Kawati, Rafael MD*; Lattuada, Marco MD†; Sjöstrand, Ulf MD, PhD*; Guttmann, Josef PhD‡; Hedenstierna, Göran MD, PhD†; Helmer, Alois MD§; Lichtwarck-Aschoff, Michael MD, PhD*§
Partial obstruction of the endotracheal tube (ETT) is a notorious problem for ventilator-dependent patients (1), and detecting narrowing of the ETT or the circuit is essential with respect to patient safety and prevention of prolonged mechanical ventilation resulting from ETT-induced increased work of breathing. Narrowing translates to increases in peak airway pressure (Ppeak); consequently monitoring of Ppeak has become an integral part of the alarm system of all modern respirators. There are drawbacks, however. With adult ETTs and during volume-controlled ventilation, an increase in the relatively low inspiratory flow rate will induce minor increases in Ppeak only. Not until the flow (V̇) gets turbulent and hence, disproportionately more pressure is required for driving gas, would Ppeak reach a clinically relevant level. A potential alternative to Ppeak is monitoring the V̇ signal of passive expiration (2). Any obstruction acts as a brake on expiratory V̇, increasing the time required for emptying the lung. This can be described in terms of the expiratory time constants (τΕ) of consecutive segments of the expiratory V̇-volume plot (3). The decelerating effect of an ETT obstruction is most pronounced with the high V̇ of early expiration, increasing τΕ during early expiration and yielding a typical τE over expiratory volume pattern that is easy to recognize and independent of the inspiratory V̇pattern applied.
To test these assumptions, we calculated τΕ with three degrees of ETT obstruction in anesthetized, mechanically ventilated piglets. We hypothesized that the increasing emptying time of the lung, corresponding to increasingly higher τΕ, would indicate increasing grades of obstruction. The increase in Ppeak was assumed to be less pronounced and of clinical relevance at the most severe grade of obstruction only.
The study was conducted in accordance with the Helsinki convention for the use and care of animals. The local Ethics Committee for Animal Experimentation reviewed and approved the study. We studied 9 healthy male and female piglets of Swedish landrace breed (mean body mass 26.1 kg ± 1.4, mean ± sd). For n = 8 the probability was calculated to be 80% that the study would detect a difference in τE at a 2-sided 5% significance level (assuming that the true difference is ≥200 ms and the sd, as inferred from pilot studies, is 100 ms). To safeguard against potential experimental loss, we planned for 10 animals; however, one animal died before the experiment was finished and was excluded from the analysis.
After preparation and a recruitment procedure (tidal volume [Vt] doubled, positive end-expiratory pressure [PEEP] 10 cm H2O for 60 s), the lungs of the anesthetized piglets were ventilated (Servo 300; Siemens-Elema, Solna, Sweden) in volume-controlled mode (PEEP of 4 cm H2O, Fio2 0.4), and baseline measurements with the unobstructed ETT were performed. Thereafter, the ETT was stepwise obstructed with a graded external clamp (3 levels, defined as 2, 4, and 6 turns of the screw, respectively.
The expiratory volume-flow (V-V̇) curve was analyzed as described in detail previously (3). The V-V̇ curve was divided into 5 slices of equal volume, assuming a linear slope within each particular slice. To eliminate the effects of inertia and the opening and closing of the ventilator valves on the τE analysis, we excluded the first 5% and the last 5% of the data points. The slope of a straight line fitted to the expiratory V-V̇ curve within a slice (least squares fit) is equal to the derivative dV/dV̇, which has the dimension of time and can be directly read as the expiratory τE of that particular slice. The model assumes that the V̇-dependent resistance of the ETT is reasonably linear for narrow segments (slices) of the V-V̇ curve. The narrower the slices, the better this assumption is fulfilled, but the more pronounced is the impact of cardiogenic oscillations on the determination of τE. The choice of five slices is, therefore, a compromise dictated by the model assumption.
Ppeak and end-expiratory pressures were also recorded. Intrinsic PEEP was estimated from the difference between set PEEP and the pressure after a 5-s end-expiratory hold (4,5), and end-inspiratory pause pressure was determined also after a 5-s hold. Compliance of the respiratory system (Crs) was calculated from Vt divided by the end-inspiratory/end-expiratory pressure difference (intrinsic PEEP considered). Resistance was calculated according to the method described by Milic-Emili et al. (6).
Compressing a plastic ETT with a clamp creates a nonconcentric obstruction. With exactly the same setup used in the animal experiments, we therefore studied the impact of nonconcentric narrowing on the pressure-V̇ (P-V̇) curve in a physical model at Fio2 0.21 and 0.4. The resulting P-V̇ curves were compared to those of native ETTs with known inner diameter (ID) (7), which allowed for testing the assumption that the external compression, despite being nonconcentric, would render the P-V̇ curve more curvilinear, thus simulating the effects of a strictly concentric ETT narrowing.
