Self measured and ambulatory blood pressure in assessing the ‘white-coat’ phenomenon

The so-called 'white-coat effect' (i.e. the alerting reaction and pressure rise induced in the patient during a visit by a physician) is probably one of the most annoying problems affecting the clinical evaluation of hypertensive patients through conventional clinic blood pressure measurements by the Riva-Rocci/Korotkoff technique 1–4. Ambulatory blood pressure monitoring (ABPM) and self blood pressure monitoring (SBPM) at home offer clear advantages compared to conventional office blood pressure readings 4–17, at least when accurately validated devices are used 18,19 and appropriate reference values are considered 20–22. These advantages include not only avoidance of the alerting reaction associated with a 'white coat effect', but also a higher reproducibility of the average blood pressure values recorded, the lack of a significant placebo effect, the possibility to monitor blood pressure repeatedly over time, the collection of 'objective' measurements, free from digit preference and observer bias, and a better prediction of outcome compared to 'casual' office blood pressure readings. In particular, SBPM at home offers the possibility of obtaining repeated measurements at different times of the day for several days, weeks, months and even years, which may allow a multiple daily assessment of the blood pressure effects of treatment over a wide time interval. The higher reproducibility of the average values derived from repeated SBPM also carries practical advantages when calculating the sample size of a trial, allowing for a lower number of subjects to be recruited. Such detailed and robust information can be obtained at a lower cost compared to ABPM, and its handling is now made much easier by recent technological progress in data storage, data printing and data tele-transmission 4. Among these advantages, the possibility offered by both ABPM and SBPM for obtaining blood pressure measurements free from interference by the white coat effect has probably been the most important reason for their progressively increasing clinical use. This feature has also led to the proposal of considering the difference between office and either home blood pressure or daytime average ABP as surrogate 23 measures of the 'white-coat effect' 24. Because of its simple implementation, such an indirect approach has been employed in several studies assessing the frequency and/or the clinical relevance of the white-coat phenomenon. This approach is still being proposed despite recent evidence indicating that the surrogate quantification of the 'white-coat effect' derived from the difference between clinic and home or daytime average blood pressure values is unable to faithfully reflect the 'real' white-coat effect, as directly assessed through continuous blood pressure recordings obtained immediately before and during the physician's visit 25–29. Virtually all the evidence obtained in clinical trials on the existence and on the diagnostic and prognostic significance of the white-coat phenomenon has been based on the comparison between clinic blood pressure and average daytime ABP 24,30–34. More recently several investigators have acknowledged that SBPM at home may represent a cheaper alternative method of assessing out-of-clinic blood pressure 15,35. However, the indirect evaluation of the 'white-coat' phenomenon based on the clinic–home rather than on the clinic–daytime blood pressure difference may lead to different surrogate quantifications of the 'white-coat effect' because, although often similar, home and ambulatory daytime blood pressure are not identical. Investigating this problem and performing a more general comparison of conventional blood pressure, ABP and automated SBPM at home in the clinical management of the hypertensive patient, has been the goal of the ongoing Treatment of Hypertension according to Home or Office Blood Pressure (THOP) trial 4. In this issue of the journal, Den Hond et al. 36 focus on the analysis of baseline THOP data, aiming to specifically address whether, and to what extent, SBPM at home can be a reliable alternative to ABPM in the surrogate quantification of the 'white-coat effect' and in the assessment of 'white-coat hypertension'. In their study, Den Hond et al. 36 first compared the surrogate 'white-coat effect' quantified by considering home blood pressure and an assessment of the same phenomenon obtained by taking daytime ABP into account (i.e. clinic–home blood pressure difference versus clinic–daytime ABP difference). Second, they investigated the respective ability of persisting differences between clinic blood pressure levels and either home SBP or daytime ABP average values to diagnose 'white-coat hypertension'. Using the former approach, when calculating the clinic–home and clinic–daytime average blood pressure differences, the authors dealt with an indirect (and indeed poorly accurate) measurement of blood pressure changes related to the physician's visit (`white-coat effect'). Using the latter approach, by assessing whether clinic blood pressure on the one hand, and home and ambulatory blood pressure on the other, are above or within normal limits, the authors focused on differences in blood pressure levels, with the aim of diagnosing