Background. Apnea of prematurity remains among the most commonly diagnosed conditions in the Newborn Intensive Care Unit and may prolong hospital stays in some infants. Because survival of extremely premature infants has improved markedly, the natural history of apnea in this population needs to be reassessed.
Objective. To document the natural history of recurrent apnea and/or bradycardia events in infants delivered at 24 to 28 weeks' gestation.
Methods. Medical records of all infants delivered at 24 to 28 weeks' gestation admitted to the Brigham and Women's Hospital Newborn Intensive Care Unit between January 1989 and March 1994 were reviewed to document the clinical course of apnea of prematurity. Subjects were included in the study sample if they were discharged home from the Brigham and Women's Hospital or after transfer to an affiliated hospital. Recordings of apnea and/or bradycardia events were based on nursing observations of monitor alarms and assessment of the infant's condition.
Results. Of 457 eligible infants, 226 were included in the study sample and stratified by gestational age at birth assigned by the attending neonatologist. The time to resolution of recurrent apnea/bradycardia events was longer with lower gestational age at birth. Apnea/bradycardia events were frequently observed beyond 36 weeks' postconceptional age in all gestational age groups. The incidence of apnea persisting beyond 38 weeks postconceptional age was significantly higher in the 24- to 27-week infants combined compared with the 28-week infants.
Conclusions. Apnea of prematurity frequently persists beyond term gestation in infants delivered at 24 to 28 weeks' gestational age. These persistent apnea and/or bradycardia events may contribute to prolonged hospitalization. Programs to promote earlier discharge of premature infants should take into account the variability in resolution of apnea and specifically address management of persistent apnea.
- extremely low birth weight infant
- control of breathing
- physiologic monitoring
- discharge planning
Apnea of prematurity is among the most commonly diagnosed conditions in the newborn intensive care unit (NICU) and may prolong the hospital stay of some infants. Resolution of recurrent apnea and bradycardia episodes and completion of an “apnea-free” period generally are considered to be preconditions for discharge of premature infants without a home cardiorespiratory monitor.1,2The evaluation and management of premature infants who suffer from persistent apneic and/or bradycardic events after they are otherwise ready for discharge home remain controversial, and practice varies considerably among neonatologists.
Nearly two decades have passed since Henderson-Smart's study documented the incidence and duration of recurrent apnea in >25 000 term and premature infants.3 In the 249 infants diagnosed with recurrent apneic episodes, the incidence and duration of apnea increased with decreasing gestational age at birth. Apneic spells ceased in 92% of the infants by 37 weeks' postconceptional age (PCA), and >98% were apnea-free by 40 weeks. The high mortality rate limited the number of extremely premature infants studied; indeed, the data sample included only 19 infants delivered at <28 weeks' gestation. Despite the small number of extremely premature infants included, these observations continue to influence clinical expectations for the time course to resolution of apnea of prematurity, and consequently length of hospital stay, in this population.
Because survival of extremely premature infants has improved markedly, the natural history of apnea in this population needs to be reassessed. With increasing financial pressures to decrease length of hospital stay, improved documentation of the time course to resolution of apnea events is essential to the development of strategies for earlier hospital discharge. The objective of this study was to document the natural history of recurrent apnea and/or bradycardia events in infants delivered at 24 to 28 weeks' gestation.
Medical records of all infants delivered at 24 to 28 weeks' gestation admitted to the Brigham and Women's Hospital NICU between January 1989 and March 1994 were reviewed to document the clinical course of apnea of prematurity. Subjects were included in the study sample if they were discharged home from the Brigham and Women's Hospital NICU or after transfer to the Beth Israel Hospital Special Care Nursery, an affiliated level II nursery during the study period. Infants were excluded if they did not survive to discharge or were transferred to another level III NICU after birth. Infants were stratified by the gestational age assigned by the attending neonatologist at birth. Surfactant therapy was available throughout the study period.
