Compliance Changes with Age

The pressure–volume relationships in the brain are expressed by the equation C ~ dV/dp, where C represents compliance and dV represents the change in volume that accompanies a change in pressure (dp) [Ryder et al., 1933]. Compliance is the ability of the brain to accommodate increases in intracranial volume without changes in ICP. As ICP rises, compliance decreases. In the healthy brain, compliance is along the horizontal segment of the curve, whereas in a damaged, swollen brain, compliance is located along the vertical segment of the curve.

The Monroe–Kellie doctrine is not as consistently applicable to infants as it is to adults [Wiegand and Richards, 2007]. In the developing brain, space-occupying lesions can expand locally in the relatively pliable, unmyelinated brain, exerting localized pressures that are not uniformly transmitted to the entire brain. The cranial cavity is not a rigid container in infants, as it can continue to enlarge by separation of the cranial sutures. The overall slope of the pressure–volume curve is steeper in infancy than in older children (Figure 77-2). An increase in intracranial volume by 10 mL is not likely to cause as much of an increase in ICP in an adolescent as in an infant [Beaumont et al., 2001]. This steeper volume–pressure curve persists in infancy until the point of separation of the cranial sutures. Thereafter, the infant’s cranial cavity is able to accommodate relatively more volume than is the adult skull.

Fig. 77-2 Pressure–volume curves of an infant and a 14-year-old, generated by injecting and withdrawing cerebrospinal fluid into the ventricular system.

Note that the curve is steeper in the infant than in the 14-year-old; this reflects the infant’s lesser ability to buffer increases in intracranial volume. ICP, intracranial pressure; V, volume.

(Adapted with permission from Shapiro K, Morris WJ, Teo C. Intracranial hypertension: Mechanisms and management. In: Cheek WR, Marlin AE, McLone DJ, et al., eds. Pediatric neurosurgery. Surgery of the developing nervous system. Philadelphia: WB Saunders, 1994;310.)

Effects of Intracranial Hypertension on Cerebral Perfusion

The first study without sedation of normal cerebral blood flow (CBF) in childhood was published in 1957 [Kennedy and Sokoloff, 1957] and used femoral artery and jugular vein cannulation with nitrous oxide inhalation and serial arterial blood gas sampling to derive CBF. The mean CBF (106.4 ± 3.3 mL/100 g/min) in the 9 subjects (ages 3–11 years) was significantly higher (p <0.001) than that measured in a group of 12 adults (60.1 ± 2.6) under the same conditions. Ethical considerations have limited subsequent studies in healthy children using such invasive procedures. Noninvasive ways of measuring cerebral blood flow volume (CBFV, measured in ml/min) to estimate CBF (i.e., adjusting CBFV data for the mean brain weight of each age range) also suggest decreasing rates of CBF from ages 3 (approximately 60), to 18 (approximately 50 mL/100 g/min) years following a transient increase at age 6 approximately 70 [Schöning and Hartig, 1996]. After a peak at approximately 6 years of age, values decrease before reaching adult values by age 19 [Chiron et al., 1992]. Increased ICP reduces CBF. This relation between ICP and blood flow is a function of changes in systolic (S) and diastolic (D) arterial pressure (AP) expressed as: CPP = MAP – ICP, or CPP = [(SAP + 2DAP)/3] – ICP (CPP, cerebral perfusion pressure; MAP, mean arterial pressure) [White and Venkatesh, 2008].

Cerebral Autoregulation

CPP reflects the vascular pressure gradient across the cerebral beds, and is determined by both the CPP and cerebrovascular resistance (CVR), according to the formula CBF = CPP/CVR [Harper, 1965]. Under normal conditions, CBF can be maintained as a constant of approximately 50 mL/100 g/minute over a wide range (60–150 mmHg in the healthy brain) of CPP [Steiner and Andrews, 2006; Cunningham et al., 2005] (Figure 77-3). When autoregulation is preserved, increasing blood pressure leads to a progressive compensatory decrease in the caliber of cerebral resistance, thereby maintaining a constant CBF by increasing CVR [Fog, 1938]. As CVR increases, cerebral blood volume decreases and ICP falls, if autoregulation is intact [Ursino and Di Giammarco, 1991]. When CPP falls below the lower limit of autoregulation (LLA) (40–50 mmHg), CBF decreases as the autoregulatory reserve is exhausted [Lassen, 1959], and the system becomes pressure-passive. The severity of the resulting ischemic injury will depend on the degree and duration of decreased CBF, with ischemic injury beginning at approximately 18–20 mL/100 g/min [Astrup et al., 1981; Zauner et al., 2002]. Conversely, when compliance is impaired and shifted to the steep segment of the pressure-volume curve, autoregulation is also impaired, and even minor increases in CBF – for example, because of hypercarbia, use of pressors, or fluid boluses – can lead to hyperemia and significant rises in ICP.

