Elevated maternal anxiety in the NICU predicts worse fine motor outcome in VLBW infants

Abstract

Aim:

The literature is sparse with respect to the impact of elevated maternal psychological distress in the neonatal intensive care unit (NICU) on later neurodevelopmental outcome in VLBW infants. The objective of this study is to examine the impact of elevated maternal distress, defined as elevated depression, anxiety and perinatal-specific post-traumatic stress, during the NICU hospitalization on VLBW infant ND outcome at 20 months corrected age (CA).

Methods:

This was a prospective study of 69 mothers and their VLBW infants recruited in 2011–2012. Elevated maternal distress was collected by maternal questionnaire in the NICU. Elevated depression was quantified with the Center for Epidemiological Studies-Depression Scale, anxiety with the Spielberger State-Trait Anxiety Inventory and perinatal-specific post-traumatic stress with the Modified Perinatal Posttraumatic Stress Disorder Questionnaire. VLBW infant ND outcome was assessed using the Bayley Scales of Infant & Toddler Development-III at 20 months CA. Regression analyses determined the impact of elevated distress on ND outcome after adjusting for infant medical and maternal sociodemographic variables.

Results:

After controlling for infant and maternal covariates, elevated maternal anxiety in the NICU predicted lower fine motor scores at 20 months CA.

Conclusion:

Elevated maternal anxiety in the NICU is associated with adverse ND outcome in VLBW infants in the 2nd year of life. NICU-based support services may help mothers’ quality of life and VLBW infant outcome.

1. Introduction

Mothers of very low birth weight (VLBW, birth weight < 1500 g) infants hospitalized in the neonatal intensive care unit (NICU) experience acute emotional vulnerability stemming from the unexpected and often traumatic nature of the preterm birth [1], and immediate separation from their critically ill infant. The stress of parenting a critically ill infant during the two to three month NICU hospitalization impacts mothers’ emotional well-being as they face uncertainty about their VLBW infants’ increased risks for life-long health and neurodevelopmental (ND) difficulties [2]. Relative to mothers of full-term infants in the first three months of the post-partum period,¹ mothers of VLBW infants hospitalized in the NICU demonstrate disproportionately higher rates of elevated maternal psychological distress, typically defined as elevated symptoms of depression, anxiety and perinatal-specific post- traumatic stress (PPTS) [3–8]. Fourteen to 63% of mothers of VLBW infants hospitalized in the NICU endorse elevated rates of depression, 43–55% endorse elevated rates of anxiety, and 24–25% endorse elevated PPTS. Rates for all three aspects of psychological distress for mothers of full-term infants in the first three months of the post-partum period are significantly lower, with elevated depression endorsed by 10–15%, elevated anxiety endorsed by 43–55% and elevated PPTS by only 3% of mothers.

Although little is known about the effect of elevated maternal distress in the NICU on VLBW infants’ ND outcome in early childhood, the negative effect of elevated post-partum depression and, to a lesser extent, PPTS and anxiety, on ND outcome for children who were born full-term is well established [9–13]. Elevated levels of all three aspects of distress are hypothesized to impact full-term children’s ND outcome by altering mother-child relationships and mothers’ ability to engage in sensitive parenting, defined as parenting characterized by emotionally and developmentally synchronized interactions [14–17]. More specifically, altered relationships and insensitive parenting negatively impacts mothers’ ability to help their child regulate attention and arousal, participate in contingent play, and engage in scaffolded, structured play, all of which promote ND outcome [11,14–17]. The first three months of the post-partum period is recognized as a critical time for establishing mother-child relationships. Thus, elevated maternal distress occurring within this time frame has the potential to have a stronger negative influence on both the developing mother-child relationship and the child’s ND outcome than other periods of development [17,18].

The critical first three-months of the post-partum period coincides with the average length of NICU hospitalization for VLBW infants. It is also the time period in which mothers of VLBW infants experience the highest rates of elevated psychological distress [3–8]. Furthermore, for VLBW infants this same time period also coincides with gestational weeks (24–40) uniquely associated with rapid pre-oligodendroglial, microglial, axonal, subplate neuronal, cortical and cerebellar growth and “enhanced vulnerability” of the brain [19]. To our knowledge, in the past 15 years [20] only Zelkowitz et al. [21] has assessed elevated maternal psychological distress during VLBW infants’ NICU hospitalization. This seminal study revealed that elevated maternal anxiety during the NICU period significantly predicted lower composite cognitive and language score in infants at 20 months corrected age (defined as age adjusted for prematurity, CA). Zelkowitz et al. [21] only examined the association between one psychological distress variable (e.g., anxiety) during the NICU hospitalization and ND outcome, not all three variables (e.g., depression and PPTS also).

