Pediatric Pulrnonology 14:233-238 (1992)
Response to External Obstruction in Preterm Infants With Apnea C.J. Upton, YRCP, A.D. Milner, YD,
FRCP,
and G.M. Stokes, MPhll
Summary. A proportion of preterm infants respond to an external airway obstruction by becoming apneic. We have studied 23 infants (median birthweight, 1.14 kg; gestation, 29 weeks) on 80 occasions, to determine the time course of the response and its relationship with spontaneous apnea occurrence. Upper airway flow was measured with a face mask and pneumotachograph, and a tap was turned intermittently to produce an occlusion. A total of 380 occlusions were analyzed. The infants became apneic during the obstruction on 72 occasions (19%), and after the obstruction on 122 occasions (32%). Both of these events were significantly more common than immediately prior lo t h e obstruction, when apnea occurred on 29 occasions (8%). Of the apneas following occlusion 57% were central in type. The point in the respiratory cycle at which obstruction occurred had no effect on the production of apnea. An obstruction score was calculated for each study. This was the mean of the number of apneas during and after each obstruction, expressed as a percentage of the number of obstructions per study. This score was positively correlated with the number of spontaneous apneas recorded. Obstruction score rose from a mean of 20% during days 0-1 4, to 34% during days 15-28, and thereafter it declined. This pattern may be relevant to the time course of apnea in susceptible D 1992 wey-Liss, Inc infants. Pediatr Pulmonol. 1992; 14:233-238. Key words: Obstruction score; time course of apnea; inductive plethysmography;pneumotachography.
INTRODUCTION
was measured at the mouth and nose during tidal breathing, an artefact, produced by the transmission of the cardiac impulse up the patent airway, distorted the airflow trace. This was particularly obvious during apneic episodes. Absence of this cardiac artefact was noted during some central apneas, as well as mixed and obstructive episodes. All apneic episodes with a cardiac artefact present on the flow trace began at end-expiration and if the artefact persisted until the end of the episode, the first respiratory effort always produced an inspiratory tidal exchange. In the absence of the artefact, apnea could begin at other points in the respiratory cycle and in over
The behavioral and ventilatory responses to external airway obstruction vary widely in infancy. Swift and Emery found that 3% of term infants fail to achieve oral breathing within 2 5 seconds of nasal obstruction, and postulated that obstruction may lead to apnea and possibly sudden death in some premature and susceptible infants. That such an obstruction occasionally results in apnea in preterm infants, has been demonstrated by our group when measuring lung volumes in a total body plethysmograph.’ In a group of ten preterm infants studied on 15 occasions, closure of a shutter connected to a face mask produced brief apnea, but this response was never seen in term infants. Stark and Thach also demon- From the Department of Neonatal Medicine, City Hospital, Nottingstrated that neck flexion may produce airway obstruction ham, United Kingdom. and that in a 29 week premature infant this resulted in profound apnea and bradycardia. They subsequently Received February 13, 1992; (revision) accepted for publication July demonstrated that this may occur spontaneously in pre- 8, 1992. term infants and postulated that such a response may be T h i s study was supported by a grant from the Medical Research Counimportant in the pathophysiology of mixed a ~ n e a as , ~ cil of Great Britain. obstructed breaths preceded the central element in 14 of 36 mixed apneas. Mixed apnea is undoubtedly common Presented at the Paediatric Research Society Meeting, Edinburgh. in preterm infants5 and is important because it becomes March, 1990. more prevalent as apnea duration lengthens .6*7 Address correspondence and reprint requests to Dr. C.J. Upton, DeWe have previously shown that airway closure may partment of Neonatal Medicine, City Hospital, Nottingham, NG5 occur during apparently central apnea. When airflow 1PB. U K .
’
0 1992 Wiley-Liss, Inc.
234
Upton et al.
half of these the first breathing effort did not result in a was defined as cessation of breathing effort for 10 secresumption of flow. Therefore, it appeared that the pres- onds or less if associated with bradycardia of 90 beats/ ence of the cardiac artefact was a reliable mcasurc of minute or less. lnfants were then studied in their own cot or incubator, upper airway patency. It can provide a means of gauging the degree of patency of the airway in the absence of lying in the right lateral position, while asleep, within 1 frank respiratory effort against a closed airway. By ana- hour of a feed. Sedation was not given. A Bennett face lyzing its use in a much larger number of apneas than mask was used in measuring upper airway flow, as inpreviously, we have recently been able to confirm its fants are not obligatory nose breathers. I 2The seal with the face was helped by the application of petroleum jelly. usefulness as an indicator of upper airway patency.' The aim of this study was to analyze the apneic re- In order to eliminate dead space, a bias flow of air of 3 sponse of preterm infants to external obstruction in more L/min entered one port of the mask. The other port was detail. By looking at the cardiac artefact during the apneic conncctcd to a Fleisch 0 pneumotachograph to record the episodes following occlusions we could analyze the ef- baby's respiratory flow. The pressure gradient along the fects of external obstructions on intrinsic airway patency. pneumotachograph was measured by a Validyne MP4S We also wanted to examine the time course of the re- differential pressure transducer (range, ? 0 . 9 cmH,O) sponse, studied serially in a group of preterm infants, and and a transducer amplifier (P.K. Morgan no. 076). A to see how it related to spontaneous apnea frequency. two-way tap was situated close to the mask, interposed Another aim was to analyze how obstructions at different between the mask and both the bias flow of air and the points in the respiratory cycle affected the response. The pneumotachograph. By turning the tap, the mask inlet Hering-Breuer inflation reflex is known to cause apnea in and outlet were blocked, to provide an effective external the newborn" and it has been postulated that it may be obstruction. A blow-off valve set at 10 cmH,O pressure important in the apneic response to external o b ~ t r u c t i o n . ~was included in the bias air flow line, in order to prevent If this were so, obstruction at any point, other than end- the pressure building up in the rest of the system during expiration would tend to maintain lung volume above tap closure. Infants were studied for approximately half an hour, functional residual capacity, and would be expected to be during which time air flow and the other variables meamore potent at producing apnea. sured during the polygraphic study were continuously recorded. Spontaneous apneas were recorded during the MATERIALS AND METHODS first 10-15 minutes, and then intermittent tap closure was Preterm infants born at 32 weeks gestation or less were performed in order to study the response to obstruction. included in the study and they were not preselected for Our aim was to achieve