Portal de Periódicos da CAPES

Relationships of Intraosseous and Systemic Pressure Waveforms in a Swine Model

2014; Wiley; Volume: 21; Issue: 8 Linguagem: Inglês

10.1111/acem.12432

1553-2712

Robert A. De Lorenzo, John A. Ward, Bryan S. Jordan, Chris E. Hanson,

Cardiac Arrest and Resuscitation

Despite some focus on the use of intraosseous (IO) catheters to obtain laboratory samples, very little is known about the potential for obtaining other forms of clinical data. Largely unstudied is the relationship between IO pressures (IOPs) and systemic hemodynamic pressures such as mean arterial pressure (MAP) and central venous pressures (CVP). The objective was to explore the relationship between hemodynamic parameters (blood pressures) measured through an IO catheter and intravascular catheters placed in the arterial and central venous circulation. Eight pigs (Sus scrofa) weighing 30 to 45 kg were sedated with a short-acting agent, intubated with a cuffed endotracheal tube, and anesthetized with 2% to 3% isoflurane. Intravascular catheters were placed into the femoral or carotid artery and the femoral or jugular vein for MAP and CVP measurements. IO catheters, 15 mm for the sternum and 25 mm for the long bones, were placed percutaneously into the proximal tibia, proximal femur, proximal humerus, right proximal ulna, and/or sternum. Pressures were recorded during normotension, hypotension, and hypertension. Averaged across all eight animals, the means (ranges) for baseline systemic pressures were as follows: MAP = 66.5 (55.6 to 76.7) mm Hg, tibia IOP = 17.4 (9.3 to 34.5) mm Hg, femur IOP =18.4 (3.3 to 33.1) mm Hg, humerus IOP = 15.7 (2.8 to 28.9) mm Hg, ulna IOP = 16.0 (7.9 to 25.6) mm Hg, sternum IOP = 5.7 (–0.5 to 47.9) mm Hg, and CVP = 2.7 mm Hg (–3.3 to 7.9) mm Hg. The best median correlation occurred between femur IOP and mean MAP (r = 0.65). The four highest correlations between IOP and MAP were associated with mean femur IOP. Only one IO site had a correlation coefficient over 0.50 for CVP. The long bones tended to correlate better with the MAP and the sternum tended to correlate better with the CVP. Nonlinearity was observed in the actively rising pressure phases, which can be explained by a hysteresis model. The relationship between IOP and MAP or CVP is variable by site, with the MAP and CVP tending to be estimated by the femur and sternum, respectively. The relationship to actively rising pressures is nonlinear and a hysteresis model is proposed to explain the phase change. Further experimentation is needed to refine the IOP relationship to the MAP and CVP and assess the potential of these measurements to provide clinically relevant information. A pesar de ciertas aproximaciones en el uso de catéteres intraóseos (IO) para la obtención de muestras de laboratorio, se sabe poco sobre su potencial para obtener otros tipos de datos clínicos. Una cuestión poco estudiada es la relación entre la presión intraósea (PIO) y las presiones hemodinámicas sistémicas como la presión arterial media (PAM) y la presión venosa central (PVC). Explorar la relación entre los parámetros hemodinámicos (presiones arteriales) medidos a través de un catéter IO y de catéteres intravasculares colocados en la circulación venosa central y arterial. Ocho cerdos (Sus scrofa) con pesos entre 30 y 45 kg, se sedaron con un fármaco de corta duración de acción, se intubaron con un tubo endotraqueal con balón y se anestesiaron con isoflurano al 2%-3%. Los catéteres IO, de 15 mm para el esternón, y 25 mm para huesos largos, se colocaron percutáneamente en la tibia proximal, el fémur proximal, el húmero proximal, el cúbito proximal derecho y/o el esternón. Las mediciones promedio de los 8 animales, la media y los rangos de las presiones sistémicas iniciales fueron: MAP = 66,5 mm Hg (rango de 55,6 a 76,7 mm Hg), PIO en tibia = 17,4 mm Hg (9,3 a 34,5 mm Hg), PIO en fémur = 18,4 mm Hg (3,3 a 33,1 mm Hg), PIO en húmero: = 15,7 mm Hg (2,8 a 28,9 mm Hg), PIO en cúbito = 16,0 mm de Hg (7,9 a 25,6 mm de Hg) y PIO en esternón = 5,7 mm de Hg (-0,5 a 47.9 mm Hg), y PVC = 2,7 mm de Hg (-3,3 a 7,9 mm de Hg). La mejor correlación de las medianas se produjo entre la PIO del fémur y la media de PAM (r = 0,65). Los cuatro mayores correlaciones entre la PIO y la PAM se asociaron con la media de la PIO del fémur. Sólo una localización IO obtuvo un coeficiente de