Abstract
The aim of this investigation was to find a time segment in which average blood pressure (BP) has the best correlation with 24‐hour BP control. A total of 240 patients with full ambulatory BP monitoring (ABPM) were included; 120 had controlled BP (systolic BP [SBP] ≤135 mm Hg and diastolic BP [DBP] ≤85 mm Hg) and 120 had uncontrolled BP (SBP >135 mm Hg and/or DBP >85 mm Hg). Each ABPM was divided into 6‐ and 8‐hour segments. Evaluation for correlation between mean BP for each time segment and 24‐hour BP control was performed using receiver operating characteristic curve analysis and Youden's index for threshold with the best sensitivity and specificity. The mean BP in the following segments showed the highest area under the curve (AUC) compared with average controlled 24‐hour BP: SBP 2 am to 8 am (AUC, 0.918; threshold value of 133.5 mm Hg, sensitivity−0.752 and specificity−0.904); SBP 2 pm to 10 pm (AUC, 0.911; threshold value of 138.5 mm Hg, sensitivity−0.803 and specificity−0.878); and SBP 6 am to 2 pm (AUC, 0.903; threshold value of 140.5 mm Hg, sensitivity−0.778 and specificity−0.888). The time segment 2 pm to 10 pm was shown to have good correlation with 24‐hour BP control (AUC >0.9; sensitivity and specificity >80%). This time segment might replace full ABPM as a screening measure for BP control or as abbreviated ABPM for patients with difficulty in performing full ABPM.
Blood pressure (BP) measured by ambulatory BP monitoring (ABPM) is more closely associated with target organ damage, and is considered the most accurate method to evaluate true BP.1, 2 In addition, ABPM is the only technique to evaluate nighttime BP and early morning surge.1 Although ABPM is an important tool in evaluation and follow up of hypertensive patients, its use in clinical practice may be limited by availability, cost, and patient inconvenience. For accurate measurements, the patient is required to keep the cuff on the arm during the entire measurement period (usually 24 hours) as well as having to wear the monitor unit on the waist (by a belt or strap) during the day and keep it at the bedside at night.3
Because of these disadvantages, an easier and shorter method is required. Some studies compared various techniques for clinical BP measurements with full ABPM.3, 4 Ernst and colleagues5, 6 demonstrated the accuracy of a shortened ABPM session of 6 hours in classifying BP as controlled or uncontrolled. Due to the above difficulties in performing full ABPM and the current knowledge about optional shortened ABPM, we aimed to find a time segment during which average BP has the best correlation with 24‐hour BP control.
Patients and Methods
Study Population
In this cross‐sectional study, data were extracted from ABPM studies performed on hypertensive patients from our institute during 2010–2011. The patients were defined as hypertensive if they met one of the following criteria:
Recorded diagnosis of hypertension on their referral.
Recorded diagnosis of hypertension in their electronic medical file.
Recorded current utilization of thiazide diuretics.
Group Classification
According to the European Society of Hypertension/European Society of Cardiology Task Force on the Management of Arterial Hypertension Guidelines, normal mean 24‐hour BP is considered systolic BP <130–125 mm Hg and diastolic BP <80 mm Hg.7
In our study, we grouped the ABPM studies into 2 groups with a cut point of 135/85 mm Hg4, 8:
Controlled BP by ABPM: 24‐hour BP defined as mean 24‐hour SBP <135 mm Hg and mean 24‐hour DBP <85 mm Hg.
Uncontrolled BP by ABPM: 24‐hour BP defined as mean 24‐hour SBP ≥135 mm Hg and/or mean 24‐hour DBP ≥85 mm Hg.
Due to power calculation of 0.8, 120 ABPM studies were enrolled in each group.
ABPM Study
The ABPM studies were performed using the Oscar 2 system (SunTech Medical, Morrisville, NC) device. For each ABPM study, the patient was provided with the appropriate‐sized BP cuff as determined by arm circumference according to the American Heart Association guidelines for BP measurement.9 BP readings were obtained every 20 minutes during daytime (6 am to 10 pm) and every 30 minutes during nighttime (10 pm to 6 am). Only ABPM studies with at least 70% of expected measurements were included.
