Systemic sclerosis in sub-Saharan Africa: a systematic review

Systematic studies on connective tissue disorders are scarce in sub-Saharan Africa. Our aim was to analyse the published clinical data on systemic sclerosis (SSc) in sub-Saharan Africa. A systematic review was carried out in accordance with the PRISMA guidelines. We screened the Embase, PubMed and African Health Sciences databases for literature published until March 2018. Searches produced 1210 publications. After abstract and full-text screenings, 91 publications were analysed, and epidemiological information and clinical features extracted. Publications were mostly publications case reports (36%), cross-sectional studies (26%) and case series (23%) and came predominantly from South Africa (45%), Nigeria (15%) and Senegal (14%). A total of 1884 patients were reported, 66% of patients came from South Africa. The patients were between 4 and 77 years old; 83% of patients were female. Overall, 72% had diffuse SSc. Raynaud´s phenomenon was reported in 78% and skin ulcerations in 42% of patients. Focal skin hypopigmentation was common and telangiectasia not frequent. Interstitial lung involvement was reported in 50%, pulmonary hypertension in 30%, heart involvement in 28% of patients. Oesophageal reflux was observed in 70% and dysphagia in 37% of patients. Antinuclear antibodies were positive in 65% of patients. Anti-centromere autoantibodies (9.2%) and RNA polymerase 3 antibodies (7.1%) were rare and anti-fibrillarin most frequent (16.5%). SSc presentations in sub-Saharan Africa differ from those reported in Europe and America by a frequent diffuse skin involvement, focal skin hypopigmentation and a high prevalence of anti-fibrillarin autoantibodies.


Introduction
With substantial advances in the prevention and treatability of infections, the non-communicable diseases have now emerged as an important contributor to the global disease burden and become a significant driver of health care costs in low and middle-income countries. Among the noncommunicable diseases, some musculoskeletal disorders such as systemic sclerosis (SSc), are characterized by large unmet medical needs and growing proportions of global morbidity and mortality.
Systemic sclerosis (SSc) is a chronic, autoimmune disease with a multisystem vasculopathy and an immense increase of fibrous tissue in affected organs. Even in high-income societies, SSc has a poor prognosis, as about 55% of affected patients die as a direct consequence of their disease [1]. In America, African descendants are known to have a particularly high SSc incidence and also a more severe disease course than Caucasians [2]. Various studies have identified differences in genetic disease associations between Caucasians and African Americans. Two seminal observations were however made in Africa itself, namely the first description of the prominent visceral involvement in SSc, coining the term 'progressive systemic sclerosis' [3] and the association of SSc with silica dust exposure [4].
Although the epidemiology, clinical phenotype and evolution of SSc organ involvement are intensively researched in economically advantaged countries, further robust data of SSc presentations in sub-Saharan Africa are largely lacking, prompting us to undertake a systematic review.

Methods
Information sources: we searched the Embase, PubMed, and African Health Sciences databases. Additionally, conference abstracts were retrieved by targeted searches of relevant websites, e.g. the African League against Rheumatism, the South African Rheumatism Arthritis Association, and the proceedings of the World Scleroderma Congresses. Reference lists of retrieved publications were also screened for publications not identified by the above search strategies. Search terms were defined (Table 1)  Study selection and data collection: two reviewers (VKJ and JE) independently screened the titles and abstracts. Patients were eligible for the analysis, if they were diagnosed with SSc by the reporting physician and resided within sub-Saharan Africa. Disagreement between the reviewers was resolved by consensus. In a second step, inclusion into this review was based on the full text of the remaining publications (JE), cross-checked by another author (VKJ). Data was extracted by one author (JE) and checked by another author (VKJ). In case of disagreement, a consensus was sought by discussion between authors.
Data items: 'SSc patients' were defined as patients diagnosed with SSc by the reporting physician and included if they were residing in sub-Saharan Africa. Information extracted included study methodology, patient demographics, and disease presentations. The prevalences of symptoms and organ involvement were calculated as percentages based on the denominator of the number of patients with available information, rather than the total number of patients included.

