OUP user menu

Orthostatic hypotension: a new classification system

B.M.T. Deegan , M. O’Connor , T. Donnelly , S. Carew , A. Costelloe , T. Sheehy , G. ÓLaighin , D. Lyons
DOI: http://dx.doi.org/10.1093/europace/eum177 937-941 First published online: 25 August 2007


Aims Orthostatic hypotension (OH) is a common condition, which is defined as a reduction in systolic blood pressure of ≥20 mmHg or diastolic blood pressure of ≥10 mmHg within 3 min of orthostatic stress. Utilizing total peripheral resistance (TPR) and cardiac output (CO) measurements during tilt-table testing (Modelflow method), we classified OH into three categories, namely arteriolar, venular, and mixed. The principle defect in arteriolar OH is impaired vasoconstriction after orthostatic stress, reflected by absence of the compensatory increase in TPR. In venular OH, the predominant defect is excessive reduction in venous return, reflected by a large drop in CO after orthostatic stress with marked tachycardia. Mixed OH is due to a combination of both these mechanisms.

Methods and results We analysed haemodynamic parameters of 110 patients with OH and categorized them as arteriolar, venular, or mixed. Significant differences between the groups were found for the magnitude and time to reach nadir of the systolic blood pressure drop post-head-up tilt. The mixed OH category had the largest systolic blood pressure reduction (42.5, 31.9, 53.3 mmHg, P < 0.001) and the longest nadir time (18.6, 20, 30.7 s, P = 0.002).

Conclusion This is a practical classification tool and when validated physiologically, this system could be useful in directing treatment of OH.

  • Orthostatic hypotension
  • Blood pressure
  • Cardiac output
  • Total peripheral resistance
  • Syncope
  • Tilt-table testing


Orthostatic hypotension (OH) is defined in the joint consensus statement of the American Autonomic Society and the American Academy of Neurology, as a reduction in systolic blood pressure of ≥20 mmHg or a reduction in diastolic blood pressure of ≥10 mmHg within 3 min of undergoing orthostatic stress.1 It is a clinical condition that can result in transient cerebral hypoperfusion after a change in posture, with symptoms such as dizziness, weakness, blurred vision, or syncope. OH is a common condition with a prevalence reported as high as 30% in elderly, home-dwelling subjects.2

Blood pressure is a function of blood flow and vascular resistance. Blood flow is equivalent to cardiac output (CO). Resistance to blood flow occurs predominantly in the peripheral vessels—total peripheral resistance (TPR). This leads to the following equation: Embedded Image

Upon assumption of upright posture, gravity causes a downward displacement of between 500 mL to 1 L of blood in the lower limbs and abdomen,3 causing a reduction of venous return to the heart and a CO drop of ∼20%.4,5 Control centres in the medulla act to compensate for the resulting drop in arterial blood pressure by increasing sympathetic and reducing parasympathetic nervous system output resulting in reflex tachycardia and an increase in TPR. In a healthy subject, orthostatic stabilization is normally achieved within 1 min of standing.

A failure of any of these systems can lead to an episode of syncope, dizziness or falls. We classified the patient database into three categories based on changes in TPR and CO on during head-upright tilt-table testing (HUT) as follows.

Arteriolar dysfunction

Patients in this category experienced a drop in TPR during HUT. This led to a drop in blood pressure despite an increase in CO. Changes in TPR during orthostatic stress can be measured during tilt using the Finometer Modelflow software (see Table 1 and Figure 1).

Figure 1

Arteriolar orthostatic hypotension.

View this table:
Table 1

Classification of orthostatic hypotension

Arteriolar failure↓/↔TPR and ↑/↔CO
Venular failure↑TPR and ↓CO
Mixed arteriolar/venular failure↓/↔TPR and ↓COa
  • a>10% change.

Venular dysfunction

Patients in the venular category exhibited a drop in CO on HUT, leading to a drop in blood pressure despite the compensatory increase in TPR (see Table 1 and Figure 2).

Figure 2

Venular orthostatic hypotension.

Mixed dysfunction

Arteriolar and venular dysfunction can occur simultaneously in any individual to cause OH (see Table 1 and Figure 3).

Figure 3

Mixed orthostatic hypotension.


