1. Introduction
Systemic lupus erythematosus (SLE) is a chronic, heterogeneous autoimmune connective tissue disease characterised by numerous visceral manifestations and the presence of autoantibodies [1]. The management of SLE relies on standardised clinical and laboratory monitoring and the judicious use of biomarkers to guide therapeutic decisions. The updated EULAR guidelines emphasise these principles and the need for decision‐aiding tools applicable in routine practice [2].
In addition to the autoantibodies that constitute its immunological signature, SLE is characterised, as the disease progresses, by the expression of numerous markers of inflammation, including C‐reactive protein (CRP). In SLE,CRP exhibits a distinctive pattern: it generally remains moderately elevated during isolated autoimmune flares, whereas a marked increase is more suggestive of a concurrent infection or certain inflammatory conditions such as vasculitis [3, 4]. In several hospital cohorts, CRP thresholds around 24 mg/L have been proposed as practical indicators of infection in febrile lupus patients, to be interpreted within the clinical context [5].
In sub‐Saharan Africa, and particularly in West Africa, the clinical presentation of SLE occurs in an environment with a high burden of infection and where access to specialist care remains unequal, which complicates the interpretation of elevated CRP levels. Epidemiological reviews show that infection and cardiovascular complications contribute significantly to excess mortality in SLE, whilst, in practice, the early differentiation between a flare‐up and an infection remains a major challenge for patient safety [6]. In Ghana, a prospective cohort study demonstrated the influence of the APOL1 genotype on kidney damage and prognosis, whilst in Nigeria, a recent multi‐centre study detailed the clinical and biological characteristics in hospital settings [7, 8].
In Abidjan, hospital‐based research highlights the burden of disease and the impaired quality of life among patients followed at the Cocody University Hospital, underscoring the value of simple tools for clinical decision‐making [9].
In malaria‐endemic areas, malaria often causes a significant increase in the CRP. This elevation can complicate the assessment of a febrile lupus patient presenting with elevated CRP. This reality reinforces the value of combined approaches (clinical, targeted microbiology, procalcitonin and simple ratios) to correctly interpret elevated CRP [10].
In this context, it is important to identify the factors contributing to elevated CRP levels in lupus patients, to distinguish between those related to infection, disease activity or other factors, and to draw conclusions regarding the implications for diagnosis and treatment adjustment. Our objective was therefore to analyse the factors that may be associated with elevated serum CRP levels in patients with SLE at the Cocody University Hospital.
2. Materials and Methods
This was a cross‐sectional analytical study based on clinical and laboratory data from patients with SLE. The medical records of patients hospitalised or seen in routine outpatient consultations at the Rheumatology Department of the Cocody University Hospital in Abidjan were systematically reviewed over a 4‐month period from March 2025 to June 2025.
The study included adult women (aged ≥18 years) with a confirmed diagnosis of SLE according to international criteria (ACR/EULAR) who had given their consent for the secondary use of their data. Excluded were all patients receiving treatment with corticosteroids and/or immunosuppressants prior to the biological assessment, as well as those whose records were incomplete and did not allow the primary data extraction, patients with an active infection at the time of the study, known liver disease, pregnancy, or another concomitant inflammatory or autoimmune disease. An exhaustive non‐probability sampling was carried out by consecutively including all records meeting the eligibility criteria during the study period, resulting in a total of 62 patients.
A single clinical and biological assessment point was retained for each woman, including the CRP measurement. Treatment information documented in the medical records was collected for descriptive purposes. The study did not include repeated CRP measurements before and after treatment.
Data collection was carried out using an anonymous questionnaire to ensure the confidentiality of the information. The variables included socio‐demographic characteristics (age and body mass index) and medical history. Clinical and disease‐related parameters included the SLE Disease Activity Index (SLEDAI) [11], pain assessment using a visual analogue scale (VAS) [12], the level of clinical severity, disease status (flare‐up or remission) and response to treatment.
CRP levels were measured by immunoturbidimetry on a Cobas c6000 automated analyser (Roche Diagnostics) using the Tina‐quant CRP IV (CRP4) reagent. This method, based on the agglutination of latex particles coated with anti‐CRP antibodies, offers a measurement range of 0.6–350 mg/L, a detection limit of 0.3 mg/L and a quantification limit of 0.6 mg/L. The recommended reference value is <5 mg/L, but each laboratory must verify its transferability.
In the present study, a CRP level >6 mg/L was considered positive [13]. Routine laboratory data, including a complete blood count and tests for specific autoantibodies (antinuclear antibodies [ANAs], anti‐double‐stranded DNA antibodies, etc.), were included in the analysis [14].
