Respiratory responses of the domestic fowl to low level carbon dioxide exposure

Research in Veterinary Science /986,40, 99-/04

Respiratory responses of the domestic fowl to low level carbon dioxide exposure L. S. ANDERSON, M. GLEESON, A. L. HAIGH, V. MOLONY, Department of Veterinary Physiology, Royal (Dick) School of Veterinary Studies, University of Edinburgh, Edinburgh EH9 IQH

Respiratory function in the restrained, conscious domestic fowl was measured using a non-invasive technique and computer aided analysis to examine respiratory flow, and in-dwelling arterial catheters to monitor blood carbon dioxide levels. The effects of two low ranges of inspiratory carbon dioxide (0' 2 to 1· 0 per cent and O' 25 to 2· 25 per cent) were studied, simulating levels of carbon dioxide that may occur in commercial poultry units and representing a mild environmental stress for the birds. A linear increase in minute volume with inspiratory carbon dioxide was observed, due primarily to increases in tidal volume. Arterial carbon dioxide (Paco2) tension also rose as the inspiratory carbon dioxide concentration was raised, but higher inspiratory carbon dioxide levels were required to affect significantly blood carbon dioxide concentration than to modify respiratory parameters. Variation was observed in the individual bird's response to carbon dioxide (bird X carbon dioxide interaction), suggesting that resting values of the respiratory parameters measured were important in determining the bird's ventilatory response to carbon dioxide.

al 1974, Davies and Dutton 1975, Powell et al 1978, Brackenbury et al 1982). However the changes in inspiratory carbon dioxide concentration used were relatively large (greater than 2 per cent carbon dioxide), there being few reported effects of changes in carbon dioxide concentration at lower levels. Osborne and Mitchell (1977, 1978) observed increases in minute volume as inspiratory Pco, was elevated from 0 to 21 mm Hg in steps of 7 mm Hg, though arterial carbon dioxide tension (paeo2) and [H't la remained unchanged. In this study, changes in respiratory function occurring during exposure to levels of inspiratory carbon dioxide in the ranges O·2 to I . 0 per cent (I ·4 to 7·0 mm Hg Peo2) and O· 25 to 2' 25 per cent (I' 7 to 15'8 mm Hg Peo 2) , and concomitant changes in Paeo2 , have been examined. Methods and materials

Animals Six female White Leghorn strain domestic fowl (Gallus domesticus) reared from day-old chicks to 15 weeks of age, were used. The birds were housed individually at 22 ± 1°C, with humidity and day length uncontrolled. Rec"ords of behaviour and bodyweight were kept, the latter ranging from I '08 to 1'19 kg.

RESPIRATORY function in domestic fowl may be affected by the levels of environmental pollutants present in commercial poultry units, before the appearance of clinical symptoms. A technique capable of detecting slight changes in respiratory function, with a potential application in subclinical diagnosis, has been developed. The technique consists of measurement of respiratory flow, using non-invasive methods and online computer analysis to yield various respiratory parameters relating time and volume, and associated data for further statistical analysis. The sensitivity of the system in detecting statistically significant changes in respiratory function has been established in studies of the effects on conscious birds of exposure to low levels of carbon dioxide, a potential environmental pollutant. Various studies of the effects of inspiratory carbon dioxide on avian respiration have been reported (Ray and Fedde 1969, Jones and Purves 1970, Bouverot et

Measurement of respiratory flow A non-invasive technique was used to obtain a respiratory flow signal for each bird. The bird was confined for one hour in a body plethysmograph (Bouverot et al 1974, Brackenbury et al 1982) comprising a perspex box with detachable lid, which enclosed the body of the bird, and an airtight seal (two greased flexible rubber collars) around the hole in the lid for the bird's neck. A second, head chamber could be attached to the lid of the box so that the bird could be exposed to gas mixtures of varying composition. A pneumotachograph (Fleisch number 0) was connected to a single outlet in the box, and coupled to a micromanometer (MDC FCOO2; Furness 99

100

L. S. Anderson, M. Gleeson, A. L. Haigh, V. Molony

Controls) which sensed the pressure differential across the pneumotachograph. The micromanometer output (mm HP) could be displayed on an oscilloscope and fed to a tape recorder (S.E. 8-4 portable; S.E. Labs [EMIl) for off-line computer analysis, or directly into a microcomputer (Cromemco System III microcomputer, Z80 CPU, 64K), for on-line analysis.

