Pulmonary Mechanics and the Work of Breathing in the Semi-Aquatic Turtle, Pseudemys Scripta

It has been shown in mammals that for any given level of alveolar ventilation (VA) there is a combination of tidal volume (VT) and breathing frequency (f) which minimizes the mechanical power required for ventilation (Otis, Fenn & Rahn, 1950). More power is required to ventilate the lungs if breathing frequency is above or below this ‘optimal’ rate for a constant level of ventilation. The spontaneous breathing pattern of many mammals corresponds very closely with their predicted optimal pattern (Christie, 1953; Agostoni, Thimm & Fenn, 1959; Crosfill & Widdicombe, 1961; Yamashiro, Daubenspeck, Lauritsen & Grodins,. 1975).

In reptiles, normal breathing at rest is intermittent, unlike mammals where breathing is continuous. The reptilian pattern is characterized by single breaths or bursts of breaths separated by breath-holds of varying duration (McCutcheon, 1943 ; Gans & Hughes, 1967; Naifeh al. 1970; Gans & Clark, 1976; Glass & Johansen, 1976). The primary variable that changes if breathing is stimulated by hypoxia, hypercapnia or increased temperature is the breath-hold period (Glass & Johansen, 1976; Milsom & Jones, 1980; Benchetrit & Dejours, 1980), while the response to exercise in lizards is characterized by increases in tidal volume (Wood & Lenfant, 1976; Cragg, 1978). Very little is known about the relationship between the work of breathing and the breathing patterns or respiratory response of intermittent breathers.

The static compliances of the lung and total respiratory system have been measured in a few reptile species, and Perry & Duncker (1978, 1980) have attempted to correlate resting breathing patterns with these static compliances. Such analysis, however, is limited since respiration is an active rather than a static process. An understanding of dynamic as well as static pulmonary mechanics is essential for assessing the mechanical factors influencing breathing patterns.

Only one such study has been performed on a non-homeothermic species. Milsom & Vitalis (1984) have shown that, in the Tokay gecko, both the dynamic and static mechanics are most strongly influenced by the mechanics of the body cavity and chest wall. The unicameral lungs of this species are relatively large and compliant and offer little resistance to air flow. Most of the work required to ventilate the lungs is used to overcome elastic forces in the chest wall alone, regardless of ventilation frequency. Similar values are reported in the literature for the body wall compliance of lizards and mammals of similar size. The highly developed, alveolar lungs of mammals, however, are very stiff by comparison with reptile lungs (Crosfill & Widdicombe, 1961; Perry & Duncker, 1978; Milsom & Vitalis, 1984) so that pulmonary mechanics in mammals are more strongly influenced by the properties of the lungs. In mammals, work is required to overcome both elastic and non-elastic forces during breathing (Agostoni, 1970; Crosfill & Widdicombe, 1961; Otis et al. 1950). Despite these differences, there is still a combination of tidal volume and frequency in the gecko which gives minimum values for the mechanical power required for each level of alveolar ventilation.

The aim of the following experiments is to extend such studies of static and dynamic pulmonary mechanics to the turtle, Pseudemys scripta. The multicameral lungs of the turtle are less complex than mammalian lungs but more complex than those of lizards. The bony carapace of the turtle should make the body wall very stiff. Given this unique combination of features, it is of interest to see how the differences in the architecture of the lungs and body wall are reflected in the mechanics of the respiratory system and to what extent dynamic mechanics can be used in predicting likely patterns of ventilation in intermittent breathers.

