three of the 18 infection-related deaths (17%) were due to TB, similar to the observations of Mody et al. [30] in hospitalized SLE patients in Kwa-Zulu Natal. There is mounting evidence that patients with SLE are at increased risk for TB especially, extra-pulmonary disease [31]. In a recent retrospective review, we found that 13.6% of all SLE and lupus-like patients at our hospital contracted TB during the course of their follow up, of whom 26.4% had extrapulmonary disease [32]. In a country where HIV infection is endemic, we found only two patients with HIV-related deaths, one of whom had TB. Bearing in mind that this study covers a period in the late 1980s and early 1990s when HIV infections were less prevalent, it is likely that the impact of the HIV epidemic is likely to grow. Patients with SLE are at increased risk of infections due to both intrinsic factors, such as functional asplenia and complement abnormalities, and drug therapy, in particular CS and immunosuppressive agents, especially cyclophosphamide [33]. In the present study, the only significant risk factor for infection-related deaths was CNS disease and there was a trend towards an association with HD-CS use. The protective effect of azathioprine observed in this study probably relates to its steroid-sparing effect of the drug, although we did not specifically examine the cumulative doses of CS in patients who had received azathioprine vs patients who had not received the drug. We were unable to discern a bimodal pattern of causes of death. Of particular note is that there were no late deaths attributable to atherosclerotic cardiovascular disease and cancer, which are now the leading causes of late deaths in SLE in industrialized countries. Some of the possible reasons for this difference include a low prevalence of both conditions in the general black population [34], relatively short follow-up period, relatively young age and relatively few male patients. An obvious limitation of this study is the large number of patients (28.3%) lost to follow-up and our inability to establish a cause of death in 27.7% of the known deaths. Both these issues may have an impact on the interpretation of the results. It has been suggested that interpretation survival data is problematic when loss to follow-up rates exceed 20% [6]. We were also unable to quantify the cumulative dose of CS and other immunosuppressive agents, which may have a bearing on the risk of infectionrelated deaths. Other limitations include the variation in clinical expertise of the clinicians and standard of care over the study period and the absence of socio-economic data and global disease activity scores to relate to outcome. Notwithstanding these limitations, we believe our findings provide further evidence that lupus is not a rare disease in South Africa, that survival is not as good as in industrialized countries, and that causes of death resemble those documented in other developing countries. There is an obvious need to improve early detection of disease, particularly nephritis, to reduce mortality.

