- 學位論文
慢性腎臟病進展的定量化流行病學模型
Quantitative Epidemiological Models for Chronic Kidney Disease Progression
若您是本文的作者，可授權文章由華藝線上圖書館中協助推廣。
摘要
研究背景和目的 慢性腎臟病是一種具有世界性，和非傳染性特色的疾病，它的盛行率有越來越高的趨勢。而且，慢性腎臟病會導致較高的死亡率和失能率，所以照顧慢性腎臟病變成是世界政府的財政負擔和嚴重的健康照護問題。台灣在2014年時，有著有著全世界最高的透析盛行率和發生率。慢性腎臟病的高盛行率可能是主要的原因。然而，關於台灣慢性腎臟病相關的流行病學統計資料是有所不足的。因此，我門設計此研究來探討台灣慢性腎臟病的樣貌，慢性腎臟病進展的危險因子和轉移機率預測方程式。最後，我們還試著去估算慢性腎臟病在不同期別之間的年度轉移機率，也就是慢性腎臟病的自然病史。 材料與方法 本論文所研究的資料來自參與基隆闔家歡篩檢計劃的民眾。首先，描述台灣慢性腎臟病的盛行率與發生率。之後，利用加速失效模式(Accelerated failure time model)來評估影響慢性腎臟病狀況轉換的危險因子。影響腎絲球過濾率從≥60到59-30 mL/min/1.73m2 為起動因子，而影響eGFR 59-30到<30 mL/min/1.73m2是為促進因子。同時建立慢性腎臟病進展的預測方程式。最後，還利用五階段馬爾可夫模式來描述慢性腎臟病的自然病史（第一階段：腎絲球過濾率從≥60 mL/min/1.73m2，第二階段：腎絲球過濾率從59-30 mL/min/1.73m2，第三階段：腎絲球過濾率從<30 mL/min/1.73m2，第四階段：接受透析，第五階段：全因死亡）。 結果 第一部分：慢性腎臟病的盛行率與發生率 預估所有慢性腎臟病(1-5期)的盛行率是15.46%，而慢性腎臟病3-5期盛行率是9.06%。預估所有慢性腎臟病的發生率是27.21/每1000人年，而慢性腎臟病3-5期的發生率是16.89/每1000人年。在次族群分析中，老年人，男生，高血壓，糖尿病，代謝代謝症候群，和蛋白尿者，有較高的慢性腎臟病盛行率和發生率。此外，慢性腎臟病3-5期盛行率與發生率的比值顯示為5.37 (代表平均處於這個時期的時間為5.37年)。男生，年紀較輕，和代謝代謝症候群者，有較低的盛行率與發生率比值（與他們相對的族群比較）。 第二部分：慢性腎臟病的啟動因子和促進因子。 慢性腎臟病的獨立啟動因子為年紀大(風險比率: 1.08；95％信賴區間: 1.07-1.09)，糖尿病(風險比率: 1.49；95％信賴區間: 1.23-1.81)，代謝症候群分數 (風險比率: 1.13；95％信賴區間: 1.0-1.19),，蛋白尿 (風險比率: 1.16；95％信賴區間: 1.10-1.22),，高尿酸血症 (HR: 1.12；95%CI: 1.08-1.17)，與較高的低密度膽固醇(風險比率: 1.15；95%CI: 1.02-1.30) ，和腎絲球過濾率 (風險比率: 0.94；95%CI: 0.93-0.95)。慢性腎臟病的獨促進因子為冠狀動脈疾病(風險比率: 1.57；95%CI: 1.04-2.36)，代謝症候群分數(風險比率: 1.31；95%CI: 1.12-1.53),，蛋白尿(風險比率: 1.48；95%CI: 1.31-1.67，高尿酸血症(風險比率: 1.11 ; 95%CI: 1.02-1.22),，與貧血(風險比率: 0.84；95%CI: 0.75-0.95), 和腎絲球過濾率(風險比率: 0.89；95%CI: 0.87-0.91)。此外，我們針對慢性腎臟病的發生與慢性腎臟病的進展，發展出兩套預測方程式，可以估算出慢性腎臟病狀態轉移的絕對發生機率。 第三部分：慢性腎臟病的隨機過程馬可夫鍊模式 我們成功地估計出慢性腎臟病的自然病史。每年腎絲球過濾率從≥60 到 59–30 mL/min/1.73m2的進展機率是0.0169。 從eGFR 59–30到<30 mL/min/1.73m2的進展機率是0.0259。從eGFR <30 mL/min/1.73m2 到需要接受透析的進展機率是0.0988。男性相對約女性，有較有較高的慢性腎臟病進展機率和死亡率。整體而言，平均停留在腎絲球過濾率從59–30 mL/min/1.73m2（慢性腎臟病第三期）的時間是5.84年，平均停留在腎絲球過濾率從<30 mL/min/1.73m2（慢性腎臟病第4–5期）的時間是2.99年。不同性別，糖尿病，和蛋白尿會表現出不同的平均停留時間。 結論 台灣有高的慢性腎臟病盛行率與發生率。我們的研究找出了慢性腎臟病進展的相關獨立因子，同時並建立慢性腎臟病進展的預測方程式。這些結果可以讓我們對慢性腎臟病的進展能有更加的瞭解，有助於在慢性腎臟病的照護中，針對危險族群發展出特異化醫療。而發展出的危險預估方程式也可以輕易地被整合入報告系統，可以早期警示醫師，將高危險群病患轉介至腎臟科，接受多學科的全人照護。此外，慢性腎臟病的平均停留時間可以被當成慢性腎臟病介入政策的評量指標。最後，我們還利用隨機馬可夫模式，發展出慢性腎臟病的進展病史（近似自然病史），這讓我們對慢性腎臟病的進展有更近一步的了解。也有利於未來慢性腎臟病介入計畫的成本效益的分析。
並列摘要
