“Arterial stiffness is not associated with changes in the circadian pattern of blood pressure in patients with type 1 diabetes mellitus and cardiovascular autonomic dysfunction”

Abstract

Introduction

Cardiovascular autonomic neuropathy (CAN) associates an abnormal circadian pattern in blood pressure (BP) regulation that might be aggravated by the coexistence of arterial stiffness. We aimed to evaluate the effect of arterial stiffness in the circadian rhythm of BP in patients with type 1 diabetes and CAN.

Methods

Cross-sectional study including 56 consecutive patients with type 1 diabetes and CAN, with (n = 28) or without (n = 24) arterial stiffness as defined by an ankle-brachial index above 1.2. CAN was diagnosed by BP and heart rate responses to active standing and cardiovascular autonomic reflex tests. Absence of nocturnal decrease in BP–“non-dipping” pattern– was defined by a daytime to nighttime decrease in mean BP smaller than 10%.

Results

The study’s subjects mean age was 40 ± 11 years-old, their mean duration of diabetes was 22 ± 10 years, and their mean A_1c was 7.9 ± 1.5%. A “non-dipping” pattern was observed in 28 patients (54%) regardless of the presence or absence of arterial stiffness. Age, waist circumference, body mass index, and A_1c, were introduced as independent variables into a multiple regression analysis. The stepwise model (R²: 0.113, p = 0.016) retained only A_1c levels (β: ‒ 0.333, 95% confidence interval [CI]: −3.10 to −0.33) as significant predictor of the percentage of nighttime decrease in mean BP.

Conclusions

A non-dipping pattern in BP is very common in patients with type 1 diabetes presenting with subclinical CAN and is associated with a poorer metabolic control. On the contrary, coexistence of arterial stiffness is not associated with abnormalities in circadian BP regulation.

Introduction

Diabetes is a favorable scenario for disturbed blood pressure (BP) variability, which might mediate the harmful effects of cardiovascular autonomic neuropathy (CAN) on cardiovascular risk. ¹ Absence of the physiological nighttime decrease in BP – non-dipping pattern – occurs in 39–65% of patients with type 1 diabetes (T1D), and associates increased risk for macro- and microvascular complications, including diabetic nephropathy and retinopathy.^2-4

CAN may play a major pathogenic role in this association. ¹ Cardiovascular autonomic dysfunction is commonly found in patients with T1D, appearing even during early stages of the disease.^5-7 In these subjects, the non-dipping BP pattern is associated with abnormal heart rate (HR) variability ⁸ and with CAN severity. ⁹ Moreover, CAN is associated with arterial stiffness (AS) – a well-recognized risk factor for vascular disease among young patients with T1D ¹⁰ – independently of other classic cardiovascular risk factors.^11-14 Earlier reports from our group suggested that the compliance of peripheral arteries is related to CAN among young adults with T1D who maintain acceptable glycemic control, and that such an association persisted even after adjusting for the presence of other relevant cardiovascular risk factors including peripheral artery disease. ¹²

A reduced arterial compliance could result into abnormal BP variability. ¹⁵ 24-hour ambulatory blood pressure monitoring (ABPM) reveals that one in every four patients with T1D may have masked hypertension, and AS is increased in those individuals. ¹⁶ In the same line, AS might be closely related with a non-dipping pattern in young patients with hypertension or high normal BP, as recently reported. ¹⁷

However, the possible link between AS, abnormalities in the circadian pattern of BP of patients with T1D, and asymptomatic cardiovascular autonomic dysfunction has not been completely elucidated. We hypothesized that: (i) AS may worsen the circadian BP pattern in patients with T1D and asymptomatic CAN; and (ii) hyperglycemia and subclinical inflammation may be inversely associated with the magnitude of nocturnal BP dipping among middle-aged adults with T1D and subclinical CAN. Hence, we aimed to investigate the relationship between AS, as measured by ankle-brachial index (ABI), and ABPM recordings in patients with T1D and asymptomatic CAN.

Methods

Study design

We conducted a cross-sectional study including 56 consecutive patients with T1D and diagnosis of subclinical CAN who regularly attended the diabetes outpatient clinic of an Academic Hospital from Madrid, Spain. The patients were select from a larger cohort (clinicaltrials.gov NCT02910271) designed to address the subclinical atherosclerosis profile of a population with T1D. ¹⁸ A detailed description of this trial is reported elsewhere. ¹⁸ The study protocol conformed to the ethical guidelines of Hospital Universitario Ramón y Cajal ethics committee (date of approval: 22 January 2016; protocol ID: 464/15).

Study population

The diagnosis of T1D required an age at onset of diabetes under 30 year-old, previous episode of ketoacidosis and/or diabetic autoimmunity (autoantibodies against glutamic acid decarboxylase 65 [GAD], IA2 and zinc transporter 8), and mandatory use of insulin for survival following the American Diabetes Association criteria. ¹⁹ All patients included here had a diagnosis of CAN, as defined below. Exclusion criteria were: (i) Normal cardioautonomic assessment; (ii) symptomatic intermittent claudication according to Edinburgh Claudication Questionnaire; ²⁰ (iii) previous diagnosis of peripheral artery disease; (iv) end-stage renal disease; (v) ongoing pregnancy; and (vi) diagnosis of types of diabetes mellitus other than T1D.

