Hypertension is a common condition with increasing age [1,2], but, so far, data on the association between a single blood pressure measurement and mortality in very old age are surprisingly conflicting [3–13]. In a large meta-analysis of observational data from 61 studies , both high SBP and DBP were independently predictive of cardiovascular mortality in all age groups. However, this association attenuated considerably with age. In contrast, several observational studies in exclusively very old populations have shown that low, and not high blood pressure, is associated with increased mortality in very old age [4–13]. Following a steady rise during earlier life, SBP decreases in very old age [5,14–17]. There is limited information about the clinical significance of this decreasing SBP in the individual over time. Decreasing trend in SBP in very old age has been associated with an increased mortality risk, but data from these studies are not consistent [5,14,15,17–19].
In view of these conflicting data on the associations of different blood pressure values and trends over time with mortality, we investigated the association between SBP trends and mortality in a cohort of nonagenarians from the Leiden 85-plus Study. The present study evaluates the independent contributions of both current SBP value at age 90 and the trend in preceding SBP measurements to the prediction of mortality in nonagenarians.
PARTICIPANTS AND METHODS
The Leiden 85-plus Study is a prospective population-based study among 85-year-old inhabitants of the city of Leiden (the Netherlands). Study design and cohort characteristics have been described in detail [20,21]. In brief, between September 1997 and September 1999, 705 people in the 1912–1914 birth cohort reached the age of 85 years and were eligible to participate. No exclusion criteria were used. Fourteen people died before enrollment; a total of 599 (87%) people gave informed consent and were enrolled. At baseline and yearly up to the age of 90 participants were visited at their place of residence. Figure 1 presents the flowchart of the present study, approved by the Medical Ethics Commission of the Leiden University Medical Center.
In this study, the cohort of interest consists of all participants who were alive at age 90 and for whom blood pressure was measured at age 90 years (n = 271).
Blood pressure and blood pressure categories
Each year up to age 90 years blood pressure was measured twice with an interval of 2 weeks. Blood pressure was measured using a mercury sphygmomanometer, in seating position, after at least 5 min of rest without having performed vigorous exercise during the preceding 30 min. The systolic value was measured at the onset of phase 1 of the Korotkoff sounds. For the analysis, the mean of the two systolic values for each year was used.
As there was no evidence that the relation between SBP and mortality was U-shaped, the SBP values at age 90 years were dichotomized into categories, ‘low’ and ‘high’ SBP. Low SBP was defined as a SBP of 150 mmHg or less and high as a SBP more than 150 mmHg. This cut-off was based on the target value used in studies assessing the risks and benefits of treating hypertension in individuals aged at least 80 years [22,23].
Blood pressure trends
A linear regression of SBP against time was used to assess the individual regression coefficient (β) of the blood pressure trend over the 5 preceding years (90–85 years) for each participant. Participants were divided into tertiles based on the beta from their individual model to describe the trend in SBP. The lowest tertile was considered as a ‘decreasing trend in SBP’; the highest tertile was considered as an ‘increasing trend in SBP’; and the average tertile was considered as an ‘average trend in SBP’.
All participants were followed for mortality for 5 years from age 90 years onward. Dates of death were obtained from municipality records.
Sociodemographic, functional, and clinical characteristics
At age 90 years, participants were visited at their place of residence for face-to-face interviews, to perform several tests [24–27] and to collect information on sociodemographic characteristics. Each participant's general practitioner (or, if applicable, nursing-home physician) was interviewed about the participant's medical history, using standardized questionnaires, including questions on present and past cardiovascular  (including presence of a medical history of myocardial infarction, angina pectoris, arrhythmias, or heart failure, or an ECG at age 90 revealing a prior myocardial infarction, atrial fibrillation or left ventricular hypertrophy, history of stroke, and peripheral arterial disease) and noncardiovascular morbidities (including history of chronic obstructive pulmonary disease, malignancies, arthritis, Parkinson's disease, and hip fracture). Pharmacists provided detailed information on all medication used by participants.
