Treatments for suppression of lactation

Olufemi T Oladapo; Bukola Fawole

doi:10.1002/14651858.CD005937.pub3

. 2012 Sep 12;2012(9):CD005937. doi: 10.1002/14651858.CD005937.pub3

Treatments for suppression of lactation

Olufemi T Oladapo ^1,^✉, Bukola Fawole ²

Editor: Cochrane Pregnancy and Childbirth Group

PMCID: PMC6599849 PMID: 22972088

Abstract

Background

Various pharmacologic and non‐pharmacologic interventions have been used to suppress lactation after childbirth and relieve associated symptoms. Despite the large volume of literature on the subject, there is currently no universal guideline on the most appropriate approach for suppressing lactation in postpartum women.

Objectives

To evaluate the effectiveness and safety of interventions used for suppression of lactation in postpartum women (who have not breastfed or expressed breastmilk) to determine which approach has the greatest comparative benefits with least risk.

Search methods

We searched the Cochrane Pregnancy and Childbirth Group's Trials Register (30 June 2012).

Selection criteria

Randomised trials evaluating the effectiveness of treatments used for suppression of postpartum lactation.

Data collection and analysis

Two review authors independently assessed trial quality and extracted data.

Main results

We included 62 trials (6428 women). Twenty‐two trials did not contribute data to the meta‐analyses. The trials were generally small and of limited quality. Three trials (107 women) indicated that bromocriptine significantly reduced the proportion of women lactating compared with no treatment at or within seven days postpartum (three trials, 107 women; risk ratio (RR) 0.36, 95% confidence interval (CI) 0.24 to 0.54). Seven trials involving oestrogen preparations (diethylstilbestrol, quinestrol, chlorotrianisene, hexestrol) suggested that they significantly reduced the proportion of lactating women compared with no treatment at or within seven days postpartum (RR 0.40, 95% CI 0.29 to 0.56). We found no trials comparing non‐pharmacologic methods with no treatment. Trials comparing bromocriptine with other pharmacologic agents such as methergoline, prostaglandins, pyridoxine, carbegoline, diethylstilbestrol and cyclofenil suggested similarity in their effectiveness. Side effects were poorly reported in the trials and no case of thromboembolism was recorded in the four trials that reported it as an outcome.

Authors' conclusions

There is weak evidence that some pharmacologic treatments (most of which are currently unavailable to the public) are better than no treatment for suppressing lactation symptoms in the first postpartum week. No evidence currently exists to indicate whether non‐pharmacologic approaches are more effective than no treatment. Presently, there is insufficient evidence to address the side effects of methods employed for suppressing lactation. When women desire treatment, bromocriptine may be considered where it is registered for lactation suppression in those without predisposition to its major side effects of public concerns. Many trials did not contribute data that could be included in analyses. Large randomised trials are needed to compare the effectiveness of pharmacologic (especially bromocriptine) and non‐pharmacologic methods with no treatment. Such trials should consider the acceptability of the intervention and lactation symptoms of concern to women and be large enough to detect clinically important differences in major side effects between comparison groups.

Plain language summary

Treatments for suppression of lactation

Women cannot always breastfeed after birth. Reasons may be because the infant dies or is adopted, or the mother is too ill, or for the well being of the mother or infant. HIV‐positive mothers, particularly those not on antiretroviral drugs during pregnancy, avoid breastfeeding to reduce the risk of passing on the virus to their infants. Some mothers do not breastfeed on personal or social grounds. Without an infant suckling, milk production (lactation) eventually stops of its own accord. In the meantime, women can experience breast engorgement, leakage of milk, discomfort and pain. Clinicians may provide treatment to suppress lactation and reduce these symptoms. Binding the breasts or wearing a tight brassiere, applying an infra‐red lamp, fluid and diet restrictions, external application of jasmine flower and ice packs are tried non‐drug approaches. Drug treatments include oestrogens and bromocriptine which lowers prolactin levels. However, increased risks of thromboembolism, cerebral accident and myocardial infarction have been reported with their use.

The evidence to support treatments for preventing lactation is limited. The review authors identified 62 controlled trials that randomised a total of 6428 mothers to receive the treatment under investigation, no treatment or another treatment. Twenty‐two trials did not contribute data to the meta‐analyses. The trials were generally of limited quality and most were conducted among healthy women who chose not to breastfeed for personal reasons at hospitals in industrialised countries before 1980. Half of the trials involved bromocriptine. Two trials (107 women) reported that taking bromocriptine was better than no treatment in suppressing lactation in the first week after giving birth. The 11 trials using oestrogen preparations (diethylstilbestrol, quinestrol, chlorotrianisene, hexestrol) also showed suppression of lactation. A combination of testosterone and oestrogen preparations was of some benefit in reducing symptoms in three trials (436 women). Other pharmacologic agents (clomiphene, tamoxifen, prostaglandins, pyridoxine, oxytocin, L‐dopa and homeopathic preparation) were tested in single small trials. Generally, side effects were poorly reported and no case of thromboembolism was recorded among trials that included it as an adverse treatment outcome. Most of the drugs tested are currently not available or registered for suppressing lactation. No trials compared non‐drug approaches with no treatment and none of the included trials provided reliable data on women’s satisfaction with the treatment.

Background

Description of the condition

Indications for lactation suppression

For many years, the importance of breastfeeding to both infants and mothers has been emphasised by healthcare providers and various strategies have been employed to promote it globally. In spite of the well‐known advantages of breastfeeding (for example, infant protection against diarrhoeal morbidity and mortality), there are instances when the well being of the mother or infant requires suppression of lactation. Suppression of lactation becomes essential when breastfeeding is no longer required (as in the events of perinatal death and infant adoption) or when the mother is too ill to breastfeed (as in cases of severe obstetric morbidity). Besides medical indications, some mothers in circumstances where alternatives to breastfeeding exist may seek lactation suppression on personal or social grounds. It is estimated that over 30% of women in the United States and United Kingdom do not breastfeed their infants, while a larger proportion discontinue breastfeeding within two weeks of childbirth (Hamlyn 2002; Ryan 2002). Although physiologic cessation of lactation eventually occurs in the absence of physical stimulus such as infant suckling, a variable proportion of women experience moderate to severe milk leakage and discomfort, before lactation ceases. Up to two‐thirds of non‐breastfeeding women experience moderate to severe engorgement and breast pain when no treatment is applied (Spitz 1998). Almeida and Kitay (Almeida 1986) showed that breast engorgement was responsible for puerperal fever in 13.3% of 75 non‐breastfeeding mothers. However, the prevalence, characteristics and health implications of these symptoms have not been well described in the literature.

Lactation suppression and prevention of vertical transmission of HIV

Unlike in the 1970s, when a social reason was the most common indication for lactation suppression (Eastham 1976), the need for complete avoidance of breastfeeding by HIV‐positive mothers to reduce the risk of vertical transmission of HIV has offered a more compelling reason in the last decade. Postnatal transmission through breastfeeding accounts for one‐third to one‐half of all cases of vertical HIV transmission worldwide, with an estimated 16.2% rate of transmission for infants of women untreated with antiretroviral drugs during pregnancy (Nduati 2000). Although breastfed infants of HIV‐positive mothers who receive antiretroviral treatment during pregnancy are less likely to be infected with HIV (Wiktor 1999), the risk is further reduced when such infants are fed with substitutes of breastmilk (Shaffer 1999). Therefore, as the global prevalence of HIV continues to rise, the need for supervised inhibition of lactation may likely become increasingly relevant, especially in developed countries where safe alternatives of infant feeding are available. The symptoms associated with physiologic cessation of lactation may further compromise the physical and emotional status of the HIV‐positive mothers and an effective method of suppressing lactation is desirable to avoid additional morbidity.

Description of the intervention

Non‐pharmacologic methods of lactation suppression

Interventions to suppress lactation in non‐breastfeeding women have evolved for centuries. Healthcare providers have used different non‐pharmacologic approaches to suppress lactation and relieve the associated symptoms. Before the 20th century, these approaches included breast binding or strapping, emptying of the breast by massage, fluid and diet restrictions and application of external products such as belladonna ointment to the breast and nipples. Later, the avoidance of tactile breast stimulation and application of external agents such as cabbage leaves, jasmine flower and ice packs were included. Although many of these methods are still in use today, data on their efficacy are few and inconclusive. A review by Spitz et al (Spitz 1998) showed that up to one‐third of women may experience severe breast pain for most of the first postpartum week when these methods of lactation suppression are employed.

Pharmacologic methods of lactation suppression

In the 1960s, oestrogen preparations given alone or in combination with androgens were demonstrated to be effective in 40% to 100% of women (Llewellyn‐Jones 1968; Senior 1969) but their reported association with a high rate of rebound lactation (resurgence of lactation following cessation of treatment) and increased risk of thrombosis and pulmonary embolism discouraged their use (Jeffcoate 1968). After it was demonstrated that postpartum lactation depends primarily on pituitary prolactin secretion, the synthetic dopamine agonist and strong prolactin inhibitor bromocriptine was introduced in 1972. Its efficacy in the suppression of postpartum lactation is well documented (Bhardwaj 1979; Dewhurst 1977; Duchesne 1981; Van der Heijden 1991). It is, however, associated with some unpleasant side effects and requires administration for about 10 to 14 days to prevent rebound lactation. It has also been implicated in serious puerperal complications such as cerebral accident and myocardial infarction (Iffy 1996; Ruch 1989). In 1989, the United States Food and Drug Administration recommended against the routine use of bromocriptine for suppression of postpartum lactation, noting that while there was no clear proof of adverse effects, there were also no proven health benefits (US FDA 1989). In spite of this development, bromocriptine is still being used in many countries. Since then, many other drugs have been used for suppression of lactation, including those with recognised prolactin‐lowering activity and those with uncertain mechanism of action. These include different preparations of oestrogens, oestrogens in combination with androgens or progestogens, or both, clomiphene, pyridoxine, prostaglandin E₂, other dopamine agonists (cabergoline and lisuride) and serotonin antagonists (cyproheptadine, methysergide and methergoline). All these drugs have demonstrated variable effectiveness in the inhibition of postpartum lactation.

The search for another ergot derivative with clinical efficacy similar to bromocriptine but with better compliance and tolerability profile led to the trials on cabergoline. Various randomised studies in the late 1980s described similar effectiveness and a better side‐effect profile of cabergoline, administered as a single dose compared with the conventional bromocriptine dose in the prevention of postpartum lactation (European 1991; Giorda 1991). Recent reports suggest that a new drug, which belongs to the sulfhydryl compound, is also an effective inhibitor of lactation and breast engorgement (Akrivis 2000).

