Systemic lupus erythematosus (SLE) is a complex, multisystem chronic autoimmune disease predominantly affecting females in their childbearing years. A wide variability and combination of serologic abnormalities and clinical manifestations characterize SLE. The disease features may be influenced by several factors, including sex, age, and time elapsed since the onset of symptoms. The definite diagnosis and classification of SLE are preceded by preclinical and early clinical stages, which should be recognized to adopt preventive measures and ensure early diagnosis and therapeutic intervention aimed at preventing organ damage development that already begins to occur early in the clinical disease process. The clinical course is frequently subject to unpredictable flares of disease activity, progressive organ damage accrual, and reduced health-related quality of life. Therefore, understanding the disease course and clinical patterns is of the utmost importance to a timely diagnosis and proper management of SLE patients.

The current view of SLE pathogenesis is that environmental agents (e.g., infections, drugs, ultraviolet light) in a genetically susceptible individual trigger the activation of innate and adaptive immune responses, leading to the production of pathogenic autoantibodies. Positive feedback loops involving the innate and adaptive immune systems amplify autoimmune response during the preclinical stage of SLE [1]. As a result, there may be a prolonged preclinical phase characterized by accumulating an increasing number of autoantibody specificities. Anti-nuclear antibodies (ANAs) represent the immunologic hallmark of SLE, and they are typically present many years before the diagnosis while subjects are still asymptomatic. Although not currently part of the standard assessment for SLE, multiparametric predictive models based on genetic factors, transcriptomics (e.g., type I interferon signature), and soluble mediators (e.g., antibodies and pro-inflammatory cytokines) may be helpful for the identification of subjects at high risk of developing SLE clinical features [2,3]. Preventive measures in at-risk individuals include removing environmental modifiable risk factors claimed as potentially triggering SLE onset (Table 1) [[4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18]].

ANA-positive subjects who later develop SLE show progressive accumulation of a higher prevalence and higher titres of more specific autoantibodies, such as anti-dsDNA and anti-Sm, before the onset of SLE. Arbuckle et al. reported that anti-Ro/SSA antibodies were detected in sera of SLE patients on average 3.0 years before the clinical onset of SLE and up to 9.4 years (mean 3.7 years) before the definite diagnosis, while anti-dsDNA and anti-Sm were detected 1.2 and 0.5 years before clinical onset, respectively [19]. Isolated ANA positivity has low dependability for SLE diagnosis as it can be found in various systemic and organ-specific autoimmune diseases, viral infections, in subjects taking drugs known to induce ANA positivity (e.g., TNF-alpha inhibitors, isoniazid), even in healthy individuals and especially healthy relatives of SLE patients. Using indirect immunofluorescence (IIF) Hep-2 assay, the prevalence of low titers ANA (1:40) in the healthy population is estimated to be up to 30%, whereas it is less than 5% when higher titers of ANA (>1:160) are present [20]. A recent meta-regression analysis included 64 studies comprising 13,080 SLE patients and confirmed a reduced sensitivity and a greater specificity for increasing ANA titers using IIF-Hep2 [21]. Therefore, ANAs must be used as a sensitive screening test to narrow the population suspected of having SLE and should be combined with other autoantibodies serving as specific biomarkers (Table 2) [[22], [23], [24]]. There is no SLE diagnostic gold standard, so the diagnosis of the disease is still a clinical decision based on physician expertise and supported by biomarkers. The lag time between symptoms onset and diagnosis has progressively shortened since the implementation of autoantibody testing [25].

SLE onset is often insidious, with the classifiable disease usually developing over years. At disease onset, in up to 50% of patients, clinical and serological manifestations are insufficient to meet the classification criteria of SLE [26]. Classification differs from diagnosis, as classification criteria are standardized definitions that help researchers identify a uniform group of patients for clinical research. They are traditionally not intended to capture all the possible patients but to create homogeneous cohorts for scientific purposes [27]. On the other hand, diagnostic criteria recognize the heterogeneity of the disease, intending to identify as many people with the condition as possible. Given the lack of diagnostic criteria, SLE diagnosis relies on the judgment of a trained physician, and it requires ruling out potential lupus mimickers, including other connective tissue disease (CTD), infections (e.g., Parvovirus B19, Epstein Barr virus, Leishmania), and malignancies (e.g., lymphoma). However, a set of classification criteria with a good combination of sensitivity and specificity may serve as a framework to aid clinicians in the diagnostic process (Table 3) [22,23].