All the patients included had a diagnosis of CAN according to cardiovascular autonomic reflex tests. We divided our cohort into two groups according the ankle-brachial index (ABI) results: (i) study group: including patients with an ABI > 1.2, that was assumed as an equivalent of AS and, (ii) the control group: defined by all patients with an ABI ≤ 1.2.

Clinical, anthropometric, and biochemical variables

We reviewed the medical records of the subjects including clinical parameters related to T1D, current medications, smoking status cardiovascular risk factors, microvascular complications – including any T1D‐related eye disease; neuropathy considered as any T1D‐related neurological complication; and nephropathy considered as any T1D‐related kidney disease –, and macrovascular complications such as coronary and cerebrovascular disease. Current smoking was defined as having smoked cigarettes within 30 days preceding the survey. Individuals who regularly smoked, but were not current smokers were considered past smokers. Only individuals who had never smoked were considered non-smokers.

Patients underwent a complete physical examination including measurements of waist circumference, height, and weight measured in light clothes and without shoes. We calculated body mass index (BMI, kg/m²) and categorized patients according to their BMIs.

Diagnosis of diabetic nephropathy required increased urinary albumin–to–creatinine ratio (UACR) as measured in a random spot urine collection, considering as a whole those individuals with microalbuminuria (UACR 30–300 mg/g) or macroalbuminuria (UACR >300 mg/g). All patients were assessed for diabetic peripheral neuropathy ¹⁸ by means of a detailed clinical history (Neuropathy Symptoms Score) and clinical tests for large-fiber function, protective sensation, and detection of feet at risk for ulceration by a 128-Hz tuning fork for vibration perception, ankle reflexes, and a 10-g monofilament test. ²¹

Renal function, serum lipids profiles, A_1c, and circulating proinflammatory profiles were analyzed. Fasting blood and urine samples served to measure creatinine (immunoturbidimetry, Abbott Architect system, Abbott Laboratories), serum lipid profiles (Abbott Architect system), and A_1c (high performance ion-exchange chromatography, HA-8160 analyzer, A. Menarini diagnostics). The serum inflammation panel included measurements of high sensitivity C reactive protein (hs-CRP, ultra-high sensitivity latex-based immunoassay, CRP Vario, Sentinel diagnostics), erythrocyte sedimentation rate (ESR, EDTA‐anticoagulated tubes Ves‐Matic Cube 80, DIESSE), homocysteine (Abbott Architect system), and fibrinogen (Symex CS-5100 hemostasis system, Siemens Healthcare Diagnostics).

To assess insulin sensitivity, we used the estimated glucose disposal rate equation (eGDR) that relies on routine clinical measures: A_1c, presence of hypertension, and waist circumference. ²²

Assessment of cardiovascular autonomic function

All the patients included had a diagnosis of CAN according to cardiovascular autonomic reflex tests (CARTs), ²³ following the recommendations of the American Diabetes Association consensus statement on standardized measurements in individuals with diabetes. ²⁴ All patients were abstinent from beta-blocker when assessed for CAN. We used a modification of the Ewing score ¹² to rate the presence of CAN, which scored HR variability to deep breathing, Valsalva’s maneuver, and orthostatism, as well as the response of BP to active standing. These responses were categorized as normal (0 points), borderline (0.5 point), or abnormal (1 point). A composite score ≥1 was considered diagnostic of CAN.^12,23 We classified CAN as early or mild when the Ewing score was between 1 and 2, or as definite when the score was ≥2. ¹²

Between 7:00 and 9:00 a.m., and after resting in supine for at least 10 min in a space with stable temperature, we assessed HR variability using a Monitor OneDx® System device (Qmed, Inc., Eatontown, NJ). HR response to deep breathing was estimated by calculating the ratio of the maximum and minimum HRs during six cycles of paced deep breathing Expiration/Inspiration (E/I) ratio. HR response to Valsalva’s maneuver (VAL ratio) was assessed by calculating the ratio of the longest R-R interval after the maneuver to the shortest interval during or shortly after the maneuver. HR response to orthostatism was calculated as the ratio of the longest R-R interval (found at about beat 30) to the shortest interval (found at about beat 15) after standing up (30:15 ratio). ¹²

Adrenergic innervation was assessed by the changes in BP and HR 5 min after active standing, from the values recorded while resting in supine. Orthostatic hypotension was defined by a fall in response to standing >20 mmHg for systolic BP or >10 mmHg for diastolic BP. ²⁴ HR at resting was measured by palpating the radial pulse and counting the number of beats during 60 seconds. Resting tachycardia was defined by a HR >100 beats per minute. ²⁴

Before obtaining cardiovascular autonomic function studies, we assayed serum glucose in all participants to rule out hypoglycemia. No patient had a serum glucose value below 70 mg/dL, which is the glycemic threshold for epinephrine release. ²⁵

