Arch Gynecol Obstet DOI 10.1007/s00404-015-3692-3
REVIEW
Safety of insulin analogs during pregnancy: a meta-analysis ShiShi Lv1 • JiYing Wang1 • Yong Xu1
Received: 9 January 2015 / Accepted: 13 March 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Objective The objective of this study was to assess the safety of four insulin analogs (aspart, lispro, glargine, and detemir) for the treatment of diabetes in pregnancy. Methods We searched Embase, Pubmed, and the Cochrane Central Register for Controlled Trials database through May 31, 2014. All articles were reviewed by two independent researchers, and if a discrepancy was noted, a third researcher was consulted. Results data were summarized by RevMan 5.2 software. Results Our search resulted in the retrieval and screening of 3519 studies. Of those, 24 studies met the eligibility criteria; the studies reported on a total of 3734 women with pre-gestational or gestational diabetes during pregnancy. The use of lispro was associated with lower rates of neonatal jaundice (RR = 0.63) and severe maternal hypoglycemia (RR = 0.33) than regular insulin. Lispro use was also associated with higher birth weight (WMD = 116.44) and an increased incidence of large for gestational age (LGA) births (RR = 1.42) compared with regular insulin. Rates of cesarean section and macrosomia were similar in pregnant women treated with aspart and regular insulin. Birth weights and rates of severe maternal hypoglycemia, respiratory dysfunction syndrome, and neonatal intensive care unit admission were similar after pregnant women were treated with glargine and NPH insulin. Rates of LGA, macrosomia, and neonatal hypoglycemia were similar after pregnant women were treated with detemir and NPH insulin. & Yong Xu [email protected] 1
Department of Endocrinology, Affiliated Hospital of Luzhou Medical College, Luzhou 646000, Sichuan, China
Conclusions Aspart, glargine, and detemir are safe treatment options for diabetes during pregnancy; these insulin analogs did not increase complications for the mothers or fetuses in our study. However, lispro was related to higher birth weight and increased rate of LGA in neonates. More high-quality randomized controlled trials are needed to clarify the best treatment options for diabetes during pregnancy. Keywords
Insulin analogs Pregnancy Meta
Introduction From 1999 to 2005, the proportion of pregnant women with pre-gestational diabetes increased from 10 to 21 % [1], and the prevalence of gestational diabetes is 9.2 % in America [2]. The Hyperglycemia and Adverse Pregnancy Outcomes study confirmed that maternal hyperglycemia is associated with a higher rate of neonatal complications, including high birth weight, increased rate of cesarean section, and elevated risk of neonatal hypoglycemia [3]. Insulin is the preferred hypoglycemic agent during gestation, but there are limitations to the use of regular insulin, including the risk of hypoglycemia, and the possibility of glycemic fluctuations. Insulin analogs overcome the limitations of regular insulin and have been widely used in diabetes. However, studies of insulin analogs during pregnancy for pre-gestational and gestational diabetes report conflicting results. For example, no safety issues were identified in a study of lispro use in pregnancy [4], but the authors observed a higher rate of large for gestational age (LGA) births and higher birth weights with lispro use [5, 6]. Most evaluations of insulin analogs in pregnancy are small
123
Arch Gynecol Obstet
retrospective studies and the results differ. As such, there is a shortage of evidence for the safety of insulin analogs during pregnancy. The purpose of our meta-analysis was to investigate the role of rapid-acting insulin analogs (aspart and lispro) and long-acting insulin analogs (glargine and detemir) in pregnancy. We also aimed to explore whether adverse outcomes are associated with the use of insulin analogs during pregnancy and to provide evidence for the clinical management of diabetes during pregnancy.
Methods Eligibility criteria To be included in our analysis, a study met the following criteria: (1) it was a study of pregnant women who had pregestational or gestational diabetes; (2) study included two groups, one group use regular insulin or NPH and the other use insulin analog (aspart, lispro, glargine, or detemir); (3) it reported fetal or maternal outcomes in both control and experimental groups; and (4) it was a prospective, retrospective, or randomized study. A study was excluded from our analysis if (1) it did not supply human data; (2) it did not clearly specify the initiation or duration of treatment with insulin analogs; or (3) it was not published in English. Data sources and searches We searched Embase, Pubmed, and the Cochrane Central Register for Controlled Trials database. Keywords used in the search included aspart or lispro or glargine or lantus or hoe-901 or optisulin or humalog or detemir or levemir or long-acting insulin analogs or short-acting insulin analogs or insulin analogs or pregnancy or gestation. All databases were searched through May 31, 2014. We collected the following information from the studies: author, publication date, the choice of human insulin and insulin analogs, whether the patients had pre-gestational or gestational diabetes, and related outcomes. We collected information on the baseline characteristics of the pregnant women, including age, body mass index (BMI), duration of diabetes, and hemoglobin A1C (HbA1C) level. We collected information on pregnancy-related complications, including pregnancy-induced hypertension, preeclampsia, severe hypoglycemia, and rate of cesarean section. We also collected information on fetal complications, including birth weight, preterm delivery (\37 weeks), neonatal hypoglycemia, respiratory distress syndrome (RDS), LGA (birth weight [90th percentile for gestational age), neonatal intensive care unit (NICU) admission, macrosomia (birth weight
123
[4000 g), neonatal jaundice, stillbirth, shoulder dystocia, and congenital malformations. Study collection and evaluation of quality In order to evaluate the quality of the studies, all articles were reviewed by two independent researchers. If there was a difference in the opinions of these two researchers, a third reviewer was consulted. We used the Newcastle–Ottawa Quality Assessment Scale standards to evaluate the quality of the cohort and case–control studies, and we used Cochrane Collaboration’s tool to evaluate the quality of the randomized controlled trial (RCT) studies, as recommended by the Cochrane Collaboration. Data synthesis and analysis We analyzed continuous data and dichotomous data with weighted mean differences (WMDs) with 95 % confidence intervals (CIs) and Mantel–Haenszel (M-H) relative risks ratios (RRs) with 95 % CIs. All analyses were performed using Review Manager (RevMan) software, version 5.2, as recommended by the Cochrane Collaboration. To assess heterogeneity, we used the Chi-square and I2 tests. I2 values of 25, 50, and 75 % indicated low, medium, and high heterogeneity, respectively. For all analyses, we use a fixed-effect model, specifically the M-H method. If studies stratified the participants according to pre-gestational or gestational diabetes status prior to analysis, we analyzed the adverse outcomes separately.