Given that mother-VLBW infant dyads are at disproportionately higher risk for both elevated maternal psychological distress during the NICU hospitalization and infant ND delay [2], understanding the associations between elevated psychological distress in mothers and their VLBW infants’ ND outcome is of great importance. The purpose of this study is to identify the associations between elevated maternal depression, anxiety and PPTS at two time points during the NICU hospitalization (within the first month of the NICU hospitalization and 2 weeks prior to NICU discharge) on ND outcome in the 2nd year of life in a contemporary cohort of VLBW infants. Based on the existing literature of full-term and VLBW infants, we hypothesize that elevated maternal depression, anxiety and PPTS at both NICU time points will be associated with lower scores on ND measures for VLBW infants.

2. Methods

This is a longitudinal, mixed-method (quantitative and qualitative) study of 69 mothers and their VLBW infants recruited from a 57 bed, level IV NICU in an urban academic center, in the United States (US), from 2011 to 2012. Qualitative results have been previously published [22,23]. Eligible mothers were identified and recruited based on their participation in a larger, NIH-funded cohort study. Additional details of the NIH-funded cohort study have been previously published [24]. This sub-study of 69 mothers, identified and recruited from the larger NIH cohort study, examined associations among a series of NICU variables, namely infant medical risk factors, maternal sociodemographic variables, maternal psychological distress levels, maternal visitation rates, and human milk exposure/dose, and, post-NICU discharge VLBW infant ND outcome and maternal psychological distress. The present study reports on NICU-based maternal and infant variables, NICU-based maternal psychological distress, and post-NICU discharge VLBW infant ND outcome.

2.1. Participants

Mothers of VLBW infants were recruited to participate if the mothers were enrolled in the aforementioned larger, NIH-funded cohort study English-speaking, ≥18 years of age, and the infants were deemed likely to survive by the attending neonatologist. Of 100 eligible dyads, 72 mothers initially signed consent, 17 refused and/or were not approached given difficulty contacting the mothers, and 11 had immediate plans to transport their infants from the NICU to a hospital closer to the mother’s home. Of the 72 mothers, 69 mothers (3 mothers lost due to sudden transfer/infant death) completed the first wave of maternal distress questionnaires (Time 1, T1) and 64 completed the second wave of distress questionnaires (Time T2). The relevant institutional review board (IRB) approved this study. Written informed consent was obtained for all participants. Given privacy requirements of the IRB, no data were available for mothers who did not sign consent for participation.

2.2. Design

Infant and maternal data from the NICU hospitalization were obtained from collaboration with the larger NIH-funded cohort study via medical chart review conducted upon VLBW infant NICU discharge. During the NICU hospitalization, maternal self-report distress questionnaires were administered 2–4 weeks after birth (T1 mean28.1 days). Questionnaires were repeated on one occasion, approximately 2 weeks prior to discharge (T2 mean 14.8 days prior to discharge). Variability existed for the interim days between T1 and T2 (Mean = 50.6, Median = 49, SD = 26.23, range = 7–101). Mothers with extreme T1 and T2 interim day values (more or < 1.5 SD from the mean) did not have extreme, or outlier (more or < 1.5SD from mean) distress values. Post-NICU assessment of VLBW ND outcome occurred at routine clinical visits to the Neonatal High Risk Follow-up Clinic, a multidisciplinary clinic that monitors growth, neurologic, and neurodevelopmental status of infants cared for in the NICU.

2.3. Measures

2.3.1. Infant medical variables

The following were collected as a part of the larger NIH project and were accessed by this study: infant birthweight (BW) in grams, gestational age (GA) in weeks, sex, length of time on a mechanical ventilator in the NICU (days), presence of late-onset sepsis (presumed or confirmed), presence of necrotizing enterocolitis (stage 2 or 3) [25], presence of severely abnormal head ultrasound (defined as presence of intraventricular hemorrhage grade III or IV, periventricular leukomalacia, or hydrocephalus requiring shunt), presence of any abnormal head ultrasound (defined as presence of intraventricular hemorrhage grades I–IV, periventricular leukomalacia, or hydrocephalus requiring shunt), average daily dose of human milk (HM) received by infant, receiving any HM at time of NICU discharge, length of hospitalization (days) were gathered from the extensive prospective NIH database.

