Association Between In-Hospital Pacifier Use and Breastfeeding Continuation and Exclusivity: Neonatal Intensive Care Unit Admission as a Possible Effect Modifier

Abstract

Background:

Breastfeeding dyads frequently use pacifiers during the birth hospitalization, but the relationships between this exposure and breastfeeding continuation and exclusivity remain unclear.

Materials and Methods:

In this secondary analysis of cross-sectional survey data from the CDC Pregnancy Risk Assessment Monitoring System (PRAMS) from 10 U.S. states (AK, AR, CO, ME, MN, NJ, NY, OR, TX, and VT) from 2009 to 2011, we assessed to what extent pacifier use during the birth hospitalization is independently associated with any and exclusive breastfeeding ≥10 weeks.

Results:

A total of 37,628 mothers who were chosen by random birth certificate sampling completed surveys at ∼4 months postpartum. Adjusting for multiple pro-breastfeeding hospital practices and maternal and infant demographic characteristics, pacifier exposure during the birth hospitalization was independently associated with decreased odds of breastfeeding ≥10 weeks (adjusted odds ratio [aOR]) 0.71, 95% confidence interval [CI] 0.63–0.80, p < 0.0001) and exclusive breastfeeding ≥10 weeks (aOR 0.70, 95% CI 0.63–0.79, p < 0.0001) among infants admitted to the well-baby nursery, but not among those admitted to the neonatal intensive care unit (NICU).

Conclusions:

In this population study of mothers from 10 U.S. states, pacifier use during the birth hospitalization was associated with decreased odds of breastfeeding and exclusive breastfeeding ≥10 weeks among well-newborn, but not NICU-admitted infants. Pacifier use may be a marker rather than a cause of breastfeeding difficulties, but prospective, randomized studies are needed to help clarify this. Future studies exploring pacifier exposure and breastfeeding outcomes should account for NICU admission as an effect modifier.

Introduction

The effect of early pacifier exposure on the duration of any and exclusive breastfeeding is a hotly debated topic. Babies and parents across many cultures enjoy pacifiers, but medical professionals in the U.S. counsel against them in the early postpartum period. The American Academy of Pediatrics (AAP) Section on Breastfeeding recommends avoiding pacifier exposure until breastfeeding is fully established, at about 3–4 weeks of age.¹ This is both in an effort to avoid “nipple confusion,” whereby a breastfed infant will become accustomed to the artificial nipple of a pacifier or bottle and will struggle to latch at breast, as well as to prevent parents from using artificial nipples to pacify an infant who needs to breastfeed to stimulate maternal milk production.

One previous study found a temporal association between a unit policy to restrict the distribution of pacifiers to breastfeeding infants and a decrease in exclusive breastfeeding during the birth hospitalization.² This result was in contrast to multiple other observational studies, a meta-analysis of cross-sectional and cohort studies, and one randomized controlled trial showing a negative association between pacifier exposure and ongoing exclusive breastfeeding.^3–6 However, two systematic reviews have examined the results of the available randomized trials that examine this topic, and both have concluded that the available high-quality evidence does not definitively show a negative effect of pacifiers on exclusive breastfeeding duration.^7,8

Despite this lack of consensus, hospitals are encouraged to restrict access to pacifiers during the birth hospitalization as part of the World Health Organization (WHO) and The United Nations Children's Emergency Fund (UNICEF)'s “Ten Steps to Successful Breastfeeding,” a bundle of evidence-based practices shown to increase breastfeeding initiation and duration.^9,10 Hospitals and birthing facilities seeking Baby-Friendly designation must adhere to these hospital pro-breastfeeding practices.¹¹

We utilized population-level data from the CDC Pregnancy Risk Assessment Monitoring System (PRAMS) to assess the extent to which pacifier exposure during the birth hospitalization is independently associated with any and exclusive breastfeeding for ≥10 weeks. In doing so, we also examined which pro-breastfeeding hospital practices are independently associated with any and exclusive breastfeeding for ≥10 weeks.

Materials and Methods

Institutional Review Board approval

This study was reviewed by the University of Iowa Institutional Review Board and was granted a waiver of consent, as data were deidentified.

