Hypercapnia and surgical site infection: a randomized trial.

BACKGROUND: Tissue oxygenation is a strong predictor of surgical site infection (SSI). Mild intraoperative hypercapnia increases peripheral, gastrointestinal, and splanchnic tissue oxygenation and perfusion. Hypercapnia also has anti-inflammatory effects. However, it is unknown whether hypercapnia reduces SSI risk. We tested the hypothesis that mild intraoperative hypercapnia reduces the risk of SSI in patients having colon resection surgery. METHODS: With institutional review board approval and subject consent, patients having elective colon resection (e.g. hemicolectomy and low-anterior resection) expected to last >2 h were randomly assigned to intraoperative normocapnia (PE'CO2 ≈ 35 mm Hg; n=623) or hypercapnia ( PE'CO2 ≈ 50 mm Hg; n=592). Investigators blinded to group assignment evaluated perioperative SSI (Center for Disease Control criteria) for 30 postoperative days. SSI rates were compared. RESULTS: Patient and surgical characteristics were comparable among the groups. The SSI rate for normocapnia was 13.3%, and for hypercapnia, it was 11.2% (P=0.29). The Executive Committee stopped the trial after the first a priori determined statistical assessment point because of much smaller actual effect compared with the projected. However, because the actual difference found in the SSI rates (15-16%) were within the 95% confidence intervals (CIs) of the projected relative difference of 33% (95% CI -43 to +24%), our results cannot be considered as 'no difference', and cannot exclude a Type II error. Time to first bowel movement was half-a-day shorter in the hypercapnia group. CONCLUSIONS: Mild hypercapnia appears to have little or-possibly-no ability to prevent SSI after colon resection. Other strategies for reducing SSI risk should thus take priority.

CO(2) rebreathing: an undergraduate laboratory to study the chemical control of breathing.

The Read CO2 rebreathing method (Read DJ. A clinical method for assessing the ventilatory response to carbon dioxide. Australas Ann Med 16: 20-32, 1967) provides a simple and reproducible approach for studying the chemical control of breathing. It has been widely used since the modifications made by Duffin and coworkers. Our use of a rebreathing laboratory to challenge undergraduate science students to investigate the control of breathing provided 8 yr of student-generated data for comparison with the literature. Students (age: 19-22 yr, Research Ethics Board approval) rebreathed from a bag containing 5% CO2 and 95% O2 (to suppress the peripheral chemoreflex to hypoxia). Rebreathing was performed, and ventilation measured, after hyperventilation to deplete tissue CO2 stores and enable the detection of the central chemoreflex threshold. We analyzed 43 data sets, of which 10 were rejected for technical reasons. The mean threshold and ventilatory sensitivity to CO2 were 43.3 ± 3.8 mmHg and 4.60 ± 3.04 l·min(-1)·mmHg(-1) (means ± SD), respectively. Threshold values were normally distributed, whereas sensitivity was skewed to the left. Both mean values agreed well with those in the literature. We conclude that the modified rebreathing protocol is a robust method for undergraduate investigation of the chemical control of breathing.

Control of abdominal muscles.

Abdominal muscles serve many roles; in addition to breathing, especially at higher levels of chemical drive or at increased end-expiratory lung volumes, they are responsible for, or contribute to, such protective reflexes as cough, sneeze, and vomiting, generate the high intra-abdominal pressures necessary for defecation and parturition, are active during postural adjustments, and play an essential role in vocalization in many species. Despite this widespread involvement, however, their control has, with rare exceptions, received little attention for two major reasons. First, in most anesthetized or decerebrate preparations, they are relatively inactive at rest, in part because the position of the preparation (supine or prone with abdomen supported), reduces lung volume and, therefore, their activity. Second, unlike phrenic motoneurons innervating the diaphragm, identification of motoneurons to a particular abdominal muscle is difficult. At the lumbar level, a given motoneuron may innervate any one of the four abdominal muscles; at the thoracic level, they are also intermixed with those innervating the intercostals. The two internal muscles, the internal oblique and the transverse abdominis, respond more to increases in chemical or volume-related drive than the two external muscles, the rectus abdominis and external oblique; the basis for this differential sensitivity is unknown. Segmental reflexes at the thoracic and lumbar levels are sufficient to activate abdominal motoneurons in the absence of descending drive but the basis for these reflex effects is also unknown. Neuroanatomical experiments demonstrate many more inputs to, and outputs from, the nucleus retroambigualis, the brainstem region in which the premotor neurons are located, than can be accounted for by their respiratory role alone. These other connections likely subserve activities other than respiration. Studies of the multifunctional roles of the abdominal muscles, on the basis of recent work, hold considerable promise for improving our understanding of their control.

