QJM vol. 97 no. 11 © Association of Physicians 2004; all rights reserved.
Hypertension in diabetes: is there a place for age-adjusted centile cut-offs in those aged <50 years?
From the Wolverhampton Diabetes Centre, New Cross Hospital, Wolverhampton, UK
Received 27 November 2004 and in revised form 24 June 2004
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
Background: Conventional cut-offs for hypertension are arbitrary, and vascular complications can occur below these values, particularly in diabetes.
Aim: To evaluate systolic blood pressure (SBP) distribution in diabetes and control populations, comparing hypertension prevalence using either conventional cut-offs (140 and 160 mmHg) or age-adjusted centile (75th and 90th) cut-offs derived from the control population.
Methods: We compared 2521 individuals with diabetes to 5809 controls, after excluding those on anti-hypertensives and with established vascular disease in both groups.
Results: Diabetic individuals were older (mean ± SD 56 ± 16 vs. 43 ± 16years, diabetes vs. controls), and had higher BMI (29 ± 5 vs. 24 ± 4 kg /cm2) and SBP (145 ± 23 vs. 131 ± 18 mmHg, all p < 0.001). These factors were adjusted for in subsequent analysis. SBP rose with age and was also significantly higher in the diabetes group. In diabetes, conventional cut-offs indicated less hypertension in those aged < 50 years, compared to age-adjusted centile cut-offs. In age bands 18–29, 30–39, 40–49, 50–59, 60–69 and >70 years of the diabetes group, 24%, 33%, 43%, 62%, 70% and 74%, respectively exceeded 140 mmHg, compared to 35%, 44%, 43%, 45%, 40% and 27% exceeding the control-derived 75th centile value.
Discussion: The use of control-derived age-adjusted cut-offs may provide an alternative approach to define hypertension in diabetes that may be of particular relevance to younger patients, although this would require validation against outcomes.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Hypertension is an important risk factor in the development of vascular complications, effective treatment reducing the risk.1,2 Whether systolic or diastolic blood pressure confers greater risk has been a matter of debate, but mounting evidence from prospective trials indicates that systolic blood pressure (SBP) is more strongly related to vascular events, including death.3,4
There is strong established evidence on the effectiveness of anti-hypertensive intervention using discrete cut-offs.1,2 Nevertheless, the relationship between blood pressure levels and vascular risk is continuous and curvilinear, so the definition of when raised blood pressure is hypertension is arbitrary, whether diabetes is present or not.5 It therefore follows that vascular complications can occur at blood pressure values lower than those designated as hypertension. The evidence for this is established in people with diabetes,6,7 and may also exist in the general population.8
Blood pressure rises with age,9,10 and the use of age-adjusted blood pressure nomograms to define hypertension overcomes an important factor contributing to SBP. Such nomograms are widely used in defining hypertension in children, but not in the adult population. Furthermore, the generally held clinical impression that the prevalence of hypertension is greater among people with diabetes than in those without, has not been consistently confirmed by epidemiological studies.11 Even then, any relationship is complex and not fully explained by age and BMI.
The aim of our study was thus to evaluate the SBP distribution in an adult population with diabetes (compared to a control population) using two different definitions for hypertension: either the conventional SBP cut-offs (140 and 160 mmHg) or by 75th and 90th centile values (C75c and C90c) derived from an age-adjusted control population.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
The diabetes population was drawn from the Wolverhampton district diabetes register. This is a community-based register recording demographic, clinical and biochemical characteristics of patients at every annual review. Of the 6485 adults (>18 years) with diabetes, 2521 were selected, after excluding those on anti-hypertensive therapy and/or with known ischaemic heart disease, cerebrovascular disease, elevated serum creatinine or blindness. The control population was derived from the UK Health and Lifestyle Survey database, held at the University of Essex, UK.12 From a cohort of 10 000 individuals, 5809 were selected after applying similar exclusion criteria as above, as well as excluding people with diabetes. The exclusions were done to aid the definition of well-matched groups, excepting the presence of diabetes. Thus we avoided the confounding of SBP values by anti-hypertensive therapy and established vascular disease.
In both population study groups, trained professionals using automatic blood pressure measuring instruments took seated, resting (5 min) blood pressure measurements and the lowest of multiple readings obtained was used. In the diabetes group a Dinamap XL was used, compared to an Accutorr in the control group. Both monitors meet the standards of the American Association for the Advancement of Medical Instrumentation and the more comprehensive protocol of the British Hypertension Society. For this study, we compared two arbitrarily chosen centile values derived from the control population (75th and 90th) to two discrete thresholds (140 and 160 mmHg) for SBP.