Assessment of peripheral vascular disease and carotid ultrasound examination

PAD was assessed non-invasively following a two-step approach. ¹² First, we used the test recommended for screening, the ABI, by Doppler ultrasonography using a Minidop ES 8-Mhz sonographic device (Hadeco, Inc., Kawasaki, Japan). All subjects with an ABI >1.2 were subsequently explored to corroborate or rule out the diagnosis of PAD using a combination of toe-brachial index (TBI) and peripheral arterial Doppler ultrasonography. An ABI >1.2 was assumed as an equivalent of AS. ²⁶ Although both pulse wave velocity (PWV) and augmentation index (AIx) are validated methods for assessing AS, they can be time consuming. Furthermore, ABI has been inversely related to AIx in patients without PAD in a previous report. ²⁷ The AIx has traditionally been considered an indication of the extent of wave reflection, and increased wave reflection has been interpreted as an indicator of increased AS. ²⁸ A TBI <0.7 together with loss of the waveform triphasic pattern, visualization of atherosclerotic plaques or appropriate diagnostic criteria in arterial Doppler ultrasonography suggested subclinical PAD. We comprehensively described these procedures in a previous report. ¹²

All patients showing an ABI >1.2 underwent carotid ultrasound examination. Carotid artery intima-media thickness (cIMT) was calculated for both common carotid arteries in each patient, and the mean of such measurements was used for analysis. We used a Toshiba Nemio model SSA-550A Basic Diagnostic Ultrasound System (Toshiba Medical System S.A., Alcobendas, Madrid, Spain) with a 7.5-MHz probe for these assessments. We also scanned common carotid, internal carotid, external carotid, and vertebral arteries scanned for the presence of carotid plaques (CP), defined as IMT ≥1.5 mm and protruding into the lumen.

Blood pressure measuring and 24-hour ambulatory blood pressure monitoring

Measurements of BP were taken in the morning before breakfast and after the subject had rested 10 min. Systolic and diastolic BP values at resting were measured in a silent room three times using a stethoscope and an aneroid sphygmomanometer placed at the height of heart in the non-dominant arm. ²⁹ The average of the three measurements was used as an estimation of office systolic and diastolic BP readings.

All studied patients underwent 24 h ABPM. The device (A&D TM2430EX oscillometric device. A&D Co., Ltd., Tokyo, Japan) was set to obtain BP readings at 30-minute intervals during the day (between 6:00 a.m. and 11:59 p.m.) and at 60-minute intervals during the night (between 12:00 a.m. and 05:50 a.m.). Individuals were instructed to continue with their normal daily activities during the day. Following the European Society of Hypertension guidelines, ³⁰ correct ABPM required valid measurements (i) in at least 70% of expected measurements during the 24-h period; (ii) 20 awake and seven asleep measurements; and (iii) at least two daytime and one nighttime measurement per hour. A “dipping” pattern required a decrease of at least 10% in mean BP during the nighttime period with respect to daytime recordings; “non-dipping” pattern was defined as a nocturnal mean BP fall from 1 to 9% from daytime values, while the absence of any decrease in mean BP during nighttime indicated “reverse dipping”. ³¹

Sample size calculation

For the current study, we used the online sample size and power calculator provided by the Institute Municipal d’Investigació Mèdica from Barcelona, Spain, version 7.12 (https://www.inim.cat/ofertadeserveis/sofware-public/granmo/). Setting alpha at 0.05 and beta at 0.2 for a one-sided test, the inclusion of at least 28 T1D patients with CAN and AS (cases) and 24 similar patients without AS (controls) would allow to recognize a double prevalence of non-dipping pattern in cases compared with controls as statistically significant, taking into account the 28% overall prevalence of non-dipping pattern in patients with T1D and CAN already published by Duvnjack et al. ⁶

Statistical analysis

Data are shown as the means ± standard deviation and 95% confidence interval (95CI) [lower limit to upper limit] for continuous variables, and counts (%) for categorical variables. For continuous variables, we checked normality using the Kolmogorov-Smirnov test, and ensured normality as needed by applying logarithmic transformation. We applied nonparametric tests to variables that did not follow the normal distribution even after transformation. To compare continuous variables, Student t and Mann Whitney U tests for unpaired comparisons, as appropriate. Comparisons of discrete variables among study subgroups used χ², Fisher’s exact test, or logistic binary regression. We used Spearman correlation analysis to evaluate the association between ABI, CAN indexes, and 24-hour ABMP recordings. To perform Spearman correlation, we used the highest ABI results, as reported in our previous report. ¹² Finally, full and stepwise (probability for entry ≤0.05, probability for removal ≤0.10) multiple linear regression models were performed to ascertain the main determinants of the decrease in nocturnal mean BP expressed in percentage of daytime values, introducing as independent variables those showing statistically significant correlations in univariate analyses. The level of statistical significance was set a p value <0.05. The PASW Statistics 18.0 software package (SPSS, Chicago, IL) was used for all calculations.

Results

Study population characteristics

Data from 56 patients with T1D and CAN were reviewed, but four patients were excluded and submitted to full evaluation by the Department of Vascular Surgery because of a reduced ABI (≤0.9). Hence, a total of 52 patients were finally included in the study. Their demographic and clinical characteristics are detailed in the Table 1. All patients were under chronic intensive insulin treatment with 43 patients (83%) receiving multiple insulin dose regimens, and nine patients (17%) using continuous subcutaneous insulin infusion.