Why it is important to do this review

It is clear from this background that the evolution of lactation suppressants is not over. Thus, while the need for specific medical prevention of lactation in non‐nursing mothers is being questioned from time to time, many clinicians still apply some kind of treatment. Besides, clinicians appear unclear about the most appropriate method for suppressing lactation when an intervention is indicated. In view of the numerous approaches of lactation inhibition and the continuous search for new drugs, it becomes necessary to synthesise previous research findings to determine the most effective intervention to suppress lactation in non‐breastfeeding mothers. An ideal method would be one that has close to 100% efficacy, with minimal or no side effects and good acceptability profile. A systematic review of the previous studies was therefore conducted to understand whether further trials on new approaches or drugs, or both, are justified.

Objectives

The objective of this review was to evaluate the effectiveness and safety of interventions used for suppression of lactation in postpartum women (who have not breastfed or expressed breastmilk) to determine which approach has the greatest comparative benefits with least risk.

Methods

Criteria for considering studies for this review

Types of studies

All published randomised trials evaluating the effectiveness of treatments used for suppression of postpartum lactation. We excluded studies that evaluated the effectiveness of interventions after establishment of lactation (e.g. women who have breastfed or expressed breastmilk). We excluded quasi‐randomised trials (e.g. allocation by date of birth or hospital record number).

Types of participants

Postpartum women (who have not breastfed or expressed breastmilk) with indication(s) for suppression of lactation, irrespective of parity and mode of delivery.

Types of interventions

We assessed pharmacologic (drug) and non‐pharmacologic (breast binding or strapping, firm breast support, fluid restriction, application of ice packs and external products) interventions specifically aimed at suppressing lactation after childbirth.

We considered the following comparisons in this review:

any pharmacologic treatment versus no treatment or placebo;
any non‐pharmacologic treatment versus no treatment or placebo;
comparison of two different non‐pharmacologic treatments;
comparison of non‐pharmacologic versus pharmacologic treatments;
comparison of two different pharmacologic treatments;
comparison of two different pharmacologic combinations;
comparison of different doses of the same agent.

We excluded studies without any of the above comparisons.

Types of outcome measures

Primary outcomes

Failure to suppress lactation as indicated by breast pain, engorgement and or milk secretion (or as described by trial authors) at or within seven days postpartum and at or within 14 days postpartum. In trials where data for two or more breast symptoms or signs were reported, data for failure to suppress lactation were derived from the least suppressed of these symptoms or signs.
Minor adverse events including nausea, vomiting, headache, dizziness and major adverse events including thromboembolism, myocardial infarction and maternal death.
Acceptability of the treatment to the woman.

Secondary outcomes

Rebound lactation (resurgence of lactation after cessation of suppressant).
Percentage of women who require a second line drug or method, or both, to achieve lactation suppression.
Percentage of women who require analgesics to relieve breast pain or discomfort.

Search methods for identification of studies

Electronic searches

We searched the Cochrane Pregnancy and Childbirth Group’s Trials Register by contacting the Trials Search Co‐ordinator (30 June 2012).

The Cochrane Pregnancy and Childbirth Group’s Trials Register is maintained by the Trials Search Co‐ordinator and contains trials identified from:

monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL);
weekly searches of MEDLINE;
weekly searches of EMBASE;
handsearches of 30 journals and the proceedings of major conferences;
weekly current awareness alerts for a further 44 journals plus monthly BioMed Central email alerts.

Details of the search strategies for CENTRAL, MEDLINE and EMBASE, the list of handsearched journals and conference proceedings, and the list of journals reviewed via the current awareness service can be found in the ‘Specialized Register’ section within the editorial information about the Cochrane Pregnancy and Childbirth Group.

Trials identified through the searching activities described above are each assigned to a review topic (or topics). The Trials Search Co‐ordinator searches the register for each review using the topic list rather than keywords.

We did not apply any language restrictions.

Data collection and analysis

Selection of studies

Two review authors independently assessed for inclusion all the potential studies identified as a result of the search strategy. Disagreements were resolved through discussion.

Data extraction and management

We designed a form to extract data. For eligible studies, both review authors independently extracted the data using the agreed form. We resolved discrepancies through discussion. We entered data into Review Manager software (RevMan 2011) and checked them for accuracy.

When information regarding any of the above was unclear, we attempted to contact authors of the original reports to provide further details. There was no masking of authors or journals.

Assessment of risk of bias in included studies

We independently assessed risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We resolved any disagreements by discussion.

(1) Random sequence generation (checking for possible selection bias)

We described for each included study the method used to generate the allocation sequence in sufficient detail to allow an assessment of whether it should produce comparable groups.

We assessed the method as:

low risk of bias (any truly random process, e.g. random number table; computer random number generator);
high risk of bias (any non‐random process, e.g. odd or even date of birth; hospital or clinic record number);
unclear risk of bias.

(2) Allocation concealment (checking for possible selection bias)

We described for each included study the method used to conceal the allocation sequence in sufficient detail and determined whether intervention allocation could have been foreseen in advance of, or during recruitment, or changed after assignment.

We assessed the methods as:

low risk of bias (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);
high risk of bias (open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth);
unclear risk of bias.

(3) Blinding (checking for possible performance and detection bias)

We described for each included study the methods used, if any, to blind study participants and key personnel (including outcome assessors) from knowledge of which intervention a participant received. Blinding was assessed separately for different classes of outcomes.

We assessed the methods as:

low risk of bias (blinding of participants and key study personnel including outcome assessors ensured, and unlikely that the blinding could have been broken; participants and key study personnel not blinded but outcome assessors blinded);
high risk of bias (no blinding or incomplete blinding; likely that blinding of participants and key study personnel could have been broken);
unclear risk of bias.

(4) Incomplete outcome data (checking for possible attrition bias through withdrawals, dropouts, protocol deviations)

We described for each included study, and for each class of outcomes, the completeness of data including attrition and exclusions from the analysis. We stated whether attrition and exclusions were reported, the numbers included in the analysis at each stage (compared with the total randomised participants), reasons for attrition or exclusion where reported, and whether missing data were balanced across groups.

We assessed the methods as:

low risk of bias (e.g. no missing data; missing data for less than 20% of randomised participants; balanced missing outcome data across groups; reason for missing data unrelated to true outcome; appropriate imputing of missing data; intention‐to‐treat analysis);
high risk of bias (e.g. reason for missing data related to true outcome; missing data for more than 20% of randomised participants; unbalanced missing outcome data across groups; ‘as treated’ analysis);
unclear risk of bias.

(5) Selective reporting bias

We described for each included study how we investigated the possibility of selective outcome reporting bias and what we found.

We assessed the methods as:

low risk of bias (where it is clear that all of the study’s prespecified outcomes and all expected outcomes of interest to the review have been reported);
high risk of bias (where all the study’s prespecified outcomes have not been reported; one or more reported primary outcomes were not prespecified; outcomes of interest are reported incompletely and so cannot be used; study fails to include results of a key outcome that would have been expected to have been reported);
unclear risk of bias.

(6) Other sources of bias

We described for each included study any important concerns we had about other possible sources of bias.

We assessed whether each study was apparently free of other problems that could put it at risk of bias and indicated this as:

low risk of other bias;
high risk of other bias;
unclear whether there is risk of other bias.

Measures of treatment effect

Dichotomous data

For dichotomous data, we presented results as summary risk ratio with 95% confidence intervals.

We carried out statistical analysis using RevMan 2011. We used fixed‐effect meta‐analysis for combining data in the absence of significant heterogeneity if trials were sufficiently similar. Where there was significant heterogeneity, we used a random‐effects model.

Dealing with missing data

For included studies, we noted levels of attrition. We planned but could not explore the impact of including studies with high levels of missing data in the overall assessment of treatment effect using sensitivity analysis as few trials which addressed different interventions had missing data greater than 20%.

For all outcomes we carried out analyses, as far as possible, on an intention‐to‐treat basis i.e. we attempted to include all participants randomised to each group in the analyses. The denominator for each outcome in each trial was the number randomised minus any participants whose outcomes were known to be missing.

Assessment of heterogeneity

We used the I² statistic to measure heterogeneity among the trials in each analysis. We planned but could not explore substantial heterogeneity (exceeding 50%) by subgroup analysis as few studies reported the criterion (gestational age) that was pre‐specified in the protocol.

Data synthesis

We carried out statistical analysis using the Review Manager software (RevMan 2011). We used fixed‐effect inverse variance meta‐analysis for combining data where trials are examining the same intervention, and the trials’ populations and methods are judged sufficiently similar. Where we suspected clinical or methodological heterogeneity between studies sufficient to suggest that treatment effects may differ between trials, we used random‐effects meta‐analysis. For studies with results in formats that cannot be included in meta‐analysis, findings were succinctly reported as described by the trialists without attempting to derive a summary effect estimate.

Where substantial heterogeneity was identified in a fixed‐effect meta‐analysis, we repeated the analysis using a random‐effects method.

Subgroup analysis and investigation of heterogeneity

There was no subgroup analysis as few studies reported the criterion (gestational age) that was pre‐specified in the protocol.

Sensitivity analysis

We planned but could not conduct sensitivity analyses to explore the effect of trial quality as most of the trials were generally of low and similar methodological quality and in addition addressed different interventions.

Results

Description of studies

Out of 132 potentially relevant studies considered for inclusion in this review, 62 trials involving 6428 women met our inclusion criteria, with outcome data on efficacy variables for interventions available for 6137 women. Twenty‐two trials did not contribute data that could be included in meta‐analyses (Berrebi 2001; Bhardwaj 1979; Biggs 1978; Binns 1967; Caballero‐Gordo 1991; Cooke 1976; Defoort 1987; Hutchison 1981; Kremer 1990; Kulski 1978; Mann 1971; Martinez 1994; McNicol 1972; Menczer 1969; Mizuno 1990; Niebyl 1979; O'Donoghue 1977; Paggi 1975; Phillips 1975; Swift 2002; Van der Heijden 1991; Winter 1964). Two trials (Caballero 1996; Varga 1974) are awaiting further assessment. All of the included trials are relatively small (the largest trial having 800 women) and 42 of them have fewer than 100 participants. The majority (52/62) of the trials were conducted in industrialised countries and all but six trials were single‐centre studies conducted in a private, general or university hospital. Thirty‐eight of the trials were conducted before 1980 and only four after 2000. Seven of the trials were published in languages other than English.