2.1. Treatment of the Dependent Variable (Prior to the Statistical Analysis Section)
For statistical reasons, the dependent variable corresponding toCRP concentration was treated as a categorical variable. CRP levels were divided into three categories: <6 mg/L (normal), 6–20 mg/L (moderately elevated) [15, 16] and >20 mg/L (highly elevated) [5, 16, 17].
Participants were classified into two age groups: 18–45 years (period of high incidence corresponding to childbearing age) and > 45 years (beyond the peak period, including late‐onset forms). This categorisation is based on the epidemiology of SLE, which peaks between 15 and 45 years of age, and on reviews defining late‐onset forms as those occurring from the age of 50 onwards, as well as on local case series confirming the existence of cases in people over the age of 45 [6, 18–21].
2.2. Statistical Analysis
Quantitative variables were expressed as means ± standard deviation, and qualitative variables were expressed as percentages. Relationships between variables were explored using univariate and then multivariate analyses. For the multivariate analysis, a multinomial regression (CRP <5/5–20/>20 mg/L) was considered, but instabilities (near‐separation, limited sample sizes) led to dichotomising CRP at 20 mg/L (≥20 vs. <20 mg/L), a threshold validated in the literature [5, 17].
The outcome was therefore modelled using binary logistic regression on complete cases, with categorical variables coded using a reference category. The final model included age, BMI, proteinuria ≥0.5 g/24 h, elevated serum creatinine and SLEDAI score. Sensitivity analyses confirmed the robustness of the model, the results of which are presented as adjusted odds ratios (ORs) with 95% confidence intervals (CI).
From this adjusted model, average marginal effects (AMEs) were calculated to estimate the average change in probability of CRP ≥ 20 mg/L associated with each factor. The model’s performance was assessed using the area under the ROC curve (AUC). All analyses were performed using SPSS Statistics v29 (IBM Corp.) and Python 3.11 (statsmodels 0.14.2).
Correlation plots were generated using GraphPad Prism v10.2.0 (392). The significance threshold was set at p < 0.05.
2.3. Ethical Considerations
The study was carried out in accordance with the ethical principles set out in the Declaration of Helsinki. It was approved by the Ethics Committee of SIHIO‐TS (Ivorian Society of Haematology, Immunology, Oncology, and Blood Transfusion) under Reference Number 53_2025‐SIHIOTS was published on February 14, 2025. The secondary use of data was authorised by the participants’ informed consent, with complete anonymisation prior to the analysis.
3. Results
3.1. Basic Descriptive Characteristics
The study involved 62 women. Table 1 presents the main clinical, therapeutic and laboratory characteristics of the study population. The majority of patients reported no history of a first lupus flare‐up (83.9%).
The articular form was the most common type of flare‐up (29.0%), although nearly half of the cases were not specified in the records (45.2%). Regarding nutritional status, 41.9% of patients had a normal BMI and 38.7% were overweight. Renal manifestations were observed in 32.3% of patients.
The most common clinical presentations were cutaneous and mucosal lesions and joint pain (32.3%, respectively). In terms of treatment, synthetic antimalarials were used in 74.2% of patients, whilst immunosuppressants were prescribed in 19.4%. Clinical improvement was noted in 51.6% of patients.
Proteinuria ≥0.5 g/24 h and hypercreatininaemia were observed in 37.1% and 32.3% of patients, respectively. Antiphospholipid syndrome was found in 9.7% of cases. ANAs had a median of 1185 (2.8–1280), whilst anti‐dsDNA and anti‐Sm had median values of 77 (8–600) and 19.5 (2–320), respectively.
Anti‐SSA and anti‐SSB antibodies showed medians of 71.5 (1–640) and 12 (2–640) (Tables 1 and 2).
3.2. Distribution of CRP Concentrations According to Socio‐Demographic Factors
The mean age was 35.45 ± 13.50 years, ranging from 18 to 67 years. The mean BMIwas 25.0 ± 5.19 kg/m2, ranging from 16.49 to 43.12 kg/m2. CRP had a mean concentration of 28.20 ± 21.21 mg/L (median: 23.93 mg/L, range: 0.89–98.36 mg/L).
According to stratification by age group, 74.2% of participants were aged 18–45 years and 25.8% were aged 46–67 years. Regarding occupation, 16.1% were unemployed, 17.7% were students, 11.3% were engaged in informal work and 54.8% were in formal employment. No significant association was observed between CRP categories and age, BMI or occupation (p > 0.05) (Table 3).
