The prevalence, subtypes and associated factors of hyperuricemia in lupus nephritis patients at chronic kidney disease stages 1-3

Deep learning model to predict lupus nephritis renal flare based on dynamic multivariable time-series data

OBJECTIVES

To develop an interpretable deep learning model of lupus nephritis (LN) relapse prediction based on dynamic multivariable time-series data.

DESIGN

A single-centre, retrospective cohort study in China.

SETTING

A Chinese central tertiary hospital.

PARTICIPANTS

The cohort study consisted of 1694 LN patients who had been registered in the Nanjing Glomerulonephritis Registry at the National Clinical Research Center of Kidney Diseases, Jinling Hospital from January 1985 to December 2010.

METHODS

We developed a deep learning algorithm to predict LN relapse that consists of 59 features, including demographic, clinical, immunological, pathological and therapeutic characteristics that were collected for baseline analysis. A total of 32 227 data points were collected by the sliding window method and randomly divided into training (80%), validation (10%) and testing sets (10%). We developed a deep learning algorithm-based interpretable multivariable long short-term memory model for LN relapse risk prediction considering censored time-series data based on a cohort of 1694 LN patients. A mixture attention mechanism was deployed to capture variable interactions at different time points for estimating the temporal importance of the variables. Model performance was assessed according to C-index (concordance index).

RESULTS

The median follow-up time since remission was 4.1 (IQR, 1.7-6.7) years. The interpretable deep learning model based on dynamic multivariable time-series data achieved the best performance, with a C-index of 0.897, among models using only variables at the point of remission or time-variant variables. The importance of urinary protein, serum albumin and serum C3 showed time dependency in the model, that is, their contributions to the risk prediction increased over time.

CONCLUSIONS

Deep learning algorithms can effectively learn through time-series data to develop a predictive model for LN relapse. The model provides accurate predictions of LN relapse for different renal disease stages, which could be used in clinical practice to guide physicians on the management of LN patients.

Association between serum uric acid level and systemic lupus erythematosus kidney outcome: An observational study in Southern Chinese population and a meta-analysis

OBJECTIVE

The study aimed to explore the effect of serum uric acid (SUA) level on the progression of kidney function in systemic lupus erythematosus (SLE) patients.

METHODS

A total of 123 biopsy-proven lupus nephritis (LN) patients were included in this retrospective observational study. Cox proportional hazard regression analyses as well as restricted cubic spline analyses were performed to identify predictors of renal outcome in LN patients. We also performed a systematic review and meta-analysis for SUA and overall kidney outcomes in SLE patients.

RESULTS

Based on the laboratory tests at renal biopsy, 72 (58.5%) of the 123 patients had hyperuricemia. The median (IQR) follow-up duration was 3.67 years (1.79-6.63 years), and a total of 110 (89.4%) patients experienced progression of LN. Increased serum uric acid level, whether analyzed as continuous or categorical variable, was associated with higher risk of LN progression in Cox proportional hazard regression model (hazard ratio [HR]: 1.003, 95% confidence interval [CI]: 1.001-1.005; HR: 1.780, 95% CI: 1.201-2.639, respectively). This relationship maintained in women (HR: 1.947, 95% CI: 1.234-3.074) but not men (HR: 2.189, 95% CI: 0.802-5.977). The meta-analysis showed a similar result that both continuous and categorical SUA were positively associated with the risk of kidney function progression in LN (weighted mean difference [WMD]: 1.73, 95% CI: 0.97-2.49; odds ratio [OR]: 1.55, 95% CI: 1.20-2.01, respectively).

CONCLUSIONS

Our study found overall and especially in women that higher SUA in LN patients were associated with increased risk of renal progression. Meta-analysis yielded consistent results. Future studies are required to establish if uric acid can be used as a biomarker for risk assessment and/or as a novel therapeutic target in SLE.

Peripheral CD4 +CD8 + double positive T cells: A potential marker to evaluate renal impairment susceptibility during systemic lupus erythematosus

Lupus nephritis (LN) has a high incidence in systemic lupus erythematosus (SLE) patients, but there is a lack of sensitive predictive markers. The purpose of the study was to investigate the association between the CD4 +CD8 + double positive T (DPT) lymphocytes and LN. The study included patients with SLE without renal impairment (SLE-NRI), LN, nephritic syndrome (NS), or nephritis. Peripheral blood lymphocyte subsets were analyzed by flow cytometry. Biochemical measurements were performed with peripheral blood in accordance with the recommendations proposed by the National Center for Clinical Laboratories. The proportions of DPT cells in the LN group were significantly higher than that in the SLE-NRI group ( t=4.012, P<0.001), NS group ( t=3.240, P=0.001), and nephritis group ( t=2.57, P=0.011). In the LN group, the risk of renal impairment increased significantly in a DPT cells proportion-dependent manner. The risk of LN was 5.136 times (95% confidence interval, 2.115-12.473) higher in cases with a high proportion of DPT cells than those whose proportion of DPT cells within the normal range. These findings indicated that the proportion of DPT cells could be a potential marker to evaluate LN susceptibility, and the interference of NS and nephritis could be effectively excluded when assessing the risk of renal impairment during SLE with DPT cell proportion.