The clinical condition of still-not-classifiable SLE has been defined in several ways, including borderline, latent, intermediate, probable, potential, possible, and, most frequently, incomplete SLE (iSLE). The term iSLE describes a condition characterized by clinical manifestations and serologic abnormalities specific to SLE (e.g., malar rash, hemolytic anemia, anti-dsDNA, anti-Sm) but insufficient for classification [28]. Patients with iSLE and classifiable SLE have comparable genetic loads of SLE risk loci, suggesting similar genetic susceptibility, while phenotypical differences may be influenced by gene-environment interactions [29]. iSLE does not identify a benign form of the disease in terms of severity or duration, or suggest an inevitable progression to classifiable SLE [27]. The definition of iSLE can partly overlap with that of undifferentiated CTD (UCTD), an entity characterized by serological and clinical features insufficient to meet the classification criteria of a specific CTD. According to Mosca et al. [30], there are two types of UCTD: a) stable UCTD, with signs and symptoms that remain stable over time (for at least three years); b) evolving UCTD, which evolves into a definite CTD including SLE in 20–60% of cases. Recognizing the UCTDs that correspond to iSLE and are at risk of evolving to classifiable SLE has clinical implications for treatment and follow-up [31].

About 10–50% of iSLE patients will progress to definite SLE, mostly within five years since onset, with iSLE patients usually being older at diagnosis than patients with definite SLE [31]. A recent meta-analysis helped to define further the clinical and serologic manifestations linked to progression to classifiable SLE (Table 4). [32]. A retrospective study reported that hydroxychloroquine (HCQ) treatment could delay the progression from iSLE to SLE and decrease the repertoire and expression levels of autoantibodies present at diagnosis [33]. New and more robust evidence is expected from the randomized, placebo-controlled, double-blind clinical trial SMILE (Study of Anti-Malarials in Incomplete Lupus Erythematosus) [34]. The SMILE trial enrolled patients with iSLE, defined by ANA (at least 1:80) and 1 or 2 additional SLICC criteria, to evaluate whether HCQ is effective in preventing or delaying the progression to SLE within 24 months and to provide further insights into the appropriate target population. Besides treating patients with HCQ, progression-preventive strategies should include lifestyle changes and removing modifiable environmental risk factors (Table 1), including smoking habits and unprotected sunlight exposure.

Although not classifiable as suffering from SLE using a set of validated criteria, patients with iSLE do not necessarily have a milder disease, and in many cases they are affected with an initial yet full-blown SLE and must be diagnosed and treated accordingly. Interestingly, the set of clinical and serologic features that would make a patient with UCTD or iSLE considered to be a full-blown SLE was sufficiently consistent across studies and included renal involvement, acute cutaneous manifestations, thrombocytopenia, autoimmune hemolytic anemia, seizure, anti-dsDNA, anti-Sm, and hypocomplementemia [32,35,36]. Moreover, a multicentre international study identified a set of clinical manifestations and serologic abnormalities helping to early distinguish SLE, irrespective of the fulfillment of classification criteria, from diseases mimicking SLE [24]. Besides mucocutaneous involvement with the typical malar rash (OR 15.0; 95%CI 8.4–26.6) and urine analysis abnormalities (e.g., proteinuria, haematuria, pyuria, and casts) due to kidney involvement (OR 17.0; 95%CI 4.1–70.4), also serositis (OR 6.6; 95%CI 3.5–12.3), synovitis (OR 3.8; 95%CI 2.6–5.4) and fever (OR 3.3; 95%CI 2.1–5.1) showed a statistically significant association with SLE diagnosis. Features less common in early SLE than in SLE mimickers included Raynaud's phenomenon, sicca symptoms, dysphagia, and fatigue [24]. These results provided information for the EULAR/ACR 2019 SLE classification criteria that achieved the highest combination of sensitivity (96.1%) and specificity (93.4%) compared to the ACR 1997 criteria (93.4% specificity and 82.8% sensitivity) and the SLICC 2012 criteria (96.7% sensitivity and 83.7% specificity) [23]. Several additional studies assessed the accuracy of the 2019 EULAR/ACR classification criteria, reporting both high sensitivity (87.3–97.4%) and high specificity (87.8–97.3%) [[37], [38], [39], [40]]. Good receiver operating characteristics were confirmed in the sub-cohorts of early SLE (diagnosis less than 12–36 months) with 87.3–92.8% sensitivity and 87.8% specificity [35,38,40]. Notably, a recent single-center retrospective study suggested that 8.6–20.1% of patients diagnosed as having early SLE are not correctly classified using the EULAR/ACR 2019, SLICC 2012, and ACR 1997 criteria individually, while the combined use of all three sets of criteria ensured the classification of 97% of patients [39].

