Case report

A 5½ month old Irish white male infant of unrelated parents was admitted electively to the Rotunda paediatric unit for overnight sleep studies because of a history of apparent life threatening events. The infant was born by a spontaneous vaginal delivery at term after a normal pregnancy. Birth weight was 3030 g. Apgar scores were 9 at one minute and 10 at five minutes, and there were no neonatal or early infancy problems. At 8 weeks old he was found, in the early hours of the morning while sleeping, to be apnoeic, limp, and apparently lifeless, requiring initial shaking and subsequent mouth to mouth resuscitation. Immediately after this episode he was admitted to the local hospital where observation and investigation revealed no abnormality.

He was discharged home after three days on a home apnoea monitor. He had similar episodes at 11 and 15 weeks, occurring at 03:30 and 04:15, each requiring vigorous resuscitation. On hospital admission after each occasion no abnormality was found.

INVESTIGATIONS AND RESULTS

On admission he was well, growing and developing normally, weight 7.68 kg (50th centile), length 64 cm (25th centile), and head circumference 42 cm (50th centile). Physical examination was normal. He had no fever, was formula feeding normally, and had a resting respiratory rate of 28 breaths/min while quietly asleep. The following investigations were normal: full blood count, urea and electrolytes, arterial blood gasses, routine urine analysis, chest xray, and electrocardiogram. The remaining tests were conducted during his overnight sleep period in the prone sleeping position from 21:00 to 08:00. Autonomic function testing was done as previously described using the heart rate and blood pressure (systolic) response to a change from the horizontal to the upright position during sleep, and the heart rate variability during clinically staged quiet sleep.6 7 An eight hour recording during sleep of respiratory movements and heart rate using a Corometrics 512 neonatal cardiorespiratory monitor (Corometrics Medical Systems, Wallingford, Connecticut, USA) showed no prolonged apnoea (> 20 seconds) or bradycardia (< 90 beats/min). However, there was a profound reduction in heart rate variability in quiet sleep, with long term heart rate variability of 6 beats/min compared with a mean (SD) value of 24 (6.2) beats/min in a group of healthy 11–13 week old Irish infants.8 Long term heart rate variability is the difference between the maximum and the minimum heart rates over 512 consecutive beats, expressed in beats/min, derived from ECG chest electrodes using a peak detecting, instantaneous, digital, event to event cardiotachometer with an accuracy of ±1%.

Blood pressure was measured using an Ohio neonatal blood pressure monitor (Madison, Wisconsin, USA) and an appropriate sized infant cuff on the right upper arm. The values of systolic blood pressure were used in expressing the results. Resting blood pressure was defined as the mean of six readings taken at one minute intervals during sleep immediately before tilt testing. We measured heart rate response over one minute, and blood pressure response at one minute following a change from the horizontal prone sleeping position to the head up vertical position (with the infant’s head and trunk at 90° from the horizontal). The infant was lifted manually into this position and remained asleep during this study. His blood pressure fell by 14% when in the vertical head up position, and this was associated with a progressive fall in heart rate from 110 to 80 beats/min. This procedure was repeated twice and yielded a similar result on one occasion (a 12% fall in blood pressure associated with a fall in resting heart rate from 115 to 95 beats/min) and maintenance of resting blood pressure with no heart rate change on the third test.

The final test was to measure the response time to nasal occlusion. This was assessed during clinically staged quiet sleep by gentle digital occlusion of the orifices of the anterior nares, noting the time taken to arousal as shown by either head movement or the establishment of an oral airway. This infant had a brisk nasal occlusion response time of 5 seconds, compared with an average of 7.8 seconds in similarly tested normal infants.9 Following completion of the sleep studies at 08:00, the infant was fed, changed, and placed to sleep in the lateral position at 09:30 attached to a Graseby MR10 monitor (Graseby Medical Co, Watford, UK). He was found at 10:20 by his mother, dead in the prone position with the alarm sounding. A necropsy examination by an experienced paediatric pathologist ascribed the cause of death to SIDS.

Discussion

This infant was electively admitted for sleep studies because of three previous, severe, sleep related apparent life threatening episodes. He was studied while sleeping prone and overnight, both being risk factors for SIDS.2 3 He was well on admission, with a normal physical and developmental examination and normal vital signs. As measured, respiratory control was normal with no prolonged apnoea and a rapid response to airway occlusion, making sleep apnoea or a deficient response to airway occlusion unlikely causes of death. Autonomic function was abnormal, with a profoundly reduced heart rate variability and a fall in blood pressure associated with progressive bradycardia on upright postural challenge. This response implies poor peripheral vasomotor tone and an inappropriate heart rate response to diminished venous return.10

In adults, unexplained syncopal episodes are often associated with poor peripheral vasomotor control, best diagnosed by a head up tilt test that causes hypotension with an associated bradycardia and may reproduce the symptoms.10 This has been called neurocardiogenic syncope (or vasovagal syncope). While the precise pathways involved are uncertain, it clearly involves an inappropriate peripheral vasomotor response to a postural challenge, by allowing blood pressure to fall while failing to develop a compensatory tachycardia in the presence of a (presumed) reduction in central venous return. A report of six infants dying from SIDS while attached to a cardiorespiratory monitor at home with a built in memory showed a progressive bradycardia to be the major terminal event in all cases, with continuation of respiratory movements until death.11In neurocardiogenic syncope the essential feature seems to be poor peripheral vasomotor tone, diagnosed by a posturally induced fall in blood pressure associated with a paradoxical, vagally mediated bradycardia secondary to reduced cardiac distension, and in some patients followed by a period of sinus arrest.10 In some adult patients a reduction—or withdrawal—of resting sympathetic tone also occurs, causing peripheral, particularly muscle bed, vasodilatation. This further reduces venous return and reinforces the neurally mediated bradycardia. Poor vasomotor tone would be exacerbated by overheating, a known risk factor for SIDS, and the reduction in venous return accentuated by the physiological anaemia, which reaches its nadir at the peak age of occurrence of SIDS. This case suggests that a neurocardiac reflex can occur in infants, may have been involved in this infant’s death, and deserves further study both in the context of infant homeostasis and SIDS.