Results A total of 24 studies met the inclusion criteria: nine observational studies used lispro, five RCT studies used aspart, eight observational studies used glargine, and one observational and one RCT studies used detemir. Together, the studies reported on 3734 subjects. Figure 1 shows the algorithm of inclusion and exclusion for the collected studies. Meta-analysis of lispro studies Nine of the studies included in our analysis comprised a total of 1561 women with pre-gestational or gestational diabetes who were treated with either insulin lispro (n = 452) or regular insulin (n = 1089) during pregnancy. Baseline characteristics of women treated with lispro and regular insulin were similar in terms of age (WMD = 0.34, 95 % CI -0.27 to 0.95), duration of diabetes (WMD = -0.23, 95 % CI -1.33 to 0.87), and
Arch Gynecol Obstet
28.78–204.11; Fig. 2b) [5–9]. In our analysis, lispro was not related to higher rates of neonatal hypoglycemia, congenital malformation, NICU admission, cesarean section rate, RDS, macrosomia, preterm delivery, or stillbirth (Table 1). Maternal outcomes A decreased risk of severe maternal hypoglycemia was observed after lispro use (RR = 0.33, 95 % CI 0.12–0.89; Fig. 2c) [5, 9]. No significant difference was observed between the rates of preeclampsia and Pregnancy-induced hypertension in the lispro and control groups (Table 1). Meta-analysis of aspart studies Fig. 1 Algorithm of study inclusion
HbA1C at conception (WMD = -0.07, 95 % CI -0.29 to 0.15). The lispro group had a lower BMI (WMD = -1.16, 95 % CI -1.91 to -0.42, I2 = 84 %). Neonatal outcomes The results showed a significantly lower incidence of jaundice with lispro use compared to regular insulin use (RR = 0.63, 95 % CI 0.44–0.90; Table 1) [6, 7, 9, 12]. Lispro was associated with a significantly higher incidence of LGA than regular insulin (RR = 1.42, 95 % CI 1.20–1.69; Fig. 2a) [5–7, 9, 10]. Overall, a higher birth weight was observed after lispro use (WMD = 116.44, 95 % CI
Six of the studies included in our meta-analysis were RCT studies evaluating aspart. These studies reported on a total of 1143 women with gestational or pre-gestational diabetes treated with either insulin aspart (n = 567) or regular insulin (n = 516) during pregnancy. Baseline characteristics were similar in the aspart and control groups in terms of age (WMD = -0.2, 95 % CI -0.72 to 0.32), BMI (WMD = 0.27, 95 % CI -0.16 to 0.71), and HbA1C at entry (WMD = -0.09, 95 % CI -0.12 to 0.05). The risk for macrosomia was not significantly increased in the aspart group (RR = 0.81, 95 % CI 0.42 to 1.58) [13, 14]. No significant difference was observed in the rate of cesarean delivery rate in women receiving aspart compared to women receiving regular insulin (RR = 1.00, 95 % CI 0.92 to 1.08) [13, 15].
Table 1 Maternal and neonatal outcomes after lispro use during pregnancy Outcomes of studies (n)
Lispro Total pts. (n)
RI or NPH Events (n)
Risk ratios (95 % CI)
Total pts. (n)
Events (n)
Neonatal hypoglycemia (4)
305
83
442
100
1.27 (0.99, 1.64)
Congenital malformations (5) NICU admission (2)
293 141
11 34
647 119
31 23
0.85 (0.43, 1.69) 1.25 (0.77, 2.01)
Cesarean section rate (8)
440
274
832
535
0.98 (0.89, 1.07)
RDS (2)
144
14
290
28
1.09 (0.59, 2.01)
Stillbirth (2)
113
2
300
5
1.04 (0.20, 5.42)
Macrosomia (5)
217
38
535
84
0.76 (0.43, 1.34)
Preterm delivery (2)
118
34
316
103
1.26 (0.92, 1.73)
Preeclampsia (4)
252
32
621
45
1.24 (0.81, 1.90)
38
280
99
396
0.63 (0.44, 0.90)
247
40
608
114
0.74 (0.53, 1.03)
Neonatal jaundice (4) Pregnancy-induced Hypertension (4)
Calculated relative risk ratios (RR) of fetal/neonatal and maternal outcomes of interest (M-H fixed-effect model) M-H Mantel–Haenszel, NICU neonatal intensive care unit, NPH neutral protamine Hagedorn, RDS respiratory dysfunction syndrome, RI regular insulin
123
Arch Gynecol Obstet
Fig. 2 Comparison of neonatal and maternal outcomes after exposure to insulin lispro therapy or human insulin therapy in pregnancy. a Large for gestational age birth. b Birth weight. c Severe maternal hypoglycemia
Meta-analysis of glargine studies Our analysis included eight studies evaluating glargine, which comprised a total of 702 women with gestational or pre-gestational diabetes treated with either glargine (n = 331) or NPH insulin (n = 371) during pregnancy. Baseline characteristics of women treated with glargine and NPH were similar in terms of age (WMD = -1.24, 95 % CI -1.86 to -0.61), duration of diabetes (WMD = 1.15, 95 % CI 0.71–1.60), BMI (WMD = -0.52, 95 % CI -1.80 to 0.76), and HbA1C at conception (WMD = -0.15, 95 % CI -0.54 to 0.23).
glargine was not related with higher rates of neonatal hypoglycemia, NICU admission, RDS, neonatal jaundice, LGA, preterm delivery, malformations, or macrosomia (Table 2). Maternal outcomes Glargine was associated with a tendency for reducing the risk, but the difference was not statistically significant (RR = 0.39, 95 % CI 0.11–1.40; Fig. 3b) [17, 19, 20]. Women in the glargine and control groups experienced a similar rate of pregnancy-induced hypertension, preeclampsia, or shoulder dystocia (Table 2).
Neonatal outcomes Meta-analysis of detemir studies Three studies in our meta-analysis stratified their participants by pre-gestational or gestational diabetes status prior to analysis [17, 20, 21]. In order to maintain the baseline characteristics in the two groups, we collected data on pre-gestational and gestational diabetes separately. There were no significant differences in birth weight observed between the glargine and NPH groups (WMD = 9.98, 95 % CI -22.62 to 42.58; Fig. 3a) [16–19, 21, 22]. Overall,
123
Our analysis included two studies evaluating detemir: one was an RCT and the other was a case–control study. The two studies comprised a total of 326 women treated with either insulin detemir (n = 160) or NPH insulin (n = 166) during pregnancy. The rate of LGA was reported in both studies, and no increased risk was observed (RR = 0.85, 95 % CI 0.67–1.08) [23, 24]. The studies also reported the
Arch Gynecol Obstet
Fig. 3 Comparison of neonatal and maternal outcomes after exposure to glargine therapy or regular insulin therapy during pregnancy. a Birth weight. b Severe maternal hypoglycemia
Table 2 Maternal and neonatal outcomes after using glargine during pregnancy Outcomes of studies (n)
Glargine Total pts. (n)
RI/NPH Events (n)
Total pts. (n)
Risk ratio (95 % CI) Events (n)
Neonatal hypoglycemia (7)
304
57
346
63
Neonatal jaundice (6)
272
58
314
59
0.95 (0.70, 1.29) 1.02 (0.74, 1.40)
LGA (4)
165
57
216
75
0.99 (0.76, 1.30)
Macrosomia (4) Pregnancy-induced
157 85
37 3
198 107
39 3
1.16 (0.78, 1.73) 1.26 (0.25, 6.20)
Preeclampsia (5)
257
21
299
33
0.69 (0.39, 1.20)
Preterm delivery (4)
149
24
192
43
0.81 (0.51, 1.26)
Shoulder dystocia (2)
107
1
98
7
0.19 (0.04, 1.02)
Malformation (6)
237
17
271
23
0.77 (0.42, 1.42)
24
261
15
288
1.73 (0.96, 3.10)
110
274
107
307
0.94 (0.82, 1.09)
Hypertension (2)
RDS (6) NICU admission (7)
Calculated relative risk ratio (RR) of fetal/neonatal and maternal outcomes of interest (M-H fixed-effect model) M-H Mantel–Haenszel, NPH neutral protamine Hagedorn, LGA large for gestational age birth weight, RDS respiratory dysfunction syndrome, NICU neonatal intensive care unit
123
Arch Gynecol Obstet
incidence of neonatal hypoglycemia and demonstrated no significantly increased risk of this complication (RR = 0.76, 95 % CI 0.43–1.34) [23, 24]. Detemir was not related with a higher rate of LGA or neonatal hypoglycemia.
