Breastfeeding and risk classification of medications used during hospitalization for delivery: 2015 Pelotas Birth Cohort

REV BRAS EPIDEMIOL 2020; 23: E200026 ABSTRACT: Objective: To classify the drugs used during childbirth in relation to risks in breastfeeding, by using different sources of information and determining their disagreements. Methods: Cross-sectional study, within the 2015 Pelotas Birth Cohort. Information about the use of drugs was collected, classified and compared regarding risk according to: 1) Brazil Ministry of Health Manual (MS), 2) World Organization (WHO), 3) Newton and Hale’s classification and 4) American Academy of Pediatrics (AAP). Results: A total of 1,409 mothers participated, and they had used 14,673 medicines, with 143 different drugs, of which 28 showed discordant classification with regard to breastfeeding risk. These 28 drugs included the following: morphine (64%), classified by AAP and WHO as compatible and as judicious use use by MS and Newton and Hale; hyoscine (23%), classified as judicious use by MS and compatible (A) by AAP; and metoclopramide (18%), classified as compatible by MS, of effects unknown (D) by AAP, and should be avoided according to WHO. Of the total drugs, 49.7% were classified as compatible during breastfeeding. Almost all women used oxytocin (97.4%), followed by lidocaine (75%), ketoprofen (69%), cephalothin (66%) and diclofenac (65%), which were classified as compatible. Conclusion: There was extensive use of drugs by mothers in labor during admission, most of the drugs being classified at the same risk and almost half classified as compatible with breastfeeding. However, there was disagreement between the sources for 19.6% of the drugs analyzed, which could endanger the infant’s health or leave doubts about the use of the drug or breastfeeding.


INTRODUCTION
The use of medications in the immediate postpartum period is very common, occurring in almost all postpartum women 1 , which may be a reason for not breastfeeding. The World Health Organization (WHO) recommends breastfeeding for at least 2 years of age and exclusively until the sixth month 2 , as breastfeeding is related to several nutritional, immunological, cognitive, psycho-affective, economic and social advantages 3 , but breastfeeding time in Brazil and in several other countries does not meet this recommendation 4 . According to Victora et al. 5 , if breastfeeding was extended to an almost universal level, it could prevent 823,000 deaths of children under 5 years each year and 20,000 deaths from breast cancer, with the additional benefit of saving 300 billion dollars.
Care should be taken when making a decision about whether a medication taken during breastfeeding is safe, where the risk to the baby and the benefit to the mother need to be assessed 5 . The oral bioavailability of drugs in the gastrointestinal tract of infants also needs to be considered, since many drugs are not absorbed by their gastrointestinal system. Greater amounts of medications can be transferred to mother's milk in the colostral phase, but only minimal quantities are transferred to the infant, due to the limited volume of colostrum. On the other hand, with mature milk, there is a larger volume, but less drugs are transferred to breast milk because of the tight junctions between cells 7 . Characteristics related to the drug, such as molecular weight, liposolubility, binding to plasma proteins and half-life, are also very important to determine the amount that reaches maternal milk 3 . RESUMO: Objetivo: Classificar os medicamentos usados durante o parto quanto aos riscos na amamentação, utilizando diferentes fontes e verificando suas discordâncias. Métodos: Estudo transversal inserido na coorte de nascimentos de Pelotas de 2015. Coletaram-se informações sobre o uso de medicamentos, classificando-os quanto ao risco de acordo com: manual do Ministério da Saúde (MS), Organização Mundial da Saúde (OMS), classificação de Newton e Hale e Academia Americana de Pediatria (AAP). Resultados: Participaram 1.409 mães, utilizando 14.673 medicamentos, sendo 143 fármacos diferentes, dos quais 28 tiveram classificação de risco na amamentação discordante. Entre aqueles com classificação discordante estão morfina (64%), classificada pela AAP e OMS como compatível e pelo MS e por Newton e Hale como criterioso; hioscina (23%), criterioso pelo MS e compatível (A) pela AAP; e metoclopramida (18%), compatível pelo MS, de efeitos desconhecidos (D) pela AAP e evitado de acordo com a OMS. Do total de medicamentos, 49,7% foi classificado como compatível com a amamentação. Quase a totalidade das mulheres utilizou ocitocina (97,4%), seguida de lidocaína (75%), cetoprofeno (69%), cefalotina (66%) e diclofenaco (65%), classificados como compatíveis. Conclusão: Houve amplo uso de medicamentos pelas mães durante a internação para o parto, a maioria deles classificada no mesmo grau de risco, e quase a metade classificada como compatível com a amamentação, porém houve discordância entre as fontes para 19,6% dos medicamentos analisados, o que pode colocar em risco a saúde do lactente ou deixar dúvida quanto ao uso do medicamento ou à prática da amamentação. It is known that the use of medications by mothers can influence the success of breastfeeding. Studies suggest that more than 50% of women in the postpartum period need to use at least one medication 8 . Taking into account the fact that these women may interrupt breastfeeding or not adhere to treatment for fear of exposing the baby to the drug through breast milk 9 , it is believed very important to understand the risks of medications for the mother and the baby, and knowing the existing risk classifications as well.