a short obstruction of between 5 the presence of apnea. The 23 babies, 13 boys and 10 and 6 seconds on five or more occasions. Apneic epigirls, studied on a total of 80 occasions, were clinically sodes and external obstructions were then played on to a stable in air at the time of study. Their median birth Gould chart recorder for manual analysis. weight was 1.14 kg (range, 0.71-1.70 kg) and their gesThe number of spontaneous apneas of 5 seconds duratational age, 29 weeks (range, 25-32 weeks). Median tion or longer was recorded. Additionally, the number of day of study was day 16 (range, 2-55 days) at a median apneas of 2 seconds or longer was recorded during each post-conceptional age of 32 weeks (range, 26-36 weeks). obstruction, as well as any apnea starting during the 10 Of the studies, 38 were performed while the infant was seconds following the obstruction. In order to rule out the receiving methylxanthine treatment for apnea. Prior to possibility of chance association between external obentering the study, six infants had received mechanical struction and apnea, the numbers of episodes of 2 seconds ventilation for a period of up to 10 days, but none had or longer, which occurred or terminated during the 10 respiratory symptoms at the time of the study. Two in- seconds prior to the obstruction, were also recorded. All apneas were classified into central, mixed, or obfants had ultrasound evidence of intraventricular hemorrhage, one with parenchymal changes. Neither of these structive episodes by comparing the two channels of respiratory effort with the airflow trace. Apnea was regarded infants were symptomatic. Polygraphic recordings were made prior to the main as central if there was no respiratory effort on the impedpart of the study, in order to assess apnea frequency at ance and inductance traces. It was regarded as mixed if that time. Thoracic impedance, abdominal respiratory there was at least one obstructed breath and a central inductive plethysmography, electrocardiogram (ECG). element of at least 3 seconds. If respiratory efforts continand oxygen saturation (Sa,?) were recorded on to tape ued, and/or the pause between the efforts was less than 3 (Racal, Store 4) and subsequently analyzed manually. seconds, obstructive apnea was diagnosed. By analysis of whether the cardiac impulse was transFull details have been given elsewhere. I ' Median duration of these recordings was 3.96 hours (range, 2.19 to mitted up the airway, we were also able to detect whether 5.29). For the purposes o f this part of the study, apnea the upper airway was patent during apnea. For spontane-
External Obstruction in Preterms
ous apneas, and those following obstruction, the presence or absence of this cardiac artefact was recorded. However, it was not actually possible to assess patency during the external obstructions. Tap closure itself interrupted the passage of air to the pneumotachograph and caused loss of signal on the aifflow trace. For each external obstruction, the point in the respiratory cycle at which the obstruction was induced was recorded from analysis of the flow trace. This was classified as either mid-inspiratory, end-inspiratory,mid-expiratory, or end-expiratory. In order to assess the variability of the response with age and apnea frequency, an obstruction score was calculated for each study. This was the mean of the number of apneas during and after the obstruction, per study, expressed as a percentage of the number of obstructions in that study. Results were statistically analyzed using the MannWhitney U test, x2 test, and simple regression analysis. Approval for the study was given by the Nottingham Ethics Committee and informed parental consent was obtained.
235
ECG
.
I . I l .
FLOW
1 1 1
I
RIP
-02
10 s a
Fig. 1. An external obstruction, demonstrated by the loss of signal on the airflow trace (FLOW). There is no change in the infant’s respiratory pattern on the respiratory inductive plethysmography trace (RIP).
RESULTS ECG A total of 398 obstructions are recorded. A mean of five obstructions (range, 3 to 7) per study was produced, with a mean duration of 5.5 seconds (range, 4-9.5s). Arousal, with gross body movements, occurred in association with 18 obstructions. These were impossible to interpret and so were omitted from further analysis, leaving 380 obstructions. Apnea was not seen (Fig. 1) in 230 of the 380 obstructions, or 60% of the total. On 72 occasions (19%) apnea RIP occurred during the obstruction (Fig. 2). This usually happened as soon as the tap was turned, but occasionally one or two obstructed breaths were taken first. More common, though, was apnea following relief of the obstruction, which happened with 122 obstructions (32%), SaO, as seen in Figure 3. On 44 occasions (1 1%) apnea oc10 sec curred both during and after the obstruction. Commonly this was a continuous episode, but sometimes a few Fig. 2. An external obstruction, during which the infant ceases respiratory effort on respiratory inductive plethysmography breaths were taken on relief of the obstruction, with ap- all (RIP). Respiratory effort resumes promptly on relief of the obnea occurring again soon afterwards. struction. These events happened more frequently than would be expected by pure chance. Short pauses during the 10 seconds prior to obstruction occurred on 29 occasions closure, as demonstrated by the absence of cardiac arte(8%). This was statistically less frequent than both the fact, was present in 91 episodes (75%). Therefore, folnumber during and after the obstruction (P = 0.0001 for lowing an external obstruction, airway closure was seen both; x2 test). The apneas before obstruction were also to occur in roughly half of apparently central apneas. The point in the respiratory cycle at which obstruction shorter, with a mean duration of 3.3 s (SD 2 I .6s), compared to 6.6 s (SD 2 5.3 s) afterwards (P C 0.0001, was induced had no significant effect on the production of apnea, either during or after the obstruction. Table 1 Mann-Whitney U test). Of the 122 apneas following obstruction, 69 (57%) shows the relationship between respiratory phase at obwere central, 33 (27%) mixed, and 20 (16%) obstructive, struction and the presence or absence of apnea following as determined by conventional criteria. However, airway it. The results for apnea during the obstruction revealed a
236
Upton et al.
ence. Mean obstruction score was 16%in 41 studies with
ECG
analysis, between the number of spontaneous apieas and the obstruction score (r = 0.53, P = 0.0001). When obstruction score is compared with respect to age, then it seems to vary with postnatal, rather than postconceptional age. The obstruction score is significantly lower in the first 2 weeks of life than in the third and fourth weeks. RIP The mean score in the 36 studies during the first 2 weeks 1 was 20%, compared to 34%.in 23 studies during weeks 3 and 4 (P < 0.03,Mann-Whitney U test). Although there was a tendency for the score to later fall, being a mean 26% at weeks 5 and later, this did not reach statistical significance. -0. lowTreatment with methylxanthines did not appear to affect the obstruction score. The mean was 22% Fig. 3. An external obstruction followed by the infant develop ing central apnea. The airway remains patent, as demonstrated while receiving, and 28% while not receiving theophylby the presence of the cardiac artefact on the airflow trace. line (NS).
