Abstract
BACKGROUND:
Increased cardiorespiratory events with bradycardia and oxygen desaturation have been reported in very low birthweight (VLBW) infants following stressors such as immunizations. These events are difficult to quantify and may be mild. Our group developed an automated algorithm to analyze bedside monitor data from NICU patients for events with bradycardia and prolonged oxygen desaturation (BDs) and used this to compare BDs 24 hours before and after potentially stressful interventions.
METHODS:
We included VLBW infants from 2012–2017 with data available around at least one of four interventions: two-month immunizations, retinopathy of prematurity (ROP) examinations, ROP therapy, and inguinal hernia surgery. We used a validated algorithm to analyze electrocardiogram heart rate and pulse oximeter saturation data (HR, SpO2) to quantify BD events of HR < 100 beats/minute for≥4 seconds with oxygen desaturation < 80%SpO2 for≥10 seconds. BDs were analyzed 24 hours before and after interventions using Wilcoxon rank-sum tests.
RESULTS:
In 354 of 493 (72%) interventions, BD frequency stayed the same or decreased in the 24 hours after the event. An increase of at least five BD’s occurred in 17/146 (12%) after immunizations, 85/290 (29%) after ROP examinations, 4/33 (12%) after ROP therapy, and 3/25 (12%) after hernia surgery. Infants with an increase in BDs after interventions had similar demographics compared to those without. More infants with an increase in BDs following immunizations were on CPAP or caffeine than those without.
CONCLUSIONS:
Most VLBW infants in our cohort had no increase in significant cardiorespiratory events in the 24 hours following potentially stressful interventions.
Introduction
Episodes of bradycardia and desaturation (BDs) associated with central, obstructive, or mixed apnea are common among preterm VLBW infants. These events diminish over time, and the resolution of significant BD events is an important maturational milestone impacting length of NICU stay [1, 2]. Understanding factors that influence the occurrence or recurrence of BDs might lead to preemptive strategies to mitigate these spells.
The most common physiologic stressor reported to increase cardiorespiratory spells in otherwise healthy VLBW infants is immunizations [3–5]. Concern about apnea has led to delayed immunization of preterm infants both in and beyond the NICU stay [6–8]. Another common procedure reported to be associated with increased cardiorespiratory events is retinopathy of prematurity (ROP) examinations [9–11]. Less common but not rare procedures potentially precipitating increased apnea include ROP treatment and inguinal hernia surgical repair [12, 13]. Reports of increased events after these stressors have a great deal of variability, in part due to different definitions and timing of events in relation to the stressor, and the method of quantifying them. Another important consideration is that individual infant responses vary widely. For example, many reports indicate that only a small fraction of infants experience an increase in cardiorespiratory spells following immunizations [14, 15]. Clinical interventions and therapies at the time of the stressor may also impact changes in events, such as level of respiratory support and administration of caffeine. For major interventions such as ROP treatment and hernia surgery, the requirement for anesthetic, sedative, and analgesic agents increase the risk of post-procedure apnea for some infants [16]. The occurrence of these events can have an impact on immediate clinical care and on discharge planning. Therefore, we sought to quantify events in our population of VLBW infants.
Quantification of cardiorespiratory events in most NICUs relies on nursing documentation, which has been shown to be inaccurate [17, 18]. Our group has collected years of NICU bedside monitor vital sign data and developed automated algorithms for quantifying episodes of bradycardia and desaturation [19, 20]. In the current analysis, we focused on clinically significant events, including both bradycardia and desaturation, with a relatively low and long desaturation threshold of SpO2 < 80%for at least 10 seconds. Using this definition, we compared the number of BD events in the 24 hours before and after two-month immunizations, retinopathy of prematurity (ROP) examination and therapy, and inguinal hernia surgery.
Methods
Patient cohort
The University of Virginia Institutional Review Board approved this retrospective study with a waiver of consent. We included infants with birth weight < 1500 grams, gestational age < 33 weeks, and available HR and SpO2 data surrounding the stressors admitted to the NICU at the University of Virginia from 2012–2017 and discharged home. We excluded infants if they had major chromosomal or congenital anomalies, were transferred to UVA beyond 32 weeks PMA, or had severe chronic lung disease (CLD). For this study, we defined severe CLD as requiring high-level respiratory support (ventilator or nasal continuous positive airway pressure (CPAP) beyond 36 weeks PMA. We limited the study to infants without severe CLD in order to limit the heterogeneity of lung disease within the cohort. Clinical and demographic data were collected from the electronic medical record and a NICU database.
