If anything can be said about blood pressure without argument it is that, however meticulous we are with regard to measurement, we are unlikely to record the same value in the same patient on a second (or third) occasion. Blood pressure is extremely variable; from minute to minute, hour by hour, week by week and even differs with the seasons [1,2]. Indeed, it is remarkable how we are able to manage the care of individual patients when recorded differences as great as 30 mmHg in systolic blood pressure can be observed on separate occasions without any change in therapy [3]. Every guideline published emphasizes the importance of careful measurement under standardized conditions but, even within the confines of a randomized controlled trial, patients may be categorized differently (i.e. hypertension or normotension) on repeat measurement [4].

An awareness of the potential for an interaction between the healthcare professional and the patient (alerting reaction, white-coat effect) [5] has increased the use of 'out of office' devices for measuring blood pressure. Ambulatory blood pressure monitors remain expensive and are used exclusively by doctors, but the use of electronic 'home monitors' has mushroomed throughout the world; with and without medical supervision. It is claimed that some 30 million have been distributed in Japan alone [6]. All such devices have the ability to record unlimited numbers of blood pressure readings without the need for the involvement of a doctor or nurse [7]. The observation that the results derived from both ambulatory blood pressure monitoring and self blood pressure monitoring at home better predict clinical outcome than any 'office' measure of blood pressure [8] is almost certainly a testament to the fundamental mathematical fact that reproducibility is enhanced by averaging multiple measurements (the standard deviation of the difference in mean values on separate occasions is at least halved).

The problem arises when trying to decide the optimum number of readings required to ensure a stable blood pressure average that will allow robust clinical decisions to be made. This decision is easier with ambulatory blood pressure monitoring because, even when using a sampling frequency as low as one measure every 30 min, the 24-h average blood pressure value is based on at least 48 measurements, which may be enough to guarantee its reliable and reproducible assessment [9]. When considering home blood pressure monitoring, clearly, one reading is not enough, and 1000 might be excessive. International guidelines vary with respect to the advice offered to clinicians but all emphasize the need to obtain multiple measurements and they often suggest a rejection of the first day's readings (higher than subsequent measurements) [7,10]. At this point, it is worth emphasizing that the results obtained for groups will always be found wanting for the one patient in whom a decision has to be made regarding whether to treat or not to treat. It is axiomatic that, if blood pressure is close to a treatment threshold, different management plans are more likely to be generated on separate occasions than if blood pressure is well outside such cut-off points.

In this issue of the journal, Kawabe et al. [11] have attempted to address this point with their study on home monitoring of blood pressure. They studied a large group of individuals on two occasions, several months apart. The work was carried out using an automated Omron device that is not currently listed on either the British Hypertension Society website (http://www.bhsoc.org ) or on the website supported by members of the European Society of Hypertension Working Group on Blood Pressure Monitoring (http://www.dableducational.com ), and the statements regarding its validation are limited.

All subjects were invited to perform three blood pressure measurements twice a day (on rising and before going to bed) for 1 week and the first day's results were excluded to allow for acclimatization to the technique. A priori, they considered a range of blood pressure comparisons to establish the 'best' way to use self blood pressure monitoring. They looked at many possible permutations, the mean of the first or second measurements at each sitting, the mean of the first and second measurements (or second and third) and finally the mean of all three measurements. They looked at the same set of blood pressure readings taken in the morning and in the evening separately. They refer to the work of Dr Imai who, in contradiction to many hypertension societies, has consistently argued (with supporting evidence) that minimal numbers of measurements may be required in order to obtain a robust home blood pressure value [12].

Most of the subjects in the present study were male and the authors made no attempt to adjust therapy between the two visits. It is not clear how much the drug therapy changed but the relatively low levels of blood pressure recorded suggest that major differences are unlikely. The authors acknowledge from the outset that theirs is a population of predominantly normotensive individuals and thus extrapolation to patients with hypertension may be difficult. We doubt that the fundamentals would differ materially however.

The results obtained are complicated but, in general, blood pressure on the two separate occasions 6 months apart changed insignificantly for the group as a whole with a good correlation between the first and second set of readings. The biggest difference was seen for comparisons of the means of the first reading at each sitting, and this probably reflects the expected greater variability of 'one off' blood pressure readings. Blood pressure appeared to be more variable before going to bed (as measured by the standard deviation of the difference between the two time points) and evening blood pressure was a little lower than morning blood pressure. The results for the group as a whole hide the individual differences for each subject, however. Although the prevalence of hypertension did not change, many individuals changed category when considering blood pressure measurements taken on the two occasions. When considering home blood pressure taken in the morning, for all the five sets of measurements (average of first, of second, of first and second, of second and third and of all three blood pressure readings), approximately 20% of the subjects who were diagnosed as being hypertensive at the time of the first occasion fell into categories other than hypertension on the second occasion 6 months apart. Such a discrepant diagnostic classification was even more evident when considering evening home blood pressure measurements; with approximately 40% of the subjects who were diagnosed as hypertensive on the first occasion falling into other diagnostic categories on the second occasion.

