• epidemiology;
  • ethnicity;
  • geographic distribution;
  • systemic lupus erythematosus


  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix I: Tulup Study Group


Clinical features of systemic lupus erythematosus (SLE) have been described from different geographical regions around the world. However, data from North African countries, including Tunisia, are scarce.


The aim of this retrospective multicenter study was to analyze demographic, clinical, laboratory features and outcome of SLE in Tunisia throughout 14 Departments of Internal Medicine and to compare them with those of other ethnic and geographic groups.


Seven hundred and forty-nine cases of SLE were recorded (American College of Rheumatology criteria) during a 17-year period (1989–2006). They were 676 women and 73 men with an average age at SLE onset of approximately 30.66 years. Our Tunisian patients were characterized by a high frequency of photosensitivity (67.6%), malar rash (68.7%), renal involvement (49.5%) and anti-Sm antibodies (44.8%). Infections were the main complications. Fifty-six (7.5%) patients died during the study period.


Potential limitations and biases in our study need discussion. Specific recruitment of patients in tertiary referral centers may be the source of selection bias and adding to the frequency of moderate or even severe diseases. The therapeutic management and outcome monitoring were heterogeneous due to the fact that patients were evaluated by different doctors. However, this study remains the most representative of Tunisian SLE patients recruited from all parts of Tunisia.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix I: Tulup Study Group

Systemic lupus erythematosus (SLE) is a chronic multisystem autoimmune disease with a wide spectrum of clinical and immunological abnormalities.

Marked ethnic differences in the prevalence and severity of SLE have been reported.[1, 2] Although previously considered as a rare disease, SLE now appears to be relatively common in certain populations.

Cohort studies among North Africans and particularly Tunisian SLE patients are exceptional. To our knowledge, only four studies dealing with demographic, clinical and laboratory features of SLE Tunisian patients have previously been reported.[3-6] However, these studies focused on nonrepresentative groups of the Tunisian population. So, no systematic review can be made to determine the incidence and characteristics of SLE in Tunisia.

Therefore, TULUP (for TUnisian LUPus) aimed to analyze demographic, clinical, laboratory features and outcome of SLE in Tunisia through a multicentric cohort of Tunisian patients observed in 14 Departments of Internal Medicine, and to compare them with those of other ethnic and geographic groups.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix I: Tulup Study Group

Study population

The TULUP study is a collaborative effort between 14 Departments of Internal Medicine in Tunisia (Appendix I) sponsored by the Tunisian Society of Internal Medicine. The TULUP cohort consists of a retrospective chart of all consecutive SLE cases seen at these centers, fulfilling at least four of the revised American College of Rheumatology (ACR) criteria[7] and followed-up between 1989 and 2006. The patients included were diagnosed at any time during the study period; those diagnosed prior to this period were excluded for lack of information and heterogeneous methods of antibody detection. Patients transferred between these centers have been captured and data collected only once. Some patients had already been included in other Tunisian studies.[3-6]


Clinical, laboratory and management data of the SLE patients were recorded according to a pre-defined protocol, including variables from the following domains.

  • 1.
    Demographic features: age at onset, gender, delay of diagnosis, race, personal (including obstetric) and family histories.
  • 2.
    Clinical features: disease onset, clinical manifestations attributable to SLE at onset and at the course of the disease, ACR criteria, co-morbidities and treatments.

Disease onset was defined by the time when the first symptoms of SLE occurred as recalled by the patient and recorded retrospectively in the record. The main SLE clinical manifestations analyzed in this study were defined according to the American Rheumatism Association Glossary Committee.[8]

  • 3.
    Laboratory and Immunologic features.

Routine laboratory exams were performed at diagnosis in all cases, including blood count, blood creatinine and calculated creatinine clearance rate, 24-h proteinuria, urinary sediment (hematuria, leukocyturia), aspartate aminotransferase, alkaline phosphatase, total bilirubin and direct bilirubin, serum protein electrophoresis and erythrocyte sedimentation rate.

Nephrotic syndrome was defined as proteinuria > 3 g/day. Nephritic syndrome was defined as hematuria, hypertension, oliguria, edema and/or reduced creatinine clearance.

Antibodies were investigated in different laboratories. Anti-nuclear antibody sera were screened by immunofluorescence using Hep-2 cell line. Two methods were used to assess anti-double-stranded DNA (anti-dsDNA) by immunofluorescence against Crithidia luciliae and/or enzyme-linked immunoabsorbent assay (ELISA) procedure. Antibodies to extractable nuclear antigens (anti-Sm, Rnp, SSA and SSB), anti-cardiolipin and anti-β2 glycoprotein I were identified by a standardized ELISA or immunodot. Lupus anticoagulant was detected by measuring the effect of patients' plasma on the kaolin clotting time of normal plasma. Complement and rheumatoid factor were assessed by nephelometry.

