Benefit and harm of intensive blood pressure treatment: Derivation and validation of risk models using data from the SPRINT and ACCORD trials.
Benefit and harm of intensive blood pressure treatment: Derivation and validation of risk models using data from the SPRINT and ACCORD trials.
复制标题
DOI：
10.1371/journal.pmed.1002410
复制
发表时间：
2017-10
期刊：
影响因子：
15.8
通讯作者：
Hayward RA
中科院分区：
文献类型：
--
作者：
Basu S;Sussman JB;Rigdon J;Steimle L;Denton BT;Hayward RA
强化血压（BP）治疗可以避免心血管疾病（CVD）事件，但也可能导致一些严重的不良事件。我们试图开发和验证风险模型，用于预测 CVD 事件和强化血压治疗的严重不良事件的绝对风险差异（风险增加或风险降低）。第二个目的是测试与传统的向后变量选择方法相比，弹性网络正则化的统计方法是否会改进用于预测绝对风险差异的风险模型的估计。 Cox模型源自SPRINT试验数据，并在ACCORD-BP试验数据上进行验证，以估计CVD事件和严重不良事件的风险；该模型包括强化血压治疗和对强化治疗的异质反应的术语。然后使用 Cox 模型来估计强化治疗中每个人 CVD 事件概率的绝对降低（益处）和严重不良事件概率的绝对增加（危害）。我们将弹性网络正则化方法与传统的向后变量选择方法进行了比较，该方法使用重复的内部交叉验证来选择变量并在存在共线性的情况下估计系数。来自 9,069 名具有完整协变量数据的 SPRINT 参与者的数据用于模型开发，来自 4,498 名具有完整数据的 ACCORD-BP 参与者的数据用于模型验证。参与者接受强化治疗（目标收缩压<120毫米汞柱）与标准治疗（<140毫米汞柱）。评估了两项复合主要结局指标：(i) CVD 事件/死亡（心肌梗死、急性冠状动脉综合征、中风、充血性心力衰竭或 CVD 死亡），以及 (ii) 严重不良事件（低血压、晕厥、电解质异常、心动过缓或急性肾损伤/衰竭）。通过弹性网络正则化选择的 CVD 模型包括交互项，表明年龄较大、黑人种族、较高的舒张压和较高的血脂与强化治疗带来的更大的 CVD 风险降低益处相关，而当前吸烟则与较少的益处相关。通过弹性网络正则化选择的严重不良事件模型表明，男性、当前吸烟、他汀类药物的使用、肌酐升高和血脂较高与强化治疗发生严重不良事件的风险较高相关。预测效益最高亚组中的 SPRINT 参与者需要治疗 (NNT) 的人数为 24，以在 5 年内预防 1 次 CVD 事件/死亡（绝对风险降低 [ARR] = 0.042，95% CI：0.018, 0.066；P = 0.001），预测效益中间亚组的参与者的 NNT 为 76（ARR = 0.013， 95% CI：-0.0001，0.026；P = 0.053），最低亚组的风险没有显着降低（ARR = 0.006，95% CI：-0.007，0.018；P = 0.71）。预测伤害最高的亚组中，诱发 1 次严重不良事件所需伤害 (NNH) 的数量 (NNH) 为 27（绝对风险增加 [ARI] = 0.038，95% CI：0.014、0.061；P = 0.002），预测伤害中间亚组的 NNH 为 41（ARI = 0.025，95% CI：0.012， 0.038；P < 0.001），最低亚组的风险没有显着增加（ARI = -0.007，95% CI：-0.043，0.030；P = 0.72）。在 ACCORD-BP 中，预测获益最高亚组的参与者的绝对 CVD 风险显着降低，但与 SPRINT 相比，总体 ACCORD-BP 参与者样本偏向于预测获益较少且预测风险较高的参与者。通过传统向后选择选择的模型与弹性网络模型具有相似的识别CVD绝对风险差异的能力，但正确识别严重不良事件绝对风险差异的能力较差。该分析的一个关键限制是 ACCORD-BP 试验的样本量有限，该试验扩大了 2 型糖尿病患者 ARI 的置信区间。此外，不可能机械地解释生理关系，解释模型捕获的异质治疗效果，因为该研究是观察性二次数据分析。我们发现，预测模型可以帮助识别 SPRINT 和 ACCORD-BP 中在接受强化血压治疗的 CVD 事件/死亡中 ARR 较低与较高的参与者亚组，以及在严重不良事件中 ARI 较低与较高的参与者。 Sanjay Basu 及其同事利用两项大型临床试验的数据（显示血压治疗效果的异质性）研究了治疗益处和危害的风险如何因人而异。众所周知，血压升高是心血管及相关疾病的主要危险因素。高血压的强化治疗（旨在保持收缩压低于或等于 120 毫米汞柱）可以避免心血管疾病事件，但也可能带来一些严重不良事件的风险。我们试图创建风险计算器来估计个体患者从强化治疗中获益和受损的机会。我们还试图测试称为弹性网络正则化的统计方法（旨在减少过度拟合并提高外部有效性）是否会改善对绝对风险降低或增加的风险模型的估计。我们根据强化血压治疗 SPRINT 试验的个体参与者数据（N = 9,069，具有完整的协变量数据）开发了心血管事件和严重不良事件的统计模型，并根据强化血压治疗的 ACCORD-BP 试验的个体参与者数据（N = 4,498，具有完整的协变量数据）对其进行了验证。我们使用这些模型来计算接受强化血压治疗的个体发生 CVD 事件的概率的绝对降低（益处）和发生严重不良事件的概率的绝对增加（危害）。我们发现，这些模型可以识别心血管事件绝对风险降低高和低的群体，同样，可以识别严重不良事件绝对风险增加高和低的群体。 SPRINT 和 ACCORD 研究中的一些参与者所在的组中预测绝对风险降低程度较高，而预测绝对风险增加程度较低，反之亦然。我们还发现，与传统的向后变量选择相比，使用弹性网络正则化的统计方法提高了识别严重不良事件绝对风险增加高组和低组的能力。我们提供了一个在线风险计算器以及统计代码，以便将该方法应用于其他试验数据集。本研究得出的模型有助于确定 SPRINT 和 ACCORD-BP 中 CVD 事件绝对风险降低程度较低与较高的参与者亚组，以及严重不良事件绝对风险增加较低与较高的参与者。未来，随着越来越多的随机对照试验提供个体参与者数据，用于个性化治疗的获益和危害风险计算器可能会变得更加普遍。研究表明，通过使用弹性网络正则化方法，可以改进严重不良事件的风险计算，该方法涉及严格的交叉验证，并在结果的风险因素（如心血管疾病风险因素）相关时提高模型稳定性。该研究的局限性包括招募 2 型糖尿病患者的强化血压治疗试验 (ACCORD-BP) 的样本量有限，并且是二次数据分析，无法为观察到的治疗效果的异质性提供机制解释。
Intensive blood pressure (BP) treatment can avert cardiovascular disease (CVD) events but can cause some serious adverse events. We sought to develop and validate risk models for predicting absolute risk difference (increased risk or decreased risk) for CVD events and serious adverse events from intensive BP therapy. A secondary aim was to test if the statistical method of elastic net regularization would improve the estimation of risk models for predicting absolute risk difference, as compared to a traditional backwards variable selection approach. Cox models were derived from SPRINT trial data and validated on ACCORD-BP trial data to estimate risk of CVD events and serious adverse events; the models included terms for intensive BP treatment and