Interventions and general unit guidelines
Immunizations
Two-month immunizations are generally administered as close as possible to 2 months chronologic age, irrespective of GA, with parental assent. The vaccinations recommended at this age include a combination vaccine with acellular pertussis, diphtheria, tetanus, inactivated polio, and hepatitis B (Pediarix), as well as vaccines against Streptococcus pneumoniae and Haemophilus influenzae type B [21]. We routinely give these immunizations all at once unless a parent requests to give them individually over 2–3 days. In these cases, we used the date and time of Pediarix administration as the timestamp for the event. Acetaminophen may be given as needed to treat or prevent elevated temperature or perceived pain in the 24 hours following immunizations.
ROP examinations
We assessed BDs around the ROP exam closest to 34 weeks PMA. No pharmacologic sedation agents are given. Cyclomydril drops are administered about one hour before the exam.
ROP therapy
Infants with severe ROP undergo either laser photoablation or, more commonly, intravitreal bevacizumab. Infants typically receive intravenous fentanyl, midazolam, and atropine before the procedure. Respiratory support may be increased transiently during and immediately following the procedure as needed.
Inguinal hernia surgery
The standard of care in the years studied was to surgically repair inguinal hernias just prior to NICU discharge, once an infant was otherwise ready to go home. General anesthesia with endotracheal intubation was used for all cases, and extubation typically occurred in the operating room or the NICU when spontaneously breathing well. Caffeine is given as needed, either by the anesthesiologist in the operating room or the NICU if apnea is either noted or anticipated. The first-line analgesia is with acetaminophen, and some infants received fentanyl for postoperative pain.
Heart rate and SpO2 analysis
All UVA NICU patients have bedside monitor vital sign data stored via the BedMaster system (Hillrom’s Medical Device Integration solution, Chicago, IL; formerly Excel Medical, Jupiter, FL). HR from the electrocardiogram was collected at 0.5 Hz. SpO2 from pulse oximetry with Masimo technology and 8 second averaging time was collected at 0.5 Hz. We analyzed BDs using methods and definitions from our prior work [20, 22]. An automated algorithm quantified episodes of bradycardia (HR < 100 beats/minute for at least 4 seconds) and desaturation (SpO2 < 80%for at least 10 seconds). We defined a BD event as bradycardia and desaturation occurring concurrently and join events as one if the HR or SpO2 rise above and back below the threshold within 4 or 10 seconds, respectively. For categorical analyses, we dichotomized infants as those with and without an increase in BD events by five or more in the 24 hours following the intervention.
Statistics
We summarized continuous variables as mean and standard deviation and categorical variables as number and percent for the patients included in each stressor analysis. We compared the number of BD events in the 24 hours before and after each intervention using Wilcoxon rank-sum tests. In addition, we compared the clinical characteristics of infants with and without an increase in at least five more BD events after compared to before each intervention.
Results
Patient population and clinical characteristics at the time of interventions
In the years of the study, 674 VLBW infants were admitted to the UVA NICU and 292 (43%) met inclusion criteria of surviving to discharge, not having significant congenital or chromosomal anomalies, not requiring high respiratory support at 36 weeks PMA, and having stored vital sign data. Table 1 shows the demographic and clinical information of infants with HR and SpO2 data available in the 24 hours before and after each intervention.
Characteristics of patients analyzed for each intervention
Characteristics of patients analyzed for each intervention
Continuous variables are presented as mean (standard deviation) and categorical variables as number (percent). Abbreviations: ROP, retinopathy of prematurity; CPAP, continuous positive airway pressure.
Two-month immunizations, ROP exams, and ROP therapy occurred around 34 to 35 weeks PMA for the infants analyzed, while hernia surgeries occurred closer to term corrected age (mean PMA 38 weeks). Of those analyzed near ROP therapy, two infants underwent laser photoablation, 31 received intravitreal bevacizumab, and over half were still on CPAP and receiving daily caffeine. Only two infants were still receiving daily caffeine at the time of hernia surgery, but four received a bolus dose of caffeine either in the operating room or on return to the NICU to prevent or treat postoperative apnea. About 20%of infants were still on CPAP and 40%on daily caffeine at the time of two-month immunizations and ROP exams. Ten infants received immunizations over two days instead of all at once.
On average, major BD events did not increase significantly after any of the interventions. In fact, in 354 of 493 (72%) interventions, infants had fewer or no BDs in the 24 hours after compared to before each procedure. Of 109 infants with zero BDs before the intervention, 75 (70%) continued to have zero BDs in the 24 hours following the intervention.
