Abstract
The clinical expression of systemic lupus erythematosus (SLE) is influenced by genetic and environmental factors and therefore varies between ethnicities. Information on the epidemiology of SLE in Brazil is scarce and practically limited to studies conducted in socioeconomically developed regions (South and Southeast). The objective of this study was to describe the clinical and immunological aspects and outcome of a cohort of patients with SLE treated at a university hospital in northeastern Brazil and compare patterns related to age at onset: childhood (cSLE), adult (aSLE), and late (lSLE). A random sample of 414 records (women: 93.5%) were reviewed. The mean age at SLE onset and the mean disease duration were 28.9 ± 10.9 years and 10.2 ± 6.6 years, respectively. Most patients had aSLE (n = 338; 81.6%), followed by cSLE (n = 60; 14.5%) and lSLE (n = 16; 3.9%). The female/male ratio was 6.5:1 in cSLE and 16.8:1 in aSLE; in lSLE, all patients were female (p = 0.05). During follow-up, the cSLE group presented higher rates of nephritis (70% vs. 52.9% vs. 12.5%; p = 0.0001) and leuko/lymphopenia (61.7% vs. 43.8% vs. 56.2%; p = 0.02). No significant differences were found for anti-dsDNA, anti-Sm, and antiphospholipid antibodies. Treatment with immunosuppressants was significantly more common, and higher doses of prednisone were used, in cSLE. The prevalence of cardiovascular diseases were more frequent in lSLE (p = 0.03). No significant differences were found between the three groups with regard to mean damage accrual (SDI), remission, and mortality. Although cSLE presented higher rates of nephritis and leuko/lymphopenia, more frequent use of immunosuppressants and higher prednisone doses than aSLE and lSLE, the three groups did not differ significantly with regard to damage accrual, remission, and mortality.
Introduction
Systemic lupus erythematosus (SLE) is a multifactorial, chronic autoimmune disorder which may involve several organs and systems. Genetic, hormonal and environmental factors have been implicated in the etiopathology of SLE. Although onset may occur at any age, SLE most often affects women in childbearing age. Nevertheless, 15–20% of SLE patients are diagnosed during childhood or adolescence; 1 some authors have suggested the disorder is more aggressive at this age.2–5 The prevalence of late-onset SLE varies in the literature, probably because of ethnic factors and the use of different cutoff values in different studies. 6 Patients with late-onset SLE tend to have a poorer prognosis despite the lower incidence of severe organ involvement and the more benign clinical course, possibly because of age-related circumstances, such as a greater number of comorbidities and more extended exposure to cardiovascular risk factors. 6
Information on the epidemiology of SLE in Brazil is scarce and practically limited to studies conducted in the socioeconomically developed southern and southeastern regions of the country,7–12 or to large Latin American cohort studies.5,13 Brazil covers a vast expanse of land, with a diversity of climatic conditions and a high rate of racial miscegenation; thus, studies conducted in one region may not be representative of the country as a whole. SLE expression is known to be influenced by genetic and environmental factors and is therefore likely to vary among ethnic groups. The more homogeneous a study population is with regard to genetic, socioeconomic and environmental background, the greater the internal validity of the findings. Moreover, comparisons between studies conducted on cohorts with different backgrounds can help understand the disease and improve management.
The purpose of this study was to describe the clinical and immunological aspects and outcome of a cohort of patients with SLE treated at a university hospital in northeastern Brazil and compare patterns related to age at onset of SLE (childhood, adult, and late).
Materials and methods
Patients
Central electronic patient registration with ICD codes (International Statistical Classification of Diseases) is not available at our institution (Walter Cantídio University Hospital/HUWC, Federal University of Ceará, Brazil). Therefore, between February 2010 and December 2012 a task force of physicians and medical students conducted an active search for SLE patients attending the rheumatology, internal medicine, nephrology, or pediatrics services at any time. Patients hospitalized or attending HUWC outpatient services during the period of active search were also included. The information sources included (i) manual and electronic medical files for patients hospitalized or treated in the outpatient setting at any time, (ii) electronic databases of the central pharmacy at HUWC which dispenses medication for the treatment of SLE (antimalarials, azathioprine, mycophenolate mofetil, cyclophosphamide, rituximab), (iii) files of the hemodialysis and kidney transplantation services, and (iv) the principal author’s personal database (the principal author has conducted and supervised research involving SLE patients since the implementation of the HUWC rheumatology service in 1990). The search resulted in an initial sample of 830 records.