Phrenic motoneuron discharge during sustained inspiratory resistive loading.

I determined whether prolonged inspiratory resistive loading (IRL) affects phrenic motoneuron discharge, independent of changes in chemical drive. In seven decerebrate spontaneously breathing cats, the discharge patterns of eight phrenic motoneurons from filaments of one phrenic nerve were monitored, along with the global activity of the contralateral phrenic nerve, transdiaphragmatic pressure, and fractional end-tidal CO2 levels. Discharge patterns during hyperoxic CO2 rebreathing and breathing against an IRL (2,500-4,000 cmH2O.1-1.s) were compared. During IRL, transdiaphragmatic pressure increased and then either plateaued or decreased. At the highest fractional end-tidal CO2 common to both runs, instantaneous discharge frequencies in six motoneurons were greater during sustained IRL than during rebreathing, when compared at the same time after the onset of inspiration. These increased discharge frequencies suggest the presence of a load-induced nonchemical drive to phrenic motoneurons from unidentified source(s).

Phrenic afferents and ventilatory control at increased end-expiratory lung volumes in cats.

The role of phrenic afferents in controlling inspiratory duration (TI) at elevated end-expiratory lung volume (EEV) has been studied in pentobarbital-anesthetized, spontaneously breathing cats with intact vagi. Responses to increases in EEV, induced by imposition of an expiratory threshold load (ETL) of 10 cmH2O, were monitored before and after section of cervical dorsal roots C3-C7. The immediate (first-breath) effect of application of ETL was a prolongation of both TI and expiratory duration (TE). After 10 min of breathing against the ETL, average TI returned to control values but TE remained prolonged. Abolishing feedback from the diaphragm did not affect these responses. When steady-state responses to ETL were compared with those elicited by inhalation of 5-6% CO2 in O2, changes in EEV had, on average, no independent effect on respiratory drive (rate of rise of integrated phrenic activity), although phrenic activity increased greatly in some cats despite little or no change in arterial partial pressure of CO2. These data indicate that diaphragmatic receptors do not contribute to either the immediate (first-breath) or steady-state responses of phrenic motoneurons to increases in EEV in intact cats.

Bronchomotor responses to hypoxia and hypercapnia in decerebrate cats.

Decerebrate animals are often used in investigations of the control of breathing because anesthesia-induced depression of respiratory reflexes is absent. We therefore investigated the level of tone and responsiveness of airway smooth muscle in seven decerebrate, paralyzed, and ventilated cats. Specifically, we measured the changes in pulmonary resistance (RL) and dynamic pulmonary compliance (CLdyn) in response to hypoxia and hypercapnia. All cats responded to hypoxia (approximately 10% O2 in N2) with significant increases (mean 49%, range 5-156%) in RL from a mean control value of 0.0197 +/- 0.0081 (SD) cmH2O.ml-1.s. During inhalation of 5% CO2 in O2, RL increased significantly (mean 59%, range 16-135%) from a mean control value of 0.0190 +/- 0.0056 cmH2O.ml-1.s. Decreases in CLdyn during hypoxia and hypercapnia were much smaller, averaging -9 and -11%, respectively. After atropine was administered, average control RL fell 50%, from 0.0269 to 0.0134 cmH2O.ml-1.s (P < 0.05; n = 4). Hypoxic and hypercapnic gas mixtures did not affect pulmonary mechanics after atropine was administered. In three cats, oscillations of RL were synchronized to phrenic activity but only at low respiratory frequencies (approximately 12 cycles/min), indicating that airway smooth muscle responded slowly to vagal input. Pentobarbital sodium, like atropine, reduced control RL in three cats. These cats lost their bronchoconstrictor response to hypercapnia but had augmented responses to hypoxia compared with preanesthetic responses. We conclude that decerebrate cats possess resting bronchomotor tone and retain their responsiveness to hypoxia and hypercapnia. Thus the decerebrate cat is a useful model for studying the control of tracheobronchial smooth muscle.