Table 1.

Demographic and clinical characteristics of all patients and as a function of the presence or absence of arterial stiffness, defined by an ankle-brachial index >1.2.

	All patients	Arterial stiffness		p
Variable	(n = 52)	Yes (n = 28)	No (n = 24)
Demographics
Age (years)	40 ± 11 [37 to 43]	43 ± 11 [39 to 47]	36 ± 10 [32 to 40]	0.023
Sex (male)	28 (54) [41 to 67]	17 (61) [42 to 76]	11 (50) [28 to 65]	0.283
Duration of diabetes (years)	22 ± 10 [19 to 25]	25 ± 10 [21 to 29]	19 ± 10 [15 to 23]	0.013
Comorbid conditions
Microangiopathy, n (%)	20 (39) [26 to 52]	14 (50) [32 to 67]	6 (25) [12 to 45]	0.065
- Polyneuropathy	5 (10) [26 to 52]	4 (14) [6 to 46]	1 (4) [1 to 20]	0.358
- Nephropathy	9 (17) [9 to 30]	8 (29) [15 to 47]	2 (20) [2 to 26]	0.086
- Retinopathy	14 (27) [17 to 40]	9 (32) [18 to 51]	5 (21) [10 to 41]	0.359
Macroangiopathy, n (%)	3 (6) [2 to 16]	3 (11) [4 to 27]	0 (0) [0 to 14]	0.099
Hypertension, n (%)	17 (33) [22 to 46]	11 (39) [21 to 54]	6 (25) [15 to 49]	0.376
Smoking, n (%)	25 (48) [35 to 61]	12 (43) [27 to 61]	13 (54) [35 to 72]	0.416
Mean BMI (kg/m2)	27 ± 5 [25 to 28]	27 ± 4 [25 to 28]	26 ± 5 [24 to 29]	0.761
Overweight/Obesity, n (%)	21 (40)/10 (19) [28 to 54]/[11 to 32]	13 (46)/6 (21) [30 to 64]/[10 to 40]	8 (33)/4 (17) [18 to 53]/[7 to 36]	0.611
Body fat percentage (%)	26 ± 10 [23 to 29]	27 ± 9 [23 to 30]	25 ± 11 [20 to 30]	0.569
Waist circumference (cm)	89 ± 15 [85 to 93]	91 ± 13 [86 to 96]	88 ± 17 [81 to 95]	0.523
Biochemical variables
Fasting plasma glucose (mg/dL)	176 ± 80 [154 to 198]	158 ± 72 [130 to 186]	197 ± 84 [162 to 232]	0.059
Total cholesterol (mg/dL)	169 ± 29 [152 to 168]	167 ± 35 [153 to 181]	172 ± 20 [164 to 180]	0.419
HDL-cholesterol (mg/dL)	53 ± 16 [49 to 57]	53 ± 15 [47 to 58]	54 ± 17 [47 to 61]	0.811
LDL-cholesterol (mg/dL)	99 ± 25 [92 to 106]	97 ± 28 [86 to 108]	101 ± 20 [93 to 109]	0.183
Triglycerides (mg/dL)	83 ± 38 [72 to 94]	86 ± 46 [68 to 104]	80 ± 24 [70 to 90]	0.890
eGFR (mL/min/1.73 m2)	91 ± 24 [84 to 98]	85 ± 22 [76 to 94]	98 ± 25 [87 to 109]	0.024
UACR (mg/g)	337 ± 2,045 [0 to 9,060]	609 ± 2,780 [0 to 16,860]	20 ± 51 [20 to 42]	0.533
A1c (%)	7.9 ± 1.5 [7.5 to 8.3]	7.7 ± 1.2 [7.2 to 8.2]	8.4 ± 1.8 [7.6 to 9.2]	0.107
eGDR (mg/kg/min)	11.5 ± 2.5 [10.8 to 12.2]	11.4 ± 2.3 [10.5 to 12.3]	11.6 ± 2.7 [11.5 to 12.7]	0.546
Inflammatory biomarkers
hs-C reactive protein, mg/L	3.8 ± 3.7 [2.8 to 4.8]	3.6 ± 3.2 [2.4 to 4.8]	4.0 ± 4.3 [2.2 to 5.8]	0.926
ESR, mm/h	17 ± 17 [12 to 22]	19 ± 22 [10 to 28]	15 ± 9 [11 to 19]	0.923
Homocysteine, μmol/L	12 ± 20 [6 to 18]	10 ± 3 [8 to 12]	15 ± 30 [2 to 28]	0.307
Fibrinogen, mg/dL	366 ± 85 [342 to 390]	371 ± 100 [332 to 410]	360 ± 62 [334 to 386]	0.955
Concomitant medications
Insulin therapy with MDI, n (%)	43 (83) [70 to 91]	21 (75) [57 to 87]	22 (92) [74 to 98]	0.113
Total insulin dose (IU)	49 ± 20 [43 to 55]	49 ± 18 [42 to 56]	49 ± 22 [40 to 58]	0.840
Insulin dose (IU/kg/day)	0.64 ± 0.23 [0.58 to 0.70]	0.62 ± 0.22 [0.53 to 0.71]	0.65 ± 0.25 [0.54 to 0.76]	0.633
Antiaggregant therapy, n (%)	4 (8) [3 to 17]	3 (11) [4 to 27]	1 (4) [1 to 20]	0.615
Statin therapy, n (%)	27 (52) [39 to 65]	18 (64) [46 to 79]	10 (42) [25 to 61]	0.054
Antypertensive therapy, n (%)	17 (33) [22 to 46]	11 (39) [24 to 58]	6 (25) [12 to 45]	0.131
- ACEI OR ARB, n (%)	15 (88) [66 to 97]	9 (82) [52 to 95]	6 (100) [61 to 100]
- Diuretics, n (%)	4 (24) [10 to 47]	2 (18) [5 to 48]	2 (33) [10 to 70]
- Beta-blockers, n (%)	2 (12) [3 to 34]	2 (18) [5 to 48]	0 (0) [0 to 40]
- Calcium channel blockers, n (%)	2 (12) [3 to 34]	1 (9) [2 to 38]	1 (17) [3 to 56]
Monotherapy	12 (71) [47 to 87]	8 (73) [43 to 90]	4 (66) [30 to 90]	0.826
Dual therapy	4 (24) [10 to 47]	2 (18) [5 to 48]	2 (33) [10 to 70]
Triple therapy	1 (6) [1 to 27]	1 (9) [2 to 39]	0 (0) [0 to 40]
Once/twice daily doses	14 (82)/3 (18) [59 to 94]/[6 to 41]	10 (91)/1 (9) [62 to 98]/[2 to 38]	4 (67)/2 (33) [30 to 90]/[10 to 70]	0.168
24-hour ambulatory blood pressure recordings
Mean SBP (mmHg)	129 ± 13 [125 to 133]	129 ± 13 [124 to 134]	119 ± 17 [112 to 126]	0.024
Mean DBP (mmHg)	76 ± 14 [72 to 80]	78 ± 16 [72 to 84]	74 ± 10 [70 to 78]	0.280
Mean HR (bpm)	81 ± 13 [77 to 85]	80 ± 12 [75 to 85]	82 ± 14 [76 to 88]	0.761
Non-dipping of BP	28 (54) [41 to 67]	13 (46) [30 to 64]	15 (63) [43 to 79]	0.304
Mean decrease in nighttime SBP (%)	7.6 ± 7.6 [5.4 to 9.7]	9.1 ± 8.7 [5.7 to 12.5]	5.9 ± 6.0 [3.2 to 8.6]	0.125
Mean decrease in nighttime DBP (%)	11.7 ± 8.4 [9.4 to 14]	12.8 ± 7.9 [9.7 to 15.9]	10.4 ± 9.1 [6.6 to 14.2]	0.330

