Predictors of Stress Exposure in Hospitalized Preterm Infants

Abstract

Background:

Stress exposure in the neonatal intensive care unit (NICU) is associated with poor outcomes in preterm infants. However, factors predicting subsequent NICU stress exposure have not been identified.

Purpose:

To characterize NICU stressors experienced by preterm infants during the first two weeks of life and identify demographic, perinatal, and institutional variables associated with stress exposure.

Methods:

A secondary analysis of data from a non-experimental, prospective study was conducted using data from 60 very preterm infants born 28- to 31-weeks gestational age. Stress exposures during the first two weeks of life, operationalized as number of invasive procedures, were characterized by type and quantity for each infant using data extracted from electronic health records. Associations between number of invasive procedures and demographic, perinatal, or institutional variables were analyzed using linear regressions with robust standard errors.

Results:

Preterm infants experienced, on average, 98 (SD=41.8) invasive procedures. Of these invasive procedures, nasal and/or oral suctioning episodes (58.1%) followed by skin-breaking procedures (32.6%) were most frequent. Differences in the number of invasive procedures were found for maternal race; infants born to Black mothers experienced fewer total invasive procedures than infants born to White mothers. Number of invasive procedures also varied across NICUs.

Implications for Practice and Research:

Preterm infant stress exposure differed by maternal race and NICU, consistent with research findings of differential treatment of diverse infants. Further research is needed to understand the reasons for these differences and to identify best practices to standardize neonatal care.

Background and Significance

Very preterm infants (i.e. infants born before 32-weeks gestational age [GA]) experience significant stress exposure during their hospitalization in the neonatal intensive care unit (NICU). Separation from parents, noxious environmental exposures, medical procedures, and nursing care contribute to the preterm infant’s cumulative stress burden. Cumulative stress burden during this sensitive developmental stage can affect brain development^1–3 cognitive performance^1,4, motor development¹, and behavioral outcomes⁵. The detrimental effects of stress exposure in preterm infants persist through early childhood; in former preterm infants, high neonatal stress exposure was directly related to abnormal white matter microstructure⁴, lower intelligence scores⁴, and behavioral problems⁵ at school-age. Given the associations between early stress exposure and neurodevelopment, research to provide a better understanding of the nature of preterm infant stress exposure is necessary to develop targeted interventions to support healthy development. Thus, the purpose of this analysis was to increase understanding of the stress exposures experienced by preterm infants during the first two weeks of life, a reportedly sensitive period during which stress exposure has particularly profound effects on early neurobehavior^6,7.

Literature Review

Sensitive periods of brain development are time periods during which neurologically expected experiences shape the normal formation of neuronal circuits, but unexpected or adverse experiences can have especially lasting influences on brain circuitry and neurodevelopment⁸. Although few researchers have investigated associations between the timing of stress exposures and neurodevelopment, sensitive periods during which NICU stress exposures have a more profound effect on outcomes have been identified. Stress exposures occurring early during preterm infants’ NICU hospitalizations have more profound, negative effects on thalamic development and subsequent neurodevelopment at three years corrected age than those occurring later during hospitalization, especially among extremely preterm infants⁹. Among preterm infants born between 28- and 33-weeks GA, higher stress exposure during the first 12 days of life was associated with less mature neurobehavior at 36- to 38-weeks postmenstrual age compared to infants with lower stress exposure⁶. Similarly, cortisol levels measured during the first week of life, indicative of NICU stress exposure during that time, correlate with lower attention in preterm infants prior to NICU discharge⁷. The first postnatal weeks also coincide with the greatest stress exposure for preterm infants, as the number of highly intrusive stressors (e.g. skin-breaking procedures) decreases by 80% after the first two weeks of life⁶. Thus, the preterm infant’s stress burden may be highest during the early postnatal period, which may also represent a sensitive period for the effect of stress exposure on early neurobehavior.

