Inuenza Virus Vaccination in Pediatric Nephrotic Syndrome Induces a Signicant Reduction of Relapse and Inuenza Virus Infection: A Nationwide Survey

Although vaccination may precipitate relapses of nephrotic syndrome (NS) in children with idiopathic NS, no data are available regarding NS activity regarding inuenza (u) virus infections and NS relapses after receiving inactivated u vaccines. We conducted a nationwide study of children aged 6 months to 15 years with idiopathic NS to assess the relationship between NS relapse, u vaccination, and u infections. We used a multivariate Poisson regression model (MPRM) to calculate the risk ratio (RR) for u infection and for NS relapse in children with and without u vaccination. Data of 306 children were assessed. The MPRM in all 306 children showed a signicantly lower RR for u infection (RR: 0.21, 95% condence interval [CI]: 0.11–0.38) and for NS relapse (RR: 0.22, 95% CI: 0.14–0.35) in children receiving u vaccination compared with unvaccinated children. In an additional MPRM only among 102 children receiving u vaccination, they had a signicantly lower risk for NS relapse during the post-vaccination period (RR: 0.31. 95% CI: 017–0.56) compared with the pre-vaccination period. Although our study was observational, based on the favorable results of u vaccinations regarding u infections and NS relapse, the vaccine may be recommended for children with NS.


Introduction
Idiopathic nephrotic syndrome (NS) requires long-term follow-up to manage relapse. Over half of children develop frequently relapsing NS (FRNS) or steroid-dependent NS (SDNS) 1 . To prevent NS relapse, these patients require prolonged steroid therapy and immunosuppressive agents 2,3 . These children have been the state of immunocompromised, in which they have the potential for recurrent or severe infections.
Children with chronic renal disease have a higher risk for severe in uenza ( u) infections than healthy children [4][5][6] . Administering u vaccines is effective in reducing the prevalence of u infections and the risk of severe u infection. KDIGO guidelines suggest that children with NS should receive u vaccination to reduce the risk of serious infections 7 . Pediatric patients with NS can be protected against u infection using an inactivated in uenza vaccine, even if they are being treated with glucocorticoids or other immunosuppressive drugs 4 . However, as an immunogenic stimulus, immunizations have been reported to trigger NS relapse in children [8][9][10][11][12][13] . Although inactivated u vaccines could precipitate NS relapse 14,15 , available data regarding the relative risk of NS relapse related to u vaccines have been scarce until recently. We previously evaluated NS relapse after 104 children with NS received an inactivated subunitantigen in uenza virus vaccine 16 . Compared with the NS relapse rate during the pre-vaccination period as a control, the rate within 1 month after receiving u vaccination was slightly but non-signi cantly increased (1.19 vs. 1.23 times/person-year, risk ratio: 1.04, 95% con dence interval: 0.82-1.89, p = 0.88).
Limitations of our previous research included that it was a retrospective study in a single tertiary center and included no data of u infections. Further studies using a larger sample at multiple centers were needed to investigate not only NS relapses but also u infections after in uenza vaccination.
Here, we report the ndings of a nationwide cohort study among Japanese children with idiopathic NS in which we focus on the relationships among NS relapses, u vaccines, and u infections.

Participant characteristics.
Overall, 105 of 388 institutions responded to the third questionnaire, and we collected data for 306 children. Figure 1 shows a ow chart of the study design and study populations.
We divided the children into two groups; those who received u vaccination and those who did not receive u vaccination. These children's clinical characteristics according to both the second and third questionnaires are shown in Table 1.
Flu infection. Compared with children who received u vaccination, the u infection rate among children without u vaccination was signi cantly higher (12.7 vs. 25.4%, p=0.01) than in their vaccinated counterparts. The diagnostic method of u infections and the in uenza strain were evaluated in 62 children (68 u infection cases). All 13 infected children who received u vaccination and 54 of 55 infected children who did not receive u vaccination were diagnosed using rapid antigen detection tests and nasopharyngeal swab samples; 6 vaccinated and 25 unvaccinated children had in uenza A infection, and 7 vaccinated and 29 unvaccinated children had in uenza B infection. Only one unvaccinated patient was clinically diagnosed with u infection. Figure 2 shows that the seasonal distribution of u infections during the 2015/16 season was from January to May, and the highest number of u infections was in February.
Results of multivariate analysis to evaluate the risk ratio for u infection. Table 3 shows the results of multivariate analysis in the 306 children who received and did not receive u vaccination.
The multivariate Poisson regression model (MPRM) showed statistical evidence of a higher risk ratio for all u infections in February than in other months (risk ratio: 16.58. 95% con dence interval: 10.23-26.89), and a lower risk ratio for u infections in children with older age at the onset of NS (risk ratio: 0.90. 95% con dence interval: 0.84-0.97) and in children who received u vaccination (risk ratio: 0.21. 95% con dence interval: 0.11-0.38) (Table 3a). According to u antigen, the data of risk ratio for in uenza B infection were similar to those for all in uenza strains (Table 3c); however, the age at onset of NS was not signi cantly related to in uenza A infection (Table 3b). Flu infection in children with NS was not signi cantly related to sex or history of steroid-resistant NS (SRNS).