2.3.2. Maternal sociodemographic, reproductive, psychological history variables

The following were collected as a part of the larger NIH project and were accessed by this study: maternal age, education level, unemployment status, race/ethnicity, relationship status, insurance status, eligibility for Women, Infant and Children (WIC) program (a government subsidized program for families with incomes below state poverty standards), primipara status, history of fetal loss and history of prior preterm delivery.

Psychological history was collected via self-report and medical chart review. Information about mothers’ history of psychological diagnosis prior to birth of a preterm infant, history of involvement with the state’s Child Protective Services, and history of drug abuse during pregnancy was collected from medical chart review of psychosocial evaluation/history conducted by NICU social workers. None of the mothers were enrolled in therapy immediately prior to delivering their VLBW infant. Access to social support also was identified as part of demographic data collection by asking mothers to identify the total number of caretakers available for the infant, including the possibility of family and friends beyond child’s other parent.

2.3.3. Maternal psychological distress measures

The Center for Epidemiological Studies-Depression Scale (CES-D) [26] is a 20-item measure that assesses the frequency of depressive symptoms. The CES-D has moderate test-retest reliability over 6–8 weeks (r_xx = 0.59) and has been commonly used in this study’s population with good internal reliability (α = 0.81–0.90)[4,6]. CESD clinical cutoff scores used by previous investigations (CESD > 16) [4,6] were used to define elevated depressive symptomatology.

The Spielberger State-Trait Anxiety Inventory (STAI) [27], is comprised of a 20-item subscale of “state” or acute/situational anxiety and 20-item subscale of “trait” or generalized subscale. The STAI has demonstrated test-retest reliability (r_xx) of 0.65–0.75 over 8 weeks, good internal consistency among mothers of VLBW infants (α = 0.89) [4] and is frequently used with this population [4,6,21]. STAI-State cutoff score used by previous investigations (State scores > 40) [6] was used to define elevated anxious symptomatology. State and trait subscales for this sample were high correlated (r > 0.80, p < 0.0001), which indicates high correlation between anxiety state and trait for this sample of mothers.

The Modified Perinatal Posttraumatic Stress Disorder Questionnaires (PPQ-M) [28,29] is a 14-item questionnaire designed to measure frequency of PPTS that is well-validated in the NICU [28,29] with good test-retest reliability over 2–4 weeks (r_xx = 0.92) [28, 29] and good internal consistency (α = 0.90). The PPQ-M’s clinical cutoff score of 19 was used to define elevated perinatal-specific PTSD symptomatology[29].

2.3.4. Relevant psychological distress covariates

The Life Events Checklist (LEC) [30] is a 17-item scale that quantifies cumulative lifetime exposure to potentially traumatic events prior to the birth of a preterm infant. The LEC is well-cited, has temporal stability ranging from 0.84 to 0.37 and correlates well to other measures of post-traumatic stress [30].

2.3.5. Neurodevelopmental outcome

Infant ND outcome at 20 months CA was assessed by the normed, standardized Third Edition of the Bayley Scales of Infant and Toddler Development (Bayley-III) [31]. The scales quantify neurodevelopment in youth from 1 to 42 months of age and are considered the “gold standard” for ND assessment of children born preterm. The scales are comprised of the Cognitive, Language and Motor Indices (raw scores converted to standard scores, M = 100, SD = 3) and the Cognitive, Receptive Language, Expressive Language, Fine Motor and Gross Motor Subscales (raw scores converted to scaled scores, M = 10, SD = 3).

In addition to the aforementioned continuous Bayley-III scores, frequency of subnormal Bayley-III scores, defined as standard or scaled scores > 1 SD below mean, and severely abnormal Bayley-III scores, defined as standard or scaled scores > 2 SD below mean were calculated. However, Bayley-III scores quantified as subnormal or severely abnormal were not included as potential ND outcome variables in statistical analyses. The overall sample and prevalence of subnormal or severely abnormal scores was determined to be too low to provide statistically powerful, stable statistical results. Statistical analyses with categorical or dichotomous outcome variables, such as logistic regression, are known to require larger sample sizes to achieve statistical power and stable statistical results than statistical analyses with continuous outcome variables [32]. As such, only continuous Bayley Indices and Subscale scores were included as potential ND outcome variables in analyses.

