Blood Gases: Technical Aspects and Interpretation
David J. Durand, MD Nick A. Mickas, MD
Arterial blood gas measurements are the gold standard by which the adequacy of oxygenation and ventilation are assessed. Arterial blood gas values can be directly measured from indwelling arterial catheters or estimated from intermittent peripheral artery punctures, arterialized capillary bed samples, and central venous blood samples. Continuous noninvasive monitoring devices, particularly pulse oximeters and transcutaneous carbon dioxide monitors, play an essential role in the respiratory management of critically ill newborns by giving ongoing estimates of blood gas values. The relative advantages and disadvantages of these techniques are discussed below.
Techniques for Obtaining Blood Samples
The most accurate arterial blood gas values are obtained from indwelling arterial catheters. Although it is possible to manage a sick newborn without arterial access, the presence of an arterial catheter often simplifies care significantly. It not only allows the accurate measurement of arterial blood gases without disturbing the patient but also allows direct measurement of arterial blood pressure and provides a route for obtaining other blood samples.
Umbilical Artery Catheters
Umbilical artery catheters are the preferred route for arterial access in most intensive care nurseries, particularly for infants in the first few days of life. They usually can be quickly and easily placed with small risk of complications. The umbilical arteries are readily accessible during the first several days of life and often can be cannulated in patients as old as 2 weeks.
An umbilical catheter should be flexible, nonkinking, radiopaque, transparent, and nonthrombogenic and should have an end hole but no side hole.1 There are two common catheter sizes, 3.5 French and 5.0 French. Some feel that the larger catheter should be used whenever possible to minimize problems with thrombus formation within the catheter, making it less prone to “clotting off.” Others feel that the smaller catheter is better because it minimizes the changes in aortic blood flow that occur when a catheter is in place. Because almost no published data is available about the relative merits of the two catheter sizes, the decision about which catheter size to use is usually based on personal preference. Our usual approach is to use a 3.5-French catheter in infants weighing less than 1500 g, and a 5.0-French catheter in infants who weigh more than 1500 g.
The procedure for cannulation of the umbilical vessels can be seen in an on-line video published by the New England Journal of Medicine; it is described in a 2008 article in the Journal.2
Prior to insertion, the catheter is attached to a three-way stopcock and syringe containing a heparinized saline solution and then flushed thoroughly. When the catheter has been inserted and is functioning adequately, the stopcock should be attached to a continuous infusion of heparin-ized fluid and to a pressure transducer. Care should be taken in stabilizing stopcock connections to minimize the possibility of accidental disconnection.
The catheter is inserted while the infant is under a radiant warmer or in a heated isolette where the infant’s temperature can be maintained and the vital signs monitored. The infant’s legs should be loosely restrained, and it may be helpful to also loosely restrain the arms. The insertion of the catheter should be done under sterile conditions, after the umbilical cord is cleaned with povidone iodine or chlorhexidine. A sterile umbilical tie is then placed around the lower portion of the cord and tied loosely with a single knot. The tie is placed so it can be either tightened if bleeding occurs when the cord is cut, or loosened if it prevents passage of the catheter. Next, the cord is cut approximately 0.5 cm above the skin. Cutting the cord with a scalpel in a single cut, rather than with a sawing motion, results in a flat umbilical surface from which the umbilical arteries usually protrude. The two thick-walled arteries and the single, larger thin-walled vein can easily be identified.
The most important step in the insertion of an umbilical arterial catheter is dilation of the arterial lumen. Failure to dilate carefully is the most common cause of catheter insertion failure. The goal of dilation is to open the lumen enough to allow smooth catheter passage without tearing the intima of the vessel. If the catheter tip tears the intima and creates a “false lumen” within the vessel, it will not reenter the lumen and successful catheter passage is nearly impossible. The dilation of the vessel should begin by placing one arm tip of a small forceps into the lumen. When using forceps with teeth, great care should be used to avoid shearing the intima. If done gently, the vessel will dilate, allowing both arms of the forceps to be placed into the lumen (Fig. 17-1). Once both arms have been placed, they can be slowly spread, gradually dilating the vessel to
Figure 17-1 Umbilical stump with two umbilical arteries and one vein. A small forceps is used to gently dilate one artery.
the caliber of the catheter. As the vessel lumen dilates, the forceps should be advanced with the goal of dilating at least 5 to 8 mm of the vessel. Once the vessel has been adequately dilated, the catheter can be inserted. It is easier to pass the catheter if the vessel is stabilized with one or two small curved forceps. Usually, the catheter passes smoothly. When the catheter meets significant resistance, it usually means that the catheter has dissected through the intima, and has created a false lumen within the wall of the vessel. When this occurs, the catheter should be removed. Forcing the catheter at this point is more likely to result in damage to the vessel or perforation of the peritoneum than to success.