Renal biopsy specimens in patients with lupus nephritis (in cases of proteinuria > 0.5 g/24 h, hematuria, leukocyturia and/or renal failure) were routinely processed for standard analysis by light (LM) and immunofluorescence microscopy (IFM). They were analyzed and classified according to the 1982 World Health Organization (WHO) classification for lupus glomerulonephritis.[9]

Two-dimensional echocardiography with measurement of pulmonary artery pressure was performed for the diagnosis of cardiac involvement in case of clinical symptoms. Respiratory manifestations were diagnosed by chest radiography, pulmonary function testing and chest computed tomography scan.

Neurological manifestations were investigated by the study of cerebrospinal fluid and magnetic resonance imaging brain and/or spinal cord in patients with neuro-psychiatric symptoms.

  • 4.
    Disease activity was determined using the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI).[10] SLEDAI was calculated at the SLE diagnosis, 6, 12 months after and at the last consultation.
  • 5.
    Outcome: disease duration, flares (defined by an increase in SLEDAI score and/or a new onset manifestation), complications, death and causes of death were assessed. Follow-up visits were performed at 3 ± 1 month intervals.

Only Charts having sufficient data were included in this study according to the same standard form.

Statistical analysis

Data were entered and analyzed using Epi info version 3.2.2 ( Descriptive analyses were performed using chi-square tests (or Fisher's test) for proportions and student's t-test for the parametric data.

The Kaplan–Meier analysis was performed to evaluate the cumulative survival in our patients.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix I: Tulup Study Group

A total of 781 records were analyzed. Only 749 records of Tunisian SLE patients were included in this study. Thirty-two records were excluded because of incomplete data and/or follow-up. All patients were of Tunisian descent. The patients originated from all parts of Tunisia (Fig. 1). Fourteen patients (1.8%) were Blacks, the remaining others were Caucasians or Asians.


Figure 1. Geographic distribution of Tunisian patients in this study.

Download figure to PowerPoint

Demographic characteristics of our patients are described in Table 1).

Table 1. Demographic characteristics of patients in this study
  1. a

    Mean disease duration between the onset symptom and the diagnosis.

Gender, female (%)676 (90.2)
Female-to-male ratio9.26
Mean age at SLE onset30.66 ± 13.31 years (2–74 years)
Delay of diagnosisa15.26 months

Familial history of autoimmune diseases was noted in 44 patients (5.8%). SLE was reported in first-degree relatives of 26 patients (3.5%). Predisposing factors were identified in 189 patients (25.3%) at disease onset. The most frequent were infections (40%), sun exposure (28%) and pregnancy (16%).

The presenting symptoms in 63% of patients (473 patients) were fever with arthralgia and skin manifestations (first symptoms reported at the onset of the disease).

Table 2 summarizes the frequencies of the most common initial manifestations at SLE onset.

Table 2. Frequencies of SLE initial manifestations in this study
Clinical manifestationsn (%)
Arthralgia353 (47.12)
Arthritis138 (18.42)
Skin involvement338 (45.12)
Fever268 (35.78)
Hematologic abnormalities159 (21.22)
Serositis147 (19.26)
Nephritis94 (12.5)
Neuropsychiatric manifestations35 (4.76)

Cumulative frequencies of SLE clinical manifestations seen at any time in the course of the disease are shown in Table 3.

Table 3. Cumulative frequencies of SLE clinical manifestations in this cohort
Clinical featuresn (%)
Skin involvement612 (81.7)
Photosensitivity507 (67.6)
Malar rash515 (68.7)
Discoid lupus89 (11.9)
Mouth ulcers175 (23.3)
Arthralgia653 (87.1)
Arthritis419 (55.9)
Myositis21 (2.8)
Cardiac involvement239 (31.9)
Pericarditis200 (26.7)
Myocarditis30 (4)
Endocarditis16 (2.1)
Pulmonary involvement194 (26)
Pleurisy173 (23)
Interstitial lung disease13 (1.7)
Pneumonia5 (0.66)
Alveolar haemorrhage3 (0.4)
Haematological involvement608 (81)
Lymphopenia464 (61.9)
Autoimmune anaemia550 (73.4)
Thrombocytopenia160 (21.4)
Nephritis371 (49.5)
Neuropsychiatric manifestations277 (37)
Seizure63 (8.4)
Psychosis130 (17.3)
Venous thrombosis50 (6.7)