The whole day of 24 hours was divided into 3 time segments of 8 hours each and into 4 time segments of 6 hours each. The time segments included the following times: 2 pm to 10 pm, 10 pm to 6 am, 6 am to 2 pm, 2 pm to 8 pm, 8 pm to 2 am, 2 am to 8 am, and 8 am to 2 pm. For each segment, the mean SBP, DBP, and pulse pressure (PP) were extracted.
Clinical Data
For each patient, the following clinical data were included: age, sex, presence of diabetes mellitus, antihypertensive drug treatment, smoking habits, low‐density lipoprotein (LDL) level, creatinine level, presence of ischemic heart disease, and cerebrovascular disease.
Statistical Analysis
Student t test for independent samples and chi‐square test were used for comparison of the main variables by controlled and uncontrolled hypertensive groups as appropriate.
For sensitivity analysis and testing predictability of ABPM, time intervals receiver operating characteristic (ROC) curve analysis was used. Youden statistic was used to identify the optimal cut point of each time interval of ABPM. Youden's J statistic Y = sensitivity – (1‐specificty).
A 2‐tailed P value ≤.05 was considered significant. The statistical analysis was performed using SPSS version 18, SPSS Inc, Chicago, IL). The study received the approval of the ethics committee of our institute.
Results
Basic Characteristics: Controlled vs Uncontrolled Hypertension According to 24‐Hour ABPM
A total of 240 ABPM studies were included, 120 with controlled and 120 with uncontrolled 24‐hour ABPM (Table 1). There was no significant difference in their demographic and risk factors, including age (56.86±8.42 vs 55.31±8.97, respectively; P=.16), sex (52.5% men vs 46.7% men, respectively; P=.36), diabetes mellitus (32.5% vs 22.5%, respectively; P=.083), and LDL levels (99.83±30.57 vs 107±28.9, respectively; P=.067). As expected, there was a significant difference in 24‐hour BP measurements between the controlled vs uncontrolled group: 24‐hour SBP/DBP (125.72±7.63/72.43±6.71 vs 147.49±11.49/85.08±9.75, respectively; P<.01/.01). There were also prominent differences between the groups in BP measurements in both daytime and nighttime values.
Table 1.
Basic Characteristics of Study Population: Controlled vs Uncontrolled Hypertension
| Controlled Hypertension (n=120) | Uncontrolled Hypertension (n=120) | P Value | |
|---|---|---|---|
| Age, y | 56.86 (8.42) | 55.31 (8.97) | .167 |
| Men/women, % | 52.5/47.8 | 46.7/53.3 | .366 |
| Diabetes mellitus,No. (%) | 39 (32.5) | 27 (22.5) | .083 |
| Smokers, No. (%) | 23 (19.2) | 23 (19.2) | 1 |
| LDL, mg/dL | 99.83 (30.57) | 107 (28.9) | .067 |
| Antihypertensive medications, No. | 2.1 (0.87) | 2.09 (0.7) | .945 |
| 24‐h mean SBP | 125.72 (7.63) | 147.49 (11.49) | <.001 |
| 24‐h mean DBP | 72.43 (6.71) | 85.08 (9.75) | <.001 |
| Daytime mean SBP | 129.1 (8.64) | 149.31 (11.86) | <.001 |
| Daytime mean DBP | 74.69 (7.2) | 86.44 (10.46) | <.001 |
| Nighttime mean SBP | 118.42 (16.76) | 139.34 (12.82) | <.001 |
| Nighttime mean DBP | 66.13 (8.02) | 78.43 (9.76) | <.001 |
Abbreviations: DBP, diastolic blood pressure; LDL, low‐density lipoprotein; SBP, systolic blood pressure.
Time Segment With the Best Correlation to Controlled and Normal 24‐Hour BP
In order to find which time segment has the best correlation with 24‐hour ABPM, we performed ROC curve analysis of SBP, DBP, and PP of each time segment (2 pm–10 pm, 10 pm–6 am, 6 am–2 pm, 2 pm–8 pm, 8 pm–2 am, 2 am–8 am, and 8 am–2 pm) against 24‐hour controlled BP (24‐hour SBP <135 mm Hg and 24‐hour DBP <85 mm Hg) and against 24‐hour normal BP (24‐hour SBP <130 mm Hg and 24‐hour DBP <80 mm Hg) (Table 2).