Processing of the respiratory flow signal The computer was programmed to convert the incoming analogue flow signal to digital form, the input channel being serviced by an interrupt system that allowed the signal sampling speed to be set in 64 Ilsec units. The analogue/digital convertor could resolve the signal to 20 mv in the range ± 2· 54 v; the signal typically oscillating ± 1· 5 v, With this type of input system, the digitised signal could not be transferred directly to disc, nor held in raw form in memory without rapid memory overflow. On-line reduction of the data held temporarily in a onedimensional array was thus necessary, performed as a queued background job while the interrupt system was inoperative. The baseline voltage was specified by the user, to allow for any offset or drift, as were threshold voltage levels on either side of the baseline (in 20 mv steps), for discriminating noise in the signal. Conditions set were such that the signal, as a digital representation of the signal voltage level, was summed from the baseline as it passed through the threshold, on either side of zero, summing being stopped as it returned through the threshold to baseline. The time taken for the half cycle was obtained from the number of data elements occurring during the voltage summing, this being a function of the sampling speed and thus time. The summed voltages and times for both positive and negative half cycles (ie, inspiratory and expiratory phases of a breath respectively) were stored in four data arrays. Up to 500 consecutive breaths could be held in memory during the on-line processing, then written to a datafile on disc.

Calculation of respiratory parameters A separate program accessed the datafile, read the four arrays into memory, then filtered out any spurious data produced by the movement of the bird in the plethysmograph during restless periods. Spurious breaths were recognisable by setting conditions to ignore two consecutive half breaths of the same polarity, or distinct, atypical summed voltages or times. The summed voltages were then converted to volumes (ml) using stored calibration constants, and time units expressed in seconds, thus yielding

inspiratory volume (VI)' expiratory volume (VE), inspiratory time (T I) and expiratory time (TE)' Up to 200 breaths obtained from the respiratory record, specified by the user as successive breaths or to be selected from the record at random, were used to calculate other parameters relating time and volume on an individual breath basis (total volume [VI + VEl, total time [TTl, tidal volume [VTI, frequency lfl, minute volume [V], mean inspiratory flow [MIFI, volume ratio [VI/VEl, time ratio [TI/TTl). For all variables the results were analysed to produce means (x), standard deviations (s), number of breaths used (n), summed values (Ex) and sums of squares (Ex2), and these values passed to a line printer. In addition, the mean breath shape was displayed graphically, together with frequency histograms of the flowrates occurring during processing of the respiratory flow signal.

Calibration The relationship between pressure and flowrate was established by statically measuring the pressure differential across the pneumotachograph at a range of known flow rates. The relationship was found to remain linear over the range encountered during measurement of resting hens, ie, below that rate giving turbulent flow, being described by an expression of the form Y = a + bX. The analogue to digital convertor within the computer was calibrated regularly, both in steady state, by applying known voltages to the input, and dynamically, using a waveform generator. Dynamic calibration of the whole system was achieved by attaching a respiratory pump (Minature Ideal Pump; Searle BioScience) whose volume and frequency could be adjusted, to the sealed plethysmograph. Computed values for time and volume were checked against the pump settings.

Statistical analysis From the results obtained for .each bird at a range of carbon dioxide levels, an analysis of variance was performed on the measured variables, having ascertained that there was homogeneity of variance, by using F tests on the variance ratios for different birds. The breath-to-breath (within bird) variation was employed as a pure error term in the analysis design, and used to test the significance of the other sources of variation. The individual breaths were randomly selected from the breath record to ensure that each breath was truly independent from any other. If any test of significance or confidence limit was to be valid the sample units had to be random, and successive breaths may well be correlated in some way, period-

Respiratory responses of the domestic fowl TABLE 1: Example of the analysis of variance for inspiratory volume. using 10 random breaths from six birds exposed to five levels of carbon dioxide in the range 0-2 to 1·0 per cent Source of variation

CO2 Birds Birds x CO2 Error

SS

df

MS

F

Prob

208·243 914·398 524·072 4262·968

4 5 20 270

52·061 182·880 26·203 15·789

3·30 11·58 1·66

0·05·0·01 <0·01 0·05·0·01

101

analyser (lL213; Instrumentation Laboratories) was used to measure Pco, of the samples, the instrument being recalibrated between each sample with known gas mixtures and buffers.