Recordings of apnea and/or bradycardia events were based on nursing observations of monitor alarms and documented assessment of the infant's clinical condition at the time of the alarm. All infants had cardiac and respiratory impedance monitoring; in general, low heart rate alarms were set at 100 beats per minute in infants with a PCA (defined as the sum of gestational age at birth and postnatal age in weeks) <34 to 35 weeks and at 80 beats per minute thereafter. Apnea alarms were set at 20 seconds. Some infants were monitored concurrently by pulse oximetry. Nursing documentation included whether apnea alone, bradycardia alone, or both apnea and bradycardia was observed and the presence or absence of skin color change (or oxygen desaturation), as well as the need for intervention to interrupt the recorded event. The need for intervention (eg, tactile stimulation, supplemental oxygen, or bag and mask ventilation) during an apnea and/or bradycardia event was based on nursing assessment of the infant's condition at the time of the monitor alarm. Events that did not require intervention were noted as “self-resolved.”
To assess the influence of recurrent apnea and/or bradycardia episodes on length of hospital stay, we recorded other clinical conditions that might have influenced the decision to discharge an infant home, including adequate temperature control in an open crib, time to full nipple feedings, use of methylxanthines, and discharge with a home cardiorespiratory monitor. We also recorded medical conditions that might have affected the time course to resolution of apnea, including the diagnosis of chronic lung disease (CLD) (defined as supplemental oxygen requirement and abnormal chest radiograph at 36 weeks' PCA), and severe abnormalities on head ultrasound examination (grade 3 or 4 intraventricular hemorrhage and/or periventricular white matter disease). One-way analysis of variance and χ2statistical analyses were performed to compare the different gestational age groups; differences were considered significant ifP < .05. To assess the influence of a diagnosis of CLD or severe head ultrasound abnormalities on the duration of apnea, we performed a multivariate linear regression predicting PCA at last apnea/bradycardia event, controlling for gestational age at birth and a diagnosis of CLD or severe head ultrasound abnormalities.
Clinical guidelines for the management of apnea of prematurity in effect for the duration of the study period recommended that all premature infants with recurrent apnea and/or bradycardia episodes have a minimum of five “spell-free” days off methylxanthine therapy before discharge home. Clinical guidelines did not include recommendations for pneumocardiogram recordings before discharge; home cardiorespiratory monitor use, discontinuation of or discharge to home on methylxanthine therapy, and time of discharge were determined by the attending physician.
During the study period, 788 infants with a gestational age of between 24 and 28 weeks were admitted to the NICU. Excluded from the analysis were infants who died (21%) and infants transferred to other level III NICUs immediately after birth (21%). Of the 457 eligible infants, 226 (49.5%) were included in the study sample. The remaining infants were transferred to nurseries at community hospitals before discharge home. Details of the study population are presented in Table1.
All infants in the study population were diagnosed with apnea of prematurity and treated with a methylxanthine. The PCA at discontinuation of methylxanthine therapy was slightly higher in the 24- to 27-week infants compared with the 28-week infants; this difference was statistically significant only for the 24- and 26-week infants (P < .05) (Table 1).
As expected, the time to resolution of recurrent apnea/bradycardia events was longer with lower gestational age at birth (P < .01) (Fig1). Recurrent apnea/bradycardia events were frequently observed beyond 36 weeks' PCA in all gestational age groups (Fig 1). Apnea/bradycardia events persisted beyond 36 weeks' PCA more frequently in infants delivered between 24 and 27 weeks' gestational age compared with those delivered at 28 weeks' gestational age (P < .05) (Table2). The incidence of apnea/bradycardia events persisting beyond 38 weeks' PCA was significantly higher in both the 24- and 26-week infants compared with the 28-week gestational age infants (P < .05) (Table 2). Only the 24-week infants continued to show a higher incidence of persistent apnea beyond 40 weeks' PCA (P < .05) (Table2). To account for possible sample size effects, the 24- to 27-week infants were analyzed together as a single group and compared with the 28-week infants. The 24- through 27-week infants combined also had a higher frequency of apnea/bradycardia events persisting beyond 38 weeks', but not beyond 40 weeks', PCA compared with the 28-week infants (P < .03).
