采访撷要

　　血压变异与动态血压测量

　　血压变异尚不能用于临床实践
　　血压变异有很多种类，包括24小时内的血压变异、每日的长期血压变异、季节变异、每次就诊时的血压变异、每次心跳之间的血压变异和血压变化速度的变异。对血压变异的研究刚刚起步，很多问题没有解答，如为什么高血压患者血压变化的速度比血压正常者更迅速。这些现象的临床意义尚未被阐明。我们有关于某些类型血压变异的数据，但不是全部类型，血压变异相关证据比血压值的证据少，仍仅处研究阶段，用于临床还存在很多问题。目前，已有一些观察性研究显示24小时血压平均值的标准差有重要预后价值，但在另外一些试验中这种血压变异的意义则不明显。这一领域最大的缺陷是缺乏治疗期间评价血压变异意义的试验。在所有大规模试验中，研究者们没有测量动态血压，或者即使测量了动态血压，也是在没有随机对照的小样本的亚组中进行，样本量太小不足以对心血管事件进行有意义的分析。
　　NICE指南没有考虑血压变异
　　NICE指南并没有考虑到血压变异，而只关注白天血压的平均水平。我认为动态血压应该被视作常规检查。首先，夜间血压具有预后价值，非常重要，在一些情况下，夜间血压的预后价值大于白天血压。其次，指南认为在1小时内只测量一个血压值即可，所以测得的血压值总数是14次。但是指南并没有指出这种建议的依据是如何从研究中搜集得到的；以前动态血压会测量多次，达到70~90次。我认为采取与之前不同的方式收集证据，并假设更少的血压测量次数不影响血压评估的做法值得质疑。对于就诊时血压很高而白天血压正常者，NICE指南将其归为白大衣性高血压，并认为患者血压正常。这将排除介于真正高血压和正常血压之间的中度风险患者。这类患者群心血管风险更高，十年间发生糖尿病、确诊高血压以及左心室肥厚的几率更大，同时合并多种代谢性危险因素。