correlación mayor a 0,50 para la PVC. La PIO en los huesos largos tendieron a correlacionarse mejor con la PAM y el esternón tendió a correlacionarse mejor con la PVC. Se observó una correlación no lineal en las fases de elevación activa de las presiones, que puede explicarse mediante un modelo de histéresis. La relación entre PIO y PAM o PVC es variable según el lugar, con una tendencia del fémur y el esternón a estimar la PAM y PVC, respectivamente. La relación con el aumento de las presiones de forma activa es no lineal, y se propone un modelo de histéresis para explicar el cambio de fase. Se necesita más experimentación para refinar la relación entre la PIO con la PAM y la PVC, y valorar el potencial de estas mediciones para proporcionar información clínicamente relevante. Intraosseous (IO) access was first conceived in the 1940s and reemerged in the late 20th century as a means to obtain vascular access in pediatric patients in whom intravenous (IV) access was difficult.1 It is now widely accepted for use in adults as well and has emerged as an important adjunct to emergency vascular access.1 It is now considered an acceptable route for infusing fluids and medications. Less well studied is its role in diagnosis and monitoring. Some physiologic parameters obtained by IO catheter sampling, such as hemoglobin, glucose, and partial pressure of carbon dioxide (PCO2), have been shown experimentally to correlate with the same parameters obtained by traditional venous and arterial sampling.2 Despite some focus on the use of IO catheters to obtain laboratory samples, very little is known about the potential for obtaining other forms of clinical data. Also less studied is the relationship between IO and systemic mean arterial pressure (MAP) and central venous pressure (CVP). As early as the 1950s it was observed that a pulsatile waveform could be measured in the IO space, and it was seemingly in phase with the systemic arterial pressure.3, 4 More recently, Frascone and colleagues5, 6 in two preliminary reports showed the pulsatile waveforms associated with IO pressure (IOP) may mimic the arterial pressure waveform and are associated with respiratory variation and volume status. The IOP is reported generally to be a fraction of systemic arterial pressure, usually in the range of 10% to 30%. However, the true relationship between IOP and systemic vascular pressures remains undefined. The significance of a possible correlation between IO and arterial blood pressures is important. A relationship, if noted, would provide the clinician with an alternative option to traditional invasive vascular monitoring. It would also underscore the importance of a dedicated IO site for vascular monitoring. It is also possible that the IO site used (i.e., tibia, femur, humerus) may affect the correlation to intravascular blood pressures. Besides the long bones, the sternum, clavicle, and calcaneus have also been used as potential sites for IO insertion. A smaller bone or more distal site may show more or less correlation, but this has yet to be determined. The primary goal of this study was to explore the relationship between hemodynamic parameters (blood pressures) measured through IO catheters and intravascular catheters placed in the arterial and central venous circulation. These correlations were examined under normal, hypotensive, and hypertensive conditions. Secondary analyses were to explore whether the IO site used (e.g., tibia, femur, humerus) affected the correlation. This was a hypothesis-generating observational study using a Sus scrofa model under controlled laboratory conditions. This study was conducted in compliance with the Animal Welfare Act, the implementing Animal Welfare Regulations, and in accordance with the principles of the Guide for the Care and Use of Laboratory Animals. The study was approved in advance by the institutional animal care and use committee. We used eight pigs (S. scrofa) weighing 30 to 45 kg, each of which was placed supine on the table, sedated with short-acting agents (tiletamine and zolazepam), and intubated with a cuffed endotracheal tube. Anesthesia was maintained with 2% to 3% isoflurane. Ventilator settings were set to deliver 100% oxygen at 10 to 12 breaths/min with a tidal volume of 6 mL/kg. Settings were adjusted as necessary to maintain an end tidal mean (±SD) PCO2 of 40 (±5) mm