Design

We analyzed cross-sectional survey data from the CDC PRAMS from 10 U.S. states from 2009 to 2011. The PRAMS surveillance system utilizes birth certificate registries to provide data from a stratified, systematic sample of ∼100 to 250 new mothers from each participating state per month. To ensure that small but high-risk populations within individual states are represented in large enough numbers for analysis, each state has a slightly different stratification scheme. The survey results are weighted to account for this deliberate over-sampling of certain populations as well as noncoverage and nonresponse.

The PRAMS survey instrument is administered to participants by mail beginning at 2–4 months postpartum, with three attempts made to contact mothers by mail before a study administrator then contacts mothers by telephone. The PRAMS dataset consists of results from these questionnaires as well as deidentified birth certificate data. Additional information about PRAMS methodology can be found on the CDC-PRAMS website.¹²

For our study, we analyzed data provided by mothers from 10 U.S. states that administered an optional survey question about hospital practices to which they may or may not have been exposed during the birth hospitalization, including pacifiers. This question read:

“This question asks about things that may have happened at the hospital where your new baby was born. For each item, circle Y (Yes) if it happened or circle N (No) if it did not happen.

a. Hospital staff gave me information about breastfeeding

b. My baby stayed in the same room with me at the hospital

c. I breastfed my baby in the hospital

d. I breastfed in the first hour after my baby was born

e. Hospital staff helped me learn to breastfeed

f. My baby was fed only breast milk at the hospital

g. Hospital staff told me to breastfeed whenever my baby wanted

h. The hospital gave me a breast pump to use

i. The hospital gave me a gift pack with formula

j. The hospital gave me a telephone number to call for help with breastfeeding

k. My baby used a pacifier in the hospital.”

The CDC only releases PRAMS data from states that meet a certain minimum response rate: for 2009–2011 this rate was set at >65%.¹² The 10 states included in our dataset were: Alaska, Arkansas, Colorado, Maine, Minnesota, New Jersey, New York, Oregon, Texas, and Vermont.

Breastfeeding for the purposes of this study included both feeding at breast and/or feeding of expressed breast milk, and it was defined by a mother answering yes to the question: “Did you ever breastfeed or pump breast milk to feed your new baby after delivery?” Any breastfeeding ≥10 weeks was defined by a mother's response to the two questions: “Are you still currently breastfeeding or feeding pumped milk to your new baby?” and “How many weeks or months did you breastfeed or pump milk to feed your baby.” Breastfeeding for 10 weeks is the longest duration that can be assessed using CDC PRAMS data, given the timing of survey completion.

Exclusive breastfeeding ≥10 weeks was determined based on the WHO definition of exclusive breastfeeding, using mothers' responses to the question, “How old was your new baby the first time he or she drank liquids other than breast milk (such as formula, water, juice, tea, or cow's milk?)” This question included the answer choice, “My baby has not had any liquids other than breast milk.”

Statistical analysis

To account for the complex PRAMS survey design, we utilized SAS Complex Survey Version 9.3 (SAS Institute, Research Triangle Park, NC) to perform weighted analyses. We performed Wald chi-square analysis to determine those demographic characteristics associated with any or exclusive breastfeeding ≥10 weeks. We analyzed infants cared for in the well nursery/mother baby care unit separately from those who received neonatal intensive care.

We then used multivariable logistic regression to model the effect of the independent variable of in-hospital pacifier use on the dependent variables of any breastfeeding ≥10 weeks and exclusively breastfeeding ≥10 weeks among infants admitted to the well nursery and those admitted to the neonatal intensive care unit (NICU). We adjusted for exposure to each of the hospital breastfeeding support practices assessed by PRAMS as well as those demographic factors found to be significantly associated (p < 0.01) with any or exclusive breastfeeding ≥10 weeks among well-nursery or NICU-admitted infants with the Wald chi-square testing: maternal age, maternal education, maternal race, parity, mode of delivery, maternal marital status, Medicaid insurance, maternal pre-pregnancy body mass index (BMI) category, prematurity, and state of residence.

