糖尿病併慢性腎病變病人使用HMG-CoA 還原酶抑制劑與末期腎衰竭的風險

Abstract

摘要 背景： 糖尿病慢性腎病變為末期腎衰竭的高危險群，過去的研究顯示HMG-CoA還原酶抑制劑(Statins)可以減少慢性腎病變患者的蛋白尿並且可以減緩腎功能的惡化。2010年，英國一個以大規模族群為對象的世代研究顯示出statin類降膽固醇藥物會增加急性腎衰竭的風險。直到目前為止，探討statin類降膽固醇藥物使用在慢性腎病變病患與末期腎衰竭關係的研究仍然相當有限。因此，本研究的目的為使用大型全民健康保險資料庫來探討statin類降膽固醇藥物使用在糖尿病併慢性腎病變患者身上是否有延緩末期腎衰竭的效果。 材料與方法： 本研究為一回溯性世代研究，採用的是2000年健保資料庫糖尿病世代之抽樣檔。在2002年154347糖尿病患者中，總共有21719 位糖尿病併慢性腎病變患者。2002年有使用statin藥物的病患歸在statin使用組，2002年沒有使用statin藥物的病患歸在非使用組。Statin 使用組的index date為2002年第一次開立statin處方的第一天，非使用組的index date為依據Statin組在慢性腎病變診斷後至第一次開立statin的時間間距分布，在慢性腎病變診斷後，選取一相對應的index date。排除2002年患有癌症及愛滋病患者。此外，兩組研究對象在index date之前有使用過statin或已經為末期腎衰竭患者也皆被排除在外。結果有13272糖尿病併慢性腎病變患者，其中statin使用者有1637位，非使用者有11635位。 使用病人在index date前一年的共病症以及醫療資源的利用情形來建立propensity score，經propensity score配對後，statin使用者與非使用者各1637位。以Cox比例風險回歸模型(Cox proportional hazards regression model)來計算使用組與非使用組之間發生末期腎衰竭風險比與死亡的風險比(hazard ratio)。排除追蹤不滿一年發生末期腎衰竭與死亡的病患，看分析結果是否與原先一致作為敏感性試驗。

結果： 以末期腎衰竭為結果之分析：在statin使用組追蹤3303.4人年，年發生率為0.058；非使用組追蹤3749.9人年，年發生率為0.041。statin使用組與非使用組發生末期腎衰竭的風險比為1.36(1.10-1.68)，敏感性試驗的風險比為1.34 (1.07-1.68)。 以死亡率為結果之分析：在statin使用組追蹤3447.9人年，年發生率為0.037；非使用組追蹤3926.6人年，年發生率為0.071。statin使用組與非使用組發生死亡的風險比為0.51 (0.41-0.63)，敏感性試驗的風險比為0.45 (0.33-0.61)。 結論： 本研究的結果顯示statin類降膽固醇藥物使用在糖尿病併慢性腎病變患者會降低死亡的風險，但另一結果在末期腎衰竭之發生呈現風險上升部分，仍有一些immortal time等之偏差，需要進一步探討。

Abstract Background: Diabetic patients with chronic kidney disease were at high risk of end stage renal disease. Previous studies showed that HMG-CoA reductase inhibitors (statins) could reduce proteinuria in patients with chronic kidney disease and slow the progression of renal function deterioration. In 2010, one population based cohort study in England revealed that statin increased the risk of acute renal failure. Till now, few studies investigate the effect of statin on the progression of chronic kidney disease to end stage renal disease。Therefore, our aim is try to identify whether statins could slow the progression of chronic kidney disease to end stage renal disease in diabetic patients by using big BNHI (Bureau of National Health Insurance) claim database. Materials and Methods: This is a retrospective cohort study using 2000 National Health Insurance diabetes cohort database. We identified 21719 diabetic patients with chronic kidney disease in 2002. Patients who used statin in 2002 were classified as statin user group and those who did not use statin in 2002 were classified as non-user group. The first statin prescription date in 2002 was defined as index date in statin users. We assigned a day which was chosen by the same distribution in statin users group as index date in non-users. Patients who had cancer or AIDS in 2002 were excluded. Patients who used statins or had end stage renal disease before index date were excluded, too. Total 13272 diabetic patients with chronic kidney disease were analyzed in our study. There were 1637 statin users and 11635 patients were non-users. Patients’ comorbidities and medical resource utilization were used to construct propensity score. Statin user group and non-user group are matched by propensity score. After matching for propensity score, there were 1637 statin users and non-users. Cox proportional hazard model was used to discover the hazard ratio of end stage renal disease and all cause mortality between these 2 groups. Sensitivity analysis was done by excluding patients with follow up time less than one year to test the robustness of study result. Result: End stage renal disease outcome analysis: The follow-up person-years in statin users were 3303.4 and the annual incidence rate was 0.058. The follow-up person-years in non-users were 3749.9 and the annual incidence rate was 0.041. The end stage renal disease hazard ratio between these 2 group was 1.36(1.10-1.68). Sensitivity analysis showed that the hazard ratio is 1.34 (1.07-1.68). Mortality outcome analysis: The follow-up person-years in statin users were 3447.9 and the annual incidence rate was 0.037. The follow-up person-years in non-users were 3926.6 and the annual incidence rate was 0.071. The mortality hazard ratio between these 2 group was 0.51(0.41-0.63). Sensitivity analysis showed that the hazard ratio is 0.45 (0.33-0.61). Conclusion: Our study showed that statins decrease the risk of mortality in diabetic patients with chronic kidney disease. On the other hand, the result of increasing risk of end stage renal disease in statin users needed further investigation because of residual immortal time bias.