1
-
scan-
TABLE 1-Presence or Absence of Apnea Following Obstruction, Divided With Respect to Respiratory Phase at Obstruction"
DISCUSSION
This study has confirmed that apnea is a common event in association with external airway obstruction. Our preRespiratory phase vious study2 has been criticized, as the apnea seen with occlusion may have been a chance association. I 3 The Mid-inspiratory 34 20 temporal relationship seen previously between obstruc24 12 End-inspiratory Mid-expiratory 76 25 tion and apnea has always made this unlikely. The current End-expiratory 1 24 65 study confirms that respiratory pauses during and after occlusion are significantly more common than those beTotal 258 122 forehand. Additionally, those following the external ob'Differences are. not statistically significant (x2test). struction were longer than those preceding it. The mechanism by which the apneic response to obstruction occurs is not clear. As previously discussed, if similar, non-significant distribution. The length of the the Hering-Breuer inflation reflex is important, one obstruction also made no significant difference as to would expect occlusion at points in the respiratory cycle other than end-expiration to be more potent at producing whether apnea was induced. The mean obstruction score per study was 25% (range, apnea. No such variation with the respiratory cycle was 0-90%). The median number of spontaneous apneas re- seen and we, therefore, feel the Hering-Breuer inflation corded during the airflow study was 2 (range, 0-19). As reflex is not a vital part of this response. Trigeminal reflexes, which may be affected by applithe total time of study varied slightly and the time during which the mask was actually in position was affected by cation of a face mask, clearly play a role in the control of any intermittent episodes of arousal, the apnea frequency respiration.14 Application of a face mask in preterm inwas additionally calculated from the polygraphic record- fants can result in hypoventilation as measured by endings performed prior to the airflow study. Median apnea tidal CO,, but frank apneic episodes were not recorded.Is frequency was 2 significant episodes per hour (range, Indeed, in more mature infants application of a face mask may increase minute ventilation. l 6 In our study the mask 0-17). If the obstruction score is compared to apnea fre- had been applied for many minutes prior to the obstrucquency, there is no significant difference between the 39 tion; therefore any acute effects of mask application were studies with 2 unlikely to influence the response. Theoretically, the inepisodedhr (mean, 23 vs. 27%; NS). If, however, studies terruption of airflow to the trigeminal area with obstrucare divided with respect to the number of spontaneous 5 tion, and its subsequent reintroduction, may have stimusecond apneas during the airflow study, there is a differ- lated reflex changes. That these should result in apnea, Number without apnea after obstruction
Number with apnea after obstruction
External Obstruction in Preterms
even in susceptible infants, seems unlikely with our present knowledge of such reflexes.14 The intercostal-phrenic inhibitory reflex would appear to be a more likely mechanism for the production of apnea. Rib cage distortion, produced by both external pressure on the chest wall’7 and respiratory loading,’* has been shown to inhibit inspiratory effort in infants. Such a reflex has been used to explain the inspiratory shortening in response to occlusion seen in small preterm infants, as opposed to more mature infants, who exhibit inspiratory prolongation. ‘9*20 In susceptible infants such a reflex may inhibit inspiration sufficiently to produce apnea. Such a hypothesis is attractive, as it would help to explain why premature infants with more compliant chest walls show this response, whereas term infants do not. The fact that airway closure was seen in 75% of apneas following external obstruction would suggest that upper airway patency is affected by the occlusion. Airway closure itself may result in an inhibition of respiratory efforts’ and may therefore also be important in the production of apnea following airway occlusion. That such an external obstruction may result in intrinsic airway closure is highly likely on the basis of current evidence. Extrathoracic airway patency is maintained actively in life mainly by the actions of the genioglossus and geniohyoid.” Genioglossus activity is potentiated by negative pressure within the upper airway, such as occurs during inspiration, but decreases in response to positive pressure, which may be produced by external obstruction.22 Therefore, an external obstruction may produce a fall in genioglossus activity, resulting in intrinsic airway closure and subsequent apnea. However, whether an occlusion produces a negative or positive pressure in the pharynx depends on the point in the respiratory cycle at which the occlusion occurs. Occlusion at end-expiration would create a negative pressure, which should potentiate genioglossus activity. In practice this has been demonstrated by Carlo et al. ,23 who measured submental electromyograms as a reflection of genioglossus activity in ten preterm infants whose airways were occluded at end-expiration. Therefore, rather like the Hering-Breuer reflex, one would expect that occlusions at other points in the respiratory cycle would be more potent at producing airway closure and subsequent apnea. This was not seen to occur in this study, making it more doubtful that these factors are important in the apneic response to external obstruction. However, care needs to be taken in the interpretation of this data. It has been shown that, not surprisingly, the response of genioglossus differs in spontaneous mixed apneas and in external obstruction^.^^ Another important factor may be the activity of airway receptors, sensitive to flow or pressure. In animal experiments by Abu-Osba et a1.” flow up and down the upper airway resulted in the maintenance of pharyngeal patency
237
by increased activity of the genioglossus. This was abolished by topical anesthesia of the upper airway, suggesting that there is feedback from airway receptors to maintain genioglossus activity. Therefore, when flow ceases as a consequence of an obstruction, genioglossus activity drops and airway collapse may occur. Negative pressure produced by an external occlusion during inspiration may also have an effect on pressure sensitive receptors in the upper airway, resulting in an inhibition of thoracic inspiratory muscles.26 Clearly this may be another factor predisposing to apnea as a consequence of external obstruction, but once again the effect would be expected to be detected more often in occlusions during inspiration, which we have not seen. In spite of all these factors, Cohen and HendersonSmart found that the airways of infants were remarkably stable in response to occlusion. l 3 However, using their method of comparing esophageal and mask pressures during an occlusion, they could only detect obstructed breaths and they were not able to determine whether airway closure had occurred in the absence of respiratory effort against a frank obstruction. We have examined closure per se, by analysis of the transmission of the cardiac artefact up the airway. The relevance of the apneic response to occlusion to the pathophysiology of apnea is unclear. Although the obstruction score in our study correlated well with the number of spontaneous apneas, it did not relate to apnea frequency measured during the initial polygraphic recordings. The response to obstruction occurred more frequently than by pure chance, and therefore we do not believe that the relationship between the obstruction score and number of apneas detected by airflow is coincidental. Rather, it seems that the apneic response to occlusion may be important in the prolongation of initially short respiratory pauses, during which airway closure may occur. Obstructed breaths preceded the central element in 38% of mixed apneas in one study4 and we found a similar proportion in the spontaneous mixed apneas recorded here.9 It seems likely, then, that the response is relevant to the production of mixed apnea. The fact that it did not correlate with apnea frequency measured by polygraphic recordings may have been because we did not measure airflow during this time, and a number of mixed apneas may have been missed. Alternatively, the response may change rapidly, for example with sleep state,” and only reflect the apnea frequency over a short period of time, as measured during our airflow study. The apnea frequency measured by pneumograms is also notoriously variable over a short space of time” and may not accurately reflect the frequency even only 4 hours later. The time course of the response is also interesting. In our original study, older preterm infants seemed more likely to respond to occlusion with apnea2 and we have confirmed that the response is not prominent in the first
238
Upton et al.