Discussion Many women worry about the safety of insulin analogs during pregnancy. In our meta-analysis, we investigated the safety profiles of currently available insulin analogs (both short- and long-acting) during pregnancy in order to provide evidence for the management of diabetes with insulin analogs during pregnancy. We included nine observational studies of lispro. Compared to regular insulin, lispro was associated with a lower rate of severe maternal hypoglycemia during pregnancy. This result can be attributed to the characteristics of lispro, including the fact that it is rapid-acting and rapidly metabolized. However, we were not able to include enough RCTs that evaluated maternal hypoglycemia in our metaanalysis to confirm this result; more investigation is needed to confirm the effects of lispro. Compared with regular insulin, lispro was associated with a significantly higher birth weight and an increased rate of LGA in our analysis. In another meta-analysis, Edson et al. [25] reported no increased risk of adverse outcomes with lispro use during pregnancy compared with the use of human insulin. The meta-analysis by Blanco et al. [26] demonstrated that lispro was associated with a higher incidence of LGA. This latter result is in accordance with our findings, but we included more studies in our analysis [6–9, 12, 26]. Further, Blanco’s study [26] only included pregnant women with type 1 diabetes, but our analysis included women with type 1, type 2, and gestational diabetes. There were no significant differences in the maternal baseline characteristics of age, duration of diabetes, and HbA1C between the lispro and control groups. There was a trend toward a lower HbA1C in the lispro group, but this difference was not significant. The lispro group had a lower BMI than the control group. Therefore, the results cannot be mediated by metabolic control, and perhaps, the findings are related with the use of lispro itself. This finding may be related to insulin-like growth factor (IGF), which exists in the placenta and increases fetal growth. Elhddad et al. [27] affirmed that IGF and the IGF receptor were higher in LGA neonates than in normal-weight neonates. Another study revealed that lispro has a higher affinity than regular insulin for the insulin-like growth factor-1 (IGF-1) receptor [28]. Lispro has substantial homology with IGF-1, which raises the possibility of increased growth in fetuses of pregnant women treated with lispro [29]. However, Boskovic et al. [30] verified that
123
commonly administered doses of lispro do not cross the placenta. We included eight RCTs of aspart. The pregnancy adverse of insulin aspart in pregnancy was not clear, except for macrosomia and cesarean delivery rate. Aspart and lispro are rapid-acting insulin analogs, and the results related to their use were identical in our analysis. Aspart has a similar affinity for the IGF-1 receptor compared to regular insulin, but lispro has a higher affinity for the IGF-1 receptor [28]. But we did not observe an increased risk of macrosomia with lispro and aspart use during pregnancy. Possibly, the results of the lispro analysis can be attributed to other factors known to affect fetal growth, such as smoking habits [11], but such factors were not included in the studies in our analysis. In 2008, guidelines for the treatment of gestational diabetes in Britain demonstrated the safety of aspart and lispro in gestational diabetes [31]. The American Association of Endocrinology recommends that women with diabetes who are treated with insulin and desiring to conceive should be treated with insulin lispro or insulin aspart instead of regular insulin [32]. The results of our analysis suggest that insulin aspart is not likely to adversely affect fetal development. Our results also demonstrated that lispro is associated with significantly higher birth weight and an increased incidence of LGA, which may increase the risk of nerve and muscle damage during delivery. More studies are needed to verify these results. We included a total of eight observational studies of glargine. Glargine has no high peak function, so it can be beneficial in pregnant women with diabetes. In pregnancy, it is imperative to achieve good glycemic control and avoid hypoglycemia to reduce pregnancy complications. In our study, we observed no increased incidence of complications with glargine use. We calculated the adverse outcomes in pre-gestational and gestational diabetes separately, but there was no difference between the groups. Pollex et al. [33] confirmed that glargine does not increase the risk of pregnancy-related complications, which is in accordance with our findings. In our study, we observed no significant differences between the glargine and control groups in the risks of LGA and increased birth weight. Our analysis revealed that glargine did not increase fetal growth. A previous study revealed that glargine has a 6.5fold higher affinity for the IGF-1 receptor than regular insulin [28], and an in vitro study confirmed that glargine stimulates mitogenic action to a greater extent than insulin [34]. This has raised the concern that the use of glargine could affect fetal growth by crossing the placenta, but Pollex et al. [35] demonstrated that glargine cannot cross the placenta to influence fetal growth. We included two studies of insulin detemir in our analysis. We observed no differences between the detemir and NPH groups in the risks of neonatal hypoglycemia and
Arch Gynecol Obstet
LGA. The heterogeneity was not substantial (I2 = 45 %), which can possibly be attributed to the study style and the small number of studies included in our analysis. Callesen et al. [36] reported on a total of 113 women with type 1 diabetes treated with either glargine (n = 67) or detemir (n = 46) in pregnancy. In the article observed, a lower rate of LGA in infants born to women treated with glargine compared to those treated with detemir. But there is no high-quality RCTs of glargine in pregnancy. The American Association of Endocrinology recommends that women with diabetes who successfully used long-acting insulin analogs (glargine or detemir) before conception may continue with this therapy during pregnancy. However, glargine is not approved for use in pregnancy by the Food and Drug Administration [32]. The European Medicines Agency permits the use of detemir in pregnancy. In our study, glargine was not associated with an increased rate of pregnancy-related complications. There are limitations to our analysis. We included a small number of studies, and most of them were observational in design, as studies with lispro and glargine are all observational. Patients are not randomized in an observational study, so a subject selection bias may be present in our analysis. Retrospective studies confer another limitation, as data may be gathered in an inconsistent manner. Also, we excluded non-English articles from our study, and insufficient data in the studies did not allow statistical adjustment of the combined comparisons. In addition, pregnancy complications can be affected by characteristics that were not measured, such as socioeconomic status and race. Little information about those factors was available for the participants in our analysis. Shankardassk et al. [37] demonstrated that adverse perinatal outcomes are related to lower socioeconomic position, and the fact that the economics of a society affect pregnancy outcomes is widespread [38]. In addition, parental education and income influence pregnancy outcomes [39]. Household income is weakly related to the risk of preterm birth [37]. Other factors, such as weight gain during pregnancy, affect pregnancy complications, and Sciffres et al. [40] reported that excess maternal weight leads to a higher incidence of LGA in obese and overweight women with diabetes compared to normal-weight pregnant women with diabetes. Some of these factors were partially reported in some of the studies in our analysis, but there was not enough information for us to calculate risks. In conclusion, we did not observe a higher rate of pregnancy-related complications with aspart, detemir, or glargine during pregnancy. Lispro was associated with lower rates of neonatal jaundice and severe maternal hypoglycemia and slightly higher risks of increased birth weight and LGA. Since all of the studies using lispro and glargine included in our analysis were case–control or
cohort studies, the results of our analysis need to be verified by high-quality RCTs. Conflict of interest We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service, and/or company that could be construed as influencing the position presented in, or the review of, the manuscript entitled, ‘‘Safety of insulin analogs during pregnancy: a meta-analysis.’’