Results are presented as means ± SD for continuous data and % for categorical variables. The significance of the difference in means between groups was determined by one-way analysis of variance, and the difference in proportions by the 
2 test. Stepwise multivariate logistic regression analysis was used to find the subset of risk variables that best explained the presence of hypertension. Where significant variables were identified, these were adjusted for in subsequent statistical analyses. Statistical analyses were performed using version 10 of the Statistical Package for Social Sciences (SPSS UK) and were considered significant at the 5% level. Approval from the Local Research and Ethics committee was obtained for this study.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
The results of the baseline characteristics between the diabetes and control population are summarized in Table 1. The diabetes group was significantly older, heavier and had more males than the control group (p < 0.001). These factors were adjusted for in subsequent analysis. The majority of the diabetes population had type 2 diabetes (1841/2521, 73%).
|
Overall, SBP was significantly higher in the diabetes group than in controls (145 ± 23 mmHg versus 131 ± 18 mmHg, F = 191.8, p < 0.001). The overall mean SBPs were 143 ± 21 vs. 135 ± 17 in males (p < 0.001) and 147 ± 25 vs. 128 ± 19 in females (p < 0.001). Using conventional SBP cut-offs, 58% vs. 28% (diabetics vs. controls) exceeded 140 mmHg and 25% vs. 8% exceeded 160 mmHg (
2 = 790.4, p < 0.001). For males alone (diabetics vs. controls), 54% vs. 34% and 22% vs. 8% exceeded 140 and 160 mmHg, respectively (p < 0.001) and for females (diabetics vs. controls), 61% vs. 22% and 29% vs. 7%, respectively (p < 0.001). On multivariate stepwise regression analysis (F = 833.9, p < 0.001), age (R2 = 0.23) was identified to be the most significant independent factor contributing to SBP. Additional significant factors were BMI (R2 = 0.28, including age), gender (R2 = 0.29, including age and BMI) and diabetes (R2 = 0.30, including, age, BMI and gender). Accounting for these factors, ethnicity did not significantly contribute to SBP. Thus, diabetes had a significant independent effect upon SBP. However, within the diabetes group, only age, BMI and duration of diabetes were significant independent factors influencing SBP (R2 = 0.19, F = 178.5, p < 0.001) but not type of diabetes, gender or ethnicity.
To account for the significant variables, the diabetes and control groups were divided into comparable decade age-bands, and the results further analysed by gender, with statistical weighting for BMI (Table 2). Mean SBP rose significantly from 125 mmHg in the 18–29-year-olds to 153 mmHg in the >70-year-olds in the control group (F = 323.7, p < 0.001) and from 130 mmHg to 156 mmHg, respectively, in the diabetes group (F = 71.4, p < 0.001). At each of the age bands (overall and by gender), the SBP of the population with diabetes significantly exceeded that of the control group (p < 0.05) and the proportion exceeding the two conventional SBP cut-offs followed a similar pattern.
|
The 75th and 90th centile SBP values (C75c and C90c) of the control population in various age bands, together with gender breakdown, are shown in Table 3. The distribution of SBP by age bands in the diabetes group, its relationship to conventional cut-offs of 140 and 160 mmHg and to the age-adjusted C75c and C90c derived from the control group are shown in Figure 1.
|
|
|
The patterns of hypertension prevalence by age in the diabetes group, based on definitions using conventional cut-offs (140 and 160 mmHg) and the C75c and C90c derived from the control group, are shown in Figure 2 (also see Table 3). While the prevalence of hypertension using conventional cut-offs continuously rose with age, the prevalence defined by age-adjusted centile cut-offs only rose until 40 years, beyond which it first levelled off, and then fell with increasing age. Comparing C75c to 140 mmHg as definitions for hypertension in diabetes, the prevalence of hypertension was higher with the former until the age of 50 years, after which the pattern reversed. A similar picture emerged when C90c and 160 mmHg were compared, although the point of intersection shifted to the next decade. There were no differences in the patterns observed when sub-analysed by gender. In the younger cohort (< 50 years), in particular, the centile cut-offs identified 7% more patients as potentially hypertensive, compared to the conventional cut-offs.
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Our paper is intended to explore the potential to expand the understanding of the definition of hypertension in a population with diabetes with particular emphasis on those aged < 50 years. We do not intend to challenge the robust data on hypertension definition and intervention that currently exists using conventionally defined cut-off values.
As observed in previous studies,9,10 SBP rose with age in our diabetes and the control populations. Our diabetes population had significantly higher SBP than controls at every age band, even when sub-analysed by gender. The increased incidence of hypertension in diabetes has not been consistently confirmed by epidemiological studies.13–16 Furthermore, methodological problems, including the poor definition of characteristics that might influence blood pressure make interpretation of such studies difficult.11 We have accounted for such confounding factors and for the differences in composition of the groups, thus minimizing any impact this may have on our results, by sub-analysis by age bands and statistically weighting for BMI, and excluding those with established vascular disease or on anti-hypertensive therapy in both groups.