14. Haddad GG, Mellins RB. The role of airway receptors in the control of respiration in infants: A review. J Pediatr. 1977; 911281-286. 15. Chernick V, Avery ME. Response of premature infants with periodic breathing to ventilatory stimuli. J Appl Physiol. 1966; 21:43W0. 16. Dolfin T, Duffty P, Wilkes D, England S, Bryan H. Effects of a face mask and pneumotachograph on breathing in sleeping infants. Am Rev Respir Dis. 1983; 128:977-979. REFERENCES 17. Knill R, Bryan AC. An intercostal-phrenic inhibitory reflex in human ncwbom infants. J Appl Physiol. 1976; 401352-356. I . Swift PGF, Emery JL. Clinical observations on response to nasal 18. Knill R, Andrews W, Bryan AC, Bryan MH. Respiratory load occlusion in infancy. Arch Dis Child. 1973; 48:947-951. compensation in infants. J Appl Physiol. 1976; 40357-361. 2. Milner AD, Saunders RA, Hopkin 1E. Apnoea induced by airflow 19. Thach BT, Frantz ID, Adler SM, Taeusch HW. Maturation of obstruction. Arch Dis Child. 1977; 52:379-382. reflexes influencing inspiratory duration in human infants. J Appl 3. Stark AR, Thach BT. Mechanisms of airway obstruction leading Physiol. 1978; 45:203-21 I. to apnea in newborn infants. J Pediatr. 1976; 89:982-985. 4. Thach BT, Stark AR. Spontaneous neck flexion and airway ob- 20. Gerhardt T, Bancalari E. Apnea of prematurity: 11. Respiratory reflexes. Pediatrics. 1984; 74:63-66. struction during apneic spells in pretem infants. J Pediatr. 1979; 21. Brouillette RT, Thach BT. A neurornuscular mechanism main94:275-281. taining extrathoracic airway patency. J Appl Physiol. 1979: 5 . Dransfield DA, Spitzer AR. Fox WW. Episodic airway obstruc461772-779. tion in premature infants. Am J Dis Child. 1983; 137:44143. 6. Butcher-Puech MC, Henderson-Smart DJ, Holley D, Lacey JL, 22. Mathew OP, Abu-Osba YK, Thach BT. Influence of upper airway pressure changes on genioglossus muscle respiratory activity. J Edwards DA. Relation between apnoea duration and type and Appl Physiol. 1982; 5 2 : 4 3 8 4 4 . neurological status of pretenn infants. Arch Dis Child. 1985; 23. Carlo WA, Miller MJ, Martin RJ. Differential response of respira60:953-958. tory muscles to airway occlusion in infants. J Appl Physiol. 1985; 7. Muttitt SC, Finer NN, Tierney AJ, Rossmann J. Neonatal apnea: 59:847-852. Diagnosis by nurse versus computer. Pediatrics. 1988; 82:71324. Gauda EB, Miller MJ, Carlo WA, DiFiore JM, Martin RJ. Genio720. glossus and diaphragm activity during obstructive apnea and air8. Milner AD, Boon AW, Saunders RA, Hopkin IE. Upper airways way occlusion in infants. Pediatr Res. 1989; 26583-587. obstruction and apnoea in preterm babies. Arch Dis Child. 1980; 25. Abu-Osba YK, Mathew OP, Thach BT. An animal model for 55:22-25. airway sensory deprivation producing obstructive apnea with post9. Upton CJ, Milner AD, Stokes GM. Upper airway patency during mortem findings of sudden infant death. Pediatrics. 198 I;h8:79& apnoeaof prematurity. Arch Dis Child. 1992; 67:419-424. 801. 10. Cross KW, Klaus M, Tooley WH, Weisser K. The response of the 26. Thach BT, Schefft GL, Pickens DL, Menon AP. Influence of newborn baby to inflation of the lungs. J Physiol. 1960; 151:551upper airway negative pressure reflex on response to airway occlu565. sion in sleeping infants. J Appl Physiol. 1989; 67:745)-755. 1 1 . Upton CJ, Milner AD, Stokes GM. Combined impedance and inductance for the detection of apnoea of prematurity. Early Hum 27. Frantz ID, Adler SM, Abroms IF, Thach BT. Respiratory response to airway occlusion in infants: Sleep state and maturation. J Dev. 1990; 2455-63. Appl Physiol. 1976;41:634-638. 12. Miller MJ, Martin RJ, Carlo WA, Fouke JM, Strohl KP, Fanaroff AA. Oral breathing in newborn infants. J Pediatr. 1985; 107:465- 28. Hunt CE, Brouillette RT, Liu K , Klemka L. Day-to-day pneumogram variability. Pediatr Res. 1985; 19:174-177. 469. 13. Cohen G, Henderson-Smart DJ. Upper airway stability and apnea 29. Miller HC, Behrle FC, Smull NW. Severe apnea and irregular respiratory rhythms among premature infants. Pediatrics. 1959; during nasal occlusion in newborn infants. J Appl Physiol. 1986; 23:676-685. 6011511-1517.