References 1. Lawrence JM, Contreras R, Chen W et al (2008) Trends in the prevalence of preexisting diabetes and gestational diabetes mellitus among a racially/ethnically diverse population of pregnant women, 1999-2005. Diabetes Care 31:899–904 2. DeSisto CL, Kim SY, Charma AJ (2014) Prevalence estimates of gestational diabetesmellitus in the United States, Pregnancy Risk Assessment Monitoring System (PRAMS), 2007–2010. Prev Chronic Dis 19(11):E104 3. HAPO Study Cooperative Research Group, Metaqer BE, Lowe LP et al (2008) Hyperglycemia and adverse pregnancy outcomes. N Engl J Med 358:1991–2002 4. Edson EJ, Bracco OL, Vambergue A et al (2010) Managing diabetes during pregnancy with insulin lispro:a safe alternative to human insulin. Endocr Pract 16:1020–1027 5. Lapolla A, Dalfra MG, Spezia R et al (2008) Outcome of pregnancy in type 1 diabetic patients treated with insulin lispro or regular insulin: an Italian experience. Acta Diabetol 45:61–66 6. Durnwald CP, Landon MB (2008) A comparison of lispro and regular insulin for the management of type 1 and type 2 diabetes in pregnancy. J Matern Fetal Neonatal Med 21:309–313 7. Banerjee M, Bhattacharya A, Hughes SM et al (2009) Efficacy of insulin lispro in pregnancies complicated with pregestational diabetes mellitus. Pract Diabetes Int 26:366–370 8. Aydin Y, Berker D, Direktor N et al (2008) Is insulin lispro safe in pregnant women: does it cause any adverse outcomes on infants or mothers? Diabetes Res Clin Pract 80(44):4–448 9. Garcı´a-Domı´nguez M, Herranz L, Hillman N et al (2011) Use of insulin lispro during pregnancy in women with pregestational diabetes mellitus. Med Clin 137:581–586 10. Cypryk K, Sobczak M, Pertyn´ska-Marczewska M et al (2004) Pregnancy complications and perinatal outcome in diabetic women treated with Humalog (insulin lispro) or regular human insulin during pregnancy. Med Sci Monit 10:PI29–PI32 11. Evers IM, de Valk HW, Mol BW et al (2002) Macrosomia despite good glycaemic control in type1 diabetic pregnancy results of a nationalwide study in the Netherlands. Diabetologia 45:1484–1489 12. Colatrella A, Visalli N, Abbruzzese S et al (2013) Comparison of insulin lispro protamine suspension with NPH insulin in pregnant women with type 2 and gestational diabetes mellitus: maternal and perinatal outcomes. Int J Endocrinol 2013:151975 13. Balaji V, Balaji MS, Alexander C et al (2012) Premixed insulin aspart 30 (BIAsp 30) versus premixed human insulin 30 (BHI 30) in gestational diabetes mellitus: a randomized open-label controlled study. Gynecol Endocrinol 28:529–532 14. Balaji V, Balaji MS, Alexander C et al (2010) Premixed insulin aspart 30 (Bia sp 30) vs premixed human insulin 30 (Bhi 30) in gestational diabetes mellitus—a pilot study. J Assoc Physician India 58:99–101 15. Hod M, Damm P, Kaaja R et al (2008) Fetal and perinatal outcomes in type 1 diabetes pregnancy: a randomized study
123
Arch Gynecol Obstet
16.
17.
18.
19.
20.
21. 22.
23.
24.
25.
26.
27.
comparing insulin aspart with human insulin in 322 subjects. Am J Obstet Gynecol 198:186.e1–186.e7 Poyhonen-Alho M, Ronnemaa T, Saltevo J et al (2007) Use of insulin glargine during pregnancy. Acta Obstetricia et Gynecologica 86:1171–1174 Fang YM, MacKeen D, Egan JF et al (2009) Insulin glargine compared with Neutral Protamine Hagedorn insulin in the treatment of pregnant diabetics. J Matern Fetal Neonatal Med 22:249–253 Egerman RS, Ramsey RD, Kao LW et al (2009) Perinatal outcomes in pregnancies managed with antenatal insulin glargine. Am J Perinatol 26:591–595 Imbergamo MP, Amato MC, Sciortino G et al (2008) use of glargine in pregnant women with type 1 diabetes mellitus: a casecontrol study. Clin Ther 30:1476–1484 Negrato CA, Rafacho A, Negrato G et al (2010) Glargine vs. NPH insulin therapy in pregnancies complicated by diabetes: an observational cohort study. Diabetes Res Clin Pract 89:46–51 Price N, Bartlett C, Gillmer M (2007) Use of insulin glargine during pregnancy: a case–control pilot study. BJOG 114:453–457 Smith JG, Manuck TA, White J et al (2009) Insulin glargine versus neutral protamine hagedorn insulin for treatment of diabetes in pregnancy. Am J Perinatol 26:57–62 Imbergamo MP, Amato MC, Amodeo R et al (2012) Use of basal insulin analog detemir in pregnant women with type1 diabetes:a case-control retrospective study. Giorn Ital Ostet Ginecol 34:491–498 Hod M, Mathiesen ER, Jovanovic L et al (2014) A randomized trial comparing perinatal outcomes using insulin detemir or neutral protamine Hagedorn in type 1 diabetes. J Matern Fetal Neonatal Med 27:7–13 Edson EJ, Bracco OL, Vambergue A et al (2010) Managing diabetes during pregnancy with insulin lispro :a safe alternative to human insulin. Endocr Pract 16:1020–1027 Blanco CG, Ballesteros AC, Saladich IG et al (2011) Glycemic control and pregnancy outcomes in women with type 1 Diabetes mellitus using lispro versus regular insulin: a systematic review and meta-analysis. Diabetes Technol Ther 13:907–911 Elhddad AS, Lashen H (2013) Fetal growthin relation tomaternal and fetal IGF-axes: a systematic review and meta-analysis. Acta Obstet Gynecol Scand 92:997–1006
123
28. Kurtzhals P, Scha¨ffer L, Sørensen A et al (2000) Correlations of receptor binding and metabolic and mitogenic potencies of insulin analogs designed for clinical use. Diabetes 49:999–1005 29. Wilde MI (1997) Mc Tavish D. Insulin lispro: a review of its pharmacological properties and therapeutic use in the management of diabetes mellitus. Drugs 54:597–614 30. Boskovic R, Knie B, Feig DS et al (2003) Transfer of insulin lispro across the human placenta. Diabetes Care 26:1390–1394 31. The Guideline Development Group (2008) Management of diabetes from preconception to the postnatal period: summary of NICE guidance. BMJ 336:714–717 32. Blumer I, Hadar E, Hadden DR et al (2013) Diabetes and Pregnancy: an endocrine society clinical practice guideline. J Clin Endocrinol Metab 98:4227–4249 33. Pollex E, Moretti ME, Koren G et al (2011) Safety of Insulin glargine use in pregnancy: a systematic review and meta-analysis. Ann Pharmacother 45:9–16 34. Sciacca L, Cassarino MF, Genua M et al (2010) Insulin analogues differently activate insulin receptor isoforms and post-receptor signalling. Diabetologia 53:1743–1753 35. Pollex EK, Feig DS, Lubetsky A et al (2010) Insulin glargine safety in pregnancy: a transplacental transfer study. Diabetes Care 33:29–33 36. Callesen NF, Damm J, Mathiesen JM et al (2013) Treatment with the long-acting insulin analogues detemir or glargine during pregnancy in women with type 1 diabetes:comparison of glycaemic control and pregnancy outcome. J Mater Fetal Neonatal Med 26:588–592 37. Shankardass K, O’campo P, Dodds L et al (2014) Magnitude of income- related disparities in adverse perinatal outcomes. BMC Pregnancy Childbith 14:96 38. Blumenshine P, Egerter S, Barclay CJ et al (2010) Socioeconomic disparities in adverse birth outcomes: a systematic review. Am J Prev Med 39:263–272 39. Mortense LH (2013) Socioeconomic inequality in birth weight and gestational age in Denmark 1996-2007: using a family-based approach to explore alternative explanations. Soc Sci Med 76:1–7 40. Scifres CM, Feqhali MN, Althouse AD et al (2014) Effect of excess gestational weight gain on pregnancy outcomes in women with type1 diabetes. Obstet Gynecol 123:1295–1302