The term “early SLE” has been used as an alternative to iSLE to describe subjects with preclinical and clinical features consistent with SLE but not fulfilling classification criteria [41]. The meaning of the term early SLE is changing to denote a subject whose diagnosis of SLE was made recently with respect to the onset of symptoms, independently from classification (Fig. 1) [24,42]. This paradigm shift is driven by the growing awareness that early recognition of symptoms and subsequent early therapeutic intervention can prevent organ damage development and accrual. Nevertheless, the reported median lag between the onset of symptoms and diagnosis of SLE is still too long (≥2 years), mainly driven by the time between the first report of symptoms to a doctor and the assessment by a rheumatologist [43,44]. Some clinically actionable initiatives, such as developing red flags and implementing tools to screen patients for referral, may help reduce referral times from primary care providers or non-rheumatology specialties to rheumatologists (Table 5). For example, the SLE Risk Probability Index (SLERPI) is a clinician-friendly algorithm enabling risk prediction for diagnosis and exhibited high accuracy in early SLE subjects [42], even those not fulfilling any definite criteria set [45]. Pending further validation in different settings, the SLERPI may be helpful as a screening tool in subjects showing non-specific serological features (e.g., isolated ANA) with high-yield clinical manifestations (e.g., malar rash or proteinuria>500mg/24 h), in those suffering from multiple clinical manifestations but no immunological abnormalities or when specific autoantibodies (e.g., anti-dsDNA, anti-Sm) concur with a single clinical feature (e.g. thrombocytopenia or autoimmune hemolytic anemia).

Administrative database analysis showed that SLE patients diagnosed within six months of symptom onset had less severe disease, lower flare rates, and hospitalizations than those with a longer delay in the diagnosis [46]. In addition, observational studies showed that the achievement of remission or lupus low disease activity state (LLDAS) within six months since diagnosis, and their maintenance in the following 12 months, independently predicted lower damage development in newly diagnosed SLE patients [47,48]. Therefore, early therapeutic intervention is crucial when considering that organ damage develops from the very early stage of the disease, and up to 22% of SLE patients one year after diagnosis have at least one item of damage evaluated by the SLICC/ACR damage index [49,50]. Moreover, patients with early damage within one year since diagnosis have twice the risk of accruing damage, and the mortality rate is approximately three times higher than those without damage [50,51]. Implementing an integrated approach from the early stages of the disease by treating to target remission or LDA, adding HCQ as background treatment, addressing comorbidities, and minimizing prednisone use below 5 mg/day may help to reduce early damage development in newly diagnosed SLE patients [49,52].

Several definitions of early SLE have been proposed with respect to the time elapsed since symptom onset, ranging from <6 months to <36 months, with no general agreement [35,39,42,46,47]. Moreover, whether one of these timeframes represents a window of opportunity to increase the chance of achieving SLE remission and preventing damage development after treatment initiation has not been thoroughly investigated.