week or two of life. Although the existence of a “honeymoon period” as such is unlikely, such a trend may reflect the time course of apneic episodes, as apnea tends to become more problematical at the end of the first week of life. 2y
Response to External Obstruction in Preterm Infants With Apnea C.J. Upton, YRCP, A.D. Milner, YD,
FRCP,
and G.M. Stokes, MPhll
Summary. A proportion of preterm infants respond to an external airway obstruction by becoming apneic. We have studied 23 infants (median birthweight, 1.14 kg; gestation, 29 weeks) on 80 occasions, to determine the time course of the response and its relationship with spontaneous apnea occurrence. Upper airway flow was measured with a face mask and pneumotachograph, and a tap was turned intermittently to produce an occlusion. A total of 380 occlusions were analyzed. The infants became apneic during the obstruction on 72 occasions (19%), and after the obstruction on 122 occasions (32%). Both of these events were significantly more common than immediately prior lo t h e obstruction, when apnea occurred on 29 occasions (8%). Of the apneas following occlusion 57% were central in type. The point in the respiratory cycle at which obstruction occurred had no effect on the production of apnea. An obstruction score was calculated for each study. This was the mean of the number of apneas during and after each obstruction, expressed as a percentage of the number of obstructions per study. This score was positively correlated with the number of spontaneous apneas recorded. Obstruction score rose from a mean of 20% during days 0-1 4, to 34% during days 15-28, and thereafter it declined. This pattern may be relevant to the time course of apnea in susceptible D 1992 wey-Liss, Inc infants. Pediatr Pulmonol. 1992; 14:233-238. Key words: Obstruction score; time course of apnea; inductive plethysmography;pneumotachography.
INTRODUCTION
was measured at the mouth and nose during tidal breathing, an artefact, produced by the transmission of the cardiac impulse up the patent airway, distorted the airflow trace. This was particularly obvious during apneic episodes. Absence of this cardiac artefact was noted during some central apneas, as well as mixed and obstructive episodes. All apneic episodes with a cardiac artefact present on the flow trace began at end-expiration and if the artefact persisted until the end of the episode, the first respiratory effort always produced an inspiratory tidal exchange. In the absence of the artefact, apnea could begin at other points in the respiratory cycle and in over
The behavioral and ventilatory responses to external airway obstruction vary widely in infancy. Swift and Emery found that 3% of term infants fail to achieve oral breathing within 2 5 seconds of nasal obstruction, and postulated that obstruction may lead to apnea and possibly sudden death in some premature and susceptible infants. That such an obstruction occasionally results in apnea in preterm infants, has been demonstrated by our group when measuring lung volumes in a total body plethysmograph.’ In a group of ten preterm infants studied on 15 occasions, closure of a shutter connected to a face mask produced brief apnea, but this response was never seen in term infants. Stark and Thach also demon- From the Department of Neonatal Medicine, City Hospital, Nottingstrated that neck flexion may produce airway obstruction ham, United Kingdom. and that in a 29 week premature infant this resulted in profound apnea and bradycardia. They subsequently Received February 13, 1992; (revision) accepted for publication July demonstrated that this may occur spontaneously in pre- 8, 1992. term infants and postulated that such a response may be T h i s study was supported by a grant from the Medical Research Counimportant in the pathophysiology of mixed a ~ n e a as , ~ cil of Great Britain. obstructed breaths preceded the central element in 14 of 36 mixed apneas. Mixed apnea is undoubtedly common Presented at the Paediatric Research Society Meeting, Edinburgh. in preterm infants5 and is important because it becomes March, 1990. more prevalent as apnea duration lengthens .6*7 Address correspondence and reprint requests to Dr. C.J. Upton, DeWe have previously shown that airway closure may partment of Neonatal Medicine, City Hospital, Nottingham, NG5 occur during apparently central apnea. When airflow 1PB. U K .
’
0 1992 Wiley-Liss, Inc.
234
Upton et al.
half of these the first breathing effort did not result in a was defined as cessation of breathing effort for 10 secresumption of flow. Therefore, it appeared that the pres- onds or less if associated with bradycardia of 90 beats/ ence of the cardiac artefact was a reliable mcasurc of minute or less. lnfants were then studied in their own cot or incubator, upper airway patency. It can provide a means of gauging the degree of patency of the airway in the absence of lying in the right lateral position, while asleep, within 1 frank respiratory effort against a closed airway. By ana- hour of a feed. Sedation was not given. A Bennett face lyzing its use in a much larger number of apneas than mask was used in measuring upper airway flow, as inpreviously, we have recently been able to confirm its fants are not obligatory nose breathers. I 2The seal with the face was helped by the application of petroleum jelly. usefulness as an indicator of upper airway patency.' The aim of this study was to analyze the apneic re- In order to eliminate dead space, a bias flow of air of 3 sponse of preterm infants to external obstruction in more L/min entered one port of the mask. The other port was detail. By looking at the cardiac artefact during the apneic conncctcd to a Fleisch 0 pneumotachograph to record the episodes following occlusions we could analyze the ef- baby's respiratory flow. The pressure gradient along the fects of external obstructions on intrinsic airway patency. pneumotachograph was measured by a Validyne MP4S We also wanted to examine the time course of the re- differential pressure transducer (range, ? 0 . 9 cmH,O) sponse, studied serially in a group of preterm infants, and and a transducer amplifier (P.K. Morgan no. 076). A to see how it related to spontaneous apnea frequency. two-way tap was situated close to the mask, interposed Another aim was to analyze how obstructions at different between the mask and both the bias flow of air and the points in the respiratory cycle affected the response. The pneumotachograph. By turning the tap, the mask inlet Hering-Breuer inflation reflex is known to cause apnea in and outlet were blocked, to provide an effective external the newborn" and it has been postulated that it may be obstruction. A blow-off valve set at 10 cmH,O pressure important in the apneic response to external o b ~ t r u c t i o n . ~was included in the bias air flow line, in order to prevent If this were so, obstruction at any point, other than end- the pressure building up in the rest of the system during expiration would tend to maintain lung volume above tap closure. Infants were studied for approximately half an hour, functional residual capacity, and would be expected to be during which time air flow and the other variables meamore potent at producing apnea. sured during the polygraphic study were continuously recorded. Spontaneous apneas were recorded during the MATERIALS AND METHODS first 10-15 minutes, and then intermittent tap closure was Preterm infants born at 32 weeks gestation or less were performed in order to study the response