Impact of disease duration and cardiovascular dysautonomia on hypertension in Parkinson's disease

Abstract

The authors evaluated the association of Parkinson’s disease (PD) duration with hypertension, assessed by office measurements and 24‐hour (ambulatory) monitoring, in 167 patients. Hypertension was evaluated through both office and ambulatory blood pressure (BP) measurements. Among participants (mean age 73.4±7.6 years; 35% women), the prevalence of hypertension was 60% and 69% according to office and ambulatory BP measurements, respectively (Cohen's k=0.61; P<.001). PD duration was inversely associated with hypertension as diagnosed by office measurements (odds ratio [OR], 0.92; 95% confidence interval [CI], 0.86–0.98) but not by ambulatory monitoring (OR, 0.94; 95% CI, 0.81–1.01). Ambulatory BP patterns showed higher nocturnal BP among patients with long‐lasting disease. In conclusion, ambulatory BP monitoring improves the detection of hypertension by 15% in PD, compared with office evaluation. The likelihood of having hypertension does not decrease during the PD course; rather, BP pattern shifts towards nocturnal hypertension.

1. Introduction

Cardiovascular diseases and Parkinson's disease (PD) are common, and frequently coexist in the older population. These conditions predispose to disability and often trigger hospital admissions and death.1 However, the relationship between PD and cardiovascular burden is quite elusive. For example, low blood pressure (BP) and low serum lipids have been associated with higher risk of PD.2 Conversely, high BP and smoking habit have been sometimes reported as protective against PD.3, 4 Accordingly, some authors have found a more favorable cardiometabolic profile, including a lower prevalence of hypertension, in PD patients as compared with general populations. In addition, a negative association between prevalent hypertension and PD duration has been previously described.5

PD is characterized by peculiar BP abnormalities. Autonomic dysfunction accounts for some of the most frequent and disabling non‐motor features of PD, such as postural hypotension, postprandial hypotension, and supine/nocturnal hypertension, determining a high variability of BP values during 24 hours.6, 7 Notwithstanding, in clinical practice, the assessment of hypertension relies on office BP measurements, which do not take into account such a high circadian variability.

In order to verify the previously reported negative association between hypertension and PD duration, we investigated the relationship between PD duration and prevalent hypertension, verified through both office measurements and 24‐hour (ambulatory) monitoring. We also assessed the agreement between the two methods and the profile of the circadian BP pattern across different disease duration stages.

2. Methods

2.1. Participants

In this cross‐sectional study we evaluated all PD patients consecutively admitted to the geriatric day hospital of the Catholic University of Rome, Italy, between 2013 and 2015. Unwillingness to take part in the study was the only exclusion criterion. Diagnosis of PD was based on the United Kingdom Parkinson's Disease Society Brain Bank criteria.8 Of 175 patients, three refused to participate, and five were excluded for incomplete data, thus leaving a sample of 167 participants. All participants provided written informed consent. Our study complies with the Declaration of Helsinki.

2.2. Definitions of hypertension

2.2.1. Office assessment

All participants underwent BP measurement after being quietly seated for at least 5 minutes on a chair, with their arms supported at the level of the heart. According to European Society of Cardiology criteria, hypertension was defined as a BP ≥140/90 mm Hg or by treatment with antihypertensive agents. BP was measured three times during the visit, in standard conditions, according to international guidelines recommendations9; the average value was considered for the diagnosis. A manual sphygmomanometer was used, with appropriately sized aneroid cuffs. Systolic and diastolic BPs were recorded as Korotkoff phase I and phase V, respectively. Use of any angiotensin‐converting enzyme inhibitors (ACEIs), angiotensin receptor blockers (ARBs), calcium channel blockers (CCBs), β‐blockers (BBs), thiazides (TZs), clonidine, and α‐blockers was considered for the diagnosis of hypertension.9

2.2.2. Ambulatory monitoring

All participants underwent 24‐hour BP monitoring using a validated and regularly calibrated oscillometric portable monitor. In keeping with European Society of Cardiology guidelines, BP recordings started in the morning (between 8 am and 10 am) and lasted for 24 hours. BP was measured every 15 minutes during the daytime and every 30 minutes during the night. To minimize unsuccessful measurements, the aneroid cuff was applied to the arm presenting less motor symptoms. A 24‐hour monitoring period was considered reliable and included in the analyses only when ≥70% of the measurements were valid, with a minimum of 20 measurements for daytime and seven for nighttime. During monitoring, participants were encouraged to perform their usual activities, paying attention to stop every movement and to relax the arm by the side during cuff inflation. All activities, events, and symptoms, as well as bedtime and awaking, were properly reported on a dedicated form. Systolic and diastolic BP values were provided for the whole session (24 hours), for the daytime and overnight. Hypertension was defined as a mean BP over the 24 hours >130/80 mm Hg or by antihypertensive drug treatment.10

