Mennella JA, Pepino MY. Biphasic effects of moderate drinking on prolactin during lactation. Alcoholism: Clinical and Experimental Research. 2008; 32: 1,899-1,908. Epub: 2008 Aug 18.

To explore the effects of alcohol on suckling-induced prolactin levels and milk yield, and to determine the effects of alcohol on basal prolactin and prolactin response during lactation.

• Female
• Lactating
• Exclusive breastfeeding of two- to five-month-old infants
• No resumption of menstruation
• Not pregnant throughout the study
• Blood glucose level less than 99mg per dL at the beginning of the study
• The University of Pennsylvania Office of Regulatory Affairs reviewed and approved the study. Participants gave written informed consent.

• Non-lactating female
• Non-exclusive breastfeeding of two- to five-month-old infants
• Resumption of menstruation
• Pregnant
• Obese
• Smoker
• Anemic
• Alcohol-dependent
• Lifetime alcohol abstainer
• Previous gastric bypass surgery
• On any medication including oral contraceptives
• Participation in only one of the four test sessions.

Recruitment

Recruitment procedures were not described.

Design

• The double-blind, two-by-two within-person, quasi-randomized trial consisted of four conditions:
  • Alcohol vs. no alcohol
  • Breast pumping once vs. twice
• Sessions were one week apart. The order of testing was counterbalanced so that the pumping condition alternated every two sessions, and the beverage alternated every session so that two weeks passed between alcohol conditions.

Blinding Used

Study nurses were blinded to the experimental conditions and participants were blinded to the alcohol conditions.

Intervention

• Participants completed the procedures at the same time for each session because prolactin concentration has a circadian rhythm. Women were asked not to drink alcohol for the three days prior to each session, and completed a pregnancy test prior to each session. During each session, participants were not allowed to sleep, talk about babies or food or watch television (because of possible exposure to babies or food) because of the possible effects on their hormones.
• After an overnight fast (verified by a fingerprick blood glucose assay), women completed baseline pumping to minimize discomfort. A nurse then inserted an intravenous line and the women acclimated in a private room for 45 minutes (needle pricks increase prolactin for 30 minutes).
• Blood samples were drawn at regular intervals before and after beverage consumption (-40, -25, -10, 25, 37, 39, 41, 43, 45, 47, 49, 51, 65, 80, 95, 110, 125, 140, 142, 144, 146, 148, 150, 152, 154, 156, 170, 185, 200 and 215 minutes). The beverage was divided into two equal volumes, each of which was to be consumed during consecutive five-minute intervals. The alcoholic beverage had a 0.4g per kg dose of ethanol in orange juice (15% v/v). The control beverage was an equal volume of orange juice. For both conditions, 3ml of alcohol was pipetted onto the cup surface to serve as a smell and taste mask.
• During the pump-once condition, women pumped 140 to 156 minutes after drinking. During the pump-twice condition, women pumped at 30 to 51 minutes and again at 140 to 156 minutes. The breast being pumped was alternated every two minutes for 16 minutes. Participants rated each breast's degree of fullness before and during each pumping session. Blood alcohol concentration (BAC) was also measured at fixed intervals during each session.
• After data collection finished at 215 minutes, subjects were provided lunch and were not allowed to leave the testing center until they had a BAC of zero.

Statistical Analysis

• BAC and prolactin (PRL) data preparation:
  • Peak BAC, time-to-peak BAC and alcohol disappearance rates (β₆₀; g per L per hour) were calculated using linear regression lines calculated within the apparent linear portion of the descending limb of each subject's BAC time curve. The first value was taken 30 minutes after peak BAC and all subsequent readings 0.20g per L or higher.
  • From each of the four sessions, peak PRL (mcg per L) and area under the PRL time curve (AUC; mcg per minute per L) was calculated for two time periods, 35 to 140 minutes and 140 to 215 minutes. AUC was determined using the trapezoidal method. PRL at 25 minutes was used as the baseline for delta scores for AUC at 35 to 140 minutes. PRL at 140 minutes was used as the baseline for delta scores for AUC at 140 to 215 minutes. Linear regression lines were calculated within the apparent linear portion of the ascending limb of the PRL rise for each time period for each beverage and pumping condition; correlation coefficients ranged from -0.75 to -1.00. One of six slopes could not be determined for one woman and three of six could not be determined for another.
• Hypothesis testing:
  • P<0.05 was used for all analyses
  • Slopes of the PRL rise were compared to determine whether alcohol consumption altered basal and suckling-induced PRL responses during the ascending and descending BAC
  • The effects of PRL response between feeds was determined by examining the effects of one pumping session on a later one. Alcohol effect modulation was also explored.
  • Separate repeated measures ANOVAs were run, using beverage and pumping conditions as well as time as within-subjects factors. Post-hoc Fisher's Least Significant Difference analyses were conducted for significant findings.
  • Secondary analyses explored effects of alcohol on milk yield, milk fat content and maternal perceptions of breast fullness. Repeated measures ANOVAs were used from the first pump session.
  • Correlations were used to identify associations between PRL, alcohol pharmacokinetics, and lactational performance.

Timing of Measurements

Blood samples were drawn before and after beverage consumption (-40, -25, -10, 25, 37, 39, 41, 43, 45, 47, 49, 51, 65, 80, 95, 110, 125, 140, 142, 144, 146, 148, 150, 152, 154, 156, 170, 185, 200 and 215 minutes). During the pump-once condition, women pumped 140 to 156 minutes after drinking. During the pump-twice condition, women pumped at 30 to 51 minutes and again at 140 to 156 minutes.