Participants

Most of the trials included healthy postpartum women who elected not to breastfeed for personal reasons and a few recruited women who could not breastfeed as a result of stillbirth or child adoption. Participants were women who delivered vaginally at term in most cases. Giorda 1991 randomised only women who were delivered by caesarean section while Piya‐Anant 2004 randomised healthy HIV‐positive puerperal women. Exclusion criteria were not specified in many of the trials. In the few trials that specified exclusion criteria, these include abnormal findings that are relevant to prolactin secretion, liver disorders, agalactia (inability to lactate), previous breast surgery, tumour of the pituitary gland, use of drugs that might interfere with prolactin secretion and unwillingness to participate in the study. Recruitment of women was not limited to any parity group in any of the trials and parity ranged between one and eight in studies where it was reported.

Interventions

Pharmacologic treatments were compared with placebo in 29 trials and another pharmacologic treatment in 30 trials. No trial was included for the comparison of non‐pharmacologic treatments with placebo. Three trials compared non‐pharmacologic approaches [Jasmine flower (Shrivastav 1988), tight brassiere with intermittent infrared application (De Gezelle 1979), breast binding (Bhardwaj 1979)] with pharmacologic treatments. Only one trial compared two non‐pharmacologic treatments (breast binding versus wearing of support bra) (Swift 2002).

The classes of pharmacologic agents evaluated in these trials include ergot derivatives (bromocriptine, lisuride, methergoline, cabergoline and terguride); synthetic oestrogen preparations (quinestrol, diethylstilbestrol, ethinyl estradiol, chlorotrianisene and hexestrol); antioestrogenic preparations (tamoxifen, clomiphene and cyclofenil); oxytocics (intranasal oxytocin); androgen preparations (testosterone propionate); combined oestrogen and androgen preparations (testosterone and estradiol esters); combined oestrogen, progestogen and androgen preparations; dopamine agonists/precursors; prostaglandins and pyridoxine. Thirty‐one trials involved bromocriptine. Bromocriptine was compared with placebo in nine trials (Bhardwaj 1979; Biggs 1978; Cooke 1976; Dewhurst 1977; Hutchison 1981; Kulski 1978; Rolland 1973; Walker 1975; Weinstein 1976), other pharmacologic treatments in 20 trials (Biggs 1978; Boes 1980; Defoort 1987; England 1988; European 1991; Fischer 1995; Giorda 1991; Kremer 1990; Nilsen 1976; O'Donoghue 1977; Piya‐Anant 2004; Purkayastha 1991; Scapin 1982; Steenstrup 1977; Thorbert 1983; Utian 1975; Van der Heijden 1991; Varga 1972; Venturini 1981; Yuen 1977) and non‐pharmacologic methods in three trials (Bhardwaj 1979; De Gezelle 1979; Shrivastav 1988).

Most of the pharmacologic agents were given orally and a few were given by intramuscular injections. No agent was given intravenously. Dosage varied widely between trials in both amount and duration of treatment. Except in two trials (England 1988; Giorda 1991), treatments were commenced shortly after delivery (less than 12 to 24 hours). Of the drugs tested, only seven are currently included in the World Health Organization model list of essential medicines. These are ethinyl estradiol, clomiphene, tamoxifen, pyridoxine, oxytocin, testosterone and L‐dopa. None of these drugs was listed for lactation suppression.

Outcomes

The definition of lactation suppression was not consistent among the trials. The primary outcome measures and the descriptions used in most of the trials were suppression of milk secretion or leakage, breast engorgement and/or breast pain. Some trials did not describe what was meant by suppression of lactation and only referred to it as such. For most of the trials, failure to achieve suppression of between one and three breast symptoms or signs (milk secretion or leakage, breast pain, breast engorgement) was presented as a measure of treatment failure. Other outcome measures prespecified for the review such as major and minor side effects, rebound lactation, use of analgesics and need for secondary treatment to achieve suppression were poorly reported in many of the trials. The method of outcome assessments varied widely across studies. In 15 trials, breast symptoms or signs were rated on a visual analogue scale or an ordinal scale of zero to between three and six to describe increasing severity of symptoms or signs (e.g., none, mild, moderate and severe). Most trials used dichotomous variables, or ones that could be dichotomised, to describe clinical efficacy of treatments while 10 trials described outcome measures in terms of means of the breast symptom 'scores', mean number of days or mean change in the degree of breast symptoms. There was no evidence that any of these scoring systems were previously validated. Assessments of clinical efficacy were based on physical examination of the breast by the clinicians or women in the trials, or both, while in the hospital and subsequently by the women at home after hospital discharge. Women were asked to document breast symptoms on a questionnaire or data card that was collected at a specified follow‐up period. The duration and the frequency of outcome assessment also varied widely between studies. Follow‐up varied from 72 hours to eight months postpartum although most assessments were conducted during the first one or two weeks.

As a result of the diversity of the interventions and the method, duration and frequency of outcome assessments, few trials could be included in meta‐analyses for each comparison. It was not possible to conduct sensitivity analyses based on trial quality as most trials were generally at high risk of bias (e.g., allocation concealment was considered adequate in 8.1% of the included trials, all of which evaluated different interventions).

Risk of bias in included studies

Overall, the risk of bias for most reports was uncertain as they contained little methodological description (Figure 1 and Figure 2). Details for each trial are given in the Characteristics of included studies table.

'Risk of bias' graph: review authors' judgments about each risk of bias item presented as percentages across all included studies.

'Risk of bias' summary: review authors' judgments about each risk of bias item for each included study.

Allocation

The risk of selection bias was uncertain for most trials as generation and concealment of allocation sequence were considered adequate in only a few of the included studies. Generation of allocation sequence was considered adequate in seven trials (computer randomisation in three trials (Kremer 1990; McNicol 1972; Piya‐Anant 2004) and use of table of random numbers in four trials (Giorda 1991; Kulski 1978; Niebyl 1979; Swift 2002). Other studies did not describe the method by which random allocation sequence was generated and only reported that participants were "randomised" or "randomly allocated" into treatment groups. Allocation concealment was considered adequate in only five (8.1%) of the included studies (Biggs 1978; Defoort 1987; Kremer 1990; MacDonald 1976; Watson 1969). The majority did not provide adequate information to permit appropriate judgement on allocation concealment, while three trials (Bravo‐Topete 2004; Rolland 1973; Van der Heijden 1991) did not conceal their allocation sequences.

Blinding

With regard to blinding of participants, key study personnel and outcome assessors, we considered 26 trials to be at low risk and eight trials at high risk of bias for the main outcome (lactation) of the review. The risk of bias was uncertain for 28 trials, principally as a result of unclear description of blinding of outcome assessors. Eight of these trials only described their blinding method as "double blinded" with no further information to permit appropriate judgements. Of the 27 trials that addressed adverse events as outcome measures, we considered 12 to be at low risk of bias, five at high risk of bias and the rest at uncertain risk of bias.

Incomplete outcome data

For the main outcome (lactation), we considered 47 trials to be at low risk of attrition bias, seven trials at high risk of bias and eight trials at uncertain risk of bias. Loss to follow‐up or exclusions from the initial cohort entered into the trials varied between 0% and 46.5%. For adverse events, 28 trials adequately addressed incomplete outcome data. It was unclear whether incomplete data were adequately addressed in 34 trials; 31 of these did not address adverse events as outcome measures.

Selective reporting

We considered 36 trials to be at low risk of selective reporting bias, 19 trials at high risk and seven at uncertain risk of bias. In most of the reports at high risk of bias, the trialists did not include results for adverse events especially thromboembolism in spite of prior studies implicating oestrogen preparations in this complication.

Other potential sources of bias

Forty‐one trials were apparently free of other problems that could put them at a risk of bias. We considered six trials to be at high risk of other sources of bias. These sources of bias included extreme imbalance in the number of participants across groups (Caballero‐Gordo 1991), statistical analyses by supplier of tested intervention (Cruttenden 1971; Hutchison 1981, McNicol 1972) and prolonged data collection period involving 25 study personnel and outcome assessors (Winter 1964).

Effects of interventions

Primary outcomes

Pharmacologic treatment versus no treatment or placebo

Ergot derivatives versus no treatment or placebo (comparison 1)

Five trials (206 women) compared bromocriptine with placebo during the first postpartum week (Dewhurst 1977; Hutchison 1981; Rolland 1973; Walker 1975; Weinstein 1976), although only three reported data that could be used in meta‐analysis. These three trials (107 women) (Dewhurst 1977; Rolland 1973; Weinstein 1976) indicated that bromocriptine reduced the risk of failure of lactation suppression during the first seven days postpartum when compared with placebo (risk ratio (RR) 0.36, 95% confidence interval (CI) 0.24 to 0.54) (Analysis 1.1). The two other trials, which did not contribute data to the analysis because they were not in a format suitable for inclusion, indicated that bromocriptine resulted in significantly less milk secretion and breast engorgement compared to placebo during the first seven days postpartum (Hutchison 1981; Walker 1975).

1.1 — Comparison 1 Ergot derivatives versus placebo, Outcome 1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days postpartum.

Six trials (258 women) reported the cumulative efficacies of bromocriptine compared to placebo over 14 days postpartum (Bhardwaj 1979; Binns 1967; Cooke 1976; Dewhurst 1977; Kulski 1978; Varga 1972), although only two reported data that could be used in meta‐analysis. These two trials (76 women) (Dewhurst 1977; Varga 1972) indicated that this ergot derivative was similar to placebo in the suppression of lactation (RR 0.18, 95% CI 0.03 to 1.08; Analysis 1.2). However, the four other trials (Bhardwaj 1979; Binns 1967; Cooke 1976; Kulski 1978) that did not provide data in a suitable format for inclusion in an analysis, indicated that bromocriptine resulted in significantly less milk secretion, breast engorgement and discomfort compared to placebo.

1.2 — Comparison 1 Ergot derivatives versus placebo, Outcome 2 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 14 days.

Melis 1988 (32 women) indicated that cabergoline reduced the risk of failure to suppress lactation compared to placebo at 14 days postpartum (RR 0.19, 95% CI 0.07 to 0.48; Analysis 1.3).

1.3 — Comparison 1 Ergot derivatives versus placebo, Outcome 3 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 14 days.

