Kristina Björklund MD, Section of Geriatrics, Department of Public Health and Caring Sciences, PO Box 609, SE-751 25, Uppsala, Sweden (fax: +46 18 6117976; e-mail: Kristina.Bjorklund@geriatrik.uu se).
Abstract. Björklund K, Lind L, Lithell H (University of Uppsala, Uppsala, Sweden). Twenty-four hour ambulatory blood pressure in a population of elderly men. J Intern Med 2000; 248: 503–512.
Objectives. The principal aim was to study ambulatory and office blood pressure in a population of elderly men. We also wanted to describe the prevalence of hypertension and investigate the blood pressure control in treated elderly hypertensives.
Design. A cross-sectional study of a population of elderly men, conducted between 1991 and 1995.
Subjects. Seventy-year-old men (n = 1060), participants of a cohort study that began in 1970.
Main outcome measures. Office and 24 h ambulatory blood pressure.
Results. Average 24 h blood pressure in the population was 133 ± 16/75 ± 8 mmHg, and daytime blood pressure 140 ± 16/80 ± 9 mmHg. Corresponding values in untreated subjects (n = 685) were 131 ± 16/74 ± 7 and 139 ± 16/79 ± 8, respectively. An office recording of 140/90 mmHg corresponded to an ambulatory pressure of 130/78 (24 h) and 137/83 mmHg (daytime) in untreated subjects. In subjects identified as normotensives according to office blood pressure (n = 270), the 95th percentiles of average 24 h and daytime blood pressures were 142/80 and 153/85 mmHg, respectively. The prevalence of hypertension, defined as office blood pressure ≥ 140/90 mmHg, was 66%. Despite treatment, treated hypertensives (n = 285) showed higher office (157/89 vs. 127/76 mmHg) and 24 h ambulatory (138/78 vs. 122/71 mmHg) pressures than normotensives (P < 0.05). Fourteen per cent of the treated hypertensives had an office blood pressure < 140/90 mmHg.
Conclusions. Our results provide a basis for 24 h ambulatory blood pressure reference values in elderly men. The study confirms previous findings of a high prevalence of hypertension at older age. It also indicates that blood pressure is inadequately controlled in elderly treated hypertensives.
Hypertension is a well known cardiovascular risk factor [1, 2]. However, blood pressure fluctuates over time, and it has been shown that these diurnal blood pressure variations are related to end-organ damage. A diminished nocturnal blood pressure (BP) decline has been associated with left ventricular hypertrophy , microalbuminuria , silent cerebrovascular damage [5, 6] and vascular dementia . Non-invasive ambulatory blood pressure monitoring (ABPM) during 24 h gives valuable information about diurnal BP changes and offers readings devoid of the white-coat effect sometimes seen in the clinic environment.
Furthermore, closer relationships have been found between hypertensive target organ damage and ambulatory BP (ABP) than conventional BP [8, 9], and prospective studies have indicated that 24 h ambulatory BP is an independent predictor of cardio- and cerebrovascular disease, including myocardial infarction and stroke [10, 11].
Although ABPM might have advantages over office BP recordings, difficulties in interpreting the results have limited the use of the method. Reference values for normal 24 h ABP have been proposed based on results from cross-sectional population studies [12–16] and a prospective study . However, both office and ambulatory BP, as well as BP variability, increase with age, and age differences need to be taken into consideration when evaluating BP in the elderly.
The prevalence of hypertension in the elderly is high, both combined systolic/diastolic  and isolated systolic hypertension [19, 20]. A reduced or absent BP decline during the night has been reported to be more common in old individuals , a circumstance that makes ABPM particularly valuable at a higher age. Significant benefits in terms of morbidity and mortality can be achieved in the elderly with adequate diagnosis and treatment of hypertension . However, a more widespread use of ABPM in the elderly requires thorough investigations of normal 24 h BP levels in this age group, as well as prospective studies in order to improve criteria for hypertension based on 24 h ABPM. The major aim of this investigation was therefore to create a basis for reference values by establishing characteristics of 24 h ABP, including nocturnal BP patterns, in a large population of elderly men. In addition, we wanted to describe the prevalence of hypertension and evaluate the BP control in this population.
