Hypertension and proteinuria are major risk factors for cardiovascular disease (CVD) mortality in patients with type 2 diabetes. Blood pressure (BP) targets have been progressively lowered in these patients to prevent or delay the progression of nephropathy. However, no long-term population-based studies have been reported on the interaction between BP and proteinuria with respect to total and CVD mortality in patients with type 2 diabetes.
We prospectively followed 881 middle-aged type 2 diabetic patients, free of CVD events at baseline, for up to 18 years. Participants were categorized into four groups according to baseline systolic BP (<130, 130–139, 140–159 and ≥160 mmHg) and further stratified by proteinuria (≤150 or >150 mg L−1). Cox proportional hazards model was used to estimate the joint association between systolic BP and proteinuria and the risk of mortality.
During follow-up, 607 patients died including 395 because of CVD. After adjustment for confounding factors, total and CVD mortality were significantly higher in patients with proteinuria and systolic BP <130 mmHg compared with those with systolic BP between 130 and 160 mmHg. The prognosis was similar in patients with systolic BP <130 mmHg or ≥160 mmHg. Among patients without proteinuria, systolic BP <130 mmHg was associated with a nonsignificant reduction in mortality.
Type 2 diabetic patients with proteinuria and with systolic BP <130 mmHg may have an increased risk of CVD mortality. The presence of proteinuria should be taken into account when defining the target systolic BP level for the prevention of fatal CVD events in patients with type 2 diabetes.
Patients with type 2 diabetes have an approximately 2- to 3-fold increased risk of dying from cardiovascular disease (CVD) compared with the general population . Hypertension and proteinuria are major risk factors for CVD in diabetic patients [2, 3]. The coexistence of these two conditions in patients with type 2 diabetes results in higher mortality rates than the isolated occurrence of either risk factor alone . Treatment of elevated blood pressure (BP) and the management of other conventional CVD risk factors reduce both macro- and microvascular complications related to diabetes . Thus BP targets have been progressively lowered, especially among patients with hypertension and end-organ damage (e.g. retinopathy or nephropathy). However, no long-term population-based studies have been reported on the interaction between BP and proteinuria with respect to total and CVD mortality in patients with type 2 diabetes. Therefore, we investigated the impact of BP on total and CVD mortality among middle-aged Finnish type 2 diabetic patients, with or without proteinuria, in a prospective 18-year follow-up study.
Materials and methods
A detailed description of the study participants has been published previously . Overall, 1059 subjects (581 men, 478 women) with type 2 diabetes, aged 45–64 years, born and living in the Turku University Hospital district in West Finland or in the Kuopio University Hospital district in East Finland were identified through a national drug reimbursement register. Patients with type 1 diabetes, based on early onset diabetes, history of ketoacidosis and glucagon-stimulated C-peptide measurement at baseline, were excluded.
We excluded from all statistical analyses a total of 174 subjects who had possible or definite stroke, possible or definite myocardial infarction (MI) or lower-extremity amputation at the baseline examination. We additionally excluded four patients owing to missing proteinuria data from all statistical analyses involving proteinuria. Thus, the final study population comprised 881 patients with type 2 diabetes (460 men, 421 women).
The baseline examination was carried out between 1982 and 1984 during one outpatient visit at the Clinical Research Unit of the University of Kuopio or the Rehabilitation Research Centre of the Social Insurance Institution in Turku. The visit included an interview to determine history of smoking, alcohol intake, physical activity, use of medication and history of chest pain suggestive of coronary heart disease (CHD). The examinations and the methods used have been described previously in detail .
The Rose classification was used to evaluate the presence of typical angina pectoris and intermittent claudication. Whitehall changes according to Minnesota coding were used to identify ischaemic changes on the electrocardiogram (ECG) . CHD was defined as ischaemic ECG changes and typical symptoms of angina pectoris.
Medical records of patients who reported that they had been admitted to hospital for chest pain were reviewed by two investigators (M.L. and T.R.) after careful standardization of the methodology. World Health Organization (WHO) criteria for verified definite or possible MI based on chest pain symptoms, ECG changes and determination of enzyme activities were used to define previous MI . WHO criteria were also used to define previous definite or possible stroke . After a 5-min rest, BP was measured in the sitting position in the right arm with a mercury sphygmomanometer (cuff size 12.5 × 40 cm) by one of the investigators (M.L.) in East Finland and by a trained laboratory nurse in West Finland. BP was measured twice, with a 1.5-min interval, and the calculated mean was used in statistical analyses. Systolic and diastolic BP were read to the nearest 2 mmHg. The disappearance of the Korotkoff sounds (5th phase) was recorded as diastolic BP. Three series of quality-control measurements of BP were arranged during the 2-year period of the study. There were no statistically significant differences between the BP measurements in East and West Finland .