REVIEW
Safety of insulin analogs during pregnancy: a meta-analysis ShiShi Lv1 • JiYing Wang1 • Yong Xu1
Received: 9 January 2015 / Accepted: 13 March 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Objective The objective of this study was to assess the safety of four insulin analogs (aspart, lispro, glargine, and detemir) for the treatment of diabetes in pregnancy. Methods We searched Embase, Pubmed, and the Cochrane Central Register for Controlled Trials database through May 31, 2014. All articles were reviewed by two independent researchers, and if a discrepancy was noted, a third researcher was consulted. Results data were summarized by RevMan 5.2 software. Results Our search resulted in the retrieval and screening of 3519 studies. Of those, 24 studies met the eligibility criteria; the studies reported on a total of 3734 women with pre-gestational or gestational diabetes during pregnancy. The use of lispro was associated with lower rates of neonatal jaundice (RR = 0.63) and severe maternal hypoglycemia (RR = 0.33) than regular insulin. Lispro use was also associated with higher birth weight (WMD = 116.44) and an increased incidence of large for gestational age (LGA) births (RR = 1.42) compared with regular insulin. Rates of cesarean section and macrosomia were similar in pregnant women treated with aspart and regular insulin. Birth weights and rates of severe maternal hypoglycemia, respiratory dysfunction syndrome, and neonatal intensive care unit admission were similar after pregnant women were treated with glargine and NPH insulin. Rates of LGA, macrosomia, and neonatal hypoglycemia were similar after pregnant women were treated with detemir and NPH insulin. & Yong Xu [email protected] 1
Department of Endocrinology, Affiliated Hospital of Luzhou Medical College, Luzhou 646000, Sichuan, China
Conclusions Aspart, glargine, and detemir are safe treatment options for diabetes during pregnancy; these insulin analogs did not increase complications for the mothers or fetuses in our study. However, lispro was related to higher birth weight and increased rate of LGA in neonates. More high-quality randomized controlled trials are needed to clarify the best treatment options for diabetes during pregnancy. Keywords
Insulin analogs Pregnancy Meta
Introduction From 1999 to 2005, the proportion of pregnant women with pre-gestational diabetes increased from 10 to 21 % [1], and the prevalence of gestational diabetes is 9.2 % in America [2]. The Hyperglycemia and Adverse Pregnancy Outcomes study confirmed that maternal hyperglycemia is associated with a higher rate of neonatal complications, including high birth weight, increased rate of cesarean section, and elevated risk of neonatal hypoglycemia [3]. Insulin is the preferred hypoglycemic agent during gestation, but there are limitations to the use of regular insulin, including the risk of hypoglycemia, and the possibility of glycemic fluctuations. Insulin analogs overcome the limitations of regular insulin and have been widely used in diabetes. However, studies of insulin analogs during pregnancy for pre-gestational and gestational diabetes report conflicting results. For example, no safety issues were identified in a study of lispro use in pregnancy [4], but the authors observed a higher rate of large for gestational age (LGA) births and higher birth weights with lispro use [5, 6]. Most evaluations of insulin analogs in pregnancy are small
123
Arch Gynecol Obstet
retrospective studies and the results differ. As such, there is a shortage of evidence for the safety of insulin analogs during pregnancy. The purpose of our meta-analysis was to investigate the role of rapid-acting insulin analogs (aspart and lispro) and long-acting insulin analogs (glargine and detemir) in pregnancy. We also aimed to explore whether adverse outcomes are associated with the use of insulin analogs during pregnancy and to provide evidence for the clinical management of diabetes during pregnancy.
Methods Eligibility criteria To be included in our analysis, a study met the following criteria: (1) it was a study of pregnant women who had pregestational or gestational diabetes; (2) study included two groups, one group use regular insulin or NPH and the other use insulin analog (aspart, lispro, glargine, or detemir); (3) it reported fetal or maternal outcomes in both control and experimental groups; and (4) it was a prospective, retrospective, or randomized study. A study was excluded from our analysis if (1) it did not supply human data; (2) it did not clearly specify the initiation or duration of treatment with insulin analogs; or (3) it was not published in English. Data sources and searches We searched Embase, Pubmed, and the Cochrane Central Register for Controlled Trials database. Keywords used in the search included aspart or lispro or glargine or lantus or hoe-901 or optisulin or humalog or detemir or levemir or long-acting insulin analogs or short-acting insulin analogs or insulin analogs or pregnancy or gestation. All databases were searched through May 31, 2014. We collected the following information from the studies: author, publication date, the choice of human insulin and insulin analogs, whether the patients had pre-gestational or gestational diabetes, and related outcomes. We collected information on the baseline characteristics of the pregnant women, including age, body mass index (BMI), duration of diabetes, and hemoglobin A1C (HbA1C) level. We collected information on pregnancy-related complications, including pregnancy-induced hypertension, preeclampsia, severe hypoglycemia, and rate of cesarean section. We also collected information on fetal complications, including birth weight, preterm delivery (\37 weeks), neonatal hypoglycemia, respiratory distress syndrome (RDS), LGA (birth weight [90th percentile for gestational age), neonatal intensive care unit (NICU) admission, macrosomia (birth weight
123
[4000 g), neonatal jaundice, stillbirth, shoulder dystocia, and congenital malformations. Study collection and evaluation of quality In order to evaluate the quality of the studies, all articles were reviewed by two independent researchers. If there was a difference in the opinions of these two researchers, a third reviewer was consulted. We used the Newcastle–Ottawa Quality Assessment Scale standards to evaluate the quality of the cohort and case–control studies, and we used Cochrane Collaboration’s tool to evaluate the quality of the randomized controlled trial (RCT) studies, as recommended by the Cochrane Collaboration. Data synthesis and analysis We analyzed continuous data and dichotomous data with weighted mean differences (WMDs) with 95 % confidence intervals (CIs) and Mantel–Haenszel (M-H) relative risks ratios (RRs) with 95 % CIs. All analyses were performed using Review Manager (RevMan) software, version 5.2, as recommended by the Cochrane Collaboration. To assess heterogeneity, we used the Chi-square and I2 tests. I2 values of 25, 50, and 75 % indicated low, medium, and high heterogeneity, respectively. For all analyses, we use a fixed-effect model, specifically the M-H method. If studies stratified the participants according to pre-gestational or gestational diabetes status prior to analysis, we analyzed the adverse outcomes separately.
