Home > Clinical Guides > Type 2 Diabetes: National Clinical... > Glucose control: insulin therapy

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Chronic Conditions (UK). Type 2 Diabetes: National Clinical Guideline for Management in Primary and Secondary Care (Update). London: Royal College of Physicians (UK); 2008. (NICE Clinical Guidelines, No. 66.)

11Glucose control: insulin therapy

The recommendations in this guideline on thiazolidinediones (R40 to R43, chapter 10), GLP-1 mimetic (exenatide) (R44 to R46, chapter 10) and insulin therapy (R49 to R55), chapter 11) have been updated and replaced by NICE short clinical guideline 87 ‘Type 2 diabetes: newer agents for blood glucose control in type 2 diabetes’ (available at www.nice.org.uk/CG87shortguideline). This short guideline contains details of the methods and evidence used to develop the updated recommendations. Chapter 10 and 11 should be read in conjunction with the short clinical guideline 87.

11.1. Oral agent combination therapy with insulin

11.1.1. Clinical introduction

People with Type 2 diabetes with inadequate blood glucose control on oral agents have the pathogenetic problems which caused their diabetes, and still have significantly preserved islet B-cell function. There remains the possibility that medication designed to enhance insulin secretion, reduce insulin insensitivity, or otherwise improve blood glucose control might be useful in combination with insulin therapy, in improving blood glucose control, reducing insulin dose requirement, or mitigating side effects of insulin therapy.

The clinical question is which oral agents, singly or in combination, should be continued when starting insulin therapy.

11.1.2. Methodological introduction

Studies were identified which compared insulin in combination with oral hypoglycaemic agents (OHAs) with insulin monotherapy in insulin naive Type 2 diabetic patients. A Cochrane review170 was identified which included 20 RCTs in a search performed in March 2004. Ten additional RCTs were identified, five of which were excluded due to methodological limitations.171–175

Of the remaining five RCTs the treatment comparisons were:

It should be noted that the number of different drug combinations and comparisons, dosing and titration regimens limit direct comparison between the studies. Furthermore, all of the studies with the exception of one176 were open-label.

Of the five trials presented above, it can be noted that only two included a biphasic insulin arm with metformin or a sulfonylurea.64,176 Further details of the five trials in the Cochrane review, which included biphasic insulin regimens in combination with OHAs (all published between 1987 and 1998, prior to this update), are given where this data was available in the Cochrane review at the request of the GDG. These trials compared:

  • mixed insulin (25% regular, 75% protamine insulin) plus glibenclamide vs mixed insulin (25% regular, 75% protamine insulin) and placebo (N=140, Cochrane methodological quality score 2/7) (Bachman 1988)
  • mixed insulin (intermediate acting NPH plus regular insulin) twice daily and glibenclamide vs mixed insulin (intermediate acting NPH plus regular insulin) twice daily and placebo (N=20, Cochrane methodological quality score 2/7) (Gutniak 1987)
  • insulin (combination of short and intermediate acting insulin) once or twice daily plus glibenclamide vs insulin alone (combination of short and intermediate acting insulin) once or twice daily (N=27, Cochrane methodological quality score 2/7) (Ravnik-Oblak 1995)
  • mixed insulin (70% NPH, 30% soluble) at suppertime plus glibenclamide vs mixed insulin (70% NPH, 30% soluble) and placebo (N=21, Cochrane methodology score 7/7) (Riddle 1992)
  • mixed insulin (70% NPH, 30% regular human insulin) at suppertime plus glimepiride vs mixed insulin (70% NPH, 30% regular human insulin) and placebo (N=145, Cochrane methodology score 6/7) (Riddle 1998).

It is notable that some of these studies had small sample sizes and/or low methodological quality scores.

11.1.3. Health economic methodological introduction

Only one economic evaluation was identified.180 The analysis was conducted over a short time period (4 months) and intermediate outcomes were reported. For economic analysis to inform resource allocation it is important to consider the impact on final health outcomes such as mortality and morbidity.181 The incremental costs and benefits of using insulin glargine compared to conventional insulin treatment were not reported.