3.3. Distribution of CRP Concentrations According to Participants’ Medical History
Among the study population, 9.7% had a family history of autoimmune disease and 9.7% were smokers. Those with a family history of autoimmune disease all had CRP levels >20 mg/L, with no significant difference (p = 0.190). Similarly, no significant association was observed with smoking status (p = 0.144) or with the onset of a first lupus flare‐up (p = 0.421).
Finally, the distribution of flare‐up types according to CRP levels did not differ significantly (p = 0.287), despite a predominance of joint‐involving forms in the CRP >20 mg/L category (Table 4).
3.4. Distribution of CRP Concentrations According to Clinical Parameters
In this cohort, 84.5% presented with asthenia and 45.2% with arthralgia or arthritis. The pain assessment (VAS) showed that 45.2% of patients were pain‐free, 24.2% had mild pain, 17.7% had moderate pain and 12.9% had severe pain. A statistically significant association was observed with the presence of arthralgia/arthritis (p = 0.044), with a higher proportion of CRP >20 mg/L in symptomatic patients.
Similarly, the level of pain assessed by the VAS was significantly associated with CRP categories (p = 0.023), with moderate to severe forms being more frequently found in the CRP >20 mg/L category. Renal manifestations were strongly associated with CRP levels (p = 0.001), with all affected patients having CRP >20 mg/L.
However, no significant association was found between CRP categories and other clinical parameters, notably asthenia (p = 0.382) or the various clinical presentations (p = 0.982) (Table 5).
3.5. Distribution of CRP Levels According to Laboratory Parameters
Among the participants, the majority had a CRP level >20 mg/L, regardless of the laboratory parameter being considered. Nearly one‐third of patients (30.6%) with proteinuria ≥0.5 g/24 h had a CRP level >20 mg/L, with a statistically significant association (p = 0.037). The same was true for hypercreatininaemia, with a p value of 0.001.
Leukocytosis, observed in 72.6% of patients, was also strongly associated with elevated CRP, with nearly 60% having a value >20 mg/L (p = 0.001). On the other hand, thrombocytopenia (16.1%), anaemia (64.5%), the Coombs test result and antiphospholipid syndrome showed no statistically significant association with CRP concentration (p > 0.05). Thus, this table highlights a significant association between elevated CRP and three laboratory parameters: proteinuria, hypercreatininaemia and leukocytosis (Table 6).
3.6. Correlation Between CRP and the SLEDAI Score
Figure 1 shows a significant positive correlation between CRP and the SLEDAI score (r = 0.323; p = 0.012). The higher the disease activity score, the greater the tendency for CRP concentrations to increase.
3.7. Binary Logistic Regression
The model included age, BMI, proteinuria ≥0.5 g/24h, hypercreatininaemia and SLEDAI. The table below presents the adjusted ORs, 95% CI, and p values. The (CI) were wide.
After adjustment, hypercreatininaemia remained significantly associated with CRP ≥ 20 mg/L (adjusted OR = 24.8; p = 0.011), whereas the other variables showed no statistically significant associations (Table 7). The ROC curve (Figure 2) shows the probability predicted by the multivariate logistic model for identifying patients with CRP ≥ 20 mg/L. The AUC was 0.840.
The curve deviated from the line of non‐discrimination, illustrating the distribution of sensitivity as a function of 1 –specificity for the different probability thresholds.
3.8. Estimation of Marginal Effects
The predicted probabilities showed an increased likelihood of elevated CRP associated with hypercreatininaemia (+47.5 points, p < 0.001) and proteinuria >0.5 g/24 h (+31.3 points, p < 0.001). With regard to BMI, being overweight was associated with a decrease (−19.2 points, p = 0.0155), whilst obesity was associated with an increase (+14.1 points, p = 0.0043). No significant effect was observed for the age group 46–67 years vs.
18–45 years (p = 0.293). Each + 1 point of SLEDAI increases on average the probability of elevated CRP of about 1.4 points (p = 0.0092) (Table 8).
4. Discussion
The results of this study highlight a frequent elevation in CRP levels among Ivorian female patients with SLE, with many CRP values ≥20 mg/L. Descriptive analysis, logistic regression and marginal effects estimation enabled the identification of the main clinical and biological determinants associated with this systemic inflammation. Hypercreatininaemia, significant proteinuria and obesity were major factors, highlighting the central role of renal involvement and metabolism in the biological expression of lupus activity.