Protocol

Each bird in turn was placed in the plethysmograph, the head and upper neck being enclosed by a 5909·681 299 Total second chamber, and air passed through the chamber at 6 litres min -I (for carbon dioxide range O'25 to LSD= 1-96 y(2 x 15-789/60) = 1-42 (P =0-011 2· 25 per cent) or 10 litres min - 1 (range O'2 to l :0 per icities in breathing having been described by Priban cent carbon dioxide), using a compression pump and (1963). In the records used here, VI and VE were not a flowmeter with needle valve, up-line of the matched on an individual breath basis, the ratio of the chamber. The higher air flow rate reduced the carbon two variables changing so that there was overall unity. dioxide in the head chamber produced by the bird. For the three factors, breaths (10) x carbon dioxide Carbon dioxide was added to the air stream from a (five levels) x birds (six), the sources of variation in cylinder supply of pure gas via a second flowmeter the experimental design were separated into four and needle valve. After an initial 10minutes settling down period, the components, carbon dioxide (four degrees of respiratory flow signal was recorded for 10 minutes, freedom), birds (5 df), bird x carbon dioxide interaction (20 df), and an error term (270 df), giving a to obtain 100 successive breaths, at each carbon total about the mean with 299 df (Table I). The dioxide level in the ranges 0·25 to 2·25 per cent significance of the factors and interaction were (increased in steps of O' 5 per cent) and O'2 to I . 0 per derived by F tests of the variance ratio, where the pure cent (in steps of O'2 per cent). Carbon dioxide error term acted as the denominator. The least signifi- concentration in the head chamber was monitored cant difference (LSD) between mean values was with a gas analyser (Datex 'normocap' carbon dioxide calculated (at P=O'OI unless otherwise stated) from monitor), and adjusted accordingly, the gas analyser the error mean square term in the analysis, and being calibrated with 0 per cent and 4· 95 per cent used to determine significant ranges. Peco, was carbon dioxide in nitrogen (Specialty Gases). Six replicate birds were used, birds being randomly analysed as a randomised group design of analysis selected for exposure first to the wider carbon dioxide of variance. For comparison of individual mean values, range, and seven days later to the narrower range. Three of the six birds were cannulated after Student's t test and Duncan's multiple range test were used. Linear regression was performed to examine the' exposure to both ranges, and the above procedure correlations between carbon dioxide concentration repeated using a carbon dioxide range 0·2 to 2· 0 per cent. The cannula was fed out through the head and the variables measured. chamber exhaust port, to a double syringe arrangement flushed out with heparinised saline, one Surgical techniques syringe being used to retain dead space blood. One In order to obtain samples for blood gas analysis, millilitre blood samples were taken at the end of each the left carotid artery was occlusively cannulated 10 minute exposure period, for immediate blood gas under conditions of surgical cleanliness and halo- analysis. thane anaesthesia, in the caudal cervical region. The cannula was externalised behind the left auditory meatus and sealed after injection of heparinised Results saline. Antibiotic was administered to the wounds Effects of 0'25 to 2·25 per cent inspiratory carbon (Streptopen QR Cerate; Glaxo) in the form of dioxide on Vb V E, r, and TE drops. The birds were allowed to recover from the anaesthetic and returned to their usual housing. Examining first the results of the analysis of variance on the four primary parameters (Table 2), there were highly significant effects (P = 0'01) on VI, Blood gas analysis VEo T I and T E caused by raising the carbon dioxide One millilitre blood samples were removed via the concentration in steps from O'25 to 2· 25 per cent. The carotid cannula with a glass syringe during exposure means for six birds and 10 breaths per bird (Table 3) to each carbon dioxide level. Samples were stored in showed VI and VE to increase with rising carbon ice and analysed within 10 minutes. A blood gas dioxide level, the total increase in VI being 31 . 8 per

102

L. S. Anderson, M. Gleeson, A. L. Haigh, V. Molony

TABLE 2: Summary of analysis of variance - effects of five levels of inspiratory carbon dioxide on four respiratory parameters in six birds, using 10 random breaths from each bird. Two ranges of carbon dioxide used (1 =0,25 to 2·25 per cent, 2=0'2 to 1·0 per cent)