Cumulative frequency of renal involvement was 49.5%. The most frequently reported abnormalities were proteinuria above 0.5 g/24 h (97%), hematuria (42%) and decreased creatinine clearance (79 mL/min: 28%). Among the 371 patients with presumed lupus nephritis, 264 (71%) underwent renal biopsy; in the other cases renal biopsy was not done because of contraindication. Histological examination of renal biopsies were assigned to WHO classes as follows: class I (4.1%), class II (14.2%), class III (25.8%), class IV (29.6%) and class V (9.7%). In the remaining cases, this histological examination showed mixed classes represented by membranous glomerulonephritis associated with mesangial (1.1%) or proliferative glomerulonephritis (7.1%). Interstitial nephritis was seen in 6.7% of cases. Associated renal vascular microthrombi were observed in 2.7% of cases.

Neuropsychiatric manifestations were reported in 277 patients (37%). The most common symptom was headache (72%). The central nervous system (CNS) was involved in 113 patients (40.7% of the neuropsychiatric manifestations). In these patients, seizures were the most frequent manifestation (69% of CNS involvement) followed by organic brain syndrome (53%), impaired consciousness (21%), aseptic meningitis (11%) and transverse myelitis (11%). Peripheral neuropathy and cranial nerve palsy were diagnosed, respectively, in 7.9% and 11% of cases.

The auto-antibodies profile among our Tunisian SLE patients is described in Table 4.

Table 4. Autoantibody profile of this SLE cohort
Auto-antibodiesn (%)
Antinuclear734/749 (98)
Anti-dsDNA579/749 (77.3)
Anti-Sm335/749 (44.8)
Anti Rnp341/749 (45.5)
Anti-Sm/Rnp226/749 (30.1)
Anti-SSA423/749 (56.4)
Anti-SSB244/749 (32.5)
Anti-cardiolipin477/749 (63.7)
Anti-β2 glycoprotein I300/749 (40)
Lupus anticoagulant311/483 (64.5)

One to four other autoimmune diseases were diagnosed in 317 cases. Table 5 shows the frequencies of these associated autoimmune diseases.

Table 5. Prevalence of associated autoimmune disease in this cohort
Autoimmune diseasen (%)
  1. a

    Others: associated autoimmune diseases consist of type I diabetes, myasthenia gravis, vitiligo and idiopathic thromobcytopenia purpura.

Sjogren's syndrome156 (20.8)
Antiphospholipid syndrome82 (10.9)
Scleroderma55 (7.3)
Rheumatoid arthritis31 (4.1)
Myositis35 (4.7)
Dermatomyositis5 (0.6)
Hashimoto's hypothyroidism50 (6.6)
Graves/hyperthyroidism14 (1.9)
Coeliac disease26 (3.5)
Othersa78 (10.4)

Past and current medications during the period of the study were as follows: antimalarials (84.4%), non-steroidal anti-inflammatory drugs (21%), corticosteroids (87.7%) and immunosuppressive drugs (26.3%). Average initial dose of corticosteroids was 0.7 ± 0.3 mg/kg/day as deemed appropriate by the treating physician. Immunosuppressive drugs were generally administrated in case of lupus nephritis (type III or IV) or neuropsychiatric involvement. Cyclophosphamide was the most frequently given as the first immunosuppressive treatment and then replaced by azathioprine or methotrexate.

The average total follow-up was of 48.3 ± 46 months; patients were followed every 3 ± 1 months. Partial or total remission was obtained in 33.8% of cases. Thirty-six percent of patients presented one or more flares during the follow-up period.

The mean SLEDAI at the time of diagnosis was 12.7 and after 6 and 12 months of survey it decreased, respectively, to 5.3 and 4.9.

Single to multiple episodes of infection (average of 5 ± 3 episodes) with various pathogens were recorded in 45.5% of patients during the course of SLE. Bacterial infection was predominant (60.6%), followed by viral (15.9%), fungal (12.6%) and parasitic (5%) infections. Tuberculosis occurred in 5.9% of cases. The most frequent bacterial infectious locations were pulmonary and urinary tract.

Fifty-six (7.5%) patients died during this period. They were 47 women and nine men with a mean age ± SD at death of 34.7 ± 12.6 years and mean disease duration of 4 years.

