Abstract
To investigate the regional and directional compliance/distensibility of the healthy aorta.
Complete fresh porcine aortas (n = 11) were perfused ex vivo under defined haemodynamic parameters using a custom-made pulse duplicator. Both circumferential and longitudinal compliance were measured optically.
The pulse duplicator was able to perfuse the entire aorta with arbitrary haemodynamic parameters, generating a physiological pulse curve. Aortic compliance is pressure dependent, as we observed a linear relationship between pressure and distension in the range of 5-200 mmHg; however, above 200 mmHg, the porcine aorta behaved in an inelastic manner. Circumferential compliance was highest in the ascending aorta (24%/100 mmHg) but significantly (P < 0.05) decreased in both the arch (18%/100 mmHg) and the descending aorta (15%/100 mmHg). Longitudinal compliance was highest in the ascending aorta and clearly exceeded circumferential compliance. Compliance was significantly (P < 0.05) higher in the outer curvatures of the ascending aorta and the aortic arch compared with the compliance of the inner curvature at these locations (30%/100 vs 23%/100 mmHg in the ascending aorta and 20%/100 vs 9%/100 mmHg in the arch, respectively).
Longitudinal compliance of the ascending aorta, particularly the outer curvature, is predominantly responsible for the ‘Windkessel effect’. Pathological changes such as elongation and pronounced angulation of the ascending aorta increase stress on the outer curvature and may be important factors in the development of aortic dissection.
INTRODUCTION
The aortic wall serves as a blood reservoir during systole, absorbs systolic kinetic energy, emits this energy during diastole and transforms strictly pulsatile cardiac output into a more constant waveform (Windkessel effect). Material properties (i.e. elastic moduli) of the aorta have been investigated with manifold methods [1–4]; however, the reported numbers differ, most likely due to the multiplicity of methods and specimens used.
Knowledge regarding regional and directional distensibility/compliance of the aorta is limited, although it is relevant with respect to pathophysiology: the formation of an aneurysm of the ascending aorta by transverse dilatation is a well-known complication of arterial hypertension; however, less attention has been paid to elongation, the longitudinal dilatation, which is followed by pronounced angulation of the ascending aorta. These changes lead to increased wall stress as well as to changes in compliance [4]. However, the properties of the aorta change throughout life [5] and compliance most likely differs among he anatomical layers of the aorta. These changes play a crucial role in the development of Stanford type A aortic dissection. We and others have observed the entry tear in Stanford A dissection often running in a transverse rather than a longitudinal direction [6, 7]. We hypothesize that the entry tear of an aortic dissection results from an massive circumferential or, maybe more importantly, longitudinal dilatation of the aorta exceeding the compliance of the intima.
The aim of this study was to investigate regional and directional compliances of the aorta. We engineered a pulse duplicator, allowing the ex vivo perfusion of the entire aorta with arbitrary haemodynamic parameters and subsequent direct measurements of distensibility.
With respect to pathophysiology, a major drawback of our model was obvious: we used young porcine aortas; therefore, all chronic pathological changes leading to the development of an aneurysm or dissection in humans were omitted from this study.
MATERIALS AND METHODS
Preparation of aortas
Fresh porcine thoracic organs (n = 11) were purchased from a local butcher. No animals were sacrificed specifically for this study, as the organs were from regularly slaughtered pigs. At the time of slaughter, the animals were ∼6 months old and weighed ∼120 kg.
The aorta was dissected from the surrounding tissue, beginning at the aortic valve and extending to the thoraco-abdominal junction. The left ventricular outflow tract, consisting of the anterior mitral leaflet and the basal interventricular septum, was excised from the heart and sutured to a 30-mm Dacron vascular prosthesis to allow a connection to the pulse duplicator without compromising aortic valve function. All aortic branches were ligated.
Custom-made pulse duplicator for ex vivo aortic perfusion
The frame of the construction was adopted from a bicycle ergometer (Fig. 1). An electrical drilling machine (Bosch®) was used as an engine. The speed of the drilling machine (idling revolution speed 50–3000 min−1) was adjustable semi-quantitatively using a screw mechanism. A 10 : 1 speed reduction was achieved with a gear made from a bicycle chain and its rack-wheels (Shimano®). The power transmission to the actual pump was achieved with the bottom bracket and pedal from the original ergometer. The pump itself consisted of a 60-mm calibre cylinder, which had a maximum capacity of 500 ml (JAGO®). For low-resistance, high-volume flow, large valves (1.5 inch, ITAP®) were used. The inflow portion of the pump consisted of a canister that was connected to the pump by tubing that was 1.5 inches in diameter. An identical tubing was mounted at the outflow of the pump. The connection to the aorta was achieved using an adapter nipple, which reduced the internal diameter from 1.5 to 1 inch, as well as a vascular prosthesis. The distal end of the aortic specimen was connected to a 3/8-inch tubing, which provided continuity with the canister and completed the circuit. Adjustable resistance was achieved via a custom-made screw mechanism attached to the post-aortic tubing.