Little is known about the predictors of neonatal stress exposure in preterm infants. However, one type of stress exposure – painful procedures – has received some attention. For example, the number of painful procedures experienced by neonates is negatively associated with GA at birth^10–12 and birthweight^11,12; infants born at the youngest GAs and lowest birthweights experience the highest levels of procedural pain. Further, infants requiring mechanical ventilation, positive airway pressure, or supplemental oxygen experience higher numbers of painful procedures than those requiring no respiratory support¹¹. While the use of either pharmacological or nonpharmacological analgesia during painful procedures is higher for preterm infants compared to term-born infants, sicker infants and those receiving respiratory support are less likely to receive pain management interventions¹⁰. Moreover, although the use of analgesics during painful procedures differs by specific NICU¹⁰, differences in stress exposures across NICUs have not been described.

Researchers operationalize stress exposure using a variety of measures. While counts of skin-breaking procedures is the most common operational definition of stress exposure¹³, the specific procedures included as “skin-breaking” in individual studies vary and often include invasive procedures that do not actually break the skin^3,14,15. For this study, stress exposure is defined as the number of invasive procedures – that is, all procedures involving the entrance of a foreign object into the body. The specific aims were to (1) describe, by frequency and type, the invasive procedures experienced by preterm infants during the first two weeks of life, and (2) identify the important demographic, perinatal, and institutional variables associated with invasive procedures.

Methods

Study Design

Data for this secondary analysis were collected as part of a larger non-experimental, longitudinal study (hereafter referred to as the parent study) to determine the relationships among stress exposure, inflammation, and early neurobehavior in very preterm infants¹⁶. Data collected from the electronic health record (EHR) and by maternal report were used for this analysis. Demographic, perinatal, and institutional data were collected during the parent study and included as predictors in this secondary analysis.

Sample/Setting

Seventy-three infants born 28- to 31-weeks GA were enrolled in the parent study. Eligible infants were admitted or transferred into one of four Level III/IV NICUs within a single neonatal system in a large Midwest city. The Level IV NICU (NICU A) is an all-referral 90-bed unit within an academic children’s hospital with private patient rooms plus an open bay. The Level III NICUs are in general service hospitals with labor and delivery units. NICU B, located in an academic medical center, has 49 beds in an open bay. NICU C, located in a regional hospital, consists of 42 beds in private rooms with a few double rooms that are typically reserved for twins. NICU D, located in a regional hospital, consists of 23 beds primarily in an open bay with a few private rooms reserved for infants requiring special isolation. Infants were excluded from the parent study if they (1) experienced perinatal infection, (2) exhibited signs of neonatal abstinence syndrome or were born to mothers with known opioid drug abuse, (3) were born with congenital or chromosomal abnormalities, (4) were diagnosed with neurologic abnormalities (e.g. Grade III/IV intraventricular hemorrhage) known to affect neurodevelopment, (5) were born to mothers with oligohydramnios, or (6) were born small for gestational age (i.e. birthweight less than 10^th percentile)¹⁶. This analysis was performed using data from the 60 infants in the parent study whose mothers consented to secondary use of their infants’ data.

Measures

Invasive Procedures

Data for the number and type of invasive procedures experienced by infants during the first two weeks of life were retrospectively extracted from the EHR. Counts of invasive procedures were determined for each infant and included, but were not limited to, endotracheal suctioning, oral and/or nasal suctioning, insertion of feeding tubes, insertion of urinary catheters, intubation, umbilical intravenous catheter insertion, and skin-breaking procedures, defined as procedures resulting in compromised skin integrity (e.g. intravenous catheter insertion, heel lance, arterial or venous puncture, and intramuscular or subcutaneous injection). In a previous analysis of data from the parent study that aimed to determine the relationship between different measures of stress exposure and neurobehavior, suctioning was not included as an invasive procedure¹⁷. However, in this analysis, endotracheal and nasal and/or oral suctioning were added as invasive procedures because recent studies have revealed that these procedures are associated with behavioral or physiologic stress responses in preterm infants^18,19.

Demographic and Perinatal Clinical Characteristics

The following demographic data were collected from the EHR and included in this analysis: infant sex, insurance type, and admitting NICU. Mothers reported additional demographic characteristics using an investigator-developed questionnaire that included maternal race based on the racial categories specified by the National Institutes of Health, maternal marital status, maternal education, and maternal age. Perinatal clinical data were extracted from the EHR and included exposure to prenatal steroids, antibiotics, or tobacco; maternal pregnancy complications; infant GA at birth; and infant birthweight.