Oestrogen preparations versus no treatment or placebo (comparison 2)

There were 11 trials in which oestrogen preparations were compared with placebo. Seven of the trials where nine comparisons were made are presented in the data and analysis table. Diethylstilbestrol was compared with placebo in a total of five trials. Four of these trials (376 women) (Schwartz 1973; Senior 1969; Stirrat 1968; Weinstein 1976), which are included the analyses, indicate a reduced risk of treatment failure when diethylstilbestrol was compared with placebo at or within seven days postpartum (RR 0.33, 95% CI 0.12 to 0.89); Analysis 2.1.1). This analysis indicates significant heterogeneity among the trials as evident by the very high I² statistic. Although data from Winter 1964 (800 women) did not contribute to the analysis because they were not in a suitable format for inclusion, the report indicated that milk secretion, breast congestion and breast pain affected fewer women who were treated with diethylstilbestrol compared to those treated with placebo during the first postpartum week.

2.1 — Comparison 2 Oestrogen preparations versus placebo, Outcome 1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days.

Three trials (342 women) (Firth 1969; McGlone 1969; Vischi 1975) comparing quinestrol with placebo suggested that quinestrol was associated with less risk of treatment failure (RR 0.47, 95% CI 0.30 to 0.73; Analysis 2.1.2). The I² statistic of 93%, however, suggests significant heterogeneity among the trials. Schwartz 1973 (153 women) indicated that chlorotrianisene was associated with less risk of treatment failure when compared with placebo (RR 0.51, 95% CI 0.36 to 0.73; Analysis 2.1.3). Two trials (Binns 1967 (85 women); Phillips 1975 (196 women)), which did not contribute data to the analysis because they were not in a suitable format for inclusion, reported results that were in agreement with the results of this analysis (Analysis 2.1.3). However, another trial (Niebyl 1979 (99 women), which again did not provide data in a suitable format for inclusion in an analysis, indicated similarity in the risks of treatment failure between chlorotrianisene compared with placebo at day three postpartum. One trial (100 women) (Firth 1969) comparing hexestrol with placebo suggested that it was significantly associated with less risk of treatment failure (RR 0.41, 95% 0.29 to 0.57; Analysis 2.1.4), at or within seven days postpartum.

Overall, seven trials (971 women) (Firth 1969; McGlone 1969; Schwartz 1973; Senior 1969; Stirrat 1968; Vischi 1975; Weinstein 1976) with nine comparisons indicated that oestrogen preparations are associated with reduced risk of failure to suppress lactation when compared with placebo at or less than seven days postpartum (RR 0.40, 95% CI 0.29 to 0.56; Analysis 2.1).

Two small trials (Cruttenden 1971; Varga 1972) suggested similarity in the risk of treatment failure between oestrogen preparations (quinestrol and diethylstilbestrol) and placebo up to day 14 postpartum (RR 0.19, 95% CI 0.03 to 1.34; Analysis 2.2). Two trials (130 women), which did not provide data in a suitable format for inclusion in an analysis (Stirrat 1968; Weinstein 1976), reported thromboembolism as an outcome measure and reported no occurrence of this complication in women treated with diethylstilbestrol and placebo.

2.2 — Comparison 2 Oestrogen preparations versus placebo, Outcome 2 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 14 days.

Antioestrogens versus no treatment or placebo (comparison 3)

One trial (30 women) (Weinstein 1976) suggested no difference between clomiphene and placebo (RR 1.00, 95% CI 0.51 to 1.95; Analysis 3.1) when used for lactation suppression in the first postpartum week. The trial also reported no occurrence of thromboembolism in women treated with clomiphene and placebo.

3.1 — Comparison 3 Antioestrogen preparations versus placebo, Outcome 1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days.

In the Shaaban 1975 trial (140 women), tamoxifen was significantly less likely to be associated with failure to suppress lactation compared to placebo over a period of 14 days postpartum (RR 0.71, 95% CI 0.62 to 0.82; Analysis 3.2). The trial reported no occurrence of thromboembolism in women treated with tamoxifen and placebo.

3.2 — Comparison 3 Antioestrogen preparations versus placebo, Outcome 2 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 14 days.

Pyridoxine versus no treatment or placebo (comparison 4)

MacDonald 1976 and Marcus 1975 (258 women) indicated that the risk of failure to achieve lactation suppression at or within seven days postpartum was similar between pyridoxine and placebo (RR 0.98, 95% CI 0.89 to 1.09; Analysis 4.1).

4.1 — Comparison 4 Pyridoxine versus placebo, Outcome 1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days.

Combined oestrogen and androgen preparations versus no treatment or placebo (comparison 5)

Three trials (436 women) (Iliya 1966; Marcus 1975; Schwartz 1973) compared testosterone enanthate‐estradiol valerate combination with placebo for inhibition of lactation. This combination was found to significantly reduce the risk of treatment failure when compared with placebo at or within seven days postpartum (RR 0.15, 95% CI 0.10 to 0.22; Analysis 5.1).

5.1 — Comparison 5 Combined oestrogen and androgen preparations versus placebo, Outcome 1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days.

Androgen preparations versus no treatment or placebo (comparison 6)

In the comparison of testosterone propionate with placebo, a small trial (30 women) (Weinstein 1976) indicated no evidence of an association between testosterone propionate and risk of treatment failure (RR 1.13 95% 0.60 to 2.11; Analysis 6.1). There were no reports of thromboembolism in women who received testosterone propionate and placebo. Another trial that did not provide data in a suitable format for inclusion in an analysis (Biggs 1978), indicated similarity regarding breast discomfort, congestion or milk production between methyl testosterone and placebo at two weeks postpartum.

6.1 — Comparison 6 Androgen preparations versus placebo, Outcome 1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days.

Prostaglandins versus no treatment or placebo

One trial provided data for this comparison, but they were not in a suitable format for inclusion in an analysis. Tulandi 1985 compared prostaglandin E₂ with placebo in a trial involving 16 women. The trial indicated similarity in the degree of milk leakage, breast engorgement and breast pain between women who received prostaglandin E₂ and those who received placebo.

Oxytocics versus no treatment or placebo

One trial provided data for this comparison, but they were not in a suitable format for inclusion in an analysis. One trial (98 women) (Winter 1964) indicated that intranasal oxytocin was similar to placebo in the suppression of lactation symptoms.

Dopamine precursors versus no treatment or placebo

One trial provided data for this comparison, but they were not in a suitable format for inclusion in an analysis. Paggi 1975 compared L‐dopa with placebo in a trial involving 40 women. The trial indicated a lower risk of treatment failure among women who received L‐dopa compared to those that received placebo. Although the treatment was reported to last for six days after delivery, the timing of outcome assessment was not specified.

Homeopathic preparations versus no treatment or placebo

One trial provided data for this comparison, but they were not in a suitable format for inclusion in an analysis. Berrebi 2001 (71 women) suggested a lower risk of treatment failure when homeopathic preparation (with anti‐inflammatory and analgesic properties) was compared with placebo on days two and four postpartum.

Non‐pharmacologic treatments versus no treatment or placebo

No trial was included for this comparison.

Pharmacologic treatments versus non‐pharmacologic treatments (comparison 7)

Only three trials compared a pharmacologic agent with non‐pharmacologic agent. De Gezelle 1979 (90 women) was a five‐arm study, which had bendrofluazide (a diuretic), diethylstilbestrol, estradiol/testosterone ester and bromocriptine in the intervention arms and tight brassiere with intermittent application of infra‐red lamp as the control arm. The trial suggested that the risks of treatment failure at or within seven days postpartum were significantly reduced by all the studied pharmacologic treatments compared to wearing of tight brassiere and application of infra‐red lamp (Analysis 7.1). Two trials did not provide data in a suitable format for inclusion in an analysis (Bhardwaj 1979; Shrivastav 1988). Shrivastav 1988 (60 women) indicated that the risk of treatment failure was similar between women who used bromocriptine and those who applied jasmine flowers to the breasts for suppression of postpartum lactation. Bhardwaj 1979 (20 women) suggested that the risk of treatment failure was significantly reduced by bromocriptine compared to breast binding when used over a period of 14 postpartum days.

7.1 — Comparison 7 Pharmacologic treatment versus nonpharmacologic treatment, Outcome 1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain.

Comparison of two non‐pharmacologic treatments

One trial provided data for this comparison, but they were not in a suitable format for inclusion in an analysis. Swift 2002 indicated that breast binding was associated with higher risk of treatment failure compared to use of a well‐fitting support bra.

Comparison of two pharmacologic treatments (comparison 8)

Bromocriptine versus other pharmacologic treatment (at ≤ seven days postpartum)

Bromocriptine versus oestrogen preparations: bromocriptine was compared with diethylstilbestrol (Nilsen 1976, 38 women; Steenstrup 1977, 41 women), ethinyl estradiol (Piya‐Anant 2004; 230 women), and chlorotrianisene (Utian 1975; 31 women). No significant difference was demonstrated in the risks of failure to suppress lactation in any of the trials in the first postpartum week (RR 0.58, 95% CI 0.25 to 1.38; Analysis 8.1.1). One trial (O'Donoghue 1977), which did not provide data in a suitable format for inclusion in an analysis, indicated that the percentage of women showing failure of lactation suppression was significantly greater in women receiving quinestrol compared with those who received bromocriptine at seventh day postpartum.

8.1 — Comparison 8 Comparison of two pharmacologic treatments, Outcome 1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days postpartum.

Bromocriptine versus other ergot derivatives: two trials (Scapin 1982, 40 women; Fischer 1995, 150 women) comparing bromocriptine and methergoline and Venturini 1981 (38 women) comparing it with lisuride, did not indicate any difference in the risks of treatment failure in the first postpartum week (RR 1.12, 95% CI 0.37 to 3.42; Analysis 8.1.2).

Bromocriptine versus prostaglandins: England 1988 (43 women) compared bromocriptine with prostaglandin E₂ for suppression of lactation. The trial suggested no significant difference between the risks of treatment failure of the two agents at or within seven days postpartum (RR 0.55, 95% CI 0.19 to 1.60; Analysis 8.1.3).     Bromocriptine versus pyridoxine: Boes 1980 (97 women) suggested no difference in the risks of treatment failure between bromocriptine and pyridoxine in the suppression of postpartum lactation at or less than seven days postpartum (RR 0.93, 95% CI 0.75 to 1.15; Analysis 8.1.4).