When faced with the finding that either 20% (morning average) or 40% (evening average) of individuals diagnosed as hypertensive in the first period of measurement changed category in the second period, the authors question the accuracy of home blood pressure measurement as a result, but it is likely that this observation reflects normal variation rather than any failure to measure an accurate blood pressure level. We would reiterate the point that, when predominantly normotensive individuals are studied, their borderline blood pressure levels (most individuals in this study) will invariably cross the admittedly somewhat arbitrary thresholds of normotension and hypertension, from time to time.

The authors interpretation of their data is that fewer blood pressure measurements can still be reproducible, but that morning blood pressure values are better in this regard than evening values. Perhaps wisely, the authors refrain from giving us a 'guideline' based upon their data.

The results of this study should be considered on the background of the lively discussion still ongoing regarding how to make the best use of self blood pressure monitoring at home in managing patients with arterial hypertension. Indeed, a number of studies have suggested that the usefulness of home blood pressure monitoring in this setting [13] does not depend only on the statistical advantages associated with the availability of repeated measurements, although there is no doubt that the reliability of average home blood pressure values increases with the increase in the number of measurements available [9,14]. Its usefulness depends also on the possibility offered by even a few home blood pressure measurements to obtain information on blood pressure levels away from the clinic setting, in the real life condition of the patient's home.

A debated issue is the actual value of the data provided on the first day of home blood pressure monitoring and when making use of a single blood pressure measurement on each sitting [10]. Although home blood pressure is devoid of a white-coat effect [15,16], higher and unstable values are usually obtained on the first home blood pressure monitoring day [17,18]. Moreover, home blood pressure tends to decline with repeated measurements, both in the same session [18] and over consecutive days [17-19], similar to what observed for clinic blood pressure measurements [8]. Thus, the choice of focusing on a single home blood pressure monitoring day, and on a single measurement per session, may not be the most appropriate one [10]. Indeed, it was suggested that the initial home blood pressure monitoring day should be considered as a training day, and that the corresponding home blood pressure values should be discarded [17-19]. The suggestion to exclude from data analysis home blood pressure values obtained on the first day of home monitoring has been implemented in the recent recommendations published by the European Society of Hypertension Working Group on Blood Pressure Monitoring [8].

Additional information on this issue has been recently provided in a study by Ohkubo et al. [20], who showed that even a single home blood pressure measurement on the first day offers prognostic information superior to that offered by the average of two clinic blood pressure measurements. This appears to suggest that the setting where blood pressure is measured may be more important than the number of times blood pressure is taken, at least when only a few readings are considered. This observation is in agreement with the results of the SAMPLE study (Study on Ambulatory Monitoring of Pressure and Lisinopril Evaluation), which showed treatment-induced changes in home blood pressure (average of only two measurements) to be significantly related to treatment-induced changes in left ventricular mass index, whereas changes in clinic blood pressure (also the average of two measurements) did not [21]. However, the study by Ohkubo et al. [20] also emphasized that, besides the measurement site, the number of home blood pressure measurements is also an important prognostic factor when considering repeated home blood pressure sessions. Again, this is in agreement with the data of the SAMPLE study in which the relationship between treatment-induced regression of left ventricular hypertrophy and the corresponding blood pressure reduction was closest when considering changes in 24-h average ambulatory blood pressure. This was likely to depend not only on the inclusion of day and night values, but also on the fact that 24-h blood pressure was based on an average of approximately 90 measurements [21].

In conclusion, the European Society of Hypertension Working Group on Blood Pressure Monitoring guidelines recommend duplicate morning and evening home blood pressure measurements to be taken daily for 7 days, and that the measurements taken on the first day should be discarded [8]. Thus, in terms of number of home blood pressure measurements, the European Society of Hypertension Working Group recommends that at least a total of 24 measurements should be averaged. The results of the study by Kawabe et al. [11] appear to suggest that morning blood pressure readings are more reliable and support the conclusions by Ohkubo et al. [20] indicating that taking fewer measurements offers similar benefits to taking several of them. However, when deciding on the minimum frequency of home blood pressure readings to be applied in individual patients in clinical practice, it is important to consider that the conclusions reached for groups of subjects may not invariably apply to an individual patient [10,22]. Although the data by obtained Kawabe et al. [11] suggest that reproducibility of home blood pressure, even when based on a limited number of measurements, appears to be good in groups of subjects, it probably is still inadequate for decision-making in the individual patient [22]. Given that home blood pressure monitoring is easily accepted by patients and has a relatively low cost, it appears thus reasonable to favour the recommendation to average more readings than the statistically reliable minimum number. Additional information is still needed, however, to determine the optimal number of home blood pressure readings to be used in the long-term follow-up of patients.

In addition, and more generally, the study by Kawabe et al. [11] provides a helpful reminder to us all that diagnostic categorization in hypertension remains difficult, even when we have multiple blood pressure measurements to guide us.