2.4. Statistical analysis

Measures of central tendency and frequencies were used to provide descriptive statistics of infant medical, maternal sociodemographic, reproductive and psychological history, elevated maternal psychological distress, and infant ND outcome variables. Bivariate correlation analyses determined which variables were correlated with attendance at 20 months CA.

Bivariate correlational analyses determined which elevated maternal psychological distress variables were associated with Bayley-III scores at 20 months CA. Identified distress and ND associations were examined in regression models. Infant and maternal variables correlated at p < 0.10 with identified maternal psychological distress variables or identified Bayley-III scores were chosen as regression covariates. A p-value of 0.10 was used as a cut-off score versus a p-value of 0.05 in order to balance the studies relatively small sample size and the ability to capture more potential variables of clinical and theoretical interest. Post-hoc power analyses and step-wise Bonferroni corrections were completed for the regression model.

3. Results

Of the 69 mothers initially enrolled, 39/69 (57%) completed the 20 month CA ND visit. Attendance rates are consistent with relevant extant literature (19, 65% follow-up) and with attendance to an urban, NICU follow-up clinic in the United States [33]. The only difference between those who did and did not attend was a higher rate of maternal unemployment (p = 0.049) for those who did attend. Maternal anxiety at T1 and T2 was not related to attendance.

3.1. Descriptive statistics

The characteristics of the sample have been detailed in previously published research [3], but are summarized briefly here (Tables 1 & 2). Overall, this is a predominately low-income, racially/ethnically diverse sample of mothers that commonly endorsed elevated maternal psychological distress during the NICU hospitalization. Rates of elevated distress are -consistent with previous literature [6,8]. Table 3 summarizes infant medical variables. On average VLBW infants were born at 27.5 weeks (SD) weighing 957 g (SD). Their mean length of NICU stay was 91 days (SD). Table 4 summarizes infant ND scores as quantified by the Bayley-III at 20 months corrected age. VLBW infants were most likely to experience language delays at 20 months CA.

Table 1.

Maternal sociodemographic, reproductive and psychological history variables.

N = 69	M ± SD [range], N (%)
Maternal age	27 ± 6 [18–40]
Maternal education (highest grade completed)	13.4 ± 2.4 [10 – 20]
Maternal unemployment	9 (13%)
WIC eligibility	46 (67%)
Maternal race
Black	38 (54%)
Non-Hispanic white	18 (26%)
Hispanic	12 (17%)
Marital status
Married and/or living with partner	32 (51%)
Partnered, not living together	20 (32%)
Single	10 (16%)
Primipara	23 (34%)
History of prior preterm birth	7 (10%)
History of fetal loss	28 (40%)
History of psychological diagnosis prior to birth of preterm infant	5 (8%)
History of involvement with child protective services	1 (2%)
History of substance use during pregnancy	2 (3%)
Number of previous traumatic events witnessed or experienced	2.75 + 2.64
Total number of infant’s caretakersa	2.27 ± 1.12 [1–5]

^aMaternal report of total number of caretakers available for infant includes family and friends beyond child’s other parent.

Table 2.

Maternal psychological distress variables.

	Time 1	Time 2
	M ± SD, N (%)	M ± SD, N (%)
Elevated depression	23 (33%)	15 (23%)
Elevated anxiety	38 (55%)	23 (36%)
Elevated PPTS	17 (25%)	16 (25%)
Mean depression (raw score)	15.20 ± 11.71	12.09 ± 10.45
Mean anxiety (raw score)	44.98 ± 12.99	36.91 ± 11.10
Mean PPTS (raw score)	12.25 ± 10.35	12.02 ± 12.41

Table 3.

Infant medical variables.

N = 69	M ± SD, N (%)
Birthweight (grams)	957 ± 243 [470–1470]
Gestational age (weeks)	27.5 ± 2 [23.2–32.30]
Male sex	34 (49%)
Severely abnormal head ultrasounda	5 (7%)
Any abnormal head ultrasoundb	29 (42%)
Sepsis (presumed or confirmed)	20 (30%)
Necrotizing enterocolitis (stage 2 or 3)	5 (7%)
Length of mechanical ventilation (days)	14.3 ± 20.2 [0–60]
Length of NICU stay (days)	91 ± 37.1 [30–179]
Average human milk dose (mL per kG throughout course of NICU hospitalization)	56.10 ± 53.27 [0–152.50]

^aDefined as presence of intraventricular hemorrhage grades III–IV, periventricular leukomalacia, or hydrocephalus requiring shunt.