Bromocriptine versus other pharmacologic treatment (at ≤ 14 days postpartum)

At day 14 postpartum, bromocriptine has similar risks of treatment failure compared to cabergoline (European 1991; Giorda 1991, 308 women; RR 1.38, 95% CI 0.93 to 2.05, Analysis 8.4.1), diethylstilbestrol (Nilsen 1976; 38 women, RR 0.30, 95% CI 0.07 to 1.30, Analysis 8.4.2) and cyclofenil (Thorbert 1983; 24 women, RR 3.50, 95% CI 0.16 to 78.19, Analysis 8.4. 8.4.3). Yuen 1977 (39 women) suggested that bromocriptine was associated with reduced risk of treatment failure compared to chlorotrianisene (RR 0.35, 95% CI 0.19 to 0.66) (Analysis 8.4.4). One trial (Biggs 1978; 32 women) that did not provide data in a suitable format for inclusion in an analysis, indicated that milk secretion, breast congestion and discomfort were significantly worse with methyl testosterone compared with bromocriptine during 14 postpartum days.

8.4 — Comparison 8 Comparison of two pharmacologic treatments, Outcome 4 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤14 days.

Bromocriptine versus other pharmacologic treatment (at > 14 days postpartum)

Two trials provided data for this comparison, but they were not in a suitable format for inclusion in an analysis (Purkayastha 1991; Van der Heijden 1991). When used for up to 21 days, Van der Heijden 1991 (30 women) indicated that the efficacy of bromocriptine in suppressing postpartum lactation was similar to that of a non‐ergot dopamine agonist CV 205‐502 (not in data and analysis table). At 28 days of assessment, Purkayastha 1991 (48 women) indicated that bromocriptine was more effective than oestradiol/testosterone ester.

Quinestrol versus other oestrogen preparations (comparison 8)

Quinestrol was compared with ethinyl estradiol (Kuku 1968), diethylstilbestrol (Bergsjo 1974) and chlorotrianisene (King 1972). These trials (208 women) suggested similarity in the risks of treatment failure between quinestrol and other oestrogen preparations in the first postpartum week (RR 0.68, 95% CI 0.32 to 1.44; Analysis 8.1.5).

Watson 1969 (99 women) indicated that quinestrol was associated with increased risk of treatment failure when compared with diethylstilbestrol over a period of one to 10 days (RR 2.84, 95% CI 1.56 to 5.18; Analysis 8.4.5).

Comparison of different dosages of the same drug

High‐ versus low‐dose quinestrol: Vischi 1975 (132 women) suggested that the risk of failure of treatment was significantly lower with 4 mg quinestrol compared with 2 mg quinestrol in lactation suppression (RR 0.51, 95% CI 0.33 to 0.81; Analysis 9.1).

9.1 — Comparison 9 High versus low dose quinestrol, Outcome 1 Failure to suppress lactation as described by milk secretion, breast engorgement or breast pain at ≤ 7 days postpartum.

Low‐ versus high‐dose terguride: Venturini 1988 (45 women) indicated that terguride 0.5 to 1 mg significantly reduced the risk of failure to suppress postpartum lactation over a period of 15 days when compared with terguride 0.25 mg (RR 0.50, 95% CI 0.29 to 0.88; Analysis 10.1). The risks of dizziness following use were similar between women who received a high dose of terguride and those who received low‐dose terguride (RR 1.55, 95% 0.07 to 35.89; Analysis 10.2).

10.1 — Comparison 10 Low versus high dose terguride, Outcome 1 Failure to suppress lactation as described by milk secretion, breast engorgement or breast pain at days 0 ‐15.

10.2 — Comparison 10 Low versus high dose terguride, Outcome 2 Side effect: dizziness.

High‐ versus low‐dose cabergoline: Bravo‐Topete 2004 (80 women) indicated that cabergoline 1 mg reduced the risk of failure to suppress lactation when compared with cabergoline 0.5 mg (RR 0.14, 95% CI 0.03, 0.59; Analysis 11.1).

11.1 — Comparison 11 High versus low dose cabergoline, Outcome 1 Failure to suppress lactation as described by milk secretion, breast engorgement or breast pain.

Long course versus short course of tamoxifen: Shaaban 1975 (65 women) indicated that long‐course tamoxifen (14 days administration) reduced the risk of failure to suppress lactation compared with short‐course tamoxifen (six days administration) when assessed over two weeks postpartum (RR 0.75, 95% CI 0.61 to 0.92, Analysis 12.1).

12.1 — Comparison 12 Long course tamoxifen versus short course tamoxifen, Outcome 1 Failure to suppress lactation as described by milk secretion, breast engorgement or breast pain at D0‐15.

Modified release bromocriptine capsule versus normal bromocriptine tablets: Mizuno 1990 (155 women) indicated similarity in the efficacy of 5 mg once a day modified release capsule and 2.5 mg twice daily normal tablets of bromocriptine when used for suppression of postpartum lactation over a period of two weeks. Data from this trial were not in a suitable format for inclusion in an analysis.

High‐versus low‐dose lisuride: Martinez 1994 (60 women) indicated similarity in the clinical efficacy of lisuride 0.6 mg per day and 0.4 mg per day when used for suppression of postpartum lactation over a period of two weeks. Again, data from this trial were not in a suitable format for inclusion in an analysis.

Comparison of two different pharmacologic combinations

One trial involving 213 women (McNicol 1972) compared a combination of oestradiol benzoate, oestradiol valerate, norethisterone acetate and testosterone enanthate with stilbestrol. Although the trial indicated similarity in the effects of the drugs compared on lactation suppression, the data were not in a format suitable for inclusion in an analysis.

Secondary outcomes

Rebound lactation

We did not include data from many trials on rebound lactation in the results because of inadequate reporting of data. One trial (40 women) (Rolland 1973) comparing bromocriptine with placebo indicated that bromocriptine increased the risk of rebound lactation but the CI was too wide to give a reliable estimate (RR 15.26, 95% CI 1.01 to 231.20, Analysis 1.4). No other trial comparing pharmacologic agents with placebo provided reliable data on rebound lactation. In the comparison of pharmacologic versus non‐pharmacologic treatment, Shrivastav 1988 found no significant difference between the risks of rebound lactation in women treated with bromocriptine compared with those who applied jasmine flower to the breasts (RR 5.00, 95% CI 0.25 to 99.95; Analysis 7.2). Stirrat 1968 (100 women) comparing diethylstilbestrol with placebo did not show any difference between the two with respect to rebound lactation (not in data and analysis table). In the comparison of bromocriptine with other pharmacologic agents, four trials (149 women) (England 1988; Steenstrup 1977; Utian 1975; Venturini 1981) suggested similarity in the risks of rebound lactation between the study and control groups in the trials (RR 0.65, 95% CI 0.39 to 1.10, Analysis 8.2).

1.4 — Comparison 1 Ergot derivatives versus placebo, Outcome 4 Rebound lactation.

7.2 — Comparison 7 Pharmacologic treatment versus nonpharmacologic treatment, Outcome 2 Rebound lactation.

8.2 — Comparison 8 Comparison of two pharmacologic treatments, Outcome 2 Rebound lactation.

Use of second line drugs or method to achieve lactation suppression

The use of second line drug or methods to achieve suppression was poorly reported in the trials. In the comparison of bromocriptine with any other pharmacologic treatment, there was no statistically significant difference between the risks of using a second line drug or method to achieve lactation suppression when bromocriptine was compared with oestrogen preparations (Utian 1975, 31 women, RR 0.31, 95% CI 0.01 to 7.15, Analysis 8.3.1) and other ergot derivatives (Scapin 1982, 40 women, RR 2.67, 95% CI 0.82 to 8.62, Analysis 8.3.2). Boes 1980 (97 women) suggested that bromocriptine was associated with reduced proportion of women in need of second line methods to suppress lactation compared to pyridoxine (RR 0.07, 95% CI 0.01 to 0.51; Analysis 8.3.3). Phillips 1975 (196 women) suggested that women who received chlorotrianisene were less likely to use supplemental or concurrent therapy (breast binders, ice bags or analgesics, or both) compared with women who had placebo at or within four days after admission (not in data and analysis table).

8.3 — Comparison 8 Comparison of two pharmacologic treatments, Outcome 3 Use of second line drug or method to achieve suppression of lactation.

Discussion

In spite of the questions on the need to apply treatment for suppressing lactation, a lot of research has gone into finding the most effective treatment. This review indicates that the emphasis for this search has been on pharmacologic treatments (compared with either no treatment or each other). The review included comparisons of orally and intramuscularly administered pharmacologic agents with placebo, other pharmacologic agents and non‐pharmacologic methods. In terms of outcome, we used the persistence of one of the three common clinical symptoms or signs of postpartum lactation (i.e., milk secretion, breast engorgement and breast pain) as evidence of failure of lactation suppression. We also explored the consistency of results obtained from this definition with those obtained from separate consideration of each lactation symptom or sign as an indicator of treatment failure. In an attempt to reduce clinical heterogeneity, we extracted results on efficacy of treatment effects within a specified postpartum period.

This review shows that the search for the most effective lactation suppressant has relied on small trials, most of which have low methodological quality. The lack of recent randomised controlled trials on this subject may imply a supposed exhaustiveness and conclusion on this research topic. It may also be attributed to the lack of motivation of potential researchers because of the increasing number of questions on the need for treatments to suppress lactation. The gross variation in the dosage, interventions, duration of treatments and outcome assessments suggests a general lack of a clear understanding of the physiology of lactation and mechanisms of action of the treatments tested among researchers.

This is the first systematic review assessing the effectiveness of all forms of treatments for lactation suppression. Assessment of methodological quality of studies was based on stringent criteria, as evident by the number of relevant trials excluded from the review. In spite of this measure, however, the robustness of the results is diminished by the fact that the majority of included studies are of uncertain methodological quality, as evident in the proportion of trials with adequate allocation concealment (8.1%) and those that blinded their outcome assessments. Blinding of outcome assessment is particularly important for this topic, considering the subjective nature of main outcome measures (secretion, congestion and pain). The extent to which outcome measures are blinded in double‐blinded trials where the women served the dual role of participant and outcome assessor is uncertain. Combining data on outcomes assessed by different observers (clinicians and women) is likely to contribute to ascertainment bias in included studies that employed such method. In addition, the validity of the results interpreted by the trialists could not be examined in included trials that reported data that were unsuitable for inclusion in the analyses table. Another major limitation of this review is the fact that a significant proportion of drugs tested in the included trials are currently no longer registered for use in most countries, either as lactation suppressants or otherwise. This may significantly limit the applicability of the findings of the review.