Hg. Body temperature was maintained in the normal range (38 to 39°C). Intravascular catheters were placed into the femoral or carotid artery and the femoral or jugular vein for MAP and CVP measurements. MAP was recorded according to the standard formula, and the electrocardiogram (ECG) was recorded along with pressure data. A calibrated data acquisition system was used for continuous collection of data. Intraosseous catheters (15 mm for sternum, 25 mm for long bone sites; EZ-IO device, Vidacare, Inc., San Antonio, TX) were placed percutaneously into the proximal tibia, proximal femur, distal humerus, right proximal ulna, or sternum. Proper placement was established by noting a loss of resistance and penetration of the bony cortex and the ability to aspirate marrow contents. Additional confirmation using fluoroscopy was performed to visualize intramedullary placement. IO patency was maintained with hourly flushes of approximately 10 mL of normal saline or to restore pulsatile pressure. Limb position was maintained by extending the limbs and securing with ties. Once baseline measurements were recorded and the animal determined to be hemodynamically stable, decreased blood pressure was induced by IV administration of nitroprusside sodium (initial dose of 8 μg/kg/min, titrated for effect with a target of 50% of baseline MAP) delivered through the central venous catheter over a 10-minute period. After allowing blood pressure to return to baseline values, increased blood pressure was induced by IV administration of norepinephrine bitartrate (8 to 12 μg/min, titrated for effect with a target of 200% of baseline MAP), delivered through the central venous catheter over a 10-minute period. Pressures were recorded continuously at 500 Hz from the start of the experiment until the MAP returned to within 10% of initial baseline values. The study endpoint was completion of the drug infusion response. All animals were euthanized at the time data collection was completed (approximately 120 minutes from baseline). As this was an exploratory study, we wanted to relate graphically and through Pearson correlation analysis the IOPs from the different anatomic sites to the MAP and CVP. Waveform tracings were timed to the ECG signal and analyzed graphically. Pressures were averaged over 5-second intervals to obtain approximately 1,440 working data points per catheter site per animal over all experimental stages (baseline, nitroprusside, and norepinephrine). Analysis was performed using SPSS version 19.0. To model the behavior that we saw in the physiologic-phase diagrams, we developed a spreadsheet in Microsoft Excel that modeled the forcing function (aortic pressure) as a square wave and the following function (IOP) as the forcing function modified by an attenuator, a rate constant for exponential growth, a rate constant for exponential decay, and an offset. We fit the function to a minimum root mean square error by iteration using the Solver add-in. Results are reported in Data Supplement S1 (available as supporting information in the online version of this paper). Mean (±SD) baseline pressures were as follows: MAP = 66.5 (±9.3) mm Hg, CVP = 2.7 (±1.5) mm Hg, tibia IOP = 17.4 (±8.2) mm Hg, femur IOP = 18.4 (±3.8) mm Hg, humerus IOP = 15.7 (±1.8) mm Hg, and sternum IOP =5.7 (±0.5) mm Hg, except for one outlier that was 47.9 (±0.8) mm Hg and was excluded from further analysis. The IOPs of the long bones were 24% to 28% of MAP and 581% to 681% of CVP. Sternum IOP averaged within 3 mm Hg of CVP. IO needles were checked for patency at baseline by observing pulsatile pressures and comparing the shapes of MAP, CVP, and IOP waveforms at a 500-Hz sample rate. Representative waveforms from the same animal are shown in Figures 1 and 2. All of the IO sites had baseline pressures less than MAP, with the closest within 48 mm Hg of MAP. All of the IO sites had baseline pressures greater than CVP, with the closest within 3 mm Hg of CVP. After normalization for amplitude (Figure 2), examination of the sternal IOP waveform in relation to the ECG signal showed the IOP peaks correspond to the v and a waves of the CVP waveform. The Pearson product moment correlation coefficients between mean MAP and IOP over all experimental stages (baseline, nitroprusside, and norepinephrine) were tabulated (Table 1). Out of 40 IO sites, 15 (40%) had overall correlations in