heterogeneous response to intensive treatment. The Cox models were then used to estimate the absolute reduction in probability of CVD events (benefit) and absolute increase in probability of serious adverse events (harm) for each individual from intensive treatment. We compared the method of elastic net regularization, which uses repeated internal cross-validation to select variables and estimate coefficients in the presence of collinearity, to a traditional backwards variable selection approach. Data from 9,069 SPRINT participants with complete data on covariates were utilized for model development, and data from 4,498 ACCORD-BP participants with complete data were utilized for model validation. Participants were exposed to intensive (goal systolic pressure < 120 mm Hg) versus standard (<140 mm Hg) treatment. Two composite primary outcome measures were evaluated: (i) CVD events/deaths (myocardial infarction, acute coronary syndrome, stroke, congestive heart failure, or CVD death), and (ii) serious adverse events (hypotension, syncope, electrolyte abnormalities, bradycardia, or acute kidney injury/failure). The model for CVD chosen through elastic net regularization included interaction terms suggesting that older age, black race, higher diastolic BP, and higher lipids were associated with greater CVD risk reduction benefits from intensive treatment, while current smoking was associated with fewer benefits. The model for serious adverse events chosen through elastic net regularization suggested that male sex, current smoking, statin use, elevated creatinine, and higher lipids were associated with greater risk of serious adverse events from intensive treatment. SPRINT participants in the highest predicted benefit subgroup had a number needed to treat (NNT) of 24 to prevent 1 CVD event/death over 5 years (absolute risk reduction [ARR] = 0.042, 95% CI: 0.018, 0.066; P = 0.001), those in the middle predicted benefit subgroup had a NNT of 76 (ARR = 0.013, 95% CI: −0.0001, 0.026; P = 0.053), and those in the lowest subgroup had no significant risk reduction (ARR = 0.006, 95% CI: −0.007, 0.018; P = 0.71). Those in the highest predicted harm subgroup had a number needed to harm (NNH) of 27 to induce 1 serious adverse event (absolute risk increase [ARI] = 0.038, 95% CI: 0.014, 0.061; P = 0.002), those in the middle predicted harm subgroup had a NNH of 41 (ARI = 0.025, 95% CI: 0.012, 0.038; P < 0.001), and those in the lowest subgroup had no significant risk increase (ARI = −0.007, 95% CI: −0.043, 0.030; P = 0.72). In ACCORD-BP, participants in the highest subgroup of predicted benefit had significant absolute CVD risk reduction, but the overall ACCORD-BP participant sample was skewed towards participants with less predicted benefit and more predicted risk than in SPRINT. The models chosen through traditional backwards selection had similar ability to identify absolute risk difference for CVD as the elastic net models, but poorer ability to correctly identify absolute risk difference for serious adverse events. A key limitation of the analysis is the limited sample size of the ACCORD-BP trial, which expanded confidence intervals for ARI among persons with type 2 diabetes. Additionally, it is not possible to mechanistically explain the physiological relationships explaining the heterogeneous treatment effects captured by the models, since the study was an observational secondary data analysis. We found that predictive models could help identify subgroups of participants in both SPRINT and ACCORD-BP who had lower versus higher ARRs in CVD events/deaths with intensive BP treatment, and participants who had lower versus higher ARIs in serious adverse events. Using data from two large clinical