Palavras
The main risk classifications of drugs in breastfeeding, and also the most cited, are those established by the Ministry of Health of Brazil (MS) 10 , American Academy of Pediatrics (AAP) 11 , WHO 12 and Newton and Hale 13 . They include information on drugs and breastfeeding, guiding health professionals on the use of medicines in this phase of a woman's life 14 ; however, these publications must be constantly updated to afford health care professionals with greater assurance in providing care for breastfeeding mothers, as new medications appear every day on the market and since information is controversial in some situations.
The objective of the present study was to classify the risks of medications used during hospitalization for delivery by breastfeeding women participating in the 2015 Pelotas Birth Cohort. Different sources of information were used, and disagreements between the sources were determined; the possible interferences with breastfeeding or baby health were described.

METHODS
We conducted a cross-sectional study from June to October 2015, which was included in the perinatal follow-up of the 2015 Pelotas Birth Cohort. More details on the 2015 cohort can be found in the methodological article 15 .
Daily, trained interviewers collected data through the mothers' medical records, after hospital discharge, regarding prescription and use of medications during the entire hospitalization period for delivery. The following information was also collected: hospitalization; type of anesthesia (spinal anesthesia; epidural; topical; general; none); and type of delivery (vaginal or cesarean).
The mothers' sociodemographic data were obtained in an interview shortly after delivery, still in the hospital, by trained interviewers. The variables analyzed were: age (categorized as: 13-19, 20-30 and 31-45 years old), economic classification, according to the Brazilian Market Research Association of Research 16 (A, B, C and D/E), self-reported skin color (white, black and brown) and education (categorized as: 0-4, 5-8, 9-11 and 12 or more years of study).
Medications used by mothers were classified according to the risk in breastfeeding by the MS manual 10 , WHO 12 , Newton and Hale's classification 13 and AAP 11 . These classifications place drugs in risk categories for use while breastfeeding, due to unwanted effects on the infant or milk production. The MS manual uses the colors green, red and yellow for the classification of drugs, which indicate, respectively, compatible use with breastfeeding, contraindicated use and do not have reliable information about its use during breastfeeding (judicious use) 10 . The AAP classifies medicines into the following groups: • Group A: compatible; • Group B: judicious use; • Group C: contraindicated; • Group D: effects unknown in infants, but require care 11 .
WHO classifies drugs as compatible use, judicious use and to be avoided during breastfeeding 17 .
Taking into account that each source uses different criteria to classify drugs, we considered them to be concordant so that the classifications could be compared: • The information was entered into a tablet and transferred to the computer, and the data were analyzed in the Stata statistical package version 12.0 (StataCorp., College Station, TX, USA). Descriptive analysis was performed. First, the prevalence of drugs in each risk class (compatible, judicious use, contraindicated) was detailed, and the absolute and relative frequencies of each medication used were then calculated. For continuous variables, we used the mean and standard deviation (SD) or median and interquartile range (IQR) when the distribution was not normal. For categorical variables, we determined the proportion of each category of the variable. The proportion of drugs with concordant classification in each risk category and the total proportion with their 95% confidence intervals were also calculated for the concordant and discordant medications.