References
1 Abu-Shakra M, Gladman DD, Urowitz MB. Mortality studies in SLE: how far can we improve survival of patients with SLE. Autoimmun Rev 2004;3:41820. 2 Uramoto K, Michet CJ Jr, Thumboo J, Sunku J, OFallon WM, Gabriel SE. Trends in the incidence and mortality of systemic lupus erythematosus. Arthritis Rheum 1999;42:4650. 3 Alzeer AH, Al-Arfaj A, Basha SJ et al. Outcome of patients with systemic lupus erythematosus in intensive care unit. Lupus 2004;13:53742. 4 Cervera R, Khamashta MA, Font J et al. Morbidity and mortality in systemic lupus erythematosus during a 5-year period. A multicenter prospective study of 1,000 patients. European Working Party on Systemic Lupus Erythematosus. Medicine 1999;78:16775. 5 Stahl-Hallengren C, Jonsen A, Nived O, Sturfelt G. Incidence studies of systemic lupus erythematosus in Southern Sweden: increasing age, decreasing frequency of renal manifestations and good prognosis. J Rheumatol 2000;27:68591. 6 Moss KE, Ioannou Y, Sultan SM, Haq I, Isenberg DA. Outcome of a cohort of 300 patients with systemic lupus erythematosus attending a dedicated clinic for over two decades. Ann Rheum Dis 2002;61:40913. 7 Doria A, Iaccarino L, Ghirardello A et al. Long-term prognosis and causes of death in systemic lupus erythematosus. Am J Med 2006;119:7006. 8 Dubois EL. Prednisone and prednisolone in the treatment of systemic lupus erythematosus. J Am Med Assoc 1956;161:42733. 9 Bernatsky S, Boivin JF, Joseph L et al. Mortality in systemic lupus erythematosus. Arthritis Rheum 2006;54:25507. 10 Houman MH, Smiti-Khanfir M, Ben Ghorbell I, Miled M. Systemic lupus erythematosus in Tunisia: demographic and clinical analysis of 100 patients. Lupus 2004;13:20411. 11 Malaviya AN, Chandrasekaran AN, Kumar A, Shamar PN. Systemic lupus erythematosus in India. Lupus 1997;6:690700. 12 Ka MM, Diallo S, Kane A et al. Systemic lupus erythematosus and lupus syndromes in Senegal. A retrospective study of 30 patients seen over 10 years. Rev Rhum Engl Ed 1998;65:4716. 13 Kasitanon N, Louthrenoo W, Sukitawut W, Vichainun R. Causes of death and prognostic factors in Thai patients with systemic lupus erythematosus. Asian Pac J Allergy Immunol 2002;20:8591. 14 Nossent JC. Course and prognostic value of Systemic Lupus Erythematosus Disease Activity Index in black Caribbean patients. Semin Arthritis Rheum 1993;23:1621. 15 Abu-Shakra M, Urowitz MB, Gladman DD, Gough J. Mortality studies in systemic lupus erythematosus. Results from a single center. I. Causes of death. J Rheumatol 1995;22:125964. 16 Molokhia M, McKeigue PM, Cuadrado M, Hughes G. Systemic lupus erythematosus in migrants from west Africa compared with Afro-Caribbean people in the UK. Lancet 2001;357:14145. 17 Bae SC, Fraser P, Liang MH. The epidemiology of systemic lupus erythematosus in populations of African ancestry: a critical review of the prevalence gradient hypothesis. Arthritis Rheum 1998;41:20919. 18 Clark IA, al-Yaman FM, Cowden WB, Rockett KA. Does malarial tolerance, through nitric oxide, explain the low incidence of autoimmune disease in tropical Africa? Lancet 1996;348:14924. 19 Hochberg MC. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1997;40:1725. 20 Jessop S, Meyers OL. Systemic lupus erythematosus in Cape Town. S Afr Med J 1973;47:2225. 21 Dessein PH, Gledhill RF, Rossouw DS. Systemic lupus erythematosus in black South Africans. S Afr Med J 1988;74:3879. 22 Walsh SJ, Algert C, Gregorio DI, Reisine ST, Rothfield NF. Divergent racial trends in mortality from systemic lupus erythematosus. J Rheumatol 1995;22:16638. 23 Alarcon GS, Friedman AW, Straaton KV et al. Systemic lupus erythematosus in three ethnic groups: III. A comparison of characteristics early in the natural history of the LUMINA cohort. LUpus in MInority populations: NAture vs. Nurture. Lupus 1999;8:197209. 24 Alba P, Bento L, Cuadrado MJ et al. Anti-dsDNA, anti-Sm antibodies, and the lupus anticoagulant: significant factors associated with lupus nephritis. Ann Rheum Dis 2003;62:55660. 25 Bastian HM, Roseman JM, McGwin G Jr et al. Systemic lupus erythematosus in three ethnic groups. XII. Risk factors for lupus nephritis after diagnosis. Lupus 2002;11:15260. 26 Cervera R, Khamashta MA, Font J et al. Morbidity and mortality in systemic lupus erythematosus during a 10-year period: a comparison of early and late manifestations in a cohort of 1,000 patients. Medicine 2003;82:299308. 27 Walsh SJ, DeChello LM. Geographical variation in mortality from systemic lupus erythematosus in the United States. Lupus 2001;10:63746. 28 Ansell SM, Bedhesi S, Ruff B et al. Study of critically ill patients with systemic lupus erythematosus. Crit Care Med 1996;24:9814. 29 Whitelaw DA, Gopal R, Freeman V. Survival of patients with SLE admitted to an intensive care unit-a retrospective study. Clin Rheumatol 2005;24:2237.

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on October 31, 2012

Rheumatology key messages


 Infections and renal failure were the main causes of death in indigent South African patients with SLE.  Nephritis was the only independent predictor of death.

Lupus in South Africans


30 Mody GM, Parag KB, Nathoo BC, Pudifin DJ, Duursma J, Seedat YK. High mortality with systemic lupus erythematosus in hospitalized African blacks. Br J Rheumatol 1994;33:11513. 31 Sayarlioglu M, Inanc M, Kamali S et al. Tuberculosis in Turkish patients with systemic lupus erythematosus: increased frequency of extrapulmonary localization. Lupus 2004;13:2748.

1491

32 Hodkinson B, Tikly M. Tuberculosis in systemic lupus erythematosus patients in South Africa. Ann Rheum Dis 2006;65(Suppl II):198. 33 Kang I, Park SH. Infectious complications in SLE after immunosuppressive therapies. Curr Opin Rheumatol 2003;15:52834. 34 Seedat YK. Ethnicity, hypertension, coronary heart disease and renal diseases in South Africa. Ethn Health 1996;1:34957.

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on October 31, 2012