About 110 patients who were diagnosed with OH during HUT were identified from our syncope database. These patients met the diagnostic criteria for OH as per the American Society guidelines. Standard HUT protocol was observed. All tilt studies were performed in a dedicated, temperature-controlled syncope laboratory in the afternoon at least 2 h post-prandially. No medication was administered during testing.

Patients were allowed to lie supine on a hydraulic tilt table for at least 5 min to establish a baseline value. They were then tilted to 70° (head up) over 15 s and maintained semi-erect for 3 min. This is a haemodynamic study only. We did not include data for underlying disease or co-morbidity.

We analysed the haemodynamic parameters of these patients with OH using Finometer data obtained during HUT testing. Phasic blood pressure measurement allows ‘beat-to-beat’ recording of blood pressure changes (The TNO Finometer, TPD—Biomedical Instrumentation, Amsterdam) using an infrared digital artery photoplethysmograph combined in a finger cuff, using the ‘volume clamp’ method.6 The Finometer also allows indirect non-invasive calculation of several cardiovascular variables (Modelflow method) including TPR, stroke volume (SV), and CO from an arterial pressure using a three element model of the arterial input impedance.7 TPR is calculated as the quotient of mean arterial pressure and CO. A three lead electrocardiograph is recorded simultaneously on the Finometer (using a SC 7000, Siemens monitor). The patients were categorized into three groups, arteriolar, venular, and mixed based on changes in TPR and CO.

Finometer records were exported to ASCII format and then analysed using Matlab (Mathworks Inc., 24 Prime Park Way, Natick, MA, USA).

Signal artifact (e.g. due to hand movement) was automatically detected and removed from systolic blood pressure, TPR, CO, and HR traces, using a customized filtering algorithm.

Data extraction algorithms were developed and applied to the systolic blood pressure traces for the automatic calculation of:

  • magnitude of blood pressure drop to nadir within 3 min of head-up tilt,

  • rate of blood pressure drop to the nadir (decay rate), and

  • time for blood pressure to drop to the nadir (decay time)

  • measured from 90–10% of the blood pressure drop post-tilt.

Changes in systolic blood pressure drop, decay rate, decay time, and heart rate were analysed for significant differences between arterial, venular, and mixed groups. With continuous haemodynamic measurements, small fluctuations can occur over time. For the purposes of classification, a change of >10% in TPR or CO was taken to be significant.

Statistical analysis

For statistical analysis the distribution of all data was first tested for normality. For normally distributed data, analysis of variance (ANOVA) was used to test for significant differences between the three groups. Post hoc multiple t-test comparison was then performed to determine which groups were significantly different and a Bonferroni correction was applied to adjust for the problem of inflating the type-I error. For non-normal data the Kruskal–Wallis test was used to test for statistical difference between the three groups. For pair-by-pair comparison Mann–Whitney U tests were used. A P-value ≤0.05 was used as the level of statistical significance. All statistical analyses were performed using SPSS 10 for Windows.


The median values for the reduction of systolic blood pressure post-head-up tilt in the arterial, venular, and mixed categories are shown in Table 2. These data were not normally distributed (negative skew). Significant differences were noted between the three categories (non-parametric Kruskal–Wallis test, P < 0.001). When tested pair by pair the median for the arterial category was significantly different from the median of venular category (P = 0.046) and likewise for the arterial and mixed categories, and the venular and mixed categories (Mann–Whitney U test, P < 0.001).

View this table:
Table 2

Results of statistical analysis for the orthostatic hypotension categories

Number of patients275924
Age (years)a68.8 ± 18.176.5 ± 10.479.8 ± 7.9
Reduction in systolic blood pressure (mmHg)b42.5 (30.65–46.93)31.9 (28.12–40.78)53.3 (45.59–66.76)<0.001
Heart rate increase (bpm)a7.74 ± 5.699.61 ± 4.398.87 ± 6.560.311
Decay time (s)b18.6 (11.57–28.5)19.98 (11.11–35.37)39.7 (20.32–61.27)0.002
Decay rate (mmHg/s)b−1.49 (−2.3 to −0.94)−1.30 (−2.0 to −0.77)−1.03 (−2.12 to −0.69)0.356
% change TPRb−29.4 (−40.7 to −18.79)36.5 (25.15–51.42)−13.1 (−22.48 to −5.87)<0.001
% change COb15.5 (5.71–29.37)−32.5 (−41.22 to −25.18)−18.4 (−27.35 to −15.85)<0.001
% Patients reporting symptoms3347460.458
  • aValues expressed as mean ± standard deviation.