^bDefined as presence of any intraventricular hemorrhage grades I–IV, periventricular leukomalacia, or hydrocephalus.

Table 4.

Infant neurodevelopmental scores at 20 months CA.

N = 39	M ± SD, N (%)
Cognitive index	97 ± 12
> 1 SD below mean	4 (10%)
> 2 SD below mean	0 (0%)
Cognitive subscale	9.4 ± 2.3
> 1 SD below mean	4 (10%)
> 2 SD below mean	0 (0%)
Language index	85 ± 16
> 1 SD below mean	18 (46%)
> 2 SD below mean	6 (15%)
Receptive language subscale	7.1 ± 2.7
> 1 SD below mean	19 (49%)
> 2 SD below mean	3 (8%)
Expressive language subscale	7.6 ± 3.0
> 1 SD below mean	14 (36%)
> 2 SD below mean	3 (8%)
Motor index	90 ± 13
> 1 SD below mean	11 (28%)
> 2 SD below mean	2 (5%)
Fine motor index	9.1 ± 2.4
> 1 SD below mean	5 (13%)
> 2 SD below mean	1 (3%)
Gross Motor Index	7.6 ± 2.3
> 1 SD below mean	8 (21%)
> 2 SD below mean	3 (8%)

3.2. Bivariate analyses

3.2.1. Elevated anxiety

Elevated anxiety at T1 was significantly correlated with worse fine motor scores at 20 month CA (r = −0.40, p = 0.02; Fine Motor Score Mean = 9.13 ± 2.41). In additional bivariate analyses, elevated anxiety at T1 was correlated with lower infant birth weight (r = −0.24, p = 0.055) and presence of severely abnormal head ultrasound (HUS) (phi = 0.22, p = 0.09). Elevated anxiety at T2 was not significantly associated with any Bayley scores at 20 months CA.

3.2.2. Elevated depression and PPTS

Elevated depression and PPTS at T1 and T2 were not significantly correlated with any Bayley-III scores at 20 months CA. No further analyses were conducted for these distress variables.

3.2.3. Neurodevelopmental covariates

Given the identified association between elevated anxiety at T1 and fine motor scores at 20 months CA, additional bivariate analyses identified which infant medical and maternal variables correlated with fine motor scores (p < 0.10) and were therefore entered as covariates in regression analyses. Worse fine motor scores at 20 months CA were correlated with younger gestational age at birth (r = 0.40, p = 0.019), smaller infant birth weight (r = 0.35, p = 0.028), presence of severely abnormal HUS (r = −0.36, p = 0.027), and lower maternal educational achievement (r = 0.28, p = 0.09).

3.3. Regression analyses

3.3.1. Elevated anxiety

After controlling for relevant covariates, only presence of a severely abnormal HUS and elevated anxiety at T1 significantly predicted worse fine motor scores at 20 months CA (Table 5). Both of these findings persisted after step-wise Bonferroni corrections. Infants’ birth weight was excluded from the regression analysis given concerns for multi-collinearity with gestational age at birth (r < 0.60, p < 0.0001). Post-hoc power analyses revealed adequate statistical power for this regression (achieved statistical power > 0.80).

Table 5.

Multiple regression predicting fine motor score at 20 months CA.

	Unstandardized B	Unstandardized B 95% CI	Beta	p-value
Step 1
Gestational age at birth (weeks)	0.31	(−0.02, 0.64)	0.27	0.065
Maternal education (highest grade completed)	0.24	(−0.00, 0.48)	0.27	0.052
Severely abnormal head Ultrasound (yes/no)	−5.61	(−9.65, – 1.57)	− 0.39	0.008
Step 2
Elevated maternal anxiety (yes/no)	−1.60	(−2.94, – 0.27)	− 0.33	0.02

Final F (4,29) = 7.29, p ≤0.001; R² = 0.50; Step 2 ΔR² = 0.10.

To account for attrition and attendance to the 20-month appointment, we ran additional regressions including maternal unemployment status; unemployment did not significantly predict fine motor scores nor did its inclusion change the significance of any other predictor variables or the overall regression.