Our results suggest that the use of control-population-derived age-adjusted SBP centile cut-offs may provide an alternative understanding of hypertension in diabetes. Using such centile values to define hypertension in our diabetes population, we observed a greater frequency of hypertension in the younger cohort (< 50 years) and lower frequency in the elderly subjects compared to the conventional SBP cut-offs.
While the role of hypertension in the pathogenesis of diabetic complications17 is well understood, the fundamental question of how to define hypertension is still unclear. The relationship between blood pressure values and vascular risk is continuous, and therefore current definitions for hypertension are discrete and arbitrary. There is however clear evidence for the effectiveness of intervention with anti-hypertensives using discrete cut-offs to define hypertension.1,2,18,19 The application of centile cut-offs identified fewer subjects as hypertensive in the elderly. However, there are many studies in this group which have proven the benefit of anti-hypertensive intervention using conventional cut-offs,20 and it is not our aim to dispute this. We accept that, while the age-adjusted centile SBP values may overcome the impact of age, their use to predict vascular risk is completely unknown.
However, there is a pressing need for early identification of hypertension, at least in younger patients with diabetes, considering their duration of exposure to the risk of hyperglycaemia. Vascular complications can occur at blood pressure values lower than those designated as hypertensive, certainly in diabetics6,7 and possibly also in the general population.8 Furthermore, the increasing recognition of type 2 diabetes in young adulthood who can also develop serious complications as early as in the third decade of life,21,22 implies that conventional cut-offs may potentially miss some people with a need for hypertension intervention. The use of control-population-derived age-adjusted centile cut-offs, which identified 7% more patients as potentially hypertensive in the under 50 years cohort, may bring additional focus on this problem.
By using age-specific thresholds, as opposed to fixed thresholds for a variable that rises with age, our results may seem statistically unsurprising in identifying higher proportion of hypertensives in the young. However, they do not undermine the potential clinical importance to this cohort. Furthermore, the knowledge that blood pressures tend to remain in the same quartiles as individuals' age (‘blood pressure tracking’),23 and that such individuals also have unfavourable cardiovascular risk at an early stage,24 would suggest an additional role for age-specific thresholds. Clearly, trials will be needed to validate this approach. Similarly, the question as to which centile threshold would predict risk most appropriately, would also need outcome trials.
While the majority of our < 50 years group had type 1 diabetes, we did find 47% with type 2 diabetes (21%, 32% and 67% in the 18–29, 30–39 and 40–49 year bands, respectively). While the differing underlying pathogenesis of type 1 and type 2 diabetes could have influenced blood pressure distribution, we did not find type of diabetes to have a significant independent effect on blood pressure on multivariate regression analysis. Others have similarly shown time-related and other biophysical variables (age, BMI, duration) to have the highest impact on blood pressure, and indeed when matched for them, the prevalences of vascular complications were not different between the two types of diabetes.25 We therefore believe that the application of centile cut-offs may have an impact to the younger individuals with diabetes, regardless of type.
In summary, current definitions for hypertension are arbitrary and do not take into account factors such as age. There is established evidence on the effectiveness of intervention with anti-hypertensives using such discrete cut-offs to define hypertension. We present a supplementary approach to define hypertension using control-population-derived age-adjusted centile SBP values, which may be of particular use in younger patients with diabetes. Whether such centile values predict vascular risk, and whether interventions at such cut-offs improve outcome, is unknown and requires investigation.
|
Acknowledgments |
|---|
We thank South Staffordshire Medical Foundation (SSMF), UK, for their Fellowship grant support to Dr V Baskar.
|
Footnotes |
|---|
Address correspondence to Dr V. Baskar, 7 Coven Mill Close, Coven, Wolverhampton WV9 5HX. e-mail: baskar{at}doctors.org.uk
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
1. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. UK Prospective Diabetes Study Group. Br Med J 1998; 317:703–13.
2. Hansson L, Zanchetti A, Carruthers SG, Dahlof B, Elmfeldt D, Julius S, Menard J, Rahn KH, Wedel H, Westerling S. Effects of intensive blood pressure lowering and low dose aspirin in patients with hypertension: principal results of the Hypertension Optimal Treatment (HOT) randomised trial. HOT study group. Lancet 1998; 351:1755–62.[CrossRef][Web of Science][Medline]
3. Kannel WB, Gordon T, Schwartz MJ. Systolic versus diastolic blood pressure and risk of coronary heart disease: the Framingham Study. Am J Cardiol 1971; 27:335–45.[CrossRef][Web of Science][Medline]
4. Stamler J, Stamler R, Neaton J. Blood pressure, systolic and diastolic, and cardiovascular risk. Arch Intern Med 1993; 153:598–615.