Results

A total of 37,628 mothers in these 10 states were surveyed at approximately 2–4 months postpartum (median days postpartum for survey completion was 108 days with interquartile range 93–131 days). Of this total, 29,924 of the mothers delivered an infant who received care in the well-baby nursery. All demographic characteristics assessed were significantly associated with any breastfeeding and exclusive breastfeeding ≥10 weeks among infants admitted to the well nursery (Table 1). Those demographic characteristics associated with any and exclusive breastfeeding among infants admitted to the NICU are shown in Table 2.

Table 1.

Characteristics of Well-Newborn Nursery-Admitted Infants and Their Mothers

Characteristics	N^a (%)^b	Breastfeeding ≥10 weeks p^c	Exclusive breastfeeding ≥10 weeks p^d
Maternal age, years		<0.0001	<0.0001
≤19	163,814 (9)
20–24	374,796 (22)
25–29	495,503 (29)
30–34	447,371 (26)
≥35	256,687 (15)
Maternal education, years		<0.0001	<0.0001
<12	323,105 (19)
12	448,042 (26)
>12	957,641 (55)
Maternal race		<0.0001	<0.0001
White	1,295,221 (76)
Black	168,665 (10)
Other	244,752 (14)
Gestational age, weeks		0.001	0.062
≥37	1,655,035 (95)
<37	82,000 (5)
Initiated breastfeeding		<0.0001	<0.0001
Yes	1,452,015 (85)
No	256,892 (15)
Married		<0.0001	<0.0001
Yes	1,094,874 (63)
Other	642,848 (37)
Medical insurance		<0.0001	<0.0001
Medicaid	801,260 (46)
Non-Medicaid	935,406 (54)
Parity		<0.0001	<0.0001
Primiparous	685,375 (39)
Multiparous	1,052,798 (61)
Mode of delivery		0.0001	<0.0001
Vaginal	1,182,023 (68)
Other	554,097 (32)
Maternal pre-pregnancy BMI		<0.0001	<0.0001
Underweight	191,389 (11)
Normal	822,310 (47)
Overweight	384,537 (22)
Obese	339,937 (20)
State		<0.0001	<0.0001
Alaska	19,784 (1)
Arkansas	95,246 (5)
Colorado	172,950 (10)
Maine	33,247 (2)
Minnesota	174,746 (10)
New Jersey	251,222 (14)
New York (excluding NYC)	194,261 (11)
Oregon	119,961 (7)
Texas	660,537 (38)
Vermont	16,218 (1)

Number of women from sample distribution, weighted to account for deliberate survey oversampling as well as nonresponse and noncoverage.

Weighted percentages.

P-value is from the chi-square test of independence between characteristics and breastfeeding ≥10 weeks.

P-value is from the chi-square test of independence between characteristics and exclusive breastfeeding ≥10 weeks.

BMI, body mass index.

Table 2.

Characteristics of Neonatal Intensive Care Unit-Admitted Infants and Their Mothers

Characteristics	N^a (%)^b	Breastfeeding ≥10 weeks P^c	Exclusive breastfeeding ≥10 weeks P^d
Maternal age, years		<0.0001	0.158
≤19	22,084 (9)
20–24	52,070 (22)
25–29	68,492 (28)
30–34	61,272 (25)
≥35	37,390 (15)
Maternal education, years		<0.0001	0.548
<12	47,764 (20)
12	68,389 (29)
>12	123,637 (52)
Maternal race		<0.0001	0.026
White	174,089 (73)
Black	31,418 (13)
Other	32,197 (14)
Gestational age, weeks		0.013	0.568
≥37	141,234 (59)
<37	99,835 (41)
Initiated breastfeeding
Yes	196,161 (84)
No	38,146 (16)
Married		<0.0001	0.343
Yes	137,323 (57)
Other	103,819 (43)
Medical insurance		<0.0001	0.298
Medicaid	120,318 (50)
Non-Medicaid	120,858 (50)
Parity		0.482	0.257
Primiparous	105,225 (44)
Multiparous	136,082 (56)
Mode of delivery		0.910	0.452
Vaginal	119,612 (50)
Other	121,053 (50)
Maternal pre-pregnancy BMI		0.008	0.045
Underweight	29,669 (12)
Normal	100,863 (42)
Overweight	55,812 (23)
Obese	54,963 (23)
State		<0.0001	0.002
Alaska	1,971 (1)
Arkansas	12,104 (5)
Colorado	20,953 (9)
Maine	4,035 (2)
Minnesota	18,398 (8)
New Jersey	42,779 (18)
New York (excluding NYC)	28,076 (12)
Oregon	10,814 (4)
Texas	100,581 (42)
Vermont	1,597 (1)

Number of women from sample distribution, weighted to account for deliberate survey oversampling as well as nonresponse and noncoverage.

