Around 30 years ago, Sagawa and colleagues1
embarked on a systematic and detailed analysis of canine ventricular function using pressure–volume (PV) relationships. Their work led to the appreciation that such relations provided a uniquely powerful approach to quantifying heart function, particularly in vivo1-3
. Subsequent studies in large animals4
generated PV loops in real time, both under steady-state conditions and during transient reduction of inflow to the heart. This work established the methodology as the most comprehensive yet available for assessing ventricular performance independent from loading conditions, yet simultaneously quantifying load and the interaction of heart and vasculature. The most convenient way to obtain these data was the use of an impedance (or conductance) and pressure-measuring catheter, inserted to lie along the long axis of the ventricle, to provide a real-time volume signal as well as micromanometer pressure signal. This catheter was first used in large animals and humans starting in the mid-1980s. About 15 years later, technical development in miniature sensors made it feasible to apply this approach to very small mammals ()6
. This method provided simultaneous measurement of both pressure and volume signals from the intact beating mouse6-16
hearts. Despite its invasiveness, this sophisticated methodology has great potential for characterizing cardiac function in various genetically manipulated mouse models of cardiovascular disease, and testing the effects of various drugs under physiological and pathological conditions. Noninvasive methodologies for measuring cardiac function (echo and MRI) are limited by their application to steady-state conditions and reliance on motion parameters that can be influenced by loading conditions and thus lack specificity to the ventricle itself. Their advantage is that they can be repeated in the same animal and provide direct quantification of absolute volumes, whereas the conductance catheter signal is proportional to volume but must be appropriately calibrated to provide accurate absolute volume measurements.
Pressure–volume (PV) catheters and main steps of the protocol. (a) Mouse and rat PV catheters (magnified image) and working principle. (b) Flow chart indicates main procedures and important considerations of the PV protocol.
Ventricular pressure measurements have been commonly used for decades, but real-time volume measurements have historically been problematic. A technique by Baan et al
made it possible to correlate the change in ventricular volume to a change in electrical resistance of the blood pool within the LV chamber. The conductance catheter has multiple ring electrodes placed along its length (), and a high-frequency low-amplitude constant current is passed through the outer pair of electrodes to generate a local electric field between these electrodes (E4, E1). The field passes through the blood, muscle wall and surrounding structures, with field strength declining by the square of the distance from the electrodes. Electric theory indicates that if voltage potentials are measured within this field, they will be similar along planes that are perpendicular to the current field lines. The potential difference between two intervening electrodes will be inversely proportional to the amount of conductive material at that site. For the small rodent catheter, this measurement is made between two inner sensing electrodes (E2, E3), providing a time-varying signal. The resistivity of blood is about 1/3 that of the heart muscle, so the signal combines both the blood pool and chamber muscle wall. However, the latter is essentially constant12
, whereas the former varies with the cardiac cycle, so the time-varying component of the conductance signal is due to blood volume changes in the cavity.
As noted earlier, this conductance signal is itself noncalibrated. There is a fairly linear relation between absolute volume and this signal, and the slope (gain) and offset are related to the geometry of the heart and surrounding structures and their conductivity. Some have used external reservoirs filled with conductive material to mimic the heart to obtain a calibration curve (cuvette calibration; see also manufacturer's instructions). However, the accuracy of this approach may be questionable under all conditions or disease models. More accurate volume calibration can be accomplished using an independent measure of cardiac output (e.g., ultrasound flow probe) from which stroke volume is derived to calculate the gain of the signal, defined as: gain = flow probe stroke volume/conductance stroke volume. The offset is due to the fact that the ventricular cavity is not a perfect insulator, and a portion of the current leaks into the muscle. This offset must be subtracted to obtain absolute volume. It is usually estimated by the hypertonic saline dilution method5,12
The unique advantage of the PV methodology over all other available approaches to measure cardiac function is that it enables more specific measurement of the LV performance independently from loading conditions and of heart rate (for commonly used load-independent indices of systolic (Ees
–end-diastolic volume; preload-recruitable stroke work (PRSW)) and diastolic function (EDPVR), see . Ees
defines chamber end-systolic stiffness and can be a useful measure of contractile function, particularly to assess acute changes22,23
. Chronic changes in Ees
from heart disease can also reflect cardiac morphometry—that is, hypertrophy, fibrosis—and thus is not simply a reflection of ‘contractility’. The dP
–end-diastolic volume relation4
also provides a load-independent contractility index, as preload dependence of dP
is effectively reduced by using this regression. PRSW24
is a similar type of index, plotting stroke work versus end-diastolic volume for the set of load-altered loops.