Baseline summary characteristics at age 90 years are reported as medians with interquartile ranges (IQRs) or means with standard deviation (SD) for continuous variables and as numbers with percentage (%) for categorical variables. A linear regression of SBP against time was used to assess the individual regression coefficient (β) of the blood pressure trend over the 5 preceding years (90–85 years) for each participant. Differences in change in SBP between men and women were compared using the Mann–Whitney U-test. Time-to-event curves were constructed with the Kaplan–Meier method. Hazard ratios and corresponding 95% confidence intervals (CIs) were calculated in Cox proportional-hazard models, including the blood pressure category (low or high) or trend in SBP (average, decreasing, increasing) as a categorical variable. The independent contributions of both SBP value at age 90 and trend in preceding SBP measurements to the prediction of mortality were calculated in Cox proportional-hazard models, including both blood pressure category (low or high) and category of trend in SBP (average, decreasing, increasing). In the sex-adjusted model, all hazard ratios were adjusted for sex only, because the present study focused on prediction  and not cause. In the additional models, the hazard ratios were subsequently adjusted for baseline characteristics from Table 1. The fully adjusted model included adjustment for sex, place of residence (institutionalized or not institutionalized), level of education, income, Activities of Daily Living Score, history of noncardiovascular disease, Mini-Mental State Examination Score, Geriatric Depression Scale Score, Instrumental Activities of Daily Living Score, smoking status, history of diabetes, history of hypertension, use of antihypertensive drugs, total plasma cholesterol, history of cardiac or other vascular morbidity, total cholesterol, hemoglobin and body weight, all at age 90 years, and SBP at age 85 and 90 years.
As sensitivity analysis, stratified analyses were performed for sex, place of residence (institutionalized or not institutionalized), use of antihypertensive drugs, history of cardiovascular disease, and history of noncardiovascular disease, all at age 90 years. To evaluate the influence of the arbitrary choice to use the blood pressure trend over the 5 preceding years, the analyses with blood pressure trend were repeated over 1 (90–89 years), 2 (90–88 years), 3 (90–87 years), and 4 years (90–86 years), respectively.
Data analysis was performed using SPSS 17.0 for Windows (SPSS Inc., Chicago, Illinois, USA).
Of the initial cohort of 599 participants (87% response) at age 85 years, 304 individuals survived up to the age of 90 years. Blood pressure measurement at age 90 was missing for 33 (11%) participants. Therefore, the current analyses included 271 participants. Table 1 presents the baseline characteristics. At age 90 years, less than 30% of the participants were men, and almost 40% of the participants were institutionalized. Median Mini-Mental State Examination Score was 25 points. More than 60% of the study population had a history of cardiac morbidity, the majority had a history of hypertension (56%), and almost half of the population used antihypertensive drugs.
Blood pressure at age 90 and SBP trends between age 85 and 90 years
At age 90 years, median SBP was 153 mmHg. The number of participants with a low SBP (<150 mmHg) was 116 (42.8%); 155 (57.2%) had a high blood pressure.
Overall, the average change in SBP was −1.5 (SD 3.4) mmHg per year (−2.2 mmHg per year in men and −1.3 mmHg per year in women; P = 0.060), resulting in a mean decrease in SBP per participant of 7.3 (SD 29.2) mmHg over the 5 years between the ages of 85 and 90 years (range −10.8 mmHg to +7.5 mmHg per year). The tertile limits (33 and 67%) of the change in SBP were −2.9 and −0.09 mmHg/year.
SBP trends between age 85 and 90 years related to mortality after 90
From age 90 years onward, 187 (69%) of the 271 participants died during a median follow-up of 3.6 (range 0.04–5) years. Figure 2 presents the Kaplan–Meier analysis of time to all-cause mortality depending on the 5-year trend in SBP for the participants who survived up to the age of 90 years. It shows that a decreasing 5-year trend (≥2.9 mmHg/year) was associated with an increased cumulative all-cause mortality compared to an increasing or average 5-year trend. Participants with a decreasing 5-year trend in SBP had a 1.5-fold increased all-cause mortality risk compared to those with an average SBP trend (hazard ratio 1.51, 95% CI 1.07–2.15). When excluding participants who died within 3 months of follow-up (to avoid this increased risk being a reflection of terminally ill patients only), this increased mortality risk remained unchanged (hazard ratio 1.53, 95% CI 1.07–2.18). To evaluate the independent contribution of the trend in the SBP measurements between age 85 and 90 years, we analyzed mortality risk of the trends in SBP over 5 years, adjusted for SBP at age 90. Participants with a decreasing 5-year trend in SBP had an increased all-cause mortality risk (hazard ratio 1.45, 95% CI 1.02–2.06) compared to those with an average SBP trend, adjusted for SBP at age 90. There was no difference in mortality risk between participants with an average SBP trend and those with an increasing 5-year trend in SBP (hazard ratio 1.04, 95% CI 0.72–1.50; hazard ratio 1.18, 95% CI 0.81–1.71 after adjustment for SBP at age 90). After analysis of the data with a fully adjusted model, including all baseline characteristics mentioned in Table 1, the estimate for the decreasing 5-year trend in SBP versus average SBP trend was 1.08 (95% CI 0.60–1.86) and 0.79 (95% CI 0.46–1.37) for increasing 5-year trend in SBP versus average SBP trend.