Continuous and discrete variables are shown as means ± standard deviation and raw numbers (percentage), respectively. 95% CI [lower limit to upper limit] are given. Differences between groups are assessed by Student t test, χ² or Fisher exact tests as appropriate.

Abbreviations: ACEI: angiotensin-converting enzyme inhibitor; ARB: angiotensin receptor blocker; BP: blood pressure; DBP: diastolic blood pressure; eGFR: estimated glomerular filtration rate (MDRD-4 formula); eGDR: estimated glucose disposal rate; HDL: high density-lipoprotein; hs-C reactive protein: high-sensitivity C reactive protein; HR: heart rate; LDL: low density-lipoprotein; MDI: multiple doses of insulin; SBP: systolic blood pressure UACR: urinary albumin–to–creatinine ratio.

Seventeen patients (33%) were taking antihypertensive drugs: 15 (29%) were on an angiotensin-converting enzyme inhibitor (ACEI) or angiotensin receptor blocker (ARB), 4 (8%) were on diuretics, 3 (6%) on beta-blockers, and 2 (4%) on calcium channel blockers. Twelve (23%) of the patients were on monotherapy, four patients (8%) received dual therapy, and only one patient (2%) used triple drug therapy. An ACEI or ARB plus a diuretic were the most used therapy combination. Fourteen (82%) patients were on once daily antihypertensive medication (morning), while only 3 (18%) patients were taking their antihypertensive medication bis in die (morning-evening/dinner). Finally, 27 (52%) patients were on statins and 4 (8%) were taking antiaggregant drugs.

Considering all patients as a whole, total insulin dose correlated negatively with eGDR (ρ = −0.391, p = 0.003), and positively with BMI (ρ = 0.316, p = 0.018), and with waist circumference (ρ = 0.473, p < 0.001). Similarly, the BMI correlated negatively with eGDR (ρ = −0.623, p < 0.001) and positively with waist circumference (ρ = 0.837, p < 0.001).

An ABI >1.2 indicating AS was found in 28 (54%) of the 52 patients. These patients were older, had a longer duration of the disease, and higher office systolic BP than the 24 (46%) patients not showing AS (Table 1).

Among the 28 T1D patients with CAN and peripheral AS detected by ABI screening, 18 showed evidence of subclinical atherosclerosis at lower extremity and/or carotid levels using full vascular ultrasound examination. PAD was found in 14 (27%), and CP was found in 6 (12%). Two patients presented concomitant PAD and CP. No patients needed revascularization procedures. The mean cIMT among patients showing an ABI >1.2 was 0.68 ± 0.15 mm.