4. Discussion

To our knowledge, this is the first study in fifteen years to examine associations between elevated depression, anxiety, and PPTS experienced by mothers of VLBW infants during their infants’ NICU hospitalization and ND outcome in the second year of the infants’ life. Our findings reveal that elevated maternal anxiety was the only aspect of distress associated with adverse ND outcomes and that the timing of the distress, occurring during the first month of the infants’ NICU hospitalization, was significant. Although our findings revealed that maternal anxiety early in the infants’ NICU stay predicted worse fine motor skills at 20 months CA, a previous study by Zelkowitz and colleagues [21] found that combined cognitive and language developmental outcomes in the second year of life were adversely impacted by elevated maternal anxiety during the same time period [21].

Despite these seemingly incongruous results, our findings build upon and likely refine previous investigations’ results. The Zelkowitz et al. [21] investigation used the second edition of Bayley Scales of Infant and Toddler Development (Bayley-II) [34] that provided a “Mental Developmental Index (MDI),” and a “Psychomotor Developmental Index (PDI).” The MDI measured a combination of early cognitive and language development and a low score could represent a delay in language skills, cognition, or both. The MDI also included a substantial number of fine motor items. The current investigation used the 3rd Edition of the Bayley Scales (Bayley-III), which separates the MDI into four separate subscales in an attempt to isolate fine motor skills and differentiate cognitive skills from language skills: cognitive, receptive and expressive language, and fine motor. Forty-two percent of the items on the Bayley III fine motor subscale most commonly administered to children at 20 months CA have been adapted from the Bayley II MDI, 33% have been adapted from PDI, and 25% newly created for Bayley-III. Thus, because of the expansion of categories in the Bayley-III, the current fine motor results may partially confirm and likely refine the previous investigations results.

The acquisition of fine motor skills early in life is fundamental to the ability to learn how to feed and dress oneself and perform self-care. Within this context, it is known that elevated maternal anxiety early in the NICU stay is associated with less sensitive and less structured parenting among mothers of VLBW infants [35], which we hypothesize results in a style of maternal interaction that makes it more difficult for infants to acquire these specific adaptive self-care skills later in life. Therefore, because of our use of the Bayley III which isolated fine motor skills from cognitive and language development, we posit that the association between elevated maternal anxiety and altered ND outcome is unique to adaptive life skills and not distinctly related to cognitive or language development. Alternatively, there is a possibility that this result is secondary to a chance finding; however, our use of step-wise Bonferroni corrections aimed to mitigate this statistical risk.

The finding that there was no association between any elevated measure of psychological distress near the time of NICU discharge (T2) and adverse ND outcomes marks an important contribution to the field. For mothers of full-term infants, the critical time during which maternal psychological distress is known to have the greatest impact on mother-child relationships and ND outcome is the first three months of the post-partum period. However, only the first month of the post-partum period corresponding to the infants’ NICU hospitalization appears to have the greatest impact on ND outcome for mothers of VLBW infants. Earlier in VLBW infants’ NICU hospitalization the contrast between mothers’ expectations for their pregnancy, childbirth and infant, and their reality is the most striking and emotionally poignant [1]. Correspondingly, mothers’ rates of elevated psychological distress are higher earlier in the NICU hospitalization [3]. We hypothesize that the distress specifically related to mothers’ violated expectations that occurs earlier in the NICU hospitalization sets forth a style of maternal interaction that can negatively alter child outcome. Replication and extension of our findings would provide valuable information for clinicians regarding the most efficient, yet critical time window for psychological intervention for mothers of VLBW infants hospitalized in the NICU.

Neither elevated maternal depression nor elevated PPTS during the NICU period was associated with any ND outcome at 20 months CA. To our knowledge, this is the first published investigation to analyze these associations with a contemporary population of VLBW infants. The present investigation’s relatively low prevalence of elevated maternal depression during the NICU hospitalization (23–33%) may reflect a restricted range that contributes to the null findings. While some previous investigations have revealed similar rates of elevated maternal depression in the NICU (20%) [6], commonly-cited studies suggest that up to 63% of mothers with VLBW infants’ hospitalized in the NICU experience elevated rates of depression [36]. Other researchers with similar lower income, urban sample demographics that have encountered lower than expected levels of maternal depression argue that mothers’ chronic, daily stress related to socioeconomic disadvantage dominates over the acute stress of the NICU [37]. This hypothesis suggests that the present investigation’s sample of mothers from lower income, urban backgrounds are not perceiving or experiencing enough acute depressive symptoms to accurately reveal associations between elevated depression and child outcome. Interestingly, the present investigation’s rate of elevated PPTS (25%) is highly consistent with well-established prevalence rates reported in the literature (PPTS 23%) [8] and as such restricted range does not appear to offer an explanation regarding null findings with regard to this domain of distress.