Results A total of 24 studies met the inclusion criteria: nine observational studies used lispro, five RCT studies used aspart, eight observational studies used glargine, and one observational and one RCT studies used detemir. Together, the studies reported on 3734 subjects. Figure 1 shows the algorithm of inclusion and exclusion for the collected studies. Meta-analysis of lispro studies Nine of the studies included in our analysis comprised a total of 1561 women with pre-gestational or gestational diabetes who were treated with either insulin lispro (n = 452) or regular insulin (n = 1089) during pregnancy. Baseline characteristics of women treated with lispro and regular insulin were similar in terms of age (WMD = 0.34, 95 % CI -0.27 to 0.95), duration of diabetes (WMD = -0.23, 95 % CI -1.33 to 0.87), and
Arch Gynecol Obstet
28.78–204.11; Fig. 2b) [5–9]. In our analysis, lispro was not related to higher rates of neonatal hypoglycemia, congenital malformation, NICU admission, cesarean section rate, RDS, macrosomia, preterm delivery, or stillbirth (Table 1). Maternal outcomes A decreased risk of severe maternal hypoglycemia was observed after lispro use (RR = 0.33, 95 % CI 0.12–0.89; Fig. 2c) [5, 9]. No significant difference was observed between the rates of preeclampsia and Pregnancy-induced hypertension in the lispro and control groups (Table 1). Meta-analysis of aspart studies Fig. 1 Algorithm of study inclusion
HbA1C at conception (WMD = -0.07, 95 % CI -0.29 to 0.15). The lispro group had a lower BMI (WMD = -1.16, 95 % CI -1.91 to -0.42, I2 = 84 %). Neonatal outcomes The results showed a significantly lower incidence of jaundice with lispro use compared to regular insulin use (RR = 0.63, 95 % CI 0.44–0.90; Table 1) [6, 7, 9, 12]. Lispro was associated with a significantly higher incidence of LGA than regular insulin (RR = 1.42, 95 % CI 1.20–1.69; Fig. 2a) [5–7, 9, 10]. Overall, a higher birth weight was observed after lispro use (WMD = 116.44, 95 % CI
Six of the studies included in our meta-analysis were RCT studies evaluating aspart. These studies reported on a total of 1143 women with gestational or pre-gestational diabetes treated with either insulin aspart (n = 567) or regular insulin (n = 516) during pregnancy. Baseline characteristics were similar in the aspart and control groups in terms of age (WMD = -0.2, 95 % CI -0.72 to 0.32), BMI (WMD = 0.27, 95 % CI -0.16 to 0.71), and HbA1C at entry (WMD = -0.09, 95 % CI -0.12 to 0.05). The risk for macrosomia was not significantly increased in the aspart group (RR = 0.81, 95 % CI 0.42 to 1.58) [13, 14]. No significant difference was observed in the rate of cesarean delivery rate in women receiving aspart compared to women receiving regular insulin (RR = 1.00, 95 % CI 0.92 to 1.08) [13, 15].
Table 1 Maternal and neonatal outcomes after lispro use during pregnancy Outcomes of studies (n)
Lispro Total pts. (n)
RI or NPH Events (n)
Risk ratios (95 % CI)
Total pts. (n)
Events (n)
Neonatal hypoglycemia (4)
305
83
442
100
1.27 (0.99, 1.64)
Congenital malformations (5) NICU admission (2)
293 141
11 34
647 119
31 23
0.85 (0.43, 1.69) 1.25 (0.77, 2.01)
Cesarean section rate (8)
440
274
832
535
0.98 (0.89, 1.07)
RDS (2)
144
14
290
28
1.09 (0.59, 2.01)
Stillbirth (2)
113
2
300
5
1.04 (0.20, 5.42)
Macrosomia (5)
217
38
535
84
0.76 (0.43, 1.34)
Preterm delivery (2)
118
34
316
103
1.26 (0.92, 1.73)
Preeclampsia (4)
252
32
621
45
1.24 (0.81, 1.90)
38
280
99
396
0.63 (0.44, 0.90)
247
40
608
114
0.74 (0.53, 1.03)
Neonatal jaundice (4) Pregnancy-induced Hypertension (4)
Calculated relative risk ratios (RR) of fetal/neonatal and maternal outcomes of interest (M-H fixed-effect model) M-H Mantel–Haenszel, NICU neonatal intensive care unit, NPH neutral protamine Hagedorn, RDS respiratory dysfunction syndrome, RI regular insulin
123
Arch Gynecol Obstet
Fig. 2 Comparison of neonatal and maternal outcomes after exposure to insulin lispro therapy or human insulin therapy in pregnancy. a Large for gestational age birth. b Birth weight. c Severe maternal hypoglycemia
Meta-analysis of glargine studies Our analysis included eight studies evaluating glargine, which comprised a total of 702 women with gestational or pre-gestational diabetes treated with either glargine (n = 331) or NPH insulin (n = 371) during pregnancy. Baseline characteristics of women treated with glargine and NPH were similar in terms of age (WMD = -1.24, 95 % CI -1.86 to -0.61), duration of diabetes (WMD = 1.15, 95 % CI 0.71–1.60), BMI (WMD = -0.52, 95 % CI -1.80 to 0.76), and HbA1C at conception (WMD = -0.15, 95 % CI -0.54 to 0.23).
glargine was not related with higher rates of neonatal hypoglycemia, NICU admission, RDS, neonatal jaundice, LGA, preterm delivery, malformations, or macrosomia (Table 2). Maternal outcomes Glargine was associated with a tendency for reducing the risk, but the difference was not statistically significant (RR = 0.39, 95 % CI 0.11–1.40; Fig. 3b) [17, 19, 20]. Women in the glargine and control groups experienced a similar rate of pregnancy-induced hypertension, preeclampsia, or shoulder dystocia (Table 2).
Neonatal outcomes Meta-analysis of detemir studies Three studies in our meta-analysis stratified their participants by pre-gestational or gestational diabetes status prior to analysis [17, 20, 21]. In order to maintain the baseline characteristics in the two groups, we collected data on pre-gestational and gestational diabetes separately. There were no significant differences in birth weight observed between the glargine and NPH groups (WMD = 9.98, 95 % CI -22.62 to 42.58; Fig. 3a) [16–19, 21, 22]. Overall,
123
Our analysis included two studies evaluating detemir: one was an RCT and the other was a case–control study. The two studies comprised a total of 326 women treated with either insulin detemir (n = 160) or NPH insulin (n = 166) during pregnancy. The rate of LGA was reported in both studies, and no increased risk was observed (RR = 0.85, 95 % CI 0.67–1.08) [23, 24]. The studies also reported the
Arch Gynecol Obstet
Fig. 3 Comparison of neonatal and maternal outcomes after exposure to glargine therapy or regular insulin therapy during pregnancy. a Birth weight. b Severe maternal hypoglycemia
Table 2 Maternal and neonatal outcomes after using glargine during pregnancy Outcomes of studies (n)
Glargine Total pts. (n)
RI/NPH Events (n)
Total pts. (n)
Risk ratio (95 % CI) Events (n)
Neonatal hypoglycemia (7)
304
57
346
63
Neonatal jaundice (6)
272
58
314
59
0.95 (0.70, 1.29) 1.02 (0.74, 1.40)
LGA (4)
165
57
216
75
0.99 (0.76, 1.30)
Macrosomia (4) Pregnancy-induced
157 85
37 3
198 107
39 3
1.16 (0.78, 1.73) 1.26 (0.25, 6.20)
Preeclampsia (5)
257
21
299
33
0.69 (0.39, 1.20)
Preterm delivery (4)
149
24
192
43
0.81 (0.51, 1.26)
Shoulder dystocia (2)
107
1
98
7
0.19 (0.04, 1.02)
Malformation (6)
237
17
271
23
0.77 (0.42, 1.42)
24
261
15
288
1.73 (0.96, 3.10)
110
274
107
307
0.94 (0.82, 1.09)
Hypertension (2)
RDS (6) NICU admission (7)
Calculated relative risk ratio (RR) of fetal/neonatal and maternal outcomes of interest (M-H fixed-effect model) M-H Mantel–Haenszel, NPH neutral protamine Hagedorn, LGA large for gestational age birth weight, RDS respiratory dysfunction syndrome, NICU neonatal intensive care unit
123
Arch Gynecol Obstet
incidence of neonatal hypoglycemia and demonstrated no significantly increased risk of this complication (RR = 0.76, 95 % CI 0.43–1.34) [23, 24]. Detemir was not related with a higher rate of LGA or neonatal hypoglycemia.