On occasion, a catheter will travel down into the iliac artery, rather than up into the aorta. If this occurs, a second catheter can usually be inserted into the same umbilical artery, without removing the first catheter. With the first catheter lodged in the iliac artery, the second is often directed into the aorta.3
Once the catheter enters the aorta, it should be advanced to either “high position” or “low position.” The goal of both positions is to place the tip of the catheter so that it is not adjacent to the junction of the aorta and the renal, mesenteric, or celiac vessels. If a low position is chosen, the catheter tip should be between the level of the third and fourth lumbar vertebrae, safely below the renal and mesenteric arteries. If a high position is chosen, the catheter tip should be between the sixth and tenth thoracic vertebrae, above the origin of the celiac plexus. Although both positions are commonly used, several prospective randomized studies and a subsequent meta-analysis comparing low versus high catheter placement have found a greater rate of peripheral vascular complications in infants with catheters in the low position, however, most of these complications were minor.4
Several published graphs are useful for estimating the distance a catheter must be inserted to correctly place it in the lower position.5,6 The simplest method is based on the infant’s weight.7 For a 1-kg infant, the catheter should be inserted approximately 7 cm, for a 2-kg infant, it should be inserted approximately 8 cm, and for a 3-kg infant, it should be inserted approximately 9 cm. For a catheter to be placed in the high position, the formula “3 times the weight plus 9” gives a rough estimate of the required catheter insertion length in centimeters. For either method, the catheter position should be checked radiographically (Fig. 17-2).
Once correct position is confirmed, the catheter should be sutured and taped in place. We use a 3-0 or 4-0 silk suture tied in a “purse-string” around the circumference of the umbilical cord, then tied to the catheter. The catheter is then secured with a tape bridge.
Subumbilical Cutdown
If attempts to cannulate both umbilical arteries are unsuccessful, and the patient cannot be adequately managed without an umbilical catheter, the arteries can be cannu-lated via subumbilical cutdown.8 This is a surgical procedure and should not be attempted by anyone other than a surgeon who has previous experience with the technique. In our neonatal intensive care unit (NICU), this procedure has been almost entirely replaced by noninvasive monitoring and peripheral arterial catheterization.
With a subumbilical cutdown, the arteries are exposed through an incision approximately 1 cm below the umbilical stump. The subcutaneous tissues are dissected to the anterior rectus sheath, then the sheath is incised, and the rectus muscles are retracted laterally from the midline. The arteries are identified and separated from the urachus. Two sutures are placed around one artery, and a small arteriotomy is made between the sutures. The catheter is then inserted, and the distal suture is secured around the catheter and artery. The artery is tied off with the proximal suture, and the fascia and skin are closed. The position of the catheter should be confirmed radiographically.
Complications of Umbilical Artery Catheterization
Although umbilical artery catheterization is safe and well tolerated in most patients, it is important to remember that it is not without risks. Occasionally, catheter placement is associated with severe thrombotic complications, including frank gangrene and necrosis of the buttocks or leg. An echocardiographic study found intracardiac thrombi in 5% of infants with umbilical catheters.9 A more recent small study suggests that umbilical artery catheters in the first 5 days are not associated with a high risk of thrombosis.10
Infants with umbilical artery catheters in place will occasionally develop dusky or purple discoloration of their toes, presumably from microemboli or vasospasm. In some cases, warming of the contralateral leg may cause reflex vasodilation and increased perfusion in the compromised extremity. Although this is a common practice, a study in normal infants without vasospasm showed that local warming has no effect on peripheral blood flow to the contralateral heel.11 Regardless of whether there is any value in warming the contralateral foot, the compromised leg should not be warmed because of the risk that this might increase the metabolic rate of the warmed tissues, leading to increased hypoxic tissue injury. Although the majority of patients with dusky toes have adequate
umbilical artery catheter in “high” position. B, X-ray film showing umbilical artery catheter in “low” position.
perfusion and suffer no ill effects, one must always be aware of the risk that this represents potential significant vascular compromise. Failure to recognize worsening perfusion may result in necrosis and loss of a portion of the foot. If the toes remain dusky, with poor capillary filling, the catheter should be removed. Similarly, if the dusky discoloration involves more of the foot or leg, the catheter should be removed.
In rare instances, an infant with an umbilical catheter will develop blanching of the foot or part of the leg. Because blanching represents severely compromised arterial blood flow, the catheter should be immediately removed.