  • bValues expressed as median (interquartile range).

HR increase was normal and a parametric ANOVA test was used to test for significant differences in the means of the OH categories. There were no significant differences noted (P = 0.311).

Based on the presence or absence of symptoms, there was no significant difference in the presence of symptoms between the OH categories (χ2 test, P = 0.458). However, when we compared the symptomatic versus the asymptomatic group, regardless of their OH categorization, we found there was a statistically greater drop in CO in the symptomatic group (P = 0.03). For CO, the median for the asymptomatic group was −18.5% and for the symptomatic group was −28.2%.

Decay rate data were not normally distributed (negative skew). There was no significant difference in the decay rate between the OH categories (Kruskal–Wallis test, P = 0.356).

There were significant differences between the OH categories for decay time (Kruskal–Wallis test, P = 0.002). When tested pair-by-pair, the arterial and venular category medians were not significantly different but the medians for arterial and mixed (Mann–Whitney U test, P = 0.005) and venular and mixed were significantly different (Mann–Whitney U test, 0.001).


Tilt induced OH may not be the physiological equivalent of spontaneous OH. Tilt-table testing examines the orthostatic response in a maximally controlled environment. Specificity of up to 90% has been reported at tilt angles between 60° and 70°, and specificity ranging between 32 and 85%.810 In this study, we analysed the haemodynamic parameters using the Modelflow method according to three categories of OH: arteriolar, venular, and mixed. The Modelflow technique has been used to examine haemodynamic changes in vasovagal syncope.11

We found significant differences between the three categories in a number of parameters. Mean systolic blood pressure changes were statistically different between all categories. The greatest reduction in systolic blood pressure was in the mixed category. This was expected as these patients had no compensatory increase in TPR after orthostatic stress and also had a significant reduction in CO. In contrast, patients in the venular category had the smallest reduction in systolic blood pressure after tilting. In these patients, despite a marked reduction in CO, TPR increased significantly in an attempt to compensate for OH. The decay time from head-up tilt to the blood pressure nadir was significantly longer in the mixed category, and this is to be expected due to the lack of compensatory response in this category.

We believe that this classification system divides OH into logical categories based on the underlying mechanism and potential treatment options. If the predominant defect is impaired peripheral arteriolar vasoconstriction, this results in impaired elevation in TPR in response to orthostatic stress. A head-up tilt trace would show a post-tilt reduction in blood pressure with a compensatory increase in HR and CO, without compensatory TPR increase. These findings would suggest alpha-receptor abnormalities, such as autonomic failure or peripheral neuropathy. Midodrine (an alpha-1 agonist), which causes arteriolar and venous vasoconstriction, would be an appropriate choice in the treatment of arteriolar OH.

However, if the predominant defect is venular (e.g. nitrate-induced venodilation), a post-tilt reduction in blood pressure would be associated with a reduction in SV and CO with a compensatory increase in HR and TPR. This may be linked to diuretic use, hypovolaemia or venous valve incompetence. In this category, European class III pressure stockings or fludrocortisone could be an effective treatment option.

The effectiveness of interventions based on this classification system would merit further research.


A limitation of this tool is that it relies upon values for TPR and CO, which are derived using the Modelflow method.7,1217 Modelflow assumes a population average of aortic area depending on gender, age, height, and weight. Several studies have validated the Modelflow method compared to standard invasive techniques (thermodilution, pulse dye-densitometry). The Modelflow method has been shown to provide accurate estimates of SV from the intra-arterial blood pressure waveform.18,19

It should be noted that, as TPR is estimated by the Modelflow method, it is not possible to completely differentiate the venous contribution from the arterial system.11 If arterial tone is higher than venous tone, the estimated TPR does not allow us to differentiate the venous contribution from that of the arterial system.


We are proposing a novel classification system for OH utilizing TPR and CO measurements to categorize patients into arteriolar, venular, and mixed OH. This may have implications mechanistically and therapeutically. To validate this classification system, further studies are needed.


This work is part funded by an IRCSET (Irish Research Council for Science Engineering and Technology) postgraduate scholarship, and by shire pharmaceuticals groups plc.

Conflict of interest: none declared.


View Abstract