Fig. 77-3 Schema of cerebral autoregulation.

When mean arterial pressure (MAP) and cerebral perfusion pressure (CPP) are maintained in the normal range for age, cerebral blood flow (CBF) is constant. When pressure falls below the lower limit of autoregulation, or autoregulation is compromised by either increased intracranial pressure (ICP) or other metabolic derangements, the resulting fall in CBF causes ischemia and then infarction. Conversely, when MAP or CPP becomes too high, autoregulation fails and CBF becomes pressure-passive, resulting in hyperemia, increased cerebral blood volume (CBV), loss of compliance, increased ICP and cerebral tamponade.

Effects of Intracranial Hypertension on Autoregulation

Autoregulation of CBF in healthy adults is constant, between a CPP of 50 and 150 mmHg or a MAP of 60 and 160 mmHg [Paulson et al., 1990]. Under these conditions, changes in CPP or MAP have little effect on CBF. Autoregulation is compromised following TBI, and maintaining adequate CPP to overcome the effects of increased ICP is a cornerstone of medical management in the algorithms for treatment of TBI in adults [Brain Trauma Foundation, 2007]. In adults, the incidence of impaired cerebral autoregulation increased with the severity of injury, reaching 67 percent after severe TBI [Sahuquillo et al., 1998]. A similar relation is seen following pediatric TBI, ranging from 17 percent after mild injuries, up to 42 percent after severe injuries [Zauner et al., 2002; Paulson et al., 1990; Brain Trauma Foundation, 2007; Sahuquillo et al., 1998; Vavilala et al., 2004, 2007]. Following TBI, autoregulation may be compromised either unilaterally or bilaterally [Vavilala et al., 2008], and minimally or completely, and may occur without radiographic evidence of injury [Udomphorn et al., 2008]. Following severe TBI, dysfunction of autoregulation may worsen over 9 days following the insult [Tontisirin et al., 2007].

The LLA of CBF is the CPP at which vascular reactivity fails and CBF cannot be maintained despite arterial hypotension (see Figure 77-3). As CPP decreases below the LLA, the relation of CBF to CPP is passive and CBF falls linearly with CPP. The relationship of ICP to changes in the LLA is not well understood. In an atraumatic piglet model of increased ICP, increases in ICP up to 40 mmHg resulted in an increase of LLA from 29.8 mmHg in the naive animals up to 51.4 mmHg in the severe ICP group [Brady et al., 2009]. If this relation is shown to occur in children with intracranial hypertension, this implies, first, that monitoring autoregulation should be included in the management of these patients, and second, that treating acute increases in ICP with an equal increase in arterial blood pressure may be insufficient to maintain CBF as the LLA increases with increased ICP.

Loss of Autoregulation Leading to Brain Tamponade

CBF is thought to be mostly regulated at the level of precapillary cerebral resistance by a combination of glial, neurogenic, myogenic, and metabolic mechanisms [Iadecola and Nedergaard, 2007]. In 1938, Fog first reported changes in pial arteries in anesthetized cats in response to changes in blood pressure [Fog, 1938], describing a contraction of the pial vessels produced by a rise in systemic pressure, and a dilatation produced by a fall in pressure. Fog proposed that the mechanism responsible for such autoregulation was the properties inherent to the smooth muscle of the arterial wall. The relative contributions of the components of the arterial, venous, myogenic, neuronal, and glial components of the autoregulatory process are not fully understood.

As proposed by Rosner, the vasodilatory cascade theory postulates that the increase in ICP leads to a vasodilatory response [Rosner, 1986]. The resulting increase in cerebral blood volume produces an increase in vascular compliance. That is, during each cardiac cycle there is increased accommodation for blood volume within the cerebral vasculature. In parallel, intracranial compliance is decreased. In this way cerebral autoregulation is sustained by a combination of increased cerebrovascular, and decreased intracranial, compliance. When the increase in cerebral blood volume is greater than intracranial compliance can accommodate, autoregulation fails. At this point, during each cardiac cycle there is no reserve capacity in the cerebral vascular bed to accommodate variations in arterial blood volume during the cardiac cycle. As a result, CBF decreases and intracranial circulation fails, resulting in a condition of “brain tamponade.” This is the endpoint at which ICP values increase close to those of the arterial blood pressure, resulting in the absence of net flow.

History

ICP measurements were first reported in patients with a variety of intracranial lesions by Guillaume and Janny in 1951, using an electromagnetic transducer to measure ventricular fluid pressure signals [Guillaume and Janny, 1951]. Subsequently, Lundberg laid the foundation for modern ICP monitoring [Lundberg, 1960] by providing a method for ventricular cannulation that had a low risk of infection and leakage, enabled continuous recording of ICP, and was minimally traumatic. Lundberg described the following three basic patterns of physiologic changes in ICP:

Plateau or A waves: increases of ICP in the range of 540–650 mm H₂O (40–50 mmHg) that are sustained for 2–10 minutes, followed by a return to a higher baseline than before. They are indicative of a significant decrease in brain compliance. The A wave is of particular importance since these elevations in ICP may be sustained for long periods if left untreated.