Chest wall distortion and discharge of pulmonary slowly adapting receptors.

We assessed the effects of chest wall distortion, changes in lung volume, and abolition of airway smooth muscle tone on the discharge patterns of 92 pulmonary slowly adapting receptors (SAR) in decerebrate, spontaneously breathing cats. Distortion resulted from their inspiratory efforts against an occluded airway at functional residual capacity and at increased end-expiratory lung volumes. Approximately 40% of SAR increased discharge frequencies during occlusions. Modulation of SAR discharge during occlusions persisted after administration of atropine to eliminate airway smooth muscle tone. Phasic modulation of SAR discharge was eliminated during no-inflation tests after paralyzing the cats and ventilating them on a cycle-triggered pump. We conclude 1) parasympathetic modulation of airway smooth muscle tone makes no obvious contribution to SAR discharge in spontaneously breathing cats; 2) the no-inflation test (withholding of lung inflation during neural inspiration) in paralyzed and ventilated cats is a valid test for the presence of projections from SAR to medullary respiratory neurons; and 3) in the absence of tidal volume changes, distortion stimulates some SAR. Sensory feedback from receptors in the lung, not just those in the chest wall, may therefore provide information about abnormal chest wall configurations.

Discharge patterns of phrenic motoneurons during fictive coughing and vomiting in decerebrate cats.

In decerebrate, paralyzed, and ventilated cats, we recorded the activity of 100 spontaneously active phrenic motor axons during the increased phrenic discharges characteristic of fictive vomiting (FV) and coughing (FC). During control respiratory cycles, approximately one-half the neurons were recruited in the first decile of inspiration; recruitment continued throughout inspiration. During FV, the duration of phrenic discharge was halved; 20 of 26 motoneurons studied were recruited in the first decile of the burst. During FC, recruitment times did not change compared with control, although the duration of the phrenic burst doubled. Discharge frequencies increased and recruitment order of phrenic motoneurons was virtually unaffected during FC and FV. Limited recruitment of previously inactive neurons in the filaments from which we recorded was found during FV and FC. During FV, 1 previously inactive motoneuron was recruited in 16 filaments containing 25 spontaneously active motor axons. During FC, 3 new motoneurons were recruited in addition to the 64 already active in 35 filaments. Recruitment during FV and FC was absent even when recording from filaments known, on the basis of antidromic activation, to contain inactive motor axons. During FV, 10 of 26 motoneurons began their discharges with doublets (interspike interval < 10 ms); doublets occurred in only 4 of 67 motoneurons during FC. Already active phrenic motoneurons contributed to the intense phrenic activity associated with both respiratory (coughing) and nonrespiratory (vomiting) behavior by increases in discharge frequency, earlier recruitment, and doublets; the contribution of previously quiescent motoneurons remains uncertain.

Hypoxemia-induced modification of troponin I and T in canine diaphragm.

Impaired muscle function (fatigue) may result, in part, from modification of contractile proteins due to inadequate O(2) delivery. We hypothesized that severe hypoxemia would modify skeletal troponin I (TnI) and T (TnT), two regulatory contractile proteins, in respiratory muscles. Severe isocapnic hypoxemia (arterial partial pressure of O(2) of approximately 25 Torr) in six pentobarbital sodium-anesthetized spontaneously breathing dogs increased respiratory frequency and electromyographic activity of the diaphragm and internal and external obliques, with death occurring after 131-285 min. Western blot analysis revealed proteolysis of TnI and TnT, 17.5- and 28-kDa fragments, respectively, and higher molecular mass covalent complexes, one of which (42 kDa) contained TnI (or some fragment of it) and probably TnT in the costal and crural diaphragms but not the intercostal or abdominal muscles. These modifications of myofibrillar proteins may provide a molecular basis for contractile dysfunction, including respiratory failure, under conditions of limited O(2) delivery.

PCO2 affects tracheal tone during apnea in anesthetized dogs.

