Renal function in patients with AQP4 antibody-positive neuromyelitis optica spectrum disorder


 Background: Increasing evidence has demonstrated that AQP4-IgG can cause damage to kidney in AQP4 antibody-positive neuromyelitis optica spectrum disorder (NMOSD) patients. However, renal impairment in AQP4 antibody-positive NMOSD patients have not been conducted thus far. This study aimed to test and evaluate the changes of estimated glomerular filtration rates (eGFR) in NMOSD patients.Methods: We performed a cross-sectional study that recruited 93 NMOSD patients, 56 multiple sclerosis (MS) patients and 100 age- and sex-matched healthy controls (HCs), at the Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, China during November 2012 and November 2017. eGFR, including eGFRCr, eGFRCysC and eGFRCr-CysC were conducted for all study individuals.Results: NMOSD patients has mildly, but statistically lower eGFRCysC value at both the acute and stable phases than MS patients and HCs, respectively. (NMOSD vs. MS: adjusted model: p=0.034 at the acute phase, p<0.001 at the stable phase; NMOSD vs. HCs: adjusted model: p=0.001 at the acute phase, p<0.001 at the stable phase). Meanwhile, there is no significant difference of eGFRCysC between MS patients and HCs at both the acute and stable phases. In addition, for NMOSD patients with low-dose glucocorticoid therapy, eGFRCysC<90ml/min/1.73 m2 was more likely to occur at the stable phase, compared to the acute phase (OR =4.68, 95% CI 1.19 to 18.34).Conclusions: Mildly impaired renal function might occur in patients with AQP4 antibody-positive NMOSD compared to MS.


Background
Neuromyelitis optica spectrum disorder (NMOSD) is an in ammatory disease of central nervous system (CNS) associated with a highly speci c serum biomarker and pathogenic autoantibody, aquaporin-4 immunoglobulin G (AQP4-IgG) [1].Recently, increasing evidences have showed that AQP4-IgG could cause damage to the organ outside of CNS [2], such as stomach, airway, muscle, the organ of corti and kidney, which all of them were found the expression of AQP4 [3]. However, it is unknown that whether NMOSD patients would have changes of renal function or not. AQP-4, as known as the membrane bound water channel protein, plays a role in water reabsorption from renal tubular uid. In 1993, Fushimi K et.al rst discovered that the AQP-4 were expressed at the basolateral membrane of collecting duct principal cells in the kidney [4]. After that, reduced urine concentration capacity was found in AQP-4 knockout mice in 1998 [5], which this phenomenon was also noticed in NMOSD patients recently [6]. In addition, two case reports have demonstrated that AQP4-IgG could cause damage to kidney [7,8]. However, it is still unknown about the changes of glomerular ltration rate (GFR) in NMOSD patients.
GFR is universally regarded as the best overall index of renal function. And the equations that estimates GFR were widely recommended to diagnose and evaluate chronic kidney disease [9]. As one of the equations of the estimated GFR (eGFR), the Cockcroft-Gault (CG) formula depends on the patient's age, sex, body weight and serum creatinine (Cr). It can result in the underestimation of renal function, especially in female and underweight person, which might be more obvious in Asian than in non-Asian [10]. Meanwhile, the Modi cation of Diet in Renal Disease (MDRD) formula is more appropriate for patients with chronic kidney disease, since its main limitation is underestimation of measured GFR at higher levels [11]. However, the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) is recommended for general clinical use by the National Kidney Foundation and it could provide more accurate eGFR, especially at higher levels of eGFR [12]. The CKD-EPI Cr equation, which is based on serum Cr levels, is well accepted as the eGFR in clinical study. However, Cr-based eGFR is affected by patient's age, gender, race, weight, height and body surface area, and particularly by muscle mass and dietary intake [13]. In contrast, serum cystatin C (CysC) is not affected by gender and muscle mass, which provides a more accurate measure of renal function [14] and shows a higher correlation with standard measures of GFR [15]. In addition, several reports also demonstrated that eGFR based on the combination of both standardized serum Cr and CysC was more accurate than eGFR based on either marker alone [16], but was still in uenced by serum creatinine. Therefore, we chose CKD-EPI formula to estimate GFR from serum cystatin C.
To the best of our knowledge, comparison of eGFR between NOMSD and multiple sclerosis (MS) patients has not been reported yet. The objection of this study is to evaluate eGFR between AQP4 antibody-positive NMOSD and MS patients based on eGFR, estimated using Cr, CysC, and both.

