|Home | About | Journals | Submit | Contact Us | Français|
The objective of this study was to determine the agreement between cardiac output measured by central (cranial vena cava) versus peripheral (cephalic vein) venous injection of lithium chloride for lithium-dilution cardiac output (LiDCO) determination in the dog. Five dogs (2 males, 3 females), anesthetized with halothane, were used. With each dog, 12 alternating central and peripheral LiDCO measurements were made, resulting in 10 paired comparisons. A total of 50 comparisons were obtained, the cardiac output measurements ranging from 1.11 to 2.76 L/min. The LiDCO measurement from the cephalic vein was similar to that obtained from the recommended central venous site: the difference between the central and cephalic vein determinations for all measurements was 0.098 ± 0.336 L/min (mean ± 2 standard deviations). Linear regression analysis demonstrated a slope of 1.050 (95% confidence interval 0.904 to 1.196) and a y intercept of 0.005 (r = 0.902). Therefore, although the central venous site is recommended by the manufacturer, the cephalic vein can be used instead in the dog, eliminating the need for central venous catheterization and thus reducing time and expense.
Lithium-dilution cardiac output (LiDCO), a new measurement in humans (1), has recently been demonstrated to strongly agree with cardiac output measured by the thermodilution technique in horses (2), pigs (3), and dogs (4). The LiDCO measurement technique is based on indicator-dilution methodology. The 2 indicator- dilution techniques most commonly used to measure cardiac output involve thermodilution and indocyanine green dilution. All of these methods measure cardiac output by determining the amount of dilution of a known amount of indicator administered into the venous blood, allowed to mix within at least 2 cardiac chambers, and assayed from an arterial site.
The LiDCO method involves injecting lithium chloride into a central venous catheter, positioned within the cranial vena cava or the right atrium, and measuring the concentration of diluted lithium at the site of a peripheral arterial catheter with a sensor that is selective for lithium. The LiDCO computer receives the lithium concentration over time from the sensor, constructs and analyzes the dilution curve, and determines the cardiac output. The manufacturer recommends the use of a central venous catheter, with the tip in the right atrium or the cranial vena cava (1). However, the use of a peripheral venous site for injections has several advantages: less morbidity, less time for catheter placement, and less cost. In addition, there may be situations in which a central venous catheter is contraindicated, such as when the patient has a coagulopathy or an increased risk for thromboembolism.
Currently, cardiac output is infrequently measured in a general clinical setting in veterinary medicine. Efforts to simplify cardiac output measurement in general and LiDCO measurements in particular should result in enhanced use in the clinical setting. LiDCO measurement is simpler than the thermodilution technique, because it does not require the placement of a Swan–Ganz catheter. The ability to perform LiDCO measurements from a peripheral venous site would further simplify the technique.
This study was designed to determine the accuracy of LiDCO measurements with the cephalic vein used for injection. It was expected that those measurements would be lower than measurements obtained after injection via a central venous catheter owing to the greater time for the lithium chloride to disperse from the peripheral site to the arterial site.
LiDCO measurement has recently been described in the dog (4). The following is a brief description of the operation of the LiDCO measurement system. The system includes a cardiac computer (LiDCO cardiac monitor CM 31-01 computer; LiDCO Limited, London, England) that receives information from a sensor as to the concentration of lithium in the blood sample, then constructs and analyzes the dilution curve to determine the cardiac output; the sensor and its housing; tubing to and from a roller pump; the roller pump; and a disposable blood-collection bag. The housing for the sensor includes an inlet port and an outlet port. The inlet port was attached to a 3-way stopcock connected to a catheter placed in a dorsal pedal artery. The sensor was prepared as described in the operation manual. The outlet port of the sensor was attached via tubing to the blood-collection bag. The tubing passed through a flow-regulator pump. Activation of the pump withdrew blood across the sensor from the dorsal pedal artery at a constant rate and expelled the blood into the collection bag. For each cardiac output measurement, the computer required the sensor constant (provided by the manufacturer), the injected dose of lithium chloride (in millimoles), and the hemoglobin and sodium concentrations in the patient's serum.