Source of variation (1)

CO2 Birds Birds x CO2

15,45** 24,69** 5,94**

3,77** 6,57** 1·41

3'54** 3,37** 3,96**

3,20* 5,85** 4·22**

3,30* 11,58** 1,66*

1·54 9,27** 0'96

7,00** 30,44** 0·73

4,95** 28,44** 2,59**

(2)

CO2 Birds Birds x CO2 Source of variation

MIF

(1)

CO2 Birds Birds x CO2 (21 CO2 Birds Birds x CO2

21'33** 36,44** 7,87**

46,82** 29,75** 5,52**

6'17** 6,72** 1,93**

1·36 2,31* 2,17**

2,46* 12,58** 1·42

2·04 27,24** 2,97**

1·35 21,85** 2,16**

0·52 6'89** 1·19

* P=0·05 ** P=O·Ol

cent and for VE, 43 per cent. VE increased over the whole carbon dioxide range, while VI did not change after reaching a maximum at 1· 75 per cent. The relationships between T lo T E and carbon dioxide concentration were more complex. Using the calculated least significant differences to discern significant changes in mean value, T I was observed to fall between 0·25 and 0·75 per cent carbon dioxide (by 8' 1 per cent), then rise again at higher concentrations, returning to a value not significantly different from that at 0·25 per cent. T E also dropped between 0·25 and O'75 per cent (by 13· 7 per cent), but then did not change significantly at higher carbon dioxide levels. Returning to the results of the analysis, the differences between individual birds were also found to be highly significant for these four parameters, over the range of carbon dioxide concentrations used. Furthermore, for V10 T I and T E' there was a highly significant interaction between birds and carbon dioxide, ie, different birds were reacting in different ways to the change in carbon dioxide concentration with respect to these parameters. There was no interaction for VE but this may have reflected the larger error term for this variable (ie, greater breath-tobreath variation). Effects of O'2 to l- 0 per cent inspiratory carbon dioxide on Vb V E, T[ and TE For the smaller carbon dioxide concentration range, the analysis of variance (Table 2) revealed highly significant effects of increasing carbon dioxide

level from O'2 to I' 0 per cent on T I and T E (P = 0'01), a significant effect on VI (P = 0'05), and no effect on VE. Both VI and VE rose with carbon dioxide concentration, VI by 17 per cent overall and VE by 17·5 per cent (Table 3). That such an increase in VE was not significant again reflected the large pure error term for this variable. T I was observed to increase over the carbon dioxide range, by 19·3 per cent, with an overall increase for T E of II' 7 per cent. Since this second range (0' 2 to 1·0 per cent) covered part of the wider range, the results obtained for T I and T E did not agree between the two studies. This may be associated with the bird x carbon dioxide interaction, in that the initial values of these variables may dictate the type of response to carbon dioxide that occurs. Again the differences between birds were highly significant for all four variables, but the interaction between carbon dioxide and birds was significant only for VI (P=0'05) and T E (P=O·OI). It was noted that the LSDs for both volume and time in the expiratory phase were nearly double those in the inspiratory phase, reflecting the larger breath variation in expiration than inspiration. Effects on other respiratory parameters Looking at respiratory parameters derived from the above variables, the effect of carbon dioxide on VT was significant in both ranges. Combining the bird means in both ranges (n = 10, each value being the mean for six birds with 10 breaths per bird), linear regression (Fig 1) revealed a straight line relationship between VT and carbon dioxide concentration TABLE 3: Mean values for six birds (10 breaths per bird) at five inspiratory carbon dioxide levels in two studies, with associated least significant difference (LSD)

VI VE TI TE VT 'iJ MIF f

0·2

CO2 concentration (%) 0·8 1·0 0-4 0·6

14·20 13'07 0·91 1·20

15·43 13·07 1·03 1·39

15·19 14·76 1-03 1·39

LSD

16·67 15·36 1·08 1·33

1-42 2·29 0·07 0·11

16·01 14·24 14·98 14·25 13·63 401·49 358·60 398·04 387'66 411·34 14·85 15·51 15·39 13·84 15·38 28·95 27·49 27'89 30·71 28·33