Cardiovascular autonomic dysfunction

Following inclusion criteria, all patients included in this study had asymptomatic CAN as defined. CAN was categorized as early/mild in 46 subjects (88%) and definite in six (12%), but none showed orthostatic hypotension or severe disease. Among patients with AS, 25 subjects (89%) had early/mild grade of CAN, and three had definite CAN (11%). Finally, we found resting tachycardia only in one patient (4%) with a normal ABI and in two patients (7%) with AS. There were no statistically significant differences in any of these frequencies between subjects with or without AS.

Considering all patients as a whole, E/I index correlated negatively with age (ρ = −0.470, p < 0.001), onset of diabetes (ρ = −0.503, p < 0.001), and with A_1c levels (ρ = −0.293, p = 0.035). Similarly, the VAL index correlated negatively with A_1c levels (ρ = −0.323, p = 0.020). There were no differences in indexes of parasympathetic function in patients with AS compared with those showing normal ABI values (data not shown). Nonetheless, patients with concomitant AS and evidence of subclinical atherosclerosis showed a reduced 30:15 ratio compared with those subjects with AS but without atherosclerosis (4 ± 11% vs 18 ± 11%, respectively, p = 0.008) (Figure 1, panel A). In consonance, cIMT correlated negatively with VAL index (ρ = −0.466, p = 0.016).

Figure 1.

Heart rate variability to orthostatism (30:15 ratio) in patients with concomitant arterial stiffness and evidence of subclinical atherosclerosis compared with those subjects with arterial stiffness but without atherosclerosis (panel A). Heart rate (HR) change to active standing in patients with arterial stiffness and those with normal ankle-brachial index (ABI) (panel B).

Finally, regarding sympathetic function indexes, the percentage of HR change in response to active standing was reduced in patients with AS compared with those with normal ABI values (7 ± 11% vs 10 ± 11%, respectively, p = 0.025) (Figure 1, panel B).

Main determinants of abolition of the physiological nocturnal decrease in BP in 24-hour ABPM

All 52 patients had valid ABPM recordings. In our cohort of patients with T1D and asymptomatic CAN, the prevalence of a non-dipping pattern was 54% (28 patients). There were no statistically significant differences in the distribution of non-dippers among patients with AS (13 out of 28, 46%) or without AS (15 out of 24, 63%). When ABI values were analyzed as a continuous variable, no correlation with the nocturnal changes in BP were observed, either. There were not demographic, clinical or biochemical differences between the 24 dipper and the 28 non-dipper patients in our series (Table 2). As expected, non-dippers had higher nocturnal mean BP, systolic and diastolic BP during the nighttime period compared with dipper patients (Table 2).

Table 2.

Values derived of the 24-hour ambulatory blood pressure recordings in patients with dipping and non-dipping patterns.

Variable	Dipping pattern (n = 24)	Non-dipping pattern (n = 28)	p
24-hour ambulatory blood pressure recordings
Nighttime mean BP (mmHg)	74 ± 6 [71 to 77]	83 ± 8 [80 to 86]	<0.001
Nighttime systolic BP (mmHg)	104 ± 8 [101 to 108]	114 ± 11 [110 to 118]	0.001
Nighttime diastolic BP (mmHg)	57 ± 14 [51 to 63]	67 ± 7 [64 to 70]	0.001
Clinical characteristics
Age (years)	38 ± 12 [33 to 43]	42 ± 9 [38 to 46]	0.270
Duration of diabetes (years)	20 ± 10 [16 to 24]	24 ± 10 [20 to 28]	0.230
Sex (male)	15 (54) [36 to 70]	13 (57) [37 to 74]	0.833
A1c (%)	7.6 ± 1.0 [7.2 to 8.0]	8.4 ± 1.8 [7.7 to 9.1]	0.103
Microangiopathy (%)	9 (32) [18 to 51]	10 (43) [26 to 63]	0.405
Hypertension (%)	6 (25) [21 to 54]	11 (39) [16 to 51]	0.274
BMI (kg/m2)	27 ± 5 [25 to 29]	27 ± 5 [24 to 29]	0.919
Pharmacotherapy
Total insulin dose (IU)	52 ± 18 [45 to 59]	46 ± 21 [37 to 55]	0.388
Insulin dose (IU/kg/day)	0.67 ± 0.23 [0.58 to 0.76]	0.59 ± 0.23 [0.49 to 0.69]	0.255
ACEI or ARB, n (%)	6 (100) [61 to 100]	9 (82) [52 to 95]	0.770
Diuretics, n (%)	2 (33) [10 to 70]	2 (18) [5 to 48]
Beta-blockers, n (%)	0 (0) [0 to 40]	2 (18) [5 to 48]
Calcium channel blockers, n (%)	0 (0) [0 to 40]	2 (18) [5 to 48]
Monotherapy	4 (67) [30 to 90]	8 (73) [43 to 90]	0.576
Dual therapy	2 (33) [10 to 70]	2 (18) [5 to 48]
Triple therapy	0 (0) [0 to 40]	1 (9) [2 to 38]
Once/twice daily doses	6 (100)/0 (0) [61 to 100]/[0 to 40]	8 (73)/3 (27) [43 to 90]/[10 to 57]	0.459

Data are shown as counts (%) for discrete variables mean ± SD for continuous variables. 95% CI [lower limit to upper limit] are given. p value is shown for the comparison between patients with dipping and non-dipping pattern.