trials that showed heterogeneity in blood pressure treatment effects, Sanjay Basu and colleagues investigate how risks of treatment benefit and harm vary across individuals. It is known that elevated blood pressure is a major risk factor for cardiovascular and related diseases. Intensive treatment of elevated blood pressure (aimed at keeping systolic blood pressures less than or equal to 120 mm Hg) may avert cardiovascular disease events, but may also pose the risk of some serious adverse events. We sought to create risk calculators to estimate individual patients’ chances of benefit and harm from intensive treatment. We additionally sought to test whether the statistical method known as elastic net regularization, which aims to reduce overfitting and improve external validity, would improve the estimation of risk models for absolute risk reduction or increase. We developed statistical models of cardiovascular events and serious adverse events from individual participant data from the SPRINT trial of intensive blood pressure treatment (N = 9,069 with complete covariate data) and validated them on individual participant data from the ACCORD-BP trial of intensive blood pressure treatment (N = 4,498 with complete covariate data). We used the models to calculate the absolute reduction in probability of CVD events (benefit) and absolute increase in probability of serious adverse events (harm) for individuals from intensive BP treatment. We found that the models could identify groups with high and with low absolute risk reduction in cardiovascular events and, similarly, identify groups with high and with low absolute risk increase in serious adverse events. Some participants in both the SPRINT and ACCORD studies were in groups with high predicted absolute risk reduction and low predicted absolute risk increase, and vice versa. We additionally found that using the statistical method of elastic net regularization improved the ability to identify groups with high versus low absolute risk increase in serious adverse events, compared to traditional backwards variable selection. We made an online risk calculator available, along with statistical code to apply the method to other trial datasets. The models derived in this study helped identify subgroups of participants in both SPRINT and ACCORD-BP who had lower versus higher absolute risk decreases in CVD events, and participants who had lower versus higher absolute risk increases in serious adverse events. In the future, as individual participant data become increasingly available from randomized controlled trials, benefit and harm risk calculators for personalizing therapy may become more common. The study revealed that such risk calculations for serious adverse events were improved by using an elastic net regularization approach that involves rigorous cross-validation and improves model stability when risk factors for an outcome are correlated, as with cardiovascular disease risk factors. The limitations of the study include having a limited sample size in the intensive blood pressure treatment trial that enrolled people with type 2 diabetes (ACCORD-BP) and being a secondary data analysis that cannot provide mechanistic explanations for the observed heterogeneities in treatment effect.
登录
查看更多内容
影响因子：
39.2
作者：
Basu S;Sussman JB;Hayward RA
通讯作者：
Hayward RA
DOI：
10.1161/circoutcomes.117.003624
发表时间：
2017-04
期刊：
Circulation. Cardiovascular quality and outcomes
影响因子：
--
作者：
Patel KK;Arnold SV;Chan PS;Tang Y;Pokharel Y;Jones PG;Spertus JA
通讯作者：
Spertus JA
影响因子：
4.9
作者：
METZ, CE
通讯作者：
METZ, CE
影响因子：
4.2
作者：
Brown, EG;Wood, L;Wood, S
通讯作者：
Wood, S
影响因子：
7.2
作者：
Collins, Gary S.;Reitsma, Johannes B.;Moons, Karel G. M.
通讯作者：
Moons, Karel G. M.