Practical Guidance on the Use of Premix Insulin Analogs in Initiating, Intensifying, or Switching Insulin Regimens in Type 2 Diabetes

An extensive clinical dataset, based on numerous randomized clinical trials and real-life observational studies, supports the efficacy and good safety profiles of BIAsp 30 and lispro mix 25 in the initiation and intensification of insulin therapy; these data have been reviewed [10, 11].

A small number of trials compared the use of either premix insulin analogs or basal insulin analogs for insulin initiation. Systematic reviews of the available evidence [12‐14] suggest that treatment with premix insulin analogs as first-line insulin therapy results in significantly better overall glycemic control, but slightly greater risk of hypoglycemia and weight gain, compared with basal insulin. A more recent review of eight trials comparing insulin initiation with either premix analogs or basal analogs showed greater glycated hemoglobin (HbA1c) reductions, weight gain, and number of hypoglycemic episodes with the premix analogs (significance of differences not reported) [15]. The authors suggested that factors that are not addressed in clinical trials, such as complexity of regimens and the need for titration, may influence outcomes.

The most recently published trial comparing BIAsp 30 once daily (OD) with a basal analog (insulin glargine OD) for insulin initiation in T2D was OnceMix (ClinicalTrials.gov number, NCT00469092) [16]. The estimated mean reduction in HbA1c was −1.41% (15 mmol/mol) with BIAsp 30 and −1.25% (14 mmol/mol) with insulin glargine (difference [95% confidence interval]: −0.16% [−0.30; −0.02] or −2 mmol/mol [−3; −0.2]; P = 0.029). There was a significant improvement in PPG with BIAsp 30 after dinner and before bed when compared with insulin glargine, with differences of −0.52 mmol/L (P = 0.04) and −0.78 mmol/L (P < 0.01), respectively. The risk of hypoglycemia, while low in both groups, was slightly higher with BIAsp 30 for overall hypoglycemia (6.5 vs. 4.8 episodes/patient-year; P = 0.034) and nocturnal hypoglycemia (1.1 vs. 0.5 episodes/patient-year; P = 0.003). However, the proportion of patients who achieved an HbA1c level <7% with no hypoglycemia was the same in the two groups (20.0% with BIAsp 30 and 19.4% with insulin glargine). Weight change did not differ between the groups.

DURABLE (ClinicalTrials.gov number, NCT00279201) compared lispro mix 25 twice daily (BID) with insulin glargine OD in previously insulin-naïve patients [17]. Treatment with lispro mix 25 resulted in slightly greater reductions in HbA1c at 24 weeks: −1.8 ± 1.3% (−20 ± 14 mmol/mol) vs. −1.7 ± 1.3% (−19 ± 14 mmol/mol) (P = 0.005). The lispro mix group also had more weight gain: 3.6 ± 4.0 vs. 2.5 ± 4.0 kg (P < 0.0001) and higher rates of overall hypoglycemia (28.0 ± 41.6 vs. 23.1 ± 40.7 episodes/patient-year; P = 0.007), but lower rates of nocturnal hypoglycemia (8.9 ± 19.3 vs. 11.4 ± 25.3 episodes/patient-year; P = 0.009). A follow-up study over 24 months showed durability of glycemic control was longer with lispro mix 25 [18].

A meta-analysis published in 2011 [19] compared premix insulin analogs with basal–bolus therapy based on three trials: PREFER (ClinicalTrials.gov number, NCT00605020), which compared BIAsp 30 BID with insulin detemir plus insulin aspart at mealtimes [20]; a study comparing lispro mix 50 (containing 50% soluble insulin lispro and 50% protamine-crystallized insulin lispro) three times daily (TID) with insulin glargine plus insulin lispro TID [21]; and the 3-year follow-up of the 4T study, comparing BIAsp 30 BID, insulin aspart TID, or insulin detemir OD or BID as initial regimens [22]. In 4T, if glycemic control was inadequate, lunchtime insulin aspart was added to BIAsp 30 BID, bedtime basal insulin was added to insulin aspart TID, and insulin aspart TID was added to basal insulin. Most of the patients in the prandial and basal groups switched to basal–bolus therapy [23]. Based on these three trials, the authors concluded that patients treated with a basal–bolus regimen had a higher chance of reaching their HbA1c goal (odds ratio [95% confidence interval]: 1.75 [1.11; 2.77]), with no difference in incidence of hypoglycemia or weight gain between the two regimens [19]. The results have to be interpreted cautiously, as the trial populations included insulin-naïve patients as well as patients already receiving basal insulin at the start of the treatment periods.