Low versus high SBP at age 90 years
Participants with a low SBP at age 90 had an increased all-cause mortality risk compared to those with a high SBP (hazard ratio 1.57, 95% CI 1.18–2.09). To evaluate the independent contribution of SBP at age 90 years, we analyzed mortality risk of low versus high SBP at age 90 years adjusted for SBP trends. Participants with a low SBP at age 90 had a 1.6-fold increased all-cause mortality risk compared to those with a high SBP, adjusted for SBP trends (hazard ratio 1.62, 95% CI 1.21–2.20). After analysis of the data with a fully adjusted model, including all baseline characteristics mentioned in Table 1 and SBP trends, participants with a low SBP at age 90 had a 1.5-fold increased all-cause mortality risk compared to those with a high SBP at age 90 years (hazard ratio 1.47, 95% CI 0.90–2.40).
SBP trends stratified by low and high SBP at age 90
To investigate the effect of the trend in SBP in conjunction with SBP at age 90, we analyzed mortality risk of the trends in SBP over 5 years dependent on SBP at age 90 years. Participants with a high SBP at age 90 and an average 5-year trend were considered the reference group. Participants with a low SBP at age 90 in combination with a decreasing trend in SBP in the preceding 5 years had a more than doubled all-cause mortality risk compared to the reference group (sex-adjusted model hazard ratio 2.39, 95% CI 1.48–3.88, fully adjusted model hazard ratio 1.62, 95% CI 0.74–3.52; Table 2). This was most prominent in participants with an already low SBP at age 85 years, at the onset of the decrease in SBP (Appendix Table a1, http://links.lww.com/HJH/A205).
As sensitivity analyses for the association between a decreasing 5-year trend in SBP and mortality, the analyses were stratified for sex, place of residence, use of antihypertensive drugs, history of cardiovascular disease, and history of noncardiovascular disease. Overall, decreasing 5-year trend in SBP predicted an increased mortality risk in the different subgroups (Table 3). The effect was stronger in women than in men (hazard ratio 1.67, 95% CI 1.08–2.58 and hazard ratio 1.18, 95% CI 0.66–2.11, respectively), although there was no significant interaction between the mortality risk associated with sex and the risk associated with the trend in SBP between the ages of 85 and 90 years (P = 0.12). The effect was stronger in participants who were institutionalized than in those living independently (hazard ratio 1.87, 95% CI 1.10–3.19 and hazard ratio 1.30, 95% CI 0.81–2.09). The interaction between mortality risk associated with place of residence and the risk associated with the trend in SBP between the ages of 85 and 90 years was significant (P = 0.01). The average change in SBP was −1.1 mmHg for independently living participants and −2.3 mmHg in those who were institutionalized (Mann–Whitney U-test, P = 0.01). Although the average change in SBP was different between participants without a history of cardiovascular disease and those with a history of cardiovascular disease (−0.9 and 1.8 mmHg, respectively; P = 0.04), there was no difference in association between a decreasing 5-year trend in SBP and mortality. The same sensitivity analysis was performed for the association between low SBP and mortality. Low SBP at age 90 predicted an increased mortality risk in all different subgroups (Table 3).
As an additional sensitivity analyses, we analyzed mortality risks associated with trends in SBP over 1–4 years (average, decreasing, increasing). A decreasing trend in SBP in the preceding years was still associated with an increased all-cause mortality risk when the trend was assessed over 4 years, but not when they were assessed over 1, 2, or 3 years.
