Treating Diabetes During Breastfeeding

Insulin

Insulin is a normal component of human milk. Milk insulin concentrations averaged almost double the maternal fasting serum insulin in one study at 2 weeks postpartum and were a third higher at 4 months postpartum, implying active transport of insulin into milk. Exogenous insulin is used in both type 1 and type 2 diabetes and has been detected in human milk, including biosynthetic insulin. Insulin is not absorbed by the infant, but may have beneficial local activity in the gastrointestinal tract. Of interest to milk banks and neonatal intensive care units is that Holder pasteurization reduces the concentration of endogenous insulin in milk by about half.

Breastfeeding appears to improve postpartum glucose tolerance in mothers with gestational diabetes and reduce the insulin requirement of patients with type 1 diabetes if they are breastfeeding. The cause is thought to be the use of glucose in milk synthesis. The decrease in insulin requirement is quite variable among studies. In one study, insulin requirements were lower than prepregnancy dosage only during the first week postpartum: 54% of prepregnancy dosage on day 2, and 73% on day 3 postpartum. On day 7 postpartum, insulin dosage returned to prepregnancy requirements. However, another study found that dosage requirements did not return to prepregnancy values for up to 6 weeks in some mothers. Other studies have found decreases in postpartum insulin requirements of 13–34%. Close monitoring of blood glucose is particularly important in the postpartum period in nursing mothers using insulin. ¹¹

Metformin

Metformin is the drug of choice for initial treatment of type 2 diabetes for most patients. ⁹ Data from well-conducted studies indicate that metformin levels in milk are low and infants would receive <0.5% of their mothers' weight-adjusted dosage. Although the dose in milk is low, metformin is sometimes detectable in low levels in the serum of breastfed infants. Milk metformin levels are relatively constant during maternal metformin use, so timing of breastfeeding with respect to the drug administration times is of little benefit. One sizeable prospective study found no adverse effects in breastfed infants whose mothers were taking metformin for polycystic ovary syndrome. Although it does not lower plasma glucose, metformin should be used with caution while nursing newborn and premature infants.

Dipeptidyl peptidase-4 inhibitors

The dipeptidyl peptidase-4 inhibitors include alogliptin, linagliptin, saxagliptin, and sitagliptin. No information is available on their excretion into milk or use in lactation. However, most have relatively long half-lives, low protein binding, and good oral absorption, so might pass into milk and to the infant. Unlike some of the other new antidiabetic drugs, the warning in their labeling is to use with caution rather than to avoid altogether. This represents a judgment on the part of the Food and Drug Administration of the risk associated with their use. Saxagliptin has a shorter half-life than the other drugs and linagliptin is >80% protein bound, so these might be better choices among drugs in this class for nursing mothers. These drugs can lower serum glucose occasionally if used with a sulfonylurea, so monitoring infant serum glucose is advisable if symptoms appear.

Glucagon-like peptide-1 receptor agonists

Glucagon-like peptide-1 receptor antagonists include dulaglutide, exenatide, liraglutide, lixisenatide, and semaglutide. These are small peptide drugs that must be administered daily or weekly by subcutaneous injection. Most are inherently long-acting, persisting in plasma for days after an injection. Several have been associated with thyroid C cell tumors. The nondepot form of exenatide is possibly the best choice, with a short elimination half-life and no tumorigenicity warning, but it must be injected daily. Current labeling recommends against nursing with all of these drugs, although there are no reports of their use in breastfed infants.

Sodium–glucose cotransporter 2 inhibitors

Sodium–glucose cotransporter 2 inhibitors, which increase glucose excretion in the urine, include canagliflozin, dapagliflozin, empagliflozin, and ertugliflozin. These drugs have no human lactation information and are not recommended in their respective package inserts because of the theoretical risk of injury to the developing kidney. There are some important pharmacokinetic differences among them, though, especially in their protein binding. Canagliflozin is 99% bound to maternal plasma proteins, meaning that only 1% of the drug in the mother's bloodstream is available for excretion into milk. Dapagliflozin is 91% bound (9% available), empagliflozin is 86% bound (14% available), and ertugliflozin is 94% protein bound (6% available). It is unlikely that any of these drugs would be present in milk in large quantities, but canagliflozin would seem to be the preferred agent among them based on its high protein binding and an incomplete oral bioavailability of 65% in adults. Peak serum concentrations occur about 2 hours after a dose, so peak milk levels would be expected to occur 3 to 4 hours after a dose.

Sulfonylureas

These older drugs have fallen out of favor in the treatment of type 2 diabetes because they can cause hypoglycemia and weight gain. The advantage for nursing mothers is that information exists on the excretion into breast milk for two of the currently used second-generation agents, and some infant safety information is available. The best information is available for glyburide. Among 11 women taking glyburide 5 or 10 mg/day, glyburide was undetectable in milk at several times after the dose. Likewise, two women taking glipizide also had undetectable levels in milk. Three exclusively breastfed infants had normal blood glucose levels while mothers were taking one of these drugs. Glimepiride, the other second-generation sulfonylurea, does not have any information during breastfeeding, and first-generation agents are no longer recommended. ¹⁰ Checking the breastfed infant's blood glucose is advisable during maternal therapy with sulfonylureas.

Thiazolidinediones

This older class includes pioglitazone and rosiglitazone. Both of these drugs are >99% bound to plasma proteins, so amounts in milk are expected to be very low and unlikely to affect the infant. Pioglitazone labeling recommends using the drug with caution during breastfeeding, whereas rosiglitazone is labeled as contraindicated during nursing. Although the reason for this distinction is not obvious, if one of these drugs is needed pioglitazone would be the preferred agent because of the labeling.

Alpha-glucosidase inhibitors

Acarbose and miglitol comprise this class, which prevents glucose absorption from the gastrointestinal tract. Less than 2% of a dose of acarbose is absorbed from the mother's gastrointestinal tract, and only 0.02% of a 100 mg oral dose (0.2% of the weight-adjusted maternal dosage) of miglitol is excreted into breast milk. In addition, these drugs are expected to be poorly absorbed by the infant. It is unlikely that these drugs would affect the infant through breast milk, although no reports of their use by nursing mothers have been published.

Colesevelam

This bile acid-binding resin is occasionally used in type 2 diabetes and lowers A1C to about the same extent as other low-potency antidiabetic agents.^9,10 It is completely nonabsorbable after oral administration, so it would be a good choice for nursing mothers, especially those needing concurrent cholesterol-lowering therapy. No reports of its use by nursing mothers have been published.

Meglitinides

Repaglinide and nateglinide are >98% protein bound, so unlikely to appear in milk in large amounts. Both are labeled as “not recommended” during breastfeeding, which seems to be more strict that “use with caution” and less strict than “do not use.” These drugs can cause hypoglycemia and are probably not good alternatives in nursing mothers until more data become available.

Pramlintide

Pramlintide is an amylin derivative given by subcutaneous injection. Pramlintide has a short elimination half-life and a high molecular weight, so if it is excreted in human milk, it would probably be in low concentrations. Furthermore, because pramlintide is a peptide, it is most likely digested by the nursing infant, rather than systemically absorbed. However, this drug is very rarely used because of the numerous daily injections of the drug and insulin that must be given with its use. It can also cause hypoglycemia, so monitoring of the infant's serum glucose is recommended.