Current status of knowledge
Study selection: we retrieved 1090 publications from Embase, 531 from PubMed and 14 from other sources ( Figure 1). After removing duplicate publications, 1210 publications were left for screening. The selection process finally included 91 papers in this review ( Figure 1). Study characteristics: the 91 publications were published between the years 1945 [3] and 2018 [5], described a total of 1884 SSc patients and were carried out in 17 of the 48 sub-Sahara African countries (Figure 2), predominantly in South Africa (41 publications [45]), Nigeria (14 publications [15]) and Senegal (13 publications [14]).
The extent of the skin fibrosis in terms of the mRSS [98] was only reported in two case reports [17,21] with scores of 18 and 23 and one case-control study in which the average mRSS was 24.7 [83] Table 3. Anti-fibrillarin autoantibodies were the most frequent SSc-specific ANA, followed by antitopoisomerase 1 autoantibodies. RNA polymerase 3 and anti-centromere autoantibodies were infrequent; in one study of 73 black South African patients none of the patients had anti-centromere antibodies [42].
There was a high prevalence of antiphospholipid antibody specificities (anti-beta-2 glycoprotein-1, anti-phospholipid and anti-cardiolipin). These data were replicated in a Senegalese cross-sectional study that systematically analysed the association of anti-phospholipid antibodies with SSc, the lupus anticoagulant was found in 58% of patients [64]. The persistence of these autoantibodies at repeat testing was however not analysed and a clinical correlation with the antiphospholipid antibody syndrome not made [64]. More than 70% of patients in sub-Saharan SSc were reported to have diffuse cutaneous SSc. In the United States, diffuse SSc was reported in 60% of African Americans [103]. The high proportion of diffuse SSc in sub-Saharan Africans may in part be explained by economic factors situation and the organization of the health service in which patients with overt and extensive skin involvement are more likely to seek medical attention, while those with more subtle features do not, or are not referred for specialist care. Several papers [12,46,53,54,57] reported skin hypopigmentation frequently providing a 'salt and pepper' appearance of the skin. This feature is associated with anti-fibrillarin autoantibodies and seems to be much more than in the USA and Europe [104].
The overall prevalence of RP was 78% in our analysis, but the prevalence varied substantially across different sub-Saharan countries; the larger studies reported prevalences between 57% [41] and 96% [90] which is lower than the prevalence of 96% reported from EUSTAR [105]. Moreover, RP as the initial complaint of SSc occurred in only 16% of sub-Saharan patients as compared to 87% of the mostly Caucasians in EUSTAR [106]. This marked difference in RP prevalence might be due to a number of reasons. Firstly, differences in climatic conditions which modulate this vascular complication, with warmer climates in most sub-Saharan countries compared to European countries, is one explanation. An analysis of the EUSTAR database however did not show a significant relationship between prevalence of RP and ambient temperatures [101]. Studies in the US have not revealed differences in RP prevalence between African-Americans and other racial group [107]. Another possible reason might therefore be that RP symptoms are less likely to trigger a medical visit or are considered worth noting in low income countries. Finally, the difficulty of detecting the blue and red skin colour changes in darkly pigmented individuals may explain in part not only the low RP prevalence, but also the low prevalence of telangiectasia. In a recent South African study of cutaneous lupus, black patients were less likely to have the typical erythematous malar rash compared non-black patients [108].
Anti-topoisomerase 1 autoantibodies were detected in about one quarter of African American SSc patients [2,103]. The prevalence of antitopoisomerase 1 autoantibodies in our study was somewhat lower, but the prevalence of the other ANA specificities similar with studies of African-Americans in which ANA were present in 87% of patients [103], anti-fibrillarin in 19% [2] and anticentromere in about 10% [103] and 7% [2] respectively.
With regards to the causes of death, about two thirds of all SSc patients appear to die from cardiorespiratory failure secondary to interstitial lung disease, primary or secondary PAH (Mohammed Tikly, personal communication). This systematic review could however not mirror these observations due to lack of detail and follow-up. Further research is clearly needed.
An interesting fact is that SSc in combination with HIV was reported very rarely. The combination with HIV was reported in 7% of patients with known HIV-status. It is unclear if SSc improves during the course of untreated HIV infection similar to the observation in some of the other connective tissue diseases, and if HIV-induced CD4-positive T-lymphocyte loss may account for the low prevalence of some SSc manifestations [109]. This review has systematically analysed considerable patient numbers in heterogeneous study designs. A likely bias result on the one hand from unbalances the various sub-Saharan countries from which reports are available, and differences in health care access in these countries on the other hand. It should also be noted that almost half of the patients were not formally reported to be classified as SSc according to international criteria particularly the recently formulated ACR/EULAR criteria which are more sensitive. Lastly but not least, it must be noted that not all observations can necessarily be attributed to black patients, as exemplified by South Africa, which has a high proportion of other ethnicities.

Conclusion
SSc is no rarity in sub-Saharan countries but its presentation differs from that in Caucasians. Diffuse skin involvement and focal skin hypopigmentation seem to appear more frequently.
Prevalence of anti-fibrillarin autoantibodies seem to be higher likewise. More research must be undertaken to reduce bias and to elucidate the effects of tuberculosis and HIVinfection on SSc presentation and outcome

What is known about this topic
• The presence of SSc has been described in a variety of sub-Saharan countries; • Robust data of SSc serology and organ manifestations in sub-Saharan Africa are lacking.

Competing interests
The authors declare no competing interests.  Table 1: overview of the search terms used to retrieve publications Table 2: clinical features of the 1884 SSc patients Table 3: autoantibody specificity in 1145 patients with available autoantibody data Figure 1: flow chart of paper collection and selection Figure 2: origin of publications and patients