In addition to the later resolution of apnea/bradycardia episodes, the PCA at discharge of infants delivered between 24 and 27 weeks' gestation was higher compared with the discharge PCA of the 28-week gestation infants (Table 1) (P < .05). The PCAs at discharge for the 24- to 27-week gestation infants were similar (Table 1).
A diagnosis of CLD was correlated by multivariate linear regression with a longer persistence of recurrent apnea/bradycardia events in all gestational age groups (P < .05) (Table3). No correlation between the presence of severe head ultrasound abnormalities and persistence of recurrent apnea/bradycardia events was observed for any of the gestational age groups. The later resolution of recurrent apnea/bradycardia events in the most premature infants was related to the higher incidence of CLD in these infants (Table 1). When corrected for a diagnosis of CLD, there remained a strong trend for a later PCA at last apnea/bradycardia event in the 24-week infants compared with the 28-week infants (P = .06). A similar, but weaker, trend was evident for the 25- through 27-week infants.
The time course to resolution of recurrent apnea and/or bradycardia episodes followed a predictable pattern in most infants. Apnea/bradycardia events requiring nursing intervention were the first to disappear, followed by self-resolved episodes with both observed apnea and bradycardia, and finally transient, self-resolved episodes of bradycardia without observed apnea (Fig2). These self-resolved bradycardia episodes without observed apnea persisted for 7.4 ± 1.5 days after the last event requiring nursing intervention (Fig 2). The postnatal day that infants accomplished both full nipple feeds and adequate temperature control in an open crib was strongly correlated with the last postnatal day with a documented apnea/bradycardia event requiring nursing intervention (P < .001) (Fig3). Infants remained hospitalized for 16.4 ± 1.3 days after they were otherwise ready for discharge home (ie, on full nipple feeds and in an open crib) in part awaiting resolution of these self-resolving bradycardia events and completion of an event-free period (Fig 3).
To assess whether the decision to discharge infants on a home cardiorespiratory monitor affected length of hospital stay, we examined use of home monitors in the study population. Overall, 35 (15.4%) of the study subjects were discharged home on a monitor; no significant differences in home monitor use for the different gestational age groups were observed (Table 1). However, the PCA at discharge was higher for infants sent home on a monitor compared with those who were not (41 ± 2.6 vs 38 ± 2.6 weeks, respectively;P < .001, unpaired t test). Of the 35 infants discharged home on a monitor, 10 also were discharged on methylxanthine therapy. No infant was discharged on a methylxanthine without a home monitor.
This study documents the time to resolution of recurrent apnea/bradycardia events in infants delivered at a gestational age of between 24 and 28 weeks. We found that the duration of recurrent apnea/bradycardia episodes in this population increased with decreasing gestational age at birth, consistent with Henderson-Smart's data on a cohort of infants delivered later in gestation.3 In contrast with earlier observations, however, recurrent apnea and bradycardia events in extremely premature infants frequently persisted beyond term gestation, especially in infants delivered at the youngest gestational ages. These events often occurred after infants were otherwise ready for discharge home, suggesting that persistent apnea/bradycardia events contribute to longer hospital stays in extremely premature infants.
The pathogenesis of apnea in premature infants is multifactorial.4 Immaturity of central respiratory control is widely accepted as a causative role.4 In addition, a variety of environmental, metabolic, and cardiorespiratory events may precipitate apnea in premature infants,4-6 suggesting enhanced vulnerability of their brainstem respiratory centers to central and peripheral inhibitory mechanisms. Maturation also affects central and peripheral chemoreceptor responses to hypercarbia and hypoxia.6,7 Approximately half of all apneic spells in premature infants involve some degree of upper airway obstruction, indicating that respiratory muscle output and coordination also are disturbed.8-10 The resolution of recurrent apneic spells involves the maturation of these multiple aspects of respiratory control. It is not surprising, therefore, that the time required for apnea of prematurity to resolve can be quite variable across different gestational ages, as we observed. These maturational processes appear to be most delayed in infants delivered at the youngest gestational ages. Our data also suggest that maturation of respiratory control occurs in parallel with other measures of physiologic maturity, including coordination of suck and swallow and regulation of temperature.