　　《国际循环》：就目前临床上关于血压变异性和血压高低的问题来说，为什么血压升高的程度仍然是最重要的？我们该如何关注血压变异性？
　　Dr Mancia：有很多原因。首先，血压变异性有很多种类型，比如，24小时内的血压变异性。血压在白天和夜间都是有很大变异的。在夜间，血压普遍偏低但是除此之外在24小时内还有极短期持久峰或长时持久峰。然后就是每日的长期血压变异性，季节变异性和每次就诊时的血压变异性。这是一种非常复杂的现象。我们才刚刚开始研究血压变异性，而且还有很多问题仍然没弄清楚。还有每次心跳之间的血压变异性和血压变化速度的变异。例如，从研究中我们知道，高血压患者血压变化的速度要比血压正常的人更迅速。这些现象的临床意义还没有被阐明。我们有关于某些类型血压变异性的数据，但不是全部类型。关于血压变异性的所有证据比经典血压，临床血压的证据要少得多，我们很早前就有很多试验数据来证实临床血压的重要性了。所以不能将血压变异性和血压水平来做对比。一个是已经被研究了很多年并已经应用于临床实践，而另一个还仍然是研究的方法并且要想应用于临床实践还存在很多问题。就考虑到24小时血压变异性而言，可以进行综合测量，也就是24小时血压平均值的标准偏差，并且已经有一些观察性研究显示这种血压变异性有重要的预后价值。也就是说，对于一个已知的平均24小时血压水平而言，那些多年来血压标准偏差最大的患者将更有可能发生心血管事件。在一些研究中，在进行多变量分析时这种变异性和平均24小时血压相比似乎是一个非常重要的因素，但是并非所有的研究都得出这样的结果。在另外一些试验中这种变异性的重要性就很微小。这一领域最大的缺陷是没有试验在治疗期间评价血压变异性的意义。目前为止还没有研究对此进行分析。没有研究探索治疗期间的动态血压，这是一个很大的缺陷。在所有的大规模试验中，研究者们没有动态血压测量，或者即使最近的研究测量了动态血压，也是在没有随机对照的小样本的亚组中进行的，患者太少以至于不能对心血管事件进行有意义的分析。这就是目前的现状。每次就诊时的血压变异也非常有意思，具有一些特别的意义。例如，我们已经知道很多关于24小时血压变异的机制，因为从1968年到1970年间就已经做了很多研究。我们知道交感神经系统参与其中；知道很多机制。而对于就诊血压变异而言，我们真的一无所知。它可能是技术性原因比如每次就诊时测量血压存在误差，或者是因为在一次就诊时所测的血压正是低谷时的血压而下次就诊时测得的血压又是峰值的血压，或者是介于低谷和峰值之间的血压。关于就诊血压变异的第一组数据是再6、7年前收集到的，结果显示的是就诊血压变异更大的患者似乎更容易发生不良事件，特别是中风。INVEST试验是一项纳入了2300例曾经发生冠脉事件的高血压患者的大规模临床试验，在INVEST试验的分析中我研究了这些数据并将患者根据就诊率分为从小于25% 到大于75%四组，就诊期间患者血压控制在140/90mmHg以下。当把这四组进行对比并对治疗4年来的平均临床血压进行校正后，发现那些就诊频率不高的患者比就诊频率高的患者更容易发生不良事件，特别是中风（但是并不是全部是中风）。2年前由Rothwell 教授及其同事发表于Lancet杂志的研究也用一种非常严谨的方式对此相同的问题进行了或多或少的研究。研究显示在整个治疗期间都存在平均临床血压的标准偏差，而且那些标准偏差或离散系数更大的患者比那些标准偏差小的患者不良事件发生率更高。这样血压标准偏差变异的意义似乎与就诊过程中血压控制的意义一致。那这里的问题是什么呢？这是一个典型的事后分析。像我们在INVEST试验和Rothwell教授在ASCOT 和UK-TIA试验中所做的分析一样，你永远不能肯定的是当你把这两组进行非随机的对比时，即使对所有的变量都进行校正后所得出的差异是由于一些其它的原因所致还是由血压变异性所致。不管怎么说校正实际上是统计学手段；这种手段不能保证各组在开始时就被全面比较。另一个问题是为了计算每次就诊的变异性，你需要计算就诊次数。但是在计算就诊次数时还不能包括初始给药的滴定期，因为在给药滴定期时我们治疗性地改变了血压，因此还必须从滴定期结束后大概治疗稳定的时间开始计算。这意味着研究中早期发生不良事件的患者只有后半部分试验能记录他们的资料。如果每6个月就诊一次，那么在18个月时发生不良事件的患者就不能入选，因为他们可能只有两次数据记录并且不能从这两次数据中计算出标准偏差。这就是关于血压变异性研究的另一个缺陷，因为你需要排除试验早期的部分资料。尽管如此，研究的结果还是显示血压变异性有一些重要性。在Rothwell等的分析中，他们认为这是一个非常重要的因素但是我认为还是应该对血压变异性进行研究。
　　International Circulation: With regard to the variability of blood pressure versus the severity of blood pressure， in the clinical right now， why is the severity still most important and how much attention should be given to variability?
　　Dr Mancia： There are many reasons. First of all， there are different types of blood pressure variability， for example， the variability of blood pressure within a 24-hour period. Blood pressure is extremely variable throughout the day and night. During the night， blood pressure is systematically lower but in addition to that there are very short-lasting peaks or longer-lasting peaks within the 24-hour period. Then there is the longer-term variability from day to day， seasonally and visit-to-visit variability in blood pressure. It is a very complex phenomenon. We are beginners in studying blood pressure variability and there is still a lot to be learned. There is also beat-to-beat variability and variability in the speed of blood pressure changes. For example， it is known from our studies， that hypertensive patients have faster blood pressure changes than normotensive people. The clinical value of all these phenomena has not been documented. We have some data on some types of variability but not all types. The overall evidence is much less than the evidence on classical blood pressure， clinic blood pressure， for which we have data from trials from way back. So you cannot compare the two phenomena. One is something which has been studied for many years and used in clinical practice; the other one is still a method of research and is a lot more difficult to use in clinical practice. As far as 24-hour variability is concerned， you can have a comprehensive measurement which is the standard deviation of the 24-hour blood pressure mean and there are several observational studies that show that this variability has prognostic value and importance. That is to say， for a given level of mean 24-hour blood pressure those with the greatest blood pressure standard variation over the years tend to have a greater incidence of cardiovascular events. In some studies， when multi-variant analysis was done， it looked like it was quite an important factor compared to mean 24-hour blood pressure but not in all studies. In some other studies， it was there but it was of more marginal importance. The big limitation of this is that no trial has been done on the value of blood pressure variability during treatment. No study has looked at this. No studies have looked at ambulatory blood pressure during treatment; this is a big limitation. In all of the big trials， either they didn’t have ambulatory blood pressure measurement or recently， they had， but it was a small subgroup with non-randomized comparisons and too few patients to have a meaningful analysis of cardiovascular events. This is the situation. Visit-to-visit variability is quite interesting and in a somewhat different position. For example， for 24-hour variability， we know quite a lot about the mechanisms involved because studies have been done since 1968 and 1970. We know the sympathetic nervous system is involved; we know a lot of things. For visit-to-visit variability， we really do not know. It could just be the technical fact that there are errors in measuring blood pressure between visits or the fact that at one visit， blood pressure is measured at trough and at the next visit measured at peak or in between. The first sets of data were collected six or seven years ago， and what was found was that it looks like patients having greater visit-to-visit blood pressure variations tend to have more events， particularly stroke. In an analysis of the INVEST study， which was big trial of about 23000 hypertensive patients with a history of coronary events， I looked at these data and divided the patients into four groups according to the percent of visits at which blood pressure was controlled (<140/90mmHg) from <25% of the visits to >75%.When these four groups are compared and data was adjusted for mean clinic blood pressure throughout the four years of treatment， those for which control was less frequent had more events compared to those whose control was more frequent， particularly stroke (but not always stroke). Data was published in The Lancet about two years ago by Professor Rothwell and co-workers and he looked at more-or-less the same thing in a very elegant fashion. There is the standard deviation of the mean clinic blood pressure throughout the treatment period， and those having a greater standard deviation or co-efficient of variation， had again a greater incidence of events compared to those having a lower standard deviation. Again it looks like consistency of blood pressure control throughout the visit is of value. So what are the problems here? This is a typical post hoc analysis. With what we did in INVEST and what Professor Rothwell did from ASCOT and the UK-TIA study， you can never be sure that when you compare these two groups， non-randomized， that the differences are due to some other factors rather the variability of blood pressure despite adjustment for all variables. Adjustments are really statistical gimmicks anyway; they cannot guarantee the groups are fully comparable at the beginning. One other problem is that in order to calculate visit-to-visit variability， you need a number of visits. You can not include the visits during the titration period because during titration， we change blood pressure therapeutically， so you have to take visits from the end of the titration period onwards when presumably treatments are stable. This means that patients with an early event cannot be included and only the late part of the trial can be included. If you have visits every six months， patients who have an event at 18 months are not candidates of course because you only have two values maybe and you cannot calculate standard deviation from two values. This is another limitation because you need to exclude the early part of the trial. Despite this， the data suggests that this can be of some importance. In the analysis of Rothwell et al， the suggestion was that this was quite an important factor but I think this still has to be investigated.