Study design and samples
This cross-sectional study was conducted according to the principles expressed in the Declaration of Helsinki and approved by the Medical Ethics Committee of the Third A liated Hospital of Sun Yat-sen University. All study participants gave written informed consent for research and publication.  [17], respectively. Experienced neurologists identi ed patients with NMOSD/MS at the acute phase(Ap-NMOSD/Ap-acute) according to following criteria: worsening of existing neurologic symptoms with an objective change on neurologic examination that lasted for more than 24 hours or a new onset of neurologic symptoms, symptoms and signs attributed to NMOSD/MS instead of other causes, and onset preceded by at least 30 days of clinical stability [18]. Patients with NMOSD/MS at the stable phase(Sp-NMOSD/Sp-MS) meet the following criteria: oral administration of stable-dose immunosuppressive drugs for relapse prevention (Sp-NMOSD patients: the stable-dose of immunosuppressant drug and low dose of prednisone treatment, ranging from 5 to 10 mg/d; Sp-MS patients: the stable-dose of immunosuppressant drugs).
All patients were scored using the Expanded Disability Status Scale (EDSS) [15] and tested for AQP4 antibody using cell-based assay. Among these 93 NMOSD patients at the acute phase, 46 of them had never received glucocorticoid or other disease-modifying immunosuppressive therapy (Ap-NMOSD-nonG) at least three months prior to the admission, and the other 47 patients still had low-dose prednisone treatment (5-10 mg/d) and stable dose of immunosuppressant drugs for relapse prevention (Ap-NMOSD-G). Data at the stable phase were collected from 59 NMOSD with low-dose glucocorticoid therapy and 39 MS patients without steroids therapy. In addition, we also recruited 100 age-and sex-matched healthy Chinese Han controls (HCs) from the Department of Medical Examination Center. None of the study subjects had urinary tract infections, stone of kidney or urinary tract, obviously abnormal thyroid function or smoking history.

Assessment of eGFR
Serum Cr and CysC levels were measured in all NMOSD and MS patients before pulse steroid therapy on the next day after admission. The estimated glomerular ltration rate (eGFR), including eGFR Cr , eGFR CysC and eGFR Cr−CysC , were calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula [12].

Statistical analysis
Statistical analyses were performed using SPSS 23. Chi-square and Fisher's exact test were used for categorical variables. For continuous variables, Mann-Whitney U test or student T test was used for comparisons among two groups. For comparisons among three groups, Kruskal-Wallis test or one-way Analysis of Variance (ANOVA) was used and then adjusted by Bonferroni. Comparisons using adjusted models were based on linear regression models. Results were considered signi cant if p < 0.05.

Results
Baseline Characteristics Table 1 summarizes the baseline patient characteristics. According to the core clinical characteristics during the acute phase, 50 (53.76%) NMOSD patients had acute myelitis, 30(32.26%) patients had optic neuritis, and 13 (13.98%) patients had other core clinical characteristics (acute brainstem syndrome or other intracranial lesions). NMOSD patients have longer disease durations than MS patients (p = 0.020). The EDSS scores of NMOSD patients were signi cantly higher than MS patients (p < 0.001). The results are presented in Table 2 and Fig  In adjusted models, the eGFR CysC level of all NMOSD patients were statistically lower than MS patients at the acute phase (adjusted p = 0.002). There was no difference of eGFR CysC level between NMOSD patients and HCs. The eGFR Cr−CysC levels of NMOSD patients were also statistically lower than MS patients in the adjusted models (adjusted p = 0.023). Additionally, statistically lower of CysC level was observed in all Ap-NMOSD patients than Ap-MS patients (adjusted p = 0.019).
To eliminate the effect of the usage of low-dose corticosteroids, we chose 46 Ap-NMOSD-nonG patients and found that in both the unadjusted and adjusted models, the eGFR CysC level of Ap-NMOSD-nonG patients was statistically lower than Ap-MS patients (unadjusted p = 0.010, and adjusted p = 0.034, Table 3 and Fig. 2B). Compared to HCs, the eGFR CysC level of Ap-NMOSD-nonG patients was also statistically lower than HCs in the adjusted models (adjusted p = 0.001). In addition, eGFR Cr−CysC level was statistically lower in Ap-NMOSD-nonG patients than Ap-MS patients in the adjusted models (adjusted p = 0.011).