NS relapse.
In total, 100 children (32.7%) had NS relapse between 1 May 2015 and 31 April 2016, and there were 190 total NS relapses ( Table 2). The proportion of children with NS relapse and the number of NS relapses was slightly higher in children who were not vaccinated than their vaccinated counterparts, but these were not statistically signi cant. There was no seasonal distribution of NS relapse ( Figure 3).
Results of multivariate analysis to evaluate the risk ratio for NS relapse.
In Table 4, the MPRM for 306 children that received and did not receive u vaccination showed a signi cantly lower risk ratio for NS relapse in children who received u vaccination (risk ratio: 0.22. 95% con dence interval: 0.14-0.35), and a higher risk ratio for NS relapse in children with a history of FRNS or SDNS or SRNS (risk ratio: 2.80. 95% con dence interval: 1.90-4.11). Sex and age at onset of NS were not independent risk factors for NS relapse in this multivariate analysis. Table 5 shows the results of the MPRM conducted among the 102 children who received u vaccination to evaluate the risk of NS relapse during the period after u vaccination in comparison with before u vaccination. Taking glucocorticoids at the time of u vaccination was a risk factor signi cantly associated with NS relapse (risk ratio: 3.17. 95% con dence interval: 1.74-5.80). Compared with the prevaccination period, a signi cantly lower risk for NS relapse during the post-vaccination period (risk ratio: 0.31. 95% con dence interval: 017-0.56). There was no signi cant relationship between NS relapse and sex or being on cyclosporine or mizoribine at the time of u vaccination.

Discussion
In the present study, we conducted a nationwide cohort study comprising 306 children with idiopathic NS, and showed that children who received u vaccination had signi cantly fewer u infections and NS relapses than those who did not receive u vaccination. Additionally, the former group showed a signi cantly lower NS relapse risk during the post-vaccination period compared with the pre-vaccination period.
In this cohort, we examined the relationship among u vaccination, u infections, and NS relapses. Multivariate analysis revealed a signi cantly lower risk ratio for NS relapse in children who received u vaccination. It has long been considered that vaccination itself can precipitate NS relapse as an immunogenic stimulation; however, the in-depth mechanism of immunogenic pathogenesis has not been elucidated. Several clinical studies have clinically suggested that the varicella vaccine 9,10 , meningococcal C conjugate vaccine 11 , and hepatitis B virus vaccine 12  We found that being on glucocorticoids when receiving u vaccination was a risk factor signi cantly associated with NS relapse among children with NS. Pediatric patients treated with steroids at the time of u vaccination would means on the state soon after the last NS relapse. Thus, NS relapses might be easy to trigger in children who are treated with steroids at the time of receiving u vaccination owing to taking glucocorticoids being a marker of recent NS relapse.
The present study has several limitations. First, bias existed between children with and without u vaccination, with a signi cant difference in history of FRNS among children. We could not address these biases completely, even with adjustment in multivariate analysis, because this was an uncontrolled study. Each physician may have adopted the type of NS as a selection criterion in their decision to recommend u vaccination to patients. Regardless, we evaluated many in uencing factors for u infections and NS relapses so as to adjust for confounding as much as possible. Nevertheless, we should recognize this difference as confounding in the present analysis and consider that the results regarding a lower risk for NS relapse in children who receive u vaccination cannot be applied to all children with NS. Moreover, we expect that some unrecognized bias existed between children with and without u vaccination. Second, we could not collect data of all events during the observation period. With further details, including data of whether fever, rhinorrhea, coughing, and vomiting resulting from both bacterial and viral infections triggered NS relapse, we might be able to conclude that u vaccination could prevent NS relapse in terms of direct management of various infections. Third, we had no data regarding the precise interval between the timing of u vaccination and NS relapses in each individual. Because the side effects of the inactivated subunit-antigen u vaccine generally occur within 2-4 weeks of administration, NS relapses within 1 month after vaccination were possibly triggered by vaccination itself. However, differences in the state of NS activity during the pre-and post-vaccination periods are worth evaluating to determine whether there is a signi cantly lower risk for NS relapse in the post-vaccination period as compared with the pre-vaccination period. Finally, vaccine policies in each center are not uniform. Some facilities have a policy of administering inactivated subunit-antigen u vaccines to children with NS depending on their steroid dose, and other centers have different policies depending on the period from onset of NS or last NS relapse.
In conclusion, although our study was observational, based on the favorable results of u vaccination against u infection and NS relapse, the vaccine may be recommended for children with NS regardless of immunosuppressant use at the time of vaccination.