About half of the trials included compared drugs with placebo. In spite of its popularity, only five trials involving 206 women comparing bromocriptine with placebo met the inclusion criteria. All these trials used uncertain methods of generating allocation sequence and allocation concealment, none had outcome assessment blinded and all were conducted over two decades ago. It can thus be concluded that there is weak evidence that bromocriptine is better than nothing for suppressing the symptoms of lactation during the first seven days postpartum. Contrary to a number of case reports, there is insufficient evidence from this review to indicate whether or not bromocriptine is associated with an increased risk of major side effects (notably thromboembolism, myocardial infarction and maternal death) in the first postpartum week.

Although there is small evidence that oestrogen preparations may be better than nothing in the first postpartum week, none of the tested agents is presently available in the market. In addition, available data on their risks of major side effects are insufficient to make a reliable conclusion on their safety when used for lactation suppression. Of all the pharmacological agents compared with placebo, only bromocriptine is still registered for use in most countries although not necessarily for lactation suppression. It needs to be stated that bromocriptine appeared to be the gold standard in the 1970s and 1980s as shown by the number of trials where it was tested. The possibility of publication bias therefore exists in the face of seemingly 'overwhelming' evidence. This review did not show sufficient evidence to indicate if other pharmacologic agents (clomiphene, tamoxifen, prostaglandins, pyridoxine, oxytocin, L‐dopa and homeopathic preparation) are useful in suppressing the symptoms of lactation postpartum, as they are all based on individual small trials. However, the combination of testosterone and oestrogen preparations appear to be somewhat effective in suppressing the symptoms of lactation.

Evidence on the comparative effectiveness of pharmacologic agents and non‐pharmacologic methods is based on three trials, each testing different non‐pharmacologic methods (Bhardwaj 1979; De Gezelle 1979; Shrivastav 1988). It can be concluded from their results that there is not enough evidence to indicate which of the approaches is better than the other. The comparative effectiveness of jasmine flower with that of bromocriptine can only be useful once there is strong evidence that bromocriptine is better than placebo. There is currently no evidence on whether non‐pharmacologic methods are better than placebo in lactation suppression.

There are mixed views on the subject of lactation suppression in most current obstetric textbooks, although the need for applying some kind of treatment is seldom disputed. Most authors refer to research findings indicating the effectiveness of previously tried approaches, many of which were excluded from this review as a result of low methodological quality and high potential for bias (seeCharacteristics of excluded studies). Generally, non‐pharmacologic approaches such as use of a well‐supporting brassiere and avoidance of nipple stimulation, are often recommended based on the presumed safety and effectiveness of these methods. This policy, however, is not supported by the findings of this review. While the methodological limitations of conducting high‐quality trials involving non‐pharmacologic methods are understandable, they should not compromise the need to provide evidence‐based guidelines on their application.

It is unlikely that the findings from this review would inform any change in the recommendations of the United States' Food and Drug Administration on the routine use of bromocriptine for suppression of postpartum lactation, as the evidence indicating its effectiveness is weak even though the review also did not show any clear evidence of adverse effects.

Three issues related to the data and studies included in this review need to be discussed for a better understanding of the implications for practice and research regarding lactation suppression. The review addresses treatments for lactation suppression in postpartum women who do not desire to breastfeed their infant right from birth. Therefore, the findings are not directly applicable to women who had initiated breastfeeding but later wish to discontinue, or to women who lactate due to other pathology (e.g., hyperprolactinaemia). It is possible that the effectiveness as well as the side‐effect profile of tested agents remote from delivery may be different from that in the immediate postpartum period. Secondly, participants in the included trials were healthy postpartum women (including healthy HIV‐positive mothers) and it is uncertain if similar results, especially regarding side effects, would be found in postpartum women with a higher baseline risk of morbidity, e.g., those who are ill, or taking medications, or both, including antiretroviral drugs. It is important to note that this category of mothers for whom suppression of lactation may have particular health benefits, is likely to have increased risk of side effects, including drug interactions, which were not explored in the studies included in this review. Lastly, a fundamental question that is yet to be answered is whether postpartum women desire treatment for lactation suppression and, if so, which of the symptoms women are most concerned about. This would definitely go a long way in better interpretation of the effectiveness of the tested approaches. It is clear from this review that women's views about the treatment received was not a priority, as only four trials with unreliable data explored the acceptability profile of the tested approach.

With respect to side effects, there is insufficient evidence to show that pharmacologic agents are more associated with a higher risk of major adverse effects (notably thromboembolism) compared with no treatment or each other. Several case reports have been published on thromboembolism, myocardial infarction and cerebral angiopathy following the use of bromocriptine, although a causal relationship has not been established. It is interesting to note that, in spite of these concerns, emphasis was not laid on reporting these side effects in trials conducted subsequently. However, these findings should be interpreted against the background of our review of only published data as exploration of unpublished data, especially adverse outcome data from drug company trials, may provide a clearer picture.

Authors' conclusions

Implications for practice.

There is weak evidence that some pharmacologic treatments (most of which are currently unavailable to the public) are better than no treatment for suppressing the symptoms of lactation in the first week postpartum. There is currently no evidence to show that non‐pharmacologic approaches are more effective than no treatment. Presently, there is insufficient evidence to address the issue of side effects of the pharmacologic and non‐pharmacologic methods that are employed for suppressing lactation. When women desire treatment for suppressing lactation, consideration may be given to bromocriptine where it is still registered for such use in healthy mothers with no predisposition to major side effects of public concern. In spite of its importance, there is inadequate evidence to comment on the acceptability of approaches for suppressing postpartum lactation to women.

Implications for research.

For settings in which application of some forms of treatment for suppressing lactation is the norm, there is a need for well conducted and large randomised controlled trials to compare the effectiveness of pharmacologic treatment, notably, bromocriptine to no treatment. Future research should also focus on comparison of non‐pharmacologic approaches with no treatment since they presently do not appear to have any safety concerns for the public. The most important symptoms that concern women who desire not to breastfeed should be studied in large observational studies to ensure that effective treatment is sought by researchers on the inclination of clients' needs. Such studies should also be large enough to detect clinically important differences between interventions with respect to major side effects that have been reported through less rigorous research. Priority should be given to studying fewer dose regimens and cheap approaches in view of women in low‐resource countries.

What's new

Date	Event	Description
23 July 2012	New search has been performed	Search updated. Sixteen additional trials included (four previously excluded trials are now included but they do not contribute any data to the review (Bhardwaj 1979; Binns 1967; Kulski 1978; Van der Heijden 1991). Two trials added to Studies awaiting classification. No change to conclusions.
23 July 2012	New citation required but conclusions have not changed	Review updated.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days postpartum	3	107	Risk Ratio (IV, Fixed, 95% CI)	0.36 [0.24, 0.54]
1.1 Bromocriptine versus placebo	3	107	Risk Ratio (IV, Fixed, 95% CI)	0.36 [0.24, 0.54]
2 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 14 days	2	76	Risk Ratio (IV, Random, 95% CI)	0.18 [0.03, 1.08]
2.1 Bromocriptine versus placebo	2	76	Risk Ratio (IV, Random, 95% CI)	0.18 [0.03, 1.08]
3 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 14 days	1	32	Risk Ratio (IV, Fixed, 95% CI)	0.19 [0.07, 0.48]
3.1 Cabergoline versus placebo	1	32	Risk Ratio (IV, Fixed, 95% CI)	0.19 [0.07, 0.48]
4 Rebound lactation	1	40	Risk Ratio (IV, Fixed, 95% CI)	15.26 [1.01, 231.20]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days	7	971	Risk Ratio (IV, Random, 95% CI)	0.40 [0.29, 0.56]
1.1 Diethylstilbestrol versus placebo	4	376	Risk Ratio (IV, Random, 95% CI)	0.33 [0.12, 0.89]
1.2 Quinestrol versus placebo	3	342	Risk Ratio (IV, Random, 95% CI)	0.47 [0.30, 0.73]
1.3 Chlorotrianisene versus placebo	1	153	Risk Ratio (IV, Random, 95% CI)	0.51 [0.36, 0.73]
1.4 Hexestrol versus placebo	1	100	Risk Ratio (IV, Random, 95% CI)	0.41 [0.29, 0.57]
2 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 14 days	2	103	Risk Ratio (IV, Random, 95% CI)	0.19 [0.03, 1.34]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days	1	30	Risk Ratio (IV, Fixed, 95% CI)	1.0 [0.51, 1.95]
1.1 Clomiphene versus placebo	1	30	Risk Ratio (IV, Fixed, 95% CI)	1.0 [0.51, 1.95]
2 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 14 days	1	140	Risk Ratio (IV, Fixed, 95% CI)	0.71 [0.62, 0.82]
2.1 Tamoxifen versus placebo	1	140	Risk Ratio (IV, Fixed, 95% CI)	0.71 [0.62, 0.82]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days	2	258	Risk Ratio (IV, Fixed, 95% CI)	0.98 [0.89, 1.09]
1.1 Pyridoxine versus placebo	2	258	Risk Ratio (IV, Fixed, 95% CI)	0.98 [0.89, 1.09]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days	3	436	Risk Ratio (IV, Fixed, 95% CI)	0.15 [0.10, 0.22]
1.1 Testosterone oenanthate + oestradiol valerate	3	436	Risk Ratio (IV, Fixed, 95% CI)	0.15 [0.10, 0.22]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain at ≤ 7 days	1	30	Risk Ratio (IV, Fixed, 95% CI)	1.13 [0.60, 2.11]
1.1 Testosterone propionate versus placebo	1	30	Risk Ratio (IV, Fixed, 95% CI)	1.13 [0.60, 2.11]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to suppress lactation as indicated by milk secretion, breast engorgement or breast pain	2		Risk Ratio (IV, Random, 95% CI)	Subtotals only
1.1 Bromocriptine versus any nonpharmacologic treatment	2	96	Risk Ratio (IV, Random, 95% CI)	0.31 [0.16, 0.61]
1.2 DIethylstibestrol versus any nonpharmacologic treatment	1	36	Risk Ratio (IV, Random, 95% CI)	0.19 [0.07, 0.49]
1.3 Oestradiol + testosterone esters versus any non pharmacologic treatment	1	36	Risk Ratio (IV, Random, 95% CI)	0.46 [0.28, 0.76]
1.4 Bendrofluazide versus any nonpharmacologic treatment	1	36	Risk Ratio (IV, Random, 95% CI)	0.68 [0.48, 0.94]
2 Rebound lactation	1	60	Risk Ratio (IV, Fixed, 95% CI)	5.0 [0.25, 99.95]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to suppress lactation as described by milk secretion, breast engorgement or breast pain at ≤ 7 days postpartum	1	132	Risk Ratio (IV, Fixed, 95% CI)	0.51 [0.33, 0.81]
1.1 Quinestrol 4 mg versus Quinestrol 2 mg	1	132	Risk Ratio (IV, Fixed, 95% CI)	0.51 [0.33, 0.81]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to suppress lactation as described by milk secretion, breast engorgement or breast pain at days 0 ‐15	1	45	Risk Ratio (IV, Fixed, 95% CI)	0.5 [0.29, 0.88]
1.1 Terguride 0.5‐1.0 mg vs terguride 0.25 mg	1	45	Risk Ratio (IV, Fixed, 95% CI)	0.5 [0.29, 0.88]
2 Side effect: dizziness	1	45	Risk Ratio (IV, Fixed, 95% CI)	1.55 [0.07, 35.89]
2.1 Terguride 0.5‐1.0 mg vs terguride 0.25 mg	1	45	Risk Ratio (IV, Fixed, 95% CI)	1.55 [0.07, 35.89]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to suppress lactation as described by milk secretion, breast engorgement or breast pain at D0‐15	1	65	Risk Ratio (IV, Fixed, 95% CI)	0.75 [0.61, 0.92]
1.1 Tamoxifen (14 days course) vs tamoxifen (6 days course)	1	65	Risk Ratio (IV, Fixed, 95% CI)	0.75 [0.61, 0.92]