An economic model was constructed based upon the UKPDS outcomes model to inform the GDG with regard to choice of glitazones or exenatide as third-line therapy in comparison to other third-line options. This is presented in appendix C available at www.rcplondon.ac.uk/pubs/brochure.aspx?e=247

11.1.4. Evidence statements

Glycemic control

Overall the data seems to suggest that patients receiving a combination treatment with insulin NPH or pre-mixes) and metformin or a sulfonylurea showed significantly lower HbA1c levels when compared to those treated with insulin monotherapy. FPG values were not consistently assessed by most of the studies.

Table 11.1. HbA1c

Insulin dose

A Cochrane review170 reported that insulin–OHA combination therapy was associated with a significantly lower insulin dose compared to insulin monotherapy. An RCT176 reported the same trend for the combination of insulin and metformin.

Well-being and quality of life

The few studies that objectively assessed well-being, quality of life or treatment satisfaction did not report significant differences between insulin–OHA combination and insulin monotherapy. However, there was a trend towards higher levels of satisfaction for patients in the combination group (especially those receiving metformin).

Hypoglycaemia

Non-significant differences in the incidence of hypoglycaemic events between insulin–OHA and insulin monotherapy were reported across most of the studies identified. However, a higher number of hypoglycaemic events were observed in patients receiving monotherapy with biphasic insulin regimens (e.g. NPH 30/70).

Table 11.2. Hypoglycaemic events

Weight gain

It was observed across most of the studies that treatment with insulin and other OHA (especially metformin) was associated with significantly less weight gain when compared with insulin monotherapy.

Only one study147 comparing the combination of BIAsp 30 plus pioglitazone with BIAsp monotherapy showed a greater weight gain in patients treated with the combination therapy.

Other adverse events

Overall, no significant differences in frequency or severity of AEs were found for patients receiving insulin alone or combination therapy regimens. However, one study147 found that more patients experienced product-related AEs in the biphasic aspart 30/70 plus pioglitazone group (28%) compared with patients receiving biphasic insulin aspart 30/70 monotherapy (20%). The combination group was also associated with a higher proportion of patients experiencing peripheral edema (6%) compared with aspart monotherapy (0%).

11.1.5. From evidence to recommendation

The new evidence continued to support the view that metformin should be continued when starting insulin therapy. The evidence was stronger than previously for sulfonylureas, for acarbose if used, and also for the thiazolidinediones. For sulfonylureas the situation was further complicated by much of the newer data coming from use with basal insulin regimens, while there was more uncertainty and concern over use with biphasic insulin (pre-mix) regimens due to risks of hypoglycaemia and the risk this might worsen achieved blood glucose control. Positive advice was tempered by concerns that the combination might cause excessive weight gain, and it was not possible to conclude whether this was clinically significant or otherwise a concern to the individual with Type 2 diabetes.

The cost and cost-effectiveness issues of continuing thiazolidinediones were considered at the time of review of the health economic modelling, although this issue was not specifically addressed by the modelling. Being high cost, it was unclear that the thiazolidinediones could give cost-effective health gains when continued at the time of starting insulin. However, it was noted that some people (often markedly obese) get a combination of reductions of insulin doses from high levels together with markedly improved blood glucose control when thiazolidinediones were added to insulin therapy.

RECOMMENDATIONS

R47.

When starting basal insulin therapy:

R48.

When starting pre-mixed insulin therapy (or mealtime plus basal insulin regimens):

  • continue with metformin
  • continue the sulfonylurea initially, but review and discontinue if hypoglycaemia occurs.
R49.

Consider combining pioglitazone with insulin therapy for:

Warn the person to discontinue pioglitazone if clinically significant fluid retention develops.