If perfusion to the limb does not immediately improve with withdrawal of the catheter, the infant should be evaluated for possible severe thrombotic complications. Evaluation in this case usually includes some combination of ultrasound or Doppler assessment, or even angiography. Both systemic vasodilators and topical vasodilators have been described as having some efficacy in this situation.12,13 When a significant clot is identified, there may be a role for treatment with tissue plasminogen activator, either infused directly into the effected vessel or systemically.9,14 The potential advantages of thrombolytic therapy must be weighed against the theoretical risks of such therapy, particularly in the infant with a preexisting intracranial hemorrhage that could potentially extend. Unfortunately, there is little literature available regarding the optimal approach to infants with severe vascular obstruction.
The incidence of infection associated with umbilical artery catheters appears to be lower than the incidence of infections associated with central venous catheters. However, as with all central catheters, meticulous care must be taken to maintain sterility during catheter insertion, and during subsequent withdrawal of blood from the catheter. A Cochrane review suggests that there is inadequate data to recommend either for or against routine antibiotic use in infants with umbilical catheters in place.15
Some centers avoid feeding infants with an umbilical artery catheter in place because of a theoretical concern that the catheter may interfere with mesenteric blood flow. A recent study evaluating practice in United States intensive care nurseries (US NICUs) revealed that 79% of respondents prescribe small-volume enteral feeds in infants with umbilical catheters, and that over 50% prescribe larger enteral feedings.16 There are studies evaluating blood flow with umbilical arterial catheters in place, including a recent study specifically measuring superior mesenteric artery flow, showing no impact on either mesenteric flow or blood flow velocity when feeding with an umbilical catheter in place.17,18
One of the most concerning side effects of umbilical artery catheters is the effect of blood sampling on cerebral blood flow. At least two studies have suggested that the routine blood sampling alters cerebral hemodynamics and oxygenation.19,20 This effect seems to be less with lower-position catheters than with high-position catheters. Markedly slowing the rate of withdrawal to 40 seconds seems to prevent the change in cerebral blood flow.21 Although it is unknown whether these changes in cerebral hemody-namics have any long-term effects, it seems advisable to be cautious about rapidly withdrawing from or infusing into any umbilical catheter.
There is little published data on which to base decisions about how long an umbilical artery catheter can remain safely in place. In some institutions, they are usually removed within several days. Other institutions maintain them for as long as 3 weeks. As with all therapies, the potential risks of umbilical artery catheterization must be balanced against the potential advantages for each infant.
Other Indwelling Catheter Sites
In approximately 10% of infants, umbilical artery catheter-ization is unsuccessful. In these cases, percutaneous can-nulation of a peripheral artery may be the best alternative. Percutaneous arterial cannulation is also the best option for infants who no longer have an umbilical artery cannula but still require arterial access. Other techniques, such as umbilical artery or peripheral artery cutdown, are more difficult to perform and involve more risk to the patient. Although percutaneous cannulation of a peripheral artery is technically challenging, especially in infants weighing less than 1 kg, cannulation of the radial, ulnar, dorsalis pedis, or posterior tibial artery is often possible. One should avoid cannulating the temporal artery because cerebral emboli and stroke have been reported in patients with temporal artery catheters.22,23
If the radial artery is to be cannulated, an Allen’s test should be performed to ensure ulnar artery patency. Conversely, if the ulnar artery is to be cannulated, radial artery patency should be assessed. Begin the Allen’s test by gently squeezing the hand to empty it of blood. Apply pressure to both the radial and ulnar arteries, then remove pressure from the hand and the artery that will not be cannulated. If the entire hand flushes and fills with blood, it is safe to proceed with cannulation.
The artery can be localized by either palpation or transil-lumination. If the radial or ulnar artery is to be cannulated, the hand should be restrained in mild hyperextension. We usually administer an analgesic dose of morphine or fen-tanyl to the infant before beginning the cannulation. Local anesthesia with lidocaine is less effective, and leaves a wheal over the area where one needs to feel the pulse.
The insertion site should be cleaned prior to proceeding with an iodine or chlorhexidine solution. The radial artery is usually most easily cannulated at the point of maximal pulsation over the distal portion of the radius, proximal to the superficial palmar branch of the artery. In this position, the artery lies between two tendons, superficial and lateral to the median nerve (Fig. 17-3).
The catheter can be used either dry or flushed with a heparinized saline solution. The catheter and needle are advanced at an angle of approximately 30 degrees until the vessel is entered and a pulsatile blood return is encountered. The needle is held stationary and the catheter is threaded into the artery. The needle is then withdrawn.
An alternative technique is to puncture the artery through both the anterior and posterior walls, then withdraw the needle. The catheter is then withdrawn until its tip reenters the vessel lumen and a brisk blood return is obtained, at which point it is threaded into the vessel. We have found that in some cases where there is blood return, but the catheter cannot be advanced, insertion of a small guide wire through the catheter into the vessel lumen will help guide the catheter into the vessel.