to obstruction. included in the study and they were not preselected for Our aim was to achieve a short obstruction of between 5 the presence of apnea. The 23 babies, 13 boys and 10 and 6 seconds on five or more occasions. Apneic epigirls, studied on a total of 80 occasions, were clinically sodes and external obstructions were then played on to a stable in air at the time of study. Their median birth Gould chart recorder for manual analysis. weight was 1.14 kg (range, 0.71-1.70 kg) and their gesThe number of spontaneous apneas of 5 seconds duratational age, 29 weeks (range, 25-32 weeks). Median tion or longer was recorded. Additionally, the number of day of study was day 16 (range, 2-55 days) at a median apneas of 2 seconds or longer was recorded during each post-conceptional age of 32 weeks (range, 26-36 weeks). obstruction, as well as any apnea starting during the 10 Of the studies, 38 were performed while the infant was seconds following the obstruction. In order to rule out the receiving methylxanthine treatment for apnea. Prior to possibility of chance association between external obentering the study, six infants had received mechanical struction and apnea, the numbers of episodes of 2 seconds ventilation for a period of up to 10 days, but none had or longer, which occurred or terminated during the 10 respiratory symptoms at the time of the study. Two in- seconds prior to the obstruction, were also recorded. All apneas were classified into central, mixed, or obfants had ultrasound evidence of intraventricular hemorrhage, one with parenchymal changes. Neither of these structive episodes by comparing the two channels of respiratory effort with the airflow trace. Apnea was regarded infants were symptomatic. Polygraphic recordings were made prior to the main as central if there was no respiratory effort on the impedpart of the study, in order to assess apnea frequency at ance and inductance traces. It was regarded as mixed if that time. Thoracic impedance, abdominal respiratory there was at least one obstructed breath and a central inductive plethysmography, electrocardiogram (ECG). element of at least 3 seconds. If respiratory efforts continand oxygen saturation (Sa,?) were recorded on to tape ued, and/or the pause between the efforts was less than 3 (Racal, Store 4) and subsequently analyzed manually. seconds, obstructive apnea was diagnosed. By analysis of whether the cardiac impulse was transFull details have been given elsewhere. I ' Median duration of these recordings was 3.96 hours (range, 2.19 to mitted up the airway, we were also able to detect whether 5.29). For the purposes o f this part of the study, apnea the upper airway was patent during apnea. For spontane-
External Obstruction in Preterms
ous apneas, and those following obstruction, the presence or absence of this cardiac artefact was recorded. However, it was not actually possible to assess patency during the external obstructions. Tap closure itself interrupted the passage of air to the pneumotachograph and caused loss of signal on the aifflow trace. For each external obstruction, the point in the respiratory cycle at which the obstruction was induced was recorded from analysis of the flow trace. This was classified as either mid-inspiratory, end-inspiratory,mid-expiratory, or end-expiratory. In order to assess the variability of the response with age and apnea frequency, an obstruction score was calculated for each study. This was the mean of the number of apneas during and after the obstruction, per study, expressed as a percentage of the number of obstructions in that study. Results were statistically analyzed using the MannWhitney U test, x2 test, and simple regression analysis. Approval for the study was given by the Nottingham Ethics Committee and informed parental consent was obtained.
235
ECG
.
I . I l .
FLOW
1 1 1
I
RIP
-02
10 s a
Fig. 1. An external obstruction, demonstrated by the loss of signal on the airflow trace (FLOW). There is no change in the infant’s respiratory pattern on the respiratory inductive plethysmography trace (RIP).
RESULTS ECG A total of 398 obstructions are recorded. A mean of five obstructions (range, 3 to 7) per study was produced, with a mean duration of 5.5 seconds (range, 4-9.5s). Arousal, with gross body movements, occurred in association with 18 obstructions. These were impossible to interpret and so were omitted from further analysis, leaving 380 obstructions. Apnea was not seen (Fig. 1) in 230 of the 380 obstructions, or 60% of the total. On 72 occasions (19%) apnea RIP occurred during the obstruction (Fig. 2). This usually happened as soon as the tap was turned, but occasionally one or two obstructed breaths were taken first. More common, though, was apnea following relief of the obstruction, which happened with 122 obstructions (32%), SaO, as seen in Figure 3. On 44 occasions (1 1%) apnea oc10 sec curred both during and after the obstruction. Commonly this was a continuous episode, but sometimes a few Fig. 2. An external obstruction, during which the infant ceases respiratory effort on respiratory inductive plethysmography breaths were taken on relief of the obstruction, with ap- all (RIP). Respiratory effort resumes promptly on relief of the obnea occurring again soon afterwards. struction. These events happened more frequently than would be expected by pure chance. Short pauses during the 10 seconds prior to obstruction occurred on 29 occasions closure, as demonstrated by the absence of cardiac arte(8%). This was statistically less frequent than both the fact, was present in 91 episodes (75%). Therefore, folnumber during and after the obstruction (P = 0.0001 for lowing an external obstruction, airway closure was seen both; x2 test). The apneas before obstruction were also to occur in roughly half of apparently central apneas. The point in the respiratory cycle at which obstruction shorter, with a mean duration of 3.3 s (SD 2 I .6s), compared to 6.6 s (SD 2 5.3 s) afterwards (P C 0.0001, was induced had no significant effect on the production of apnea, either during or after the obstruction. Table 1 Mann-Whitney U test). Of the 122 apneas following obstruction, 69 (57%) shows the relationship between respiratory phase at obwere central, 33 (27%) mixed, and 20 (16%) obstructive, struction and the presence or absence of apnea following as determined by conventional criteria. However, airway it. The results for apnea during the obstruction revealed a
236
Upton et al.
ence. Mean obstruction score was 16%in 41 studies with
ECG
analysis, between the number of spontaneous apieas and the obstruction score (r = 0.53, P = 0.0001). When obstruction score is compared with respect to age, then it seems to vary with postnatal, rather than postconceptional age. The obstruction score is significantly lower in the first 2 weeks of life than in the third and fourth weeks. RIP The mean score in the 36 studies during the first 2 weeks 1 was 20%, compared to 34%.in 23 studies during weeks 3 and 4 (P < 0.03,Mann-Whitney U test). Although there was a tendency for the score to later fall, being a mean 26% at weeks 5 and later, this did not reach statistical significance. -0. lowTreatment with methylxanthines did not appear to affect the obstruction score. The mean was 22% Fig. 3. An external obstruction followed by the infant develop ing central apnea. The airway remains patent, as demonstrated while receiving, and 28% while not receiving theophylby the presence of the cardiac artefact on the airflow trace. line (NS).