Background and objectives Chronic kidney disease (CKD) is a major noncommunicable disease and has become a global public health problem with an increasing prevalence. Caring for patients with CKD has been shown to present financial and medical burdens owing to high mortality and morbidity. In 2014, Taiwan has the highest incidence and prevalence of end-stage renal disease, requiring renal replacement therapy. CKD may contribute to this burden. However, the current data on the epidemiologic features of CKD in Taiwan are incomplete. Therefore, we designed this study to elucidate the epidemiologic pictures of CKD, the risk factors for CKD progression and the annual transition rate between CKD stages. Materials and Methods Subjects from Keelung Community-based Integrated Screening (KCIS) Program were enrolled since 1999 to 2009. We reported prevalence and incidence rate of CKD stages and tried to estimate the risk factors for CKD state transition using accelerated failure time model. The initiator and progressor were defined as the factor affecting the eGFR from eGFR ≥60 to 59–30 and from eGFR 59–30 to <30 mL/min/1 respectively. Moreover, a five-state Markov process was used to describe the Clinical history of CKD stages (state 1: eGFR ≥60 mL/min/1.73 m2, state 2: eGFR 59–30 mL/min/1.73 m2, state 3: eGFR <30 mL/min/1.73 m2, state 4; Receiving dialysis, and state 5: all-cause death). Results Part I: The prevalence and incidence of CKD stages The participants’ mean age was 47.7 ± 15.4 years. The estimated prevalence was 15.46% for total CKD and 9.06% for CKD stages 3–5. The incidence was 27.21/1000 person-years (PY) for total CKD and 16.89/1000-PY for CKD stages 3–5. Older patients, males, and those patients with comorbidities of diabetes mellitus (DM), hypertension, and metabolic syndrome (MetS) exhibited higher prevalence and incidence rates than their opposing counterparts. Moreover, the average dwelling time (ADT) of CKD stages 3–5 was 5.37 years (95% confidence interval (CI): 5.17–5.57). Males and those with comorbidities of DM or MetS had shorter ADTs in CKD stages 3–5 than their opposing counterparts. Part II: The independent initiators and progressors of CKD The independent initiators of CKD were old age (HR, 1.08; 95% CI, 1.07–1.09), diabetes (HR, 1.49; 95%CI, 1.23–1.81), metabolic syndrome scores (HR, 1.13; 95%CI, 1.08–1.19), proteinuria (HR, 1.16; 95%CI, 1.10–1.22), hyperuricemia (HR, 1.12; 95%CI, 1.08–1.17), higher low-density lipoprotein level (HR, 1.15; 95%CI, 1.02–1.30), and low eGFR level (HR, 0.94; 95%CI, 0.93–0.95). The independent progressors of CKD were coronary artery disease (HR, 1.57; 95%CI, 1.04–2.36), metabolic syndrome scores (HR, 1.31; 95%CI, 1.12–1.53), proteinuria (HR, 1.48; 95%CI, 1.31–1.67), hyperuricemia (HR, 1.11; 95%CI, 1.02–1.22), low hemoglobin level (HR, 0.84; 95%CI, 0.75–0.95), and low eGFR level (HR, 