Background <p>Genetic analysis, crucial in determining treatment strategies for steroid-resistant nephrotic syndrome (SRNS), can be performed in limited facilities and requires a long time. Predicting the presence or absence of genetic variants by clinical and pathologic features is preferable.</p> Methods <p>In this multicenter, retrospective study, we compared the clinical or pathologic features between the patients with and without genetic variants in children with SRNS and evaluated the efficacy of immunosuppressive treatment and long-term kidney outcomes.</p> Results <p>Fifty-three patients in 17 institutes were included, and 11 patients (21%) showed genetic variants. Two patients with a family history of nephrotic syndrome harbored genetic variants. Serum albumin level at onset was significantly lower in patients without genetic variants (<i>p</i> = 0.001). The receiver operating characteristic curve analysis showed that a cutoff value of serum albumin level of 2.3 g/dL at onset had a sensitivity and specificity of 82% and 90%, respectively, in predicting genetic variants. Patients with asymptomatic proteinuria at onset were more likely to harbor genetic variants (<i>p</i> = 0.05). None of the pathologic features was significantly different between the two groups. Mesangial proliferation and diffuse foot process effacement were observed more in patients without genetic variants, although statistically insignificant. Immunosuppressive treatment was less effective, and the 5-year kidney survival was poorer (31% and 78%, <i>p</i> = 0.03) in patients with genetic variants than in those without genetic variants.</p> Conclusions <p>Higher serum albumin levels at onset can predict the presence of genetic variants. Pathologic features might have limited utility in predicting them.</p> Graphical Abstract <p></p>