1
-
scan-
TABLE 1-Presence or Absence of Apnea Following Obstruction, Divided With Respect to Respiratory Phase at Obstruction"
DISCUSSION
This study has confirmed that apnea is a common event in association with external airway obstruction. Our preRespiratory phase vious study2 has been criticized, as the apnea seen with occlusion may have been a chance association. I 3 The Mid-inspiratory 34 20 temporal relationship seen previously between obstruc24 12 End-inspiratory Mid-expiratory 76 25 tion and apnea has always made this unlikely. The current End-expiratory 1 24 65 study confirms that respiratory pauses during and after occlusion are significantly more common than those beTotal 258 122 forehand. Additionally, those following the external ob'Differences are. not statistically significant (x2test). struction were longer than those preceding it. The mechanism by which the apneic response to obstruction occurs is not clear. As previously discussed, if similar, non-significant distribution. The length of the the Hering-Breuer inflation reflex is important, one obstruction also made no significant difference as to would expect occlusion at points in the respiratory cycle other than end-expiration to be more potent at producing whether apnea was induced. The mean obstruction score per study was 25% (range, apnea. No such variation with the respiratory cycle was 0-90%). The median number of spontaneous apneas re- seen and we, therefore, feel the Hering-Breuer inflation corded during the airflow study was 2 (range, 0-19). As reflex is not a vital part of this response. Trigeminal reflexes, which may be affected by applithe total time of study varied slightly and the time during which the mask was actually in position was affected by cation of a face mask, clearly play a role in the control of any intermittent episodes of arousal, the apnea frequency respiration.14 Application of a face mask in preterm inwas additionally calculated from the polygraphic record- fants can result in hypoventilation as measured by endings performed prior to the airflow study. Median apnea tidal CO,, but frank apneic episodes were not recorded.Is frequency was 2 significant episodes per hour (range, Indeed, in more mature infants application of a face mask may increase minute ventilation. l 6 In our study the mask 0-17). If the obstruction score is compared to apnea fre- had been applied for many minutes prior to the obstrucquency, there is no significant difference between the 39 tion; therefore any acute effects of mask application were studies with 2 unlikely to influence the response. Theoretically, the inepisodedhr (mean, 23 vs. 27%; NS). If, however, studies terruption of airflow to the trigeminal area with obstrucare divided with respect to the number of spontaneous 5 tion, and its subsequent reintroduction, may have stimusecond apneas during the airflow study, there is a differ- lated reflex changes. That these should result in apnea, Number without apnea after obstruction
Number with apnea after obstruction
External Obstruction in Preterms
even in susceptible infants, seems unlikely with our present knowledge of such reflexes.14 The intercostal-phrenic inhibitory reflex would appear to be a more likely mechanism for the production of apnea. Rib cage distortion, produced by both external pressure on the chest wall’7 and respiratory loading,’* has been shown to inhibit inspiratory effort in infants. Such a reflex has been used to explain the inspiratory shortening in response to occlusion seen in small preterm infants, as opposed to more mature infants, who exhibit inspiratory prolongation. ‘9*20 In susceptible infants such a reflex may inhibit inspiration sufficiently to produce apnea. Such a hypothesis is attractive, as it would help to explain why premature infants with more compliant chest walls show this response, whereas term infants do not. The fact that airway closure was seen in 75% of apneas following external obstruction would suggest that upper airway patency is affected by the occlusion. Airway closure itself may result in an inhibition of respiratory efforts’ and may therefore also be important in the production of apnea following airway occlusion. That such an external obstruction may result in intrinsic airway closure is highly likely on the basis of current evidence. Extrathoracic airway patency is maintained actively in life mainly by the actions of the genioglossus and geniohyoid.” Genioglossus activity is potentiated by negative pressure within the upper airway, such as occurs during inspiration, but decreases in response to positive pressure, which may be produced by external obstruction.22 Therefore, an external obstruction may produce a fall in genioglossus activity, resulting in intrinsic airway closure and subsequent apnea. However, whether an occlusion produces a negative or positive pressure in the pharynx depends on the point in the respiratory cycle at which the occlusion occurs. Occlusion at end-expiration would create a negative pressure, which should potentiate genioglossus activity. In practice this has been demonstrated by Carlo et al. ,23 who measured submental electromyograms as a reflection of genioglossus activity in ten preterm infants whose airways were occluded at end-expiration. Therefore, rather like the Hering-Breuer reflex, one would expect that occlusions at other points in the respiratory cycle would be more potent at producing airway closure and subsequent apnea. This was not seen to occur in this study, making it more doubtful that these factors are important in the apneic response to external obstruction. However, care needs to be taken in the interpretation of this data. It has been shown that, not surprisingly, the response of genioglossus differs in spontaneous mixed apneas and in external obstruction^.^^ Another important factor may be the activity of airway receptors, sensitive to flow or pressure. In animal experiments by Abu-Osba et a1.” flow up and down the upper airway resulted in the maintenance of pharyngeal patency
237
by increased activity of the genioglossus. This was abolished by topical anesthesia of the upper airway, suggesting that there is feedback from airway receptors to maintain genioglossus activity. Therefore, when flow ceases as a consequence of an obstruction, genioglossus activity drops and airway collapse may occur. Negative pressure produced by an external occlusion during inspiration may also have an effect on pressure sensitive receptors in the upper airway, resulting in an inhibition of thoracic inspiratory muscles.26 Clearly this may be another factor predisposing to apnea as a consequence of external obstruction, but once again the effect would be expected to be detected more often in occlusions during inspiration, which we have not seen. In spite of all these factors, Cohen and HendersonSmart found that the airways of infants were remarkably stable in response to occlusion. l 3 However, using their method of comparing esophageal and mask pressures during an occlusion, they could only detect obstructed breaths and they were not able to determine whether airway closure had occurred in the absence of respiratory effort against a frank obstruction. We have examined closure per se, by analysis of the transmission of the cardiac artefact up the airway. The relevance of the apneic response to occlusion to the pathophysiology of apnea is unclear. Although the obstruction score in our study correlated well with the number of spontaneous apneas, it did not relate to apnea frequency measured during the initial polygraphic recordings. The response to obstruction occurred more frequently than by pure chance, and therefore we do not believe that the relationship between the obstruction score and number of apneas detected by airflow is coincidental. Rather, it seems that the apneic response to occlusion may be important in the prolongation of initially short respiratory pauses, during which airway closure may occur. Obstructed breaths preceded the central element in 38% of mixed apneas in one study4 and we found a similar proportion in the spontaneous mixed apneas recorded here.9 It seems likely, then, that the response is relevant to the production of mixed apnea. The fact that it did not correlate with apnea frequency measured by polygraphic recordings may have been because we did not measure airflow during this time, and a number of mixed apneas may have been missed. Alternatively, the response may change rapidly, for example with sleep state,” and only reflect the apnea frequency over a short period of time, as measured during our airflow study. The apnea frequency measured by pneumograms is also notoriously variable over a short space of time” and may not accurately reflect the frequency even only 4 hours later. The time course of the response is also interesting. In our original study, older preterm infants seemed more likely to respond to occlusion with apnea2 and we have confirmed that the response is not prominent in the first
238
Upton et al.