whether a sustained or a 'white-coat' hypertension occurred 32,37. The measurements available at baseline in 247 patients of the THOP trial showed that the difference between office and home blood pressure was larger than the difference between office and ambulatory daytime blood pressure, leading to a different surrogate quantification of the 'white-coat effect', according to whether office blood pressure was compared to home or daytime blood pressure. However, despite their quantitative discrepancies, the difference between office and daytime blood pressure was significantly correlated with the difference between office and home blood pressure values. When focusing on blood pressure levels, and by taking ABP as a reference, the specificity of SBPM to detect 'white-coat hypertension' was 88.6% and its sensitivity was 68.8%. The conclusions made by the authors were that, although the surrogate measures of 'white-coat effect' based, respectively, on daytime ABP and on SBPM are correlated, they identify different subsets of patients with 'white-coat hypertension', and suggest that these techniques have a supplementary, rather than a competitive, role in the diagnostic approach to hypertensive patients. Home and ambulatory blood pressure in the indirect assessment of 'white-coat effect' and 'white-coat hypertension' Although home and ambulatory blood pressure share several common features, given that both methods provide blood pressure values outside of a clinic and in the usual daily environment of the individual patient, they are also characterized by important differences because SBPM is performed under strictly standardized conditions of posture and activity (usually after a few minutes of sitting rest) and mostly at home, whereas ABP is measured under dynamic conditions, when subjects are involved in their usually variable daily activities either at work, at home and even during sleep 35,38. Thus, it cannot be assumed a priori that these two methods are completely interchangeable with one another, and studies aimed at testing this possibility, as was the case for the one carried out by Den Hond and colleagues, are indeed needed. Surrogate assessment of the 'white-coat effect' The magnitude of the surrogate 'white-coat effect' based on the above reported indirect methods is heavily dependent on the number of clinic visits considered for the calculation of the difference between clinic and out-of-clinic blood pressure. Most, if not all, of the progressive decline of blood pressure on repeated clinic measurements occurs in the first three visits 39,40. Therefore, it is recommended that clinic blood pressure measurements obtained on at least three visits should be considered for a reliable surrogate assessment of the 'white-coat effect' 32. Furthermore, this surrogate 'white-coat effect' is generally greater when clinic blood pressure is taken by a physician rather than a nurse 3,40. Although, in the THOP study, clinic blood pressure was measured only by physicians, and on only two visits, the magnitude of the surrogate 'white-coat effect' assessed by the difference between clinic and daytime ABP (12.3/8.6 mmHg) was similar to that reported in previous studies 37,41. However, it could be argued that this surrogate 'white-coat effect' might have been of a smaller magnitude had the investigators decided to assess clinical blood pressure on an additional (third) visit. Interestingly, in the THOP study, the magnitude of the surrogate 'white-coat effect' assessed by the clinic–home blood pressure difference was 5.0/3.5 mmHg (systolic/diastolic) higher compared to that assessed by considering ABPM 36. This difference resulted in a higher proportion of 'white-coat' hypertensives detected by SBPM at home (15.8%) than by ABPM (7.7%) 42. These data are not in accordance with the results of a recent review of those studies dealing with this issue, which did not show any consistent difference in the magnitude of the surrogate 'white-coat effect' assessed by considering home or ambulatory blood pressure measurements 37. On the other hand, in studies where home and ambulatory blood pressure values were compared in their ability to indirectly detect the same magnitude of surrogate 'white-coat effect' 10,43–45 and the same proportion of clinic hyper-reactivity 10,45 were found by the two methods. Moreover, when the surrogate evaluations of the 'white-coat effect' based on the differences of clinic blood pressure with SBP and ABP, respectively, were obtained in hypertensive patients with evidence of left ventricular hypertrophy, the former approach gave smaller values than the latter, at variance from the results of Den Hond et al. 36 because, in these patients, home blood pressure was higher than ambulatory daytime blood pressure 41. The protocol for the assessment of home and ambulatory blood pressure may significantly affect the reliability and the magnitude of the average out-of-clinic blood pressure values considered, and thereby the derived surrogate assessments of 'white-coat effect'. The number of days over which SBP at home is obtained has been shown to be important in this setting, and the average value of duplicate morning and evening blood pressure measurements on 3 days has been suggested as the minimum requirement for a reliable assessment of representative