11.2. Insulin therapy

11.2.1. Clinical introduction

Blood glucose control deteriorates inexorably in most people with Type 2 diabetes over a period of years, due to a waning of insulin production.55 In these circumstances oral glucose-lowering therapies can no longer maintain blood glucose control to targets and insulin replacement therapy becomes inevitable. Insulin deficiency is however only relative, not absolute, as there is still considerable endogenous insulin secretion occurring in response to the insulin insensitivity that is also usual in people with Type 2 diabetes. This means that the insulin regimens used in Type 1 diabetes (a condition of absolute insulin deficiency) may not be those needed in people with Type 2 diabetes.

The clinical question is which of the various pharmaceutical types of insulin, and in what combinations, are optimal for the management of Type 2 diabetes, both when initiating insulin and as insulin deficiency further progresses over the years.

11.2.2. Methodological introduction

Biphasic insulin preparations vs NPH

A limited number of clinical studies were identified which compare pre-mixes with NPH insulin.

There were three relevant RCTs. One study182 compared biphasic insulin aspart 30/70 and NPH insulin in a population of 403 patients with a follow-up of 16 weeks. The other study183 compared the combination of insulin aspart 30/70 and metformin with the combination of NPH insulin and metformin in a population of 140 patients with a follow-up of 12 weeks. The third study, a cross-over trial, compared a preprandial and basal regimen with insulin lispro and NPH, with a basal only regimen with twice daily NPH in 30 patients spending 12 weeks in each arm before cross-over.184

Differing populations, dosing and titration regimens may limit direct comparison between studies.

Biphasic human insulin preparations vs biphasic analogue preparations

A limited number of clinical studies were identified which compare biphasic analogue preparations with biphasic human insulin preparations.

One Cochrane review and meta-analysis was identified on this question.185 This review was excluded as 88% of the included studies were judged to be of limited methodological quality. Eight studies in Type 2 diabetics had been identified and six studies in Type 1 and Type 2 diabetics. Of the studies included in the meta-analyses on HbA1c and hypoglycaemic episodes outcomes, only one study published post-2001 was included in each analysis.

Two RCTs were identified comparing once daily biphasic insulin analog formulation (insulin aspart containing 30% soluble insulin aspart and 70% insulin aspart crystallised with protamine) with human pre-mixed insulin (30% regular, 70% NPH insulin).186,187

The study by Boehm187 was an extension RCT of Boehm186 comparing the long-term efficacy of these two formulations. An additional RCT compared three times daily biphasic insulin analog formulation (insulin aspart containing 30% soluble insulin aspart and 70% insulin aspart crystallised with protamine) with once daily human pre-mixed insulin (30% regular, 70% NPH insulin).188 One RCT compared a three times daily biphasic insulin analog formulation (50% insulin lispro and 50% neutral protamine lispro suspension) with once daily human pre-mixed insulin (30% regular insulin and 70% NPH).189

One RCT compared patients on metformin plus either once daily biphasic insulin analog formulation (insulin aspart containing 30% soluble insulin aspart and 70% insulin aspart crystallised with protamine), NPH insulin or human pre-mixed insulin (30% regular, 70% NPH insulin).183 Another RCT compared a biphasic insulin analogue (insulin aspart containing 30% soluble insulin aspart and 70% insulin aspart crystallised with protamine) with a daily basal-bolus regimen with insulin aspart before meals and evening human isophane insulin NPH).190 All studies were on patients with Type 2 diabetes except for one that included patients with Type 1 and Type 2 diabetes.186

Three open-label, single dose RCTs with methodological limitations were not considered further.

Differing populations, dosing and titration regimens may limit direct comparison between studies.

Multiple analogue insulin injection regimens compared to basal insulin or biphasic insulin regimens

A limited number of clinical studies were identified in this specific area.

A cohort study relevant to the question191 conducted in India compared a multiple analogue insulin regimen with a pre-mix regimen in a cohort of 145 participants with a follow-up of 12 weeks.

The cohort study had the following limitations.