Custom-made pulse duplicator.
Measurement of aortic dimensions and distensibility
The aortas were filmed with a digital camera while being perfused. Intra-aortic pressures and flows were recorded simultaneously using a haemodynamic monitor (Hewlett-Packard, Merlin Model 68 s) and a flow-meter (Emtec, Sono TT). For each aorta, films and pressures were recorded in at least 10 different haemodynamic situations, and representative freeze images were generated at maximum systole and diastole from these films.
The lengths of the aortic segments and aortic diameters were measured digitally from these photographs, using ‘analysing digital images’ software (freeware) (www.globalsystemsscience.org/software/download, last accessed 12 November 2014). The measurements were performed in a blinded manner: at the time of measurement, the pressure and stage of the cardiac cycle, either systole or diastole, were unknown.
Aortic dimensions were measured at the following nine distinct points and landmarks along the course of the aorta (Fig. 2): the sinotubular junction (C1), the mid-ascending aorta (C2), the beginning of the brachiocephalic trunk (C3), the end of the brachiocephalic trunk/mid-aortic arch (C4), end of the left subclavian artery/distal aortic arch (C5), the ligamentum arteriosum (C6), the proximal descending aorta (C7), the mid-descending aorta (C8) and the thoraco-abdominal junction (C9). Using these recorded diameters (d), the circumference (C) of the aorta, which represented actual tissue distension, was calculated by C = πd.
Aortic dimensions: (A) circumferential and (B) longitudinal (The pig has just two supra-aortic branches).
With respect to the longitudinal course of the aorta, the lengths of the ascending aorta (C1–C3), aortic arch (C3–C6) and descending aorta (C6–C9) were measured. The length of each segment was measured at the outer curvature (O), central aspect (L) and inner curvature (I) of the aorta.
Calculating the distensibility of the aorta, and statistical presentation
The measured aortic dimensions recorded from each individual measuring point were plotted against the pressures (Fig. 3). The aortic dimensions were analysed within the pressure range of 5–200 mmHg: lower pressures were excluded because collapse of the aorta results in non-circular geometry, and subsequently, in unusable data: values above 200 mmHg were excluded because the aortic tissue loses its linear distensibility and elastic character (see the relevant section in results) at that pressure and above.
Pressure–diameter diagrams of 3 different aortas at C3, demonstrating the linear relationship between pressure and diameter throughout the pressure range 5–200 mmHg, and the reduced compliance observed in higher pressure ranges.
A linear regression analysis was performed for the linear section of the pressure–dimension function. Pearson's product-moment correlation coefficient was calculated to ensure the quality of the measurement. The Y-intercept of the linear equation represented the virtual dimensions of the aorta at 0 mmHg. The real dimensions of the aorta at 0 mmHg cannot be measured properly because the aorta collapses and loses its circular shape. The actual pressure-related dimensions were divided by the virtual dimension at 0 mmHg, from which we obtained the relative extension of a piece of tissue 1 cm in length. The gradient of the linear equation of these values represented the distensibility of the tissue in question.
The medians, quartiles and ranges of distensibility of all aortas (n = 11) were calculated and compared using the different measuring points. The Mann–Whitney–Wilcoxon test was performed to test for significant differences, and P ≤ 0.05 was considered statistically significant. SSPS (SSPS 20.0, IBM Corp., Amonk, NY, USA) and Excel (MS Excel 2010, Microsoft, Redmond, WA, USA) were used for all analyses and data presentation.