The static pressure—volume relationships of the respiratory system were measured on eight turtles (726·3 ± 107 g body weight) maintained at room temperature (20–22°C) and anaesthetized with 20–30 mg kg⁻¹ sodium pentobarbital (Somnotol, MTC Pharmaceuticals). In all animals, once signs of reflex activity were absent, the cloaca was sutured shut to prevent water loss from the urinary bladder which would have altered body volume and thus the mechanics of the body cavity and body wall. The trachea was cannulated via a midline incision just below the larynx which was then sewn closed. The cannula consisted of a short piece of PE 220 tubing with a side-arm for pressure measurements. Starting with the animals in a supine position and the tracheal cannula open to the atmosphere, the lungs were first inflated, then deflated and reinflated to a volume of 11 ml 100 g⁻¹ body weight with an air-filled syringe. This volume corresponds to the resting lung volume (VLR), as determined for this species by Jackson (1971), and the pressures associated with VLR in both the body compartment and lungs under these conditions are referred to as zero pressure for these static pressure-volume relationships. We were confident that this technique of deflating the lungs with a syringe removed all gas from the animal’s lungs. Unlike the mammalian alveolar lung with its highly branched bronchial tree, the multi-cameral lungs of the turtle collapse completely when deflated. Any residual gas remaining within the lung is less than 5 % of VLR. Starting from VLR, the lungs were inflated in 5- or 10-ml steps until the inflation volume approached twice VLR. At this volume the pressure increase for each further step inflation became very large and the slope of the pressure-volume curve decreased dramatically. The system was then deflated in a stepwise fashion until the lungs were completely deflated, and then reinflated to VLR. The steady-state changes in intratracheal pressure associated with this inflation, deflation, reinflation sequence were measured with a Statham P23V pressure transducer and Gould d.c. amplifier and recorded on a Gould series 2600 pen recorder.

Following these measurements, the flank of each turtle, just anterior to the pelvic girdle, was pierced with a 13 gauge hypodermic needle through which a saline-filled cannula (PE 190) was inserted. The needle was then withdrawn and the cannula secured to the flank with a purse-string suture. This cannula was connected to the physiological pressure transducer for the measurement of intra-abdominal pressure during stepwise inflation, deflation and reinflation of the body cavity achieved by inflation and deflation of the lungs, starting from VLR, as described above.

In three animals, following the measurements outlined above, the plastron was removed with a necropsy saw, the pelvic girdle was carefully removed from the shell, and all the viscera including the digestive tract, liver and heart were carefully dissected from the lungs. Once the viscera had been removed, the lungs were inflated to VLR and the steady-state pressures associated with inflation, deflation and reinflation of the isolated lungs were recorded. The exposed lungs were swabbed periodically with saline throughout this procedure to prevent drying.

This series of experiments was performed on a further six animals with a mean body weight of 553 ± 81 g, anaesthetized and prepared as described for measurement of the static pressure-volume relationships. In the present series, the tracheal cannula was connected to a Harvard Small Animal Respirator (Harvard Inc., Millis, MA, USA) with a pneumotachograph (Fleish 00) placed in the airflow line leading from the pump to the tracheal cannula. The pressure drop across the pneumotachograph screen during tracheal airflow was measured with a Validyne differential pressure transducer (DP103-18) and this signal was electronically integrated by a Gould integrating amplifier to give the stroke volume of the pump. All measurements of pressure, flow and volume were continuously recorded on the chart recorder and the volume and pressure signals were also plotted simultaneously on an Esterline Angus 575 X-Y plotter. When volume was plotted against pressure on xy coordinates, through a complete pump cycle, a counter-clockwise rotating hysteresis loop was produced (Fig. 3). In three animals, inflation volumes ranging from 1 to 6 ml at pump frequencies ranging from 15 to 60 cycles min⁻¹ were used to produce a spectrum of dynamic pressure—volume loops that were within the animal’s range of minute ventilation and tidal volume. The selected volumes were absolute values and were kept constant for all animals throughout the experiment. Corrections for the weight of the animal were done during the calculation of mechanical work. All cycles started with the lungs at VLR, and pump volume always started at 6 ml and was sequentially reduced to 1 ml in 1-ml steps at each pump frequency. Frequency was sequentially increased from 15 to 20 cycles min⁻¹ and then to 60 cycles min⁻¹ in steps of 10 cycles min⁻¹. In all six animals, the intra-abdominal pressure and pump stroke volume were also plotted simultaneously, as described above, producing hysteresis loops for the changes in intra-abdominal pressure which accompany changes in the volume of the body cavity during inflation and deflation of the lungs.