2.3. Covariates

Data on age, sex, smoking habit, current use of alcohol, and anthropometric parameters were collected through a dedicated questionnaire. Smoking status was reported as total lifetime cigarette‐packs per year for actual and former smokers. Body mass index (BMI) was calculated as body weight divided by squared height and expressed in kg/m². Time from the diagnoses of PD was collected and expressed in years, then categorized according to tertiles (<2, 2–6, >6 years). The clinical severity of PD was assessed through the motor section of the Unified Parkinson's Disease Rating Scale (UPDRS III).11 All participants were evaluated during their “on” period, within 2 hours from the last antiparkinsonian drug administration. The L‐dopa equivalent daily dose was obtained and indexed for body weight, transforming the antiparkinsonian drugs daily dose in L‐dopa equivalents through a validated algorithm.12 Drugs were coded according to Anatomical Therapeutic and Chemical codes. All diagnoses were verified by the physicians and coded according to the International Classification of Diseases, Ninth Edition.

2.4. Statistical analysis

Variables are reported as mean±standard deviation or frequencies (percentages), according to the tertiles of disease duration. Total lifetime cigarette‐packs per year was analyzed after log transformation because of skewed distribution. Analysis of variance was used to compare normally distributed continuous variables. The two‐tailed Fisher exact test was used for dichotomous variables. The association (odds ratio [OR], 95% confidence interval [CI]) between diagnosis of hypertension (dependent variable) and PD duration (years) was assessed in separate models of logistic regression (crude, age‐ and sex‐adjusted, and fully adjusted). Variables distributed differently (P<.05) in the three tertiles of PD duration at the univariate analysis were considered as potential confounders and were first entered into a fully adjusted regression model, along with age and sex. A second fully adjusted regression model included the same covariates plus UPDRS III, BMI, chronic kidney disease, chronic obstructive pulmonary disease, diabetes, heart failure, coronary artery disease, and hypothyroidism. Sensitivity analysis was also performed after excluding patients with conditions that might potentially be associated with hypotension in the long run (eg, heart failure and coronary artery disease). Agreement between the diagnosis of hypertension as diagnosed by the two methods was assessed using the Cohen's coefficient (k), where: k<0.40 suggests poor agreement, k included in the range 0.40–0.75 suggests fair to good agreement, and k>0.75 suggests excellent agreement. Positive percent agreement and negative percent agreement were also computed. The adjusted ambulatory BP mean values (diurnal and nocturnal) were finally compared across tertiles of disease duration, using analysis of covariance. Analyses were performed with SPSS for Windows 18.0 (SPSS Inc, Chicago, IL, USA).

3. Results

Among the 167 participants the mean age was 73.4±7.6 years, and 59 (35%) were women. As shown in Table 1, patients with longer PD duration were more likely to be treated with higher doses of dopaminergic agents. Within the subsample diagnosed with hypertension by office BP measurements, the mean number of antihypertensive drugs decreased across the tertiles of PD duration (from 1.7±0.4 to 1.0±0.3; P=.008). No other demographic, anthropometric, or clinical differences arose among the three groups.

Table 1.

Sample Characteristics By Parkinson's Disease Duration Tertiles

	Disease Duration			P Value
	Tertile I n=61 (37%)	Tertile II n=54 (32%)	Tertile III n=52 (31%)
Demographics
Age	73.7±7.3	74.9±7.7	71.7±6.9	.073
Sex (female)	22 (36)	20 (37)	16 (31)	.765
Education, y	10.3±4.9	12.3±5.1	11.0±5.1	.537
Logarithm total lifetime cigarette‐packs per y	2.5±1.5	2.5±1.7	2.1±1.6	.428
Alcohol use (actual)	33 (54)	32 (59)	30 (58)	.847
BMI, kg/m2	27.3±3.7	27.0±4.6	27.9±4.8	.551
Parkinson's disease characteristics
UPDRS III	21.9±9.4	25.1±8.8	25.4±9.9	.091
LEDD, kg	4.8±3.3	9.0±5.6	11.1±4.6	<.001
Comorbidities
Osteoarthritis	40 (66)	35 (65)	29 (56)	.505
Hypertension (office assessment)	44 (72)	32 (59)	26 (50)	.049
Hypertension (24‐h assessment)	49 (80)	37 (69)	29 (56)	.019
Chronic kidney disease	15 (21)	21 (39)	13 (25)	.173
Diabetes	14 (23)	7 (13)	7 (14)	.267
COPD	11 (18)	6 (11)	3 (6)	.131
Hypothyroidism	11 (18)	3 (6)	10 (15)	.131
Gastritis	9 (15)	11 (20)	7 (14)	.584
Coronary artery disease	5 (8)	5 (9)	6 (12)	.831
Heart failure	3 (5)	2 (4)	2 (4)	.938
Medications
BP‐lowering drugs (count)a	1.7±0.9	1.4±0.8	1.0±0.8	.008

Figures are given as mean±standard deviation (SD) or count and percentage (%).