Abbreviations: ACEI: angiotensin-converting enzyme inhibitor; ARB: angiotensin receptor blocker; BMI: body mass index; BP: blood pressure; DBP: diastolic blood pressure; glycated hemoglobin, A_1c; SBP: systolic blood pressure.

VAL index correlated inversely with nighttime mean HR (ρ = −0.294, p = 0.036). Postural test (30:15 ratio) correlated directly with the decrease in nighttime diastolic BP with respect to daytime expressed as percentage (ρ = 0.308, p = 0.032). The percentage decrease in mean BP during the nighttime was also similar between subjects with or without AS (Figure 2).

Figure 2.

Decrease in mean blood pressure (BP) during the nighttime in subjects with arterial stiffness compared with those showing a normal ankle-brachial index (ABI).

Regarding inflammatory biomarkers, the changes in nighttime mean systolic BP with respect to daytime values negatively correlated with ESR (ρ = −0.287, p = 0.043), whereas the changes in nighttime mean diastolic BP values negatively correlated with fibrinogen levels (ρ = −0.318, p = 0.023).

Considering all patients as a whole, mean daytime systolic BP directly correlated with age (ρ = 0.292, p = 0.036), waist circumference (r = 0.331, p = 0.018), and BMI (r = 0.348, p = 0.036). Similarly, average daytime mean BP directly correlated with waist circumference (r = 0.329, p = 0.018) and BMI (r = 0.294, p = 0.034). In the same way, nighttime mean BP directly correlated with waist circumference (r = 0.304, p = 0.032) and BMI (r = 0.337, p = 0.016). The decrease in nighttime mean BP expressed as percentage of daytime values positively correlated with age (r = 0.277, p = 0.049) and negatively with A_1c levels (r = −0.318, p = 0.023).

Age, waist circumference, BMI, and A_1c levels were introduced as independent variables into a multiple regression model. The stepwise model (R²: 0.113, p = 0.016) only retained A_1c levels (β: −0.333, 95CI: −3.103 to −0.334) as significant determinant of the percentage of decrease in nocturnal mean BP in T1D patients with asymptomatic CAN (Figure 3).

Figure 3.

Results of a multiple regression model introducing the nighttime decrease in blood pressure (BP) expressed as percentage of daytime values as dependent variable, and age, waist circumference, body mass index, heart rate variability to orthostatism (30:15 ratio), and glycated hemoglobin (A_1c) levels as independent variables.

Discussion

Our study assessed the relationship between the presence of AS and circadian changes in BP under ambulatory conditions in individuals with T1D and subclinical CAN for the first time. The present study provides new insights to published research about the potential association between autonomic dysfunction, AS, and circadian blood pressure regulation in patients with T1D. In our study, we were not able to corroborate any significant association between the presence of AS and non-dipping pattern in 24-hour BP in our sample of adult patients with T1D and subclinical CAN. However, a non-dipping pattern in our cohort was associated with worse metabolic control, regardless of the presence of SA, highlighting the importance of the glycemic control in these patients.

Our results show that the non-dipping pattern in 24-hour BP regulation being a frequent condition in patients with T1D presenting evidence of CAN, even at a subclinical stage. ¹² This finding is in agreement with the well-documented role of cardiovascular autonomic dysfunction on the development of abnormalities in BP variability.^3,32 The physiopathology underlying the relationships between AS, CAN, and circadian rhythms in BP is not straightforward. The autonomic nervous system regulates BP variability, but individual’s features, environmental conditions, and drugs may also modulate circadian changes. Likewise, the autonomic nervous system, by means of vessel blood flow and BP control, plays a significant role in the development of AS. ¹⁴ On the other hand, vascular aging-induced AS increases vascular angiotensin 1-receptor levels. Angiotensin one enhances sympathetic activity disrupting autonomic function, which results in an abnormal circadian pattern of BP. Furthermore, reduced compliance of large arteries may compromise the sensitivity of arterial baroreceptors resulting in an abnormal BP variability.^16,17,33,34 In keeping with this, increased BP variability has been associated with moderate AS in the aorta, but not with carotid AS. ¹⁵ A study in 548 patients presenting with high-normal BP or hypertension, the presence of AS was closely related with the non-dipping pattern in BP ¹⁷ In short, AS would seem to be related with both cardiovascular autonomic dysfunction and circadian abnormalities in BP.^11,16,35,36