Further studies have been published since 2011, some only as abstracts at the time of the panel meeting. In PARADIGM (ClinicalTrials.gov number, NCT00548808), lispro mix 25 OD, BID, or TID, as needed, was compared with insulin glargine plus one, two, or three injections of insulin lispro, as needed. Glycemic control, weight change, hypoglycemia, the number of injections, and the increase in insulin dose were all similar between groups [24]. A phase IV study, GALAPAGOS (ClinicalTrials.gov number, NCT01121835), compared BIAsp 30 OD or BID with insulin glargine OD plus insulin glulisine OD if needed (‘basal-plus’). With BIAsp 30, 52.6% of patients achieved HbA1c <7% compared with 43.2% of those receiving basal-plus (P = 0.005). However, rates of hypoglycemia were higher with BIAsp 30: 2.9 vs. 1.2 episodes/patient-year for overall hypoglycemia, and 1.0 vs. 0.4 episodes/patient-year for nocturnal episodes (both P < 0.01) [25]. Finally, the LanScape study (ClinicalTrials.gov number, NCT00965549) compared BIAsp 30 OD with insulin glargine OD plus insulin glulisine OD at the main meal [26]. Reduction in HbA1c did not differ between the two regimens. There was no difference between overall hypoglycemia rates, but there were fewer nocturnal events with BIAsp 30 OD (3.6 vs. 5.7 events/patient-year; P = 0.02).

Factors influencing the choice of either premix insulin analogs or basal–bolus regimens for intensification in a primary care setting have recently been reviewed [27]. The authors of this study found inconclusive evidence and a lack of direct comparisons, and pointed out that clinical trials do not necessarily reflect real-world patients. In their view, GPs know their patients well and are in a good position to select the appropriate regimen for their patients (e.g., premix or basal-plus/basal–bolus therapy). However, GPs need sufficient time and support to accomplish this task.

Switching from a basal–bolus insulin regimen to premix insulin analogs is a relatively uncommon scenario. While basal–bolus therapy is considered the ‘gold standard’ for patients with advanced T2D, two groups of patients may need to switch to premix insulin analogs either BID or TID. These are patients who are unable or unwilling to cope with the complexity of a basal–bolus regimen, and patients who commence treatment with basal–bolus therapy in hospital (as occurs routinely in Australia, in accordance with guidelines from the Australian Diabetes Society [28]) and no longer require such an intensive regimen following discharge.

There is minimal published evidence on how to make this switch. Some evidence is available from a subgroup of the observational A₁chieve study (ClinicalTrials.gov number, NCT00869908), in which patients who were inadequately controlled on antidiabetic medication started or were switched to either BIAsp 30, insulin aspart, or insulin detemir [29]. The subgroup consisted of patients who were inadequately controlled on basal–bolus therapy using insulin glargine (n = 240) or neutral protamine Hagedorn insulin (n = 784) and who switched to BIAsp 30, mostly BID [30]. At 24 weeks, mean HbA1c decreased by approximately 2.5% (27 mmol/mol) and 1.9% (21 mmol/mol), respectively (P < 0.001 in both groups). The proportions reporting overall, major, or nocturnal hypoglycemia were significantly reduced (P < 0.05 in all cohorts) after switching to BIAsp 30. While this was not a randomized trial, the results do suggest that selected patients inadequately controlled on a basal–bolus regimen can benefit by switching to a premix insulin analog. The authors speculate that the improved results may have arisen from better therapy adherence due to the simpler regimen with BIAsp 30.

To our knowledge, the only previously published recommendations for switching from basal–bolus therapy to premix insulin analogs were included in Turkish guidelines, not available in English, on the use of BIAsp 30 in T2D [31]. These suggested that patients should be switched to BIAsp 30 BID or TID in preference to OD.

