Introduction

High blood pressure (BP) in patients with diabetes mellitus type 1 (T1DM) is currently considered one of the principal risk factors for the development and progression of micro-vascular complications and cardiovascular disease. It is estimated that between 35 and 75% of the complications of diabetes are due to the coexistence of elevated BP. Controlling BP decreases morbid-mortality not only in patients with T1DM but also in the general population.1, 2 It is becoming progressively more frequent to employ ambulatory monitoring of BP (ABPM). The technique can detect sub-clinical alterations of BP levels, such as the non-dipper or masked hypertension. Further, ABPM has high reproducibility, reflects the patient’s alert reaction response and the outcomes correlate better with target organ involvement and cardiovascular morbido-mortality.3 Our group has found an elevated prevalence of masked hypertension (31.8%) in patients with T1DM related, essentially, with a chronic deterioration of metabolic control.4

In recent years, greater inflammatory activity has been described in hypertensive subjects, expressed by an increase in the levels of C-reactive protein, homocysteine, interleukin-6 (IL-6) and tumor necrosis factor. However, what still remains to be clarified is whether the pro-inflammatory state is the cause or the consequence of the hypertension.5, 6, 7 Further, the majority of studies have been conducted with the conventional measurement of BP, with all the limitations that this implies.6

Conversely, elevated levels of pro-inflammatory cytokines have been confirmed in early stages of T1DM in different models, which suggest a potential implication in the auto-immune pathogenesis.9, 10, 11 Similarly, in patients with type 2 diabetes and in other related conditions, hyperglycemia not only increases oxidative stress,12, 13, 14 the levels of pro-inflammatory cytokines15, 16 and adhesion molecules but also inhibits vasodilatation (as measured by nitric oxide) and produces a pro-coagulant effect.17 Although there are several studies on the relationships between these biomarkers of vascular damage and type 2 diabetes and metabolic syndrome18, 19 there is a dearth of data on patients with T1DM, and the potential implications of the clinical or sub-clinical alterations in BP are not, as yet, known.

The objective of the present study was to evaluate the relationship of early tensional disorders (masked hypertension and non-dipper pattern) with inflammatory parameters (interleukin-6 (IL-6), tumor necrosis factor and vascular endothelial growth factor (VEGFα)) and markers of endothelial damage (vascular adhesion molecule (VCAM), intercellular adhesion molecule (ICAM) and plasminogen activator inhibitor-1 (PAI)) in normotensive patients with T1DM.

Subjects and methods

Study population

This was an observational, cross-sectional study that included 85 patients with T1DM, clinically normotensive and with consistently negative screening for micro-albuminuria who were selected consecutively from the diabetes outpatient clinic of the Endocrinology and Nutrition Unit of the Puerta del Mar University Hospital (Cadiz, Spain). Inclusion criteria were: (1) T1DM of >1 year duration; (2) age between 18 and 45 years; (3) absence of hypertension (defined as systole ⩾130 mm Hg and/or diastole ⩾80 mm Hg in more than 1 determination and/or receiving anti-hypertension medication for other reasons); (4) absence of micro-albuminuria (defined as urine albumin levels <30 mg/24 h and, as well, albumin/creatinine ratio in morning urine <30 μg mg−1 creatinine) and renal insufficiency (levels of creatinine >1.2 mg dl−1 and glomerular filtration rate <90 ml min−1); (5) absence of pregnancy, lactation or presence of concomitant diseases; and (6) written informed consent to participation in the study.

Procedure

The study protocol was approved by the Ethics Committee of the Puerta del Mar University Hospital and conformed to the principles of the Declaration of Helsinki. On the day of insertion of the ABPM device, the following data were recorded: age, sex, years of diabetes duration, weight, height and body mass index. To evaluate stable metabolic control, historical HbA1c levels were calculated as the mean of all the measurements registered in the patient’s clinical notes (at least 1 measurement every 6 months). With respect to the retinopathy screening, a retinography was performed with a non-mydrasis TOP CON camera. The results were evaluated by the same clinical ophthalmology specialist for all the participants. Fasting blood samples were taken into tubes with EDTA as anticoagulant. Plasma was isolated by centrifugation (6 min; 4000 r.p.m.), distributed into aliquots and stored at −20 °C until required for batched analyses.