The static compliances of the total system, body wall and the lungs alone were taken as the slopes of the linear portion of the respective inflation curves, beginning at the origin. Static compliances were standardized to weight and to VLR. The VLR standardized values were used to calculate the percentage of total work required to overcome the elastic forces within the lungs and the body cavity. Pressure-volume loops were analysed to determine the mechanical work required to overcome elastic and non-elastic forces during each pump cycle. Fig. 1 shows a schematic pressurevolume loop with arrows indicating the direction of change in pressure and volume for inspiration (I) and expiration (E). For each pressure-volume loop recorded in the present experiments, a perpendicular line was drawn from the volume axis to the point of peak pressure and volume (line BC in Fig. 1). The area contained within the pressure—volume loop (AlCEA = non-elastic work of inflation and deflation) and the area within AICBA (total work of one ventilation cycle) were measured and corrected for weight with an Apple II microcomputer and graphics tablet. During pump ventilation of the anaesthetized animals, expiration was always passive, and non-elastic work for each cycle in these experiments was calculated as half the area of the pressure-volume loop. The work required to overcome the elastic forces in the respiratory system was calculated by subtracting the non-elastic work from the total work in each cycle. These calculations of total, elastic and non-elastic work for each cycle were made on loops obtained for pump ventilation of the intact system and ventilation of the body cavity alone. The work required to ventilate the lungs alone was then calculated by subtracting the work required to ventilate the body cavity from that required to ventilate the total system. Dynamic compliance, the slope of the line connecting the two points of zero flow on the pressure-volume loop, was estimated by calculating the ratio (ΔV) and the corresponding change in pressure (ΔP).

The static pressure-volume curves for inflation and deflation of the total system, body cavity and lungs of a 639-g turtle are shown in Fig. 2. All curves are generally sigmoid in shape, indicating that in all cases increasing volume ultimately leads to a disproportionate increase in pressure as the elastic components of the system approach their limits of extensibility. All curves show a pronounced hysteresis between inflation and deflation.

The static compliances derived from the slopes of the mid-range portions of the various inflation curves originating from V and zero pressure are given in Table 1 for all animals examined. The mean.compliance of the total system (CT) was 8·4±0·7 ml cmH₂O⁻¹ kg⁻¹, that of the lungs alone (CB) was 35·l±0·8mlcmH₂O⁻¹ kg⁻¹ and that of the body cavity alone (CL) was l1·6±1·4mlcmH₂O⁻¹ kg⁻¹. These values are compared to published values for other reptiles and grouped according to lung type as classified by Perry & Duncker (1978) (Table 2).

The percentages of the total work required to overcome elastic forces arising from the body and the lungs on inflation of the intact respiratory system are shown in Table 3. In most reptiles the majority of the elastic recoil resisting inflation arises from the body wall, while the highly compliant lungs offer little resistance to inflation. The one exception is the chameleon, where forces are divided almost equally between the lung and body. In mammals, on the other hand, the majority of the work done to overcome elastic forces during inflation is required to expand the relatively stiff lungs, while only 25 % of the total elastic work is required to expand the tissues of the body wall.

The dynamic pressure-volume relationships of the intact respiratory system of a 750-g animal are shown in Fig. 3 while those obtained for the inflation of the body cavity alone in a 468-g animal are shown in Fig. 4. The curves in Figs 3A and 4A illustrate the effects of increasing stroke volume at a constant ventilation frequency, while Figs 3B and 4B illustrate the effects of increasing ventilation frequency at a constant pump volume. Note that whereas the pressure-volume relationships of the intact system are frequency-dependent, those of the body cavity alone are not. The relationships illustrated in Figs 3 and 4 for the total system and body compartment alone hold true for all pump frequencies at all inflation volumes.