Weighted percentages.

P-value is from the chi-square test of independence between characteristics and breastfeeding ≥10 weeks.

P-value is from the chi-square test of independence between characteristics and exclusive breastfeeding ≥10 weeks.

Well-newborn nursery infants

Using multivariable logistic regression (controlling for the hospital practices listed in Table 3, as well as maternal age, race, education, pre-pregnancy BMI category, Medicaid status, marital status, primiparity, mode of delivery, gestational age, and state of residence), pacifier exposure during the birth hospitalization was independently associated with decreased odds of any breastfeeding ≥10 weeks among infants admitted to the well-newborn nursery (adjusted odds ratio [aOR] 0.71, 95% confidence interval [CI] 0.63–0.80, p < 0.0001). Hospital practices associated with increased odds of any breastfeeding ≥10 weeks among infants admitted to the well-newborn nursery are shown in Table 3 and included: breastfeeding during the birth hospitalization (aOR 2.34, 95% CI 1.68–3.25, p < 0.0001), exclusive breast milk feeding (aOR 2.17, 95% CI 1.91–2.48, p < 0.0001), breastfeeding in the first hour (aOR 1.26, 95% CI 1.10–1.46, p < 0.0001), breastfeeding on demand (aOR 1.22, 95% CI 1.02–1.45, p = 0.029), and receiving a phone number for breastfeeding help at discharge (aOR 1.22, 95% CI 1.02–1.45, p = 0.026). Interestingly, use of a breast pump during the birth hospitalization was associated with decreased odds of any breastfeeding ≥10 weeks (aOR 0.80, 95% CI 0.70–0.91, p = 0.0009), as was staff providing a mother with breastfeeding information (aOR 0.67, 95% CI 0.49–0.92, p = 0.014) and staff helping the mother learn to breastfeed (aOR 0.69, 95% CI 0.58–0.83, p < 0.0001).

Table 3.

Hospital Practices Predictive of Any and Exclusive Breastfeeding ≥10 Weeks Among Infants Admitted to the Well-Newborn Nursery

	Any breastfeeding ≥10 weeks			Exclusive breastfeeding ≥10 weeks
Hospital practice	Adjusted OR ^a	95% CI	p	Adjusted OR ^a	95% CI	p
Staff gave breastfeeding information	0.67	0.49–0.92	0.014	0.95	0.69–1.30	0.732
Baby stayed in the same room with mother	0.89	0.73–1.08	0.237	0.90	0.71–1.12	0.345
Breastfed in the first hour after delivery	1.26	1.10–1.46	0.001	1.13	0.97–1.32	0.116
Staff helped mother learn to breastfeed	0.69	0.58–0.83	<0.0001	0.78	0.66–0.93	0.005
Mother breastfed baby in the hospital	2.34	1.68–3.25	<0.0001	0.95	0.65–1.38	0.777
Baby fed exclusively breast milk in hospital	2.17	1.91–2.48	<0.0001	4.90	4.26–5.65	<0.0001
Mother advised to breastfeed on demand	1.22	1.02–1.45	0.029	0.99	0.80–1.22	0.904
Hospital gave mother a breast pump to use	0.80	0.70–0.91	0.0009	0.83	0.73–0.95	0.006
Hospital gave mother a gift pack with formula	0.90	0.78–1.03	0.112	0.83	0.73–0.94	0.005
Hospital gave a phone number to call for breastfeeding help	1.22	1.02–1.45	0.026	1.06	0.88–1.29	0.519
Baby used a pacifier in the hospital	0.71	0.63–0.80	<0.0001	0.70	0.63–0.79	<0.0001

Referent group is infants cared for in the well-baby nursery who were not exposed to the practice.

Adjusted for the listed hospital practices, maternal age, education, race, pre-pregnancy BMI category, marital status, Medicaid status, primiparity, mode of delivery, infant gestational age, and state of residence.