Methods
Study design and populations.
The Japanese Pediatric Survey Holding Information of Nephrotic Syndrome (JP-SHINE) study, a nationwide cohort study, was established by the Japanese Study Group of Renal Disease in Children. The rst survey was sent on 4 April 2014 to 1860 institutions throughout Japan, including all universities and children's and general hospitals with more than 20 beds. Patients were eligible if they were newly diagnosed with idiopathic NS between 1 January 2010 and 31 December 2012, were between 6 months and 15 years old, and if they had been treated for up to 3 years during this time. Patients with congenital nephrotic syndrome or nephrotic syndrome secondary to nephritis were excluded. The rst questionnaire, which was designed to record the number of children with idiopathic NS at each institution and their basic information, was sent to 1860 institutions. Of these, 1050 institutions (56.5%) responded to the rst questionnaire 20 . The second questionnaire regarding patients who ful lled the eligibility criteria was sent to 388 institutions and recorded information including patients' basic characteristics, renal biopsy, idiopathic NS complications, steroid therapy side effects, and prognosis. We collected data for 999 patients 21 . In the present study, we recorded age when the patient received the rst inactivated subunitantigen u vaccine, the total number of u vaccinations received, total number of u infections, total number of NS relapses, rituximab (RTX) use at the rst u vaccination, and immunosuppressant use at the rst u vaccination (cyclosporine, mycophenolate mofetil, mizoribine, cyclophosphamide, tacrolimus) between 1 May 2015 and 31 April 2016.

De nitions
In the present study, the de nitions of general conditions in pediatric NS were according to clinical guidelines issued by the Japanese Society for Pediatric Nephrology 22,23 . Idiopathic NS in children was de ned as hypoalbuminemia (serum albumin levels ≤ 2.5 g/dL) and severe proteinuria (≥ 40 mg/h/m 2 in pooled nighttime urine or an early morning urine protein creatinine (Cr) ratio > 2.0 g/g Cr). Complete remission was de ned as a urine protein creatinine ratio < 0.2 g/g Cr or ≤ negative protein on early morning urine dipstick testing for 3 consecutive days. SSNS was de ned as complete remission in < 4 weeks after starting daily prednisolone therapy. NS relapse was de ned as ≥ 3 + protein on early morning urine dipstick testing for 3 consecutive days. FRNS was de ned as two or more relapses within 6 months after initial response or four or more relapses within any consecutive 12-month period. SDNS was de ned as two consecutive relapses occurring during steroid therapy or within 2 weeks of treatment cessation.
SRNS was de ned as the absence of complete remission after a 4-week course of oral prednisolone 60 mg/m 2 /day. RTX therapy duration was de ned as the period from the day of rituximab administration to the day of B-cell recovery (CD19 + B-cell count ≥ 1% of total lymphocytes).
In uenza virus infection was diagnosed using rapid antigen detection tests with a nasopharyngeal swab sample or was clinically diagnosed by an attending pediatrician. In Japan, children younger than 13 years old generally receive u vaccination twice per year and those over 13 years old are vaccinated once per year. In children who received two vaccinations in the same year, we de ned the " rst u vaccine" as their Ethics.
This study was conducted according to the principles of the Declaration of Helsinki and following the ethical guidelines for Medical and Health Research Involving Human Subjects of the Ministry of Health, Labour and Welfare in Japan. Yokohama City University Hospital's central ethics review board approved this study (approval number: 1509001). This ethics committee clearly stated that the researchers did not need to obtain informed consent, because complete data in this study were collected from patient medical records. In accordance with this statement, informed consent was not required from patients or their parents in this study. According to the guidelines and the institutional ethics review board for the patients' bene t, the protocol was displayed publicly in a poster at each hospital, and each patient or their family had the opportunity to refuse to be included in this analysis.

Statistical analysis.
We used t-tests for continuous values and the chi-square and Fisher's exact test for categorical values. All data were expressed as mean ± standard deviation or number (percentage). We used a MPRM for children who received or did not receive u vaccination to calculate the risk ratio for u infection, adjusted for receiving or not receiving u vaccination, sex, age at NS onset, history of FRNS or SDNS or SRNS, u infection in February or another month, and the risk ratio for NS relapse adjusted for whether the patients received u vaccination, sex, age at NS onset, and history of FRNS or SDNS or SRNS. To evaluate the risk of NS relapse after u vaccination, we used the MPRM only among children who received u vaccination to calculate the risk ratio for NS relapse, adjusted for the post-vaccination period or pre-vaccination period, sex, and whether the patient was taking glucocorticoids or various immunosuppressants at u vaccination. A P-value < 0.05 was considered statistically signi cant. We used SAS software package for Windows, release 9.4 (SAS Institute Inc. Cary, NC, USA) to perform all statistical analyses.
Data are available upon reasonable request.
Declarations 25. Yoshikawa, N. et al. A multicenter randomized trial indicates initial prednisolone treatment for childhood nephrotic syndrome for two months is not inferior to six-month treatment. Kidney Int, 87, 225-232 https://doi.org/http:// (2015).        Figure 1 Flow chart of study design and populations Of 388 institutions, 105 responded to the third questionnaire and data were collected for 306 children.