B waves: brief elevations of a moderate degree (range of 272–408 mm H₂O [20–30 mmHg]); related to fluctuations in respiration and also indicative of decreased compliance.

C waves: small-amplitude fluctuations related to intracranial transmission of arterial pulses. ICP is most reliably monitored by the placement of an intraventricular catheter that is connected to a pressure transducer and a recording device or continuous strip chart.

Indications for Intracranial Pressure Monitoring

Following TBI, the decision to monitor ICP may be based on established criteria or expert opinion [Adelson et al., 2003a], and this is the best-established condition for acute monitoring in children. In other conditions in which ICP may be increased in children (e.g., meningitis, metabolic disorders, acute liver failure, stroke, brain tumors, cardiac arrest, hypoxic-ischemic encephalopathy, near-drowning), data are lacking both on any effects on outcome, and on the optimum range of ICP in these conditions. In these cases, the indications for ICP monitoring are subject to the judgment of the medical team.

Methods of Intracranial Pressure Monitoring

The clinical neurologic examination is the first, and essential, tool used to detect increases in ICP. In patients at high risk for increased ICP with conditions for which there are limited or no clinical guidelines (meningitis, acute liver failure, metabolic disorders, diabetic ketoacidosis), the decision whether or not to monitor ICP invasively will often need to be based on evidence for a decline in the patient’s neurologic examination. In contrast, the criteria for ICP monitoring in severe pediatric TBI have been published [Adelson et al., 2003a].

ICP monitors are classified both anatomically (ventricular, parenchymal, subdural, or extradural) and by mechanism (fluid-coupled or non-fluid-coupled) [Luerssen, 1997]. The external ventricular drain (EVD) remains the gold standard for reliable and accurate ICP monitoring [Figureaji, 2010; Padayachy et al., 2010]. This is a fluid-coupled system that provides real-time measurement of ICP, allows CSF drainage to control ICP, and can be recalibrated in situ. The device is zeroed at the level of the foramen of Monro and must be re-zeroed if the level of the bed is changed.

Other devices are used as alternatives for monitoring ICP, depending on the technical challenges of inserting an EVD into small ventricles. Placement of an intraventricular catheter can be technically difficult when severe brain swelling leads to compression of the ventricles or, in some instances, to ventricular displacement. Under these conditions, an intraparenchymal fiberoptic or electronic strain gauge system may be used. Subdural and extradural monitors cannot be calibrated in situ and are less reliable than intraparenchymal monitors [Gelabert-Gonzalez et al., 2006]. The most commonly used of these devices are the Codman (Codman, Raynham, MA, USA) microsensor [Koskinen and Olivecrona, 2005], the Camino (Integra Neurosciences, Plainsboro, NJ, USA) [Gelabert-Gonzalez et al., 2006], the Spiegelberg ICP sensor (Spiegelberg KG, Hamburg, Germany) [Lang et al., 2003], and more recently the Raumedic (Raumedic AG, Germany) ICP sensor [Citerio et al., 2008]. The risks associated with the use of ventricular catheters include hemorrhage and infection [Lozier et al., 2002]. The overall risk of ventriculitis is 1–5 percent [Lyke et al., 2001], and is associated with placement for more than 5 days or fluid leak around the catheter.

Given the potential risks associated with invasive ICP monitoring and the uncertain benefit for many pediatric disorders, the prospect of noninvasive monitoring is attractive. Several such methods have been proposed, including transcranial Doppler (TCD), changes in optic nerve sheath diameter, estimated middle-ear endolymph pressure, and visual-evoked potentials. None of these approaches is in routine clinical use [Wiegand and Richards, 2007]. The design of the TCD instrument is based on the Doppler effect to detect changes in frequency of a wave when the transmitter and receiver are in motion relative to one another. The TCD probe functions as the transmitter and receiver, and the wave is reflected back to the probe by moving blood. This was first reported in 1982 as being used to assess the vessels of the circle of Willis [Aaslid and Nornes, 1982]. TCDs can be used in children to estimate ICP, CBF, and CPP [Figureaji et al., 2009; Udomphorn et al., 2008], and to measure autoregulation [Tontisirin et al., 2007; Vavilala et al., 2005], but are not routinely used for estimation of ICP in clinical practice.