CI, confidence interval; OR, odds ratio.

Using the same predictor variables to explore the outcome of exclusive breastfeeding ≥10 weeks among infants admitted to the well nursery, pacifier use was associated significantly with decreased odds of exclusive breastfeeding ≥10 weeks (aOR 0.70, 95% CI 0.63–0.79, p < 0.0001). Exclusive breast milk feeding during the birth hospitalization was associated with increased odds of exclusive breastfeeding ≥10 weeks (aOR 4.90, 95% CI 4.26–5.65, p < 0.0001). Factors associated with decreased odds of exclusive breastfeeding ≥10 weeks included: staff helping a mother learn to breastfeed (aOR 0.78, 95% CI 0.66–0.93, p = 0.005), hospital giving the mother a breast pump to use (aOR 0.83, 95% CI 0.73–0.95, p = 0.006), and hospital giving the mother a gift pack with formula (aOR 0.73, 95% CI 0.73–0.94, p = 0.005).

NICU-admitted infants

Using the same predictor variables to model breastfeeding ≥10 weeks among infants admitted to the NICU, pacifier use during the birth hospitalization was not associated with breastfeeding ≥10 weeks (aOR 0.91, 95% CI 0.59–1.42, p = 0.685), nor was the hospital providing the mother with a breast pump to use (aOR 1.01, 95% CI 0.66–1.55, p = 0.957). Hospital practices associated with increased odds of breastfeeding ≥10 weeks among NICU-admitted infants are shown in Table 4 and included: breastfeeding in the first hour after delivery (aOR 1.63, 95% CI 1.10–2.43, p = 0.016), infant breastfed in the hospital (aOR 2.06, 95% CI 1.49–2.85, p < 0.0001), and exclusive breast milk feeding in the hospital (aOR 1.62, 95% CI 1.20–2.19, p = 0.002).

Table 4.

Hospital Practices Predictive of Any and Exclusive Breastfeeding ≥10 Weeks Among Infants Admitted to the Neonatal Intensive Care Unit

	Any breastfeeding ≥10 weeks			Exclusive breastfeeding ≥10 weeks
Hospital practice	Adjusted OR ^a	95% CI	P	Adjusted OR ^a	95% CI	p
Staff gave breastfeeding information	1.19	0.57–2.48	0.645	1.36	0.68–2.71	0.379
Baby stayed in the same room with mother	0.87	0.60–1.25	0.436	1.06	0.71–1.58	0.784
Breastfed in the first hour after delivery	1.63	1.10–2.43	0.016	0.98	0.63–1.52	0.922
Staff helped mother learn to breastfeed	0.75	0.53–1.07	0.115	1.12	0.72–1.77	0.609
Mother breastfed baby in the hospital	2.06	1.49–2.85	<0.0001	1.30	0.91–1.85	0.150
Baby fed exclusively breast milk in hospital	1.62	1.20–2.19	0.002	4.85	3.56–6.62	<0.0001
Mother advised to breastfeed on demand	1.18	0.88–1.57	0.264	0.98	0.69–1.39	0.901
Hospital gave mother a breast pump to use	1.01	0.66–1.55	0.957	0.57	0.39–0.84	0.005
Hospital gave mother a gift pack with formula	0.75	0.54–1.02	0.070	0.74	0.53–1.04	0.087
Hospital gave a phone number to call for breastfeeding help	1.07	0.68–1.71	0.765	0.97	0.58–1.65	0.925
Baby used a pacifier in the hospital	0.91	0.59–1.42	0.685	0.88	0.57–1.36	0.574

Referent group is infants cared for in the NICU who were not exposed to the practice.

NICU, neonatal intensive care unit.

Using the same predictor variables to model the outcome of exclusive breastfeeding ≥10 weeks among infants admitted to the NICU, pacifier use was not associated significantly with odds of exclusive breastfeeding ≥10 weeks (aOR 0.88, 95% CI 0.57–1.36, p = 0.574). As shown in Table 4, among NICU-admitted infants, the hospital providing the infant's mother with a breast pump to use was associated with decreased odds of exclusive breastfeeding ≥10 weeks (aOR 0.57, 95% CI 0.39–0.84, p = 0.005), and in-hospital exclusive breast milk feeding was associated with increased odds of exclusive breastfeeding ≥10 weeks (aOR 4.85, 95% CI 3.56–6.62, p < 0.0001).