RESULTS
Evaluation of the custom-made pulse duplicator
The custom-made pulsatile pump was capable of perfusing porcine aortas reproducibly with stroke volumes of 30–100 ml/stroke at frequencies of 40–150 strokes/min, time volumes up to 10 l/min and pressures reaching 350 mmHg. The native aortic valve opened and closed with every beat as verified by endoscopic examination (data not shown). The pulse curve and pressure course (Fig. 4) were comparable with the physiological pressure curve: ∼30% of stroke time occurs during a rise in pressure (systole), and 70% occurs during a decline in pressure (diastole). This ratio of systole/diastole was independent of pulse frequency, resulting in systolic times of 0.35–0.15 s. The pressure increase was almost linear; the pressure decrease was steeper in early diastole and flattened towards its end. A dicrotic notch corresponding to the closure of the aortic valve was noted.
Pulse curve within the aortas in different haemodynamic situations.
Relationship between pressure and distension
The relationship between pressure and aortic distension (circumferential and longitudinal) was linear for pressures in the range of 5–200 mmHg. This is exemplified in Fig. 3, which shows the pressure–diameter diagrams of three different aortas at C3. At high pressures, above 200 mmHg, the distension curve flattens, and approaches maximum distension. In that pressure range, the aorta is no longer elastic. Consequently, pressure values >200 mmHg were excluded from this study. The average Pearson's linear correlation coefficient between pressure and aortic dimensions for 5-200 mmHg was 87.3 (SD 11.5) for longitudinal and circumferential measures, at all measurement points.
Circumferential compliance of the aorta
Figure 5 depicts regional differences in circumferential distensibility of the aorta. In the ascending aorta, the median circumferential distensibility was 24%/100 mmHg at C1 and C2, and 23%/100 mmHg at C3. The distensibility at all three points was significantly (C1 vs C4, C5 and C6; C2 vs C4, C5 and C6; C3 vs C4, C5 and C6; all P < 0.05) higher compared with the distensibility in the aortic arch, which was 18%/100 mmHg at both C4 and C5. In the descending aorta (C7–C9), the median distensibility was 15%/100 mmHg, which was again significantly lower compared with the ascending aorta (C1 vs C7, C8 and C9; C2 vs C7, C8 and C9; C3 vs C7, C8 and C9; all P < 0.05), but only marginally lower compared with the aortic arch (C4 vs C7, C8 and C9; C5 vs C7, C8 and C9; C6 vs C7, C8 and C9; all not significant).
Circumferential compliance of the aortas at different measuring points.
Longitudinal compliance of the aorta
Figure 6 depicts regional differences in longitudinal distensibility of the aorta. In the central aspect of the ascending aorta, the median longitudinal distensibility was 28.4%/100 mmHg. The distensibility in the outer curvature (30.3%/100 mmHg) was significantly higher (P = 0.040) compared with the inner curvature (23.5%/100 mmHg).
Longitudinal compliance of the aortas at different measuring points.
In the aortic arch, longitudinal distensibility was significantly (P = 0.004) lower compared with the ascending aorta. However, the distensibility of the outer curvature of the arch (20.2%/100 mmHg) was significantly (P = 0.007) higher compared with the inner curvature (9.1%/100 mmHg). Longitudinal distensibility was lowest in the descending aorta, as it was ∼10%/100 mmHg and there was no difference in distensibility between the outer and inner aspects of the vessel (P = 0.320; not significant).
DISCUSSION
Experimental design
To investigate the regional and directional elasticity of the aorta ex vivo, we designed and built a pulse duplicator. The device allows for perfusion of the entire vessel with adjustable physiological and supraphysiological stroke volumes, frequencies, resistances and pressures. Pulse duplicators have been designed before, i.e. the ‘Sheffield pulse duplicator’ [8], and are used for testing heart valves [9]; however, to the best of our knowledge, this is the first apparatus and experimental set-up focusing on pulsatile perfusion of the entire aorta. The apparatus is capable not only of perfusing the aorta ex vivo with physiological pressures, but is also able to generate a pulse wave very similar to the natural wave. However, in nature, the duration of systole is relatively independent of heart rate; in our model it is dependent on the pulse rate, as there is a linear relationship between the two. Further, in contrast to other available pulse duplicators, our apparatus is able to produce hypertensive conditions. Although our custom-made pulse duplicator performed satisfactorily, it must be considered a low-budget prototype and will be further developed and improved in the future.
In vivo investigation of aortic compliance is possible [3, 4], but echocardiography faces problems, such as difficult accessibility and problematic measurements of the lengths of some aortic segments. The classic approach of material science, testing tissue pieces on a tensile tester, provides exact- and reproducible results and perfectly allows testing the directional compliance [10, 11]; however, this approach neglects the 3D aortic structure and the dynamic and 3D nature of pressure impact on the vessel wall. With the ex vivo pulsatile perfusion model presented herein, we tried to tackle these problems. To avoid the influence of fixation or freezing, we exclusively analysed fresh tissue.