home blood pressure levels 10,12,46,47. The data of the THOP study confirm these findings, given that the average home blood pressure of the first 3 days did not differ from the average home blood pressure of 7 days 36. Furthermore, because the initial home blood pressure monitoring day has been shown to provide higher and unstable blood pressure values, some investigators have suggested these measurements be excluded 8,10,47. However, even if in the THOP study systolic home blood pressure of the initial monitoring day was the highest, this difference did not reach statistical significance compared to values obtained in the remaining days, and therefore the authors decided not to exclude these measurements 36. Based on these considerations, it appears that none of the above-mentioned factors could be held responsible for the lower home blood pressure compared to daytime ABP values in the THOP study, a finding that deserves further assessment. Two additional issues concerning the protocol for SBPM at home should be discussed in the light of the results reported by Den Hond et al. 36. First, the number of home blood pressure measurements performed per session has to be carefully considered because, again, a progressive decline in blood pressure is observed on repeated measurements during a given home blood pressure measurement session 12,44,47,48, a decline similar to that shown for clinic blood pressure 39,12. In the THOP study, three home blood pressure measurements were obtained per session, which may have contributed to the lower average home blood pressure obtained compared to studies using only two 10,12,43 or even a single measurement 21. Second, it appears that, in the THOP study, home blood pressure measures taken on weekends were also included. Measurements on weekends provide lower blood pressure values than measurements on work days 50. It is therefore possible that both the inclusion of triple home blood pressure measurements per session and of measurements on weekends might have contributed to lower home blood pressure values, and thereby to the higher magnitude of the SBPM-based surrogate white-coat effect in the THOP trial. Despite the difference in the magnitude of the 'white-coat effect' assessed by SBPM and ABPM, a strong correlation was found between the surrogate assessments of this phenomenon based on the two methods, with correlation coefficients of 0.74 and 0.60 for systolic and diastolic blood pressure, respectively 36. These data are in accordance with the findings of a previous study that included both treated and untreated hypertensives, where similar correlation coefficients were found between the 'white-coat effects' based on home and ambulatory blood pressure measurements (0.64 and 0.59 for systolic and diastolic blood pressure, respectively) 12. Diagnosis of 'white-coat hypertension' In untreated patients with a persistently elevated clinic blood pressure, similar to the participants in the THOP study, clinical decisions depend more on whether a diagnosis of 'white-coat hypertension' can be confirmed or excluded, making use of out-of-clinic blood pressure levels, rather than on the magnitude of the surrogate 'white-coat effect' (i.e. on the absolute difference between clinic and either home or ambulatory blood pressure; the clinical relevance of which is as yet poorly understood) 41. White-coat hypertension is characterized by persistently elevated blood pressure values in the clinical environment, and by persistently normal blood pressure levels at other times. Because of this feature, and because of the previously reported lack of correspondence between direct and surrogate measures of the 'white-coat effect' 25–29, it has been suggested that this persisting discrepancy between clinic and out-of-clinic blood pressure levels should be better termed 'isolated office hypertension' rather than 'white-coat hypertension' 50, a suggestion that was included in the last WHO/ISH Guidelines 51. ABPM is commonly regarded as the standard method for the assessment of 'white-coat hypertension', against which any new method should be tested. Use of home blood pressure has also been proposed 42,51–55, but its value in this setting has not yet been definitely established. If home blood pressure is to be applied in the assessment of the 'white-coat hypertension', the question is whether it could be used as an alternative method to ABPM, or whether it should only be employed as a screening test that requires confirmation by the standard method (i.e. ABPM). Indeed, the similarity in the proportions of white-coat hypertensives detected by home or ambulatory blood pressure monitoring in previous studies does not necessarily mean that the same patients are concordantly classified by the two methods 12. In the THOP study, disagreement between SBPM and ABPM in the diagnosis of 'white-coat hypertension' was found in 13% of participants, suggesting moderate to poor agreement between the two approaches. Previous studies comparing SBPM at home with ABPM in the diagnosis of 'white-coat hypertension' 45, or in the detection of subjects with a clinically important 'white-coat effect' 12, again yielded discordant diagnostic conclusions between the two methods in 20–25% of subjects. This between-method discrepancy may be attributable to inherent