| Lupus nephritis (LN) is one of the most common organ damages in systemic lupus erythematosus (SLE). The incidence of SLE in China is more than 70/10 million. According to biopsy, almost all of the SLE patients have different degrees of kidney damage, and 1/4 to 2/3 patients present the clinical features of kidney involvement. The severity of renal involvement directly influences the prognosis of SLE. The chronic renal failure caused by LN is still one of major causes of death in patients with SLE.Long-term proteinuria is one of the high risk factors on entering end-stage renal failure (ESRD). Control of proteinuria can reduce kidney damage and improve the prognosis of LN. Changes in urinary protein can be used to evaluate the therapeutic efficacy of LN. In the treatment of LN, we found that the reduction of urine protein in some patients is rapid, some is long-term persistence, and others gradually reduce over a long period. In order to explore the reasons for this phenomenon, we searched documents. We found that most literature is related to research on prognostic factors in patients with LN, and there is no answer on what causes patients with different changes in urinary protein and weather existence of factors lead to this change in urinary protein differences. Changes in urine protein after treatment may be owing to differences in application of the drug, drug dosage, cycles of treatment or owing to the severity of LN and different individual clinical characteristics. To some extent, history of the patients, laboratory parameters, renal pathological type can reflect the individual characteristics and severity of disease. Therefore, we designed a kind of multi-factor analysis by comparing relative factors on changes in urine protein. As the treatment can impact urinary protein changes, inclusion criteria is patients with the steroid and cyclophosphamide being first-line regimen for induced remission, precluding the simultaneous application of other immunosuppressants.Objective: To investigate relevant factors affecting changes in urinary protein after conventional treatment of patients with LN. To attempt to construct a mathematical model to forecast change tendency in urine protein after treatment in patients with LN and explore the significance of mathematical model.Methods: A retrospective analysis was done on 181 inpatients with LN from January 2007 to January 2010. All cases are hormones or hormone therapy combined with cyclophosphamide precluding the simultaneous application of leflunomide, mycophenolate mofetil (MMF) and other immunosuppressive agents. History of cases, biochemical and immunological indicators of laboratory examination, the maximum daily dose of hormones, duration of corticosteroids, whether concurrent infection, the cumulative amount of CTX were recorded. According to qualitative changes of urinary protein before and after treatment, all cases were divided into 4 groups: no less (group 1), reduction by 1 + (group 2), reduction by 2 + (group 3), reduction of 3 + (Group 4). Various factors between different groups were compared by analysis of variance and chi square test and P <0.05 is the standard to judge significant difference. Factors with significant difference are compared further by LSD method and split contingency tables chi-square test to analysis the differences between the various groups and filter out factors relating changes in urinary protein. At last, according to the probability distribution of factors, we construct a mathematical model by Bayes probability to predict tendency of urinary protein changes in patients with LN after conventional treatment.Results: No significant difference was found between groups in age, duration, URIC, C3, IGA, HDLC, LDLC, SLEDAI, the maximum dose and duration of corticosteroids, the average accumulation of CTX and gender in 181 cases of LN patients (P>0.05). Systolic blood pressure, diastolic blood pressure, 24h urinary protein, BUN, GFR, ALB, Hb, TG, TC, IGG, IGM, ESR, the number of positive ANAs, and recurrence and whether the hospital infection between the four groups were statistically significant difference (P<0.05).Further comparison shows: proportion of relapsed patients in group 1,2,3,4 reduced in order. No significant difference between group 1 and group 2, other groups showed statistically significance; In group 1 mean systolic and diastolic blood pressure are higher than other groups. Statistically significant differences were found on systolic blood pressure between Group 1 and the other groups. Diastolic blood pressure in group 1 and group 3 was significant differences. Average of 24h urinary protein in group 3 was lower than other groups, and statistically significant difference was found between group 1 and group 3, group 4 and group 3. The average of BUN in group 1 and group 2 was higher than those in Group 3 and Group 4, and statistically significant difference was found between group 1 and group 3, group 1 and group 4, group 2 and group 3, group 2 group 4. In group 1 average of GFR was lower than other groups, and statistically significant difference was found between group 1 and group 3, group 1and group 4. Group 1 has the lowest average of ALB, and the group 1 and group 3, group 1 and group 4 were significantly different. In group 3 mean Hb was higher than other groups and statistically significant differences was found between group 3 and the other groups. In group 1, TG, TC were higher than other groups, and the two indicators between group 1 and group 2, group 1 and Group 3, group 1 and Group 4 were statistically significantly different. The value of IGG and IGM increase in Group 1,2,3,4, in order and the average of IGG in group 1 and group 3, group 1 and group 4, group 1 and Group 2, group 1 and Group 4 were statistically different, respectively. Average of IGM in group 1 and group 3, group 1 and group 4 was statistically different, respectively. The highest average ESR was found in group 2, and in group 1 and group 2, group 1 and group 3, group 2 and group 3, group 2 and group 4 was statistically significant different, respectively. In group 4 there is the highest average of the number of ANAs, and between group 2 and group 3, group 3 and group 4 were significant differences, respectively. Sex, infection during hospitalization is needed to further investigation on more sample cases to make a more accurate comparison analysis.We construct a mathematical model to forecast change tendency in urine protein after treatment in patients with LN. The self-validation accuracy of the model was 71.3%.Conclusion: Changes in urinary protein in patients with LN between the individuals and the groups of patients are related with clinical features and severity of the disease: Systolic blood pressure, diastolic blood pressure, 24h urinary protein, BUN, ALB, IGG, IGM, Hb , TG, TC, GFR, and recurrence is the relevant factors on changes in urinary protein after conventional treatment in patients with LN. Higher SBP, DBP, BUN, TG, TC values, lower GFR, ALB values, and relapse in patients with conventional treatment is the risk factors affecting the reduction of urine protein. Patients with higher IGG, IGM values were advantageous on protein reduction. In patients with slow effect after conventional therapy, the 24h urinary protein and (or) Hb may predict less changes in urinary protein, while in patients with better drug reaction, this kind of relation was not found. Mathematical model based on Bayes probability can be used to forecast change tendency in urine protein after treatment in patients with LN. By puting the Mathematical model into clinical practice we can more accurately determine the patient's condition on each admission and develop individualized treatment plan and evaluate prognosis. |
PDF Full Text Request