Statistical Analysis

Measures of central tendency and variance were used to describe the occurrence of invasive procedures. Using demographic, perinatal, and institutional variables as predictors, univariate associations with number of invasive procedures were identified using linear regression models with robust standard errors to account for possible misspecification of the variance. The following dichotomous or categorical predictors were analyzed: maternal race (White, Black), infant sex (male, female), insurance type (private, public/none), maternal marital status (single or not cohabitating, married or cohabitating), maternal education (high school or less, some college, 4-year college, some graduate school), exposure to prenatal steroids (yes, no), exposure to prenatal antibiotics (yes, no), exposure to prenatal tobacco (yes, no), admitting NICU (A, B, C, or D), maternal diabetes (yes, no), maternal hypertension (yes, no), and maternal preeclampsia (yes, no). Infant GA at birth, infant birthweight, and maternal age were included as continuous variables. Statistical significance was set at α< .05, with Bonferroni-Holm adjustments made to control for multiple comparisons after 16 individual tests. The analysis described above was repeated using specific procedures (e.g. heel lance) that had a median overall count of at least 15 to identify associations between these procedures and the demographic, clinical, and institutional predictors. Stata (version 14) was used to conduct all statistical analyses.

Procedure

Infants were enrolled in the parent study during the first two weeks of life. Parents of infants eligible for the parent study were approached after the infant’s second day of life for potential enrollment. After a full discussion of study procedures, risks, and benefits, mothers provided written informed consent for their infant’s participation. The hospital’s Institutional Review Board reviewed and approved the study prior to participant enrollment. Data were included in the current analysis if mothers of infants participating in the parent study provided written informed consent for secondary analysis (N=60). This secondary analysis was approved by the hospital’s Institutional Review Board by expedited review. Demographic and clinical data were collected at the time of infant enrollment and throughout hospitalization. Mothers of enrolled infants completed a demographic questionnaire at the time of infant enrollment. Prenatal data were extracted from the infant’s EHR at the time of enrollment and included maternal medications and pregnancy complications. Infant clinical data, including type and number of invasive procedures, were similarly extracted from the infant’s EHR at the time of enrollment and throughout hospitalization.

Results

The majority of infants providing data for this analysis (N=60) were male (n=40) and born to White mothers (n=52). Infants were born at a mean GA of 30 weeks (SD = 1.2) and birthweight of 1460 grams (SD = 336). Only one mother identified as Hispanic; thus, ethnicity was not included as a predictor in the analysis. Demographic, perinatal, and institutional variables included as predictors in the analysis are presented in Table 1.

Table 1.

Demographic, Perinatal, and Institutional Variables (N = 60)

Variable	n (%), Mean (SD)b
Demographic Variables
Male sex	40 (66.7)
Private insurance (n=57)	35 (61.4)
Maternal race
White	52 (86.7)
Black	8 (13.3)
Maternal marital status  Single  Partnered, not cohabitating  Partnered, cohabitating  Married  Unknown	6 (10) 2 (3.3) 12 (20) 39 (65) 1 (1.7)
Maternal education  High school or less  Some college  4-year college  Some graduate school  Unknown	17 (28.3) 11 (18.3) 17 (28.3) 5 (8.3) 10 (16.7)
Maternal age	29.1 (7.1)b
Perinatal Variables
Infant GA at birth (weeks)	30 (1.2)b
Infant birthweight (grams)	1460.1 (336)b
Prenatal steroids  None  1–2 doses  3 or more doses  Unknown	6 (10) 42 (70) 10 (16.7) 2 (3.3)
Prenatal antibiotics (n=58)	33 (56.9)
Prenatal tobacco	9 (15)
Pregnancy complications  Diabetes (n=59)  Any hypertension  Preeclampsia	7 (11.9) 19 (31.7) 16 (26.7)
Institutional Variables
Admitting NICU (n=59)a  NICU A – Level IV  NICU B – Level III  NICU C – Level III  NICU D – Level III	10 (16.9) 18 (30.5) 23 (39) 8 (13.6)

Abbreviations: SD, standard deviation; GA, gestational age; NICU, neonatal intensive care unit.