The study was approved by the Research Ethics Committee of the School of Medicine, Federal University of Pelotas, under Certificate of Presentation for CAAE the No. 26746414.5.0000.5313.

RESULTS
The study included 1,409 women, with an average age of 27 ± 7 (13 to 45) years. Of these, 70.3% were white, 15.4% black and 14.3% brown. The majority (48.7%) belonged to economic class C, and 33.9% had 9 to 11 years of schooling.
Cesarean was the most frequent delivery form, as was spinal anesthesia, both at a rate of 67%. A detailed description of the sociodemographic and hospitalization characteristics of the sub-sample studied can be found in a previous study 18 . After analysis (not shown in the results, since it was only done for comparative purposes), it was shown that the mothers participating in this sub-study were similar to the others in the cohort with regard to sociodemographic characteristics.
The median number of medications used by mothers was 6 (IQR 25-75 3 -9), with 14,673 being the total number of drugs. The number of different drugs used was 143. Table 1 shows the percentages of drugs used there were considered compatible, judicious use or contraindicated in each of the classifications used. The number of possible medications to be classified was different between the sources, since the sources did not contain all the drugs used. most studied drugs in the category to be avoided during breastfeeding, which we identified as contraindicated (n = 9, 6.3%).
On the other hand, the most widely used drugs that showed disagreements between the classifications were: morphine (n = 905), which was classified by MS as for judicious use, by Newton and Hale as possibly compatible (L3) and by AAP and WHO as compatible; hyoscine (n = 324), classified by MS as judicious use and by AAP as compatible; and metoclopramide (n = 256), considered by MS as compatible and AAP as of unknown effects and which the WHO recommends being avoided during breastfeeding. Metronidazole and haloperidol are considered compatible with breastfeeding according to MS, and WHO recommends avoiding them, but they were respectively used by 1.6 and 0.1% of the mothers in this study. Ciprofloxacin was classified as compatible by AAP, as possibly compatible by Newton and Hale, as for judicious use by MS and as should be avoided by WHO with regard to breastfeeding (Chart 2).

DISCUSSION
The aim of this study was to classify the medications used during hospitalization for delivery regarding the risks of breastfeeding, by using four sources and determining the agreement between them. Of the total number of medications used, almost half were considered compatible with breastfeeding, while 6% were contraindicated. For one-fifth of the drugs, we found disagreements between the classifications.
One limitation of this study we can point out is the difficulty of reconciling, in the same risk category, drugs classified by different sources, as they resort to different nomenclatures to establish the risks, which led us to carry out a reclassification, to try to find concordances and divergences referring to the risk of the medication during breastfeeding. This lack of standardization and the disagreement between the classifications can result in uncertainty among health care professionals, often causing doubt about whether a medication is compatible or not with breastfeeding, which can lead to the discontinuation of breastfeeding or the medication that could be necessary.
The main factors that can explain these disagreements are: • the use of different risk criteria and classifications between the sources; • the type of evidence used by the sources to support their recommendations; • updating each source (review).
In a study by Long and Montouris 19 based on the responses to questionnaires given to 202 professionals who participated in the annual meeting of the American College of Physicians in 2003, of which 92% were physicians, it was reported that more than 50% did not know that most anticonvulsants are safe when taken during breastfeeding. Therefore, constant  update on medication safety is essential to avoid interruption of breastfeeding. Doctors should choose to prescribe medications already studied for their safety in breastfeeding 1 .
With regard to the choice of sources used, the use of the WHO classification was due to the fact that it is known worldwide and that of the Ministry of Health because the study was carried out in this country. The other two were the most cited in scientific articles; we found AAP to be the most cited. The Food and Drug Administration (FDA) classification was not mentioned here, since a new FDA classification came into effect in June 2015, Pregnancy and Lactation Labeling Rule, for use of drugs during pregnancy and lactation. Therefore, considering that it was being incorporated recently and that it did not allow us to compare it with the other classifications that were already being analyzed in the study, we chose not to use it, since it does not place drugs in risk categories, but rather exposes the risks to the infant or to milk production.