Discussion

Our study adds to the existing literature additional evidence that pacifier use in the well-newborn nursery/mother baby care unit is associated with adverse breastfeeding outcomes. Interestingly, the same was not true for pacifier use during an NICU stay. We also found the hospital providing mothers with a breast pump to be associated with adverse breastfeeding outcomes in both the well newborn and NICU populations.

Among infants admitted to the well-baby nursery, pacifier use during the birth hospitalization was associated with decreased odds of both any and exclusive breastfeeding for ≥10 weeks. However, in our adjusted models, both maternal receipt of breastfeeding information and staff helping mothers learn to breastfeed were also associated with adverse breastfeeding outcomes. Reverse causality, whereby mothers who were struggling to breastfeed during the birth hospitalization were given more information about and help with breastfeeding, is one very plausible explanation of these findings.

Use of a pacifier during a well-nursery hospitalization may, likewise, be a marker of breastfeeding difficulties and associated infant fussiness rather than a cause of breastfeeding problems. The “marker rather than cause” hypothesis is in agreement with the findings of Kramer et al. published in JAMA in 2001.¹³ This study found no association between advising parents to avoid pacifiers and early weaning in intention-to-treat analysis, but found a strong observational association between pacifier use and early weaning in the per-protocol analysis, indicating that pacifier use is a marker rather than a cause of breastfeeding difficulties.

We caution, however, that one randomized controlled trial examining timing of pacifier introduction, conducted by Howard et al., found early pacifier introduction (median 1 week compared with late pacifier introduction at 4 weeks) to lead to a slightly decreased likelihood of overall breastfeeding duration (hazard ratio 1.2, 95% CI 1.02–1.42).³ Jenik et al. published the results of another large randomized controlled trial in 2009, which found the recommendation to offer a pacifier at 15 days did not modify the prevalence or duration of any or exclusive breastfeeding at 3 months.¹⁴

In the study we present herein, we found no association between pacifier use during the birth hospitalization and breastfeeding outcomes at 10 weeks among NICU-admitted infants. This may be because infants in the NICU are more likely to be fed expressed breast milk by bottle due to separation from their mothers, and any nipple confusion from pacifier exposure is less of a detriment to ongoing breastfeeding. Additionally, the population of mothers whose infants were admitted to the NICU had a higher proportion of mothers with 12 or fewer years' education and a higher proportion of primiparous mothers, mothers who are Medicaid recipients, and unmarried mothers than the population of infants admitted to the well nursery. These and other differences between these two populations of mothers may modify the effect that pacifiers have on breastfeeding continuation and exclusivity, as it has been previously suggested that early pacifier exposure may have a positive effect on exclusive breastfeeding among vulnerable mothers.¹⁵

We also found maternal use of a breast pump during the birth hospitalization to be associated with adverse breastfeeding outcomes in both the well newborn and NICU-admitted populations. Like pacifiers, this may be secondary to unresolved confounding from breastfeeding difficulties. However, the fact that the association between breast pump use and lower odds of exclusive breastfeeding for 10 weeks was seen in both the NICU and well-baby populations makes it worthy of discussion and consideration. This finding indicates a need for additional studies of a prospective and randomized design to help inform breast pumping recommendations versus alternative therapies to aid with medically indicated supplementation, such as hand expression or short-term donor milk or formula supplementation.^16–18

Previous qualitative work has shown breast pumping to be a strong negative experience for some women,^19,20 so it is certainly possible that encouraging mothers to pump during the birth hospitalization may lead to an undesired effect of early weaning. In fact, in a randomized trial comparing early (before the onset of lactogenesis II) electric breast pumping with a control intervention of holding the pump cups to the breast without suction six times per day among women who delivered through Cesarean, Chapman et al. found that breast pumping not only did not improve milk transfer in the first 3 days postpartum, but also primiparae in the pumping group breastfed for ∼5 months less than those in the control group.²¹