Once in place, the catheter should be taped securely and connected to an infusion of heparinized saline with a T connector and a three-way stopcock. The tape securing the catheter must allow for unobstructed view of all five digits because hypoperfusion, potentially leading to ischemic necrosis, is the major complication of peripheral arterial catheters.
Figure 17-3 Anatomy of the hand demonstrating radial and ulnar arteries and surrounding structures.
Infusion of Fluids Through Arterial Catheters
Patency of both central and peripheral arterial catheters should be maintained with a heparinized solution. In most centers, the heparin concentrations range from 0.25 unit/mL to 1.0 unit/mL. It does not appear that differences in the heparin concentration affect the incidence of intra-cranial hemorrhage.24 However, care must be taken to infuse the correct concentration of heparin, because overdoses have been reported from the use of an adult concentration. Although there is no published data on safe rates at which to run fluids through arterial catheters, we usually try to run peripheral catheters at 1 mL/hr and umbilical catheters at a rate of at least 1 mL/hr. We never run either peripheral or umbilical arterial catheters at a rate of less than 1 mL/hr.
Although saline, glucose, and hyperalimentation solutions can all be infused into an umbilical artery catheter, one study suggests that infusing an amino acid containing solution of normal osmolarity causes less hemolysis than does a quarter normal saline solution.25 In contrast to umbilical arteries where we have infused a wide range of solutions, we are concerned about the irritant effects of anything other than a physiologic saline solution infused into a peripheral artery. In small infants for whom 1 ml/hr of a physiologic saline solution provides an excessive sodium load, we sometimes infuse 0.45% saline. In cases where extra base is required, we infuse sodium acetate rather than sodium chloride. Medications or blood products are never administered through a peripheral arterial catheter.
Arterial Puncture
Blood gas samples can be obtained from intermittent puncture of the radial, ulnar, temporal, posterior tibial, or dorsalis pedis arteries. In general, the femoral and brachial arteries should not be used for arterial puncture because significant thrombus formation could lead to loss of the extremity, and median nerve damage has been reported with brachial puncture.26 As noted above, an Allen’s test should be performed before puncture of the radial or ulnar artery.
After the exact location of the desired artery has been determined by transillumination or by palpation, the skin should be prepared with a povidone iodine or chlorhexi-dine solution. A 25-gauge needle is inserted in the bevel-up position at a 45-degree angle through the skin, against the direction of the arterial flow. Blood should flow into the tubing spontaneously or with gentle suction. After the needle is removed, continuous pressure should be applied to the artery for 5 minutes. If hematoma formation is prevented, multiple specimens can be obtained from the same artery.
The main drawback to arterial puncture is that the procedure can rarely be done without disturbing the patient. One study showed that venipuncture, generally regarded as less traumatic than arterial puncture, caused a 6 mm Hg decrease in Paco2, and a 17 mm Hg decrease in Pao2. 27 Although subcutaneous administration of lidocaine (without epinephrine) over the artery before arterial puncture will provide partial analgesia, most infants still become agitated during the puncture. For this reason, we rarely use arterial puncture to obtain blood gasses.
Arterialized Capillary Blood
Arterialized capillary blood can provide a crude estimate of arterial blood values. In theory, blood flowing through a dilated peripheral capillary bed has little time for O2 and CO2 exchange to occur, making capillary blood gas values approximate those in the arterial blood.
Capillary samples can be obtained from a warmed heel or from the sides of the distal phalanges. To arterialize the capillary blood, the extremity should be warmed for several minutes. Warming should be performed with exothermic chemical packs specifically designed for arterializing capillary blood, rather than with warm compresses, which provide poor control over temperature. The site should be carefully cleaned, and a small lancet should be used to puncture the skin. When obtaining blood from the heel, the puncture should be made on the medial or lateral aspect of the plantar surface. The posterior curvature should not be used (Fig. 17-4).
There are multiple technical challenges to obtaining optimal capillary blood samples. Inadequate warming of the site will result in inadequate arterialization of the blood. Excessive squeezing will cause contamination of the “arterialized” blood with venous blood or interstitial fluid. Exposure of blood to air during collection will skew the Po2 and Pco2 values. Longer-term problems associated with capillary samples include calcaneal osteochondritis and calcified heel nodules.28 These calcified nodules may persist for several months to years, but do not seem to cause permanent problems for the infant.
Capillary puncture can be done only rarely without disturbing the infant. This, plus the fact that arterialized capillary blood is not the same as true arterial blood, means that a capillary blood gas represents only an approximation of the infant’s baseline arterial blood gas status. One study and review of the literature regarding capillary blood
Figure 17-3 Anatomy of the hand demonstrating radial and ulnar arteries and surrounding structures.