The expiratory V̇ curves were assessed for intrapulmonary flow limitation with the method described by Lourens et al. (8) comparing two subsequent V̇ curves. If, with an added external obstruction, the second V̇ did not decrease, flow limitation was assumed to be present during the first condition. When V̇ differed by <5% during >50 ms (the time during which cardiogenic oscillations impacted on the V̇ curve), the respective segment of the V̇ curve was considered flow limited (8).
With the animals in the prone position the lungs were ventilated via an ETT (No. 8; Mallinckrodt, Athlone, Ireland) connected by a rigid 60-cm tubing system to a Servo 300 ventilator. Ventilatory frequency was 18–20 breaths/min, inspiration-to-expiration ratio was 1:1, and Vt was between 10 to 12 mL/kg, resulting in an Paco2 close to 5.5 kPa.
Anesthesia was induced with 100–200 mg ketamine and 1 mg/kg morphine and maintained with ketamine infusion (30 mg · kg−1 · h−1) plus morphine infusion (2 mg/h). Neuromuscular block was produced by continuous infusion of pancuronium 0.26 mg · kg−1 · h−1. Animals received sodium chloride 4.5 g/L with glucose 25 g/L at 10 mL · kg−1 · h−1 and a 5 mL/kg bolus of dextran-70 (Macrodex 70, Pharmacia Infusion AB, Uppsala, Sweden) to insure normovolemia.
Airway pressures were measured with a side port at the site of the pneumotachograph and fed to a pressure transducer. V̇ was measured with a heated Fleisch pneumotachograph placed between the ETT and the Y-piece. Signals were sampled at 200 Hz and fed to a data acquisition system (Acqknowledge, version 3.2.7, BioPac Systems, Inc, Goleta, CA).
All values are expressed as mean and 95% (lower to upper) confidence interval. Differences in variables within the groups were assessed using repeated-measures analysis of variance. Significant differences were evaluated using the Student-Newman-Keuls test. Statistical significance was assumed with P ≤ 0.05.
Despite the nonconcentric narrowing of the native ETT (ID 8 mm) created by the clamp, grade 1 obstruction produced a P-V̇ curve almost indistinguishable from an ETT with ID of 7.5 mm, grade 2 corresponded to an ID 7 mm, and grade 3 to an ID of 5 mm.
In the unobstructed tube and with grade 1 and 2 obstruction, τE was significantly higher in early and mid-expiration (slices 1–3), compared with late expiration (Table 1). With grade 2 and grade 3 obstruction, τE increased in early expiration (slices 1 and 2; P ≤ 0.05), whereas in late expiration (slice 5) τE increased significantly only in grade 3 obstruction (Table 1, Figure 1).
While τE increased with obstruction in all animals, Ppeak did not increase at all in 3 of 9 animals with grade 1 and 2 obstruction and by not more than 1.0 cm H2O in another 2 animals.
Ppeak did not increase significantly from its native level of 13 cm H2O (12.1 to 13.1) with grade 1 and 2 obstruction (Table 1, last column). With grade 3 obstruction, it increased to 20 cm H2O (16.7 to 23.2; P ≤ 0.05). Intrinsic PEEP was 2 cm H2O (0.5 to 3.5) in grade 3 obstruction.
Compared with the unobstructed tube, Crs decreased from 43 mL/cm H2O (38.8–47.8) to 37 mL/cm H2O (34.3–40.6), 38 mL/cm H2O (34.2–41.5), and 38 mL/cm H2O (32.4–42.8) in grade 1, 2, and 3 obstruction, respectively (P ≤ 0.05 to native ETT, Table 2, last column). Total effective resistance of the respiratory system (ETT included) did not increase significantly from its native level of 11 cm H2O · s · L−1 (9.3–12.6) with grade 1 and 2 obstruction. With grade 3 obstruction it increased to 29 cm H2O · s · L−1 (23.1–34.7; P ≤ 0.05).
Intrapulmonary flow limitation was not detected under any conditions, or in any animals Figure 2 shows a representative animal.
The main finding of this study was that, compared with the unobstructed ETT, the expiratory τE increased in all individual animals with ETT obstruction. The mean increase in τE was statistically significant for grade 2 and 3 obstruction. Ppeak increased with grade 3 obstruction, but with grade 1 and 2 the increase was not significant or, in individual animals, even absent.
Although the clamp created a nonconcentric narrowing, the resulting P-V̇ curve did not differ in shape from the one recorded from native ETTs with decreasing ID, the latter corresponding to a strictly concentric narrowing across the total ETT length. Therefore, grade 1 obstruction corresponded to an ID 7.5 mm, grade 2 to ID 7 mm, and grade 3 to ID 5 mm, suggesting that whatever the geometry of the narrowing, the P-V̇ curve will not differ in any fundamental aspect from the P-V̇ curve of a strictly concentric narrowing.
It was difficult to standardize the screwing of the clamp to obtain identical obstruction levels in all animals. We therefore cannot exclude that the levels of obstruction differed slightly from animal to animal. Grade 3 obstruction created an intrinsic PEEP of 2 cm H2O (0.5 to 3.5), which increased Ppeak in parallel. We assume that Ppeak would have performed even worse had intrinsic PEEP not been present.