In this prospective, population-based study in nonagenarians, we report a strong association between decreasing 5-year trend in SBP and an increased all-cause mortality risk. This observation was more pronounced in institutionalized participants. Additionally, we replicated the earlier observed association between low SBP at age 90 and an increased all-cause mortality risk. Both the 5-year trend in SBP measurements and the SBP value at age 90 contribute independently to the prediction of mortality. This applied to those with and without antihypertensive drugs at age 90 and in those with and without history of cardiovascular disease or noncardiovascular disease. The most striking observation is that a decreasing trend in SBP in the preceding 5 years in those participants with a low SBP at age 90 predicted a more than doubled mortality risk compared to participants with an average 5-year trend in SBP and high SBP at age 90. Therefore, repeated blood pressure measurement may contribute to identification of older people at risk. This implies looking at the global trend in SBP measurements over previous years when interpreting a single blood pressure measurement in very old persons in daily practice, and weighing both the absolute current value and the existing trend in SBP when making treatment or withdrawal decisions. The full potential clinical relevance of these findings may become clearer when biological explanations for the observed association have been studied.
The results of previous studies assessing the association between decreasing blood pressure and increased mortality risk in old age are not consistent [5,14,15,17–19]. In the present study, decreasing 4-year and 5-year trend in SBP were associated with an increased mortality risk, whereas, in contrast, there was no significant difference in all-cause mortality risk associated with shorter term trends (up to 3 years) in SBP. It has been suggested that declining blood pressure is associated with declining health . In the present study, the association between decreasing trend in SBP and increased mortality risk was stronger in institutionalized participants, yet this association was equally present in participants with and without history of cardiovascular disease or noncardiovascular disease. The estimated mortality risk associated with the trend in SBP did not change after adjustment for SBP at age 90, sex, place of residence (institutionalized or not institutionalized), level of education, income, Activities of Daily Living Score, and history of noncardiovascular disease (data not shown). After additional adjustments for Mini-Mental State Examination Score, Geriatric Depression Scale Score, Instrumental Activities of Daily Living Score, smoking status, history of diabetes, history of hypertension, use of antihypertensive drugs, total plasma cholesterol, body weight, hemoglobin, SBP at age 85 and 90 years, and history of cardiac or other vascular morbidity, the estimates approached unity. Although this could indicate that declining cardiovascular and cognitive health are true confounders, it is more likely that the adjustments for cardiovascular, cerebral, and cognitive characteristics were made within the causal pathway. Therefore, no interpretation regarding cause or causality was made in this study. This warrants additional studies on the underlying mechanisms. A new finding of the present study is that in nonagenarians, both the trend in preceding SBP measurements and SBP value at age 90 independently contribute to the mortality risk; decreasing trend in SBP in the preceding 5 years with a low SBP at age 90 predicted the highest mortality risk. The observed association between low SBP and an increased mortality risk is in contrast with a large meta-analysis of observational data from 61 studies ; however, it is in line with earlier studies in exclusively very old populations [4–13,15]. Some suggested that this association is caused by poor health status [4,5], whereas in the present study, the association between low SBP and increased mortality risk was equally present in participants with and without history of cardiovascular disease or noncardiovascular disease, as reported in other studies [8–12,15]. After adjustment for the baseline characteristics in the fully adjusted model, the estimated mortality risk associated with low SBP remained similar (initial model hazard ratio 1.62, 95% CI 1.21–2.20 and fully adjusted model hazard ratio 1.47, 95% CI 0.90–2.40). The observed association between SBP at age 90 years and mortality in the present study is linear until it reaches a plateau from SBP 160–170 mmHg onward. Therefore, it seems that ‘the lower the SBP, the worse’ rather than ‘the higher the blood pressure, the better’. In addition to the association between a decreasing trend in preceding SBP measurements and increased mortality, there might also be a trend for increased mortality with an increasing trend in SBP, especially when SBP at age 85 years is low (appendix Table a1, http://links.lww.com/HJH/A205). This could indicate that variability in blood pressure is also associated with increased mortality. Previous studies showed that higher variability in blood pressure is associated with an increased risk of cardiovascular events and mortality in middle-aged individuals [30–32]. Relevant data for older individuals on blood pressure variability are scarce; whether blood pressure variability is associated with increased mortality in very old age is, therefore, subject for further research.