The clinical course of SLE after diagnosis is generally characterized by periods of increased clinical disease activity (relapse or flare) alternating with periods of quiescence (remission) [53]. The duration of remission varies significantly among patients and depends on several factors, including compliance. Apart from this well recognized pattern of disease activity, early studies from the Johns Hopkins Lupus cohort also identified a “long quiescent” and a “chronically active” pattern [54]. That report was based on non-inception patients (not followed since diagnosis) and showed that the chronically active pattern was the most prominent with almost 40% of the cumulative patient years (mean follow-up of 4.5 years) [54]. About 20 years later, a study from the same center that included all patients with at least one year of follow-up concluded that only 19% will follow a chronically active disease pattern [55]. Relapsing-remitting disease was the most common pattern 50% of the patients whereas 31% followed a long quiescent course [55]. In these studies, disease activity was defined based on the Physician's Global Assessment (PGA) and the Mexican version of the SLEDAI (M-SLEDAI, excluding serology) while flares were defined as any increase in the M-SLEDAI score from the previous visit. In a study of 267 ethnically diverse inception patients (time from diagnosis to first clinic visit 3 months on average) from the Toronto Lupus Clinic who were followed for at least 10 years, approximately 70% of the patients demonstrated a relapsing-remitting pattern [53]. At baseline, these patients did not differ from individuals with prolonged remission or persistently active disease with regards to the demographic and clinical characteristics, serologic activity or initial therapeutic approach. During the first 10 years, these patients spent almost half the time in remission (mean 5.3 years) and they experienced 2–4 flares. Disease activity was defined on the basis of SLEDAI-2K; remission was defined as a clinical SLEDAI-2K = 0 and active disease as SLEDAI-2K ≥ 1.

Approximately 10% of the 267 patients ran a prolonged remission course. After the initial phase of disease activity (at diagnosis), these individuals achieved clinical remission (as defined by SLEDAI-2K) within two years from diagnosis and did not flare during the first decade of disease [53]. Most of these patients (74%) went on to demonstrate a monophasic disease course without any flare during follow-up of 18 years on average [56]. About 25% of them had severe manifestations at onset such as diffuse proliferative nephritis and neuropsychiatric lupus. Earlier reports on the patterns of disease activity in SLE failed to capture monophasic patients; the relatively short follow up (1–5 years) may account for this [54,55,57,58]. In this context, severe clinical manifestations, such as lupus nephritis, may be resolved in several years after therapy initiation [59]. Thus, in the appropriate clinical setting (i.e. when signs of partial remission are evident), it seems reasonable to allow sufficient time for remission to occur.

Persistently active disease (never or only achieving short-lived remission, i.e. less than 6 months) was reported in approximately 10% [53]. Earlier studies reported a significantly higher percentage (up to 40%), albeit with a much shorter follow-up [54,60]. The factors which determine an unfavorable disease course have not been defined. However, it is believed that poor adherence to medications is a leading cause. About 43–75% of lupus patients were reported to have poor adherence in a systematic review, whereas about 33% arbitrarily discontinue their medications after 5 years [61]. Non-adherence has been linked to higher flare rate, morbidity and hospitalizations among lupus patients [62]. Other predictors include higher disease activity at onset, as well as musculoskeletal and skin involvement [60]. The different patterns of disease activity over time are displayed in Fig. 2.

In addition to the aforementioned well-defined patterns of disease activity, the remaining 10% of patients demonstrated an unusual course with one remission period that varied from 1 to 8 years during the first decade [53]. These “hybrid” patients demonstrated a global disease activity that fell between the relapsing-remitting and the persistently active patients. Their damage accrual at the end of 10 years was similar to the relapsing-remitting group.

Disease course patterns greatly impact prognosis over time. There seems to be a linear relationship between disease activity and damage accrual over time with increasing irreversible damage in the persistently active and the relapsing-remitting patients. Even the patients who achieved prolonged remission developed some late damage [50,56,63]. Damage is related to disease activity and glucocorticoids; the latter becomes more prominent in later phases of the disease, most likely as a result of the prolonged exposure to glucocorticoids. Certain comorbidities that are directly or indirectly associated with glucocorticoids (osteoporosis, osteonecrosis and atherosclerotic cardiovascular events) are more frequently observed in the persistently active and relapsing-remitting pattern than in the patients who achieve prolonged remission [64,65]. In all relevant studies, damage was assessed with the SLICC/ACR Damage Index [66]. Disease patterns also affect mortality with persistently active disease being linked to higher mortality in the first five years after diagnosis [67].