Abbreviations: BMI, body mass index; BP, blood pressure; COPD, chronic obstructive pulmonary disease; LEDD, L‐dopa equivalent daily dose; UPDRS, Unified Parkinson's Disease Rating Scale.

^aExcluding patients without hypertension according to office assessment (n=102).

Hypertension was diagnosed in 103 (60%) patients by office BP measurements and in 118 (69%) by ambulatory monitoring. Between the two methods, the positive percent agreement was 81%, the negative percent agreement was 85%, and the Cohen's k was 0.61 (P<.001), suggesting a fair‐good agreement. As shown in Figure 1, the prevalence of hypertension, as defined by both methods, decreased across tertiles of PD duration (from 72% to 50% for office assessment and from 80% to 56% for ambulatory BP monitoring [ABPM], P<.05 for both trends). Use of ARBs, BBs, and CCBs presented a nonsignificant decreasing trend along the course of PD. Conversely, no trend was evident for ACEIs and TZs (Fig. S1).

Figure 1.

Prevalence of hypertension assessed according to different criteria across tertiles of disease duration.

According to multivariate analysis (Table 2), PD duration was negatively associated with hypertension, as diagnosed according to the office measurements (fully adjusted OR, 0.92; 95% CI, 0.86–0.98), but not associated with hypertension defined according to the ambulatory monitoring (fully adjusted OR, 0.94; 95% CI, 0.81–1.01). The sensitivity analysis, conducted after excluding patients with heart failure and coronary heart disease, lead to consistent results (data not shown).

Table 2.

Association Between Parkinson's Disease Duration (Years) and Diagnosis of Hypertension According to Logistic Regression Models

	OR	95% CI	P Value
Hypertension according to office assessment
Crude	0.93	0.88–0.98	.005
Age‐ and sex‐adjusted	0.93	0.89–0.98	.009
Fully adjusted model 1	0.93	0.87–0.98	.013
Fully adjusted model 2	0.92	0.86–0.98	.016
Hypertension according to 24‐h monitoring
Crude	0.95	0.90–1.00	.054
Age‐ and sex‐adjusted	0.95	0.90–1.01	.069
Fully adjusted model 1	0.94	0.89–1.01	.072
Fully adjusted model 2	0.94	0.88–1.01	.103

Abbreviations: CI, confidence interval; OR, odds ratio. Model 1 adjusted for age, sex, and L‐dopa equivalent daily dose (kg).

Model 2 adjusted as model 1 plus: body mass index, Unified Parkinson's Disease Rating Scale III, chronic kidney disease, chronic obstructive pulmonary disease, diabetes, heart failure, coronary artery disease, and hypothyroidism.

Figure 2 shows the diurnal and nocturnal pattern, as assessed through 24‐hour BP monitoring, of systolic and diastolic BP according to tertiles of PD duration. In particular, nocturnal systolic BP showed increasing adjusted mean values with longer PD duration (from 115.8±2.5 mm Hg to 125.0±2.7 mm Hg, P<.05). A nonsignificant positive trend was finally observed for nocturnal diastolic BP.

Figure 2.

Diurnal and nocturnal blood pressure (BP; ambulatory monitoring) according to tertiles of disease duration. * P<.05. BP‐adjusted means and standard error (analysis of covariance). Adjusted for age, sex, Unified Parkinson's Disease Rating Scale III, L‐dopa equivalent daily dose (kg), and number of BP‐lowering drugs.

4. Discussion

According to our results, ABPM improves the detection of hypertension by 15% in PD patients as compared with office measurement. The agreement of these methods is not excellent. Prevalent hypertension is not inversely associated with PD duration, as previously suggested.

Along with osteoarthritis, hypertension was the most common comorbidity in our sample, as it was detected in 60% to 70% of participants, depending on the evaluation method. Previous studies, indeed characterized by a wide methodological heterogeneity, reported figures varying from 14% to 73%.5, 13, 14 To our knowledge, this is the first study to assess and compare the prevalence of hypertension with two different approaches in PD patients.

We found a good, not excellent, agreement between office and ambulatory BP assessments. Previous studies concluded that age and hypertension itself are the main determinants of the scarce agreement between the different methods.15, 16 We believe that the high circadian BP variability, which is a common finding in PD, might explain the nonoptimal agreement in our study.