A strength of the present study is that we used a population-based sample of nonagenarians. There were no exclusion criteria for participation in the Leiden 85-plus Study: 87% of the general population aged 85 years living in Leiden participated and there was a complete follow-up for mortality. Furthermore, a blood pressure pattern was established over 5 years, with annual per-protocol measurements within the study. Also, use of the mean SBP values of two separate measurements annually, reduced regression to the mean. The present study focuses on prediction in clinical practice; SBP, being the most used blood pressure value in clinical practice, is therefore, the relevant blood pressure value to investigate. The choice of a cut-off value used for SBP at age 90 and the use of tertiles to describe the trends in SBP can be considered arbitrary. However, as there are no clinical cut-offs for these values and trends in very old age, it seems sensible to adhere to the target value used in studies assessing the risks/benefits of treating hypertension in individuals aged at least 80 years [22,23]. Furthermore, there was no evidence for a U-shaped relationship between mortality and SBP at age 90 years, indicating that dichotomization of the blood pressure range can be considered correct. In addition, our relatively small number of participants reduced the statistical power to detect differences in mortality risk in subgroup analyses, creating the possibility of type-II errors. Due to a lack of power, we were unable to analyze the effects of declining SBP and low SBP on cause-specific mortality. Finally, as our study population included only strong survivors, thus a selected study population, we cannot extrapolate our conclusions to younger age groups.
In conclusion, at age 90 years, both the trend in SBP over the preceding 5 years and current SBP value independently contribute to all-cause mortality risk; the combination of a decreasing trend in SBP in the preceding 5 years and low SBP at age 90 predicts the highest mortality risk. Therefore, in very old age, all available preceding SBP measurements contribute to a more accurate prediction of mortality and should be taken into account. In addition, repeated blood pressure measurement may contribute to identification of older people at risk.
R.K.E.P. contributed to the analysis and interpretation of the data, drafting of the article, critical revision of the article, and statistical analysis.
W.deR., S.P.M., W.J.J.A., and J.W.B. contributed to the analysis and interpretation of the data, drafting of the article, and critical revision of the article.
A.J.M.deC., R.G.J.W., and J.G. contributed to the study concept and design, acquisition of data, analysis and interpretation of the data, drafting of the article, and critical revision of the article.
J.G. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
All authors gave their final approval for publication.
The Leiden 85-plus Study was partly funded by an unrestricted grant from the Dutch Ministry of Health, Welfare and Sports (1997–2001). R.K.E.P. received an unrestricted grant from the SBOH, employer of general practitioner trainees, a foundation funded by the Dutch Ministry of Health, Welfare and Sports (2010–2013). All researchers worked independently of the funders.
All funding sources were independent and had no influence on the study design; the collection, analyses, and interpretation of our data; the writing of this report; or the decision to submit the article for publication.
Conflicts of interest
There are no conflicts of interest.
Reviewer's Summary Evaluation Reviewer 1
What is new: A higher mortality in very old subjects is observed with low BP levels especially in those with a decreasing trend in SBP between 85 and 90 years. Interestingly, this association was stronger in institutionalized individuals i.e. the most frail elderly subjects.
Clinical relevance: These data can be explained by the fact that in very old frail subjects, a decrease in BP is often the result of co-morbidities and other conditions associated with poor prognosis.
Conclusions: In these subjects, BP evolution over time can provide important information and BP values should always be integrated within the individual's general health status.
1. de Craen AJ, Gussekloo J, Teng YK, Macfarlane PW, Westendorp RG. Prevalence of five common clinical abnormalities in very elderly people: population based cross sectional study. BMJ
2. Aguado A, Lopez F, Miravet S, Oriol P, Fuentes MI, Henares B, et al. Hypertension in the very old; prevalence, awareness, treatment and control: a cross-sectional population-based study in a Spanish municipality. BMC Geriatr
3. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R. Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies. Lancet
