Identification of new urinary risk markers for urinary stone patients using logistic model and multinomial-logit model
Sources of Funding: None
Introduction
The risk assessment of urinary stone disease so far has been conducted mainly on urinary biochemistry, but it has not been fully utilized for prevention of recurrence. We have clarified that inflammation is involved in the formation process of calcium oxalate (CaOx) stones. In this study, we tried to extract new stone risk factors by statistically analyzing urinary biochemistry and urine inflammation related factors._x000D_ _x000D_
Methods
The subjects were male (20-79 years old) who visited Nagoya City University Hospital, excluding patients with history of abnormal urinalysis, history of tumor and collagen disease, patients taking immunosuppressants and steroids, and divided into two groups, normal group (48 cases) without history of stone and stone group who experienced CaOx stone (22 cases of the first time group, 40 cases of recurrence).
Results
Comparison between the normal group and the first time group revealed that the areas under the curve (AUC) of ROC of IL-1a and IL-4 as independent factors were significantly high (1.00 and 0.87 (P <0.01 in each case), respectively), which suggested that the two factors were specific to first time patients. In the comparison between the normal group and the stone group, the AUC value increased to 0.87, 0.86 by combining IL-1a or IL-4 with GM-CSF and IL-1b (both P <0.01) respectively, and the values were not as high as the discrimination between the normal and the first time groups. In the comparison of the three groups (normal, first time and recurrence group), discrimination ability by multinomial logit model using IL-4, GM-CSF, IL-1b and IL-10 including urinary Mg was the highest (prediction accuracy: 82.6%).
Conclusions
IL-4, IL-1a, GM-CSF, IL-1b and IL-10 were identified as urinary inflammation related factors that can accurately distinguish normal subjects and urinary stone patients. These factors are related to the activity of macrophages and neutrophils and it was suggested that combining with urinary biochemistry data could be an index to more clearly evaluate the risk of urinary stone formation.
Funding
None
Teruaki Sugino
Rei Unno
Kazumi Taguchi
Shuzo Hamamoto
Ryosuke Ando
Keiichi Tozawa
Takahiro Yasui