Duration of Physiologic Derangement and Outcome

Using age-specific values, Jones and colleagues examined the relation between the “dose” of abnormal CPP (duration of exposure to abnormal pressure) and favorable or unfavorable outcome [Jones et al., 2003]. The authors first constructed a “table of derangement” of core physiological values, based on their review of published literature, and organized these data by “bins” of single-year increments. They excluded children younger than 1 year and used a maximum age of 15 years. For example, this resulted in raised ICP being defined as ≥6 mmHg for ages 1–4 years, ≥8 mmHg for ages 7–8, and above the patient’s age for years 10–15. The lower limit for deranged ICP ranged from ≥6 mmHg for ages 1–4 years to ≥15 mmHg at age 15. For CPP, the thresholds ranged from ≤45 mmHg at age 1 to ≤58 mmHg at age 15. These values were then used in a prospective study of children with head injury admitted to a single center. Invasive monitoring was used as clinically indicated. Using these thresholds, the investigators calculated the percentage of time of “deranged physiology” of the total monitoring duration. The percentage duration of low CPP, but not raised ICP, was significantly associated both with mortality and with likelihood of poor (nonindependent) outcome.

Application of Receiver Operating Characteristic Curves to Determining Thresholds

As part of a larger prospective study of CPP and ICP thresholds affecting outcome following severe TBI, Chambers and colleagues recorded ICP and CPP, noting duration from 6.3 to 173 hours in 84 children (age 3 months to 16 years) every 2 minutes, and assessed outcome at 6 months using the Glasgow Outcome Scale score [Chambers et al., 2001]. Over each 1-hour interval, they calculated the maximum ICP and minimum CPP as a summary measure. They calculated Receiver Operating Characteristic (ROC) curves for ICP_max and CPP_min, and dichotomized outcome to independent (good recovery or moderate disability) and dependent (severely disabled, vegetative, or dead). These results suggested that the ICP_max threshold for independent outcome was 35 mmHg, and for CPP_min 43–45 mmHg.

In a later multicenter study of 235 children of ages 2–16 years with head injury requiring ICP monitoring, this group examined the association of mean ICP and CPP in the first 6 hours of monitoring with outcome [Chambers et al., 2005]. The cohort was divided into three groups – ages 2–6, 7–10, and 11–16 years. Three types of ICP monitoring devices were used among the five centers, and the mean values for ICP and CPP over the first 6 hours of monitoring were used for analysis. For each age group, there were significant differences in outcome (dichotomized as independent or poor outcome using the Glasgow Outcome Scale score at 12 weeks’ recovery). In each age group, the mean ICP was above 20 mmHg in patients with a poor outcome. The ICP value for differentiating poor and good outcome was different among the three age groups. For the youngest group, sensitivity was 100 percent at 37 mmHg, in contrast to 24 mmHg for ages 7–10 years and 29 mmHg in the 11–16-year group. For all three groups, a mean CPP of 50 mmHg or lower was 100 percent sensitive for predicting good or poor outcome.

These data suggest that there are age-dependent differences in ICP and CPP within the first 6 hours after TBI, and that these differences are associated with outcome. Nevertheless, these data do not indicate whether treating ICP or CPP improves outcome, nor what the optimal strategies for such management should be.

Area Under the Curve Analyses of Intracranial Pressure and Cerebral Perfusion Pressure Insults

A number of methods have been proposed to calculate the “dose” of ICP and the effect of this dose on outcome. In an early report, Marmarou and colleagues calculated the proportion of hourly ICP readings above 20 mmHg [Marmarou et al., 1991]. In this study of 428 patients in the Traumatic Coma Data Bank with ICP monitoring, the age range was 1–85 years, with 33 percent of cases between 20 and 30 years, but the pediatric cases are not analyzed separately. Others have used number of events, cumulative duration of events, and area under the curve (AUC) [Hemphill et al., 2005]. In a single-center study of 81 TBI patients aged 16–79 years, Elf and colleagues counted the number of insults above a specific ICP level (goal ≤20 mmHg; standard insult >25 mmHg; severe insult >35 mmHg) [Elf et al., 2005]. In this study, the investigators calculated the amount of a secondary insult as the time spent within the threshold level divided by the “good” (i.e., reliable) monitoring time for that patient. Although mortality was 11 percent in this study and only 7.5 percent had a good recovery at 6 months, less than 5 percent of time was spent with ICP above 25 mmHg, and this calculated variable was not associated with outcome. In contrast, Resnick and colleagues followed 38 patients with severe TBI for a year after the insult [Resnick et al., 1997]. They evaluated peak ICP, mean ICP, and the percentage of time spent with an elevated ICP, and found no association with outcome.

The validity of AUC as a management and predictive tool for relating ICP or CPP to outcome has been evaluated in a number of studies, some including children. A retrospective study by Vik and colleagues, of patients with severe TBI or Glasgow Coma Scale score above 8 with dilated pupils, included pediatric patients (age range 1–82 years), who were not analyzed separately [Vik et al., 2008