Of the four classifications studied, the one with the highest number of medications was that of MS (n = 112), followed by the WHO classification (n = 65). The Newton and Hale and AAP classifications showed fewer drugs (38 and 49, respectively). Looking at the recommendation for use during breastfeeding, the WHO classification had the highest number of contraindicated drugs, appearing to be the most rigorous classification.
Regarding the nomenclature used by the classifications mentioned, all with the exception of Newton and Hale, are very clear and objective. In the case of Newton and Hale, it was difficult to see the difference between L2 (probably compatible) and L3 (possibly compatible). Therefore, to be able to compare the classifications and considering greater safety for the infant, we chose to group L2 and L1 (compatible). L3 was considered as judicious use like in the other classifications, and L4 (possibly dangerous), grouped with L5 (contraindicated).
In the discussion of the results, emphasis was placed on medications that showed disagreements in the study classifications and/or whose effects were considered important. The drugs whose disagreements were greater and that had a higher prevalence of use were morphine, hyoscine, metoclopramide, promethazine, povidone-iodine, atropine and ephedrine.
The narcotic pain reliever morphine (n = 905) is excreted in low concentrations in breast milk 20,21 . WHO and AAP consider its use compatible with breastfeeding 11,12,17 . MS 8 and Newton and Hale 13 consider it to be for judicious use. Although AAP classifies this drug as safe, it recommends that its serum level be measured in babies 11 , a procedure that is difficult in clinical practice 21 . Morphine can cause significant effects on suckling, apnea, bradycardia and cyanosis 13,17 and sedation in infants 10 . Prolonged treatment, at high doses, can increase its plasma concentration, leading to greater levels in breast milk, so the risk-benefit of breastfeeding should be evaluated 13,17,22 . MS recommends that breastfeeding should be stopped when there is maternal dependence on opioids 10 .
Hyoscine (n = 324) is classified as judicious use by MS 10 and compatible with breastfeeding by AAP 11 . Although babies are sensitive to anticholinergic effects 22 , AAP considers hyoscine safe, perhaps due to the fact that it is excreted in breast milk in insignificant amounts, with no side effects reported in infants 10,21 .
REV BRAS EPIDEMIOL 2020; 23: E200026 Metoclopramide (n = 256) is classified as compatible with breastfeeding according to MS 10 and contraindicated by WHO 12 , while AAP 11 classifies it as judicious use. It is excreted in breast milk, accumulating after repeated doses. This drug stimulates the release of prolactin in women with reduced or inadequate milk production 10,17,23 . MS considers the drug compatible when used for a short period. When used for more than four weeks, agitation, sedation and extrapyramidal manifestations have been described in the nursing mother 10 , and infants may experience problems in neural development 17 and symptoms of gastrointestinal discomfort 21 . Metoclopramide apparently does not pose a risk for the infant with the use of maternal doses of up to 45 mg per day 21 .
Promethazine (n = 189) is considered by WHO 12 and MS 10 to be compatible with breastfeeding when taken in a single dose, but Newton and Hale 13 consider it as judicious use. When repetitive doses are used, some adverse effects should be monitored, such as drowsiness 10,17 , dry mouth, extrapyramidal symptoms and risk of apnea 13 in infants. In lactating women, milk flow may be reduced 10 .
As for povidone-iodine (n = 153), repetitive topical applications of the solution should be avoided, as it can be absorbed and concentrate in breast milk, reaching a toxic level in the infant. Therefore, its chronic use is not recommended and it is recommended to monitor the baby's thyroid function 10 . AAP 11 considers the substance to be compatible with breastfeeding and MS 10 for judicious use, while WHO 12 contraindicates its use. It is advisable to monitor infants for hypothyroidism 17 , as it was observed that breastfed newborns of mothers who used this medication had a 25 to 30-fold increase in the rate of congenital hypothyroidism compared to bottle-fed babies 21 . MS reports that there may be an increase in iodine level in breast milk and, consequently, changes in the neonatal concentration of thyroid stimulating hormone 10,21 . This increase in iodine in breast milk causes a strong odor from the baby's skin 11,21 .