However, one prospective study of 704 breastfeeding mothers comparing the effect of hospital gift bags containing formula versus breastfeeding information alone versus breastfeeding information plus a manual breast pump on breastfeeding exclusivity found improved exclusive breastfeeding rates through 12 weeks among both the pump and breastfeeding information groups compared with the formula group, and the pump group fared no worse than the information group. Additionally, 97% of the 127 mothers in the pump group reported being grateful for the manual pump.²²

The results of our study must be interpreted in the context of its limitations. While the results are representative of the populations within the included states, our findings may not extend to populations in other U.S. states or other countries. Variations in hospital-level criteria that determine which infants are cared for in the well-baby nursery versus the NICU may confound the results and cannot be adjusted for using the PRAMS dataset. As mentioned previously, the retrospective, observational design of this study allows us to show associations but not causation. We provide observational population-level data that support a need for additional prospective research to examine the effect of pacifiers and breast pumps on breastfeeding continuation and exclusivity.

Conclusion

In this population study of mothers from 10 U.S. states, pacifier use during the birth hospitalization was associated with decreased odds of any and exclusive breastfeeding ≥10 weeks among well-newborn but not NICU-admitted infants. Use of a breast pump during the birth hospitalization was associated with adverse breastfeeding outcomes. Pacifier and breast pump use may both be markers of rather than causes of breastfeeding difficulties, but prospective, randomized studies are needed to help clarify this. Future studies exploring the effect of early pacifier exposure on later breastfeeding outcomes should take into account NICU admission as an effect modifier. Physicians and hospital staff should use caution when recommending use of a pacifier or breast pump during the birth hospitalization, as the true risks of these interventions are unknown. Our work supports the current recommendation of not offering pacifiers to breastfeeding infants in the first 3–4 weeks after birth, but suggests that this may not be necessary for infants in the NICU setting.

Footnotes

Acknowledgments

The authors thank the Stead Family Department of Pediatrics at the University of Iowa for providing funding for the software used for analysis and for travel to present this work in abstract form at the Pediatric Academic Societies' annual meeting in Baltimore, MD, in May 2016. They also thank the PRAMS working group: Alabama—Qun Zheng, MS; Alaska—Kathy Perham-Hester, MS, MPH; Arkansas—Mary McGehee, PhD; Colorado—Alyson Shupe, PhD; Connecticut—Jennifer Morin, MPH; Delaware—George Yocher, MS; Florida—Kelsi E. Williams; Georgia—Chinelo Ogbuanu, MD, MPH, PhD; Hawaii—Jane Awakuni; Illinois—Theresa Sandidge, MA; Iowa—Sarah Mauch, MPH; Louisiana—Amy Zapata, MPH; Maine—Tom Patenaude, MPH; Maryland—Diana Cheng, MD; Massachusetts—Emily Lu, MPH; Michigan—Patricia McKane; Minnesota—Judy Punyko, PhD, MPH; Mississippi—Brenda Hughes, MPPA; Missouri—Venkata Garikapaty, MSc, MS, PhD, MPH; Montana—JoAnn Dotson; Nebraska—Brenda Coufall; New Hampshire—David J. Laflamme, PhD, MPH; New Jersey—Ingrid Morton, MS; New Mexico—Eirian Coronado, MPH; New York State—Anne Radigan-Garcia; New York City—Candace Mulready-Ward, MPH; North Carolina—Kathleen Jones-Vessey, MS; North Dakota—Sandra Anseth; Ohio—Connie Geidenberger PhD; Oklahoma—Alicia Lincoln, MSW, MSPH; Oregon—Kenneth Rosenberg, MD, MPH; Pennsylvania—Tony Norwood; Rhode Island—Sam Viner-Brown, PhD; South Carolina—Mike Smith, MSPH; Texas—Tanya Guthrie, PhD; Tennessee—Ramona Lainhart, PhD; Utah—Laurie Baksh, MPH; Vermont—Peggy Brozicevic; Virginia—Christopher Hill, MPH, CPH; Washington—Linda Lohdefinck; West Virginia—Melissa Baker, MA; Wisconsin—Katherine Kvale, PhD; Wyoming—Amy Spieker, MPH; CDC PRAMS Team, Applied Sciences Branch, Division of Reproductive Health.