Table 2.
Time Segment by Controlled and Normal 24‐h BP: Area Under the Curve
| Test Result Variable for Time Segments | 24‐h SBP <135 mm Hg and DBP <85 mm Hg | 24‐h SBP <130 mm Hg and DBP <80 mm Hg | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Area | Standard Error | Asymptotic Sig. | Asymptotic 95% Confidence Interval | Area | Standard Error | Asymptotic Significance | Asymptotic 95%Confidence Interval | |||
| Lower Bound | Upper Bound | Lower Bound | Upper Bound | |||||||
| Mean SBP 2 am–8 am | 0.918 | 0.018 | <.001 | 0.884 | 0.953 | 0.910 | 0.021 | <.001 | 0.869 | 0.951 |
| Mean DBP 2 am–8 am | 0.834 | 0.026 | <.001 | 0.783 | 0.884 | 0.848 | 0.025 | <.001 | 0.799 | 0.897 |
| Mean SBP 10 pm–6 am | 0.880 | 0.021 | <.001 | 0.838 | 0.922 | 0.884 | 0.023 | <.001 | 0.838 | 0.929 |
| Mean DBP 10 pm–6 am | 0.811 | 0.027 | <.001 | 0.758 | 0.865 | 0.834 | 0.027 | <.001 | 0.781 | 0.887 |
| Mean SBP 2 pm–10 pm | 0.911 | 0.018 | <.001 | 0.875 | 0.947 | 0.926 | 0.017 | <.001 | 0.893 | 0.958 |
| Mean DBP 2 pm–10 pm | 0.814 | 0.028 | <.001 | 0.759 | 0.868 | 0.848 | 0.025 | <.001 | 0.799 | 0.897 |
| Mean SBP 6 am–2 pm | 0.903 | 0.020 | <.001 | 0.865 | 0.941 | 0.914 | 0.020 | <.001 | 0.875 | 0.953 |
| Mean DBP 6 am–2 pm | 0.806 | 0.028 | <.001 | 0.750 | 0.862 | 0.831 | 0.026 | <.001 | 0.780 | 0.883 |
Table 2 summarizes the area under the curve (AUC) of the various BP measurements in each time segment against 24‐hour BP controlled/normal. The segments with AUC >0.9 are the time segments for SBP during: 2 pm to 10 pm, 2 am to 8 am, and 6 am to 2 pm for both controlled and normal 24‐hour BP (Figure 1). The AUC for PP in all time segments was lower than 0.8.
Figure 1.
Receiver operating characteristics curve analysis of the predictive power of systolic blood pressure (SBP) in 3 time segments on blood pressure (BP) control. Time segment 2 pm to 10 pm. (A1) Prediction of control of 24‐hour SBP <135 mm Hg and diastolic BP (DBP) <85 mm Hg (confidence interval [CI], .875–947; P<.001). (A2) Prediction of control of 24‐hour SBP <130 mm Hg and DBP <80 mm Hg (CI, .893–958; P<.001). Time segment 6 am to 2 pm. (B1) Prediction of 24‐hour control of SBP <135 mm Hg and DBP <85 mm Hg (CI, .865–941; P<.001). (B2) Prediction of control of 24‐hour SBP <130 mm Hg and DBP<80 mm Hg (CI, .875–953; P<.001). Time segment 2 am to 8 am. (C1) Prediction of control of 24‐hour SBP <135 mm Hg and DBP <85 mm Hg (CI, .884–953; P<.001). (C2) Prediction of control of 24‐hour SBP <130 mm Hg and DBP <80 mm Hg (CI, .869–951; P<.001).
Time Segment With the Best Correlation to Normal Nighttime BP
Our next step was to find whether a short daytime segment (6 or 8 hours) had good correlation with normal nighttime ABPM (Table 3). The ROC curve analysis of SBP, DBP, and PP of the above‐mentioned time segments was performed against normal nighttime BP (SBP <120 mm Hg and DBP <70 mm Hg). The highest AUC was for SBP during the 8‐hour daytime segments 2 pm to 10 pm and 6 am to 2 pm (0.870 and 0.879, respectively). Two 6‐hour time segments that include both wake and sleep periods, 8 pm to 2 am and 2 am to 8 am, had high AUC for SBP (0.902 and 0.949, respectively).