0.89; 95%CI, 0.87–0.91). Furthermore, two risk prediction functions were also built for the absolute risk prediction of CKD state transition. Part III: The stochastic Markov model of CKD The annual progression rate was 0.0169 (95% CI, 0.0164–0.0173) from eGFR ≥60 to 59-30 mL/min/1.73m2, was 0.0259 (95%CI, 0.0240–0.0278) form eGFR 59-30 to <30 mL/min/1.73m2, was 0.0988 (95% CI, 0.0902–0.1075) from eGFR <30 mL/min/1.73m2 to dialysis. The man had higher progression rate for the movement from eGFR ≥60 to eGFR 59–30 mL/min/1.73m2 than the woman. The ADT of eGFR 59–30 mL/min/1.73m2 (CKD stage 3)was 5.48 years (95%CI, 5.62–6.07) and was 2.99 years (95%CI, 2.77–3.25) for eGFR <30 mL/min/1.73m2 (CKD stages 4–5). The ADT varied by age, gender and comorbidities. Conclusion The prevalence and incidence of CKD in Taiwan are high. We ascertained the independent initiators and progressors of CKD in our study. The results are useful to understand the association between factors and the state transition of CKD, which can be beneficial for developing specialized CKD care programs for the risky population. Also, these two prediction functions can be easily integrated into the reporting system to early alert the physicians to transfer the risky patients to receive a nephrologist-based multidisciplinary care. Moreover, the ADT in CKD can be used as an indicator for evaluating the CKD policy. Finally, a stochastic mode of CKD was successfully established to elucidate the approximated natural history of CKD in Taiwan, which this can offer an updated understanding of the CKD progression and also can be applied to the cost-effectiveness analysis of CKD intervention.
參考文獻
1. Hsu, C.C., et al., High prevalence and low awareness of CKD in Taiwan: a study on the relationship between serum creatinine and awareness from a nationally representative survey. Am J Kidney Dis, 2006. 48(5): p. 727-38.
2. Centers for Disease, C. and Prevention, Prevalence of chronic kidney disease and associated risk factors--United States, 1999-2004. MMWR Morb Mortal Wkly Rep, 2007. 56(8): p. 161-5.
3. Hwang, S.J., J.C. Tsai, and H.C. Chen, Epidemiology, impact and preventive care of chronic kidney disease in Taiwan. Nephrology (Carlton), 2010. 15 Suppl 2: p. 3-9.
4. Tin, A., et al., Patterns of Kidney Function Decline Associated with APOL1 Genotypes: Results from AASK. Clin J Am Soc Nephrol, 2016. 11(8): p. 1353-9.
5. Saran, R., et al., US Renal Data System 2016 Annual Data Report: Epidemiology of Kidney Disease in the United States. Am J Kidney Dis, 2017. 69(3 Suppl 1): p. A7-A8.