Methods	Randomised trial.
Participants	41 women in whom inhibition of lactation was planned. Indications included late abortion or fetal death, missed abortion, previous mastitis and adoption. Women were included in the study if staying in the hospital for at least 5 days. Exclusion criteria were not specified. Setting: a university hospital in Norway.
Interventions	Quinesterol tablet 4 mg (on day 0) and if necessary, an additional 4 mg tablet on day 4 (n = 23) versus diethylstilbestrol 5 mg tablet thrice daily for 5 days (n = 18).
Outcomes	Milk secretion, breast pain, additional oestrogen medication, symptoms suggestive of mastitis and thrombosis.
Notes	Quinestrol experimental tablets were supplied by Apothekernes Laboratorium for Special preparater, Oslo, Norway.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of sequence generation not stated. Participants were divided into 2 groups "by random allocation".
Allocation concealment (selection bias)	Unclear risk	No information about allocation concealment to permit judgement.
Blinding (performance bias and detection bias)   Lactation	High risk	No information on blinding of participants, study personnel and outcome assessors. Differences in the dosage regimen between the 2 interventions suggested no blinding.
Blinding (performance bias and detection bias)   Adverse events	High risk	No information on blinding of participants, study personnel and outcome assessors. Differences in the dosage regimen between the 2 interventions suggested no blinding.
Incomplete outcome data (attrition bias)   Lactation	Low risk	2 participants (4.9%) were lost to follow‐up after leaving the hospital (1 from the quinestrol group and the other from the diethylstilbestrol group). Outcome data were available for all participants for the 5 days in the hospital.
Incomplete outcome data (attrition bias)   Adverse events	Low risk	Outcome data were available for all participants for the 5 days in the hospital.
Selective reporting (reporting bias)	Low risk	Comparison of outcome measures in the 'Methods' and 'Results' sections of the report indicated no evidence of selective outcome reporting.
Other bias	Low risk	The study appears free of other sources of bias.

Methods	Double‐blind randomised trial.
Participants	90 women not wishing to breastfeed. Setting: a general practitioner maternity unit at Bridgnorth, Shropshire, United Kingdom.
Interventions	Chlorotrianisene 24 mg tablets versus placebo. Two tablets thrice daily for 4 days starting from the day of delivery.
Outcomes	Milk leakage, breast pain and congestion; side effects. The assessment of degree of severity of breast symptoms was on a 4‐point scale. Observations were made daily for a minimum of 6 and usually 7 or more days until the symptoms ceased.
Notes	Data were not included because the results were not presented in a usable form.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of sequence generation not stated. "The order of administration (of the interventions) was random."
Allocation concealment (selection bias)	Unclear risk	Insufficient information about allocation concealment to permit judgement. "One of each consecutive pairs of numbered bottles contained active chlorotrianisene tablets and the other dummy tablets".
Blinding (performance bias and detection bias)   Lactation	Low risk	Participants, study personnel and outcome assessors were blinded. "None of those involved in the trial knew at the time whether the patient was taking active or dummy tablets. Assessment of record cards was made, before breaking the code, by an independent statistical adviser who was not directly involved in the trial."
Blinding (performance bias and detection bias)   Adverse events	Low risk	Participants, study personnel and outcome assessors were blinded. "None of those involved in the trial knew at the time whether the patient was taking active or dummy tablets. Assessment of record cards was made, before breaking the code, by an independent statistical adviser who was not directly involved in the trial."
Incomplete outcome data (attrition bias)   Lactation	Low risk	5 (5.6%) of the women initially recruited into the trial "were prematurely discharged from the hospital and were therefore excluded from the within‐pair comparison."
Incomplete outcome data (attrition bias)   Adverse events	Low risk	5 (5.6%) of the women initially recruited into the trial "were prematurely discharged from the hospital and were therefore excluded from the within‐pair comparison."
Selective reporting (reporting bias)	Low risk	Comparison of outcome measures in the 'Methods' and 'Results' sections of the report indicated no evidence of selective outcome reporting.
Other bias	Low risk	The study appears free of other sources of bias.

Methods	Method of randomisation not stated. Participants, clinicians and outcome assessors were blinded.
Participants	31 postpartum non‐breastfeeding women who were free from metabolic or surgical conditions which might interfere with the absorption, metabolism or excretion of the drugs, absence of concurrent medication. Setting: a university hospital in Cape Town, South Africa.
Interventions	Oral bromocriptine 2.5 mg twice daily for 14 days (n = 16) versus chlorotrianisene 24 mg twice daily for 14 days (n = 15).
Outcomes	Milk production, breast congestion and side effects (blood pressure changes, pulse rate, rebound lactation, 24 hour urinary output and thromboembolism.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of sequence generation not stated.
Allocation concealment (selection bias)	Unclear risk	No information about allocation concealment to permit judgement.
Blinding (performance bias and detection bias)   Lactation	Unclear risk	Study participants and personnel were blinded to interventions. Uncertain whether outcome assessors were blinded.
Blinding (performance bias and detection bias)   Adverse events	Unclear risk	Uncertain whether outcome assessors were blinded.
Incomplete outcome data (attrition bias)   Lactation	Low risk	7 (18.4%) of the 38 participants who entered the trial were lost to follow‐up. Drop‐out rates were similar for both groups.
Incomplete outcome data (attrition bias)   Adverse events	Low risk	Drop‐out rates were similar for both groups.
Selective reporting (reporting bias)	Low risk	Comparison of outcome measures in the 'Methods' and 'Results' sections of the report indicated no evidence of selective outcome reporting.
Other bias	Low risk	The study appears free of other sources of bias.