In SLE, hepatic CRP production is generally inhibited by type I interferon [4, 22]. On the other hand, a marked increase in CRP, particularly above ~20–24 mg/L, should primarily raise suspicion of a concurrent infection, whilst bearing in mind that active serositis, arthritis or nephritis may also cause a marked elevation [4, 5, 23]. In our cohort, CRP correlated positively with the SLEDAI score, suggesting that it reflects, at least in part, the overall disease activity [5, 24].
The significant association with renal manifestations, arthralgia/arthritis and pain intensity reinforces this role.
Multivariate analysis confirms the importance of renal parameters in explaining systemic inflammation. Hypercreatininaemia proved to be the strongest predictor (OR 24.8; p = 0.011) and increased on average the probability of a CRP ≥ 20 mg/L of 47.5 points (p < 0.001). Proteinuria ≥0.5 g/24 h, although not significant in the adjusted model (OR = 2.08; p = 0.385), nevertheless increased this probability by 31.3 points (p < 0.001).
In SLE, CRP may rise during active renal involvement, but its specificity is limited, hence; the, value of combined approaches (ESR/CRP ratio, PCT, NLR/PLR) to distinguish between infection and flare‐up [5, 23, 25]. The SLEDAI score was not significant in the regression analysis (OR = 1.02; p = 0.709) but its marginal effect was modest and positive (+1.4 probability points per unit, p = 0.007) [4, 22, 26].
In the study population, analysis of marginal effects revealed a differential effect of adiposity on CRP: being overweight reduced the likelihood of elevated CRP (–19.2 points; p = 0.015), whilst obesity increased it (+14.1 points; p = 0.004). This pattern is consistent with the literature indicating that an increase in BMI is associated with higher CRP concentrations, including in young adults aged 17–39 (indicating low‐grade systemic inflammation) [27], and that BMI is a major determinant of CRP levels from childhood onwards (the best predictor in a multiple regression analysis) [28]. Conversely, data on weight loss show that losing weight reduces CRP levels [29].
Finally, the ROC curve for the predictive score showed an area under the curve of 0.84, indicating excellent discriminatory performance, in line with the standards reported for clinical models applied to SLE [5].
These results highlight the importance of combining logistic regression with marginal effect estimation to identify the most relevant determinants of systemic inflammation in SLE, particularly the role of renal involvement and obesity.
Elevated CRP levels, particularly those of 20 mg/L or higher, should prompt an immediate consideration of infection. However, our study has shown that kidney disease and obesity may also be associated with elevated CRP levels. Therefore, in practice, the clinical context, proteinuria, creatinine levels and complete blood count should be taken into account for a more accurate interpretation.
This study has several limitations. Its cross‐sectional and single‐centre nature, as well as the relatively small sample size, limit the generalizability of the results and do not allow for the establishment of temporal or causal relationships.
The SLICC/ACR Damage Index (SDI) was not collected in a standardised manner when the database was compiled. The available information, particularly regarding the irreversibility and duration of organ damage, was insufficient to reliably reconstruct this score, thus making it impossible to assess the influence of cumulative organ damage on CRP concentrations.
Furthermore, only one clinical and laboratory assessment was considered for each patient, and CRP levels were not measured repeatedly over time. Consequently, changes in CRP levels before and after treatment could not be assessed. Therefore, the potential influence of treatment on CRP concentrations cannot be excluded.
Finally, certain factors likely to influence CRP levels, notably subclinical infection, nutritional status or other unmeasured inflammatory factors, could not be fully taken into account. Larger, multicenter, longitudinal studies incorporating a standardised assessment of the SDI and repeated CRP measurements would be required to confirm these results.
Despite these limitations, the rigour of the data collection, the use of a multivariate model and the estimation of marginal effects lend considerable strength to the conclusions, whilst highlighting the need for larger‐scale longitudinal studies.
5. Conclusion
In this study, elevated CRP levels were primarily associated with renal manifestations, notably proteinuria and elevated serum creatinine levels, as well as disease activity, as assessed by the SLEDAI. These results suggest that CRP could serve as a complementary biomarker in the assessment of patients with SLE, particularly in the presence of renal involvement.
However, given the cross‐sectional nature of the study and the absence of repeated CRP measurements, these results should be interpreted as associations and do not allow for an assessment of the influence of treatments on the progression. Larger‐scale longitudinal studies, incorporating repeated CRP measurements and a standardised assessment of the SDI, are necessary to confirm these results.
Funding
The authors declare that no specific funding was received for this research.
Disclosure
All authors read and approved the final manuscript.
Conflicts of Interest
The authors declare no conflicts of interest.