Due to the logistic difficulty of reviewing the medical records of 830 SLE patients, we decided to use only half the records in the in-depth analysis. The records were arranged in increasing numerical order (reflecting the order of hospital admission) and divided into sets of 50. The approach resulted in 16 sets of 50 records and 1 set of 30 records. In each set, the records were numbered from 1 to 50 (or from 1 to 30). Finally, half the records in each set were selected by drawing lots. To be included in the analysis, patients should be diagnosed with SLE according to the criteria of the American College of Rheumatology (ACR), 14 have no associated autoimmune disorder (rheumatoid arthritis, polymyositis, dermatopolymyositis, systemic sclerosis), and have been followed by the HUWC for at least one year or until their death. Records which did not meet the criteria for analysis were replaced by records drawn at random from the same group. The demographic and clinical data collected from the medical records were organized in standardized forms and transferred to an electronic database (Stata v. 9.0) for subsequent statistical analysis. The search in the medical records was conducted from January 2013 to September 2014. HUWC is a public university hospital and tertiary-level referral facility. Most users originated from socioeconomically underprivileged communities in Fortaleza (the state capital), the hinterland of Ceará, and other states in northeastern Brazil.
Study parameters
Demographic data (age at diagnosis, age at last evaluation, gender, race, and level of schooling), clinical data (SLE manifestations observed at any time during the course of the disease, time from diagnosis to the last assessment or death, and SLE medication administered at any time during the course of the disease), and immunological data indicating positivity for SLE at any time during the course of the disease (antinuclear antibodies (ANA), IgG/IgM anticardiolipin antibodies (aCL), lupus anticoagulant (LA), anti-dsDNA, and anti-Sm antibodies) were retrieved from the medical records using a standardized form. Lupus nephritis was diagnosed according to the following criteria: 24-hour proteinuria >500 mg and/or urinalysis with protein >2+ (associated with hematuria and/or cylindruria) in the absence of other causes and/or renal biopsy compatible with lupus nephritis. Patients were diagnosed with arterial hypertension when presenting at least two readings ≥130/85 mmHg or when using anti-hypertensive medication. Events were considered thrombotic when confirmed by the presence of a thrombus in any vascular site on imaging. Patients with ANA titers ≥1:160 (regardless of the pattern) and aCL titers ≥40 units of IgG and/or IgM were considered positive. Positivity for LA was confirmed with standard tests. Provided sufficient information was available, damage was assessed upon the patient’s last consultation using the Systemic Lupus International Collaborating Clinics/ACR damage index (SDI). 15 The following outcome parameters were considered in the study: chronic kidney disease (serum creatinine ≥1.2 mg/dL for >3 months), disease remission evaluated upon the last consultation (absence of any clinical manifestation or laboratory finding indicating active disease), disease activity upon the last patient assessment (presence of any clinical manifestation or laboratory finding compatible with active disease), and death.
Patients diagnosed with SLE were distributed into three groups according to age at diagnosis: childhood-onset SLE (cSLE), age ≤18 years; adult-onset SLE (aSLE), age >18 to 50 years; and late-onset SLE (lSLE), age >50 years. The cutoff age between childhood and adult onset (18 years) was based on a recent study. 16
The study protocol was approved by the HUWC Research Ethics Committee and filed under number 051.05.11.
The data were managed with the software Stata v. 9.0. Categorical variables were analyzed with the chi-square test, while continuous variables were submitted to one-way ANOVA. P-values less than or equal to 0.05 were considered significant.
Results
Demographic, clinical, and immunological characteristics of a cohort of patients with systemic lupus erythematosus
SD = standard deviation.
Total number of evaluated patients.
Clinical and immunological characteristics of a cohort of patients with systemic lupus erythematosus according to age at onset
cSLE = childhood-onset SLE; aSLE = adult-onset SLE; lSLE = late-onset SLE; SD = standard deviation. *Total number of evaluated patients.