the range from 0.53 to 0.85. The best median correlation occurred between femur IOP and mean MAP (r = 0.65). The four highest correlations between IOP and MAP were associated with mean femur IOP. Only one IO site had a correlation coefficient over 0.50 for CVP. The long bones tended to correlate better with the MAP and the sternum tended to correlate better with the CVP in six of eight animals. In one animal, the sternum also correlated to the MAP. The nitroprusside response of MAP ranged from 39% to 71% of baseline with an average of 58%. The norepinephrine mean response ranged from 170% to 262% of baseline with an average of 215%. Representative responses from one animal are shown in Figure 3. Descriptive statistics for baseline, nitroprusside, and norepinephrine MAP responses for all animals are shown in Data Supplement S2 (available as supporting information in the online version of this paper). To further explore the relationship between MAP and IOPs over time, we constructed a phase diagram (Data Supplement S3, available as supporting information in the online version of this paper) and fitted the mean MAP to selected IOP measurements (Data Supplements S4 and S5, available as supporting information in the online version of this paper). Details of these results are provided in Data Supplement S1. Because IO catheterization is typically used when standard vascular access techniques are difficult or impossible, we wanted to explore the possibility of using IO catheters to provide an alternative monitoring capability when vascular access is otherwise unavailable or undesirable. Arterial and central venous catheterization both carry nontrivial inherent risks that may equal, or be even greater than, the risks of IO catheterization. Thus, clinicians would gain an additional option when selecting an invasive monitoring technique. Interestingly, Nakai and colleagues7 suggest that acute postoperative blood loss may correlate well with IOP and be a better predictor of clinical status than arterial blood pressure. We sought to characterize the relationship between IOP and MAP or CVP. Our model showed a weak correlation with CVP at the sternum and MAP at the femur. After normalization for amplitude, examination of the sternal IOP waveform suggests a relationship to the CVP waveform with v and a waves present. In 1964, Azuma studied 51 rabbits using various vasopressors to manipulate the systemic blood pressure.8 He found the IOP had a degree of correlation to the systemic blood pressure, although the IOPs in his model were notable for "undulations" or cyclical changes not reflected in the arterial blood pressures. He also showed that ligation of the femoral artery or vein proximal to the IO site rapidly resulted in reduction or increase of IOP, respectively. This study established the dependence of IOP on arterial input and venous outflow. Wilkes and Visscher,9 in 1975, used a dog model to measure femoral IOPs and found them to be approximately 17% of femoral arterial pressures under the fairly narrow range of pressures they studied. Of note, they surmised that the IOP was related to the nutrient vein end pressure. Unfortunately, "a relatively large number of animals were lost to the study because of technical difficulties." In 1979, Tondevold and colleagues,10 also using a canine model, demonstrated that the systemic MAP correlates well to the IOP, so long as the MAP is greater than 81 mm Hg. Below this value, the IOPs drop disproportionally compared to the MAP. The technique used to obtain IO access in this and all previous studies required drilling and the creation of bone membranes, and is markedly different than current IO needles in clinical use. Following the early investigations on intramedullary pressure waveforms, Herzig and Root11 showed a small but definite pulse pressure when they examined a cat model. Interestingly, they noted that systemic hemorrhage resulted in a lowering of IOPs, as did vasopressor infusions, the latter apparently due to local vasoconstriction. Unfortunately, the study had limitations including wide variations between animals relative to collection of baseline data, principally IOPs. Kiaer et al.12 showed a similar result in a model of rabbits. In a preliminary report Frascone and colleagues5 show the pulsatile waveforms associated with IOP may mimic the arterial pressure waveform. In