We expected to see an association between greater pulmonary or central nervous system dysfunction with a more prolonged persistence of apnea. A diagnosis of CLD was associated with later resolution of recurrent apnea/bradycardia events in all gestational age groups (Table 3). No such relationship was observed, however, with the presence of severe head ultrasound abnormalities. The later PCA at resolution of apnea/bradycardia events in the 24- to 27-week infants compared with the 28-week infants was in part related to the higher incidence of CLD in the more premature infants. However, when corrected for a diagnosis of CLD, we observed a continued effect of gestational age at birth on the PCA at last apnea/bradycardia event in the infants delivered at 24 weeks. The lack of an observed effect of gestational age at birth with later resolution of apnea/bradycardia events when corrected for a diagnosis of CLD in the 25- to 27-week infants may be a function of the sample size at these gestational ages. Alternatively, it is possible that the maturation of peripheral and central respiratory control necessary for resolution of recurrent apnea/bradycardia events is comparatively delayed only in the most premature infants. The lack of an association between severe head ultrasound abnormalities with longer persistence of apnea suggests that the central nervous system disturbances we examined have a minimal impact on maturation of central respiratory control.
Similar to Henderson-Smart's study,3 we used nursing observations of monitor alarms and the infant's clinical condition to document the occurrence of apnea and bradycardia episodes. Standard cardiac and respiratory impedance monitors, which were used to monitor all subjects, may not adequately record all apnea events, especially those with an obstructive component.11,12 Although the nursing staff in our NICUs are trained to observe the infant's condition during a monitor alarm and record the presence of respiratory effort and cyanosis, it is possible that some events were described inadequately. It is likely, however, that these limitations would underestimate rather than overestimate the frequency and duration of recurrent apnea. In addition, nursing documentation of monitor alarms is often the only description of apnea and bradycardia events available in most NICUs to use in formulating discharge decisions for individual infants.
We documented apnea with associated bradycardia as well as transient self-resolved bradycardia episodes without observed apnea. These transient episodes of bradycardia were the last events to disappear in most infants (Fig 2). The etiology and physiologic significance of these self-resolved bradycardia events remain unclear.13-15 Episodes of bradycardia in apnea of prematurity are most often associated with apneas of longer duration and may be precipitated by respiratory efforts against an obstructed airway.16-18 It has been hypothesized that bradycardia occurring during these longer apneic events may be a reflex response to the consequent hypoxemia and lack of lung inflation.17,18Some authors have argued, however, that the onset of bradycardia in apneic episodes is too rapid to be initiated by hypoxemia, and that they rather represent a primary, vagally mediated cardiac rhythm disturbance.16,19 Transient self-resolved episodes of bradycardia without associated central or obstructive apnea also may occur in convalescent premature infants before and after discharge and in healthy term infants up to 6 months of age.14,15 Thus, these events may be a normal reflex response that should not prompt additional evaluation or monitoring.
It is likely that some of the transient bradycardia episodes we observed were associated with obstructive apnea that remained undetected.11 Some also may have been primary vagally mediated events without associated apnea. The predictable pattern of resolution of apnea and bradycardia events we observed suggests that these transient self-resolved bradycardias are part of the continuum of apnea of prematurity (Fig 2). The risk, if any, of subsequent, more severe apnea/bradycardia events in premature infants with a history of recurrent apnea who continue to demonstrate these transient episodes of bradycardia has not been studied adequately. Because of uncertainty as to etiology and clinical significance, our practice has been to await resolution of all monitored events, including these transient episodes of bradycardia, before discharging an infant home. This practice likely delayed discharge in some patients.