The study population consisted of 1060 men, participants in a longitudinal population-based health survey regarding cardiovascular risk factors that was started in Uppsala, Sweden, in 1970. The population has been described by Byberg et al. . A total of 2322 men participated in the baseline examination at 50 years of age. Between 1991 and 1995, at an age of around 70, the men were invited to a re-investigation in which 1221 out of 1681 eligible subjects participated (72.6%). At this time the investigation included office and ambulatory BP measurements. Valid office and ambulatory recordings as well as information about antihypertensive medication were obtained in 1060 subjects, whose median age was 71 years (range 69.4–74.1). All participants gave written informed consent, and the study was approved by the local ethics committee.
Height was measured to the nearest whole cm, and body weight to the nearest 0.1 kg. The body mass index (BMI) was calculated as the ratio of the weight in kg to the square of the height in m2.
Blood pressure measurements
Office BP was measured in the right arm with a sphygmomanometer using the appropriate cuff size. The recordings were made to the nearest 2 mmHg twice after 10 min supine rest, and the mean of the two measurements was used for the analyses. Systolic and diastolic BPs were defined as Korotkoff’s phases I and V, respectively. The resting heart rate was measured as supine pulse rate during 1 min. Ambulatory BP was recorded using the Accutracker 2 equipment (Suntech Medical Instruments, Inc., Raleigh NC, USA). The device was attached to the subject’s non-dominant arm by a skilled lab technician, and BP was measured during 24 h starting at 11.00 h. BP recordings were made every 20 min between 06.00 and 23.00 h, and every 20 or 30 min between 23.00 and 06.00 h. Data were edited by omitting all readings of zero, all heart rate readings < 30, diastolic BP readings > 170 mmHg, systolic BP readings > 270 and < 80 mmHg, and all readings where the difference between systolic and diastolic BPs was less than 10 mmHg. Systolic and diastolic BPs were given by the auscultatory device. Mean arterial pressure (MAP) was calculated according to the formula MAP = DBP + 1/3 (SBP – DBP). Using short fixed time intervals, we defined daytime as 10.00 h to 20.00 h and night-time as midnight to 06.00 h in accordance with Staessen et al. . Nocturnal BP decline was assessed by the ratio between night and day BP. This ratio was multiplied by 100, expressing night-time BP as a percentage of the daytime level.
A self-administered questionnaire concerning medical background and ongoing treatment was administered under the supervision of a specially trained nurse. A positive answer to the question ‘Do you receive treatment for high blood pressure?’ and a statement of an antihypertensive drug as ongoing treatment were required for the subjects to be considered as treated hypertensives in this study. The duration of antihypertensive treatment was 8.5 ± 7.8 years (mean ± SD).
Hypertension status was defined according to the subject’s office BP recordings. An office BP ≥ 140 and/or ≥ 90 mmHg was considered hypertension in untreated individuals, meeting the Joint National Committee VI standards for hypertension .
Isolated systolic hypertension (ISHT) was defined as an office systolic BP ≥ 159 mmHg and a diastolic BP below 90 mmHg. In this study, the term ‘hypertensives’ will include combined systolic/diastolic as well as isolated systolic hypertensives unless stated otherwise.
We classified normotensives as subjects with an office BP < 140/90 mmHg and not taking antihypertensive medication. Ninety subjects were taking drugs with antihypertensive properties for reasons other than hypertension (post-MI, congestive heart failure). These subjects were not included in any of the groups in the following analyses.
Distributions were tested for normality by Shapiro-Wilk’s W-test. We used anova to calculate differences in means between subgroups, and Bonferroni’s correction in post-hoc analysis of more than two groups. Student’s paired t-test was used for paired comparisons, and two-tailed significance values with P < 0.05 were regarded as significant.