^aone infant was transferred from a non-participating hospital and was not included in the analysis of NICU.

^bMean (SD).

Frequency of Invasive Procedures

During the first two weeks of life, infants experienced a mean of 98 (SD = 41.8) invasive procedures. There was high variance among infants with infants experiencing between 19 and 195 invasive procedures during the first two weeks of life. Oral and/or nasal suctioning episodes were the most frequent invasive procedures (M=63.4, SD=36), followed by skin-breaking procedures (M=25.8, SD=13.9; Table 2). Of the skin-breaking procedures, a specific subset of invasive procedures, heel lances were most common (M=17.4, SD=11.1). All infants experienced at least one skin-breaking procedure and one feeding tube insertion. Additionally, at least 50% of infants experienced at least one episode of nasal and/or oral suctioning and one intubation attempt (Table 2).

Table 2.

Invasive Procedures Experienced by Infants (N=60)

Procedure	Mean Count (SD)	Min – Max	Median	Infants (%)
Oral or nasal suctioning	63.4 (36)	0 – 122	74	59 (98.3)
Skin-breaking procedures	25.8 (13.9)	11 – 105	24	60 (100)
Heel lance	17.4 (11.1)	7 – 86	15	60 (100)
Intravenous catheter insertion	5 (4.9)	0 – 19	4	48 (80)
Arterial puncture	1.8 (1.7)	0 – 7	1	46 (76.7)
Intramuscular injection	1 (0.1)	1 – 2	1	60 (100)
PICC insertion	0.3 (0.6)	0 – 2	0	14 (23.3)
Intradermal injection	0.08 (0.6)	0 – 5	0	1 (1.7)
Venipuncture	0.07 (0.4)	0 – 3	0	2 (3.3)
Lumbar puncture	0.05 (0.3)	0 – 2	0	2 (3.3)
Subcutaneous injection	0.02 (0.1)	0 – 1	0	1 (1.7)
Feeding tube insertion	4.1 (1.6)	2 – 8	4	60 (100)
Endotracheal suctioning	2.4 (11.2)	0 – 80	0	8 (13.3)
Intubation	1.3 (2)	0 – 10	1	32 (53.3)
Umbilical catheter insertion	0.7 (0.8)	0 – 4	0	29 (48.3)
Rectal medication/tube insertion	0.3 (2.2)	0 – 17	0	3 (5)
Urinary catheter insertion	0.03 (0.2)	0 – 1	0	2 (3.3)

Abbreviations: SD, standard deviation, Min, minimum; Max, maximum; PICC, peripherally inserted central catheter

Associations with Demographic, Perinatal, and Institutional Variables

Among demographic variables, there were differences in the number of invasive procedures by maternal race, maternal education, and maternal age. Infants born to Black mothers experienced, on average, 53.7 (95% CI [23.2, 84.1], p=.001) fewer invasive procedures than infants born to White mothers. Maternal education was also associated with number of invasive procedures with infants born to more highly educated mothers experiencing fewer invasive procedures (F(3,46) = 3.22, p=.03), with the largest difference observed between infants whose mothers had a high school education or less and those born to mothers with a 4-year college degree. Finally, infants experienced, on average, 1.6 (95% CI [0.3, 3], p=.02) fewer procedures for each one-year increase in maternal age. There were no associations between number of invasive procedures and infant sex, insurance type, or maternal marital status. After adjusting for multiple comparisons, only maternal race remained significantly associated with the number of invasive procedures. To identify differences between infants born to Black mothers and infants born to White mothers that might account for differences in invasive procedures, associations between maternal race and illness severity were analyzed using infant GA at birth and infant birthweight as proxies for illness severity. Linear regressions with robust standard errors revealed no differences in GA at birth (p=.39) or birthweight (p=.84) between infants born to Black mothers and infants born to White mothers.