Fig. 5 shows the changes in dynamic compliance of the total respiratory system, as a function of pump frequency, for three pump volumes. Statistically, the dynamic compliance is independent of pump volume over the ranges recorded, although there is a slight trend for the system to behave in a stiffer fashion at higher inflation volumes. The dynamic compliance is, however, strongly frequency-dependent: for any given volume, as frequency increases, the dynamic compliance falls. The decreases in dynamic compliance indicate that the system becomes stiffer with increasing frequency: thus, for any given volume, as frequency increases, the pressure difference generated by inflation also increases. This apparent stiffening of the respiratory system with increasing ventilation frequency is reflected in the work required to produce each breath. Calculated values of the mechanical work required to produce each breath for the total respiratory system, measured from the pressurevolume loops as described above, are shown in Fig. 6. For a constant frequency, increases in VP due to increases in VP result in a linear rise in W on these log-log curves, indicating an exponential increase in W. If VP is increased by holding volume constant and increasing frequency, W also rises but in a non-linear fashion. Thus the rise in W, with increasing levels of f at constant VP, is small at f < 30 cycles min⁻¹, and rises sharply with f > 30 cycles min⁻¹. The total work required to produce each breath can be subdivided into the work required to overcome elastic and non-elastic forces (Otis et al. 1950). Fig. 7 shows that, whereas the work required to overcome both elastic and non-elastic forces is roughly equal at low pump volumes and frequencies, the work required to overcome non-elastic forces becomes the major component of the total work at larger stroke volumes and higher pump frequencies.

The work required to produce a single breath can also be analysed in terms of the work required to inflate the lung and the work required to inflate the body compartment, since the pressure within the respiratory system is the sum of the pressures contributed by the lung and body wall (Agostoni, 1970). Fig. 8 shows the contribution of the lung and body wall to the total work required to produce a single breath at pump volumes of 2 and 6 ml with increasing pump frequencies. The work required to overcome the forces within the body wall are volume-dependent and increase with stroke volume. This component of the total work, however, is relatively frequency-independent. As a consequence, rises in the total work required to produce each breath as VP increases are a result of frequency-dependent increases in the work required to inflate the lungs.

The values of the work required to produce each breath from Fig. 6 are replotted in Fig. 9 as minute work (Ẇ), the product of the work required to produce each breath and the number of breaths produced each minute (f). For any given level of minute pump ventilation, Ẇ initially falls as f increases and VP decreases until Ẇ reaches a minimum. Further increases in f and decreases in VP produce a rise in Ẇ. The absolute levels of Ẇ increase and the shape of this curve becomes accentuated at higher levels of ⁠.

For values ranging from 100 to 300 ml min⁻¹, Ẇ is minimum at frequencies between 35 required to and 45 cycles min⁻1 Minute work is analysed in terms of the work overcome elastic and non-elastic forces for a of 200 ml min⁻¹ in Fig. 10B. As pump frequency increases, the minute work done to overcome elastic forces drops sharply and then stabilizes or rises very slightly as f continues to increase. The work required to overcome non-elastic forces increases slowly over the entire frequency range. The minute work required to overcome both elastic and nonelastic forces sum to give a resultant ‘U’ shape to the total minute work curvé. This effect becomes more pronounced as increases.

In Fig. 10A total minute work is broken down into the components required to ventilate the lungs and the body cavity. The minute work required to overcome the forces within the body cavity for a of 200 ml min⁻¹ steadily declines as f increases. The minute work required to overcome the forces within the lungs as pump frequency increases, however, declines initially and then rises, causing the total minute work curve to be ‘U’ shaped. Again these effects are more pronounced at higher levels of ⁠.

The compliance of the total respiratory system of the turtle, Pseudemys scripta, measured in this study, is in general agreement with results obtained for the same species by Jackson (1971). He reported values for lung compliances (170ml cmH₂O⁻¹ kg⁻¹), however, which were much higher than those measured in the present study (35·1 mlcmH₂O⁻¹ kg⁻¹). This is probably because he used different methods. In Jackson’s (1971) study, the compliance of the exposed lung was determined beginning at atmospheric pressure with the lungs completely collapsed. In the present study the static inflation curve was determined beginning with the exposed lungs inflated to VLR. As a consequence, CL was determined at a greater lung volume which may have corresponded to a stiffer portion of the pressurevolume curve.