Hemodynamic parameters and indices of systolic and diastolic function derived from PV relations in mice and rats.
The following protocol describes the procedures (summarized in ) for this method (anesthesia/analgesia, intubation techniques (), surgical techniques for LV catheterization (open and closed chest approaches; ), vena cava inferior occlusion methods () and calibrations (Figs. and ) to convert the raw conductance signals to true volumes), provides movies of the key processes/steps performed in our laboratories (Supplementary Movies 1
online), presents representative examples of PV loops and various calculated useful hemodynamics indices ( and Figs. -) and gives troubleshooting advice (see TROUBLESHOOTING).
Figure 2 Intubation techniques. (a) Shows a less invasive technique without tracheal incision which requires more experience; (b) shows a more invasive, but simpler technique. The technique shown in (b) is recommended, because the tracheal tube is more secured (more ...)
Figure 3 Surgical procedures for LV catheterization. (a) Closed-chest approach: insertion of the catheter into the LV through right carotid artery (see also Supplementary Movie 2 online). Sequentially numbered panels indicate stages of procedure. Image 11 shows (more ...)
Figure 4 Occlusion techniques, aortic flow measurements and jugular vein injection. (a) Vena cava inferior occlusion techniques. (b) Aortic flow measurements. (c) Jugular vein injection (see also Supplementary Movies 4 and 5 online). Sequentially numbered panels (more ...)
Figure 5 Representative examples of rat and mouse baseline PV loops and occlusions. Examples show representative (a) rat and (b) mouse PV loops before the calibration (in RVUs) and following cuvette and saline calibrations (in microliters) obtained by closed-chest (more ...)
Figure 7 Representative example of noise and effects of filtration of volume/pressure signal on PV relations. Left column: (a) baseline PV loops with noise showing regular pattern at volume channel, (b) the same following the use of a low-pass 60-Hz digital filter (more ...)
Anesthesia, body temperature control and intubation
The following injectable or gas anesthetics/analgesics can be used (see also ref. 25
- Ketamine (50 mg kg−1)
- Ketamine (50 mg kg−1 + fentanyl 250 μg kg−1)
- Ketamine/diazepam (40–80/5–10 mg kg−1)
- Ketamine/xylazine (80–100/10 mg kg−1)
- Chloral hydrate (300–400 mg kg−1)
- Alpha-chloralose (55 mg kg−1)
- Pentobarbital sodium (40–80 mg kg−1)
- Urethane (800–1,200 mg kg−1)
- Etomidate (5–10 mg kg−1)
Inhalants (induction 3–4%, maintenance 1.5% mixed with 100% oxygen)
Analgesia: morphine (1 mg kg−1) or fentanyl (50–250 μg kg−1)
Muscle relaxant: pancuronium (2 mg kg−1)
In obese animals, considerable differences may occur in the distribution of the injectable anesthetics, similar to that in aging animals or in animals with impaired liver function. Animals with various models of heart failure and shock may be more sensitive to the cardiodepressive effects of these agents. Even a slight overdose, especially with ketamine/xylazine or pentobarbital sodium, may profoundly decrease the heart rate and cardiac function. For most injectable anesthetics, the intubation of the animals will significantly improve the hemodynamic variables (this is especially critical with pentobarbital, which markedly increases mucus secretion in the respiratory tract). With the proper use and careful optimization, it is possible to achieve reasonable results with almost all anesthetic agents, and we will show examples with ketamine/xylazine, pentobarbital sodium, urethane+etomidate+morphine and isoflurane in our protocol (see ). However, because of the ease of overdosing and decreasing heart rate with some of the above-mentioned agents, we recommend using urethane + etomidate + morphine/fentanyl or isoflurane for anesthesia. There are reports describing extreme cardiodepressive effects of various anesthetics (e.g., ketamine/xylazine or pentobarbital), but most likely a significant part of these effects may actually be attributed to the lack of proper temperature control, intubation and overdosing.
Most of the above-mentioned agents require special handling because of the drug regulation laws and possible toxic effects. Urethane is carcinogenic (avoid contact with the skin) and prolonged use in animals may lead to hemolysis, making urine red. Similarly, avoid contact with and inhalation of gas anesthetics (especially methoxyflurane). Use downdraft table to avoid exposure to waste gases.