5. World Health Organisation. International Society of Hypertension Guidelines for the Management of Hypertension. Blood Pressure 1999; 8 (Suppl. 1):3–5.
6. Knowles WC, Benett PH, Ballantine EJ. Increased incidence of retinopathy in diabetics with elevated blood pressure; a six-year follow up study in Pima Indians. N Engl J Med 1980; 302:645–50.[Abstract]
7. Poulsen PL, Hansen KW, Mogensen CE. Ambulatory blood pressure in the transition from normo to microalbuminuria. A longitudinal study in IDDM patients. Diabetes 1994; 43:1248–53.[Abstract]
8. Vasan RS, Larson MG, Leip EP, Evans JC, O’Donnell CJ, Kannel WB, Levy D. Impact of High-Normal Blood Pressure on the Risk of Cardiovascular disease. N Engl J Med 2001; 345:1291–97.
9. Landahl S, Bengtsson C, Sigurdsson JA, Svanborg A, Svardsudd K. Age-related changes in blood pressure. Hypertension 1986; 8:1044–9.
10. Acheson RM. Blood pressure in a National sample of US adults: Percentile distribution by age, sex and race. Intl J Epidemiol 1973; 2:293–301.
11. Fuller JH. Epidemiology of hypertension associated with diabetes mellitus. Hypertension 1985; 7(suppl. 2):3–7.
12. Health and Lifestyle Survey, 1984–1985 (HALS1). [http://www.data-archive.ac.uk/findingData/snDescription.asp?sn=2218].
13. Reaven PD, Barrett-Connor EL, Browner DK. Abnormal glucose tolerance and hypertension. Diabetes Care 1990; 13:119–25.[Abstract]
14. Jarrett RJ, Keen H, McCartney. Glucose tolerance and blood pressure in two population samples: Their relation to diabetes mellitus and hypertension. Intl J Epidemiol 1978; 7:15–24.
15. Drury PL. Diabetes and arterial hypertension. Diabetologia 1983; 24:1–9.[Web of Science][Medline]
16. Garcia MJ, McNamara PM, Gordon T, Kannel WB. Morbidity and mortality in diabetics in the Framingham population. Sixteen year follow up study. Diabetes 1974; 23:105–11.[Web of Science][Medline]
17. Adler AI, Stratton IM, Neil HAW, Yudkin JS, Matthews DR, Cull CA, Wright AD, Turner RC, Holman RR. Association of Systolic blood pressure with macrovascular and microvascular complications of type 2 diabetes (UKPDS 36): prospective observational study. Br Med J 2000; 321:412–19.
18. Toumilehto J, Rastenyte D, Birkenhager WH, Thijs L, Antikainen R, Bulpitt CJ, Fletcher AE, Forette F, Goldhaber A, Palatini P, Sarti C, Fagard R. Effects of calcium channel blockade in older subjects with diabetes and systolic hypertension. Systolic Hypertension in Europe Trial Investigators. N Engl J Med 1999; 340:677–84.
19. Effects of Ramipril on cardiovascular and microvascular outcomes in people with diabetes mellitus: results of the HOPE study and MICRO-HOPE sub-study. Heart Outcomes Prevention Evaluation Study Investigators. Lancet 2000; 355:253–9.[CrossRef][Web of Science][Medline]
20. MacMahon S, Rodgers A. The effects of blood pressure reduction in older patients: an overview of five randomised controlled trials in elderly hypertensives. J Clin Exp Hypertension 1993; 15:967–78.[CrossRef]
21. Rosenbloom AL, Joe JR, Young RS, Winter WE. Emerging epidemic of type 2 diabetes in youth. Diabetes Care 1999; 22:345–54.
22. Dean H. NIDDM-Y in first nation children in Canada. Clin Pediatr (Phila) 1998; 37:89–96.
23. Lambrechtsen J, Rasmussen F, Hansen HS, Jacobsen IA. Tracking and factors predicting rising in ‘tracking quartile’ in blood pressure from childhood to adulthood: Odense schoolchild study. J Hum Hypertens 1999; 13:385–91.[CrossRef][Web of Science][Medline]
24. Vos LE, Oren A, Uiterwaal C, Gorissen WHM, Grobbee DE, Bots ML. Adolescent blood pressure and blood pressure tracking into young adulthood are related to subclinical atherosclerosis: The atherosclerosis risk in young adults (ARYA) study. Am J Hypertens 2003; 16:549–55.[CrossRef][Web of Science][Medline]
25. Knuiman MW, Welborn TA, McCann VJ, Stanton KG, Constable IJ. Prevalence of diabetic complications in relation to risk factors. Diabetes 1986; 35:1332–9.[Abstract]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||