Laboratory analyses

HbA1c was measured in a Cobas Integra 700 analyzer (Roche Diagnostics, Rotkreuz, Switzerland), using an immuno-turbidometric technique for completely hemolyzed anti-coagulated blood (laboratory reference range for healthy individuals: 4.5–5.7%). The lipid profile including triglycerides (mmol l−1), low-density lipoprotein cholesterol (mmol l−1) and high-density lipoprotein cholesterol (mmol l−1) were measured in a Modular DPD auto-analyzer (Roche Diagnostics) using specific enzymatic colorimetry. Plasma creatinine was measured in the Modular DPD auto-analyzer (Roche Diagnostics; normal reference range concentrations <1.2 mmol l−1). Glomerular filtration rate was calculated using the formula: glomerular filtration rate= ((140−age in years) × weight (in Kg))/(72 × creatinine (in mg dl−1)) in males and the result × 0.85) in females. Urinary albumin levels were determined using immuno-turbidometryin the Cobas Integra 700 system (Roche Diagnostics; normal values <30 mg/24 horas).20

Concentrations of the inflammatory cytokines (including IL-6, tumor necrosis factor-α and VEFG) were measured in serum and the markers of endothelial damage (VCAM, ICAM and PAI) in plasma using appropriate commercial kits (ProcartaInmunoassays Human multiplex, Affymetrix, Santa Clara, CA, USA) using the Luminex equipment that uses the Xmap technology as base, and following the manufacturer’s instructions. The coefficient of variability intra-assay was <10% and inter-assay <15% for all measured markers.

Recording of BP over 24 h

We used the SPACELABS 90207 monitor (SpacelabsInc, Washington, USA), validated by the British Hypertension Society,21 with a cuff adapted to the brachial circumference of each patient, and located on the non-dominant arm. A measurement of BP was made every 20 min of the patient’s activity periods and every 30 min during rest periods (defined by the hours established by the patient). Values of <75% of valid readings were considered invalid. The following recordings were considered as pathological: (a) average BP systolic ⩾130 mm Hg or diastolic ⩾80 mm Hg during the 24 h period and of activity, or average of BP systolic ⩾120 mm Hg or diastolic ⩾70 mm Hg during rest periods; (b) Nocturnal decrease of BP systolic or diastolic <10% relative to the diurnal mean (non-dipper pattern).

Statistical analyses

The data were codified, introduced into the database and analyzed using SPSS for Windows (version 15.0; International Business Machines Corporation (IBM), Armonk, NY, USA). Descriptive analyses of the qualitative variables were expressed as frequencies and percentages. Quantitative variables were expressed as the mean, s.d. for those variables that followed normal distribution, and median and range for those variables that did not follow a normal distribution. Comparison of quantitative variables between independent groups was with the Student t-test or the Mann–Whitney U test for non-parametric variables. Comparison of qualitative variables was with the χ2 test. The relationship between any two quantitative variables was tested with the Pearson’s or Spearman’s correlation coefficient. Multivariate analysis was performed with binary logistic regression models and linear regression analysis. All the significant values refer to two-tailed test, and significance was set at P<0.05.

Results

Clinical and BP parameters

Included were 89 patients with T1DM. They had normo-albuminuria, were normotensive and without hypotensive medication. Excluded from the final analyses were four patients who presented with ABPM with <75% of valid readings. The principal characteristics of the 85 patients finally included in the study are summarized in Table 1.