After 12-h fast, blood was collected at 08.00. Glycosylated haemoglobin (HbA1) level (reference range in nondiabetic subjects 5.5–8.5%) was determined by affinity chromatography (Isolab, Akron, OH, USA). Levels of serum total cholesterol, HDL cholesterol and triacylglycerol were determined using standard laboratory methods . Coomassie brilliant blue was used to measure total urinary protein concentration from a morning spot urine specimen . The interassay coefficient of variation was 7% at protein levels of 100 and 250 mg L−1 and 3% at 600 mg L−1 in this study. Plasma creatinine level was determined by the Jaffe method. We used the Cockcroft-Gault (CG) equation to estimate glomerular filtration rate (GFR): 1.23 × (140 − age (years)) × weight (kg) × 0.85 (if female)/plasma creatinine (μmol L−1). Body surface area (BSA) in m2 was calculated as 0.007184 × height (cm)0.725 × weight (kg)0.425 and used to correct the CG equation for BSA: CG-determined estimated GFR × (1.73 m2 per BSA) [12, 13].
Follow-up continued for 18 years until 1 January 2001. Information on vital status of all participants and copies of death certificates of subjects who had died before the end of follow-up were obtained from the national Cause-of-Death Register (Statistics Finland). All death certificates of participants were reviewed by two authors (A.J. and S.L.). In the final classification of causes of death, hospital records and autopsy records were also reviewed if available. The study endpoints were total mortality, CVD mortality [International Classification of Diseases 9th revision (ICD-9) codes 390–459] and CHD mortality (ICD-9 codes 410–414).
The Ethics Committees of the Turku University and Turku University Central Hospital, and the University of Kuopio approved the study. Informed written consent was obtained from all participants.
All statistical analyses were performed using pasw statistics (version 18.0; SPSS Inc., Chicago, IL, USA) and sas (version 9.2; SAS Institute Inc., Cary, NC, USA). Data for continuous variables are expressed as mean ± SD or median (interquartile range) and categorical variables as percentage. Baseline characteristics were compared using analysis of variance for continuous variables or the Kruskal–Wallis test when appropriate, and the chi-square test for categorical variables. Owing to a skewed distribution, total triacylglycerol was analysed after logarithmic transformation. Participants were categorized according to systolic BP into four groups (<130, 130–139, 140–159 and ≥160 mmHg) based on the recommendations of the European Society of Hypertension for hypertension management . Furthermore, participants were stratified according to the occurence of proteinuria: no proteinuria (≤150 mg L−1) or borderline/clinical proteinuria (>150 mg L−1). Our cut-off level for proteinuria was based on the observation that proteinuria >150 mg L−1 significantly predicts total and CVD mortality rate, and therefore indicates patients at high risk of total and CVD mortality as well as incipient renal impairment. A similar cut-off level was used in a previous study of the same population . Moreover, we decided to use this cut-off point instead of the more commonly used value of 300 mg L−1 because the latter would have resulted in a relatively small number of cases, especially in the proteinuria groups, thus limiting the statistical analysis.
Statistical analyses showed a significant interaction between systolic BP and proteinuria with respect to mortality. No interaction was observed between diastolic BP and proteinuria. Therefore, we used the Cox proportional hazards model to evaluate the effect of systolic BP in different proteinuria groups on total and CVD mortality with the <130 mmHg group as the reference. Unadjusted and adjusted hazard ratios and their 95% confidence intervals were calculated. Adjustment was made for age, gender, duration of diabetes, area of residence (East or West Finland), total cholesterol, use of alcohol (user versus nonuser), smoking (smoker versus nonsmoker), HDL cholesterol, total triacylglycerol (log), HbA1, diabetes medication (diet alone, oral drugs, insulin), physical activity, diastolic BP, BP medication, body mass index (BMI), estimated GFR, presence of retinopathy and CHD. Kaplan–Meier procedure was used to evaluate the associations between different levels of systolic BP with/without proteinuria and both total and CVD mortality. There were no interactions between gender and systolic BP or gender and proteinuria. Therefore, men and women were combined in all statistical analyses. P < 0.05 was considered to be statistically significant.
