Impairment of white matter microstructure and structural network in patients with systemic lupus erythematosus

Abstract

Objective

The study aimed to investigate the damage of white matter (WM) microstructure and structural network in patients with systemic lupus erythematosus (SLE) using diffusion tensor imaging.

Methods

Tract-based spatial statistics (TBSS) were used to compare the difference in WM fractional anisotropy (FA) between SLE and HCs groups. The differences in WM networks between groups are compared using graph theory. The correlation between clinical data and SLE abnormal WM structure and network was analysed.

Results

The sample included 140 SLE patients and 111 healthy controls (HCs). Due to data missing, excessive head movement amplitude, failure of quality control and other reasons, 127 cases of SLE (103 females, mean age 29.84 years (SD 7.00), median years of education 12.00, interquartile range(9.00,15.00) and a median course of disease (month) 12.00, interquartile range (3.00,24.00)) and 102 cases of HCs (76 females, mean age 30.63 years (SD 7.24), median years of education 15.00, interquartile range(12.00,16.00)) were finally included in the study. The FA values of 5 clusters involving the right retrolenticular part of the internal capsule (RLIC), the genu of corpus callosum (GCC), the body of corpus callosum, the splenium of corpus callosum (SCC), were significantly lower in the SLE group compared to the HCs (P < 0.05 with threshold-free cluster enhancement corrected). The SLEDAI showed a negative correlation with FA in GCC, and HAMD showed a negative correlation with FA in SCC and right RLIC (P < 0.05). Regarding network indicators, Cp, E_glob, and E_loc were significantly decreased, while Lp was significantly increased in the SLE group. The degree centrality (DC) of 6 brain regions and the E_nodal of 17 regions were significantly lower in the SLE group. SLEDAI showed a negative correlation with the area under the curve (AUC) of DC and E_nodal in the left inferior frontal gyrus triangular (q < 0.05 with false discovery rate corrected), while MMSE showed a positive correlation with the E_nodal in the left hippocampus (P < 0.05).

Conclusion

The study concludes that changes in WM microstructure and its structural network may contribute to the development of severe neuropsychiatric symptoms in SLE patients. These changes may be the basis of brain damage that leads to the development of NPSLE from SLE without major neuropsychiatric manifestations.