DISCUSSION
Allergic reaction to insulin has been uncommon since the introduction of human recombinant insulin preparations instead of animal insulin; the frequency ranges from 0.1% to 3% during insulin treatment [
1].
The exact prevalence of anaphylaxis and allergy to insulin during pregnancy is unknown, but the condition is very rare during pregnancy due to altered immune reactions [
2,
3]. Pregnancy is often considered a period of pure immunosuppression because of the autoimmune disease improvement and the increased susceptibility to influenza infections during pregnancy. During pregnancy, T cells, B cells, and natural killer cells are transcriptionally downregulated by estrogen and progesterone [
2].
Multiple insulin components-such as pharmaceutical formula additives (protamine, metacresol, or zinc) and latex-and insulin itself can act as the allergen [
1].
The mechanisms of insulin allergy can be of three types: hypersensitivity I, III, and IV. The IgE-mediated type I hypersensitivity is the most common, usually localized at the injection site and rarely causes anaphylaxis [
4]. IgG-mediated type III immune complex-type hypersensitivity and immune T cell-mediated delayed-type type IV hypersensitivity occur infrequently [
5].
IgE-mediated symptoms occur immediately after insulin administration and biphasically followed by a more sustained induration and generalized reaction 4 to 8 hours later. Type III reactions occur at the injection site as subcutaneous nodules after 2 to 6 hours after insulin administration [
4]. Type IV reactions occur usually as a reaction to the retarding agents in insulin preparations [
6].
For accurate diagnosis, skin prick test, patch test, intradermal test, specific IgE and occasionally, skin biopsy can be used. However, history taking is the most important [
4].
In our two cases, eosinophil, IgE and IgG levels were measured and a skin biopsy performed. The IgE level was elevated, IgG level was normal, eosinophilia was observed and skin biopsy showed eosinophilic infiltration. Our two insulin allergy cases were considered type I hypersensitivity because of the elevated IgE. The first case was a more severe form with accompanying systemic symptoms, such as dyspnea and cough, compared with the second case, that showed localized skin lesions only.
Various treatment options are available for managing allergic reactions to insulin [
1]. Antihistamines and corticosteroids can be used for symptom relief. Changing the insulin to different oral hypoglycemic agents or insulin preparations is also important [
1]. Additionally, use of insulin as a continuous subcutaneous infusion can be helpful. Subcutaneous insulin desensitization is a specific immunotherapy that increases insulin dose progressively under close observation [
1,
4]. In our cases, the symptoms were improved after changing the insulin regimen to another oral agent or preparation due to subtle differences in antigenicity.
In case 1, the patient had only 3 weeks to childbirth and presented with insulin antibodies. Therefore, we changed the insulin to metformin and glibenclamide although the drugs were classified as category B and C in pregnancy, respectively. The American College of Obstetrics and Gynecology and the UK National Institute of Health and Care Excellence recommend that either metformin or glibenclamide can be used to treat gestational diabetes [
7,
8].
In case 2, the allergic reaction to insulin occurred when insulin lispro and insulin detemer were used. The symptoms became aggravated after injection of insulin lispro; thus, the insulin lispro was considered the allergen. The patient's laboratory findings showed no insulin antibodies; a previous case where insulin lispro caused localized allergic reactions was reported, but insulin aspart was less immunogenic [
9]. Therefore, insulin lispro was changed to aspart, and no allergic reaction occurred after the insulin regimen was changed. Subsequently, the symptoms were resolved.
Insulin lispro has subtle structural differences from human insulin (substitution of proline with lysine in position 28 and lysine with proline in position 29 of the B chain) and includes inactive ingredients such as glycerol, phenol, metacresol, or zinc [
4,
10]. Conversely, insulin aspart differs from insulin lispro structurally (a single substitution of proline with aspartic acid in position B28) and in doses of metacresol in preparation (lispro 3.15 mg/dL, aspart 1.72 mg/dL) [
10,
11,
12,
13], which may have caused a difference in allergic reaction. Case 2 is interesting because allergy to insulin was thought to be caused by insulin lispro, which has a low immunogenic potency and is usually administered to patients who suffer from insulin allergy [
14].
In summary, we reported two cases of allergic reaction to human insulin in patients with gestational diabetes. These cases present interesting and unique aspects because insulin allergy in gestational diabetes is extremely infrequent owing to its immunosuppressive state [
2]. To our knowledge, less than 10 cases of allergy to insulin in gestational diabetes have been reported globally [
15,
16,
17,
18,
19,
20,
21,
22] and this is the first report of insulin allergy in gestational diabetes in Korea. Therefore, we report two cases of allergic reaction to insulin in patients with gestational diabetes and a review of the literature.