Among perinatal predictors, maternal diabetes, infant GA at birth, and infant birthweight were associated with number of invasive procedures. Infants born to mothers with diabetes experienced, on average, 42.3 (95% CI [10.3, 74.3], p=.01) fewer invasive procedures than infants born to mothers without diabetes. Additionally, for each one-day increase in infant GA at birth or 100-gram increase in birthweight, infants experienced, on average, 1.5 (95% CI [0.3, 2.8], p=.02) or 3.2 (95% CI [0.8, 5.7], p=.01) fewer invasive procedures, respectively. However, none of these associations were significant after adjusting for multiple comparisons. There were no relationships between number of invasive procedures and maternal hypertension, maternal preeclampsia, or exposure to prenatal steroids, antibiotics, or tobacco.

Finally, the association between admitting NICU and number of invasive procedures was analyzed. There were significant differences in the number of invasive procedures (F(3,55) = 20.8, p<.0005) among the four NICUs included in the analysis. The association remained significant after adjustment for multiple comparisons. To identify differences in the patient populations that might account for differences in invasive procedures among the NICUs, associations between NICU and maternal race and illness severity using infant GA at birth and infant birthweight as proxies for illness severity were analyzed. Fisher’s exact test revealed no associations between NICU and maternal race (p=.26). Linear regressions with robust standard errors revealed no associations between NICU and infant GA at birth (p=.19) or infant birthweight (p=.53).

Further Analysis of Maternal Race and Admitting NICU

Differences by maternal race and admitting NICU were examined for procedures with a median overall number of at least 15 (i.e. oral and/or nasal suctioning, total skin-breaking procedures, heel lance). Linear regression models revealed differences in oral and/or nasal suctioning and heel lances between infants born to White and Black mothers but no difference in total skin-breaking procedures (Figure 1). Infants born to Black mothers experienced, on average, 44.3 (95% CI [16.9, 71.8], p=.002) fewer episodes of oral and/or nasal suctioning and 4.8 (95% CI [0.9, 8.7], p=.02) fewer heel lances than infants born to White mothers. Additionally, there were differences among admitting NICUs in nasal and/or oral suctioning (F(3, 55) = 77.1, p<.0005) and the total number of skin-breaking procedures (F(3, 55) = 7.1, p=.0004) but not heel lances (Figure 2).

Figure 1: Differences in number of invasive procedures by maternal race^a (N = 60).

^a Invasive procedures with a median count of at least 15 were included in the analysis: oral/nasal suctioning; total skin-breaking procedures; and heel lances, a subset of skin-breaking procedures.

Figure 2: Differences in number of invasive procedures by admitting NICU^a (N = 59).

NICU, neonatal intensive care unit; A=Level IV all-referral NICU in an academic children’s hospital with both private rooms and open bay; B=Level III NICU in an academic hospital with open bays; C=Level III NICU in a regional hospital with mostly private rooms; D=Level III NICU in a regional hospital with open bays. ^a Invasive procedures with a median count of at least 15 were included in the analysis: oral/nasal suctioning; total skin-breaking procedures; and heel lances, a subset of skin-breaking procedures.

Discussion

In this secondary analysis, preterm infants experienced many invasive procedures during the first two weeks of life in the NICU. High stress exposure among preterm infants, operationalized as painful procedures, have been reported by other researchers. In a systematic review including 3,156 neonates (preterm and term-born infants within the first 28 days of life), Cruz et al. found that neonates experienced an average of 7.5–17.3 painful procedures per day of hospitalization and that heel lance, suctioning, and venipuncture were the most common procedures¹¹. Other researchers report that preterm and term-born neonates experience a mean of 98 (SD=78) painful procedures during the first two weeks of life and that the most common painful procedures are suctioning, heel lance, and adhesive removal¹⁰. Although the number of procedures experienced by infants in the current study were similar to those in earlier reports, Carbajal et al. included infants born between 24- and 42-weeks GA and operationalized painful procedures broadly to include 44 different procedures such as skin-breaking procedures, invasive procedures, chest physiotherapy, adhesive removal, and others¹⁰. The differences in operational definitions of stress exposure among studies underscore the importance of specifying included procedures so that comparisons can be made across studies and effects on infant outcomes can be more clearly delineated in future research. Importantly, “acceptable” stress exposure thresholds, above which specific adverse neurodevelopmental outcomes are more likely, have not been identified but may be important for improving future clinical practice.