Antimalarials were used by over 70% of our patients, with no significant difference between the groups. However, mycophenolate mofetil was used more frequently in cSLE (38.3%) than in aSLE (22.9%) or lSLE (0%) (p = 0.003). Cyclophosphamide pulse therapy was used by half the patients with cSLE, 37% of patients with aSLE, and 18.7% of patients with lSLE (p = 0.04). The use of azathioprine and methotrexate was similar in the three groups. Rituximab was only used by patients with aSLE (n = 4). Treatment with corticoids and antimalarials only was more frequent in lSLE (37.5%) than in aSLE (22.4%) or cSLE (16.6%), but the difference was not significant. The mean daily dose of prednisone at the last consultation was lower in lSLE than in cSLE or aSLE (5.5 ± 2.7 mg vs. 10.9 ± 17.4 mg vs. 9.4 ± 13.9 mg, respectively) (p = 0.005).
Among the observed comorbidities, cardiovascular diseases were significantly more common in lSLE (28.6%) than in aSLE (13.6%) or cSLE (5.1%) (p = 0.02). In the three cSLE patients affected with cardiovascular disease, the pathology manifested during adult life: congestive heart failure diagnosed at 30 years of age (diabetic, hypertense, and dyslipidemic patient), coronary artery disease diagnosed at 32 years of age (hypertense and dyslipidemic patient with chronic use of corticoids at ≥ 10 mg/day), and aortic and mitral valve disorder diagnosed at 38 years of age (severely hypertense patient with left ventricle hypertrophy). The four lSLE patients with cardiovascular disease experienced acute myocardial infarction (n = 2) or ischemic cerebrovascular accident (CVA) (n = 2). In the aSLE group, 44 of 322 evaluated patients were affected with rheumatic/congenital/functional valve disorder (n = 14), CVA (n = 11), coronary artery disease (n = 8), myocardiopathy (n = 7), heart failure (n = 5), pulmonary hypertension (n = 3), aortic or cerebral aneurism (n = 1 each), and congenital cardiopathy (n = 1). Some patients presented more than one type of cardiovascular disease. The groups did not differ significantly with regard to the incidence of antiphospholipid antibody (aPL) syndrome, hypothyroidism, thrombosis, neoplasia, arterial hypertension, and diabetes, although the latter tended to be more frequent in lSLE (13.3%) than in cSLE (0%) and aSLE (8.2%) (p = 0.06).
Thrombotic manifestations were observed in seven cSLE patients (11.6%): deep vein thrombosis (DVT) (n = 1), mesenteric artery thrombosis (n = 1), CVA (n = 1), pulmonary embolism (PE) (n = 1), portal vein thrombosis (n = 1), renal vein thrombosis (n = 1), and DVT + PE (n = 1). Five of these tested positive for aPL. The remaining two were not tested, but one had DVT associated with nephrotic syndrome while the other had renal vein thrombosis following renal transplantation. In the aSLE group, 41 of 338 evaluated patients (12.3%) displayed thrombotic manifestations: DVT (n = 20), CVA (n = 7), PE (n = 2), arteriovenous fistula thrombosis (n = 2), popliteal artery thrombosis (n = 3), renal vein thrombosis (n = 1), DVT + PE (n = 2), DVT + CVA + superior vena cava thrombosis (n = 1), mesenteric artery thrombosis + acute myocardial infarction (n = 1), and thrombosis of unspecified type (n = 2). Twenty-five of these 41 patients (61%) tested positive for aCL and/or LA, 5 (12.2%) tested negative for aCL and LA, 6 (14.6%) were submitted to only one of these tests, the result of which was negative, and 5 (12.2%) were not tested for aPL. Two of the five patients with thrombosis and negativity for aCL and LA presented nephrotic syndrome. In two other patients, thrombosis was central venous catheter-related (site of thrombosis). Two lSLE patients experienced thrombotic events (12.5%): one with DVT + CVA, aCL positivity and LA negativity, and one with CVA associated with cardiac arrhythmia and negativity for aPL. Malignancy was only observed in one cSLE patient (aged 15 years upon SLE diagnosis): cervical cancer diagnosed at 28 years of age and anal cancer at 37 years of age. The patient had been treated for lupus nephritis with cyclophosphamide pulse therapy at a cumulative dose of 16.6 g. Eleven aSLE patients had malignant lesions of the following types/locations: stomach (n = 1), lymphoma (n = 2), anus (n = 1), liposarcoma (n = 1), breast (n = 1), skin (n = 1), lung (n = 1), ovary (n = 1), and cervix (n = 2). Only one lSLE patient presented malignancy (uterine cancer at 57 years of age).