We calculated the correspondence between clinic and ambulatory BPs by linear regression analysis. The statistical analyses were performed using the statistical software packages JMP (SAS Institute, Inc., Cary, NC, USA) or Stata (Stata Corporation, Collage Station, USA).
Ambulatory and office BPs showed normal distribution. The prevalence of hypertension in the population according to office BP was 66%. Of the 1060 participants, 285 (27%) were treated hypertensives. Amongst the untreated subjects, 362 (34%) of all men showed combined systolic/diastolic hypertension. Isolated systolic hypertension was identified in 53 individuals (5%), whereas 270 men (25%) were considered normotensive (Fig. 1). Results from office and ambulatory BP recordings, including demographic data, are presented in Table 1, averaged for the total population as well as separately for treated and untreated subjects. The office BP level exceeded ambulatory pressures, 24 h and daytime levels, in all groups except for daytime systolic BP in normotensives. However, the absolute difference between office and daytime systolic pressure was smallest in normotensive subjects. Hypertensives did, by definition, present with a higher office BP than normotensives in the cohort. Despite antihypertensive treatment, treated hypertensives showed higher office BP than normotensives and untreated hypertensives. Moreover, BMI was significantly higher in treated hypertensives when compared with normotensives and untreated systolic/diastolic hypertensives.
Table 1. Demographic characteristics, office and 24 h ambulatory blood pressure and heart rate in individuals with and without antihypertensive treatment
BP < 140/90 mmHg
BP ≥140/90 mmHg
SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure. Pressures are in mmHg. Heart rate is in beats min–1.
Values are expressed as means (SD).
P < 0.05 vs. untreated normotensives;
P < 0.05 vs. untreated normotensives and untreated systolic/diastolic hypertensives;
P < 0.05 vs. untreated systolic/diastolic hypertensives.
Office hypertensives, treated and untreated, showed a higher 24 h ABP than normotensives and these differences remained significant when the 24 h period was divided into daytime and night-time periods (Table 1).
Furthermore, subjects classified as ISHT by office BP also showed the highest systolic 24 h, day and night pressures.
Blood pressure and heart rate from the ABPM were also calculated for each hour of the 24 h time period and are shown in Figs 2 and 3. In general, a peak value was seen around 10.00–11.00 h, the time when the subjects received the ambulatory device and the recordings started. The BP thereafter stabilized during the day and a reduction was generally seen during the night. This pattern was seen in the total population (Fig. 2), as well as in normotensives and treated and untreated hypertensives (Fig. 3). The 24 h BP level was similar for hypertensives, with very small differences between treated and untreated individuals (Fig. 3).
Influence of antihypertensive treatment
The treated hypertensives were divided into subgroups according to the number of antihypertensive drugs used. Table 2 shows the results from office and ambulatory BP and heart rate measurements in the different groups.
Table 2. Blood pressure and heart rate data in untreated subjects and treated hypertensives according to the number of antihypertensive drugs used. Values are expressed as means (SD)
Number of antihypertensive drugs
0 (n = 685)
1 (n = 155)
2 (n = 104)
≥3 (n = 26)
Systolic blood pressure (mmHg)
Night/day ratio (%)
Diastolic blood pressure (mmHg)
Night/day ratio (%)
Mean arterial pressure (mmHg)
Night/day ratio (%)
Heart rate (beats min–1)
Night/day ratio (%)
Neither office nor ambulatory BP differed between the three treatment groups, whereas the night/day ratio for systolic and diastolic BP tended to be higher in the groups with two or more antihypertensive drugs. However, these differences were not significant.
The 95th percentile of the ambulatory BP distribution in untreated subjects (n = 685) was 162/88 mmHg for 24 h pressure, and 169/94 for daytime BP. In subjects identified as normotensive according to their office BP, the 95th percentiles of 24 h blood pressure were 142/80 mmHg and the corresponding daytime and night-time levels were 153/85 and 132/73 mmHg, respectively. Using the 95% level of 24 h BP in normotensives as an upper limit for normal 24 h BP, 59% of the subjects identified as hypertensives according to their office BP had a 24 h ABP below this level (Table 3).
