Who benefits from Colorectal Cancer Screening? A subgroup Analysis of the Minnes

描述（由申请人提供）： 1.在明尼苏达州结肠癌控制研究（MCCCS）中，确定粪便潜血试验（FOBT）筛查对结肠直肠癌死亡率、非结肠直肠癌死亡率和全因死亡率的影响，更新至2006年2。确定2006年更新的MCCCS中FOBT筛查对不同年龄（5岁和10岁年龄组）和性别的结直肠癌死亡率和非结直肠癌死亡率的影响。确定基线特征对MCCCS中结直肠癌死亡率和非结直肠癌死亡率的独立影响。意义：大多数指南建议在50岁时开始筛查结直肠癌的平均风险个体。其中最早和测试最多的策略是FOBT，由四项大型随机临床试验支持，这些试验确定了结直肠癌筛查的有效性。明尼苏达结肠癌控制研究在美国明尼苏达州进行。该试验招募了50至80岁的健康志愿者。这些试验没有报告按年龄和性别分层的结果。此外，尚未报告其他基线变量（如合并症和阿司匹林摄入量）对结局的影响。最近对这些试验的荟萃分析证实了结直肠癌死亡率的降低（RR 0.87; 95% CI 0.8-0.95），但也报告了非结直肠癌死亡率的绝对幅度大致相等的增加（RR 1.02; 95% CI 1.00-1.04），因此总体死亡率几乎相同（RR 1.002，95% CI 0.989-1.015）。为什么非CRC死亡率会增加尚不清楚，但一种可能性是FOBT试验包括一些从CRC筛查中受益的个体，以及其他经历净伤害的个体。因此，对FOBT试验数据进行详细的亚组分析，以区分从筛选中获益的亚组和可能受到伤害的亚组，这一点很重要。例如，70岁或75岁以上的患者可能从筛查中获得较少的益处和更多的伤害，在这种情况下，可能需要重新考虑对这一人群进行筛查。本研究的目的是评估在明尼苏达州结肠癌控制研究中，随机化后18年（1999年报告）筛查对结肠直肠癌死亡率的影响是否持续存在，或在随机化后28-30年（更新至2006年）不同，以及结肠直肠癌筛查的影响是否因年龄和性别亚组而异。我们还希望评估基线特征对这些结果的影响，如饮食，阿司匹林的使用和体重指数。项目设计和方法：该项目包括更新明尼苏达州结肠癌控制研究中截至2006年的所有参与者的生命状态和死亡原因，并计算CRC，非CRC和全因死亡率估计。我们计划在随机化5年（50-54、55-59、60-64、65-69、70-74、75-80）和10年（50-59、60-69、70-80）时按年龄类别比较筛选组和未筛选组之间的结直肠癌（CRC）死亡率和非结直肠癌死亡率。我们还将按性别亚组对上述分析进行分层。明尼苏达州结肠癌对照研究未分析或报告除年龄和性别外的患者基线数据。参与者完成了一份4页，56项的基线问卷。我们将对现有数据文件进行质量检查，以确定是否需要重新输入基线调查问卷上的所有问题或部分问题。我们有兴趣探索非甾体抗炎药的消费，高纤维饮食和BMI之间的关系，作为筛选有效性的影响因素。 公共卫生相关性： VA在研究和实施结直肠癌筛查方面投入了大量精力。结直肠癌的发病率随年龄增长而增加。它是退伍军人中最常见的癌症之一。Lieberman等人57报告称，在接受结肠镜筛查的无症状退伍军人中，结肠病变的发生率为5.9%。据估计，75%的VA患者年龄超过50岁，40%年龄超过70岁。此外，妇女占老年患者的5%，预计这一数字还会增加。最近的一项退伍军人管理局的研究呼吁注意CRC筛查的老年退伍军人与严重的共病谁是不太可能受益1。因此，重要的是要澄清筛查是否对所有亚组都有益。更好地了解要筛查的理想目标人群，将有助于集中努力，以资源有效的方式实施。

DESCRIPTION (provided by applicant): 1. To determine the effect of fecal occult blood testing (FOBT) screening on colorectal cancer mortality, non- colorectal cancer mortality and all-cause mortality in the Minnesota Colon Cancer Control Study, (MCCCS) updated to 2006 2. To determine the effect of FOBT screening on colorectal cancer mortality and non-colorectal cancer mortality by age (by 5 and 10 year age groups) and gender in the MCCCS updated to 2006 3. To determine the independent influence of baseline characteristics on colorectal cancer mortality and non-colorectal cancer mortality in the MCCCS. Significance: Most guidelines recommend starting screening for colorectal cancer at age 50 years for average risk individuals. The earliest and most tested strategy among these is FOBT, supported by four large randomized clinical trials that established the effectiveness of colorectal cancer screening. The Minnesota Colon Cancer Control Study was conducted in Minnesota, US. The trial recruited healthy volunteers between the ages of 50 to 80. The trials have not reported results stratified by age and gender. Additionally, the effects of other baseline variables, such as co-morbidities and intake of aspirin on outcomes have not been reported. A recent meta-analysis of these trials confirmed a reduction in colorectal cancer mortality (RR 0.87; 95% CI 0.8-0.95), but also reported an increase in non-colorectal cancer mortality of approximately equal absolute magnitude (RR 1.02; 95% CI 1.00-1.04) so that overall mortality