Discussion Many women worry about the safety of insulin analogs during pregnancy. In our meta-analysis, we investigated the safety profiles of currently available insulin analogs (both short- and long-acting) during pregnancy in order to provide evidence for the management of diabetes with insulin analogs during pregnancy. We included nine observational studies of lispro. Compared to regular insulin, lispro was associated with a lower rate of severe maternal hypoglycemia during pregnancy. This result can be attributed to the characteristics of lispro, including the fact that it is rapid-acting and rapidly metabolized. However, we were not able to include enough RCTs that evaluated maternal hypoglycemia in our metaanalysis to confirm this result; more investigation is needed to confirm the effects of lispro. Compared with regular insulin, lispro was associated with a significantly higher birth weight and an increased rate of LGA in our analysis. In another meta-analysis, Edson et al. [25] reported no increased risk of adverse outcomes with lispro use during pregnancy compared with the use of human insulin. The meta-analysis by Blanco et al. [26] demonstrated that lispro was associated with a higher incidence of LGA. This latter result is in accordance with our findings, but we included more studies in our analysis [6–9, 12, 26]. Further, Blanco’s study [26] only included pregnant women with type 1 diabetes, but our analysis included women with type 1, type 2, and gestational diabetes. There were no significant differences in the maternal baseline characteristics of age, duration of diabetes, and HbA1C between the lispro and control groups. There was a trend toward a lower HbA1C in the lispro group, but this difference was not significant. The lispro group had a lower BMI than the control group. Therefore, the results cannot be mediated by metabolic control, and perhaps, the findings are related with the use of lispro itself. This finding may be related to insulin-like growth factor (IGF), which exists in the placenta and increases fetal growth. Elhddad et al. [27] affirmed that IGF and the IGF receptor were higher in LGA neonates than in normal-weight neonates. Another study revealed that lispro has a higher affinity than regular insulin for the insulin-like growth factor-1 (IGF-1) receptor [28]. Lispro has substantial homology with IGF-1, which raises the possibility of increased growth in fetuses of pregnant women treated with lispro [29]. However, Boskovic et al. [30] verified that
123
commonly administered doses of lispro do not cross the placenta. We included eight RCTs of aspart. The pregnancy adverse of insulin aspart in pregnancy was not clear, except for macrosomia and cesarean delivery rate. Aspart and lispro are rapid-acting insulin analogs, and the results related to their use were identical in our analysis. Aspart has a similar affinity for the IGF-1 receptor compared to regular insulin, but lispro has a higher affinity for the IGF-1 receptor [28]. But we did not observe an increased risk of macrosomia with lispro and aspart use during pregnancy. Possibly, the results of the lispro analysis can be attributed to other factors known to affect fetal growth, such as smoking habits [11], but such factors were not included in the studies in our analysis. In 2008, guidelines for the treatment of gestational diabetes in Britain demonstrated the safety of aspart and lispro in gestational diabetes [31]. The American Association of Endocrinology recommends that women with diabetes who are treated with insulin and desiring to conceive should be treated with insulin lispro or insulin aspart instead of regular insulin [32]. The results of our analysis suggest that insulin aspart is not likely to adversely affect fetal development. Our results also demonstrated that lispro is associated with significantly higher birth weight and an increased incidence of LGA, which may increase the risk of nerve and muscle damage during delivery. More studies are needed to verify these results. We included a total of eight observational studies of glargine. Glargine has no high peak function, so it can be beneficial in pregnant women with diabetes. In pregnancy, it is imperative to achieve good glycemic control and avoid hypoglycemia to reduce pregnancy complications. In our study, we observed no increased incidence of complications with glargine use. We calculated the adverse outcomes in pre-gestational and gestational diabetes separately, but there was no difference between the groups. Pollex et al. [33] confirmed that glargine does not increase the risk of pregnancy-related complications, which is in accordance with our findings. In our study, we observed no significant differences between the glargine and control groups in the risks of LGA and increased birth weight. Our analysis revealed that glargine did not increase fetal growth. A previous study revealed that glargine has a 6.5fold higher affinity for the IGF-1 receptor than regular insulin [28], and an in vitro study confirmed that glargine stimulates mitogenic action to a greater extent than insulin [34]. This has raised the concern that the use of glargine could affect fetal growth by crossing the placenta, but Pollex et al. [35] demonstrated that glargine cannot cross the placenta to influence fetal growth. We included two studies of insulin detemir in our analysis. We observed no differences between the detemir and NPH groups in the risks of neonatal hypoglycemia and
Arch Gynecol Obstet
LGA. The heterogeneity was not substantial (I2 = 45 %), which can possibly be attributed to the study style and the small number of studies included in our analysis. Callesen et al. [36] reported on a total of 113 women with type 1 diabetes treated with either glargine (n = 67) or detemir (n = 46) in pregnancy. In the article observed, a lower rate of LGA in infants born to women treated with glargine compared to those treated with detemir. But there is no high-quality RCTs of glargine in pregnancy. The American Association of Endocrinology recommends that women with diabetes who successfully used long-acting insulin analogs (glargine or detemir) before conception may continue with this therapy during pregnancy. However, glargine is not approved for use in pregnancy by the Food and Drug Administration [32]. The European Medicines Agency permits the use of detemir in pregnancy. In our study, glargine was not associated with an increased rate of pregnancy-related complications. There are limitations to our analysis. We included a small number of studies, and most of them were observational in design, as studies with lispro and glargine are all observational. Patients are not randomized in an observational study, so a subject selection bias may be present in our analysis. Retrospective studies confer another limitation, as data may be gathered in an inconsistent manner. Also, we excluded non-English articles from our study, and insufficient data in the studies did not allow statistical adjustment of the combined comparisons. In addition, pregnancy complications can be affected by characteristics that were not measured, such as socioeconomic status and race. Little information about those factors was available for the participants in our analysis. Shankardassk et al. [37] demonstrated that adverse perinatal outcomes are related to lower socioeconomic position, and the fact that the economics of a society affect pregnancy outcomes is widespread [38]. In addition, parental education and income influence pregnancy outcomes [39]. Household income is weakly related to the risk of preterm birth [37]. Other factors, such as weight gain during pregnancy, affect pregnancy complications, and Sciffres et al. [40] reported that excess maternal weight leads to a higher incidence of LGA in obese and overweight women with diabetes compared to normal-weight pregnant women with diabetes. Some of these factors were partially reported in some of the studies in our analysis, but there was not enough information for us to calculate risks. In conclusion, we did not observe a higher rate of pregnancy-related complications with aspart, detemir, or glargine during pregnancy. Lispro was associated with lower rates of neonatal jaundice and severe maternal hypoglycemia and slightly higher risks of increased birth weight and LGA. Since all of the studies using lispro and glargine included in our analysis were case–control or
cohort studies, the results of our analysis need to be verified by high-quality RCTs. Conflict of interest We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service, and/or company that could be construed as influencing the position presented in, or the review of, the manuscript entitled, ‘‘Safety of insulin analogs during pregnancy: a meta-analysis.’’