Table 3.
Correlation Between Time Segment and Night Blood Pressure: Time Segment by SBP <120 mm Hg and Nighttime DBP <70 mm Hg: Area Under the Curve
| Test Result Variable | Area | Standard Error | Asymptotic Significance | Asymptotic 95% Confidence Interval | |
|---|---|---|---|---|---|
| Lower Bound | Upper Bound | ||||
| Mean SBP 8 am–2 am | 0.902 | 0.026 | <.001 | 0.852 | 0.953 |
| Mean DBP 8 pm–2 am | 0.852 | 0.027 | <.001 | 0.799 | 0.905 |
| Mean SBP 2 am–8 am | 0.949 | 0.016 | <.001 | 0.918 | 0.980 |
| Mean DBP 2 am–8 am | 0.884 | 0.024 | <.001 | 0.838 | 0.930 |
| Mean SBP 2 pm–10 pm | 0.870 | 0.026 | <.001 | 0.818 | 0.922 |
| Mean DBP 2 pm–10 pm | 0.791 | 0.032 | <.001 | 0.728 | 0.855 |
| Mean SBP 6 am–2 pm | 0.879 | 0.028 | <.001 | 0.824 | 0.934 |
| Mean DBP 6 am–2 pm | 0.806 | 0.032 | <.001 | 0.744 | 0.868 |
Abbreviations: DBP, diastolic blood pressure; SBP, systolic blood pressure.
Sensitivity and Specificity for SBP in Different Time Segments
The AUC in various time segments was superior for SBP compared with DBP (Table 4). The SBP in the best correlated segments (Tables 2 and 3) were chosen for calculating the cutoff point that has the optimal sensitivity and specificity. The cutoff SBP value was chosen according to the highest Youden's index calculated for the SBP in each time segment [Youden's index J statistic; Y = sensitivity – (1‐specificty)].
Table 4.
Cutoff SBP Measurements for Optimal Sensitivity and Specificity in Different Time Segments Against 24‐h ABPM and Nighttime ABPM
| Time Segment | Time Segment | AUC | SBP, mm Hg | Sensitivity | Specificity | PPV/NP | Youden's Index |
|---|---|---|---|---|---|---|---|
| 24‐h BP SBP <135 mm Hg and DBP <85 mm Hg | 2 am – 8 am | 0.918 | 133.5 | 0.752 | 0.904 | 0.88/0.78 | 0.656 |
| 2 pm – 10 pm | 0.911 | 138.5 | 0.803 | 0.878 | 0.87/0.81 | 0.681 | |
| 6 am – 2 pm | 0.903 | 140.5 | 0.778 | 0.887 | 0.87/0.79 | 0.665 | |
| 24‐h BP SBP <130 mm Hg and DBP <80 mm Hg | 2 am– 8 am | 0.910 | 124.5 | 0.892 | 0.824 | 0.92/0.78 | 0.716 |
| 2 pm – 10 pm | 0.926 | 133.5 | 0.854 | 0.824 | 0.91/0.73 | 0.678 | |
| 6 am – 2 pm | 0.914 | 132.5 | 0.905 | 0.811 | 0.91/0.80 | 0.716 | |
| Night BP SBP <120 mm Hg and DBP <70 mm Hg | 2 am – 8 am | 0.949 | 125.5 | 0.835 | 0.946 | 0.98/0.65 | 0.781 |
| 8 pm – 2 am | 0.902 | 128.5 | 0.807 | 0.893 | 0.96/0.59 | 0.700 | |
| 2 pm – 10 pm | 0.870 | 133.5 | 0.784 | 0.821 | 0.93/0.54 | 0.605 | |
| 6 am – 2 pm | 0.879 | 132.5 | 0.841 | 0.839 | 0.94/0.62 | 0.68 |
Abbreviations: ABPM, ambulatory blood pressure monitoring; AUC, area under the curve; DBP, diastolic blood pressure; PPV/NPV, positive predictive value/negative predictive value; SBP, systolic blood pressure.
