The perils of determining cardiovascular risk in relation to home blood pressure

In most guidelines, home blood pressure (BP) measurement is recommended for diagnosing hypertension when 24-h ambulatory BP monitoring (ABPM) is not available or not feasible. This recommendation is based upon evidence derived from a number of clinical outcome studies involving study participants followed over a period of years after being assessed for both office and home BP. The superiority of ABPM over home BP in the guidelines for diagnosing hypertension is based upon several factors, including the number of clinical outcome studies, the sample sizes enrolled, the number of out-of-office BP measurements and the extent to which ABPM readings are superior to office BP in predicting future cardiovascular events. In most instances ABPM comes out ahead, although more clinical outcome studies directly comparing ABPM and home BP are needed to resolve the question of which method is superior. The recent development of home BP devices which are capable of recording BP during sleep [1] may eliminate the differences between these two approaches to obtaining out-of-office BP readings. Furthermore, the use of automated office BP measurements [2] which are devoid of a white-coat effect (office BP higher than ambulatory/home BP) may reduce the superiority of ABPM and home BP over conventional office BP readings. Numerous studies have examined the possibility that individuals with white-coat hypertension (WCH) who have hypertension in the office but a normal BP using ABPM or home BP may be subject to an increased risk of experiencing a cardiovascular event. Despite these efforts, the association between WCH and an increase in cardiovascular events remains uncertain. Individual studies have sometimes reported positive findings, although the overall data are not supportive of a causal relationship. One reason for the somewhat contradictory findings is a mismatch in out-of-office BP between control normotensive study participants and those with WCH. Study participants with WCH may only be at increased cardiovascular risk when their home BP or awake ambulatory BP is higher than it is in the normotensive control group. A recent review [3] of six separate meta-analyses [4–7] from the International Database on Ambulatory BP in Relation to Cardiovascular Outcomes and International Database on Home BP in Relation to Cardiovascular Outcome reported an association between cardiovascular risk and the difference in out-of-office BP between the WCH study participants and normotensive controls. Overall, study participants with WCH tended to have a higher ambulatory or home BP compared with controls. The data from these studies exhibited a statistically significant correlation (P = 0.019 using Spearman's coefficient of correlation analysis) between the hazard ratios for WCH patients experiencing a cardiovascular event and the difference in home or ambulatory BP [3]. Only the two analyses [6,7] with the greatest differences in out-of-office BP reported a statistically significant hazard ratio. Otherwise, these International Database on Ambulatory BP in Relation to Cardiovascular Outcomes and International Database on Home BP in Relation to Cardiovascular Outcome meta-analyses involved a rigorous examination of data from community-dwelling study participants untreated for hypertension, thus enhancing the external validity of their overall findings. It is easy to understand why the literature on the possible cardiovascular risk of WCH may seem confusing. Another meta-analysis of clinical outcome studies published in the same issue [8] as the above review [3] concluded that WCH was a cardiovascular risk factor. This analysis used data from studies which contained higher out-of-office BP values in WCH versus normotensive study participants. In this instance, the authors dismissed the importance of this difference, saying that it ‘could not completely account for the significant increase in cardiovascular disease or total mortality’, without providing any detailed analysis in support of this statement. The results of the Didima Study [9], published elsewhere in the current issue, also contain data which appear to implicate WCH as a cardiovascular risk factor. However, in this instance, the authors recognized other possible explanations for this apparent association. The Didima Study is unusual in that its 662 participants were followed over a lengthy period of 19 years for all-cause mortality and fatal and nonfatal cardiovascular events in relation to home BP readings recorded at baseline. Only 34 (5.1%) study participants exhibited WCH, which in itself is quite a small number compared with much higher proportions in other studies. Nonetheless, the adjusted hazard ratio for cardiovascular mortality and morbidity for the WCH group was 2.67 compared with 1.45 and 1.53 for sustained and masked hypertension, respectively. Similar results were seen for all-cause mortality as an outcome. As this is the only study to have reported such a disproportionately high hazard ratio for WCH compared with masked or sustained hypertension, the authors sought other explanations for their findings. They noted the limited power of the study with a small sample size and only 15 endpoints in the WCH group which could explain the discrepant results, together with the difficulty in ascertaining clinical outcomes over 19 years of follow-up. They also stated that the increase in hazard ratio in the WCH group could be attributed to their mean home BP being 14/7 mmHg higher than with the normotensive controls. Although the analyses were adjusted for antihypertensive drug use, there were marked differences between groups for this factor, with 27% of WCH study participants receiving drug therapy for hypertension compared with only 6% in the control group. We now know that the cardiovascular risk of treated WCH is considerably higher than for untreated WCH [5], which could easily thwart any adjustment of the data. Thus, the finding of a higher rate of cardiovascular events in the WCH group may actually be valid, but the higher risk may have been because of the enrollment of patients with treated hypertension, being classified as having WCH. If so, these results would not be relevant to WCH confounding a diagnosis of hypertension in an untreated patient. Leaving aside any limitations of the Didima Study related to its patient population, it is remarkable that a single set of 12 home BP readings was still able to have prognostic value for both office and home BP with no additional BP data or other measures of cardiovascular risk being collected during 19 years of follow-up. This study is also an example of how difficult it is to conduct clinical research in the community over a prolonged period and still come up with meaningful results. In this study, study participants were under the care of their own physicians who made decisions regarding all aspects of hypertension management, including changes in antihypertensive drug therapy. Clinical outcomes were only ascertained once every decade, with the initial results being published in 2007 [10]. In this earlier report, the primary outcome was combined fatal and nonfatal cardiovascular events, but, because of the uncertainty of accurately ascertaining endpoints in community-dwelling study participants over 19 years, the primary outcome was changed to all-cause mortality in the final data analysis. However, Ntineri et al.[9] also included separate analyses for cardiovascular events in this final report, with the findings being similar to the new primary outcome. A number of lessons can be learned from the Didima experience. In any clinical outcome cohort study involving WCH or masked hypertension, it is important to match groups optimally, especially for factors affecting cardiovascular endpoints, when examining the risks of WCH, such as home BP or ambulatory BP. Normotensive and WCH study participants should ideally be untreated if the results are to be relevant to diagnosing hypertension. Statistical adjustment may not always remove the impact of relevant covariables on the results. Groups should be examined at baseline for any obvious discrepancies, such as the unusually low percentage of study participants with WCH. Despite one-third of participants dying during the 19 years of follow-up, the Didima Study may still have been underpowered for this outcome, resulting in type I error. The same could be said for the analyses on BP variability and cardiovascular risk, using data from the study, which produced results of limited statistical significance. If possible, cardiovascular events should be monitored more frequently than once every decade. Finally, the use of all-cause mortality will avoid missing cardiovascular outcomes which are misclassified, but does so at the expense of including end-points, such as deaths from cancer, which are unlikely to be related to the parameter of interest, in this case BP. During the past three decades, the white-coat effect experienced by 20–25% of patients who are diagnosed with high BP readings in the office has resulted in the overdiagnosis and overtreatment of hypertension. With the increased use of ABPM and home BP to make a diagnosis of hypertension combined with automated office BP for hypertension screening and follow-up, there should be a dramatic reduction in the proportion of patients misdiagnosed and overtreated because of a white-coat effect. Until proven otherwise, ABPM and home BP are the best determinants of cardiovascular morbidity and mortality related to BP partly because, by definition, they are devoid of any white-coat effect. ACKNOWLEDGEMENTS Conflicts of interest There are no conflicts of interest.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.