Clinical outcome of a cohort of patients with systemic lupus erythematosus according to age at onset
SLE = childhood-onset SLE; aSLE = adult-onset SLE; lSLE = late-onset SLE; SDI: Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index.
Total number of evaluated patients.
Discussion
The study population consisted of patients from a public tertiary-level referral center for SLE patients in a state capital of northeastern Brazil. Ours is the largest cohort of SLE patients from a single center in northeastern Brazil for which demographic, clinical, immunological, and outcome data are available. The predominance of non-white race and the overall low level of schooling reflected the poor socioeconomic status of our sample. The association between the development of SLE and unfavorable sociodemographic/environmental factors is supported by the greater prevalence of SLE in the South of the United States and other underprivileged areas.17,18
Demographic, clinical, and laboratory parameters of three Brazilian cohorts of patients with systemic lupus erythematosus
SD = standard deviation.
Malar rash/discoid rash.
There is no consensus in the literature regarding the cutoff age for cSLE and lSLE. In a meta-analysis published by Silva et al. in 2012, 16 the upper cut-off age for cSLE ranged from 14 to 20 years, and similar inconsistencies apply to lSLE. Such variations are likely to impact results and compromise comparison. Based on the published articles and issues pertaining to the provision of medical care, Silva et al. suggested 18 years as the cut-off age for cSLE and recommended adopting the term “childhood-onset SLE”. 16 Although no statistically significant differences have been observed between studies using different cut-off ages for lSLE, 26 in this study we preferred the most commonly adopted value (>50 years). In the present study, aSLE (81.6% of the cohort) was considerably more common than cSLE (14.5%) and lSLE (3.9%), matching results in the literature. Though relatively rare in our cohort, lSLE displays a highly variable frequency across cohorts worldwide (3.67–39.3%), as shown in a recent review. 6 cSLE accounts for ∼10–20% of all SLE patients.1,4,5,7,22,27 In the Latin American experience of the Grupo Latinoamericano de Estudio del Lupus (GLADEL), which included 230 patients with cSLE (considering a cutoff age of ≤18 years), the percentage was 18.%. 5 As expected, the female gender was predominant in all three groups, especially in aSLE and lSLE. Although the widely recognized female predominance of SLE during the reproductive years is less apparent in cSLE and lSLE,2,26–33 several studies have found similar gender ratios in all age groups.3,5,6,34
Nephritis and leuko/lymphopenia were significantly more frequent in cSLE. Hemolytic anemia and thrombocytopenia were also more common in cSLE, but the difference was not statistically significant. Although the literature is rather inconsistent with regard to the predominant clinical manifestations in cSLE vs. aSLE, many studies show nephritis to be more common in cSLE than in aSLE.2–4,22,27,35,36 In a study from Canada by Brunner et al. comparing 67 children and 131 adults with SLE, 2 the only significant difference was a higher nephritis rate in cSLE (78% vs. 52%; p = 0.0005). This is very close to the nephritis rates of the present study (70% vs. 52.9%; p = 0.0001), based on a sample of 60 children and 338 adults. A 2011 meta-analysis of clinical manifestations in cSLE vs. aSLE concluded that malar rash, nephritis, thrombocytopenia, hemolytic anemia, seizure, fever, and lymphadenopathy were more common in cSLE. 27 In another cohort of 28 subjects with cSLE (<16 years), 107 with aSLE (16-50 years), and 14 with lSLE (≥50 years), no significant difference in nephritis rate was observed after adjusting for baseline renal parameters, duration of nephritis, use of cyclophosphamide and initial response to treatment. 37 The inclusion in the aSLE group of subjects with onset of SLE between 16 and 18 years of age may in part explain this finding. Likewise, a Chinese study with 1790 SLE patients segregated by age at onset (<15 years, 15–50 years, and ≥50 years) revealed no age-related differences in clinical or laboratory parameters. 28 Studies segregating SLE patients into two age groups only (<50 years and ≥50 years) show lower rates of nephritis, less visceral involvement and a more benign course with less use of corticoids and immunosuppressants in the older group.6,29,38,39
Although the anti-dsDNA percentage was higher in cSLE than in aSLE or lSLE in our cohort (55.5% vs. 37.9% vs. 30%; p = 0.07), the difference was not statistically significant. The higher rates of anti-dsDNA detection among children in some studies may be due to higher nephritis rates in this age range.3,7,27,35,40 According to a meta-analysis by Livingston et al., 41 the percentage of IgG/IgM aCL is significantly higher in cSLE than in aSLE (OR = 1.66; 95% CI = 1.20–2.28). Due to the greater frequency of infectious diseases, transient non-pathogenic aCL appear to be more common in children. Antiphospholipid-related thrombotic events are rare at this age.