Most deaths were related to either SLE activity in 38.2% (renal involvement in 20.6%, neuropsychiatric involvement in 17.6% and cardiac complications in 16.2% of cases) or a superimposed infectious process (26.5%). Survival rates at 5, 10 and 15 years in our patients were respectively 85.3%, 81.9% and 80% (cf Fig. 2: Kaplan–Meier survival curve).


Figure 2. Kaplan–Meier survival curve.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix I: Tulup Study Group

This study characterizes the demographic, clinical and laboratory features of SLE in Tunisian patients based on the multicentric data of 749 cases referred to 14 Departments of Internal Medicine.

Incidence and prevalence of SLE in Tunisia as well as in other North African countries cannot be clarified for lack of national epidemiological registers.

The Tunisian population results from an important intermingling of different populations throughout history, mainly composed of Berbers, Arabs, Turkish and southern Europeans.

Regarding demographic features (Table 6), −11–21 the data of our patients were similar to those reported in other ethnic groups despite the frequency of familial history of SLE. Generally this frequency was about 10%[20] and in some studies high rates have been reported;[21] however, in our study only 3.5% of patients had relatives with SLE. Unfortunately, there are no data on familial SLE from other southern Mediterranean, African or Arab countries. Nevertheless, we thought that this frequency was likely underestimated as family history was not documented in all records.

Table 6. Review of the demographic, clinical and immunological SLE features in previously reported series
AuthorsOur seriesUthman[11]Alballah [12]Ka[13]Taylor[14]Tikly[27]Phan[15]Correa[16]Hopkinson[17]Alarcon[18]Cervera[19]Thumboo[29]
PopulationTunisiaLebanonSaudiSenegalZimbabweSouth AfricaVietnamColombiaEnglishHispanicAfro-AmericanCaucasianEuropeanMalayIndiaChinese
Number of patients7491008730311112314014770887110006328381
F/M676/7386/1478/930/030/1103/821/2 136/1166/478/1058/1392/90857/624/4354/27
Median age30.62528,5302835.138 35,134,235,442,829323031
Malar rash68.7%52%56% 61%55%61% 30%   26,4%64%26%63%
Discoid lupus11.9%19%18% 19%29%9% 10%   5,4%18%20%7%
Oral ulcers23.3%40%16%19%22%30%37%   8,9%26%25%19%
Serositis49.7%40%56,3%43%23%28%30%22%37%   12,9%22%19%21%
Positive ANA98%87%98%100%96%98,2%100% 97%98,6%97,7%97,4%96%90,5%91,7%93,4%
Positive anti-DNA77.3%50%93%78%100%66,2%57%71%54%44,3%40,9%21,1%78%86%85,8%85,6%
Anti-Sm antibodies44.8%40%44,2%36%33%3%4,5%9,5%4,4%10%34,5%27%26,1%
Anti-Ro antibodies56.4%   60,5%          
Anti-La antibodies32.5%   28,4%          
Positive anti-Rnp45.5%65,5%54%48%    13%50%39,1%38,1%
Positive anti-cardiolipin63.7%22%7%10,9%4,8%24%
Anti-β2 glycoprotein I40%             
Lupus anticoagulant64.5%             

The geographical distribution across Tunisia revealed a higher proportion of SLE patients in the north of the country (cf Fig. 1). This result may be due to the differences in the accessibility to university hospitals and in the population density according to the regions. In fact, the geographic distribution of patients is representative of the population distribution, since the Tunisian population is denser in the north, especially in the capital,and in the coastal areas.

Clinical features at the onset of the disease in our study were similar to those reported among other ethnic groups, and it is commonly reported that fever, articular manifestations and skin lesions are the most frequent presenting features.[2, 17]

The prevalence of major clinical features during the evolution of the disease in this cohort is comparable to that reported in previous studies (Table 6).

Photosensitivity (67.6%) and malar rash (68.7%) were relatively more frequent in our patients than in other ethnic groups.

The prevalence of photosensitivity in SLE is not uniform worldwide.[22] It appears most commonly in American,[18] European[17] and some Asiatic countries: Taiwan (90.9%),[23] Puerto Rico (76.9%),[24] Kuwait (48.8%)[25] and Pakistan (60%),[26] and less commonly in Black African countries.[13, 14, 27, 28]

Studies have shown that in most cases the lighter the skin type, the greater the prevalence of photosensitivity. Nevertheless, the discordance observed regarding these frequencies may be due to the fact that there are different definitions of photosensitivity in lupus.[22]