= refers to the comparison between MS and HCs
Comparisons between Ap-NMOSD-G patients and Ap-NMOSD-nonG patients were conducted to investigate the changes of eGFR levels treated by low-dose glucocorticoid. No signi cant differences were observed in the eGFR levels, CysC and Cr levels between the two study cohorts in the unadjusted model (Table  S1 and Fig. 2E). However, higher eGFR CysC levels and lower serum CysC levels were observed in the patients with glucocorticoids therapy in the adjusted model (p = 0.048 eGFR Cys , p = 0.002 CysC).

Comparisons among NMOSD, MS patients at the stable phase and HCs
In both the unadjusted and adjusted models, the eGFR CysC level of Sp-NMOSD patients was statistically lower than Sp-MS patients (unadjusted p = 0.013, and adjusted p < 0.001). Compared to HCs, the eGFR CysC level of Sp-NMOSD patients was statistically lower in both the unadjusted and adjusted models (unadjusted p = 0.001, and adjusted p < 0.001).
In addition, the eGFR Cr−CysC level of Sp-NMOSD patients were statistically lower than Sp-MS patients in both the unadjusted and adjusted models (unadjusted p = 0.042, and adjusted p < 0.001, Table 4 and Fig. 2C).

= refers to the comparison between MS and HCs
Comparisons between acute-and stable-phase patients with NMOSD.
In both the unadjusted and adjusted models, the eGFR CysC level of Ap-NMOSD-G patients was signi cantly higher than Sp-NMOSD patients (unadjusted p = 0.007, and adjusted p < 0.001, Table 5 and Fig. 2D). In addition, the eGFR Cr−CysC level of Ap-NMOSD-G patients was also statistically higher than Sp-NMOSD patients (unadjusted p = 0.034, and adjusted p < 0.001). Moreover, the mean serum CysC level of Ap-NMOSD-G patients was also signi cantly lower than the serum CysC level of Sp-NMOSD in both the unadjusted and adjusted model (unadjusted p < 0.001, and adjusted p < 0.001).  Table 5 The comparison of eGFR levels between acute-phase NMOSD-G patients and stable-phase NMOSD patients.

Variables
Ap-NMOSD-G a (n = 47) Note: NMOSD, neuromyelitisoptica spectrum disorder; Ap-NMOSD-G: NMOSD patients at acute phase received oral administration of low-dose glucocorticoid drugs and stable-dose immunosuppressant drugs for relapse prevention; Sp-NMOSD: NMOSD patients at stable phase received oral administration of low-dose glucocorticoid drugs and stable-dose immunosuppressant drugs for relapse prevention; HC, healthy control; CysC, Cystatin C; Scr, serum creatinine; eGFR Cr , creatinine based estimated glomerular ltration rate; eGFR CysC , cystatin C based estimated glomerular ltration rate; eGFR Cr−CysC , creatinine and cystatin C based estimated glomerular ltration rate; SD, standard deviation; IQR: interquartile range *** P < 0.001, ** P < 0.01, * P < 0.05 a and b : The blood samples were collected from patients before pulse steroid therapy on the next day after admission.