We hypothesized that CO2, like hypoxia and withdrawal of pulmonary slowly adapting receptor input, would cause tracheal constriction during neural apnea (absence of phrenic activity). In seven anesthetized paralyzed dogs ventilated to neural apnea, we increased arterial PCO2 (PaCO2) in steps by adding CO2 to the inspirate while keeping ventilation constant. Increases in PaCO2 caused tracheal constriction during neural apnea in all dogs; 69 +/- 26 (SD)% of the change in tracheal diameter occurred during neural apnea. Average sensitivity of tracheal diameter to CO2 was 0.44 mm/Torr PaCO2. Our data suggest that central chemoreceptor inputs to brain stem neurons controlling smooth muscle of the extrathoracic airway bypass central mechanisms generating inspiration.

Tracheal constrictor drive above the apneic threshold in anesthetized dogs.

We have previously shown that raising arterial PCO(2) (Pa(CO(2))) by small increments in dogs ventilated below the apneic threshold (AT) results in almost complete tracheal constriction before the return of phrenic activity (Dickstein JA, Greenberg A, Kruger J, Robicsek A, Silverman J, Sommer L, Sommer D, Volgyesi G, Iscoe S, and Fisher JA. J Appl Physiol 81: 1844-1849, 1996). We hypothesized that, if increasing chemical drive above the AT mediates increasing constrictor drive to tracheal smooth muscle, then pulmonary slowly adapting receptor input should elicit more tracheal dilation below the AT than above. In six methohexital sodium-anesthetized, paralyzed, and ventilated dogs, we measured changes in tracheal diameter in response to step increases in tidal volume (VT) or respiratory frequency (f) below and above the AT at constant Pa(CO(2)) ( approximately 40 and 67 Torr, respectively). Increases in VT (400-1,200 ml) caused significantly more (P = 0.005) tracheal dilation below than above AT (7.0 +/- 2.2 vs. 2.8 +/- 1.0 mm, respectively). In contrast, increases in f (14-22 breaths/min) caused similar (P = 0.93) tracheal dilations below and above (1.0 +/- 1.3 and 1.0 +/- 0.8 mm, respectively) AT. The greater effectiveness of dilator stimuli below compared with above the AT is consistent with the hypothesis that drive to tracheal smooth muscle increases even after attainment of maximal constriction. Our results emphasize the importance of controlling PCO(2) with respect to the AT when tracheal smooth muscle tone is experimentally altered.

Responses of inspiratory neurons of the dorsal respiratory group to stimulation of expiratory muscle and vagal afferents.

In decerebrate, paralyzed and ventilated cats, we monitored the intracellular responses of 30 inspiratory neurons of the dorsal respiratory group (DRG) to stimulation of vagal and expiratory muscle (internal intercostal and abdominal) afferents. We hypothesized that the inhibitory effects of stimulation of expiratory muscle afferents, previously reported, would block the excitatory responses of inspiratory neurons of the DRG to vagal stimulation. Although prolonged stimulus trains to expiratory muscle afferents caused respiratory phase-switching, single shocks or short trains elicited no responses in 17 bulbospinal neurons, excitatory responses in 6, and inhibitory responses in 2. Of the 4 propriobulbar neurons tested, 2 had inhibitory responses and 2 did not respond. In only 2 neurons, both bulbospinal, did conditioning stimuli to expiratory muscle afferents block or reduce the excitatory effects of vagal stimulation. These results suggest that interaction of vagal and expiratory muscle afferents, which might account for the absence of a change in inspiratory duration despite increased vagal afferent feedback at elevated end-expiratory lung volumes, does not occur within the DRG.

Effects of amino acids on the excitability of respiratory bulbospinal neurons in solitary and para-ambigual regions of medulla in cat.

