降血糖藥品與結腸癌復發之相關性：巢式病例對照研究

Abstract

背景 結腸癌在全球為發生率第三名和死亡率第二名的癌症，糖尿病為盛行率高的疾病，被認為是結腸癌可能的危險因子之ㄧ，也會對結腸癌的預後有不良影響，而治療糖尿病使用的降血糖藥品就其藥理機轉上也被認為可能是影響結腸癌預後的因子之一。早期（第一至三期）結腸癌雖可以予以手術切除及輔助性治療，但五年復發率仍高，因此降血糖藥品對於癌症復發的潛在影響仍是待釐清的問題。過去觀察性文獻大多以死亡作為此主題的研究終點，目前仍缺少分析不同種類降血糖藥品的使用與結腸癌復發相關性之研究。 目的 本研究欲分析結腸癌術後使用不同種類的降血糖藥品與結腸癌復發事件之相關性，並針對具顯著相關性之藥品進一步探討是否存在劑量反應關係（dose-response relationship）。 方法 本研究為巢式病例對照研究，使用台灣癌症登記資料庫2007-2015年的資料， 與全民健保資料庫全人口檔的2004-2017年度資料進行分析，納入於2007/01/01-2015/12/31期間內同時有第二型糖尿病、新診斷罹患早期結腸癌（第一至三期）且接受根除性手術的病人，於研究世代中篩選出癌症復發病人為病例組（結腸癌復發日期為index date），再使用risk-set sampling的方式以隨機方式選取出在index date尚未復發的對照組病人。病例組與對照組以1:4比例，依年齡（± 2歲）、性別、接受結腸癌手術之年份、結腸癌診斷時期別以及有無接受術後輔助性治療這五項因子進行配對後，分析兩組在從結腸癌術後至index date六個月前的期間內，降血糖藥品的暴露狀況上有無差異。探討的降血糖藥品依照藥理機轉分成八大類，以條件式羅吉斯迴歸模型計算勝算比，以評估各類降血糖藥品與結腸癌復發之相關性。此外，再進一步計算降血糖藥品之累積暴露劑量或累積暴露時間，分析是否存在劑量反應關係（dose-response relationship）。在敏感度分析中，使用不同的癌症復發定義以及以假設陣列方式（array method）評估無法測量之干擾因子（糖化血色素〔HbA1c］、肥胖〔以身體質量指數BMI表示］）對本研究的影響，檢視主要分析結果之穩健性，最後在次群組分析中探討不同族群在此相關性上是否有差異。 結果 本研究納入5,288位病人為研究世代，共找出418位結腸癌復發病人，經配對後共得到386位病例組與1,385位對照組病人。在校正了相關共變項後，顯示術後biguanides類藥品使用者相較於未使用者有顯著降低的復發危險性（adjusted OR 0.74, 95% CI 0.57-0.97）。進一步分析metformin之累積劑量時顯示，隨著累積劑量愈高，復發危險性愈低，並在累積劑量834,500 mg以上或累積期間超過一年達顯著差異。其他降血糖藥品的使用則因樣本數過少或暴露比例低，未發現與結腸癌復發顯著相關。敏感度分析結果皆與主要分析一致，而未校正干擾因子HbA1c和BMI對本研究結果影響並不大。次群組分析由於人數過少檢定力不足，無法進行結果推論。 結論 本研究發現biguanides類藥品的使用與降低結腸癌復發危險性有關，且存在劑量反應相關性，研究結果可作為未來臨床上藥品選擇的參考依據。其他降血糖藥品則受限於檢定力不足而未看出顯著相關性，建議未來研究可透過延長資料期間增加樣本數，來針對其他種類的降血糖藥品進行探討，以提供更完整的臨床參考資訊。

Background Colon cancer is the third most common cancer and the second leading cause of cancer death globally. Type 2 diabetes (T2DM), as a prevalent disease worldwide, is known as a risk factor which may lead to worse prognosis of colon cancer. In addition, the use of antihyperglycemic drugs might affect the prognosis of colon cancer. Although patients with early-stage (stage I-III) colon cancer can be treated by surgery and adjuvant therapy, there remains a high five-year recurrence rate after surgical removal. There is a need to determine the impact of antihyperglycemic drug use on the risk of colon cancer recurrence. Previous observational studies mainly focused on the survival outcome of metformin in early-stage colon cancer, and there is only limited evidence on the association between different classes of antihyperglycemic drugs and colon cancer recurrence. Objectives Our study aimed to assess the association between antihyperglycemic drug use and the risk of colon cancer recurrence after curative resection. We further examined if there is a dose-response relationship between use of antihyperglycemic drugs and colon cancer recurrence by evaluating the effect of cumulative dose and duration of use. Methods We conducted a nested case-control study using data linkage of the Taiwan Cancer Registry (2007-2015) and the National Health Insurance Research Database (2004-2017). Adult patients with T2DM who were newly diagnosed stage I-III colon cancer and underwent curative surgery between 2007/01/01 and 2015/12/31 were included. Cases with colon cancer recurrence were identified, and controls were selected by risk-set sampling. Each case was randomly matched to 4 controls on age (±2 years), gender, year of cohort entry, colon cancer stage at diagnosis, and adjuvant therapy (yes or no). The index date was assigned as the date of colon cancer recurrence in matched cases. Exposure of antihyperglycemic drugs (in eight classes) was measured between surgery and six months prior to the index date. Conditional logistic regression was used to examine the risk of colon cancer recurrence associated with the use of antihyperglycemic drugs. We further calculated the cumulative dose and cumulative duration of the exposures to determine whether a dose-response relationship exists. Sensitivity analyses with different case definitions and assessing the impact of unmeasured confounder (i.e., HbA1c, BMI) using the array method were performed to test the robustness of study results. Subgroup analyses by diabetes duration, and colon cancer stage at diagnosis were conducted to investigate differential risks in different subgroups. Results The study cohort consists of 5,288 patients, and 418 patients with colon cancer recurrence were identified. After matching, 386 cases and 1,385 controls were included. Post-surgical use of biguanides was associated with a significantly lower risk of colon cancer recurrence compared to non-users (adjusted OR 0.74, 95% CI 0.57-0.97) after adjusting for all other covariates. Further analysis of metformin revealed a dose-response relationship－the risk of colon cancer recurrence decreased as the cumulative dose increased. Patients with a cumulative duration of one year or cumulative dose of 834,500 mg had a significantly decreased risk of colon cancer recurrence. Due to a lack of statistical power or little exposure, the other classes of antihyperglycemic drugs failed to show an effect on the risk of cancer recurrence. All the sensitivity analyses yielded consistent results with the main analysis, and HbA1c and BMI values were not likely to bias our findings. Differential risks in different subgroups could not be assessed due to the limited sample size. Conclusions Our study found that the use of biguanides after surgical resection of colon cancer was associated with a decreased risk of colon cancer recurrence in a dose-response manner. We expect that the findings of this study would help to guide the choice of antihyperglycemic treatment in patients with stage I-III colon cancer and improve their clinical outcomes. However, the risks related to other classes of antihyperglycemic drugs remain undetermined due to lack of statistical power. Further studies with an extended study period and a larger sample size are needed to evaluate the effect of other antihyperglycemic drugs.