Table 3. Comparison of ambulatory blood pressure (ABP) levels regarding ability to predict office hypertension in untreated subjects (n = 685). Proposed upper limits of normal for ABP are the 95th percentile of 24 h ABP in office normotensives (142/80 mmHg) and 24 h ABP corresponding to an office BP of 140/90 mmHg (130/78). Office hypertension includes combined systolic/diastolic and isolated systolic hypertension
Office hypertension [n (%)]
Office normotension [n (%)]
24 h ABP ≥ 142/80
24 h ABP < 142/80
24 h ABP ≥ 130/78
24 h ABP < 130/78
Total [n (%)]
Office and ambulatory BPs were correlated (Fig. 4). In a regression analysis of untreated subjects, a 24 h BP of 130/78 mmHg corresponded with the upper limit of normal for clinic BP, 140/90 mmHg. Equivalent levels for daytime and night-time BP were 137/83 and 116/70 mmHg, respectively. Table 3 illustrates the magnitude of agreement between office BP and ABP in a classification of untreated subjects as hypertensive using different upper limits of 24 h ABP based on this population study.
Our findings are based on a large cohort of elderly men derived from the general population and living in a well-defined geographical area. Several previous population studies of 24 h BP have comprised subjects of different age, and both men and women, resulting in a proportionately small number of individuals in each age and gender stratum. In the Allied Irish Bank study  as well as in the PAMELA study  an increased ABP was found with rising age, in both men and women. An increased pulse pressure is commonly seen in the elderly as a result of increased arterial stiffness causing the systolic BP to rise whilst the diastolic BP plateaus . Imai et al.  found a decline in diastolic BP in Japanese males above 60 years of age, and a similar tendency was also noted in a Danish population .
Besides a longitudinal study in Japan where Ohkubo et al.  investigated the prognostic significance of 24 h ABP, current reference values for 24 h BP reflect the ABP corresponding to the upper limit of normal for office BP or the distribution of normal 24 h ABP in a population. Data from large population studies have been relatively consistent regarding 24 h BP levels in normotensive subjects [12, 13, 15, 26] and by using the 95th percentile of ABP in individuals classified as normotensives according to office BP, a working definition of 130/80 mmHg as an upper limit of normal for 24 h pressure has been proposed [27, 28]. This level is well in accordance with 134/79 mmHg, the cut-off value for 24 h ABP that best predicted cardiovascular mortality in the Japanese prospective study.
In the present cohort, the 95th percentile of the 24 h ABP distribution in office normotensives, 142/80 mmHg, showed systolic levels higher than in comparable population studies [12, 14, 15, 26], which was probably a consequence of our population being older. Some previous investigations of ABP in elderly men have, in separate analyses, included all untreated subjects in a population. We found that mean 24 h ambulatory levels in untreated hypertensives and normotensives (n = 685) in the present study, 131/74 mmHg, corresponded well with results of other studies – 123/70 , 123/74 , 126/79  and 135/74 mmHg – particularly regarding the diastolic pressure. Interestingly, mean BP values in these elderly men most closely resembled those from a Danish population study , a finding which may indicate an impact of cultural  as well as climatic factors.
The 24 h ABP that corresponded to an office BP of 140/90 was 130/78 mmHg in untreated subjects of this population, and whilst this constituted a considerably higher systolic pressure than corresponding results from the PAMELA study of the elderly (120/76 mmHg) , BP levels were again more similar to those of elderly men in the Danish MONI 10 study, i.e. 134/78 mmHg.