This analysis also revealed associations between the number of invasive procedures experienced by preterm infants during the first two weeks of life and maternal race and admitting NICU. Differences in NICU structure, process, and outcome measures have been reported by researchers²⁰, with complex racial and/or ethnic disparities noted, most often disadvantaging infants of color, especially Black infants. Infants born to Black mothers^21–23 or from disadvantaged backgrounds (e.g. low-income, low-education)^24–26 experience poorer clinical and neurodevelopmental outcomes compared to infants born to White mothers or from higher socioeconomic backgrounds. However, the relationships among infant race, caregiving, and outcomes are complex. As an example, in the post-surfactant years, respiratory distress survival, which was lower in White infants prior to widespread surfactant therapy, is now lower in Black infants, who are less likely to receive surfactant²⁰, suggesting interactions among race and caregiving practices that affect outcomes. Importantly, it is not known if differences in care, as measured by invasive procedures, represent a disadvantage for infants born to Black mothers. Theoretically, fewer procedures would reduce the stress burden experienced by the infant. However, if these procedures were needed because of particular health conditions and not provided, that would be a matter of extreme concern and a certain health disparity. Unfortunately, necessary data were not available to ascertain if needed procedures were not done.

Disparities in quality of care across NICUs have been reported and may contribute to the persistence of racial and/or ethnic disparities²⁷. In a cohort study of 117,982 very low birthweight and very preterm infants, researchers found that NICUs were segregated by race and ethnicity²⁸. Compared to White infants, Black infants were concentrated at lower-quality NICUs and Hispanic and Asian infants at higher-quality NICUs; NICU quality was defined by hospital-level measures of infant outcomes and adherence to established standards of care²⁸. Further, NICUs with the highest proportions of Black infants had higher nurse to infant ratios; due to workload, nurses working in these NICUs reported rationing care more frequently than nurses working in NICUs with lower nurse to infant ratios and a comparatively lower proportion of Black infants²⁹.

While the NICUs included in this analysis did not differ significantly in racial demographics, there were differences in numbers of invasive procedures across NICUs. All four NICUs in the study are part of the same care delivery system with identical policies for nursing care that should encourage standardized nursing practice. However, the culture of care may vary among NICUs, resulting in differences in nursing routines, workflows, and training that are not reflective of policies and may, ultimately, affect the care provided to patients. Interestingly, despite the physiologic instability that can occur with suctioning³⁰, oral and/or nasal suctioning comprised the majority of invasive procedures in our study, similar to other epidemiological studies of painful procedures in preterm infants¹¹, and differed in frequency across NICUs. Some policies for caregiving may be flexible or ambiguous, resulting in decision-making by individual clinicians based on experience, training, education, or unit culture. For example, Mann et al. found that 96% of NICU nurses and respiratory therapists used respiratory assessment to determine the need for nasopharyngeal suctioning, but 62% of these clinicians reported routinely providing nasopharyngeal suctioning with wide variations in time intervals between suctioning episodes³¹.

There are limitations to this analysis, which included a small sample with limited racial and ethnic diversity and multiple potential confounders. Additionally, because this was a secondary analysis, analyses were limited to the data that had already been collected. Thus, data pertaining to the need for various procedures or demographic data for the nurses were not available. Further, there may be important predictors of stress exposure, including a multitude of maternal and infant factors, that could not be analyzed based on inclusion and exclusion criteria for the parent study or because those data were not collected. Second, the NICUs included in this analysis are all part of a single neonatal system, which may result in underestimation of the differences among NICU practices in general. Finally, some of the procedures occurred infrequently and could not be further analyzed. More research is needed to identify best practices around stressful procedures in the NICU.

Conclusion

Preterm infants experience many invasive procedures during the early postnatal period that may affect long-term neurodevelopment. The number of invasive procedures experienced by infants may be related to non-clinical factors such as maternal race and admitting NICU, although the reasons underlying these differences are unknown and, likely, complex. Further research is needed to determine the reasons for differences in the number of invasive procedures experienced by preterm infants, develop best practices for the routine care of preterm infants, and decrease these exposures or buffer their effects.

Acronyms and Abbreviations:

GA

Gestational Age

NICU

Neonatal Intensive Care Unit

EHR

Electronic Health Record