a separate abstract, they show a relationship with respiratory variation and volume status.6 In our study, the observed pulsatile IOPs in rhythm with systemic pressures suggest that lower-frequency pressure changes will also be transmitted into the IO compartment. Differences within a bone may also affect IO properties. The canine study of Tondevold et al. also demonstrated that the systemic–IO blood pressure relationship was adequate whether the metaphyseal, diaphyseal, or epiphyseal space was cannulated.10 Although IO catheterization has been shown to work in bones lacking a medullary cavity, the effect on IOP is unknown.13 Because resistance varies across the vascular system of an animal and can vary between individuals, this likely played at least some role in the varying pressures we observed. Additionally, between IO sites, proximity to a major vessel and the size and shape of the medullary cavity influence the amplitude and shape of the waveform. It is unknown if the pressure correlation is degraded over time. In other words, it is not known if the correlation holds up for the duration the IO is in use, which is hours to days in current clinical practice. Our study concluded at approximately 2 hours and did not test beyond this time frame. Future work should focus on developing a model, linear or otherwise, that relates IOP to either MAP or CVP over a wide range of physiologic conditions. If a suitable model can be constructed and validated, it might be possible to clinically estimate MAP or CVP from selected IO sites. Our study is limited primarily by the animal model. The degree to which the physiology transfers to humans is not known, although porcine and caprine models are frequently used in IO studies because of the presumed similarity to humans. We also studied the relationship under laboratory-controlled ranges of systemic blood pressures induced primarily through vascular dilation and constriction. The applicability of the model to other physiologic derangements such as hyper- or hypovolemia is not known. It is also known that high concentrations of bradykinins injected locally into the osseous circulation can influence IOP in that bone.14 We used norepinephrine bitartrate and nitroprusside sodium, with injections through a central vein. It is likely the vasoactive agent was significantly diluted before reaching the bone vasculature. Because infusions were through central veins, it is unlikely that hydrostatic pressure from the injections played a significant role. Of course, an IO catheter serving to access the vasculature for both pressure measurements and infusions would be affected by hydrostatic pressure. Study variations in the site for measuring MAP and CVP (e.g., femoral or jugular vein) could contribute to small variations in pressure magnitudes between animals but should not affect correlations. We did not analyze other parameters such as cardiac output or serum lactate, which may indicate shock or other physiologic derangements, as we focused on the relationship between IOP, MAP, and CVP. Limb position can affect IOP but this is mitigated by the tethered limb positions in our study.15 Not all pressure relationships achieved high degrees of correlation. This could be a function of the experimental setup (e.g., clot formation in the IO catheter that degrades the recorded pressure waveform) or inherent limitations of the physiologic model relating IO and vascular pressures. Further experimentation is needed to refine the IOP relationship to the MAP and CVP and assess the potential of these measurements to provide clinically relevant information. The relationship between intraosseous pressure and mean arterial pressure or central venous pressure is variable by site, with the mean arterial pressure and central venous pressure tending to be estimated by the femur and sternum, respectively. The relationship to actively rising pressures is nonlinear, and a hysteresis model is proposed to explain the phase change. The authors are grateful to Bernard J. Rubal, PhD, for his generous help in the design and execution of the experiment and his thoughtful critique of the manuscript. The authors are grateful to Bernard J. Rubal, PhD, for his generous help in the design and execution of the experiment and his thoughtful critique of the manuscript. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.