The general characteristics of the study participants at baseline, within the four systolic BP categories, are shown in Table 1. Higher baseline systolic BP levels were significantly associated with the female gender, increased age, diastolic BP, BP medication use, total cholesterol, triacylglycerol, BMI and proteinuria and a lower prevalence of smoking. Alcohol consumption was lowest in patients with systolic BP between 130–139 and ≥160 mmHg. Among patients without proteinuria, systolic BP was <130 mmHg in 16.3%, 130–139 mmHg in 15.3%, 140–159 mmHg in 38.1% and ≥160 mmHg in 30.3%. The values among patients with proteinuria were 12.7%, 9.8%, 35.0% and 42.5%, respectively.
Table 1. Baseline characteristics and number of subjects with endpoints by systolic blood pressure
Data are expressed as the mean ± SD or median (interquartile range), unless otherwise indicated. P value obtained by analysis of variance, the Kruskal–Wallis test or chi-square test.
56.0 ± 5.69
57.1 ± 5.01
57.8 ± 5.00
59.2 ± 4.52
Diastolic BP, mmHg
76.8 ± 9.21
78.8 ± 9.22
86.0 ± 9.37
92.3 ± 11.8
Duration of diabetes, years
7.9 ± 4.9
7.7 ± 4.3
8.1 ± 3.7
8.0 ± 3.7
Total cholesterol, mmol L−1
6.34 ± 1.49
6.55 ± 1.38
6.70 ± 1.78
6.87 ± 1.76
HDL cholesterol, mmol L−1
1.19 ± 0.35
1.29 ± 0.39
1.21 ± 0.35
1.25 ± 0.37
Triacylglycerol, mmol L−1
1.89 ± 0.99
1.96 ± 1.16
2.64 ± 3.14
2.83 ± 3.27
10.0 ± 2.9
9.8 ± 2.0
9.9 ± 1.9
9.9 ± 2.3
BMI, kg m−2
28.1 ± 4.88
28.6 ± 5.38
29.6 ± 4.87
29.7 ± 5.78
Estimated GFR, ml min−1 1.73 m−2
90.1 ± 17.3
91.4 ± 19.8
91.3 ± 20.4
88.8 ± 23.0
Physical activity, METs
4.3 ± 2.2
4.2 ± 2.0
3.9 ± 1.7
3.9 ± 1.7
Area of residence, Turku, %
BP medication, %
Current smokers, %
Alcohol users, %
Retinopathy (mild/proliferative), %
CHD without MI at baseline, %
Diet alone, %
Oral drugs, %
Insulin therapy, %
Cardiovascular disease mortality
Urinary protein, mg L−1
<150 mg L−1
≥150 mg L−1
Outcome according to systolic blood pressure and proteinuria
During 10 784 patient years of follow-up, a total of 607 (68.9%) patients died, including 395 (44.8%) from CVD. The event rates of total, CVD and CHD mortality according to systolic BP in patients with or without proteinuria are shown in Fig. 1.
In Cox regression analyses, a statistically significant interaction between proteinuria and baseline systolic BP (P = 0.01) was observed for total mortality, and a borderline significant interaction (P = 0.05) for CVD mortality. For CHD mortality, the P value for interaction was 0.07 in univariate analysis and 0.10 in multivariate analysis. Owing to this interaction we examined mortality rates among systolic BP categories separately in subjects with and without proteinuria (Table 2). Systolic BP <130 mmHg has been recommended as a target level in diabetic subjects, therefore this BP group was used as a reference for comparison with all other groups .
Table 2. Relative mortality in patients in various systolic BP categories stratified by the presence of proteinuria, compared to those with systolic BP <130 mmHg
aVariables included in multivariate adjustment: age, gender, diabetes duration, HbA1, total cholesterol, smoking, alcohol, BMI, area of residence, physical activity, diastolic pressure, BP medication, type of diabetes therapy, triacylglycerol, estimated GFR, HDL cholesterol and presence of retinopathy and of CHD without MI. *P < 0.05 for the difference between patients in various systolic BP groups and those with systolic BP <130 mmHg.