The results of the present study further indicate that the compliance of the total respiratory system of this species of turtle is determined mainly by the compliance of the body wall. Only 23 % of the forces resisting inflation within this system arise from the lungs. Using the values for CT and CL reported by Jackson (1971), the elastic contribution is calculated at only 8 % of the total elastic forces resisting inflation. The percentage of elastic resistance to inflation offered by the lungs, as calculated in this study, lies well within the range of values calculated for other reptiles (8–44%), whereas the value calculated from Jackson (1971) lies at the lower extreme.

A comparison of compliance values for the total respiratory system, body wall and lungs of various species of reptiles indicates that the respiratory system of turtles is relatively stiff (Table 3). The lizards, for instance, have a system that is 2 to 30 times more compliant than the chelonian respiratory system. The relatively stiff respiratory system of the turtle can be attributed to the animal’s shell. The ribs of the turtle are fused to the dome of the carapace and are immobile. The only portions of the animal’s body capable of distension upon inflation are the flank cavities and the pectoral girdle. The lizards, on the other hand, have articulated ribs capable of moving as the lungs inflate and the volume of the body cavity increases. The multicameral lungs of the turtle, with their internal septation, are somewhat stiffer than the unicameral or paucicameral lungs of lizards (Table 3). Interspecific comparisons of compliance values normalized to body weight show that the compliance of the respiratory system of lizards is much greater than that of chelonians and that chelonians possess a more compliant system than mammals (CT for lizards is 2–30 times greater than CT for turtles which is 9–10 times greater than CT for rats). Comparisons made on a volume basis (normalized to resting lung volume) again show that lizards have the highest compliance values, but the chelonian and mammalian systems have similar values (CT for lizards is 5–23 times CT for turtles which is approximately equal to CT for rats). In mammals, however, CT is determined primarily by the lungs, which contribute 75 % of the elastic forces resisting inflation in the respiratory system. Turtles, on the other hand, have relatively compliant lungs compared to those of mammals and CT is predominantly determined by the body wall, which contributes 77% of the elastic forces within the respiratory system resisting inflation.

Perry & Duncker (1978) claim that high lung compliance in reptiles is associated with the degree of development of caudal and ventral dilations in the lung: the development of these membraneous regions is most extensive in the chameleon, followed by the Tokay gekko and savanna monitor (Table 3). The data of the present study, however, do not support this trend in the relationship between lung volume, general internal architecture of the lung and lung compliance. Whether the comparisons are made using values standardized to body weight or to resting lung volume, CT is much lower than would be predicted given the lung volume and lung architecture of this species. It may be the case that this trend is only found in the lizards or is merely a consequence of the small sample size used by Perry & Duncker (1978).

Under dynamic conditions, the compliance of the respiratory system shows frequency dependence, but little or no dependence on VP over the range studied. The frequency dependence of the dynamic compliance (C_dyn) must arise from a change in stiffness of the lungs with changing frequency, as there is no frequency dependence in C_dyn for the body cavity alone (Fig. 6). In the Tokay gekko, the respiratory system showed a frequency-dependent decrease in C_dyn which was enhanced by increasing ventilation volume at high ventilation frequencies (Milsom & Vitalis, 1984). This dependence of C_dyn on pump volume at high ventilation frequencies in the Tokay gekko was thought to be the result of the visco-elastic properties of the body wall. The absence of such an effect in the present study may be because the turtle has a shell and therefore does not experience geometric change with increasing volume, leaving the membranes of the flank cavities and the pectoral girdle as the sole visco-elastic elements. It is more probable, however, that the volume effects on the frequency dependence of C_dyn only appear at larger pump volumes. The pump volumes used in the present study were only 2–6% of resting lung volume, whereas in the study on the Tokay gekko the pump volumes used were 10–50% of resting lung volume. The dependence of C_dyn on frequency observed in the turtle means that a greater change in pressure will result for any given change in volume as frequency increases. This is reflected in the measurements of the total work breath^-1 obtained for the total respiratory system.