Discussion
The aim of the current study was to evaluate whether partial ABPM can give accurate information about 24‐hour and nighttime BP control. We found that SBP in the time segment 2 pm to 10 pm has very good correlation with both controlled (SBP <135 mm Hg and DBP <85 mm Hg) and normal (SBP <130 mm Hg and DBP <80 mm Hg) 24‐hour ABPM, as reflected by AUCs of 0.911 and 0.926, respectively, with sensitivity and specificity of 80.3% and 87.8% for SBP 138.5 mm Hg and 85.4% and 82.4% for SBP 133.5 mm Hg, respectively. Normal nighttime BP (SBP <120 mm Hg and DBP <70 mm Hg) had a fair correlation with SBP in the time segment 2 pm to 10 pm (AUC=0.87 with sensitivity of 78.4% and specificity of 82.1% for SBP 133.5 mm Hg), and good correlation with SBP in the time segment 8 pm to 2 am (AUC = 0.902 with sensitivity of 80.7% and specificity of 89.3% for SBP 128.5 mm Hg).
ABPM measurements are characterized by multiple measurements during 24 hours while the patient is engaged in every day activity. This technique overcomes observer bias and the white‐coat effect.10 It may be said that ABPM is the most reliable estimate of a patient's BP. ABPM is also superior to office BP in the prediction and correlation of target organ damage.11 Despite its importance, many patients do not undergo ABPM because of daytime inconvenience and nighttime frequent waking.12 It was also recognized that special groups of patients, including obese patients and patients with impaired renal function, are associated with less complete ABPM session results.13 Because of this limitation, alternative validated multiple measurement techniques are sought. One study used the mean of multiple automated office BP measurements in order to recognize white‐coat hypertension.4 Some studies suggested that home BP measurements are comparable to ABPM in predicting target organ damage.14 The usefulness of abbreviated ABPM was demonstrated in 1982 by Weber and colleagues. This small study (performed in only 6 patients) demonstrated a good correlation between the mean of BP measurements during a 2‐hour period and the average of 24‐hour measurements.15 This early study looked at measurements during daytime only. A larger study looking at daytime BP measurements measured a 6‐hour daytime period with full awake mean BP.16 These two studies looked at daytime measurement only, while it is well acknowledged that nighttime BP has a correlation with target organ damage. Because of the importance of BP measurements, more studies looking at nighttime as well as daytime BP measurements have been performed. A study that examined the ABPM of 254 patients with normal or borderline office BP, found that 4‐hour mean BP readings from 10 am to 10 pm differed from daytime readings by less than 2 mm Hg, and 2‐hour mean readings from 3 am to 7 am differed from mean nighttime readings by less than 1 mm Hg.17 These results led the authors to the conclusion that 4‐hour daytime measurement together with 2‐hour nighttime measurement are sufficient for BP status evaluation.
Limitations
In the current study, a fixed clock was used; therefore, there was inaccuracy regarding daytime and nighttime BP definition. Another limitation is that the mean age of the study population was younger than 60 years. In older patients, the most appropriate time segment might be different from the segment in this study.
Conclusions
A complete ABPM of 24 hours provides full information about BP control including nighttime dipping and early morning surge. However, many patients face difficulties in completing a full ABPM. Due to the importance of ABPM as a diagnostic and prognostic tool, the clinician should not avoid performing ABPM studies in these patients. An abbreviated ABPM can be done. ABPM administered between 2 pm and 2 am can give reasonable information about the patient's BP control. For patients with more profound difficulties, such as sleeping problems with the ABPM device, an 8‐hour ABPM between 2 pm and 10 pm may also be acceptable. However, in order to confirm the diagnostic accuracy and the prognostic ability of short ABPM, more prospective studies are needed.