Patients may fear initiating insulin for many reasons: fear of hypoglycemia or weight gain; concern that insulin therapy indicates that they are heading towards severe complications such as blindness, limb amputation, or kidney failure; resistance to the need for monitoring blood glucose; or a belief that starting insulin indicates a failure on their part. It is important to allay these fears [38, 39]. It is also important to select the right HbA1c target for the individual patients. If the target is set too low, patients may omit insulin to avoid hypoglycemia, at the expense of their glycemic control.

When choosing an insulin regimen for initiation, it is imperative to bear in mind the long-term progressive nature of T2D and the likely need for intensification. The patient’s ability to cope with intensification should influence the choice of initiation regimen.

Figure 1 summarizes patient characteristics that may help determine a preference for either basal insulin or premix insulin analogs. Choose the regimen that provides the best match overall to the characteristics described. Age is not shown in Fig. 1, as function and degree of frailty are more important than chronological age. If life expectancy is short, the probability of future insulin intensification is less important.

See “Box 1” for guidance on dose and titration when initiating insulin with premix insulin analogs.

See “Box 2” for guidance on the use of other glucose-lowering drugs when initiating insulin with premix insulin analogs.

Intensification of insulin therapy is as important as initiation. Regular review with appropriate dose adjustment is critical, to ensure that the patient does not continue on their initiation regimen if glycemic control is suboptimal. Intensification is required if the individual’s HbA1c level remains above the individualized target for 3–6 months without any obvious reversible reason such as a steroid course or dietary non-compliance. Insulin should also be intensified if 2-h postprandial blood glucose values are above 10 mmol/L (180 mg/dL) and there is a difference (postmeal minus premeal) of ≥3 mmol/L (≥54 mg/dL), or when the maximum dose of 40–50 units is reached on premix insulin analog OD.

Practical guidance already exists for intensifying from basal insulin only or premix insulin analog OD to premix insulin analog BID [41]. “Box 3” and Figs. 2 and 3 have been adapted from this reference. For titration and monitoring when intensifying to premix insulin analog BID, refer to the guidance above on initiation (“Box 1”).

Use of premix insulin analogs TID may be needed owing to poor control, but this regimen is used much less often than premix insulin analogs OD or BID. In the panel members’ experience, the move from two to three injections daily can represent a larger barrier to patients than the move from one to two injections; patients must also be willing to undertake sufficient self-testing. If control is not satisfactory using a premix insulin analog BID, consider referring the patient to a specialist; otherwise, refer to the existing practical guidance for intensification [41].

The ratio of short- to intermediate-acting insulin in both BIAsp 30 and lispro mix 25 will be appropriate for most patients. Different ratio premixes, such as 50% soluble rapid-acting insulin/50% protaminated insulin, are also available and can be useful for patients with specific needs (e.g., patients with very high PPG values or problems with hypoglycemia). For information on intensifying to higher ratio premix insulin analogs, refer to the published guidance [43].

Not all patients achieve successful glycemic control with long-term basal–bolus therapy. The reasons for this are many and, with further education, the regimen may prove to be appropriate. For others, switching to a premix insulin analog may be the right decision. For example, some patients are unwilling or unable to deal with complexity, or they may have tried a basal–bolus regimen but their circumstances may have changed. Others may find carbohydrate counting, dose adjustment, or the required degree of monitoring too difficult, or be unable to handle two different insulin delivery devices.

To date, there is limited published evidence concerning this change, although all practicing endocrinologists will have experience in this situation. Rather than suggesting a specific HbA1c cutoff for switching regimens, we suggest a switch to premix insulin analogs in patients who are clearly unable or unwilling to use basal–bolus therapy, or whose HbA1c has consistently remained above target despite using a basal–bolus regimen while having access to proper training and adequate information.

The default switch is to premix insulin analog BID, not TID, and the titration guidance shown in “Box 4” is our suggested approach. If the patient is transferring from an analog basal–bolus regimen, education concerning the need to resuspend the protaminated insulin is important.

Some hospitalized patients may have been placed on a basal–bolus regimen to provide a flexibility of management that is not required following discharge. Under these circumstances, switching to an alternate premix insulin regimen would be appropriate.

In such patients, it is important to be conservative, and insulin requirements may change dramatically from those required in hospital. Discharge planning is vital and expert input is desirable. For titration guidelines, see “Box 4”.