While some components of the SDI may be related to the use of corticoids (cataracts, avascular necrosis, osteoporosis, muscle atrophy/weakness) and cyclophosphamide (premature gonad failure, malignancy), some components may also be related to age (cataracts, diabetes mellitus, neoplasia, osteoporosis, stroke, and acute myocardial infarction). Although the mean SDI tends to increase over time, 42 a Canadian study found higher indices in cSLE, with corticoid-related toxicity as the main associated factor. 2 In our study, the three groups displayed similar SDI values.
The more frequent use of mycophenolate mofetil and cyclophosphamide in cSLE in the present study may be explained by the higher incidence of nephritis and greater need of therapy optimization. Other studies have also found corticoid and immunosuppressant use to be more frequent in cSLE.2,43
In our study, the only comorbidities which differed significantly between the three groups were cardiovascular diseases (more frequent in lSLE). This is not surprising since they are known to be age-related. According to other authors, while onset is more insidious in lSLE, 44 visceral involvement is smaller,23,27,30,38,39 disease activity is lower, 45 and the course is more benign.30,39 The greater frequency of age-related comorbidities and greater organ damage in elderly SLE patients (due to aging and longer exposure to classic cardiovascular risk factors) contributes to worsening the prognosis. The mortality rate was similar in our three age groups, but in a meta-analysis from 2012 the mortality risk was greater in cSLE than in aSLE. 41 Others have reported an association between disease activity and higher mortality rates in both cSLE and aSLE.46,47
Our study has a number of limitations. The use of a retrospective design, with analysis of medical records, may have underestimated the frequency of some clinical manifestations such as photosensitivity. Also, despite our best efforts and the long period covered by the active search (1974–2013), many SLE patients attending the HUWC may have escaped our attention due to the unavailability at our institution of central electronic patient registration with ICD codes. However, since the patient search involved all HUWC services treating SLE patients (including pediatrics), and the 414 medical records in the final sample were selected at random, we do not believe this limitation significantly affected the proportion of patients or disease-related clinical manifestations among the three subgroups by age at diagnosis. Moreover, considering the possibility that not all patients who died of lupus or lupus-related comorbidities (especially after a short course of disease) were identified in our search, the observed mortality rates may be underestimated. This potential bias compromised the calculation of standardized mortality; thus, only group mortality rates were analyzed in this study. Another limitation was the scarcity of records of disease activity employing well-established indices, making it impossible to make comparisons of disease activity between the groups based on such indices.
In conclusion, this is the largest cohort of SLE patients from a single center in Northeastern Brazil for which demographic, clinical, immunological and outcome data are available. Adult-onset SLE was almost six times more common than cSLE and 20 times more common than lSLE. When compared to the other two groups, cSLE affected a smaller proportion of females and displayed a higher rate of nephritis and leuko/lymphopenia, a greater frequency of immunosuppressant use and higher doses of corticoids. Cardiovascular diseases were more common in lSLE. Despite the higher rate of nephritis in cSLE and greater frequency of cardiovascular diseases in aSLE, prognosis with regard to clinical remission and mortality rates was similar in the three groups. The incidence of chronic kidney disease tended to be higher in cSLE. In view of the strongly negative impact of nephritis on the outcome and survival of SLE patients, patients with this manifestation should be diagnosed as early as possible and provided efficient treatment with minimal toxicity, especially children in view of their earlier exposure to SLE-related inflammation and adverse effects of corticosteroids and immunosuppressants. The higher rate of cardiovascular disease in aSLE points to the need for a better control of related risk factors in this patient population.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.