J Clin Hypertens (Greenwich). 2013;15:570–574. ©2013 Wiley Periodicals, Inc.23889719
References
- 1. Pickering TG, Shimbo D, Haas D. Ambulatory blood‐pressure monitoring. N Engl J Med. 2006;354:2368–2374. [DOI] [PubMed] [Google Scholar]
- 2. Conen D, Bamberg F. Noninvasive 24‐h ambulatory blood pressure and cardiovascular disease: a systematic review and meta‐analysis. J Hypertens. 2008;26:1290–1299. [DOI] [PubMed] [Google Scholar]
- 3. Viera AJ, Lingley K, Hinderliter AL. Tolerability of the Oscar 2 ambulatory blood pressure monitor among research participants: a cross‐sectional repeated measures study. BMC Med Res Methodol. 2011;11:59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Myers MG. A proposed algorithm for diagnosing hypertension using automated office blood pressure measurement. J Hypertens. 2010;28:703–708. [DOI] [PubMed] [Google Scholar]
- 5. Ernst ME, Sezate GS, Lin W, et al. Indication‐specific 6‐h systolic blood pressure thresholds can approximate 24‐h determination of blood pressure control. J Hum Hypertens. 2011;25:250–255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Ernst ME, Weber CA, Dawson JD, et al. How well does a shortened time interval characterize results of a full ambulatory blood pressure monitoring session? J Clin Hypertens (Greenwich). 2008;10:431–435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Mancia G, De Backer G, Dominiczak A, et al. 2007 Guidelines for the management of arterial hypertension: the task force for the management of arterial hypertension of the European Society of Hypertension (ESH) of the European Society of Cardiology (ESC). J Hypertens. 2007;25:1105–1187. [DOI] [PubMed] [Google Scholar]
- 8. Head GA, McGrath BP, Mihailidou AS, et al. Ambulatory blood pressure monitoring in Australia: 2011 consensus position statement. J Hypertens. 2012;30:253–266. [DOI] [PubMed] [Google Scholar]
- 9. Pickering TG, Hall JE, Appel LJ, et al. Recommendations for blood pressure measurement in humans and experimental animals: part 1: blood pressure measurement in humans: a statement for professionals from the subcommittee of professional and public education of the American Heart Association council on high blood pressure research. Hypertension. 2005;45:142–161. [DOI] [PubMed] [Google Scholar]
- 10. Staessen JA, Thijs L, Fagard R, et al. Predicting cardiovascular risk using conventional vs ambulatory blood pressure in older patients with systolic hypertension. Systolic Hypertension in Europe Trial Investigators. JAMA. 1999;282:539–546. [DOI] [PubMed] [Google Scholar]
- 11. Mancia G, Zanchetti A, Agabiti‐Rosei E, et al. Ambulatory blood pressure is superior to clinic blood pressure in predicting treatment‐induced regression of left ventricular hypertrophy. SAMPLE Study Group. Study on ambulatory monitoring of blood pressure and Lisinopril evaluation. Circulation. 1997;95:1464–1470. [DOI] [PubMed] [Google Scholar]
- 12. Degaute JP, van de Borne P, Kerkhofs M, et al. Does non‐invasive ambulatory blood pressure monitoring disturb sleep? J Hypertens. 1992;10:879–885. [PubMed] [Google Scholar]
- 13. Fravel MA, Ernst ME, Weber CA, et al. Influence of patient characteristics on success of ambulatory blood pressure monitoring. Pharmacotherapy. 2008;28:1341–1347. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Stergiou GS, Argyraki KK, Moyssakis I, et al. Home blood pressure is as reliable as ambulatory blood pressure in predicting target‐organ damage in hypertension. Am J Hypertens. 2007;20:616–621. [DOI] [PubMed] [Google Scholar]
- 15. Weber MA, Drayer JI, Wyle FA, Brewer DD. A representative value for whole‐day BP monitoring. JAMA. 1982;248:1626–1628. [PubMed] [Google Scholar]
- 16. Sheps SG, Bailey KR, Zachariah PK. Short‐term (six hour), ambulatory blood pressure monitoring. J Hum Hypertens. 1994;8:873–878. [PubMed] [Google Scholar]
- 17. Chanudet X, Chau NP, Larroque P. Short‐term representatives of daytime and night‐time ambulatory blood pressures. J Hypertens. 1992;10:595–600. [DOI] [PubMed] [Google Scholar]