Figure 1 shows the mean and individual patient hourly BDs in the 24 hours before and after two-month immunizations (n = 146) and 34-week postmenstrual age ROP exams (n = 290). BD trends for ROP treatments and hernia surgeries are not shown due to the low number of events with wide hourly variability among patients. For immunizations, there was a trend toward a slight increase in hourly events the day before, followed by a decline. Near ROP exams, the BD events frequency did not change significantly in the 24 hour periods before and after.
Hourly BDs 24 hours before and after immunizations and ROP exams. Using an automated algorithm, we quantified hourly episodes of bradycardia-desaturation (BD), with bradycardia < 100 beats/minute for at least 4 seconds and desaturation SpO2 < 80%for at least 10 seconds. We calculated a centered, 6 hour rolling average of BDs for each infant in the 24 hour window before and after each procedure. The left panel shows the number of BD events (y-axis) before and after two-month immunizations, with the time from vaccination on the x-axis, for 146 infants. The right panel shows the same but with the time from the retinopathy of prematurity (ROP) exam closest to 34 weeks PMA on the x-axis for 290 infants. Bold lines indicate group means, while the thin gray lines show the hourly BD counts for each individual patient.
Figure 2 shows the difference in the number of events in the 24 hour period after versus before each intervention for individual infants. The plot shows a small number of outliers with many more BDs either the day before or the day after. For two-month immunizations and ROP exams, we compared the patient characteristics of infants with and without a substantial increase in BDs after the procedure, which we defined as a difference of five or more (Tables 2 and 3). For immunizations, infants with at least five more BDs after were more likely to be on CPAP (Table 2).
Change in number of 24 hour BD events after interventions. For each infant, the difference in the number of bradycardia-desaturations (BDs) in the 24 hours after versus before each intervention is shown for two-month immunizations (n = 146), retinopathy of prematurity (ROP) exam nearest 34 weeks’ post-menstrual age (PMA) (n = 290), ROP therapy (n = 33), and inguinal hernia surgery (n = 25). One outlier with BD difference < –35 after immunization is not shown. For hernia repair, two outliers with BD difference < –35 and one with BD difference > 35 are not shown.
Characteristics of infant with an increase of five or more bradycardia-desaturations (BDs) in the 24 hours after immunizations compared to before and those with the same or fewer BD events
Continuous variables are presented as mean (standard deviation) and categorical variables as number (percent). *p < 0.05.
Characteristics of infant with an increase of five or more BDs in the 24 hours after Retinopathy of prematurity (ROP) exams compared to before and those with the same or fewer bradycardia-desaturation (BD) events
Continuous variables are presented as mean (standard deviation) and categorical variables as number (percent).
Infants analyzed near the time of hernia repair surgery had a median of 11 (IQR 2 –19) BDs the 24 hours before and 6 (IQR 0 –13) BDs the day after, a decrease that did not reach statistical significance due to small sample size and wide variability. Of note, ten infants were extubated in the operating room and 15 in the NICU less than 24 hours after surgery. Only three of 25 infants (12%) had an increase of five or more BDs after surgery, while six had a decrease in events after surgery. One infant had discharge delayed due to receipt of caffeine in the operating room and had an extended monitoring period but did not have any BD events.
Infants with BDs analyzed near the time of ROP therapy had a median of 8 (IQR 5 –11) events before and 5 (IQR 3 –7) after the procedure. More infants had a decrease in BDs versus an increase (10 vs. 4 out of 33). The majority of infants (26 of 33, 78%) were on supplemental oxygen before and during the procedure and eight (25%) infants had a temporary increase in respiratory support in the 24 hours following the procedure. One infant with a BD event after ROP therapy had a delayed discharge after the apnea countdown was restarted.
Some infants were included in the analyses of multiple stressors: 149 infants were included in one, 106 in two, 40 in three, and 3 in all four analyses. We did not find trends to suggest that infants with an increase in BDs after one event would be more likely to have an increase after another event. For example, of the 139 infants included in both the immunization and the ROP exam analysis (the analyses with the largest overlap), only three had an increase of five or more BDs after both interventions, 113 infants had a decrease or stable number after both interventions, and 23 infants had an increase after one intervention but not the other.
Using a conservative event definition and an automated algorithm, we found that most of the preterm VLBW infants in our cohort did not have an increase in major cardiorespiratory events after two-month immunizations, eye exams, ROP therapy, and hernia surgery.