Clinicopathological Patterns and Outcomes in Patients with Lupus Nephritis and Hyperuricemia

A limited number of large cohort studies have reported the clinicopathological characteristics and prognosis of patients with lupus nephritis (LN) and hyperuricemia (HUA). In this retrospective cohort study, 1297 LN patients were enrolled from January 1996 to December 2011 in the First Affiliated Hospital of Sun Yat-Sen University, and HUA occurred in 649 (50.04%) of these 1297 LN patients. Compared to patients without HUA, those with HUA presented with higher blood pressure and triglyceride levels, lower hemoglobin and serum albumin levels, worse renal function, more severe hematuria and proteinuria, higher lupus activity, and more positive antiphospholipid antibody. Pathologically, HUA cases presented more crescents, a higher degree of mesangial matrix, endothelial cell proliferation, and inflammatory cell infiltration. During the 52-month follow-up, the 5-year and 10-year incidence rates of renal endpoint events were 11.1% and 19.5% in the HUA group, and 8.3% and 13.8% in the non-HUA group, respectively (p = 0.073). In addition, the 5-year and 10-year mortality rates did not differ significantly between the HUA (12.0% and 18.2%) and non-HUA (12.2% and 17.5%) groups, respectively. This study verified that HUA was not an independent risk for poor clinical outcomes, and steroids that delay the deterioration of renal function did not affect the survival of these patients.

Uric acid and kidney damage in systemic lupus erythematosus

Systemic lupus erythematosus (SLE) is an autoimmune disease that affects multiple organs; lupus nephritis (LN) is one of the most severe complications of SLE. In the kidneys, an intense inflammatory reaction affects the glomeruli and tubular interstitium. Uric acid has been considered a key molecule in the pathogenesis of some conditions such as metabolic syndrome, hypertension, and kidney disease as it is produced by injured cells and promotes immune-inflammatory responses. In this regard, high serum uric acid concentrations may be involved in the activation of some inflammatory pathways, associated with kidney damage in SLE. Therefore, the purpose of this article was to review the main physiological mechanisms and clinical data on the association between serum uric acid and kidney damage in SLE. Scientific evidence indicates that hyperuricemia has the potential to be an adjuvant in the development and progression of kidney manifestations in SLE. Uric acid may promote the activation of inflammatory pathways and the formation and deposition of autoantibodies in kidneys, leading to a reduction of glomerular filtration rate. Other potential mechanisms of this association include the presence of polymorphisms in the urate transporters, metabolic syndrome, use of some medications, and other situations associated with a reduced renal excretion of uric acid.

Serum Uric Acid is Associated with Renal Prognosis of Lupus Nephritis in Women but not in Men

Lupus nephritis (LN) is a major complication of systemic lupus erythematosus. Early intervention in lupus nephritis improves prognosis. There is an association between hyperuricemia and lupus nephritis; nevertheless, the sex-specific role of uric acid in lupus nephritis remains unclear. We retrospectively analyzed 578 patients diagnosed with LN by renal biopsy. We determine the relationship of serum uric acid to progression of LN using Kaplan–Meier survival analyses and Cox proportional hazards models. The primary end point was LN progression defined as the initiation of dialysis or kidney transplantation. Men had higher mean serum uric acid levels than did women. Every 1 mg/dL increase in baseline uric acid level increased the risk of LN progression by 15.1%. The serum uric acid level was an independent risk factor for LN progression in women (hazard ratio [HR], 1.158; confidence interval [CI], 1.018–1.317; p = 0.028) but not in men (HR, 1.499; CI, 0.964–2.331; p = 0.072). Sensitivity analysis involving serum uric acid terciles generated consistent and robust results. Serum uric acid level was an independent risk factor for LN progression in women but not in men.

Polymorphism of ABCG2 Gene in Hyperuricemia Patients of Han And Uygur Ethnicity with Phlegm/Non-Phlegm Block in Xinjiang, China

Background

This study investigated the relationship between hyperuricemia (with phlegm/non-phlegm block) and ABCG2 gene polymorphism in Han and Uygur people from Xinjiang, China.

Material/Methods

We recruited 600 hyperuricemia patients with phlegm/non-phlegm block. Genomic DNA was extracted from the whole blood. Gene polymorphism was classified by SnaPshot method.

Results

The SNP loci rs2725220 and rs2231137 of the ABCG2 gene, but not rs2231142, were significantly different between patients with non-phlegm block and phlegm block (P<0.05). The rs2231142 allele G was the protective factor in Uygur hyperuricemia patients. In both Han and hyperuricemia patients, the rs2725220 allele G was a protective factor and the rs2231137 allele C was a risk factor. For non-phlegm-block hyperuricemia, the rs2231142 and rs2231137 genotypes were significantly different between Uygur and Han patients (P<0.05). The rs2231142 allele G was 1.563 times higher in the Uygur patients compared with Han, and rs2231137 allele C was 1.673 times higher in the Uygur patients compared with the Han. For phlegm-block hyperuricemia, rs2231142 allele G was 1.397 times higher in the Uygur patients compared with the Han.

Conclusions

ABCG2 gene rs2231137 with more allele C tends to be phlegm-block type and rs2725220 with more allele G tends to be non-phlegm-block type. In the Uygur hyperuricemia patients, ABCG2 gene rs2231142 with more allele G tends to be non-phlegm-block type. Allele C of rs2231137 and allele G of rs2231142 in ABCG2 gene are more likely to be found in the Uygur people.