However, although sympathetic response was lower in those participants with AS compared to their counterparts with a normal ABI, we were not able to corroborate any significant association between the presence of AS and non-dipping patterns in BP in our sample of adult patients with T1D and subclinical CAN. We speculate that the presence of cardiovascular autonomic dysfunction may mask the effects of AS on BP. A priori, the possibility exists that the presence of cardiovascular autonomic dysfunction ‒ even at subclinical stages ‒ commands the loss of the circadian pattern in BP, with AS becoming a vascular consequence of CAN. Several tests of cardiovascular autonomic dysfunction correlated with nighttime BP variability, suggesting that an early detection of subclinical cardiovascular autonomic dysfunction may help in distinguishing those patients who are at higher risk of being non-dippers, and therefore, of developing microvascular complications.^2-6,37 In this line, there is extensive evidence indicating that CAN is a factor of paramount importance in impairing dipping at night in patients with T1D.^6,8,37 However, the influence of CAN on the association between non-dipping patterns and AS has not been addressed in prior studies, despite both AS and CAN frequently coexisting in patients with diabetes. ¹² CAN may favor AS by increasing HR, because any increase in HR per se might result in AS. ³⁸ In addition, cardiovascular autonomic dysfunction may impair the elasticity of the arterial wall by modulating the vascular tone of large vessels. ³⁹ If this was the case, AS would become a secondary event in these patients, since CAN would contribute to both circadian BP deregulation and AS development.

The effect of the administration of diverse antihypertensive drugs in patients with CAN was studied. However, there is no conclusive evidence for CAN-modifying pharmacological therapy. Angiotensin-converting enzyme inhibitors and angiotensin receptor blockers (quinapril, trandolapril, losartan) had inconsistent efficacy, and β-blockers (metoprolol) were effective in one study on heart rate variability in patients with CAN. All these findings, however, are insufficient to reach recommendations on these agents with the current guidelines. ⁴⁰

Data about the impact of glycemic control and metabolic parameters on the circadian regulation in BP and cardioautonomic function of patients with T1D are very scarce. ⁷ A positive correlation between night-to-day ratios of mean BP and A_1c values has been reported, suggesting that smoothing of circadian BP variation is more frequent in patients with worse glycemic control. ⁴¹ The interplay between impaired baroreflex sensitivity as a measure of autonomic neuropathy with A_1c and daily insulin doses have been also addressed in children and adolescents with T1D. ⁴² These young patients with T1D displayed significantly less variance of BP, and baroreflex sensitivity decreased in parallel to increasing A_1c and daily insulin doses. ⁴² In children with T1D, insulin dose and A_1c were positively associated with nighttime BP recordings, and negatively with both systolic and diastolic nocturnal dipping, suggesting that both hyperglycemia and hyperinsulinemia may blunt the physiological fall in BP during nighttime by different pathways. ⁴ In the same line, the nocturnal decline in systolic and diastolic BP values appears to be lower in patients on intensive insulin therapy compared with those on a conventional insulin regimen. ⁴³ High insulin doses in the context of poor glycemic control would act as a central sympathetic activator. ⁷ Nevertheless, this possibility must be considered with caution. Insulin also has hypotensive properties ‒ a phenomenon known as ‘‘insulin-induced hypotension’’ ‒ especially in patients with T1D and CAN. ⁴⁴ To this regard, sympathetic denervation due to the presence of cardiovascular autonomic dysfunction might avoid insulin-induced sympathetic activation and vasoconstriction during hyperinsulinemia. ⁴⁵ In conceptual agreement, we did not find any significant differences in insulin doses when our patients were compared according their dipping status, a finding that is consistent with previous studies.^46-49

Finally, we also found negative associations between some inflammatory markers and the decrease in BP during nighttime in our patients with D1M. In other words, a higher level of subclinical inflammation was associated with a lower decrease in nocturnal blood pressure. This finding not surprising is in agreement with previous reports in hypertensive patients from the general population, ⁵⁰ and translates the effect of an underlying chronic inflammatory state such as diabetes on vascular endothelium and hemorreologic^51–54 parameters.

Our study is not free of limitations: (i) our relatively small sample size precludes us to rule out small differences in the primary objective; (ii) the study design does not include a control group without CAN, (iii) our study design excluded subjects with severe microvascular and macrovascular complications, favoring survivorship bias; (iv) we did no assess gold standard noninvasive measures of AS such as pulse wave velocity; (v) no participant presented with severe CAN, overt neurogenic orthostatic hypotension, or supine hypertension, all conditions with a severe impact on day-to-night rations in BP; (vi) we did not check interstitial glycemia by continuous glucose monitoring (CGM) systems for hyperglycemic or hypoglycemic fluctuations during 24h-ABPM. A very high or low glucose level may disturb body fluid homeostasis or induced adrenergic activation, respectively, and therefore, modifying 24 h BP recordings; and (vii) vascular function was not explored thoroughly in the subset of patients with normal ABI.

In conclusion, many factors may act synergistically to disrupt the circadian BP pattern in patients with T1D and CAN, even though the presence of AS does not appear to be a significant one. Chronic glycemic control may play a role on the circadian BP variability in patients with CAN. Accordingly, screening of microvascular complications in T1D should be directed towards earlier stages in which CAN may be present but end-organ damage has not occurred yet. Prospective studies are further needed in order to clarify the causal relationship, if any, between peripheral AS and loss of circadian BP pattern in patients with T1D and CAN. Also, well-planned research is warranted in order to ascertain the effect of glycemic control and exogenous insulin administration on the BP of patients with T1D, since solid evidence is lacking.

ORCID iD

Lía Nattero-Chávez https://orcid.org/0000-0002-9758-9397