4. Boshuizen HC, Izaks GJ, van Buuren S, Ligthart GJ. Blood pressure and mortality in elderly people aged 85 and older: community based study. BMJ
5. Heikinheimo RJ, Haavisto MV, Kaarela RH, Kanto AJ, Koivunen MJ, Rajala SA. Blood pressure in the very old. J Hypertens
6. Langer RD, Ganiats TG, Barrett-Connor E. Paradoxical survival of elderly men with high blood pressure. BMJ
7. Mattila K, Haavisto M, Rajala S, Heikinheimo R. Blood pressure and five year survival in the very old. Br Med J (Clin Res Ed)
8. Molander L, Lovheim H, Norman T, Nordstrom P, Gustafson Y. Lower systolic blood pressure is associated with greater mortality in people aged 85 and older. J Am Geriatr Soc
9. Oates DJ, Berlowitz DR, Glickman ME, Silliman RA, Borzecki AM. Blood pressure and survival in the oldest old. J Am Geriatr Soc
10. Rastas S, Pirttila T, Viramo P, Verkkoniemi A, Halonen P, Juva K, et al. Association between blood pressure and survival over 9 years in a general population aged 85 and older. J Am Geriatr Soc
11. Satish S, Freeman DH Jr, Ray L, Goodwin JS. The relationship between blood pressure and mortality in the oldest old. J Am Geriatr Soc
12. van Bemmel T, Gussekloo J, Westendorp RG, Blauw GJ. In a population-based prospective study, no association between high blood pressure and mortality after age 85 years. J Hypertens
13. van Hateren KJ, Landman GW, Kleefstra N, Groenier KH, Kamper AM, Houweling ST, Bilo HJ. Lower blood pressure associated with higher mortality in elderly diabetic patients (ZODIAC-12). Age Ageing
14. Lernfelt B, Svanborg A. Change in blood pressure in the age interval 70-90. Late blood pressure peak related to longer survival. Blood Press
15. Hakala SM, Tilvis RS, Strandberg TE. Blood pressure and mortality in an older population. A 5-year follow-up of the Helsinki Ageing Study. Eur Heart J
16. Landahl S, Bengtsson C, Sigurdsson JA, Svanborg A, Svardsudd K. Age-related changes in blood pressure. Hypertension
17. Tervahauta M, Pekkanen J, Enlund H, Nissinen A. Change in blood pressure and 5-year risk of coronary heart disease among elderly men: the Finnish cohorts of the Seven Countries Study. J Hypertens
18. Langer RD, Criqui MH, Barrett-Connor EL, Klauber MR, Ganiats TG. Blood pressure change and survival after age 75. Hypertension
19. Satish S, Zhang DD, Goodwin JS. Clinical significance of falling blood pressure among older adults. J Clin Epidemiol
20. van Exel E, Gussekloo J, Houx P, de Craen AJ, Macfarlane PW, Bootsma-van der Wiel A, et al. Atherosclerosis and cognitive impairment are linked in the elderly. The Leiden 85-plus study. Atherosclerosis
21. Bootsma-van der Wiel A, van Exel E, de Craen AJ, Gussekloo J, Lagaay AM, Knook DL, et al. A high response is not essential to prevent selection bias: results from the Leiden 85-plus study. J Clin Epidemiol
22. Beckett NS, Peters R, Fletcher AE, Staessen JA, Liu L, Dumitrascu D, et al. Treatment of hypertension in patients 80 years of age or older. N Engl J Med
23. Bulpitt CJ, Beckett NS, Cooke J, Dumitrascu DL, Gil-Extremera B, Nachev C, et al. Results of the pilot study for the Hypertension in the Very Elderly Trial. J Hypertens
24. Tombaugh TN, McIntyre NJ. The Mini-Mental State Examination: a comprehensive review. J Am Geriatr Soc
25. Kempen GI, Miedema I, Ormel J, Molenaar W. The assessment of disability with the Groningen Activity Restriction Scale. Conceptual framework and psychometric properties. Soc Sci Med
26. de Craen AJ, Heeren TJ, Gussekloo J. Accuracy of the 15-Item Geriatric Depression Scale (GDS-15) in a community sample of the oldest old. Int J Geriatr Psychiatry
27. Cantril H. The pattern of human concerns.
New Brunswick (NJ): Rutgers University Press; 1965.
28. de Ruijter W, Westendorp RG, Macfarlane PW, Jukema JW, Assendelft WJ, Gussekloo J. The routine electrocardiogram for cardiovascular risk stratification in old age: the Leiden 85-plus study. J Am Geriatr Soc
29. Grobbee DE, Hoes AW. Clinical epidemiology. Principles, methods, and applications for clinical research
. Boston: Jones and Bartlett Publishers; 2009.
30. Muntner P, Shimbo D, Tonelli M, Reynolds K, Arnett DK, Oparil S. The relationship between visit-to-visit variability in systolic blood pressure and all-cause mortality in the general population: findings from NHANES III, 1988 to 1994. Hypertension
31. Pringle E, Phillips C, Thijs L, Davidson C, Staessen JA, de Leeuw PW, et al. Systolic blood pressure variability as a risk factor for stroke and cardiovascular mortality in the elderly hypertensive population. J Hypertens
32. Rothwell PM, Howard SC, Dolan E, O’Brien E, Dobson JE, Dahlof B, et al. Prognostic significance of visit-to-visit variability, maximum systolic blood pressure, and episodic hypertension. Lancet