Administration of the lithium chloride involved the “park and ride” feature described in the operation manual: the dose was “parked” in a fluid line attached to the venous catheter and then flushed with 10 mL of heparinized saline into the right atrium (for a central venous injection) or the cephalic vein (for a peripheral injection) 7 s after activation of the computer, at end expiration in spontaneously breathing patients. The operation manual specifies that an ideal sensor signal should lie in the amplitude range of 0.2 to 0.8 mM, as indicated by the cardiac monitor. We used 1 mL of lithium chloride (0.15 mmol) for each cardiac output determination, which resulted in a signal amplitude of 0.6 to 0.8 mM.
This study complied with the guidelines of the Canadian Council on Animal Care and was approved by the Animal Care Committee at the University of Guelph. We used 5 young beagles (2 male, 3 female), with an average weight of 10.76 kg (range, 9.8 to 11.8 kg). They were assessed as healthy from their history and a physical examination. All instrumentation and LiDCO measurements were done with the dogs under general anesthesia.
Following premedication with butorphanol tartrate [Torbugesic, 0.4 mg/kg, given intramuscularly (IM); Ayerst Veterinary Laboratories, Guelph, Ontario], a peripheral venous catheter (Insyte-W, 22 gauge, 25 mm; Becton Dickinson, Sandy, Utah, USA) was percutaneously placed in a cephalic vein to facilitate the induction of anesthesia as well as the injection of lithium chloride. Thiopental hydrochloride (Pentothal, 20 mg/kg; Merial Limited, Iselin, New Jersey, USA) was injected into the catheter, the trachea intubated, and anesthesia maintained with halothane in oxygen (Halothane B.P.; Bimeda-MTC Animal Health, Cambridge, Ontario), with the vaporizer set at 1.5%. The dogs were allowed to spontaneously breathe throughout anesthesia. A peripheral arterial catheter (Insyte-W, 20 gauge, 40 mm) was percutaneously placed in a dorsal pedal artery, and a straight flush catheter (6 fr, 65 cm; Medi-tech, Watertown, Massachusetts, USA) was introduced via a jugular venous cutdown and advanced to the level of the 7th intercostal space (estimated level of the right atrium).
All the dogs were maintained at a normal depth of anesthesia, assessed by standard clinical parameters. Cardiac output determinations were obtained only after the dogs had achieved a stable anesthetic plane, also assessed by standard clinical parameters, at least 30 min after the beginning of halothane anesthesia. Prior to the start of cardiac output measurement, a 2-mL blood sample was obtained from the dorsal pedal arterial catheter for measurement of the hemoglobin and sodium concentrations, which were entered into the computer. Then, within 3 min, 12 cardiac output measurements were obtained, 6 using the central venous catheter (designated as central LiDCO) and 6 using the cephalic venous catheter (designated as peripheral LiDCO). The 1st measurement for each dog involved the central catheter; the injection site for the subsequent determinations was alternately selected. The sensor was changed after 6 determinations; thus, 2 sensors were used for each dog.
The 12 cardiac output measurements from each dog were considered for statistical analysis. Pairing observations obtained consecutively resulted in 5 pairs per sensor and 10 pairs per dog; each pair consisted of a central and a peripheral measurement.
Both the Bland–Altman graphical method (5) and linear regression analysis with SPSS software (SPSS, Chicago, Illinois, USA) were used to assess agreement between the 2 methods of measurement.
The cardiac outputs observed, 50 paired observations, ranged from 1.11 to 2.76 L/min. Overall, there was excellent agreement between the measurements using the 2 injection sites, as illustrated by the Bland–Altman plot (Figure 1). The bias and precision [mean of the difference between central and peripheral LiDCO ± 2 standard deviations (SDs)] was 0.098 ± 0.336 L/min.