Discussion
To the best of our knowledge, this study is the rst to compare eGFR between AQP4 antibody-positive NOMSD and MS patients. In this study, we found a slight reduction of eGFR (especially eGFR CysC ) in NMOSD patients, compared to MS patients and HCs, although the average eGFR were within the normal range in all groups.
eGFR is a universal marker of renal function [9], and decreased GFR represents the presence of glomerular small vessel disease [19]. The estimation of renal function using serum CysC levels has clinical advantages as it is not in uenced by the blood concentrations, muscle mass, gender, or age [20,21]. Therefore, to a certain degree, eGFR CysC may be superior to eGFR Cr for evaluating renal function in NMOSD. However, studies have showed that glucocorticoid administration was associated in a dose-dependent fashion with increased CysC values, leading to a dose-dependent underestimation of GFR calculations based on CysC [22]. Thus, steroid-free status may improve the accuracy of the CysC-based equations. In order to reduce potential bias of results caused by glucocorticoids treatment, we analyzed the serum CysC data from NMOSD patients who had never received steroids or other disease-modifying immunosuppressive therapy at least three months prior to the admission.
The AQP-4 is expressed in principal cells of collecting duct of kidney, where about 15% of tubular uid was reabsorbed through AQP-2, AQP-3 and AQP-4. Both AQP-4 and AQP-3 are located in the basolateral membrane of principal and responsible for the transportation of water, which is reabsorbed by AQP-2 from tubular uid into intercellular uid [23]. If either AQP-4 or AQP-3 was dysfunction, the other one might compensate to keep osmotic pressure in tubular uid in some degree. And these theories are con rmed by Chou CL et.al, who has reported the mild defect in urinary concentrating ability in AQP-4 knockout mice [5].
Furthermore, it was also discovered that NMOSD patients show a urinary concentrating defect [6]. Thus, the deposition of AQP4-IgG and the loss of AQP4 may potentially lead to damage in kidney. In our study, we found that the mildly lower eGFR was found in NMOSD, but still in the normal range. This phenomenon may be cause by the fact that AQP-4 might play a major role in concentration and dilution of renal tubules, but not in glomerular ltration. Besides, there are so many factors affecting GFR in complex ways, which needs further development.
In a recent study, higher CysC concentration was observed in renal transplant recipients on low-dose prednisone treatment (5-10 mg/d) compared to those on steroid-free immunosuppression [24]. Interestingly, in our study, we did not nd higher CysC concentrations in acute-phase NMOSD patients with low-dose glucocorticoid therapy before acute attack, compared to acute-phase NMOSD patients without glucocorticoid therapy before acute attack. The greater increase of CysC in stable-phase NMOSD with glucocorticoid therapy may be explained by a result of the decrease in the in ammatory process. Besides, 27.56% of stable-phase NMOSD patients with low-dose glucocorticoid therapy had eGFR CysC < 90 ml/min/1.73 m 2 , which implies a substantial loss of glomeruli and impaired renal reserve. In addition, no changes of renal function occurred in MS patients in this study, which was contrary to a previously reported result [25] on progressive type MS. The difference in renal function may be due to the different subtypes of MS. The enrolled MS patients in this study were relapsing-remitting MS, while the previous study was chronic (both primary and secondary) progressive MS patients.
Several limitations need to be addressed in our study. In the present study, we detected differences in eGFR, but results of all groups are in the normal range. However, these still indicated that reduced renal function has occurred in patients with AQP4 antibody-positive NMOSD compared to MS patients and HCs, suggesting that more attention should be paid to protect the renal function throughout the disease stage, especially when applying nephrotoxic drugs.
Additionally, the number of patients was relatively small and blood samples were just obtained once. However, to the best of our knowledge, this is the rst study to investigate renal function in NMOSD patients. Our hospital is the major research unit focused on this condition in China. More cases will be collected and blood samples will be obtained at different time points for the future investigation. Furthermore, additional measurement of proteinuria and BMI may improve the sensitivity and speci city of our assessment of renal function, which can be investigated in the future.

Conclusion
In conclusion, our results indicate that reduction of eGFR (especially eGFR CysC ) occurs in patients with AQP4 antibody-positive NMOSD compared to MS patients and HCs. The ndings suggest that we may need to pay attention to the protection of renal function and to use nephrotoxic drugs while treating NMOSD.