The effects of inhibitory (gamma-aminobutyric acid (GABA) and glycine) and excitatory (L-glutamate and DL-homocysteate, DLH) amino acids on the excitability of respiratory bulbospinal neurons were studied in decerebrate, paralyzed, bilaterally vagotomized, artificially ventilated cats. Unit activities were recorded extracellularly in the medulla in both the ventrolateral portion of the nucleus tractus solitarius and the para-ambigual region in the vicinity of the nucleus ambiguus (dorsal and ventral respiratory groups, respectively). All neurons were bulbospinal since they could be antidromically activated by electrical stimuli to the spinal cord. We used variations in antidromic latency (ADL) as a measure of changes in excitability of the soma. All neurons exhibited variations in ADL related to the respiratory cycle, being shortest (minimum ADL) during neural activity and longest (maximum ADL) in the silent period. Neurons whose discharge frequencies fell during application of putative inhibitory amino acids showed an increase of minimum ADL compared to control, indicating hyperpolarization. Minimum ADL, in some cells, became shorter during application of excitatory amino acids, indicating depolarization; in others, mechanisms secondary to increased neuronal firing likely obscured their effects. The transient maximum ADL usually present at the onset of the silent period was increased by excitatory amino acids and, in some units, was reduced or eliminated by inhibitory amino acids. These effects are discussed in terms of a modulation by synaptic inputs and neurotransmitters of the cumulative afterhyperpolarization which follows bursts of action potentials.

Effects of ethanol on respiratory activity in the neonatal rat brainstem-spinal cord preparation.

Ethanol (1-12 mM) added to the superfusion medium of the isolated brainstem-spinal cords of newborn rats did not affect phrenic activity but significantly reduced hypoglossal activity by 54%, 67% and 55% at 3, 6 and 12 mM, respectively. Although the reasons for the suppression of hypoglossal activity remain unknown, this preparation may be a useful model for determining why cranial motoneurons are more vulnerable than phrenic motoneurons to various agents and, more generally, how ethanol impairs neural function.

Expiratory neurones of the rostral medulla: anatomical and functional correlates.

Intracellular recordings of the activities of 16 bulbar expiratory neurones of the rostral medulla were performed in decerebrate cats. Seven of these identified neurones were intracellularly injected with horse-radish peroxidase for morphological examination. We observed 3 categories of expiratory neurones including two pharyngeal motoneurones of the retrofacial nucleus, one with an augmenting, the other with a decrementing discharge pattern. Augmenting patterns were also observed in neurones ventromedial to the retrofacial nucleus and in another located 320 micron from the ventral surface of medulla. Their possible functions are discussed in relation to their anatomical location and morphology.

Hypoxia, not hypercapnia, induces cardiorespiratory failure in rats.

Mechanical respiratory loads induce cardiorespiratory failure, presumably by increasing O2 demand concurrently with decreases in O2 availability (decreased PaO2). We tested the hypothesis that asphyxia alone can cause cardiorespiratory failure ("failure") in pentobarbital-anesthetized rats. We also tested the hypothesis that hypoxia, not hypercapnia, is responsible by supplying supplemental O2 during mechanical loading in a separate group of rats. Asphyxia (mean PaO2 and PaCO2 of 43 and 69mmHg, respectively) resulted in failure, evident as a slowing of mean respiratory frequency (133-83breaths/min) and a sudden and large drop in mean arterial pressure (71-47mmHg), after 214±66min (n=16; range 117-355min). Neither respiratory drive nor heart rate decreased, indicating that failure was peripheral, not central. Of 8 rats tested after 3h of asphyxia for the presence in blood of cardiac troponin T, all were positive. In an additional 6 rats, normocapnic hypoxia (mean PaCO2 and PaO2 were 39±2.2 and 41±3.1mmHg, respectively) caused failure after an average 205min (range 181-275min), no different from that of asphyxic rats. In the 6 rats that breathed O2 during an initially moderate inspiratory resistive load, endurances exceeded 7h (failure occurring only because we increased the load after 6h) and tracheal pressure and left ventricular dP/dt were maintained despite supercarbia (PaCO2>150mmHg). Thus, asphyxia alone can induce failure, the failure is due to hypoxia, not hypercapnia, and hypercapnia has minimal effects on cardiac and respiratory muscle function in the presence of hyperoxia.

Changes in serum fast and slow skeletal troponin I concentration following maximal eccentric contractions.