  • Although described as a prospective study, it seems to be a retrospective collection of patients’ data.
  • It did not have a placebo-controlled arm.

Only one RCT was found that partially addressed the question.192 This RCT did not directly compare multiple analogue insulin injection regimens with basal insulin or biphasic insulin regimens. The study was primarily designed to compare two different initiation treatment algorithms with insulin glargine (physician visit-base titration vs patient self-titration) in people with Type 2 diabetes suboptimally controlled on their previous antidiabetic treatment. A separate abstract reported the results for a subgroup of study participants who changed from once daily pre-mix insulin to once daily insulin glargine alone or with prandial insulin and/or oral antidiabetics (OADs). This reported baseline and endpoints values for HbA1c along with incidence of hypoglycaemia among seven groups of patients receiving different basal-bolus regimes with or without OADs.

This subgroup analysis should be interpreted with caution because:

  • there was no subgroup treatment protocol to ensure consistent management
  • there was only a historical control arm to demonstrate greater clinical efficacy of a multiple insulin regimen over a biphasic insulin regimen.

Long-acting insulin analogues (insulin glargine compared to NPH insulin, biphasic insulins or multiple daily injections)

A NICE technology appraisal (TA)193 previously reviewed the evidence available until the end of 2001 and made recommendations on the use of insulin glargine in Type 2 diabetes. This guideline updates this appraisal and the GDG considered whether the appraisal recommendations should change in the light of new evidence.

Two meta-analyses194,195 and 14 further RCTs178,196–208 were identified which compared a regimen containing insulin glargine with another insulin containing regimen in those with Type 2 diabetes. One RCT compared morning and evening administration of insulin glargine.209 One RCT compared insulin glargine with an optimised oral diabetic agent treatment arm.210

A recent meta-analysis by Horvath195 compared the long-acting insulin analogues (insulin glargine and insulin determir) with NPH insulin. Only the results of the insulin glargine and NPH comparison are considered here. In this meta-analysis six RCTs were included in the glargine and NPH comparison.196,199,211–214 A further RCT by Yokohama was mentioned in the study but not included in the meta-analysis.208

An older meta-analysis by Rosenstock194 which contained some of the same studies as the Horvath analysis combined four RCTs211–214 which compared insulin glargine once daily with NPH insulin once or twice daily (in three studies NPH insulin was administered once daily,211–213 and in the other study it was administered once or twice daily).214 Four further RCTs compared once daily insulin glargine with once daily NPH insulin.196,199,200,206 One RCT was excluded for methodological reasons.208

Eight RCTs compared insulin glargine with biphasic insulins.178,198,201–205,207 In two studies201,202 an insulin lispro mix 75/25 (75% insulin lispro protamine suspension and 25% insulin lispro) administered twice daily was compared with bedtime insulin glargine. Two further studies compared intensive mixed preprandial regimens with insulin lispro before each meal compared to once daily insulin glargine.203,205 Another study178 compared insulin glargine once daily with human pre-mixed insulin (30% regular, 70% NPH insulin) twice daily, however these groups were not directly comparable as metformin and glimepiride were given with the insulin glargine and not with the pre-mixed insulin. Three studies198,204,207 compared a once daily biphasic insulin analog formulation (insulin aspart containing 30% soluble insulin aspart and 70% insulin aspart crystallised with protamine) with once daily insulin glargine, although in one of these studies204 glimepiride was added to the glargine arm and metformin to the biphasic arm.

The study that compared morning and evening administration of insulin glargine included glimepiride in both arms.209

The review commissioned by NICE,197,215 on which previous appraisal recommendations were based, noted that in studies where insulin glargine is demonstrated to be superior in controlling nocturnal hypoglycaemia, this may only be apparent when compared with once daily NPH and not twice daily NPH. It is thus notable that no new studies were identified which compared insulin glargine with NPH insulin administered twice daily.

The range of definitions of hypoglycaemia used and differing populations may limit direct comparison between studies.