When we compared ABP levels regarding ability to predict office hypertension, the total number of correctly classified hypertensives was higher when the upper limit of ABP was derived from correspondence criteria (72%) than when using the 95th percentile of ABP in office normotensives (61%). However, these percentages are probably both underestimated since the office BP, being the mean of two BPs measured at a single occasion, was likely to have been subjected to a measurement error derived from within-subject variation as well as an independent measurement error. According to international guidelines, a diagnosis of hypertension should be based on two or more readings at repeated visits, and it is known that the BP in clinical studies tends to be higher at the first visit compared with later visits. Furthermore, a presence of white-coat hypertension will add to the disparities seen between ambulatory and clinic values. For these reasons, the prevalence of office hypertension might be somewhat overrated in this population. Given this uncertainty regarding ‘true blood pressure’ introduced by a measurement error, an underestimation of the slope of the regression line between ambulatory and office pressure can be expected, but since the cut-off value for office hypertension (140/90) is close to the mean office BP (143/82) in untreated subjects of this population, the measurement error will have little impact on the corresponding predicted ambulatory pressure value. Thus, the ambulatory BP obtained by the correspondence criteria seems to provide an acceptable measure of the upper limit of normal ABP in this population.
The study population originated in 1970 when all 50-year-old men in Uppsala county were invited to take part in a health survey. Although the remaining participants at age 70 may represent a selection of healthy subjects, we found a hypertension prevalence of 66%, similar to what has been shown in NHANES  and other health surveys [32, 33] of the elderly. A limitation of this study is the lack of data on women, and further studies of large populations of elderly women would be desirable in order to find out if this group should be treated differently from men.
The impact of hypertension as a risk factor for cardiovascular disease is a thoroughly investigated matter, and there is a substantial amount of evidence indicating that adequate control of BP reduces cardiovascular morbidity and mortality in hypertensives [34–36]. Nevertheless, several studies have reported that hypertension generally is poorly managed [37, 38]. In a study on treated hypertensive men, Berlowitz et al. found that less than 25% of the men had an office BP < 140/90 mmHg, the JNC VI criteria for hypertension.
In this population of elderly men, only 14% of the treated hypertensives showed an office BP < 140/90 mmHg, the current target BP for elderly , and 51% had an office diastolic BP < 90 mmHg. According to treatment guidelines for the elderly in Sweden at the time the study was carried out (< 160/90) , 39% of the treated subjects fulfilled the treatment goals.
Twenty-eight per cent displayed a 24 h ABP < 130/78 mmHg, a proposed upper limit of normality in this study, whereas 55% showed an isolated 24 h diastolic ABP < 78 mmHg.
Hence, systolic BP control, in particular, was inadequately achieved amongst these elderly hypertensives, and although ABP monitoring revealed somewhat more satisfactory treatment results, it was evident that only a third of the treated hypertensive population were properly controlled. These figures should, however, be interpreted with caution, as the study was not designed to assess the management of hypertension.
Various methods of analysis of BP changes between sleep and awake periods have been proposed. Besides the ‘diary’ method, where sleeping time is assessed by the individual, a short fixed time method gives a good estimate of the true sleeping period . The nocturnal BP decline can be defined in different ways. In a comparison of the ratio between night and day BP with the nocturnal BP fall, Staessen et al.  found that the ratio is less dependent on the BP level and suggested the use of night/day ratio as the preferred method. Further details on night-time BP patterns in this cohort will be covered in a separate paper.
We conclude that hypertension in this elderly population was common, and office as well as ambulatory BP control inadequate in treated elderly hypertensives, despite the well-known risk associated with a high BP. We further suggest the descriptive data of 24 h ABP in this large population of elderly men as a basis for future reference values. When guidelines are established based on appropriate reference values for 24 h BP, ABPM may be a useful tool for diagnosing hypertension and evaluating therapy in the elderly.
This study was supported by grants from the Medical Faculty at Uppsala University, the Swedish Medical Research Council (MFR, grant no. 5446), the Ernfors Fund, the Foundation for Geriatric Research, the Swedish National Association against Heart and Lung Disease, the Uppsala Geriatric Fund, the Swedish Council for Planning and Co-ordination of Research and the Thuréus Fund for Geriatric Research.