OBJECTIVES: We tested the hypothesis that fast skeletal muscle troponin I (fsTnI) concentration in serum would increase more than those of slow skeletal muscle troponin I (ssTnI) after eccentric exercise of the elbow flexors using a sensitive blood marker to track fibre specific muscle damage. DESIGN: Observational comparison of response in a single experimental group. METHODS: Eight young men (26.4±6.2 years) performed 210 (35 sets of 6) eccentric contractions of the elbow flexors on an isokinetic dynamometer with one arm. Changes in serum fsTnI and ssTnI concentrations, serum creatine kinase (CK) activity, and maximal voluntary isometric contraction torque (MVIC) before and 1, 2, 3, 4 and 14 days following exercise were analysed by a Student-Newman-Keuls multiple comparison test. The relationship between serum CK activity and fsTnI or ssTnI concentrations was determined using a Pearson's product moment correlation. RESULTS: Significant (P<0.05) decreases in MVIC and increases in serum CK activity and fsTnI were evident after exercise, but ssTnI did not change. The time course of changes in fsTnI was similar to that of CK, peaking at 4 days post-exercise, and the two were highly correlated (r=0.8). CONCLUSIONS: Increases in serum fsTnI concentrations reflect muscle damage, and it seems likely that only fast twitch fibres were damaged by eccentric contractions.

Repeated inspiratory occlusions in anesthetized rats acutely increase blood coagulability as assessed by thromboelastography.

Many of the components contributing to coagulability are enhanced by repeated episodes of hypoxia, as occurs in obstructive sleep apnea, but no one has yet measured the global hemostatic properties of blood in an animal model of this disease. Using thromboelastography, a hemostatic assay, we measured hemostasis in six pentobarbital-anesthetized rats before and after 3h of repeated inspiratory occlusions lasting 30s applied every 2 min and compared the results to those in six identically prepared rats before and after 3h of resting breathing. Rats subjected to occlusions displayed faster onset of clotting (p<0.031) and more rapid coagulation (p<0.031). Thus, repeated inspiratory occlusions acutely cause hypercoagulability in rats. Thromboelastography, a simple test of hemostasis, may help evaluate the factors responsible for this increase and, in patients with obstructive sleep apnea, the risk of future cardiovascular disease.

Precise control of end-tidal carbon dioxide levels using sequential rebreathing circuits.

Anaesthesiologists have traditionally been consulted to help design breathing circuits to attain and maintain target end-tidal carbon dioxide (P(ET)CO2). The methodology has recently been simplified by breathing circuits that sequentially deliver fresh gas (not containing carbon dioxide (CO2)) and reserve gas (containing CO2). Our aim was to determine the roles of fresh gas flow, reserve gas PCO2 and minute ventilation in the determination of P(ET)CO2. We first used a computer model of a non-rebreathing sequential breathing circuit to determine these relationships. We then tested our model by monitoring P(ET)CO2 in human volunteers who increased their minute ventilation from resting to five times resting levels. The optimal settings to maintain P(ET)CO2 independently of minute ventilation are 1) fresh gas flow equal to minute ventilation minus anatomical deadspace ventilation, and 2) reserve gas PCO2 equal to alveolar PCO2. We provide an equation to assist in identifying gas settings to attain a target PCO2. The ability to precisely attain and maintain a target PCO2 (isocapnia) using a sequential gas delivery circuit has multiple therapeutic and scientific applications.

Respiratory and stepping frequencies in conscious exercising cats.

The incidence of entrainment between respiratory and stepping frequencies has been investigated in exercising cats. Stainless steel wire electrodes were surgically implanted in the right hemidiaphragm, quadriceps, and hamstring muscles of six cats. Electromyograms were recorded while the cats were at rest or walking on a treadmill at speeds between 0.31 and 1.67 m/s (6 km/h). Autocorrelation of diaphragmatic activity showed cats to have two respiratory patterns: regular (maintained periodicity in the autocorrelogram) or irregular (flat autocorrelogram). Autocorrelation of quadriceps or hamstring activity revealed stepping frequency. Cross correlation of diaphragmatic vs. quadriceps or hamstring activities revealed the presence of entrainment between respiration and stepping. In four cats, entrainment was either very weak or absent, but the other two cats clearly showed entrainment of the two activities, even in the absence of a regular respiratory rhythm. In conclusion, at speeds up to 1.67 m/s, respiratory frequency is not tightly locked to stepping frequency in cats; however, the occurrence of entrainment in some instances clearly indicates the existence of neural circuitry linking these two pattern generators.