11.2.3. Meta-analysis

Meta-analyses were conducted (using the Cochrane Collaboration’s RevMan software) to investigate the choice of third-line therapies where more than one study was available for a comparison. Interventions considered were:

Because of the high acquisition costs of these third-line therapies, the pooled point estimates and CI of efficacy were used in a health economic model comparing these treatment options (see below. Full results are shown in appendix C available at www.rcplondon.ac.uk/pubs/brochure.aspx?e=247). The economic model was an adaptation of the UKPDS risk calculations, and in order to supply the risk factors in UKPDS, the following outcomes were sought:

Of these, the only outcome where more than one study could be pooled was HbA1c. Change in weight or BMI was not one of the risk factors in UKPDS, and so was addressed in the economic model by sensitivity analyses (see appendix C for more detail available at www.rcplondon.ac.uk/pubs/brochure.aspx?e=247).

Hypoglycaemia was not an outcome variable which could be varied in the UKPDS-based analysis. Accordingly a sensitivity analysis was performed by improving quality of life in insulins in evidence with less hypoglycaemia (see appendix C for more detail available at www.rcplondon.ac.uk/pubs/brochure.aspx?e=247).

The following studies were pooled:

None of the comparisons had significant heterogeneity but the two studies comparing glargine to human insulin196,199 had notably different baseline demographics and so a random effects analysis was used in this instance.

The comparison of biphasic analogues with human insulin showed no significant difference.

The comparison of glargine with human insulin showed no significant difference.

The comparison of biphasic analogue with glargine had a pooled weighted mean difference of 0.43% HbA1c (95% CI 0.40 to 0.46) in favour of biphasic analogues. This analysis was dominated by one large trial198 but all three trials showed significant differences in the same direction of effect, which supports the validity of the pooled result.

11.2.4. Health economic methodological introduction

Two studies were found that compared the cost-effectiveness of glargine insulin with other forms of insulin.193,216 Both studies were based on meta-analysis and used the UKPDS outcomes model to predict events and costs. However, they did not take in to account the impact on quality of life of AEs such as weight gain and vomiting.

For this guideline, an economic model was constructed based upon the UKPDS outcomes model to inform the GDG with regard to the cost-effectiveness of various third-line therapy options. This is presented in appendix C available at www.rcplondon.ac.uk/pubs/brochure.aspx?e=247

11.2.5. Evidence statements

Insulin glargine was not included in the Type 2 diabetes guideline 2002 under review. However, it was the subject of a NICE TA at that time, and the current review is an update of that.

Biphasic insulin preparations vs NPH

HbA1c

The two studies182,183 found that HbA1c levels decreased linearly and statistically significantly in both treatment groups (biphasic insulin aspart 30/70 and NPH insulin) compared to baseline values. There was not a significant statistical difference between the two interventions. Level 1+

The third study found a significantly greater reduction in HbA1c in the lispro and NPH arm than in the twice daily NPH arm (p<0.01).184 Level 1+

Fasting blood glucose/fasting plasma glucose

In patients receiving either biphasic insulin aspart 30/70 or NPH insulin, studies182,183 showed similar reductions from baseline in FBG/FPG values. There was however no statistically significant difference between the two interventions. Level 1+

Postprandial blood/plasma glucose

One study182 reported that the mean prandial glucose increment over the three main meals was significantly lower in the aspart 30/70 group than in the NPH group, (0.69 mmol/l lower; p<0.0001, between groups.) Level 1+

The other study183 found no significant differences between the groups regarding the mean values for the 8-point self-monitoring of blood glucose (SMBG) profile at week twelve. The study reported that SMBG values for before breakfast and before lunch values tended to be lower for the NPH insulin group, while after dinner and 10 pm, values tended to be higher for the NPH insulin group as compared to the biphasic insulin aspart. Level 1+

In the insulin lispro vs NPH comparison, the postprandial glucose excursion was significantly lower in the lispro arm (p<0.001).184 Level 1+