The decrease in Crs from the unobstructed baseline level of 43 mL/cm H2O to 37 mL/cm H2O with all grades of obstruction was probably caused by the recruitment maneuver before the baseline measurement. Recruitment effects on Crs fade rapidly in healthy animals. In retrospect, it would have been better to repeat the recruitment before each level of obstruction to eliminate this time-related decrease in Crs.
The inspiratory pressure that is necessary to overcome the resistance of the ETT predominantly depends on the ID of the ETT and V̇. A critical V̇ threshold (V̇crit) can be roughly estimated (9,10), below which V̇ through a tube is laminar. The narrower the ETT, the lower is V̇crit. Not until narrowing comes down to an ID at which a new V̇crit is arrived at and V̇ gets turbulent, is a sudden and pronounced increase in Ppeak observed. From the equation used to calculate the Reynolds number (9,10) it can be derived that, assuming a V̇ of 9 L/min, the cross-sectional area of an ETT with ID 8 mm can be reduced by 36% or, with a V̇ of 6 L/min, by 72%, respectively, before V̇ gets turbulent. In the current study the ventilator delivered an inspiratory V̇ of 0.17 L/s (10.2 L/min) corresponding to the V̇crit for an ETT with an ID of 7 mm (9,10). In other words, Ppeak would not increase to any clinically detectable level until the ID is narrowed to 7 mm. Even with high turbulent V̇, pronounced obstruction is necessary to increase Ppeak to a clinically relevant level. Assuming a tube-related increase in inspiratory pressure of 3 cm H2O being the intervention threshold of the clinician, this threshold is only reached at a reduction in cross-sectional area of 61% at a V̇ of 300 mL/s or of 44% at 600 mL/s or of 23% at 900 mL/s, respectively.
Although a substantial degree of narrowing is required for the low V̇ of constant flow inspiration to reach V̇crit, the situation is different during expiration.
Expiratory V̇ exceeds by far the V̇ of constant flow inflation (0.90 L/s versus 0.17 L/s in this study) (Figure 2) and, thus, makes expiration relatively more sensitive to the decelerating effects of ETT obstruction. Actually, high V̇ also prevails during the inflation with a decelerating V̇ profile (pressure controlled ventilation, PCV). Monitoring Ppeak is, however, not meaningful with PCV because airway pressures are preset. Expiratory V̇, by contrast, is independent of any inspiratory V̇ pattern applied.
Moreover, our findings support the assumption that τE detects ETT obstruction earlier and more reliably than Ppeak. With grade 1 and 2 obstruction, Ppeak did not increase at all in 3 of 9 animals and by not more than 1.0 cm H2O in another 2 animals. The average Ppeak increase (Table 1) would probably escape clinical observation, whereas with grade 3 obstruction the pressure increase was very much amplified by the intrinsic PEEP.
Are there implications for the ventilator-dependent patient despite obviously quite different experimental conditions? Smaller Vts as used in the respiratory treatment of patients (10–12 mL/kg) require smaller V̇ than in the current experiment. The lower the actual V̇, the broader the range for the inspiratory V̇ to increase before becoming turbulent. Short tracheostomy tubes would reduce resistance, and the V̇crit would be arrived at later. Changing length or the ID of the ETT or inspiratory V̇ (by changing Vt, inspiration to expiration ratio, or frequency), all change either the threshold for V̇ to become turbulent or the difference between the actual V̇ and V̇crit. To avoid depending on the inspiratory V̇ pattern we suggest to analyze and compare subsequent expiratory V̇ curves. Expiratory V̇ is driven by the alveolar–mouth pressure difference and also depends on intrapulmonary resistance. Before concluding that ETT is partially obstructed, compliance and intrapulmonary airway resistance (assessing the V̇ curve for flow limitation) must therefore be determined. To minimize potential increases in intrapulmonary resistance ketamine, which is recognized as a potent bronchodilator, was used in the current study.
Kinking of tubings or blocking of the ETT by thick secretions or similar emergencies would certainly not go undetected with Ppeak monitoring (11). Lower grades of obstruction that develop slowly and progressively might, however, be overlooked. This is of little importance during controlled ventilation because the respirator provides the pressure for overcoming the resistive forces of the respiratory system, ETT, and circuit. With any type of assisted spontaneous breathing, however, a low-grade obstruction already imposes a resistive load on the patient (12) and this extra tube-related work of breathing is difficult to detect. We regard the current study as one step toward developing a noninvasive tool for assessing the patency of the ETT as it relates to spontaneous breathing, which is crucial.
We conclude that analyzing the expiratory V̇ signal during passive expiration detects partial obstruction of the ETT more reliably, earlier, and independently of the inspiratory V̇ pattern applied than does the Ppeak. The expiratory τE over volume plot has, therefore, the potential to serve as a monitoring tool during controlled mechanical ventilation.