1-61 39·18 1'66 4·85

14·70 13·77 0·98 1·23

CO2 concentration (% I

VI VE TI TE VT

'it

MIF f

2·25 20·41 18'95 0·95 1·15

1·70 3·23 0·07 0·11

16·32 19·15 19·68 14·36 15·62 380·80 469·38 458·64 509·88 558·22 20·45 17·17 19·14 15·09 19·28 29·86 30·36 28·98 30·86 36·22

.1·35 26·65 2·36 6·78

0·25 15·49 13·24 0·95 1·31

0·75 16·01 15·24 0·87 1·13

1·25 17·46 15·18 0·97 1·17

1·75 20·60 17·70 0·99 1·23

LSD

Respiratory responses of the domestic fowl

103

30

~ ::1

I

16

14

0

121

...."...,,_ _----:,.-'-:-_ _~_=__,

,

1·4

2·0

CO2 concentration (%)

FIG 1: Effect of inspiratory carbon dioxide concentration on tidal volume (VTI. • = results from 0·2 to 1·0 per cent carbon dioxide range, 0 =0,25 to 2·25 per cent carbon dioxide. Each point is the mean value for six birds, 10 breaths per bird. LSD between mean values, in each range, denoted by bar. Line fitted by regression analysis (r = 0'974)

(r=0·974). Taking the calculated LSD for VT and interpolating on the fitted line (Y = 12·998 + 3' 055X), a change in carbon dioxide concentration of O'7 per cent was required to significantly increase VT at P=O'OI and 0·65 per cent at P=0·05. Over the range O'25 to 2· 25 per cent, carbon dioxide also had highly significant effects on minute volume and mean inspiratory flow, but no effect on respiratory frequency. V rose concomitantly with carbon dioxide, by 46·6 per cent over the whole range, while MIF rose at first, then remained constant. A significant bird x carbon dioxide interaction was observed for V, MIF and f, though the latter showed no overall changes due to carbon dioxide. Thus, each bird must have modified f in response to carbon dioxide, but in contrary ways. The lower carbon dioxide levels had no effect on V, MIF or f', but V and MIF showed significant bird x carbon dioxide interactions.

Effects of 0·2 to 2·0 per cent inspiratory carbon dioxide on arterial carbon dioxide The results of blood gas analysis revealed a significant effect on Paco 2 of increasing the ambient carbon dioxide level from 0·2 to 2·0 per cent. Using linear regression (Fig 2) on the mean values for replicate birds, a straight line relationship was observed between Paco 2 and ambient carbon dioxide concentration (r=0·861). The calculated LSD showed a change in Paco 2 of 2' 17 to 3·23 mm Hg was necessary to be significant (at P=0'05 to 0,01), and this interpolated (Y = 24' 97 + 1. 89X) as a change in ambient carbon dioxide level of 1. 2 to 1· 7 per cent. Discussion The methods employed to monitor respiratory function allowed accurate measurements of parameters to be made, and enabled the within bird variation to be partitioned out so that the effects of

24

•

~

!:-I

0'2

1~ 0

--::-,-"...-

1~ 8

--::,'";;-

~

CO2 concentration (%)

FIG 2: Effect of inspiratory carbon dioxide concentration on arterial carbon dioxide level (PaC02). Each point is the mean value for three birds. Least significant differences between mean values noted by bar. Line fitted by regression analysis (r = 0·861)

treatments, and differences between birds, could be examined. However, the need to sample breaths at random in order to estimate a pure error term could produce an imbalance in inspiratory versus expiratory volume due to variability in the latter from breath-tobreath, these parameters being equal when averaged over a number of successive breaths. The lower carbon dioxide concentrations chosen for this study may well occur in poorly ventilated commercial units, where hens are being reared intensively, and which might thus represent an environmental stress. We have shown that small changes in carbon dioxide level (0' 7 per cent) are liable to affect respiratory function in poultry, causing increases in tidal volume and hence minute volume, the relationship between VT and carbon dioxide level being linear. Such a finding is in agreement with a number of authors (Ray and Fedde 1969, Jones and Purves 1970, Powell et al 1978, Brackenbury et al 1982) who have observed increases in V in response to carbon dioxide caused primarily by a rise in VT . For all respiratory parameters examined there were significant differences between individual birds. However, the bfrd sample was homogeneous in respect of age, weight, strain, housing conditions and diet. Either poultry must be able to adopt resting values for respiratory parameters in a fairly wide range; or the birds respond differently to confinement in the plethysmograph during measurement, some perhaps more unsettled by the situation than others. The significance of a bird x carbon dioxide interaction might then be explained in terms of different 'resting' values of certain respiratory parameters, the variation in response to carbon dioxide from bird to bird being dependent on the initial value of the parameter in question. Ray and Fedde (1969), using anaesthetised unidirectionally ventilated chickens, reported increases in respiratory frequency with carbon dioxide, while Jones and Purves (1970) found no specific correlation. Brackenbury et al (1982) observed changes in