Infusion of Fluids Through Arterial Catheters
Patency of both central and peripheral arterial catheters should be maintained with a heparinized solution. In most centers, the heparin concentrations range from 0.25 unit/mL to 1.0 unit/mL. It does not appear that differences in the heparin concentration affect the incidence of intra-cranial hemorrhage.24 However, care must be taken to infuse the correct concentration of heparin, because overdoses have been reported from the use of an adult concentration. Although there is no published data on safe rates at which to run fluids through arterial catheters, we usually try to run peripheral catheters at 1 mL/hr and umbilical catheters at a rate of at least 1 mL/hr. We never run either peripheral or umbilical arterial catheters at a rate of less than 1 mL/hr.
Although saline, glucose, and hyperalimentation solutions can all be infused into an umbilical artery catheter, one study suggests that infusing an amino acid containing solution of normal osmolarity causes less hemolysis than does a quarter normal saline solution.25 In contrast to umbilical arteries where we have infused a wide range of solutions, we are concerned about the irritant effects of anything other than a physiologic saline solution infused into a peripheral artery. In small infants for whom 1 ml/hr of a physiologic saline solution provides an excessive sodium load, we sometimes infuse 0.45% saline. In cases where extra base is required, we infuse sodium acetate rather than sodium chloride. Medications or blood products are never administered through a peripheral arterial catheter.
Arterial Puncture
Blood gas samples can be obtained from intermittent puncture of the radial, ulnar, temporal, posterior tibial, or dorsalis pedis arteries. In general, the femoral and brachial arteries should not be used for arterial puncture because significant thrombus formation could lead to loss of the extremity, and median nerve damage has been reported with brachial puncture.26 As noted above, an Allen’s test should be performed before puncture of the radial or ulnar artery.
After the exact location of the desired artery has been determined by transillumination or by palpation, the skin should be prepared with a povidone iodine or chlorhexi-dine solution. A 25-gauge needle is inserted in the bevel-up position at a 45-degree angle through the skin, against the direction of the arterial flow. Blood should flow into the tubing spontaneously or with gentle suction. After the needle is removed, continuous pressure should be applied to the artery for 5 minutes. If hematoma formation is prevented, multiple specimens can be obtained from the same artery.
The main drawback to arterial puncture is that the procedure can rarely be done without disturbing the patient. One study showed that venipuncture, generally regarded as less traumatic than arterial puncture, caused a 6 mm Hg decrease in Paco2, and a 17 mm Hg decrease in Pao2. 27 Although subcutaneous administration of lidocaine (without epinephrine) over the artery before arterial puncture will provide partial analgesia, most infants still become agitated during the puncture. For this reason, we rarely use arterial puncture to obtain blood gasses.
Arterialized Capillary Blood
Arterialized capillary blood can provide a crude estimate of arterial blood values. In theory, blood flowing through a dilated peripheral capillary bed has little time for O2 and CO2 exchange to occur, making capillary blood gas values approximate those in the arterial blood.
Capillary samples can be obtained from a warmed heel or from the sides of the distal phalanges. To arterialize the capillary blood, the extremity should be warmed for several minutes. Warming should be performed with exothermic chemical packs specifically designed for arterializing capillary blood, rather than with warm compresses, which provide poor control over temperature. The site should be carefully cleaned, and a small lancet should be used to puncture the skin. When obtaining blood from the heel, the puncture should be made on the medial or lateral aspect of the plantar surface. The posterior curvature should not be used (Fig. 17-4).
There are multiple technical challenges to obtaining optimal capillary blood samples. Inadequate warming of the site will result in inadequate arterialization of the blood. Excessive squeezing will cause contamination of the “arterialized” blood with venous blood or interstitial fluid. Exposure of blood to air during collection will skew the Po2 and Pco2 values. Longer-term problems associated with capillary samples include calcaneal osteochondritis and calcified heel nodules.28 These calcified nodules may persist for several months to years, but do not seem to cause permanent problems for the infant.
Capillary puncture can be done only rarely without disturbing the infant. This, plus the fact that arterialized capillary blood is not the same as true arterial blood, means that a capillary blood gas represents only an approximation of the infant’s baseline arterial blood gas status. One study and review of the literature regarding capillary blood
Figure 17-4 Technique for obtaining arterialized capillary heel sample. Stippled sections denote correct areas for sampling.
gases concluded that capillary blood gases are “at best, only gross predictors of arterial values and, at worst, misleading assessments that may result in inappropriate management decisions.”29 We find that they are moderately useful for tracking gross changes in pH and Pco2. In an era of routine pulse oximetry, we find no value in tracking capillary Po2.