Disclosure Statement

No competing financial interests exist.

References

Section on Breastfeeding. Breastfeeding and the use of human milk. Pediatrics, 2012; 129:e827–e841.

Kair

, Kenron

, Etheredge

, et al. Pacifier restriction and exclusive breastfeeding. Pediatrics, 2013; 131:e1101–e1107.

Howard

, Howard

, Lanphear

, et al. Randomized clinical trial of pacifier use and bottle-feeding or cupfeeding and their effect on breastfeeding. Pediatrics, 2003; 111:511–518.

Buccini

, Perez-Escamilla

, Venancio

. Pacifier use and exclusive breastfeeding in Brazil. J Hum Lact, 2016; 32:NP52–60.

Lindau

, Mastroeni

, Gaddini

, et al. Determinants of exclusive breastfeeding cessation: Identifying an “at risk population” for special support. Eur J Pediatr, 2015; 174:533–540.

Karabulut

, Yalcin

, Ozdemir-Geyik

, et al. Effect of pacifier use on exclusive and any breastfeeding: A meta-analysis. Turk J Pediatr, 2009; 51:35–43.

O'Connor

, Tanabe

, Siadaty

, et al. Pacifiers and breastfeeding: A systematic review. Arch Pediatr Adolesc Med, 2009; 163:378–382.

Jaafar

, Jahanfar

, Angolkar

, et al. Effect of restricted pacifier use in breastfeeding term infants for increasing duration of breastfeeding. Cochrane Database Syst Rev, 2012; 7:CD007202.

Abrahams

, Labbok

. Exploring the impact of the Baby-Friendly Hospital Initiative on trends in exclusive breastfeeding. Int Breastfeed J, 2009; 4:11.

10.

Perrine

, Scanlon

, Li

, et al. Baby-Friendly hospital practices and meeting exclusive breastfeeding intention. Pediatrics, 2012; 130:54–60.

11.

The Baby-Friendly Hospital Initiative. Birth Gaz, 1998; 14:30.

12.

Centers for Disease Control and Prevention PRAMS Methodology. Available at www.cdc.gov/prams/methodology.htm (accessed May 6, 2016).

13.

Kramer

, Barr

, Dagenais

, et al. Pacifier use, early weaning, and cry/fuss behavior: A randomized controlled trial. JAMA, 2001; 286:322–326.

14.

Jenik

, Vain

, Gorestein

, et al. Does the recommendation to use a pacifier influence the prevalence of breastfeeding?. J Pediatr, 2009; 155:350–354.e351.

15.

Sipsma

, Nickel

, Jones

, Gilliam

. Hospital practices to promote breastfeeding: Does maternal age matter? Abstract presented at American Public Health Association Meeting, Chicago, IL, November 3, 2015. Available at https://apha.confex.com/apha/143am/webprogram/Paper336926.html

16.

Flaherman

, Gay

, Scott

, et al. Randomised trial comparing hand expression with breast pumping for mothers of term newborns feeding poorly. Arch Dis Child Fetal Neonatal Ed, 2012; 97:F18–F23.

17.

Flaherman

, Aby

, Burgos

, et al. Effect of early limited formula on duration and exclusivity of breastfeeding in at-risk infants: An RCT. Pediatrics, 2013; 131:1059–1065.

18.

Kair

, Colaizy

, Hubbard

, et al. Donor milk in the newborn nursery at the University of Iowa Children's Hospital. Breastfeed Med, 2014; 9:547–550.

19.

Flaherman

, Hicks

, Huynh

, et al. Positive and negative experiences of breast pumping during the first 6 months. Matern Child Nutr, 2016; 12:291–298.

20.

Kair

, Flaherman

, Newby

, et al. The experience of breastfeeding the late preterm infant: A qualitative study. Breastfeed Med, 2015; 10:102–106.

21.

Chapman

, Young

, Ferris

, et al. Impact of breast pumping on lactogenesis stage II after cesarean delivery: A randomized clinical trial. Pediatrics, 2001; 107:E94.

22.

Bai

, Wunderlich

, Kashdan

. Alternative hospital gift bags and breastfeeding exclusivity. ISRN Nutr, 2013; 2013:560810.