Similar frequencies of malar rash were noted in patients from South Africa,[14, 27] the Middle East[11, 12] and Asia.[15, 29] Therefore, these frequencies were greater than those observed in the European,[17, 19] Indian[29] and Senegalese[13] patients. We thought that this difference may be due to the higher levels of sunshine in Tunisia (long periods of sunshine, intense UVB) as well as in the Middle Eastern countries in comparison to Europe and the lesser dark skin of Tunisian and Asian peoples compared to Senegalese or Indians. It is possible that melanin has a protective role but we cannot confirm this hypothesis because of the scarcity of SLE studies concerning Black Africans. Alternatively, different antibody profiles may explain these differences.[22]

The overall prevalence of renal disease in this cohort is approximately 50%, somewhat higher than other studies, mainly concerning European populations[17, 19, 30] but this is probably due to the study patients being from a tertiary referral center.

Neuropsychiatric SLE had various frequencies and manifestations. A wide range of neuropsychiatric manifestations have been described, from common features such as headache and seizure, to rarer events such as psychosis.[31]

Frequencies of neuropsychiatric manifestations range from 9% to 68% (Table 6). In our study 37% of patients had neuropsychiatric involvement. Differences in the reported prevalence of SLE neuropsychiatric manifestations may be due to several factors. Most studies have been performed by retrospective chart review and on established patient cohorts with variable disease duration. Differences in demographic characteristics, socioeconomic status and selection bias among cohorts are additional potential confounders. Finally, the use of different classifications and definitions of neuropsychiatric involvement provides significant limitations.[31]

Regarding immunological features, some of our findings are not surprising; as previously described,[3-6] the frequency of anti-Sm antibodies in our Tunisian patients was particularly high (Table 6). Such frequencies were noted in South African and Saudi patients.[12, 27]

The frequency of anticardiolipin antibodies and anti-β2 glycoprotein I brought up in our series is probably explained by the fact that they were sought especially in cases of obstetric or thrombotic manifestations of antiphospholipid syndrome.

SLE is often complicated by exacerbations and flares of varying severity. The relationship between ethnicity and outcomes in SLE has gained considerable attention in the past decade. It has been postulated that the disparate course and outcome observed among patients with SLE from different populations relates to the patients' biologic and immunogenetic[1, 20] features and/or to their socioeconomic status.[7-14]

Some studies have reported the frequency of clinical features of SLE in Tunisia but are often not representative of the actual spectrum of SLE because the cohort originated from a definite geographical area and does not include patients from all parts of Tunisia.

However our study also had some limitations. In this multicenter study, patients were from different internal medicine departments. Therefore, patients were evaluated by different physicians. Furthermore, our study consisted mainly of SLE patients attending tertiary referral centers; so it can be speculated that this study includes a greater number of milder and severe forms of SLE than moderate ones. Nevertheless, our population was characterized by high frequencies of photosensitivity, malar rash and anti-Sm antibodies.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix I: Tulup Study Group

Appendix I: Tulup Study Group

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix I: Tulup Study Group

Abdallah Meya1, Abdelhafidh Nadia2, Ajili Faida2, Ayari Mahmoud3, Bahloul Zouheir4, Bahri Fethi5, Bchir Saloua1, Béji Maher6, Ben Abdallah Olfa7, Ben Dahmen Fatma8, Ben Dridi Mohamed1, Ben Fredj Ismail Fatma7, Ben Hassine Lamia9, Ben Salem Thouraya10, Boukhriss Imene9, Bouslama Kamel1, Boussema Fatma11, Chebbi Wafa12, Cherif Ouahida11, Ennafaa Mourad1, Hamza Mohsen13, Harmel Ali1, Harzallah Olfa14, Jallouli Moez4, Kaddour Neila4, Khalfallah Narjess9, Khamassi Naziha13, Khelifa Mabrouk5, Klai Rim14, Laouani Chedia7, Larbi Thara1, Louzir Bassem2, Mahjoub Sylvia14, Miled Mohamed10, Mrad Skander1, Othmani Salah2, Ouertani Dorra10, Rokbani Lilia11.

1Mongi Slim university hospital, La Marsa – Tunis; 2Military university hospital – Tunis; 3Regional hospital – Bizerte; 4Hedi Chaker university hospital – Sfax; 5Farhat Hached university hospital Sousse; 6Military hospital – Gabes; 7Sahloul university hospital – Sousse; 8Regional hospital – Nabeul; 9Charles Nicolle university hospital – Tunis; 10La Rabta university hospital – Tunis; 11Habib Thameur university hospital – Tunis; 12Tahar Sfar university hospital – Mahdia; 13Razi university hospital – Tunis; 14Fattouma Bourguiba university hospital – Monastir.