Continuous Invasive Monitoring
Over the last two decades, a number of devices have been developed for the direct intravascular or inline measurement of hemoglobin saturation, Po2, and Pco2. 30,31 However, despite their apparent advantages, these devices still have not made it into common use in most US NICUs, both because of their cost and complexity, and because of the ease of use of noninvasive technology. One study has demonstrated a reduction in the need for blood transfusions in premature infants using an inline blood gas analyzer.31
Errors in Blood Gas Measurements
Even small air bubbles in a blood gas sample can cause significant errors. Room air has a Pco2 of essentially zero and a Po2 of approximately 150 mm Hg. If air bubbles contaminate a blood gas sample, they lower the Pco2 and can either raise or lower the Po2, depending on whether the Po2 is below or above 150 mm Hg.32 One study showed that the amount of air that comes in contact with arterial blood drawn through a butterfly infusion set is enough to alter the Po2 measurement.33
Dilution of a blood sample with intravenous fluids lowers the Pco2 and increases the base deficit without affecting the pH. This effect is probably due to the diffusion of CO2 from blood into the intravenous fluid, which contains no CO2. 32,34,35 Because of the buffering capacity of the blood, the pH changes little, despite the decrease in Pco2, giving the appearance of a combined metabolic acidosis and respiratory alkalosis. Dilution of a blood gas sample with a lipid emulsion does not appear to have any effect on the blood gas measurements.36 Dry heparin does not appear to affect blood gas results.35
After blood is withdrawn from an artery, it continues to consume oxygen and produce carbon dioxide. Blood gas results may be inaccurate if the specimen is not processed promptly. Placing the sample in ice minimizes these changes. Immediate measurement, as with “point of care” techniques, also minimizes these changes.
Most blood gas analyzers measure Po2, then calculate the saturation, assuming that the blood sample is from an adult. However, if the sample contains a significant amount of fetal hemoglobin, the calculated saturation will be inappropriately low. If it is important to exactly measure the patient’s saturation, this should be done with a co-oximeter rather than with a standard blood gas analyzer.
Noninvasive Estimation of Blood Gases
The development of techniques for simply and safely obtaining continuous noninvasive estimates of blood gases was one of the most important advances in neonatal care of the last 30 years. Pulse oximeters are so ubiquitous in intensive care nurseries that many think oxygen saturation is as important a vital sign as heart rate or blood pressure. Although less widely used than pulse oximeters, with recent technologic advances, both transcutaneous monitoring and end-tidal CO2 monitoring have an important role in the management of neonates. Near infrared spectroscopy (NIRS) is a technology that is gradually moving from experimental to routine clinical use in selected infants.
Pulse Oximetry
Pulse oximeters work on the principle that saturated hemoglobin is a different color than desaturated hemoglobin, and thus absorbs light at a different frequency.37-40 A sensor, consisting of a light source and a photosensor, is placed so that the light source and photosensor are on opposite sides of an artery. As light passes through the artery and the surrounding tissues, the saturated and desaturated hemoglobin absorb different frequencies. By measuring the difference between light absorbed during systole and diastole, the amount absorbed due to arterial flow can be calculated. Then, by comparing the absorption at the two appropriate frequencies, the percentage of saturated hemoglobin can be calculated. Refinements of this system include complex algorithms for calculating exact saturation, and for separating arterial pulsations from motion artifact. The calculation of saturation is dependant on sensing light, so that ambient light striking the sensor can lead to a false reading.
In general, pulse oximeters provide excellent data about oxygenation in the physiologic range. However, the values they provide must not be accepted without care. Poor perfusion, ambient light, and motion all interfere with an adequate signal. Also, different manufacturers use different algorithms for calculating saturation, and so may give slightly different results. It is important to know that manufacturers are constantly updating the software in their devices, making many published articles on the limitations of specific devices out of date.
Pulse oximeters are dependent on an adequate arterial pulse. In situations such as shock, or if severe edema obscures the pulsatile arterial flow, the oximeter may not function reliably. Similarly, in patients on total support from venoarterial extracorporeal membrane oxygenation (ECMO) who have minimal arterial pulsations, we have found that pulse oximeters rarely function if the pulse pressure is less than 10 mm Hg.
The shape of the oxygen-hemoglobin dissociation curve (Fig. 17-5) makes it impossible for pulse oximeters to differentiate between degrees of hyperoxia. For example, a Pao2 of 80 and a Pao2 of 180 mm Hg both represent essentially 100% saturation in a preterm neonate. At least one study suggests that this is a significant limitation of pulse oximetry compared with transcutaneous oxygen monitoring, particularly in an era when avoiding hyperoxia to decrease the risk of retinopathy of prematurity is a significant concern.41 Pulse oximeters are also less accurate in the low end of the saturation range (e.g., less than 70% saturation) than in the normal physiologic range. Fortunately, this does not usually pose a clinically significant problem because the exact degree of severe desaturation is usually less important than the desaturation itself.
Figure 17-3 Anatomy of the hand demonstrating radial and ulnar arteries and surrounding structures.
