Low bone mineral density in Thai children with systemic lupus erythematosus: prevalence and risk factors

We conducted a retrospective review of medical records of patients who had onset of SLE before the age of 18 years, fulfilling at least 4 of the American College Rheumatology diagnostic criteria [20]. All patients attended the Pediatric Nephrology Clinic between 2010 and 2015 at the King Chulalongkorn Memorial Hospital, a large 1435 bed general and tertiary referral teaching hospital in Bangkok. Patients with previously diagnosed chronic kidney disease based on an estimated glomerular filtration rate declined to <60 mL/min/1.73 m² for 3 months were excluded from the study. Participants underwent DXA measurements at the hip and lumbar spine BMD and repeated DXA during follow-up as a part of their routine care in the clinic. Demographic and clinical characteristics including sex, age, weight, height, disease duration before DXA scan, history of fracture, and use of medications were recorded. The SLE Disease Activity Index (SLEDAI) score and serum level of complement component 3 (C3) were used to assess SLE disease activity. Prescribed medications including glucocorticoid, immunosuppressants, calcium, and vitamin D supplements were recorded. Pubertal status and physical activity could not be determined by this retrospective chart review. The study was approved by the Institutional Research Ethics Review Board, Faculty of Medicine, Chulalongkorn University (IRB No. 149/59) and complies with the principles outlined in the contemporary revision of the Declaration of Helsinki 1964 (World Medical Association) incorporating the most recent (2013) and earlier amendments, and international guidelines for human research protection detailed in the Belmont Report, Council for International Organizations of Medical Services Guidelines, and the International Conference on Harmonization in Good Clinical Practice. We used the cohort guidelines in the STROBE statement to ensure complete reporting of this observational study [21].

DXA had been performed to measure BMD of the lumbar spine (L1–L4) and hip as recommended by the American College of Radiology as suitable for pediatric patients [22]. BMD measurement is reported as a z score calculated based on age and sex-matched reference values. Low BMD was defined by a BMD z score ≤ −2 [15,16,17].

Demographic data is summarized using descriptive statistics and expressed as mean with SD for variables with a normal distribution, as median and interquartile ranges (IQRs) for those without normal distribution, or as a percentage for categorical variables. Data between groups were compared using either an independent t test, Mann–Whitney U test, or χ² test, as appropriate. A Wilcoxon signed-rank test was used to determine differences in mean values from the initial and the follow-up DXAs. Factors associated with low BMD were analyzed using univariate logistic regression. Multivariate logistic regression was performed for variables with P < 0.1 from univariate analysis and are expressed as an adjusted odds ratio (OR). These statistical analyses were performed using IBM Statistics for Windows (version 21). Power analysis was based on the BMD of the lumbar spine at baseline and at the last follow-up using Stata (version 14, StataCorp). For all statistical analyses, P < 0.05 in a two-tailed test was considered significant.

We considered the demographic and clinical data from medical records of a cohort of 60 pediatric patients with SLE in this retrospective observational study. The majority of the patients (52, 87%) were girls, yielding a female-to-male ratio of 6.5:1. The mean age of all patients at diagnosis was 11 ± 2.5 years. The disease duration from the diagnosis to the first DXA ranged from 2 months to 58 months and to follow-up BMD measurement ranged from 11 months to 87 months, respectively. The most common renal pathology was proliferative nephritis (52%). All patients had been receiving steroid treatment at the time of the first DXA with a cumulative steroid dose ranging from 1.9 g to 42.6 g. None of our patients had a history of fracture. Demographic information including disease activity, medications, and laboratory results are summarized in Table 1.

Of the 60 patients who had a DXA scan performed for the first time, 24 had a BMD z score ≤ −2, demonstrating a prevalence of low BMD of 40%. Only 51 patients had a follow-up DXA scan completed. At follow-up, 28 patients had a BMD z score ≤ −2, which indicates an increased prevalence of low BMD to 55%. Comparison between the first and the follow-up DXA showed a significant decrease in both hip BMD z score (median difference −0.25, 95% confidence interval [CI] −0.40 to −0.05; P = 0.016) and lumbar spine BMD z score (median difference: −0.49, 95% CI −0.69 to −0.28; P < 0.001) (Figure 1).

Figure 1

There was no significant association between low BMD and age at the time of BMD measurement, disease duration, disease activity, vitamin D level, or the intake of either immunosuppressant drugs or vitamin D supplements. Mean weight for age z score in patients with low BMD was significantly lower than in patients with a BMD z score > −2 and was significant in univariate but not in multivariate analyses. The cumulative dose of steroids in patients with BMD score ≤ −2 tended to be higher than patients with a BMD z score > −2 (adjusted OR = 1.08, 95% CI 1.0 to 1.17; P = 0.050); however, this tendency was not significant (Table 2).