Dependent Variables

• Prolactin (PRL) concentration (mcg per L):
  • 2ml blood samples were drawn into containers with EDTA, kept on ice for up to an hour and centrifuged. Plasma aliquots were stored at -70°C. Plasma PRL was measured in duplicate with a direct two-site immunoradiometric assay.
  • PRL was measured in terms of baseline concentrations as well as suckling-induced response
• Milk yield (ml): Amount of milk pumped during each session
• Milk fat content (g per L): The creamatocrit method was used, where milk was vortexed, drawn into three microhematocrit tubes and centrifuged for 10 minutes. The height of the cream layer was measured and converted to fat by the formula, fat g per L = (mean creamatocrit percentage minus 0.59)/0.146
• Perceptions of breast fullness: Self-rated at the beginning and during pumping on a four-point scale (higher numbers mean more fullness).

Independent Variables

• Alcohol consumption (0.4g per kg dose vs. none)
  • BAC (g per L) measured by subjects breating into an Alco-Sensor IV at fixed intervals
  • BAC effects were assessed separately during ascending and descending BAC
• Pumping condition (pump once vs. twice).

• Initial N: 15
• Attrition (final N): 13; two women completed only one session and were excluded from the analyses
• Age: 33.1±1.4 years
• Ethnicity: Six Caucasian, three African American, four other or mixed ethnicity
• Anthropometrics: Mean BMI of 24.7±1.1kg/m²
• Location: University of Pennsylvania Clinical and Translational Research Center.

• BAC peaked around 42 minutes post-beverage at a concentration of 0.6±0.03g per L, and with a disappearing rate of 0.14±0.01g per L per hour
• Pumping significantly increased PRL. In the non-alcohol condition, PRL increased from 54.5±6.7mcg per L at baseline to 68.0±9.3mcg per L after 10 minutes of pumping and peaked at around 30 minutes at 106.7±15.9mcg per L. PRL levels were elevated for 75 minutes after pumping.
• Alcohol increased basal PRL compared to the no-alcohol condition (P<0.0001)
• During rising BAC, alcohol significantly enhanced PRL during breast pumping. The PRL slope was also steeper vs. the no alcohol condition (P=0.07)
• During falling BAC, alcohol significantly inhibited PRL during breast pumping, regardless of whether pumping was occurring for the first or second time. Peak PRL was significantly lower (P<0.05) and the PRL area under the curve at 140 to 215 minutes was significantly smaller (P=0.05). PRL slope was not significantly associated with beverage condition (P=0.16), pumping condition (P=0.12) or the interaction between the two (P=0.95).
• PRL dynamics were affected by both alcohol and pumping conditions
  • PRL levels were higher when women pumped for the second time vs. the first time. However, there were no differences in AUC at 140 to 215 minutes and slope of PRL rise between the pump conditions in the non-alcohol conditions. There were also no differences in the PRL time response.
  • Alcohol significantly delayed PRL response to pumping from eight minutes after pumping started to 12 minutes in the pump-twice condition or 14 minutes in the pump-once condition.

*P<0.05 for for same pumping condition, control beverage; **P≤0.05 for comparison with control beverage (both pumping conditions combined); ***P=0.07 for comparison with same pumping condition, control beverage; ****P<0.05 for comparison between pump-once (beverages combined) and pump-twice (beverages combined) (data not shown).

Other Findings

• Effects of time since pumping were assessed with the first pump session on each of the four days (35 to 51 minutes post-beverage to 120 minutes post-baseline pumping for pump-twice condition or 140 to 156 minutes post-beverage to 225 minutes post-baseline pumping for pump-once condition)
• Women reported feeling more breast fullness (1.4 vs. 1.0; P<0.001), expressed more milk (90.0 vs. 62.7ml; P=0.005), and had lower milk fat content (48.0 vs. 60.3g per L; P=0.06) in the pump-once condition vs. the pump-twice condition (i.e., longer interval since baseline pumping).

• Breast pumping increased plasma prolactin concentrations in lactating women with infants age two to five months, with PRL returning to baseline in 75 minutes
• Alcohol modified PRL effects in a transient and biphasic manner
  • Alcohol increased basal PRL within the hour after drinking, and the effects were stronger for lactating women than what has been observed previously with non-lactating women.
  • During ascending BAC, alcohol enhanced the PRL response to pumping, but during descending BAC, it inhibited the PRL response
  • Alcohol also altered PRL dynamics between pumpings. If a woman pumped within an hour of consuming alcohol, PRL response was delayed by a few minutes in a subsequent pumping session 1.5 hours later.
• Alcohol's mechanism of action on PRL is not understood. Since the slope of PRL rise was steeper during pumping shortly after alcohol consumption, alcohol may be acting synergistically with PRL regulatory mechanisms. It is also unclear whether the effects are due to changes in PRL production or to changes in its metabolic clearance.

• The authors did not describe any study limitations or potential biases
• As noted under study methods, randomization was done within fixed parameters for the two conditions so the condition order is only quasi-random. The study is currently class A, but if it were considered to be non-random assignment, it would be a class C study.
• Recruitment procedures were not described. Since the study was very intensive, it is likely that only highly motivated women participated. The participants are not likely to be representative of the population, but it is unclear how much of a concern this is since the study was of biological mechanisms rather than social or behavioral phenomena. For this reason, the selection of participants was determined to be free of bias on the research design and implementation checklist.