Infusion of Fluids Through Arterial Catheters
Patency of both central and peripheral arterial catheters should be maintained with a heparinized solution. In most centers, the heparin concentrations range from 0.25 unit/mL to 1.0 unit/mL. It does not appear that differences in the heparin concentration affect the incidence of intra-cranial hemorrhage.24 However, care must be taken to infuse the correct concentration of heparin, because overdoses have been reported from the use of an adult concentration. Although there is no published data on safe rates at which to run fluids through arterial catheters, we usually try to run peripheral catheters at 1 mL/hr and umbilical catheters at a rate of at least 1 mL/hr. We never run either peripheral or umbilical arterial catheters at a rate of less than 1 mL/hr.
Although saline, glucose, and hyperalimentation solutions can all be infused into an umbilical artery catheter, one study suggests that infusing an amino acid containing solution of normal osmolarity causes less hemolysis than does a quarter normal saline solution.25 In contrast to umbilical arteries where we have infused a wide range of solutions, we are concerned about the irritant effects of anything other than a physiologic saline solution infused into a peripheral artery. In small infants for whom 1 ml/hr of a physiologic saline solution provides an excessive sodium load, we sometimes infuse 0.45% saline. In cases where extra base is required, we infuse sodium acetate rather than sodium chloride. Medications or blood products are never administered through a peripheral arterial catheter.
Arterial Puncture
Blood gas samples can be obtained from intermittent puncture of the radial, ulnar, temporal, posterior tibial, or dorsalis pedis arteries. In general, the femoral and brachial arteries should not be used for arterial puncture because significant thrombus formation could lead to loss of the extremity, and median nerve damage has been reported with brachial puncture.26 As noted above, an Allen’s test should be performed before puncture of the radial or ulnar artery.
After the exact location of the desired artery has been determined by transillumination or by palpation, the skin should be prepared with a povidone iodine or chlorhexi-dine solution. A 25-gauge needle is inserted in the bevel-up position at a 45-degree angle through the skin, against the direction of the arterial flow. Blood should flow into the tubing spontaneously or with gentle suction. After the needle is removed, continuous pressure should be applied to the artery for 5 minutes. If hematoma formation is prevented, multiple specimens can be obtained from the same artery.
The main drawback to arterial puncture is that the procedure can rarely be done without disturbing the patient. One study showed that venipuncture, generally regarded as less traumatic than arterial puncture, caused a 6 mm Hg decrease in Paco2, and a 17 mm Hg decrease in Pao2. 27 Although subcutaneous administration of lidocaine (without epinephrine) over the artery before arterial puncture will provide partial analgesia, most infants still become agitated during the puncture. For this reason, we rarely use arterial puncture to obtain blood gasses.
Arterialized Capillary Blood
Arterialized capillary blood can provide a crude estimate of arterial blood values. In theory, blood flowing through a dilated peripheral capillary bed has little time for O2 and CO2 exchange to occur, making capillary blood gas values approximate those in the arterial blood.
Capillary samples can be obtained from a warmed heel or from the sides of the distal phalanges. To arterialize the capillary blood, the extremity should be warmed for several minutes. Warming should be performed with exothermic chemical packs specifically designed for arterializing capillary blood, rather than with warm compresses, which provide poor control over temperature. The site should be carefully cleaned, and a small lancet should be used to puncture the skin. When obtaining blood from the heel, the puncture should be made on the medial or lateral aspect of the plantar surface. The posterior curvature should not be used (Fig. 17-4).
There are multiple technical challenges to obtaining optimal capillary blood samples. Inadequate warming of the site will result in inadequate arterialization of the blood. Excessive squeezing will cause contamination of the “arterialized” blood with venous blood or interstitial fluid. Exposure of blood to air during collection will skew the Po2 and Pco2 values. Longer-term problems associated with capillary samples include calcaneal osteochondritis and calcified heel nodules.28 These calcified nodules may persist for several months to years, but do not seem to cause permanent problems for the infant.
Capillary puncture can be done only rarely without disturbing the infant. This, plus the fact that arterialized capillary blood is not the same as true arterial blood, means that a capillary blood gas represents only an approximation of the infant’s baseline arterial blood gas status. One study and review of the literature regarding capillary blood
Figure 17-4 Technique for obtaining arterialized capillary heel sample. Stippled sections denote correct areas for sampling.
gases concluded that capillary blood gases are “at best, only gross predictors of arterial values and, at worst, misleading assessments that may result in inappropriate management decisions.”29 We find that they are moderately useful for tracking gross changes in pH and Pco2. In an era of routine pulse oximetry, we find no value in tracking capillary Po2.