Body weight

Two studies183,184 found non-significant differences in terms of body weight gain between the biphasic insulins and NPH. Level 1+

Adverse events

Both studies comparing insulin aspart with NPH182,183 concluded that the number and type of AEs were similar for each of the treatment groups with non-significant differences between them. Level 1+

One study182 found that in terms of incidence of hypoglycaemia, the RR was not statistically significantly different between treatments (RR=1.21 (95% CI 0.77 to 1.90), p=0.40). The other study reported that there was no significant difference between regimens for either overall or nocturnal hypoglycaemia.184 Level 1+

The other study183 found that nocturnal hypoglycaemia (midnight–6 am) was less frequently reported for patients receiving biphasic insulin aspart (seven patients) as compared to patients in the NPH insulin group (11 patients). No statistical analysis was reported. Level 1+

Lipid profile

One study184 reported changes in lipid measures between groups and found a significantly lower fasting low-density lipoprotein cholesterol (LDL-C) and LDL-C/HDL-C ratio in the biphasic insulin lispro) and NPH arm compared with twice daily NPH (p=0.035). After a standard meal both LDL-C (p=0.012) and HDL-C (p=0.004) were significantly higher in the biphasic insulin (lispro) and NPH arm compared with twice daily NPH arm. Level 1+

Table 11.3. Biphasic human insulin preparations vs biphasic analogue preparations

HbA1c

Overall, on endpoint means HbA1c levels biphasic analogue preparations were comparable to human pre-mixed insulin,183,186,187,188 as well as to a basal-bolus regimen of insulin aspart and NPH.190 Level 1+

One RCT found three times daily biphasic insulin lispro (50/50) gave a significantly greater reduction from baseline in mean HbA1c values compared with once daily pre-mixed human insulin 30/70.189 Level 1+

Fasting blood glucose

Two RCTs found no significant differences among the treatment groups on FBG.186,183 Level 1+

One RCT found that FBG was significantly increased in patients on three times daily biphasic analogue insulin compared with once daily human pre-mixed insulin.189 Level 1+

Postprandial glucose

In terms of PPG, three RCTs reported significant treatment differences in favour of biphasic insulin aspart.188,186,189 Level 1+

Bodyweight

No studies reported any significant differences between treatment groups.186,187,183,190 Level 1+

Adverse events

Studies reported similar AEs profiles for biphasic analogue insulin and biphasic human insulin.188,186,187,183,189,190 Level 1+

Hypoglycaemia

Overall, few major hypoglycaemic episodes were associated with either biphasic analogue or human insulin.188,186,183,189,190 Level 1+

A longer-term efficacy study found that during the second year of treatment significantly fewer patients in the once daily biphasic analogue insulin than the human pre-mixed insulin group experienced a major episode.187 Level 1++

No study reported any significant differences between treatments on minor or nocturnal hypoglycaemic episodes.188,186,183,190 Level 1+

Multiple analogue insulin injection regimens compared to basal insulin or biphasic insulin regimens

HbA1c

For HbA1c levels the cohort study reported that both multiple insulin regimen and pre-mix insulin regimen lowered HbA1c levels significantly compared to baseline values. Pre-mix insulin analogue fared better than the basal-bolus analogue therapy in lowering HbA1c (1.58% vs 1.16% respectively, p<0.05). Also 41% more patients in the pre-mix group could achieve target HbA1c of <7% at the end of 12 weeks (45.61% vs 32.26%). Level 2+

FPG/PPPG

Both regimes lowered FPG and postprandial plasma glucose (PPPG) levels significantly as compared to baseline. No statistical comparison was performed between groups. Level 2+

Body weight

The body weight did not change significantly in either group at the end of the study. Level 2+

Hypoglycaemia events

The percentage of patients experiencing minor hypoglycaemia was significantly lower in the pre-mix group than in the basal-bolus group at 12 weeks (16.7% vs 58.06%, p<0.05). Level 2+

Throughout the study period of 12 weeks, there were no major hypoglycaemic episodes reported in both the treatment groups. Level 2+

Subgroup analysis

The analysis of the sub-population previously receiving pre-mix insulin suggests that optimisation of basal insulin therapy with once daily insulin glargine is safe (according to the low incidence of severe hypoglycaemic events) and results in significant improvements in glycaemia control.