Interestingly, the time in clinical remission was the most important predictor of damage accrual in most relevant studies [53,54,68]. The relapsing-remitting patients who spent more than 50% of the first decade in remission had comparable damage accrual to the individuals who achieved prolonged remission (average time on remission 8.8 years) despite receiving higher cumulative glucocorticoid dose over the first 10 years [53]. Moreover, the “hybrid” patients were resembling all three groups according to their time in remission. Other investigators also showed that a 2-year remission is protective against damage in Caucasian lupus patients, whereas a sustained remission for 5 years was associated with 96% less risk of damage [68].

Since complete remission has not been strictly defined yet and is rare [69], LDA seems to be of value in daily practice and, also, in clinical trials for new medications in lupus. LDA captures a disease state that, although different than complete remission, leads to favorable long-term outcomes, such as reduced damage accrual and increased survival [70].

In this context, different definitions of low disease activity emerged during the last several years, namely minimal disease activity (MDA) [71], LDA [72] and lupus low disease activity state (LLDAS) [73]. The major differences (Table 6) consist of the level of acceptable disease activity as expressed by the SLEDAI-2K (from 1, 2 and 4, respectively) and the permissible treatment (prednisone up to 7.5 mg/day in LLDAS but not in MDA and LDA). All three definitions have been tested for their impact on long-term outcomes, such as damage accrual and mortality, and have shown that LDA status confers outcomes comparable to remission [70].

Duration and sustainability of LDA seems to be the most important factor for improved outcomes. Attainment of LDA for a short period (e.g. for a few months before the next disease flare) is not clinically meaningful in the long term. Relevant studies have shown that LDA duration of two years, at a minimum, will be translated to improved outcomes [74]. Interestingly, studies with prolonged follow-up reported that the LDA patients spent approximately 80% of the time in clinical remission [74,75]. Practically, this implies that patients had active disease in the beginning of the observation period and, then, achieved prolonged clinical remission with minor flares that could be managed with topical treatment.

The concept of LDA has not been applied to studies examining the disease activity patterns over time. It is possible that some flares of the relapsing-remitting patients will still fall into the definitions of LDA (“minor flares”) while others will be more severe (“major flares”). If this is proven, it is reasonable that the outcomes of patients with minor flares only will be closer to the patients achieving prolonged remission. On the other hand, relapsing-remitting patients with major flares will develop outcomes closer to the persistently active ones. Such data may further improve the accuracy of predictive models and enhance the ability to apply early, patient-tailored therapeutic approaches.

Of note, serology (anti-dsDNA antibodies and complement C3/C4) was not found to affect disease course patterns [[53], [54], [55]]. Serologically active clinically quiescent (SACQ) patients do not accrue more damage than patients in clinical and serological remission and, thus, do not warrant active treatment but close surveillance since they are occasionally linked to an increased flare risk [76]. Consequently, solely increased anti-dsDNA titers and/or decreased C3/C4 (without concomitant clinical activity) are accepted by most investigators for inclusion into the definition of remission [69,77]. Nevertheless, serology monitoring is indicated in periods with increased risk for disease relapse even in the absence of clinical activity.

SACQ patients are a particularly interesting subset of SLE. Initially identified in the late 1970s, such patients were intensively studied to understand the discrepancy between serological activity (increased anti-dsDNA titers and decreased levels of C3/C4) and disease activity [78]. Mechanistic studies on the nature of anti-dsDNA and anti-chromatin antibodies did not show any differences between SACQ and non-SACQ patients [79]. Moreover, the expression of interferon type I (“interferon signature”) and pro-inflammatory cytokines/chemokines did not differ between SACQ and serologically and clinically quiescent patients, implying an “autoimmune remission” status [80]. Although many patients in a large cohort may exert a temporary SACQ status, only about 6% were demonstrated to maintain this status for 2 or more consecutive years [81]. About 60% of them flared in the next three years but fluctuations in the anti-dsDNA titers and/or complement levels were not reliable predictors. Compared to clinically active individuals, SACQ patients exhibited less damage accrual, less cardiovascular events and less renal damage over 10 years of follow-up [76].