Home cardiorespiratory monitors were prescribed for 15.4% of our study population (Table 1). Home monitor use did not shorten hospital stay; rather, these infants were discharged almost 3 weeks later than infants who were not discharged home on a monitor. The rate of home monitor use in our study population was substantially lower than that reported for a sample of infants born weighing <1500 g obtained from the 1988 National Maternal and Infant Health Survey.20 In this study, almost one third of the 1010 infants included were discharged with a home monitor, which indicates widespread use of this technology in discharge planning for low birth weight infants. Surprisingly, this study also showed unexplained racial disparity in home monitor use, with a higher frequency in nonblack infants compared with black infants. The use of an apnea monitor in the home also was associated with a significantly higher risk for subsequent rehospitalization of these infants. These unexpected effects of short-term home monitoring in premature infants require additional study before advocating its use to promote earlier discharge.
Early discharge programs for premature infants recently have been described and evaluated.21-24 These early discharge programs generally have combined more intensive parental teaching with stronger community-based supports for parents of premature infants who meet established discharge criteria, which include adequate nipple feeding, temperature regulation, and absence of recurrent apnea/bradycardia episodes. Most programs have shown cost savings resulting from shorter hospital stays without an increase in postdischarge morbidity.21,23,24 Few of these studies have addressed the impact of recurrent apnea and bradycardia events on length of hospital stay, especially in the most immature infants. In one randomized trial of an early discharge program, significant reductions in length of hospital stay were observed only in infants with birth weights between 1500 and 2000 g.23 No reduction in length of hospital stay was observed in infants of <1500 g birth weight, in part because of delayed resolution of recurrent apnea/bradycardia episodes in these infants. Other studies of early discharge programs for premature infants have accomplished shorter hospital stays even in the smallest, most immature infants.21,24 In one study with data comparable to our own, the mean length of stay and weight at discharge for infants with birth weights between 500 and 750 g in the early discharge group were nearly identical to that of our 24-week infants (110 vs 107 days and 2417 vs 2544 g, respectively).24 A recent study documenting time to discharge for low birth weight infants also showed that PCA at discharge was highest for infants with the lowest birth weights.25 These data suggest that our experience with prolonged length of stay in these smallest infants is representative, and that resolution of recurrent apnea and bradycardia episodes and attainment of other measures of physiologic maturity may delay hospital discharge of extremely premature infants.
In conclusion, we have shown that apnea of prematurity frequently persists beyond term gestation in infants delivered at 24 to 28 weeks' gestational age. This observation, in part, may be secondary to the higher incidence of CLD at the earliest gestational ages. These persistent apnea and/or bradycardia events may contribute significantly to prolonged hospitalization of some extremely premature infants. Additional research on the clinical significance of transient episodes of bradycardia is necessary to help in discharge planning. Programs to promote earlier discharge of premature infants should take into account the variability in resolution of apnea in this population, and specifically address management of persistent apnea and bradycardia.
We thank Douglas K. Richardson, MD, MBA, for his help in the statistical analysis of the data.
- Received October 8, 1996.
- Accepted January 9, 1997.
Reprint requests to (E.C.E.) Department of Newborn Medicine, Brigham and Women's Hospital, 75 Francis St, Boston, MA 02115.
This work was presented, in part, at the American Pediatric Society/Society for Pediatric Research Meetings, San Diego, CA, May 1995.
- NICU =
- newborn intensive care unit •
- PCA =
- postconceptional age •
- CLD =
- chronic lung disease •
- IVH =
- intraventricular hemorrhage
- ↵American Academy of Pediatrics, American College of Obstetricians and Gynecologists. Guidelines for Perinatal Care. 3rd ed. (Freeman RK, Poland RL, eds). Washington, DC: American Academy of Pediatrics and American College of Obstetricians and Gynecologists; 1992:109
- Rigatto H,
- Brady JP
- Gerhardt T,
- Bancalari E
- Milner AD,
- Boon AW,
- Saunders RA,
- Hopkin IE
- Warburton D,
- Stark AR,
- Taeusch HW
- Poets CF,
- Stebbens VA,
- Richard D,
- Southall DP
- Kattwinkel J,
- Fanaroff AA,
- Klaus MH
- Henderson-Smart DJ,
- Butcher-Puech MC,
- Edwards DA
- Upton CJ,
- Milner AD,
- Stokes GM
- Casiro OG,
- McKenzie ME,
- McFadyen L,
- et al.
- Copyright © 1997 American Academy of Pediatrics