Table 1 Demographic, clinical and laboratory characteristics of the participants in the study
Full size table

Before inclusion, the office BP records of all patients were reviewed and those patients who had pathological levels in any predetermination were excluded. The principal BP parameters are summarized in Table 1. Despite all the patients included in the study being clinically normotensive, according to the ABPM 16 patients (18.8%) presented with pathological levels of 24-h BP systolic and diastolic, and 27 (31.8%) during the period of activity. Further, 36 (42.4%) patients showed a nocturnal decrease in BP (systolic or diastolic) of <10% (non-dipper pattern).

Bivariate analysis of the relationship between the levels of inflammatory biomarkers and the parameters of ABPM

In the relationships between the different parameters of ABPM and inflammatory markers, of note was that the levels of IL-6 and VEGF were significantly higher among the patients with diurnal systolic BP⩾130 mm Hg, or diastolic ⩾80 mm Hg. The levels of VCAM were, as well, more elevated in the patients with average pathological levels despite, in this case, not reaching statistical significance (Table 2).

Table 2 Cytokines and markers of endothelial function damage measured in ABPM
Full size table

Bivariate correlation between the BP parameters and inflammatory markers

There were significant correlations between levels of VEGF and several BP parameters: BP diastolic (r: 0.329; P=0.002) and systolic (r: 0.276; P=0.012) over 24 h; BP diastolic (r: 0.334; P=0.002) and systolic (r: 0.309; P=0.005) in the activity period (Figures 1 and 2). Similarly, a positive correlation was observed between the levels of IL-6 with the average 24 h diastolic BP (r: 0.242; P=0.027) and diastolic BP in the period of activity (r: 0.231; P=0.035).

Figure 1
figure 1

Correlation between diurnal activity systolic BP and VEGF level.

Full size image
Figure 2
figure 2

Correlation between diurnal activity diastolic BP and VEGF level.

Full size image

However, no significant correlations were found between the markers of endothelial damage (VCAM, ICAM and PAI) and the mean BP values, nor in the alterations in circadian rhythms of BP.

Multivariate analyses of the relationships between the inflammatory parameters and BP

Table 3 summarizes the results of the final binary logistic regression model for the dependent variable ‘diurnal systolic BP ⩾130 or diastolic BP ⩾80 mm Hg’ in which the independent variables included were those probable clinical significance, or those statistically significant in the bivariate analyses. Of note is that the levels of IL-6 was a risk factor for diurnal pathological levels of BP (OR=1.446; P=0.014).

Table 3 Binary logistic regression model for the dependent variable BP in the pathological active period
Full size table

To identify possible relationships between the BP and inflammatory markers, we have performed a multiple liner regression analysis (Table 4). Following adjustment for possible confounding factors, the concentrations of VEGF (β=0.217; P=0.033), duration of diabetes and male gender were observed to be independent factors influencing the levels of diurnal systolic BP.

Table 4 Multiple linear regression model with the dependent variable SBP in active period
Full size table

Discussion

A high percentage of patients with T1DM and normal BP (based on the mean BP cut-off) can present with masked hypertension, or the non-dipper pattern, when ABPM analysis is performed, implying a greater risk of complications.22, 23 However, there is no consensus regarding the candidates, the frequency of measurement, nor the limits of normality of ABPM in T1DM patients. In the general population, the levels considered appropriate for BP are <130/80 mm Hg over a period of 24 h; <135/85 mm Hg in periods of activity and <120/70 mm Hg at rest at night.3 In our study, normal levels were <130/80 mm Hg in the active period, which was the threshold proposed for BP in patients with diabetes at the time of conducting this study.24 Hence, in our cohort, 31.8% presented with diurnal hypertension and 22.4% nocturnal hypertension with ABPM, despite being considered normotensive following a single-point determination of BP.4 This population mean is clearly higher than that described in general populations (between 10 and 20%)23 and similar to that described in subjects with T1DM by other authors.25

In relation to the inflammation profile, the results of the present study indicate that there is an increase in pro-inflammatory cytokines such as IL-6 and VEGF in early stages of hypertension in patients with T1DM. Conversely, the levels of tumor necrosis factor-α, marker related to the development of micro-angiopathy complications and duration of diabetes,26 is not associated with the alterations in BP, nor with the markers of endothelial damage (VCAM, ICAM and PAI).