The timely prediction of disease course holds the potential for a tailored treatment plan with early treatment de-escalation in the prolonged remission (or monophasic) patients or prolonged maintenance therapy in the case of relapsing-remitting or persistently active patients. Genetic factors that suppress the effector arm of the immune response or enhance immune regulation (or both) may be of importance [82] whereas epigenetic variables may also play a role in the attenuation of the pathogenetic process [83]. Several soluble mediators have been implicated in disease pathogenesis and their serum levels are altered before clinical flares [84]. Black race/ethnicity and increased disease activity over the first 2 years (as expressed by the adjusted mean SLEDAI-2K) were independently associated with disease course [53]. This implies that early response predicts better outcome over time and treat-to-target strategies should aim at remission or, if not possible, a LDA state.

Although SLE may develop at any age, its peak incidence occurs during the reproductive age years [85]. There is no agreed definition of the age cut-off for late- or early-onset SLE [86]. Most studies define late-onset SLE (LSLE) as manifesting at or after 50 years of age [87,88]. A second definition of an age of 65 or more has also been proposed [89]. However, when comparing LSLE populations selected with these two different cut-off ages, no relevant differences were found [89]. There are fewer published data on LSLE in comparison to early-onset disease. However, differences in disease activity, clinical manifestations, comorbidities and morbidities have been demonstrated in diverse countries [86,[90], [91], [92], [93], [94]].

In 2–20% of patients, SLE develops after the age of 50 [88,91,92]. Female predominance decreases with age, ranging from 7:1 to 18:1 in early-onset cases to 4:1 to 7:1 occurring after 50 [87,89,95]. This reduction has been linked to variations in estrogens levels [92,95]. Patients with LSLE are mainly Caucasian [89,92].

LSLE patients often take longer to diagnose. This delay could relate to LSLE being insidious in its onset with atypical clinical manifestations, comorbidities that may obscure key symptoms, and reluctance to consider SLE occurring in the elderly [[88], [89], [90],92,96].

LSLE patients less frequently develop the characteristic manifestations of SLE of earlier onset, namely mucocutaneous, kidney and musculoskeletal disease [87,[91], [92], [93],95] (Table 7). This is possibly due to immune senescence [92]. It has been widely accepted that LSLE presents with a more insidious disease onset, less organ involvement, and a more benign disease course [91,97]. Nevertheless, Alonso et al. [87] reported the most frequent clinical manifestation in both late- and early-onset SLE patients was arthritis, with no significant difference between groups, while some authors reported more frequent musculoskeletal involvement in LSLE [86,92,97]. Choi et al. [91] reported fever, anemia, and thrombocytopenia were less frequent in LSLE. Alonso et al. found a lower frequency of seizures and psychosis in older patients [87]. The most important age-related difference seems to be a notable decrease in the incidence and severity of renal disease in LSLE [87]. LSLE patients more often develop cardiopulmonary dysfunction, notably serositis and interstitial lung disease (ILD) [92,93,98]. Linked to advancing age, tobacco use and ‘immune senescence’, patients with SLE-Sjogren's syndrome (SS) overlap syndrome could be associated with a higher risk of ILD in LSLE [98].

Riveros Frutos et al. [92] reported LSLE patients more frequently had thromboembolism, deep vein thrombosis and lupus anticoagulant positivity. However, no significant differences in thrombosis frequency between late- and young-onset SLE patients were found by Cartella et al. [99]. Furthermore, the incidence of antiphospholipid antibody syndrome does not differ significantly between late- and early-onset SLE patients [89,91,95].

Some studies found no significant difference in major organ involvement among different age of SLE onset groups [86,88,89,97]. The only significant difference reported by Padovan et al. [89] was in peripheral nervous system involvement, which was more frequent in the LSLE subgroup aged ≥65 years. The severity of SLE appears to decrease with age [88,[91], [92], [93]]. In contrast, Prevete et al. and Padovan et al. [86,89] found no difference between early- and late-onset SLE patients regarding disease activity. Organ damage seems to be more common in LSLE [[89], [90], [91]]. Although this increase is possibly due to iatrogenic effects such as osteoporosis and age-related morbidity, it challenges the interpretation of LSLE as a more benign disease [89,97].

LSLE patients had increased risks of multiple pre-existing comorbidities at diagnosis [86,92,95,97], including a higher incidence of hypertension, cerebro-vascular accidents, cardiovascular diseases, peripheral vascular diseases, cancer, osteoporosis, diabetes mellitus (DM), thyroid disease, higher body mass index, and depression [86,88,[90], [91], [92],96]. For half of LSLE patients it only took less than one year to develop any increase in the burden of comorbidities, with a higher impact (33.3% cumulative incidence) to all-cause mortality [95].