After adjustment for confounding factors, patients with proteinuria with systolic BP <130 mmHg had significantly increased (approximately 2-fold higher, P < 0.05) total and CVD mortality than those with systolic BP between 130 and 139 mmHg and increased (approximately 1.6-fold higher, P < 0.05) total and CVD mortality than those with systolic BP between 140 and 159 mmHg. Among patients without proteinuria, systolic BP <130 mmHg tended to be associated with slightly reduced CVD mortality. We observed no interaction between proteinuria and diastolic BP, and therefore only results for systolic BP are shown.
Figure 2 shows Kaplan–Meier curves of cumulative survival for patients in different systolic BP categories stratified by baseline proteinuria level. Among patients without proteinuria, those with systolic BP <130 mmHg tended to have a decreased risk of total mortality. However, among patients with proteinuria, patients with systolic BP <130 mmHg had the worst prognosis with respect to total mortality. This became evident after 2 years of follow-up.
To study whether the high risk of CVD death in patients with proteinuria and with systolic BP <130 mmHg could be explained by differences in the prevalence of baseline CHD without prior MI, we calculated the prevalence of CHD by symptoms and ECG changes in the population stratified by proteinuria and systolic BP. In patients with proteinuria, the prevalence of CHD was 28.8%, 20.0%, 15.4% and 22.4% in patients with systolic BP <130, 130–139, 140–159 and ≥160 mmHg, respectively (P = 0.18). In patients without proteinuria, the respective prevalence of CHD was 5.2%, 11.1%, 14.4% and 19.6% (P = 0.027).
To our knowledge, this is the first large population-based long-term follow-up study to investigate the impact of systolic BP on total and CVD mortality in patients with type 2 diabetes with and without proteinuria. Our main finding was that systolic BP <130 mmHg was associated with increased total and CVD mortality among patients with proteinuria compared with those with systolic BP of 130–139 or 140–159 mmHg. Indeed, patients with proteinuria with BP <130 mmHg tended to have a poorer prognosis compared even to those with BP ≥160 mmHg. Patients with systolic BP between 130 and 139 mmHg had the best prognosis in terms of total and CVD mortality. These associations were independent of conventional CVD risk factors, diastolic BP, duration of diabetes, diabetes and BP treatment, glycaemic control, estimated GFR, retinopathy and baseline CHD without MI.
We found that among patients with type 2 diabetes without proteinuria, systolic BP <130 mmHg was associated with a tendency towards slightly lower total and CVD mortality. It is noteworthy that diabetes among the study patients was poorly controlled at baseline. However, the level of glycaemic control during follow-up was unknown; therefore it is unclear how changes in glycaemic control may have influenced the prognosis of these patients. It is known that good glycaemic and BP control have additive positive effects on total and CVD mortality in patients with type 2 diabetes [5, 16], but it is not known whether this additive effect is also observed among patients with proteinuria.
Our findings indicate that patients with type 2 diabetes with proteinuria have an increased risk of mortality when systolic BP is lower than <130 mmHg, compared with higher BP levels. It is worth mentioning that this finding does not necessarily apply to incident nonfatal CVD events such as MI. There is also a possibility that the increased risk in the low BP group may be because of chance given to the limited sample size in the proteinuria group. However, our findings are likely to be valid for several reasons. First, it has been proposed that microalbuminuria and similarly proteinuria may be indicators of generalized endothelial dysfunction [17, 18] and therefore may indicate atherosclerotic disease leading to underperfusion of vital organs when BP is low, thus increasing CVD mortality. This hypothesis is supported by our analyses of baseline data showing that CHD prevalence evidenced by symptoms and ECG findings was only 5.2% among patients without proteinuria and systolic BP <130 mmHg versus 28.8% (i.e. approximately 5-fold increase) in patients with proteinuria and systolic BP <130 mmHg. Secondly, impaired left ventricular systolic function, known to be associated with chronic kidney disease and diabetes, could decrease BP as a result of reduced cardiac output [19-22]. Thus, low systolic BP in such patients may be a marker of underlying disease as the cause of worse prognosis.