These results are difficult to compare with those described in the literature due to the different characteristics of the populations studied and the different stages of hypertension evaluated. In general, several studies have measured elevated levels of C-reactive protein, E-selectin, ICAM and VCAM in subjects with diabetes, compared with healthy control individuals.27, 28 Hence, some authors postulate that HbA1c should not be considered only as an indicator of metabolic control, but also as a molecule implicated in the genesis of endothelial dysfunction.29 Similarly, in patients with diabetes, the implications of certain pro-inflammatory cytokines VCAM and ICAM have been describe in relation to the pathogenesis of retinopathy,30, 31 nephropathy32, 33 and cardiovascular disease18, 31

However, the majority of publications that assessed the associations between the different inflammatory markers and the presence of hypertension had employed the conventional measurement of BP. As such, despite the evidence that the coexistence of hypertension and diabetes increases the risk of appearance and progression of its complications,34, 35 there is no sufficient information on the significance of early sub-clinical BP alterations and the role that cytokines and the markers of endothelial damage play in their appearance. Zorena et al,36 in a cross-sectional study, showed levels of VEGF significantly higher in patients with T1DM and hypertension compared with those patients without hypertension. On the other hand, in the only prospective study published to-date in patients with T1DM follow-up for 15 years, Sahakyan et al37 showed that patients with elevated baseline levels of VCAM had a higher risk of developing hypertension. None of the other markers of inflammation analyzed (C-reactive protein, tumor necrosis factor-α, IL-6, ICAM-1 and homocysteine) were related to the presence of hypertension in this cohort.37

That there was no evidence in our study of significant differences in the markers of endothelial damage (such as VCAM) in relation to hypertension, This could be due to the patients having been diagnosed as having hypertension in the ABPM method (31.8% of the patients in our study) and had values not very far beyond the reference range of normality. These findings conflict with those of other studies.37, 38

Although in our series of patients we had demonstrated higher levels of VEGF and IL-6 in those patients with altered diurnal BP parameters, there still remains a need to clarify whether the pro-inflammatory status is a cause or a consequence of the hypertension. Although the relationship between BP and VEGF is statistically significant, even after adjusting for the other major variables, the correlation coefficient shows a low direct relationship and should be interpreted with caution. Of note is that the presence of a sub-clinical inflammatory status contributes not only to the appearance of T1DM but also to the vascular complications.27, 39, 40 In this sense, the detection of elevated levels of pro-inflammatory cytokines in very early stages of hypertension would help the hypothesis of inflammatory activity as a primary cause. Further, the grade of glycemic control was similar in both groups and, as such, obscured the potential impact of metabolic control on the levels of cytokines and endothelial adhesion molecules. On the other hand, patients who present with sub-clinical BP alterations show a more atherogenic profile (higher levels of triglycerides and lower levels of high-density lipoprotein cholesterol), placing these metabolic alterations as causal factors in the elevation of biomarkers and of the early alterations in BP.41

The principal strengths of our study result from the homogeneity of our sample and in the evaluation, for the first time, of inflammatory status and of endothelial dysfunction in patients with pre-clinical alterations in BP and T1DM, using the measurement of pro-inflammatory cytokines and VCAMs. Nevertheless, our study has some limitations. First, the intrusion of the BP measurements into the life-style of the patient can modify the levels. Second, a greater number of patients studied would have enabled a greater precision of measurement of the trends in the associations observed. Third, the cross-sectional design of the study does not enable causal relationships to be drawn between the altered BP levels measured by ABPM, and the markers of inflammation markers

In conclusion, in patients with T1DM with early alterations in BP, there is an inflammatory status that is manifested as elevated levels of Il-6 and VEGF. The physio-pathological and clinical significance of these findings would require the conduct of further long-term prospective studies, with a wide population base.