References 1. Lawrence JM, Contreras R, Chen W et al (2008) Trends in the prevalence of preexisting diabetes and gestational diabetes mellitus among a racially/ethnically diverse population of pregnant women, 1999-2005. Diabetes Care 31:899–904 2. DeSisto CL, Kim SY, Charma AJ (2014) Prevalence estimates of gestational diabetesmellitus in the United States, Pregnancy Risk Assessment Monitoring System (PRAMS), 2007–2010. Prev Chronic Dis 19(11):E104 3. HAPO Study Cooperative Research Group, Metaqer BE, Lowe LP et al (2008) Hyperglycemia and adverse pregnancy outcomes. N Engl J Med 358:1991–2002 4. Edson EJ, Bracco OL, Vambergue A et al (2010) Managing diabetes during pregnancy with insulin lispro:a safe alternative to human insulin. Endocr Pract 16:1020–1027 5. Lapolla A, Dalfra MG, Spezia R et al (2008) Outcome of pregnancy in type 1 diabetic patients treated with insulin lispro or regular insulin: an Italian experience. Acta Diabetol 45:61–66 6. Durnwald CP, Landon MB (2008) A comparison of lispro and regular insulin for the management of type 1 and type 2 diabetes in pregnancy. J Matern Fetal Neonatal Med 21:309–313 7. Banerjee M, Bhattacharya A, Hughes SM et al (2009) Efficacy of insulin lispro in pregnancies complicated with pregestational diabetes mellitus. Pract Diabetes Int 26:366–370 8. Aydin Y, Berker D, Direktor N et al (2008) Is insulin lispro safe in pregnant women: does it cause any adverse outcomes on infants or mothers? Diabetes Res Clin Pract 80(44):4–448 9. Garcı´a-Domı´nguez M, Herranz L, Hillman N et al (2011) Use of insulin lispro during pregnancy in women with pregestational diabetes mellitus. Med Clin 137:581–586 10. Cypryk K, Sobczak M, Pertyn´ska-Marczewska M et al (2004) Pregnancy complications and perinatal outcome in diabetic women treated with Humalog (insulin lispro) or regular human insulin during pregnancy. Med Sci Monit 10:PI29–PI32 11. Evers IM, de Valk HW, Mol BW et al (2002) Macrosomia despite good glycaemic control in type1 diabetic pregnancy results of a nationalwide study in the Netherlands. Diabetologia 45:1484–1489 12. Colatrella A, Visalli N, Abbruzzese S et al (2013) Comparison of insulin lispro protamine suspension with NPH insulin in pregnant women with type 2 and gestational diabetes mellitus: maternal and perinatal outcomes. Int J Endocrinol 2013:151975 13. Balaji V, Balaji MS, Alexander C et al (2012) Premixed insulin aspart 30 (BIAsp 30) versus premixed human insulin 30 (BHI 30) in gestational diabetes mellitus: a randomized open-label controlled study. Gynecol Endocrinol 28:529–532 14. Balaji V, Balaji MS, Alexander C et al (2010) Premixed insulin aspart 30 (Bia sp 30) vs premixed human insulin 30 (Bhi 30) in gestational diabetes mellitus—a pilot study. J Assoc Physician India 58:99–101 15. Hod M, Damm P, Kaaja R et al (2008) Fetal and perinatal outcomes in type 1 diabetes pregnancy: a randomized study
123
Arch Gynecol Obstet
16.
17.
18.
19.
20.
21. 22.
23.
24.
25.
26.
27.
comparing insulin aspart with human insulin in 322 subjects. Am J Obstet Gynecol 198:186.e1–186.e7 Poyhonen-Alho M, Ronnemaa T, Saltevo J et al (2007) Use of insulin glargine during pregnancy. Acta Obstetricia et Gynecologica 86:1171–1174 Fang YM, MacKeen D, Egan JF et al (2009) Insulin glargine compared with Neutral Protamine Hagedorn insulin in the treatment of pregnant diabetics. J Matern Fetal Neonatal Med 22:249–253 Egerman RS, Ramsey RD, Kao LW et al (2009) Perinatal outcomes in pregnancies managed with antenatal insulin glargine. Am J Perinatol 26:591–595 Imbergamo MP, Amato MC, Sciortino G et al (2008) use of glargine in pregnant women with type 1 diabetes mellitus: a casecontrol study. Clin Ther 30:1476–1484 Negrato CA, Rafacho A, Negrato G et al (2010) Glargine vs. NPH insulin therapy in pregnancies complicated by diabetes: an observational cohort study. Diabetes Res Clin Pract 89:46–51 Price N, Bartlett C, Gillmer M (2007) Use of insulin glargine during pregnancy: a case–control pilot study. BJOG 114:453–457 Smith JG, Manuck TA, White J et al (2009) Insulin glargine versus neutral protamine hagedorn insulin for treatment of diabetes in pregnancy. Am J Perinatol 26:57–62 Imbergamo MP, Amato MC, Amodeo R et al (2012) Use of basal insulin analog detemir in pregnant women with type1 diabetes:a case-control retrospective study. Giorn Ital Ostet Ginecol 34:491–498 Hod M, Mathiesen ER, Jovanovic L et al (2014) A randomized trial comparing perinatal outcomes using insulin detemir or neutral protamine Hagedorn in type 1 diabetes. J Matern Fetal Neonatal Med 27:7–13 Edson EJ, Bracco OL, Vambergue A et al (2010) Managing diabetes during pregnancy with insulin lispro :a safe alternative to human insulin. Endocr Pract 16:1020–1027 Blanco CG, Ballesteros AC, Saladich IG et al (2011) Glycemic control and pregnancy outcomes in women with type 1 Diabetes mellitus using lispro versus regular insulin: a systematic review and meta-analysis. Diabetes Technol Ther 13:907–911 Elhddad AS, Lashen H (2013) Fetal growthin relation tomaternal and fetal IGF-axes: a systematic review and meta-analysis. Acta Obstet Gynecol Scand 92:997–1006
123
28. Kurtzhals P, Scha¨ffer L, Sørensen A et al (2000) Correlations of receptor binding and metabolic and mitogenic potencies of insulin analogs designed for clinical use. Diabetes 49:999–1005 29. Wilde MI (1997) Mc Tavish D. Insulin lispro: a review of its pharmacological properties and therapeutic use in the management of diabetes mellitus. Drugs 54:597–614 30. Boskovic R, Knie B, Feig DS et al (2003) Transfer of insulin lispro across the human placenta. Diabetes Care 26:1390–1394 31. The Guideline Development Group (2008) Management of diabetes from preconception to the postnatal period: summary of NICE guidance. BMJ 336:714–717 32. Blumer I, Hadar E, Hadden DR et al (2013) Diabetes and Pregnancy: an endocrine society clinical practice guideline. J Clin Endocrinol Metab 98:4227–4249 33. Pollex E, Moretti ME, Koren G et al (2011) Safety of Insulin glargine use in pregnancy: a systematic review and meta-analysis. Ann Pharmacother 45:9–16 34. Sciacca L, Cassarino MF, Genua M et al (2010) Insulin analogues differently activate insulin receptor isoforms and post-receptor signalling. Diabetologia 53:1743–1753 35. Pollex EK, Feig DS, Lubetsky A et al (2010) Insulin glargine safety in pregnancy: a transplacental transfer study. Diabetes Care 33:29–33 36. Callesen NF, Damm J, Mathiesen JM et al (2013) Treatment with the long-acting insulin analogues detemir or glargine during pregnancy in women with type 1 diabetes:comparison of glycaemic control and pregnancy outcome. J Mater Fetal Neonatal Med 26:588–592 37. Shankardass K, O’campo P, Dodds L et al (2014) Magnitude of income- related disparities in adverse perinatal outcomes. BMC Pregnancy Childbith 14:96 38. Blumenshine P, Egerter S, Barclay CJ et al (2010) Socioeconomic disparities in adverse birth outcomes: a systematic review. Am J Prev Med 39:263–272 39. Mortense LH (2013) Socioeconomic inequality in birth weight and gestational age in Denmark 1996-2007: using a family-based approach to explore alternative explanations. Soc Sci Med 76:1–7 40. Scifres CM, Feqhali MN, Althouse AD et al (2014) Effect of excess gestational weight gain on pregnancy outcomes in women with type1 diabetes. Obstet Gynecol 123:1295–1302