Our observations in patients with type 2 diabetes and proteinuria are in line with those in the IDNT trial , a post hoc study in which type 2 diabetic patients with nephropathy and with systolic BP ≤120 mmHg had an increased CVD mortality rate. Again, consistent with the present findings, two studies in type 2 diabetic patients with coronary artery disease showed that the benefits of lowering systolic BP to <130 mmHg were driven mostly by a reduction in stroke incidence whereas CVD and total mortality rate were unchanged or even increased [24, 25]. In the recent ACCORD trial, lowering systolic BP to <120 mmHg in high-risk type 2 diabetic patients did not reduce mortality compared to those with a systolic BP target of <140 mmHg . Previous findings as stated before and also our present finding, suggest that one target BP level may not necessarily fit all type 2 diabetic patients. It appears that there is heterogeneity of the effects on different CVD outcomes (e.g. nonfatal events versus mortality) of intensive BP control and also of lowering BP to <130 mmHg. In addition, the effects may also vary on the basis of the presence or absence of comorbid conditions, such as proteinuria. Therefore, lowering systolic BP to <130 mmHg might not be justified for type 2 diabetic patients with proteinuria or for patients otherwise at high risk of CVD-related mortality. Furthermore, even though the evidence supporting a target systolic BP level <140 mmHg in slowing progression of kidney disease is strong, there are limited data regarding the effects of lowering BP to <130/80 mmHg [14, 27, 28]. The mode and intensity of treatment of BP should be based on the patient's individual CVD risk and possible comorbid conditions rather than on aggressive BP goals recommended by global guidelines alone. Based on a recent meta-analysis, it was concluded that, among patients with type 2 diabetes, the more aggressive target for systolic BP of <130 mmHg has to be balanced between the benefits of lowering stroke incidence and the increased risk of serious adverse events, such as life-threatening events or hospitalization, as well as the lack of benefit for cardiac, renal and retinal outcomes . Furthermore, in type 2 diabetes, proteinuria has been highlighted as an important prognostic marker for cardiac disease, in addition to conventional CVD risk factors . Thus, our results suggest that it might be justifiable to recommend higher systolic BP targets for patients with type 2 diabetes and proteinuria compared with those without proteinuria.
Our study has several strengths. First, our baseline examination was carried out in 1982–1984. As statin treatment did not become common until the second half of the 1990s, it is unlikely that lipid-lowering therapy has caused a major bias in our study. Secondly, unlike many similar epidemiological studies, data regarding glucose control, diabetes duration and mode of diabetes treatment were available. Thirdly, we used the post-glucagon C-peptide measurement to exclude patients with type 1 diabetes. Finally, to avoid a potential bias from possibly increased early mortality because of a severe disease at baseline, we excluded all subjects with possible or definite stroke, possible or definite MI or amputation at baseline.
Our study also has limitations. First, BP was only measured at baseline. Therefore, we lack data on changes in BP during follow-up. Moreover data were not available on the use of angiotensin-converting enzyme (ACE) inhibitors, which are known to improve the prognosis of patients with kidney disease. However, as the first ACE inhibitor captopril was approved by the US Food and Drug Administration in 1981, it is highly unlikely that our study participants were taking these drugs at baseline. Secondly, we measured total urinary protein concentration from spot urine samples instead of measuring urinary albumin excretion rate. Our findings do not apply to subjects with microalbuminuria. On the other hand, our cut-off point for proteinuria, 150 mg L−1, is lower than the commonly used level for clinical proteinuria (300 mg L−1). However, 150 mg L−1 was previously associated with total and CVD mortality , therefore indicating a high risk of CVD. Owing to this limitation, it is likely that our study underestimates rather than overestimates the association of BP and proteinuria with CVD outcomes and total mortality. Thirdly, without the necessary data we were not able to make assumptions about the safe systolic BP levels for patients with renal failure, It should be mentioned that at the time of our baseline examination, risk assessments and interventions for CVD in type 2 diabetes patients were not routinely performed, in contrast to the situation in clinical practice today. These factors are likely to have affected the natural history of patients with proteinuria with respect to CVD risk.
In conclusion, systolic BP <130 mmHg tended to be associated with a slight decrease in total and CVD mortality rate in middle-aged type 2 diabetic patients without proteinuria. Among patients with proteinuria, systolic BP <130 mmHg was associated with a significant increase in total and CVD mortality rate. Patients with a systolic BP level between 130 and 139 mmHg had the best prognosis with respect to total and CVD mortality. These results suggest that type 2 diabetic patients with proteinuria and with systolic BP <130 mmHg may have an increased risk of total and CVD mortality. Further investigation is needed to determine conclusively whether aggressive lowering of (systolic) BP in patients with type 2 diabetes and proteinuria would result in excess risk of CVD outcomes or death.
Conflict of interest statement
None of the authors has any conflicts of interest associated with this manuscript to declare.