ABPM has been proven the most reliable method for the diagnosis and staging of hypertension. Unlike in our study, Stergiou and colleagues, in a population‐based study reported that casual measurements of BP overestimate the diagnosis of hypertension by 7% as compared with 24‐hour assessment.15 Conversely, ABPM has been demonstrated to be a valuable approach to detect masked uncontrolled hypertension (31% of the sample) in a large sample of hypertensive patients.17, 18 In our study, 24‐hour BP monitoring constantly detected a higher prevalence rate of hypertension, as compared with office assessment across different PD duration tertiles.

The unbalance between sympathetic and parasympathetic drive, resulting from the degeneration of the peripheral nervous system, accounts for several cardiovascular dysautonomic features of PD, such as postural hypotension, postprandial hypotension, and supine/nocturnal hypertension.19 Such degenerative phenomena are strictly linked to the pathophysiology of PD, and have been found to progress along with disease duration.20 In favor of this observation, we described increasing nocturnal BP levels along the course of PD. Nocturnal hypertension, which is more prevalent in the advanced stages of the disease, might explain why 24‐hour monitoring recognized a higher number of hypertensive patients in the present study.

The relationship between BP and PD is ambiguous. High BP levels have sometimes been shown to be protective against PD.21 In particular, several studies suggested that an increased cardiometabolic burden (high glycemia, lipids, and BP) reduces the risk of PD.2, 3, 22 On the other hand, once the diagnosis is established, PD patients apparently present with a more favorable cardiometabolic profile then the general population.23 Furthermore, scanty evidence shows a negative association between high BP and PD duration.5 In our study, hypertension was inversely associated with PD duration when diagnosed by office measurements, but not when evaluated over the 24 hours, independently from demographics, clinical features, and antiparkinsonian drug use.

In clinical practice, the diagnosis of hypertension relies on office measurements, and pharmacological treatment is adjusted to this target. We found a decreasing use of BP‐lowering drugs with increasing duration of PD, which might reflect the physicians’ perception of hypertension severity. Interestingly, we did not observe any decreasing temporal trend for the drugs used as first‐line antihypertensive agents (ie, ACEIs and TZs). Therefore, office measurements do not take into account nocturnal BP that, in our study, increased with increasing PD duration. As a consequence, hypertension severity is potentially underestimated and undertreated in PD patients, especially in the later disease stages, thus putting them at higher cardiovascular risk.

In PD, the complex interplay between cardiovascular dysautonomia, aging, and dopaminergic treatment challenges physicians who treat hypertension. In the general population, nocturnal hypertension has been independently associated with increased risk of stroke, myocardial infarction, and death.24, 25 In addition, nocturnal high BP is a predictor of brain white matter lesions.26 These same brain lesions mediate the relationship between common cardiovascular risk factors and cognitive impairment in PD. Early identification and treatment of such risk factors might even slow the progression to PD dementia.27

5. Study Limitations

Some limitations of the present study should be mentioned. First, the cross‐sectional design of the study did not allow us to verify at the individual level how BP varies along the course of PD. Second, this was a single‐center study that involved a nonrandom sample of PD patients; therefore, our results may not be generalizable to other populations with PD. Third, the study participants were not naive with respect to the assessment of hypertension and its treatment. This implies that many of them already had a diagnosis of hypertension and were on pharmacological treatment. However, this allowed us to describe the treatment modalities of those with hypertension and to speculate on the perception of hypertension diagnosis and severity in PD among the prescribing physicians. Fourth, several conditions not considered in the present study, including postural hypotension and sleep apnea, might be responsible for residual confounding and may potentially affect the study results.

6. Conclusions

Hypertension is prevalent among PD patients and does not decrease during the course of the neurological disease. Rather, nocturnal BP increases along with PD duration, increasing the risk of cardiovascular events and death. For this reason, office assessment of hypertension in PD patients is not sufficient because it does not consider the extreme circadian variability of BP, which is common in PD. Ambulatory (24‐hour) BP monitoring should be routinely embedded in the screening and follow‐up of hypertension in PD patients.

Conflicts of Interest

None declared.

Supporting information

 

Vetrano, D. L. , Pisciotta, M. S. , Brandi, V. , Lo Monaco, M. R. , Laudisio, A. , Onder, G. , Fusco, D. , L′Angiocola, P. D. , Bentivoglio, A. R. , Bernabei, R. and Zuccalà, G. (2017), Impact of disease duration and cardiovascular dysautonomia on hypertension in Parkinson's disease. Journal of Clinical Hypertension, 19:418–423. doi: 10.1111/jch.12938