For any given pump volume, the work required to produce a single breath increases as frequency increases, with the slope of the rise in Ẇ being small at low frequencies but increasing sharply above 30 cycles min⁻¹. This is primarily due to an increase in the work required to overcome non-elastic forces within the lung. The work required to inflate the body wall and body cavity is relatively frequency independent and primarily required to overcome elastic forces. Thus, although at low pump frequencies the majority of work must be done to overcome elastic forces in the body wall, at higher frequencies the majority of the work is required to overcome non-elastic forces in the lungs.

For any given ventilation frequency, the work required to produce each breath increases exponentially as pump volume increases. The work required to overcome elastic forces in the body wall, body cavity and lungs accounts for approximately 50% of the total work at low pump volumes, but the work required to overcome non-elastic forces contributes the major portion to the total work of ventilation at high pump volumes because of the necessary increases in flow rates and thus flow resistance and viscous effects. The intrapulmonary bronchi and the trachea are the most probable sites for flow resistance in the respiratory system of turtles, and, even though turtles do not possess a highly branched bronchial tree as in mammals, it is not surprising that flow resistance is found in their lungs. Because of the large cross sectional area of the lower portion of the bronchial tree in mammals, little flow resistance occurs below the medium-sized bronchi, with the majority of flow resistance being in the upper airways (Pedley, Schroter & Sudlow, 1970). The minute work (Ẇ) required to maintain a constant level of minute ventilation is a product of the work required to produce each breath multiplied by the pump frequency. When Ẇ is expressed as a function of f for constant levels of the resulting curves demonstrate that there is a combination of f and VP at which the work of ventilation is minimal. At pump frequencies below this level, minute work increases as VP increases, primarily due to an increase in the work required to overcome elastic forces in the body wall. At pump frequencies above this level, minute work increases primarily as a result of an increase in the work required to overcome non-elastic (flow-resistive) forces in the lung.

The work required to inflate the lungs as a percentage of the total work of breathing increases in the sequence gecko > turtle > mammal. This hierarchy also reflects the increasing complexity of lung architecture. Given the simple structure of their lungs, the total work of breathing in the Tokay gecko is almost entirely required to overcome elastic forces arising from the body wall at all frequencies and tidal volumes (Milsom & Vitalis, 1984). In turtles, the mechanical work required to inflate the respiratory system is divided equally between overcoming elastic and non-elastic forces at low pump frequencies and volumes with almost all of these forces residing in the body wall and cavity. The work required to overcome non-elastic forces, however, predominates at high frequencies and tidal volumes, with the lungs contributing an increasing portion of this work. In mammals, the work required to inflate the respiratory system is primarily required to overcome elastic forces at low pump frequencies and non-elastic forces at high pump frequencies, but in both instances these forces reside primarily in the lungs (Otis et al. 1950). The present results and the above comparisons would suggest that, whereas differences in the structure and rigidity of parts of the body wall can have significant effects on pulmonary mechanics, far more profound effects arise in conjunction with differences in lung architecture. Although these differences were not evident from the static compliance curves, which only provide information on elastic forces, the dynamic measurements show that lungs of increasing complexity become progressively more important in determining the overall mechanics of the respiratory system. The role of mechanical factors in determining breathing patterns in turtles is complex. From the mechanical analysis one would predict that the most energetically favourable breathing pattern is one that occurs at 40 breaths min⁻¹. This prediction is based on the power curves for continuous ventilation presented in Fig. 9. Turtles, however, breath intermittently not continuously and therefore have a very low breathing frequency. The interpretation of intermittent breathing in the light of the present mechanical analysis is considered in the following study (Vitalis & Milsom, 1986).