Review of the literature in full-term infants offers several other insights regarding this study’s lack of significant associations between elevated maternal depression and elevated PPTS and ND outcome. Several investigations have revealed that the associations between elevated maternal post-partum depression and adverse ND outcome in full-term infants are most significant among mothers with chronic depression [10,11,38]. Research of trajectories of depression and PPTS among mothers of VLBW infants suggests that while elevated depression and PPTS are common during the NICU hospitalization, the vast majority of these mothers demonstrate a “resilient” or “recovered” and not a “chronically distressed,” pattern of functioning over the first two years after child birth [4,20].

Furthermore, some research suggests that there may be a fundamental flaw in the hypothesis that the association between elevated maternal post-partum depression and adverse ND outcome is explained by, or mediated by, difficulties with the mother-child relationship [11,39]. Some investigations that have revealed inconsistent associations between elevated maternal post-partum depression and worse ND outcome in full-term children have concluded that mothers with elevated post-partum depression may be capable of “good enough,” parenting interactions that allow them to parent such that their depression does not negatively impact children’s ND outcome as widely theorized [11,39]. A similar hypothesis is likely for mothers experiencing elevated PPTS [12]. It is possible that either the concept of “good enough,” parenting or the lack of chronicity of PPTS symptoms may explain the lack of association between elevated PPTS during the first month of the NICU hospitalization and worse ND outcome in this study.

This study has several limitations, including its relatively small sample size and attrition. It is possible that the study was not sufficiently powered to reveal associations between elevated depression and PPTS during the NICU hospitalization and ND outcome. Although attrition was statistically controlled for and the attrition rate is consistent with previous literature [21,33], it is important to note that rate of attrition, or non-attendance to the 20 month appointment, was not related to any measure of elevated maternal psychological distress at either time point during the NICU hospitalization. A previous investigation has revealed an association between distress during NICU hospitalization and follow-up clinic attendance [40], but these associations did not approach statistical significance in the present sample (p > 0.30). Other limitations are the study’s use of self-report measures to gauge elevated symptomatology and absence of direct assessment of mother-preterm infant interaction. Self-report measures are well-supported empirically in the literature, yet are not as rigorous as clinical diagnostic interviews or directly observable behaviors which are arguably more objective. Furthermore, recent observational assessment of mother-VLBW infant interaction reveals that while the majority of mothers of VLBW infants demonstrate sensitive, synchronous interactions typical of mothers of term infants [41], prematurity relates to controlling behavior in mothers with insensitive interaction styles [41,42]. This controlling maternal behavior associated with prematurity and its potential association with elevated maternal anxiety during the NICU hospitalization marks a high priority for future research. Finally, this study focused exclusively on psychological distress in mothers and not fathers. A recent investigation has revealed important associations between elevated post-partum depression in fathers of preterm infants and worse ND outcome [43]. Again, this marks an important consideration for future research.

The study’s use of the CES-D versus the post-partum specific Edinburgh Postnatal Depression Scale (EPDS) [44] may be considered a limitation by some. However, the CES-D has been widely used among mothers of preterm infants hospitalized in the NICU [4,6] and has a well-established cut-off score with this population [4,6]. Additionally, the EPDS is not specific to mothers of preterm infants and wording for several items can be viewed as insensitive to mothers of infants hospitalized in the NICU. Specifically, two items ask mothers to rate the presence of different aspects of distress that they are experiencing “for no [very] good reason.” Based on our clinical experiences with mothers of VLBW infants, many mothers object to these questions, or even refuse to answer these items, stating that they believe the birth of a VLBW infant and subsequent NICU hospitalization would be considered a “[very] good reason,” for experiencing psychological distress. In contrast, our research team did not encounter any objections to items on the CES-D. This study’s results, many of which were novel contributions to the field, and limitations highlight the need for replication and extension with future, larger prospective cohort investigations.

In conclusion, we have found that elevated maternal anxiety during the critical first month of the NICU hospitalization was associated with worse fine motor outcomes at 20 months CA. Maternal anxiety, which is common and an understandable experience in the context of these mothers’ realities [3,6,21], warrants additional attention to address the quality of life of these mothers and to better understand a potentially modifiable risk factor for their infant’s ND outcome.