Continuous Invasive Monitoring
Over the last two decades, a number of devices have been developed for the direct intravascular or inline measurement of hemoglobin saturation, Po2, and Pco2. 30,31 However, despite their apparent advantages, these devices still have not made it into common use in most US NICUs, both because of their cost and complexity, and because of the ease of use of noninvasive technology. One study has demonstrated a reduction in the need for blood transfusions in premature infants using an inline blood gas analyzer.31
Errors in Blood Gas Measurements
Even small air bubbles in a blood gas sample can cause significant errors. Room air has a Pco2 of essentially zero and a Po2 of approximately 150 mm Hg. If air bubbles contaminate a blood gas sample, they lower the Pco2 and can either raise or lower the Po2, depending on whether the Po2 is below or above 150 mm Hg.32 One study showed that the amount of air that comes in contact with arterial blood drawn through a butterfly infusion set is enough to alter the Po2 measurement.33
Dilution of a blood sample with intravenous fluids lowers the Pco2 and increases the base deficit without affecting the pH. This effect is probably due to the diffusion of CO2 from blood into the intravenous fluid, which contains no CO2. 32,34,35 Because of the buffering capacity of the blood, the pH changes little, despite the decrease in Pco2, giving the appearance of a combined metabolic acidosis and respiratory alkalosis. Dilution of a blood gas sample with a lipid emulsion does not appear to have any effect on the blood gas measurements.36 Dry heparin does not appear to affect blood gas results.35
After blood is withdrawn from an artery, it continues to consume oxygen and produce carbon dioxide. Blood gas results may be inaccurate if the specimen is not processed promptly. Placing the sample in ice minimizes these changes. Immediate measurement, as with “point of care” techniques, also minimizes these changes.
Most blood gas analyzers measure Po2, then calculate the saturation, assuming that the blood sample is from an adult. However, if the sample contains a significant amount of fetal hemoglobin, the calculated saturation will be inappropriately low. If it is important to exactly measure the patient’s saturation, this should be done with a co-oximeter rather than with a standard blood gas analyzer.
Noninvasive Estimation of Blood Gases
The development of techniques for simply and safely obtaining continuous noninvasive estimates of blood gases was one of the most important advances in neonatal care of the last 30 years. Pulse oximeters are so ubiquitous in intensive care nurseries that many think oxygen saturation is as important a vital sign as heart rate or blood pressure. Although less widely used than pulse oximeters, with recent technologic advances, both transcutaneous monitoring and end-tidal CO2 monitoring have an important role in the management of neonates. Near infrared spectroscopy (NIRS) is a technology that is gradually moving from experimental to routine clinical use in selected infants.
Pulse Oximetry
Pulse oximeters work on the principle that saturated hemoglobin is a different color than desaturated hemoglobin, and thus absorbs light at a different frequency.37-40 A sensor, consisting of a light source and a photosensor, is placed so that the light source and photosensor are on opposite sides of an artery. As light passes through the artery and the surrounding tissues, the saturated and desaturated hemoglobin absorb different frequencies. By measuring the difference between light absorbed during systole and diastole, the amount absorbed due to arterial flow can be calculated. Then, by comparing the absorption at the two appropriate frequencies, the percentage of saturated hemoglobin can be calculated. Refinements of this system include complex algorithms for calculating exact saturation, and for separating arterial pulsations from motion artifact. The calculation of saturation is dependant on sensing light, so that ambient light striking the sensor can lead to a false reading.
In general, pulse oximeters provide excellent data about oxygenation in the physiologic range. However, the values they provide must not be accepted without care. Poor perfusion, ambient light, and motion all interfere with an adequate signal. Also, different manufacturers use different algorithms for calculating saturation, and so may give slightly different results. It is important to know that manufacturers are constantly updating the software in their devices, making many published articles on the limitations of specific devices out of date.
Pulse oximeters are dependent on an adequate arterial pulse. In situations such as shock, or if severe edema obscures the pulsatile arterial flow, the oximeter may not function reliably. Similarly, in patients on total support from venoarterial extracorporeal membrane oxygenation (ECMO) who have minimal arterial pulsations, we have found that pulse oximeters rarely function if the pulse pressure is less than 10 mm Hg.
The shape of the oxygen-hemoglobin dissociation curve (Fig. 17-5) makes it impossible for pulse oximeters to differentiate between degrees of hyperoxia. For example, a Pao2 of 80 and a Pao2 of 180 mm Hg both represent essentially 100% saturation in a preterm neonate. At least one study suggests that this is a significant limitation of pulse oximetry compared with transcutaneous oxygen monitoring, particularly in an era when avoiding hyperoxia to decrease the risk of retinopathy of prematurity is a significant concern.41 Pulse oximeters are also less accurate in the low end of the saturation range (e.g., less than 70% saturation) than in the normal physiologic range. Fortunately, this does not usually pose a clinically significant problem because the exact degree of severe desaturation is usually less important than the desaturation itself.