Study	Reason for exclusion
Almeida 1986	This is a randomised double blind study evaluating the effect of bromocriptine mesylate on suppression of puerperal fever resulting from breast engorgement.
Bare 1960	This study was conducted to determine the efficacy and the optimum dosage of fluoxymesterone in the suppression of lactation. There was no evidence of randomisation. The patients were divided into 4 groups. Group I consisted of ward patients who received placebo and other groups were all private patients. The dosage of the intervention was administered on "a strict alternating basis".
Barns 1961	This study compared the effectiveness of a combination of testosterone enanthate and estradiol valerate in the suppression of postpartum lactation. There was no indication of randomisation.
Booker 1970	This study compared combined and sequential type of oral contraceptive pills for suppressing lactation. Alternate allocation method was used.
Brett 1971	This study compared stilbestrol and quinestrol with placebo. Although the study was described as a "double‐blind trial", there was no indication of randomisation.
Bristol 1966	This study compared the effectiveness of compression breast binder and supporting brassiere in the suppression of postpartum lactation. There was no indication of randomisation. The final study group consisted of 38 women who were "divided into two groups of 19 patients each".
Brooten 1983	This study compared 3 non‐pharmacologic measures (compression binder, standardised support bra and fluid restriction) with bromocriptine. The assignment of women to bromocriptine group was "by virtue of their physicians' treatment".
Caballero 1987	This is a double‐blind study comparing methergoline with bromocriptine in the suppression of lactation. Although the study was described as randomised, allocation of interventions was by alternation.
Canales 1977	This study evaluated the effect of clomiphene in the suppression of puerperal milk secretion and serum prolactin. There was no indication of randomisation.
Cantis 1977	This was a single‐blind study carried out in 42 women to assess the lactation suppressing activity of piribedil. There was no evidence of randomisation. Treatments were assigned to unbalanced groups of women (33 women received the active agent while 9 received placebo).
Cicinelli 1996	This study evaluated the effectiveness of nasal bromocriptine on serum prolactin.
David 1977	This study compared the effectiveness of stilbestrol with placebo on prolactin level.
De Aloysio 1988	This study compared the effect of dihydroergocristine to that of bromocriptine on prolactin secretion and postpartum lactation. All the puerperae enrolled in the study had a physiological delivery and wished to interrupt breastfeeding after at least 3 months of nursing.
De Cecco 1979	This study evaluated the effect of Lisuride on lactation and postpartum serum prolactin level. There was no indication of randomisation.
Del Pozo 1975	This study evaluated the action of methergoline on plasma prolactin and milk secretion in the first 7 postpartum days. There was no indication of randomisation and women were "divided" into 2 groups.
Duthie 1990	This randomised trial compared the efficacy, tolerance and effect on prolactin level of 4 different dosages of intramuscular bromocriptine retard. Patients were randomised into 4 equal groups. Loss to follow‐up was 10.8%. Variable numbers of women in the groups developed mild to moderate breast engorgement and milk flow prior to administration of bromocriptine.
Fleming 1977	This double‐blind trial compared the efficacy of pyridoxine, stilbestrol and a placebo in the inhibition of puerperal lactation. There was no indication of randomisation.
Foukas 1972	This double‐blind trial compared diethylstilbestrol and pyridoxine with placebo in the suppression of lactation. There was no indication of any randomised comparison between the groups. Patients were divided into 4 unbalanced groups (86, 68, 75 and 25).
Garry 1956	This study was undertaken to evaluate the effectiveness of oestrogen‐androgen preparation in the suppression of lactation. There was no indication of randomisation. The patients were divided into 2 unequal groups (100 and 50).
Gerstner 1978	This study compared bromocriptine with hexestrol for suppressing lactation. Alternate allocation method was used.
Gillibrand 1968	Data to evaluate the validity of the methods used are not available in this published correspondence that described 2 clinical trials. It has been excluded because no full publication of the study could be located.
Gopalan 1997	This randomised double‐blind study assessed the effectiveness of pyridoxine in inhibition of lactation and serum prolactin. The report is only available in abstract form.
Grant 1978	Although this trial was described as a "double blind evaluation of quinestrol, chlorotrianisene and placebo", there was no indication of randomisation.
Hale 2004	This study evaluated the effects of pseudoephedrine on milk production, plasma prolactin and breastmilk levels following a maternal dose of 60 mg in 8 breastfeeding mothers. There was no indication of randomisation.
Kaiser 1952	This study was conducted to evaluate the effect of combined oestradiol and progesterone on lactation suppression. There was no indication of randomisation.
Kalir 1975	This study was conducted to evaluate the effect of clomiphene citrate on lactation suppression. There was no indication of randomisation.
Kee 1989	This study evaluated the usefulness of serrapeptase (Dansen) in postpartum women with breast engorgement in a double blind randomised controlled trial.
King 1958	Study comparing chlorotrianisene with placebo. There was no indication of randomisation.
Kirkland 1960	This "randomised" study involving 160 women had the allocation concealment deciphered for the placebo group during the study resulting in unequal allocation of subjects to the placebo arm and consequent discontinuation of the arm. This questions adherence to the initially generated allocation sequence till the end of the trial.
Koshiishi 1971	This study evaluated the effect of a proteolytic enzyme (bromelain) on breast engorgement.
Lee 1971	This study compared chlormezanone (a non‐hormonal tranquillizer) with stilbesterol. There was no evidence of randomisation. Patients were allocated alternately into 1 of 2 groups.
Lee 1979	This study compares Ginsenocide triol with bromocriptine. The paper is only available in abstract form.
Llewellyn‐Jones 1963	This double‐blind trial compared the effectiveness of stilbestrol with placebo in the suppression of lactation. There was no indication of randomisation.
Lo Dico 1980	This study evaluated the effectiveness of pyridoxine, 2‐ Br‐ alpha‐ergocriptine and piribedil in lactation suppression. There was no indication of randomisation. Patients were "divided" into 4 uneven groups.
Louviere 1975	The primary purpose of this study was to evaluate the effectiveness of Deladumone OB in the suppression of postpartum breast engorgement and lactation. There was no indication of randomisation.
MacDonald 1965	This double‐blind trial compared stilbestrol with placebo. Alternate allocation method was used.
MacLeod 1977	This study evaluated the dose response and timing of administration of bromocriptine. There was no evidence of randomisation. Patients were divided into 3 groups.
Markin 1960	This study compared the effectiveness of 5 preparations (diethylstilbestrol, dienestrol plus methyltestosterone, conjugated equine oestrogen plus methyltestosterone, testosterone propionate plus diethylstilbestrol and testosterone enanthate plus estradiol valerate) with placebo. Alternation method of allocation was used. "Drug E" or "Drug O" was administered by the nursing staff to the patients delivering on the even‐numbered or the odd‐numbered days of the month, respectively.
Masala 1978	This study was designed to assess the effect of tamoxifen on the inhibition of puerperal lactation. There was no indication of randomisation. Treatments were assigned to 2 unbalanced groups (60: experimental group and 20: placebo group).
McLachlan 1991	This randomised double‐blind placebo controlled trial tested the efficacy of thermal ultrasound therapy as a treatment for severe postpartum breast engorgement.
Momberg 1976	This "double‐blind study" was "based on the order of admittance to the clinic and a randomisation list" suggesting the inclusion of an alternate allocation method.
Morris 1967	This study was designed to evaluate the effectiveness of quinestrol in suppressing puerperal lactation. There was no indication of randomisation.
Morris 1970	This double‐blind study evaluated the effectiveness of 3 preparations; oral chlorotrianisene in 3 different dosage strengths; an intramuscular combination of 2 steroids, testosterone enanthate and estradiol valerate; and identical placebos in the inhibition of puerperal breast engorgement, discomfort and milk secretion. Allocation of patients does not suggest randomisation. For the chlorotrianisene study, an unbalanced number of patients received the active agent and placebo (75 vs 25).
Nappi 1987	This study examined whether the side effects of bromoergocriptine could be prevented by combining bromoergocriptine treatment with the antiemetic domperidone, without affecting the prolactin lowering effect and subsequent inhibition of lactation.
Nappi 1990	This study assessed the effect of Ibopamine, a peripheral agonist on prolactin and milk production. 80 participants were admitted into the study including 30 nursing mothers. Participants were randomly "divided" into 6 groups.
Nappi 1993	This study assessed the effects of dihydroergocriptine on serum prolactin levels and lactation in postpartum women. Women were "divided" into 6 unequal groups. Although it was stated that women were "randomly assigned" to 4 groups of non‐nursing mothers and 2 groups of nursing mothers, the non‐nursing mothers received the active agents while the nursing mothers received placebo. .
Ng 1972	This study compared Quinestrol with Ablacton. Alternate allocation method was used.
Nisha 2006	This "prospective observational study" compared oral carbegoline with intramuscular injection of combined oestrogens and androgens. Women were "randomly divided" into 2 groups. The study was only available in abstract form.
Osbourne 1978	This study compared the effect of bromocriptine and quinestrol on coagulation and fibrinolysis.
Polatti 1982	This study evaluated the inhibitory effect of cyclofenil on prolactin. There was no indication of randomisation.
Poulsen 1976	This study evaluated the effect of bromocriptine on established lactation.
Primrose 1957	This study compared TACE, Premarin and methyltestosterone, stibesterol with placebo. There was no indication of randomisation.
Reisfield 1966	A "double‐blind" study using hydrochlorothiazide and an identical placebo in 100 consecutive postpartum women. There was no indication of randomisation.
Robuschi 1987	This study evaluated the effect of maternal administration of bromocriptine on fetal and maternal serum growth hormone concentrations.
Rolland 1978	This study describes 2 double‐blind studies on the effect of bromocriptine compared with placebo and an oestrogen/androgen compound. There was no indication of randomisation.
Roser 1966	This study compares the effect of testosterone enanthate and estradiol valerate (Deladumone 2X) on suppression of symptoms of postpartum lactation. There was no indication of randomisation. Women were "divided" into 2 groups.
Ryan 1962	This study evaluated the effectiveness of intranasal syntocinon compared to a placebo for relief of postpartum breast discomfort. Patients were instructed to begin assigned treatment when they first felt discomfort in their breasts. A total of 38.3% of women included in the study were excluded from the analysis as they did not experience enough discomfort to require the assignment of treatment.
Schneider 1964	Double‐blind study comparing depot types of oestrogen and androgen together with a rapidly acting oestrogen against placebo. Treatments were allocated to alternative patients.
Seppala 1975	This double‐blind study compared the effect of CB 154 (2‐Br‐alpha‐ergocriptine methane sulphonate) with diethylstilbestrol on established mammary secretion and congestion. There was no indication of randomisation. 39 patients were "divided into two groups".
Shapiro 1984	Compares bromocriptine with breast binders and analgesics for inhibiting lactation. Alternate allocation method was used.
Steele 1968	This study described the comparison of stibestrol with placebo in a double‐blind trial. Paper was presented as correspondence and full paper could not be located. There was no indication of randomisation in the published correspondence.
Stenchever 1962	There was no evidence of randomisation.
Tyson 1966	This study was designed to test the efficacy of a 3‐day course of chlorotrianisene for prevention and treatment of postpartum breast engorgement. There was no evidence of randomisation.
Van Dam 1981	This study compared the lactation‐inhibiting effects of lisuride and bromocriptine. Although the study was labelled as a "double‐blind trial", there was no indication of randomisation.
Varga 1972b	This study evaluated the lactation‐inhibitory effects of bromocriptine, diethylstilbestrol and placebo. There was no indication of randomisation.
Walker 1980	This is the same study as Walker 1975.
Willmott 1977	This double‐blind placebo controlled study was undertaken to evaluate the clinical effectiveness of bromoergocrytine in suppressing lactation and observe any side effects over 28 days. There was no evidence of randomisation.
Zuckerman 1973	This study was designed to determine the effectiveness of clomiphene in inhibiting postpartum lactation. There was no indication of randomisation. Patients were divided into 4 uneven groups (110, 26, 31 and 10).

PERMALINK

Treatments for suppression of lactation

Olufemi T Oladapo

Bukola Fawole

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Background

Description of the condition

Indications for lactation suppression

Lactation suppression and prevention of vertical transmission of HIV

Description of the intervention

Non‐pharmacologic methods of lactation suppression

Pharmacologic methods of lactation suppression

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

(1) Random sequence generation (checking for possible selection bias)

(2) Allocation concealment (checking for possible selection bias)

(3) Blinding (checking for possible performance and detection bias)

(4) Incomplete outcome data (checking for possible attrition bias through withdrawals, dropouts, protocol deviations)

(5) Selective reporting bias

(6) Other sources of bias

Measures of treatment effect

Dichotomous data

Dealing with missing data

Assessment of heterogeneity

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Participants

Interventions

Outcomes

Risk of bias in included studies

1.

2.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Primary outcomes

Pharmacologic treatment versus no treatment or placebo

Ergot derivatives versus no treatment or placebo (comparison 1)

1.1. Analysis.

1.2. Analysis.

1.3. Analysis.

Oestrogen preparations versus no treatment or placebo (comparison 2)

2.1. Analysis.

2.2. Analysis.

Antioestrogens versus no treatment or placebo (comparison 3)

3.1. Analysis.

3.2. Analysis.

Pyridoxine versus no treatment or placebo (comparison 4)

4.1. Analysis.

Combined oestrogen and androgen preparations versus no treatment or placebo (comparison 5)

5.1. Analysis.

Androgen preparations versus no treatment or placebo (comparison 6)