104

L. S. Anderson, M. Gleeson, A. L. Haigh, V. Molony

frequency with increase in carbon dioxide level that were dependent on the bird's temperature, f being initially higher at elevated ambient temperatures, and carbon dioxide decreasing frequency above a certain range of resting frequencies. They ascribed the finding of Osborne et al (1977), that raising carbon dioxide concentration decreased respiratory frequency, to a high initial value. Though lower carbon dioxide concentrations were used in our studies, there is agreement that the response to carbon dioxide is determined by resting respiratory conditions. Carbon dioxide did not affect respiratory frequency in either concentration range, yet significant bird x carbon dioxide interactions occurred. Examination of T 1 and TE for the individual birds revealed that over the carbon dioxide range 0'25 to 2'25 per cent, these parameters increased by up to 56 per cent in one bird, remained unchanged in two, and decreased by up to 36 per cent in three other birds. The largest increase with carbon dioxide occurred in that bird with the lowest values of T I and TEat O: 25 per cent carbon dioxide, and the greatest decrease when the initial values were high. Thus for T 1 at O' 25 per cent carbon dioxide of less than O' 95 seconds and TE of less than 1· 20 seconds, raising carbon dioxide concentration increased T 1 and T E. Conversely, if T 1 and T E were initially greater than these values, carbon dioxide decreased T 1 and T E• The observation that both Peco, and VT increased linearly with ambient carbon dioxide concentration does not agree with the finding of Osborne and Mitchell (1977) that there is little change in Paco 2 at Pico2 below 21 mm Hg, due to arterial homeostasis of PC02' However, it was observed that it took a larger increase in ambient carbon dioxide to alter Peco, (I . 1 to I' 5 per cent) than VT (0' 7 per cent), ie, that carbon

dioxide modified respiration before changes in arterial carbon dioxide. This may either reflect differences in sensitivity of the measurement of these variables, as a result of instrumentation or sampling, or support the finding of Osborne and Mitchell (1977) that respiratory function is affected by carbon dioxide levels insufficient to alter Paco 2 , and be attributed to the carbon dioxide sensitive intrapulmonary chemoreceptors.

Acknowledgements This work is supported by a grant from the Agricultural and Food Research Council. M. Shotter, A. D. Short and D. F. Cottrell are thanked for their help,

References BOUVEROT, P., HILL, N. & JAMMES, Y. (1974) Respiration Physiology 22,137-156 BRACKENBURY, J. H., AVERY, P. & GLEESON, M. (1982) Respiration Physiology 49,235-250 DAVIES, D. G. & DUTTON, R. E. (1975) Journal of Applied Physiology 39, 405-410 JONES, D. R. & PURVES, M. J. (1970) Journal of Physiology (London) 211, 295-309 OSBORNE, J. L. & MITCHELL, G. S. (1977) Respiration Physiology 31,357-364 OSBORNE, J. L. & MITCHELL, G. S. (1978) Respiratory Function in Birds, Adult and Embryonic. Ed J. Piiper. Berlin and Heidelberg, Springer-Verlag. pp 168-174 OSBORNE, J. L., MITCHELL, G. S. & POWELL, F. L. (1977) Respiration Physiology 30, 369-382 PRIBAN, I. P. (1%3) Journal ofPhysiology (London) 166,425-434 POWELL. F. L., FEDDE, M. R., GRATZ, R. K. & SCHEID, P. (1978) Respiration Physiology 35.349-359 RAY, P. J. & FEDDE, M. R. (1%9) Respiration Physiology 6, 135-143

Accepted June 6, 1985