limitations of the blood pressure measurement techniques employed, such as their imperfect reproducibility 7,10, as well as to inherent methodological differences, namely 35. It has also been suggested that the home blood pressure method might misdiagnose the occurrence of out-of-office hypertension in those patients who only exhibit an increase in blood pressure at work 56. The use of home blood pressure as a screening test for white-coat hypertension was first proposed in 1990 by Pickering and colleagues 50,53. On the other hand, the use of home blood pressure as a diagnostic test in the detection of 'white-coat hypertension' was first proposed in 1993 by Appel and Stason 53. In 1997, Nesbitt et al. 42 suggested that, in borderline hypertension, the decision to treat could be made solely on the basis of an elevated home blood pressure (diagnostic test). The use of SBPM as a screening test for white-coat hypertension was endorsed by an Ad Hoc panel of the American Society of Hypertension 55, by the First International Consensus Conference for Self-Blood Pressure Monitoring 16 and by several other investigators 40,46. The recent decision of the US Centers for Medicare and Medicaid Services to provide limited coverage of ABPM only in cases of suspected white-coat hypertension is also based on out-of-clinic (home) blood pressure measurement as a screening test before performing ABPM 57. The data in the THOP study allow for the diagnostic value of SBPM at home to be tested. With ABPM taken as reference method, home blood pressure was found to have high specificity and a high negative predictive value (89 and 97%, respectively), but moderate sensitivity (68%) and low positive predictive value (33%). Two previous studies have assessed the usefulness of home compared to ambulatory blood pressure in the diagnosis of sustained hypertension and 'white-coat hypertension' 42,45. The first reported high specificity and high negative predictive values (79 and 87%, respectively), but moderate sensitivity and moderate-low positive predictive value (61 and 48%) of home blood pressure in the diagnosis of 'white-coat hypertension' 45. The second study also concluded for the high specificity (93%) with low sensitivity (43%) of this approach 42. A further study, comparing home with ambulatory blood pressure in the detection of subjects with a significant 'white-coat effect' among untreated and treated hypertensives, also revealed high specificity and high negative predictive value (85% for both), but moderate sensitivity and moderate positive predictive value (57% for both) of home blood pressure in the correct diagnosis of blood pressure hyperreactivity to a clinic environment 39. There is significant agreement among the above four studies in that SBPM at home has high specificity and high negative predictive value but lower sensitivity and lower positive predictive value in the detection of 'white-coat hypertension'. Given that, for a diagnostic test, a high sensitivity and a high positive predictive value are essential requisites 58, based on these data, SBPM at home does not appear to represent a suitable candidate to be proposed as an alternative to ABPM in the diagnosis of 'white-coat hypertension' 12,45. On the other hand, for a reliable screening test, it is very important to have high specificity and high negative predictive value 58, as appears to be the case for SBPM at home. Thus, as Den Hond et al. 36 conclude, SBPM at home may be useful as a screening test for 'white-coat hypertension'. Nevertheless, for a screening test, it is also important that false negative results can be prevented at the cost of false positive ones 58. Therefore, an important limitation of home blood pressure as a screening test in the diagnosis of 'white-coat hypertension' could be its unsatisfactory sensitivity (43–68%). There are two more arguments opposing the view that decision-making should be based solely on home blood pressure and in favour of restricting this method to screening purposes. First, the prognostic significance of home blood pressure remains largely unknown 16,21,59, whereas ABP has been proven to provide additional prognostic information, over and above that provided by clinic blood pressure 38,51,59. Second, there is a potential for misreporting of home blood pressure measurements by patients, which has been shown to occur in up to 50% of cases 60,61. Although this error can be prevented by using memory-equipped devices, at present, these devices are not widely employed because of their cost. However, despite these potential problems, it should not be disregarded that SBPM at home has an overall much lower cost than ABPM. This feature, together with its wide availability and acceptability, supports the role of SBPM as a useful screening test, although some methodological difficulties should be acknowledged. In conclusion, the THOP study provides support for previous observations that SBPM at home should not be regarded as an alternative to ABPM in the detection of 'white-coat hypertension'. On the other hand, this method appears to be useful mostly as a screening test but, in this case, the physician should be aware that false negative findings might not be uncommon, and that ABPM may represent an important complementary procedure for a safer diagnosis.