Certainly, some VLBW infants do have increased cardiorespiratory events following stressors, with some combination of apnea, heart rate deceleration, or oxygen desaturation. Control of breathing matures as preterm infants approach term corrected age, but immunizations have been reported to increase apnea in some infants [3–5, 23]. The immune response to vaccines involves increased prostaglandins, an inflammatory response that may precipitate apnea [6, 24]. Most recent studies have found that only a small fraction of infants have increased spells after immunizations, with wide variation in the definition of events and period studied. Nonetheless, some NICU providers remain hesitant to administer two-month immunizations on schedule to preterm infants due to concerns of cardiorespiratory instability [6, 8]. We chose to focus on what we consider to be clinically significant events with both low heart rate and prolonged and low SpO2 (< 80%for at least 10 seconds) and found that most infants had no increase in these events after two-month immunizations. The Advisory Committee on Immunization Practices and the American Academy of Pediatrics recommend giving immunizations to preterm infants according to the same schedule as term infants and our findings support the safety of this recommendation.
Therapeutic interventions for ROP can be stressful for infants, and even an ROP eye exam has been reported to be followed by an acute increase in cardiorespiratory events [10]. In contrast, we found that the mean number of BDs by our definition decreased in the 24 hours following ROP therapy or exam. However, 85 infants (29%) had at least five more BDs the day after compared to the day before an ROP exam performed close to 34 weeks PMA. We did not find any demographic or clinical variables that differentiated these infants with an increase of five or more BDs from those without. Our infants had at least one ROP exam before 34 weeks but we did not quantify BDs around this earlier exam. Concerning ROP therapy, we found that only four infants had at least five more events in the 24 hours after the procedure. Of note, all infants received atropine as a premedication to prevent vagal heart rate decelerations during the procedure. The anticholinergic effect typically wears off within six hours but may have limited post-procedure bradycardia events as intended.
The most significant stressor we studied, inguinal hernia repair under general anesthesia, was performed at a later PMA than the other three inter-ventions (38 weeks compared to 34–35 weeks), at which point apnea is less likely. Additionally, several infants received a bolus dose of caffeine to prevent postoperative apnea, and some infants remained on mechanical ventilation for several hours after return to the NICU to allow the general anesthetic effect to wear off. There is increasing interest in delaying surgical repair of inguinal hernias until after NICU discharge when the risk of post-operative apnea is lower (HIP Trial, ClicalTrials.gov #NCT01678638) [13, 25].
Strengths of our study include the use of an automated algorithm rather than medical record documentation, and the use of a rigorous definition of cardiorespiratory events, requiring both bradycardia and significant, prolonged desaturation. We analyzed a large number of infants near immunizations and ROP exams, but relatively few near ROP therapy and hernia surgeries. We acknowledge several other limitations of our analysis. First, we cannot account for caretaker interventions before or after stressors, such as changes in respiratory support, that may impact the number or severity of cardiorespiratory spells. In our unit, nurses caring for infants undergoing ROP therapy or hernia surgery may have a lighter patient assignment or may respond more quickly to the infants’ apneic spells and avert bradycardia or prolonged oxygen desaturation. Second, we cannot account for a transient increase in oxygen given during and following these procedures, which is not always well documented and may have prevented BD events. Finally, the analyses we performed used multiple thresholds: Bradycardia and desaturation depth and duration, the before and after time frame, and an increase of five or more BDs after an intervention. Therefore, we cannot rule out changes outside of these thresholds or in larger cohorts. For example, we might have found a difference in milder BD events, in desaturation events only, or in a longer post-intervention time frame. For infants near discharge on an apnea watch, an increase of even one event could have significant implications but, for most infants, a small increase could be due to chance and unrelated to the intervention.
Conclusions
We found that planned, potentially stressful interventions did not increase major cardiorespiratory events for most VLBW infants. In fact, the overall trend was a decrease in events after two-month immunizations, ROP exams, and ROP treatment. Importantly, many of the VLBW infants analyzed had significant BD events in the day before these interventions, at around 34–35 weeks PMA, and some did have an increase in BDs, which substantiates the need for cardiorespiratory monitoring for this population.
Footnotes
Acknowledgments
None.
Conflict of interest
The authors have no conflicts of interest to declare.
Ethics statement
This study was performed in accordance with the Declaration of Helsinki. The UVA Internal Review Board for Health Sciences Research approved the study with a waiver of informed consent (protocol # 12008).
Funding
K23 HD097254 (PI: B. Sullivan); R01 HD072071 (PI: K. Fairchild).