Atropine (n = 129) is classified as compatible by WHO 12 and AAP 11 and as judicious use by MS 10 . It is excreted in breast milk in small amounts, with peak plasma concentration occurring at one hour 10 . When used at high doses, it can cause decreased lactation. Although neonates are susceptible to anticholinergic effects, AAP and WHO consider this drug to be safe 22 .
Ephedrine (n = 103) is classified as compatible during lactation by WHO 12 and AAP 11 and as judicious use according to MS 10 . It is recommended that irritability and disturbed sleep in infants be monitored 10,17 ; these effects resolve spontaneously 12 hours after discontinuation of breastfeeding 21 . In mothers, the decrease in milk production must be monitored 10 .
Methylergometrine (n = 89) is classified as compatible with breastfeeding according to WHO 12 and as judicious use by MS 10 . It is excreted in breast milk, which can cause ergotism 24 , vomiting, diarrhea and seizures in the baby 25 , and lactation and prolactin inhibition in the mother 17,24 . During the use of this drug, the infant's growth must be closely monitored 3 .
Benzodiazepine such as diazepam (n = 35) and their metabolites are excreted in breast milk 21 , but WHO classifies them as compatible with breastfeeding if taken in a single dose. In repeated doses, however, they should be avoided when possible 12 . MS 10 and AAP 11 consider the drug as judicious use. If a benzodiazepine is necessary, short-term ones should be preferred 10,17 . This medication, when used by lactating women, has adverse effects on babies, such as slowness, sedation 17 , low suction 10 , weight loss, apnea, irritability, minimal risk of central nervous system depression 26 and electrocardiogram consistent with the use of sedative drug 10 . These effects may be present if breastfeeding occurs less than eight hours after the use of the drug, as the peak concentration in maternal plasma occurs between one and two hours after administration 10 and when the infant is premature or is underweight 17,21 .
The active metabolite (oxazepam) has been found in the newborn's urine, but not in the mother's plasma or milk, even though showing a long half-life ub the baby. It is known that doses greater than 30 mg per day of diazepam should be avoided during lactation, while doses of 10 mg or less do not produce significant adverse effects in infants. However, more recently, significant effects have been noted with doses between 6 and 10 mg 22 .
In regard to metronidazole (n = 23), WHO 12 contraindicates its use, and AAP 11 considers it as judicious use, while MS 10 views it compatible with breastfeeding. There is little systemic absorption with topical use, when used in low doses, and in short-term therapy it does not interfere with breastfeeding 20 . However, if the exposure is prolonged, it presents a potential risk of toxicity 27 . This drug is excreted in breast milk and can cause some adverse effects in infants, such as diarrhea, lactose intolerance 21 and changes in the baby's intestinal flora 6 , in addition to the fact that it imparts an unpleasant taste in the milk. Therefore, it is recommended to discontinue breastfeeding for 12-24 hours after use by the mother 11,20,27 . An alternative to not discontinuing breastfeeding in this case is to collect the breast milk and store it in the refrigerator, in order to breastfeed the baby during this period 12 . Australian studies show that a third of doctors and pharmacists are unaware that metronidazole is compatible with breastfeeding when used as recommended 6 .
As future perspectives, it is important to carry out studies with the drugs used by puerperal women, analyzing their effects on milk production and on important outcomes in the baby in the short and long term, to confirm the compatibility of these substances with this period in a woman's life.

FINAL CONSIDERATIONS
It is important that the mother and health care providers know the risks of the medications taken during breastfeeding, so that it is not interrupted if the use of medications is truly necessary, or that the patient does not stop using any needed medication because of concern that it may cause some risk to the child, or that if the medication is not safe, breastfeeding must be halted or another safer therapeutic option must be sought.