Surgical procedures for LV catheterization
For drug testing and more prolonged experiments, the closed-chest approach (see Step 8A, and Supplementary Movie 2
online) is more suitable because it is less invasive and animals are more stable for a longer period. An additional advantage of this approach in various animal models (e.g., in heart failure) is that arterial pressure records can easily be obtained from the carotid artery at the start or at the end of the experiment, allowing the calculation of total peripheral resistance (TPR) later (TPR = (mean arterial pressure–mean venous pressure)/cardiac output). The carotid approach should also be used in a chronic heart failure model induced by ligation of the left anterior coronary artery because of the scar formation in the apex area. Retrograde insertion via the LV apex does have some methodological advantages even for drug testing or other protocols as proper placement of the catheter in the LV is easier to confirm, and the procedure is done very quickly. In addition, if the carotid artery is severely atherosclerotic (e.g., in ApoE mice fed with a high-fat diet), or when the aortic valve is calcified (e.g., in advanced aging models) or in transverse aortic constriction (TAC)-induced hypertrophy and heart failure models, the open chest approach (Step 8B, and Supplementary Movie 3
online) is appropriate.
Conductance catheter calibration
The conductance signal is itself noncalibrated and must be carefully converted to absolute volumes if such information is required. The primary equation relating conductance to volume is V
), where ρ
is the blood resistivity, L
is the distance between sensing electrodes, G
is the conductance (measured as a voltage and utilizing a constant current circuit, this is directly proportional to G
is the parallel conductance due to conductivity of the muscle wall and surrounding tissues and α
is a gain coefficient (volume correction/calibration factor). The simplistic model of this approach is that the electric field is as if applied from parallel plates, so the current lines are straight and parallel to the catheter shaft. The fact that we use point source electrodes means that the field lines are curved, and this introduces a value of α
that is not unity, and some nonlinearity to the volume signal calibration. This nonlinear behavior is more problematic the larger the heart, and actually, the mouse heart may be the best designed for this technology, given the small distances from the catheter that are involved. In this mammal, the relationship between catheter and Doppler-measured stroke volume, for example, is highly linear over a broad loading range12
. The small heart and local electric field distribution also have implications for the parallel conductance, which can be quite large in larger mammals, as right ventricular (RV) volumes are clearly registered in the signal. In the mouse, Gp
appears mostly due to the muscle wall, and there is little far field (i.e., RV or other chamber volume) contribution. This was tested and previously demonstrated6,12
With regards to the actual calibration procedure, there are two approaches generally taken. One is to use mock-up cylinders with known volumes, and the catheter is placed in each using fluid matching the conductivity of blood (or blood), and a calibration curve generated. In our experience, this can yield inaccurate values as the cylinders do not exactly replicate the field distribution in the mouse heart, and calibrations can vary between different types of hearts (e.g., heart failure models, hypertrophy and infarction). Rather, we prefer the direct approach where both the gain and offset of the calibration are determined. The gain is assessed by measuring the cardiac output by Doppler flow probe. This can be done with a flow-velocity probe that can be placed on the ascending (or descending) aorta (Craig Hartley, Houston company) multiplied by a measure of cross-sectional diameter or using a volume flow probe (Transonics) that provides flow without requiring area assessment. For practical reasons, we prefer the latter and use the descending thoracic aorta as the placement position. Although this clearly excludes some relevant blood flow, the anatomy of the mouse (small upper body) implies that fairly little flow is missed, and this proportion is fairly constant among animals6,12
Data acquisition, analysis, noise filtering
In our experience, sampling data at 1–2 kHz is optimal; higher rates simply increase the file size. If concomitant electrical analysis is being performed, higher digitization rates would be useful. The default setting of the PVAN in PowerLab (PV analysis program) uses 1 kHz. The data analysis using PVAN module integrated into Chart is straightforward and quick (see the manufacturer's instructions). More experienced users may also create algorithms in chart to calculate various selected parameters, but as PVAN is included with the Millar PV system, this is not required. Some users have developed their own custom software for PV analysis7,9
Noise filtering may be a critical issue under some circumstances. In general, we do not recommend using noise filters. In 99% of cases, noise (mostly appearing at volume channel) can be prevented by very simple things (see TROUBLESHOOTING for details). In most of these cases, noise at volume channel actually originates from electric network and replugging the system into stabilized circuit/outlet may solve the problem. If the noise problem cannot be solved and the noise pattern is regular, most likely one of the filters built into PowerLab/Chart may improve the signal. In these cases, filtering of volume signal may be acceptable (this will only minimally affect derived parameters). However, do not try to filter the pressure signal because it may profoundly affect derived parameters (e.g., +dP\dt) (see for representative examples). However, if any filter is applied, it should be used for all conditions.