14. Haddad GG, Mellins RB. The role of airway receptors in the control of respiration in infants: A review. J Pediatr. 1977; 911281-286. 15. Chernick V, Avery ME. Response of premature infants with periodic breathing to ventilatory stimuli. J Appl Physiol. 1966; 21:43W0. 16. Dolfin T, Duffty P, Wilkes D, England S, Bryan H. Effects of a face mask and pneumotachograph on breathing in sleeping infants. Am Rev Respir Dis. 1983; 128:977-979. REFERENCES 17. Knill R, Bryan AC. An intercostal-phrenic inhibitory reflex in human ncwbom infants. J Appl Physiol. 1976; 401352-356. I . Swift PGF, Emery JL. Clinical observations on response to nasal 18. Knill R, Andrews W, Bryan AC, Bryan MH. Respiratory load occlusion in infancy. Arch Dis Child. 1973; 48:947-951. compensation in infants. J Appl Physiol. 1976; 40357-361. 2. Milner AD, Saunders RA, Hopkin 1E. Apnoea induced by airflow 19. Thach BT, Frantz ID, Adler SM, Taeusch HW. Maturation of obstruction. Arch Dis Child. 1977; 52:379-382. reflexes influencing inspiratory duration in human infants. J Appl 3. Stark AR, Thach BT. Mechanisms of airway obstruction leading Physiol. 1978; 45:203-21 I. to apnea in newborn infants. J Pediatr. 1976; 89:982-985. 4. Thach BT, Stark AR. Spontaneous neck flexion and airway ob- 20. Gerhardt T, Bancalari E. Apnea of prematurity: 11. Respiratory reflexes. Pediatrics. 1984; 74:63-66. struction during apneic spells in pretem infants. J Pediatr. 1979; 21. Brouillette RT, Thach BT. A neurornuscular mechanism main94:275-281. taining extrathoracic airway patency. J Appl Physiol. 1979: 5 . Dransfield DA, Spitzer AR. Fox WW. Episodic airway obstruc461772-779. tion in premature infants. Am J Dis Child. 1983; 137:44143. 6. Butcher-Puech MC, Henderson-Smart DJ, Holley D, Lacey JL, 22. Mathew OP, Abu-Osba YK, Thach BT. Influence of upper airway pressure changes on genioglossus muscle respiratory activity. J Edwards DA. Relation between apnoea duration and type and Appl Physiol. 1982; 5 2 : 4 3 8 4 4 . neurological status of pretenn infants. Arch Dis Child. 1985; 23. Carlo WA, Miller MJ, Martin RJ. Differential response of respira60:953-958. tory muscles to airway occlusion in infants. J Appl Physiol. 1985; 7. Muttitt SC, Finer NN, Tierney AJ, Rossmann J. Neonatal apnea: 59:847-852. Diagnosis by nurse versus computer. Pediatrics. 1988; 82:71324. Gauda EB, Miller MJ, Carlo WA, DiFiore JM, Martin RJ. Genio720. glossus and diaphragm activity during obstructive apnea and air8. Milner AD, Boon AW, Saunders RA, Hopkin IE. Upper airways way occlusion in infants. Pediatr Res. 1989; 26583-587. obstruction and apnoea in preterm babies. Arch Dis Child. 1980; 25. Abu-Osba YK, Mathew OP, Thach BT. An animal model for 55:22-25. airway sensory deprivation producing obstructive apnea with post9. Upton CJ, Milner AD, Stokes GM. Upper airway patency during mortem findings of sudden infant death. Pediatrics. 198 I;h8:79& apnoeaof prematurity. Arch Dis Child. 1992; 67:419-424. 801. 10. Cross KW, Klaus M, Tooley WH, Weisser K. The response of the 26. Thach BT, Schefft GL, Pickens DL, Menon AP. Influence of newborn baby to inflation of the lungs. J Physiol. 1960; 151:551upper airway negative pressure reflex on response to airway occlu565. sion in sleeping infants. J Appl Physiol. 1989; 67:745)-755. 1 1 . Upton CJ, Milner AD, Stokes GM. Combined impedance and inductance for the detection of apnoea of prematurity. Early Hum 27. Frantz ID, Adler SM, Abroms IF, Thach BT. Respiratory response to airway occlusion in infants: Sleep state and maturation. J Dev. 1990; 2455-63. Appl Physiol. 1976;41:634-638. 12. Miller MJ, Martin RJ, Carlo WA, Fouke JM, Strohl KP, Fanaroff AA. Oral breathing in newborn infants. J Pediatr. 1985; 107:465- 28. Hunt CE, Brouillette RT, Liu K , Klemka L. Day-to-day pneumogram variability. Pediatr Res. 1985; 19:174-177. 469. 13. Cohen G, Henderson-Smart DJ. Upper airway stability and apnea 29. Miller HC, Behrle FC, Smull NW. Severe apnea and irregular respiratory rhythms among premature infants. Pediatrics. 1959; during nasal occlusion in newborn infants. J Appl Physiol. 1986; 23:676-685. 6011511-1517.
week or two of life. Although the existence of a “honeymoon period” as such is unlikely, such a trend may reflect the time course of apneic episodes, as apnea tends to become more problematical at the end of the first week of life. 2y
