Post-market Review Products Used in the Management of Diabetes Report to Government Stage 2: Insulin Pumps February 2015

4.5 Third literature review: observational studies in pregnant women

4.5.1 Background

Reductions in the rate of adverse outcomes can be achieved by optimal care and tight glycaemic control before conception and throughout pregnancy in women with diabetes (Pearson 2007; Bismuth 2012). A Cochrane systematic review and meta-analysis, which included published literature to July 2011, identified five RCTs involving 153 women and 154 pregnancies comparing insulin pumps with multiple insulin injections (Farrar 2007). A separate meta-analysis that included literature up to April 2006 included six trials involving 213 women (Mukhopadhyay 2007). Collectively these reviews included eight trials (one trial included women with type 1 or type 2 diabetes). No trials have been published since the Cochrane review. With regards to differences in HbA_1c, hypoglycaemia, hyperglycaemia, macrosomia (birth weight greater than 4 kg), preterm delivery, caesarean birth, perinatal mortality, fetal anomaly and worsening of diabetes complications, neither Farrar (2007) or Mukhopadhyay (2007) reported significant differences between women using insulin pump therapy and women using multiple insulin injections.

Observational evidence published up to 2007 supports the results of the RCTs. A review of observational studies published between 2002 and 2007 identified six studies which compared insulin pump therapy and multiple insulin injections (two published in full and four in abstract format) (Cummins 2010). Most studies found similar glycaemic control, maternal and fetal outcomes between the groups. There were small numbers of women on insulin pump therapy in all studies limiting the statistical power to detect the endpoints under study.

4.5.2 Methodology and limitations

This literature review assessed the evidence for the effects of insulin pump therapy versus multiple insulin injections on diabetes management, maternal pregnancy outcomes and newborn outcomes in women with type 1 diabetes who are pregnant or planning to conceive. It included all English language, observational studies published between January 2008 and October 2014.

These observational studies had a number of limitations. They generally included a low number of women and may lack statistical power to detect significant differences. Most studies did not adjust for the differences observed between the study groups at the start of the study or other known risk factors that may impact on obstetric and perinatal outcomes. A failure to achieve therapeutic goals with multiple daily injections was an indication to switch to insulin pump therapy in some studies and could bias the results towards the null effect.

4.5.3 Results

The majority of the ten retrospective cohort studies identified were small with the number of women included ranging between 14 and 113 in the insulin pump group and between 20 and 424 in the multiple insulin injection group.

The majority of studies did not show a significant difference between groups either on diabetes management, maternal pregnancy outcomes or newborn outcomes.

4.5.3.1 Diabetes management

With regards to HbA_1c control, results were variable. Six studies did not show a difference between the insulin pump and injection groups for HbA_1c at any time during pregnancy. Four studies showed a lower level of HbA_1c with insulin pump use at the start of the study (Gonzalez-Romero 2010; Bruttomesso 2011; Kallas-Koeman 2014; Neff 2014), which was sustained throughout pregnancy in three studies. The largest study which examined 387 pregnancies (Kallas-Koeman 2014), found a significant difference in HbA_1c in favour of insulin pump therapy in all three trimesters without increased risk of severe hypoglycaemia.

Most studies did not find significant differences between insulin pump and multiple insulin injection therapy for other outcomes (ketoacidosis, non-severe hypoglycaemia, severe hypoglycaemia).

4.5.3.2 Maternal pregnancy outcomes

Most studies did not find significant differences between insulin pump and multiple insulin injection therapy for the outcomes of hypertension or preeclampsia, duration of pregnancy, weight gain during pregnancy, caesarean section and preterm labour.

4.5.3.3 Newborn outcomes

Most studies did not find significant differences between insulin pump and multiple insulin injection therapy for the outcomes of hypoglycaemia in newborns, newborn weight, large-for-gestational-age infants, 5-min AGPAR score, stillbirth and congenital abnormalities.

4.6 AIHW Insulin Pump User Survey

The AIHW’s Report on Insulin Pump Use in Australia (2012) details the experiences of 5,680 insulin pump users in Australia. The survey indicates strong device satisfaction among insulin pump users with the popular benefits being good integration with the user’s lifestyle (86%), better diabetes control (83%) and convenience (71%).

The survey results also highlighted the common problems experienced by insulin pump users as well as the reasons why some users decided to discontinue using the device. The biggest issue among the insulin pump users surveyed was the cost of insulin pump consumables, with 32% of respondents indicating that they were too expensive. Other common problems included issues with relocating the cannula/tubing of the pump (16%) and others disliking wearing a pump (15%).

The most common reasons for survey respondents to have discontinued insulin pump therapy were that they did not like wearing the pump (50%), they experienced little or no improvement in diabetes control (30%), problems relocating the cannula/tubing (29%), and wanting a break (29%).

The survey found that people with type 1 diabetes living in remote and very remote areas were less likely to use an insulin pump (around 8%) than those living in major cities, inner regional and outer regional areas, which all had similar proportions of use (around 10%). This result was based on small numbers of users from remote and very remote areas (70 users or 1% of survey respondents), so it is difficult to draw firm conclusions regarding reduced access to insulin pumps for people living in these areas. The survey also indicated that insulin pump users living in remote and very remote areas had less contact with specialist diabetes doctors or diabetes educators. Around 15% of people living in rural and remote areas had last seen a diabetes specialist or educator more than six months ago, compared to 9% for those living in major cities.