was nearly identical (RR 1.002, 95% CI 0.989-1.015). Why non-CRC mortality would be increased is unclear, but one possibility is that the FOBT trials included some individuals that benefited from CRC screening, and other individuals who experienced net harm. Therefore, detailed sub-group analyses of the FOBT trial data, to separate out sub- groups that benefit from screening from those who may potentially be harmed would be important. For example, patients over age 70 or 75 might derive less benefit and more harm from screening, in which case screening this population may need to be re-considered. The goals of this research are to evaluate if effects of screening on colorectal cancer mortality seen 18 years after randomization (reported in 1999) in the Minnesota Colon Cancer Control Study persist, or are different 28-30 years after randomization (when updated to year 2006), and whether effect of screening for colorectal cancer varies in age and gender sub- groups. We also wish to evaluate the influence of baseline characteristics, such as diet, aspirin use and body mass index on these outcomes. Project Design and methods: This project involves updating the vital status and cause of death of all participants from the Minnesota Colon Cancer Control Study up to 2006 and calculating CRC, non-CRC and all-cause mortality estimates. We plan to compare colorectal cancer (CRC) mortality and non-colorectal cancer mortality rates between the screened and unscreened groups, by age categories at time of randomization of 5 year (50-54, 55-59, 60-64, 65-69, 70-74, 75-80) and 10 year (50-59, 60-69, 70-80) intervals. We will also stratify the above analyses by gender subgroups. No patient baseline data other than age and gender has been analyzed or reported from the Minnesota Colon Cancer Control Study. A 4-page, 56-item baseline questionnaire was completed by participants. We will perform a quality check on the existing data files to determine if re-entry is required for all or a subset of the questions on the baseline questionnaire. We are interested in exploring the relationship between consumption of NSAIDs, high fiber diet and BMI as effect modifiers of effectiveness of screening. PUBLIC HEALTH RELEVANCE: VA has put much effort into studying and implementing colorectal cancer screening. Colorectal cancer incidence increases with age. It is one of the most common cancers among the veteran population. Lieberman et al57 reported an incidence rate of 5.9% for colonic lesions among asymptomatic veterans undergoing screening colonoscopy. An estimated 75% of VA patients are over age 50, and 40% are over age 70. Also, women constitute 5% of veteran patients, and the number is expected is increase. A recent VA study has called attention to CRC screening in elderly veterans with serious co- morbidities who are unlikely to benefit 1. It is therefore important to clarify whether screening is beneficial in all sub-groups. A better understanding of the ideal target population to be screened would help focus efforts at resource-effective implementation.