A critical point in this study is the use of young porcine aorta and the debatable applicability of our results to humans. Pigs are regularly used as a model in cardiovascular research, and it is very probable that the results demonstrated herein are representative of the physiology of a juvenile human; however, they are certainly not comparable with the material properties of an aged and diseased aorta. Consequently, the conclusions drawn herein have more of a physiological than a pathophysiological impact.
Aortic compliance
Aortic compliance is pressure dependent: we observed a linear correlation between longitudinal and circumferential distension of the aortic wall and pressure within a physiological range. Under hypertensive conditions, the aortic wall loses its elasticity and becomes inflexible. This effect has been described before in a static model of rabbit [12] and human aortas [13] and is a result of the histological composition of the medial layer with elastin being responsible for the elastic response in the physiological range of pressure and collagen being responsible for the right end of the pressure–strain curve. The transition zone between elastic and stiff wall properties in our dynamic model of juvenile porcine aorta was ∼200 mmHg. Within the physiological pressure range, the healthy aorta clearly functions in the elastic part of the curve. However, it is very probable that the elasticity of the aorta decreases in the elderly and that the transition zone between elastic and stiff wall properties shifts towards lower blood pressures; therefore, the aorta may behave in a non-compliant manner within the physiological blood pressure range [13]. Consequently, the impact generated by systolic momentum must be absorbed by collagen fibres without the resilience of elastin.
As expected, regional elasticity was highest in the ascending aorta and decreased along its course towards the descending aorta. Remarkably, longitudinal compliance in the ascending aorta exceeded circumferential compliance, which is in line with the findings of Bergel et al. [1] and Iliopoulos et al. [11], whereas others did not report directional differences in compliance [10]. In the arch, longitudinal and circumferential distensibility seemed to be almost identical, and in the descending aorta, the circumferential distensibility seemed to be marginally higher. The longitudinal compliance of the ascending aorta seems to be predominantly responsible for the ‘Windkessel effect’ and enables relative movement of the heart during the heart cycle and allows for an effective interplay between the heart and the aorta.
Interestingly, the outer curvature of the ascending aorta and the aortic arch was more elastic compared with the inner curvature, which, to the best of our knowledge, has not been described before; however, this finding has to be confirmed by further studies. This difference in compliance may be of great physiological and pathophysiological relevance, as the ascending aorta is fixed to the surrounding tissue along its proximal (heart) and distal (brachiocephalic trunk) aspects. Consequently, strictly centrifugal systolic expansion of the aorta is impossible. As illustrated in Fig. 7, systolic elongation of the vessel subsequently results in a stronger angulation and bending of the ascending aorta, which leads to significant stress on the outer curvature. This effect is even more pronounced when the aorta becomes elongated as is frequently seen in the elderly [6, 14]. The outer curvature appears to be exposed to the greatest impact from the cardiac output jet, particularly with the increasing angulation of the ascending aorta. As depicted previously, the entry tear of aortic dissections is often observed to run in a transverse manner [6, 7]; several mechanisms are believed to be responsible for both the recurrent location and direction [2, 15].
Illustration of the theoretical changes in length of the inner and outer curvatures of the ascending aorta: presumably, the proximal and distal portions are fixed to the surrounding tissue. Note the disproportionate gain in length of the outer curvature.
In the face of the findings presented herein, we have hypothesized that the following play a pivotal role in the development of aortic dissection: (i) the overall longitudinal compliance of the ascending aorta; (ii) the particularly high regional compliance of the outer curvature and its changes throughout life and (iii) the pathological elongation of the ascending aorta.
The aorta appears to be a complex organ with distinct regional and directional differences in elasticity, designed to effectively absorb the kinetic energy of cardiac systole and to cushion the momentum of systolic impact. These regional differences in physiological behaviour may be of importance in determining the pathophysiological incidence of specific events, such as the formation of an intimal tear of aortic dissection. Further research is required to (i) reproduce our results in the healthy human and to (ii) evaluate changes in the biophysics of the ageing and diseased aorta.
Conflict of interest: none declared.
REFERENCES
Author notes
Presented at the 28th Annual Meeting of the European Association for Cardio-Thoracic Surgery, Milan, Italy, 11–15 October 2014.