Linear regression analysis of the same 50 paired LiDCO measurements demonstrated a slope of 1.050 [95% confidence interval (CI) 0.904 to 1.196] and a y intercept of 0.005 (Figure 2). The linear regression equation was y = 1.050x + 0.005 (r = 0.902). With this equation, if the LiDCO obtained with the peripheral venous injection was 1, 2, or 3 L/min, the estimated LiDCO (and 95% CI) with a central venous injection would be 1.055 (−0.715 to 1.395), 2.105 (1.765 to 2.445), or 3.155 (2.815 to 3.495) L/min, respectively.
As has previously been reported, the LiDCO is not a “gold standard”; however, in dogs, it is comparable to cardiac output measurements obtained with the thermodilution method (4). Our study demonstrates that LiDCO measurements obtained using a peripheral venous catheter in a cephalic vein are similar to those obtained using a central venous catheter, the technique recommended by the manufacturer. Using a peripheral venous catheter reduces both the time and the cost of catheter placement. Additionally, a peripheral venous catheter allows the assessment of cardiac output in patients for which a central venous catheter is contraindicated, such as those with a coagulopathy or a high risk of thromboembolism.
In our study, the bias from the Bland–Altman analysis demonstrated that there is little difference between the measurements obtained with the 2 methods, on average, but that the precision can vary, most determinations falling within ± 0.336 L/min (2 SDs). Similarly, the linear regression analysis demonstrated strong agreement, the r value being 0.902.
The Bland–Altman analysis revealed that the average difference between central and peripheral LiDCO was slight, in the order of 0.098 L/min. This difference is not likely to be clinically significant for these dogs. The peripheral LiDCO was, on average, lower than the central LiDCO, as expected, probably owing to greater dispersion of the lithium chloride prior to entering the heart. The increased dispersion would result in a longer transit across the lithium sensor and, hence, a longer indicator-dilution curve (concentration vs time). The area under this curve (AUC) is inversely related to the cardiac output; therefore, any factor that results in an increase in the transit time of the indicator across its sensor (i.e., increases the AUC) will cause a reduction in the cardiac output measurement. Several factors associated with a peripheral injection site — distance between the site and the sensor, degree of vasodilation, and venous volume — could all lead to increased dispersion of the indicator.
The ideal indicator for cardiac output measurement by an indicator-dilution technique would be one that is nontoxic. Lithium toxicity has been reported, in association with long-term administration, in the dog (6). The half-life of lithium is 21.6 h in mixed-breed dogs but 13.5 h in beagles (7,8). Toxicity would be a concern with repeated LiDCO measurement in a short period. However, a previous report suggested that as many as 34 measurements could be performed within 7 h without evidence of toxicity (9). Thus, even under conditions that require frequent LiDCO measurement, toxicity due to lithium accumulation is unlikely in the clinical setting.
Our experiment was performed on healthy dogs with a cardiovascular system presumed to be normal on the basis of history and physical examination. The potential for poorer agreement between central and peripheral LiDCO in patients with cardiovascular or circulatory derangement cannot be inferred.
In conclusion, the LiDCO measurement technique can be performed using the manufacturer-recommended central venous catheter or using a peripheral venous catheter in a cephalic vein. Cardiac output measurement has been greatly underused in veterinary medicine. Our findings should encourage its use in the clinical setting as a result of the simplicity of the LiDCO measurement method. The increased availability of cardiac output measurement data should improve our knowledge of therapeutic responses and provide more appropriate treatment, which could decrease the time that patients spend in the critical care setting.
The source of funding for this project was the Pet Trust Fund at the University of Guelph. The LiDCO system was donated by LiDCO Limited. We thank Amanda Hathway for her assistance with anesthesia.
Address correspondence and reprint requests to Dr. Douglas J. Mason, Animal Emergency and Critical Care Center, 1810 Skokie Blvd., Northbrook, Illinois 60062, USA, tel: 847-564-5775, fax: 847-564-5899, e-mail: dmason/at/ovc.uoguelph.ca
Received July 9, 2001. Accepted February 27, 2002.