LSLE patients more often have concomitant SS [87,91,92]. The subset of patients with LSLE and SS has a distinct clinical and laboratory phenotype, with a higher frequency of photosensitivity, oral ulcers, Raynaud's phenomenon, anti-Ro and anti-La antibodies. The relationship between LSLE and SS is debated, with autoimmune endocrinopathy being considered a manifestation of SLE and SS being claimed as a secondary manifestation of LSLE. Feng et al. [97] observed SS was the most common simultaneously occurring autoimmune disease, while hypertension and infection accounted for the most common non autoimmune disorders. In contrast, Choi et al. [91] found no significant difference in the prevalence of cancer, DM, thyroid disease. There was no significant difference when comparing the prevalence of hypertension in an age-matched general population. Thus, hypertension could be a consequence of the aging process instead of SLE itself [91].

Patients with LSLE have a lower frequency of anti-dsDNA, anti-nucleosome, anti-Sm, anti-RNP antibodies and lupus anticoagulant positivity. Decreased complement levels, including C3, C4 and CH50, have also been reported to be less common in LSLE [[86], [87], [88], [89],91,92,97] (Table 8). This serological profile could affect or reflect disease activity [91]. Some studies report an atypical immunological profile including higher frequencies of rheumatoid factor (RF), anti-Ro/SSA and anti-La/SSB antibodies [87,89,92]. The first could relate to RF occurring more frequently after age 65 [89]. In contrast, Padovan et al. reported significantly higher anti-dsDNA antibody levels in LSLE [89] and Wen et al. [93] significant differences in antibody profiles among SLE onset-age groups. Although SS was more common in LSLE patients, there were no differences in the prevalence of anti-Ro/SS-A and anti-La/SS-B antibodies in a Korean cohort study. This study also found no relationship between autoantibody profiles and lupus nephritis in LSLE [91]. Elevation in inflammatory markers, namely the erythrocyte sedimentation rate and C-reactive protein does not appear to differ significantly among different age-groups [91].

In some studies, the frequency of corticosteroid and hydroxychloroquine use did not differ significantly between age-at-onset groups [[88], [89], [90]]. In contrast, Feng et al. [97] found antimalarial drugs were less likely to be used in LSLE. Given the increased likelihood of comorbidities in LSLE and their correlation with adverse outcomes, strategies for preventing or mitigating the impact of comorbidities should be enhanced, including minimizing the use of glucocorticoids and the related risk of osteoporosis, diabetes, hypertension, and mood disorders. Sohn et al. [88] reported that the frequency of immunosuppressive agent use was not different between late- and early-onset patients. In contrast, the frequency of cyclophosphamide and azathioprine use has been reported as significantly lower in LSLE [88,95,97], implying lower disease activity in this group compared to younger patients. Padovan et al. [89] reported that no immunosuppressive agents were used to treat a group of 30 LSLE patients aged ≥65 years.

In a cohort of hospitalized patients, the SLEDAI-2K scores in LSLE tended to decline more significantly than in younger patients at discharge, suggesting LSLE could be more sensitive to treatment [97]. Alonso et al. [87] found no significant difference in flare frequency between early and late-onset patients.

Although Sohn et al. [88] found standardized mortality ratio of LSLE was not higher than that of the general population, LSLE patients have a lower survival probability than early-onset SLE patients [87,91,92,96,97] because of their age. Feng et al. [97] observed nearly half of the mortality among LSLE patients was due to infections, particularly pulmonary, which were more frequent in comparison to early-onset patients. Vital organ damage/SLE-related mortality did not differ significantly. Although not independently associated with deaths in LSLE, anti-Sm antibody positivity and the use of antimalarial drugs, reported to be protective factors [100], were less frequent among these patients [97]. In contrast, Cartella et al. [99] reported cardiovascular disease as the main cause of mortality in LSLE. Overall higher mortality is possibly related to the higher comorbidity burden and organ damage due to aging and longer exposure to traditional cardiovascular risk factors in LSLE [91,96]. Long-term prospective studies are lacking to better understand mortality in LSLE.