The same analysis indicates that once daily insulin glargine in combination with prandial therapies (prandial insulin and/or OADs) offers additional glycaemic benefits.

Long-acting insulin analogues (insulin glargine compared to NPH insulin, biphasic insulins or multiple daily injections)

NB. Glargine and its comparators are often used in these studies in combination with OAD medications. For simplicity, references to these drugs are not included in the evidence statements unless they differ between the two groups.

Table 11.4. Insulin glargine vs NPH insulin

Table 11.5. Insulin glargine vs biphasic insulins

None of the studies194–196,199,200,206 reported differences between the insulin glargine and NPH groups in terms of proportion of patients achieving target HbA1c, insulin dose, body weight, daytime hypoglycaemia or AEs. One study found a significantly greater reduction in the mean HbA1c at endpoint in the insulin glargine arm.206 Five studies194–196,199,206 found significant risk reductions in overall risk of hypoglycaemia with insulin glargine compared to NPH insulin (one only in the first 12 weeks)196 while the shorter study found no difference.200 Five studies194,195,199,200,206 reported significant risk reductions in terms of nocturnal hypoglycaemia with insulin glargine compared to NPH insulin. Additionally, FPG values were significantly lower at endpoint in the glargine groups in two studies196,214 but showed no significant difference in the shorter study.200 Level 1+

Seven studies198,201–205,207 reported better HbA1c outcomes with the insulin mixes compared to insulin glargine. The other study found significantly higher reductions in HbA1c with insulin glargine from baseline, however insulin glargine was combined with OAD drugs which were not received by the insulin mix group.178 With respect to decreases in FBG from baseline results, they were less consistent. Statistically significant decreases in FBG were reported in insulin glargine groups compared to the insulin mix groups in four studies,178,201,202,205 although three studies did not find a significant difference.203,204,207 Insulin doses were higher in the insulin mix groups in all studies.178,198,201–205,207 In five studies the insulin mix groups had significantly increased body weight from baseline compared with insulin glargine.198,201,202,205,207 Two studies found no significant difference in body weight change between the groups178,203 and the remaining study204 reported a greater weight increase in the insulin glargine and glimepiride group than in the biphasic insulin analogue and metformin group although they did not report if this was statistically significant. In terms of hypoglycaemia, one study found no significant difference202 in overall hypoglycaemia rates, while the remaining studies178,198,201,203–205,207 found overall hypoglycaemia rates were better with insulin glargine than insulin mixes. For nocturnal hypoglycaemia, two studies reported no significant difference between the groups,201,203 another found higher rates in the glargine group202 and two others found significantly reduced rates in that group compared to the insulin mix group.178,207 Only one study reported daytime hypoglycaemia rates and these were found to be significantly higher in the insulin mix group.202 No significant differences between the groups were reported in terms of AEs.178,198,201,202,204,205,207 Level 1+

Morning vs evening administration of insulin glargine

Standl et al.209 compared insulin glargine delivered at different times of the day to determine the impact on glycaemic control and rates of hypoglycaemia. It was found that morning and evening administration of glargine was equivalent with respect to the incidence of nocturnal hypoglycaemia. Similar improvements in HbA1c, FBG and the proportion of patients achieving an HbA1c of less than 7% was demonstrated in the two arms of the study, without any difference in the incidence of AEs. Level 1+

Insulin glargine vs oral therapy

Gerstein et al.210 compared the addition of insulin glargine to current treatment with the intensified oral glucose-lowering therapy. HbA1c outcomes were reported to be significantly better in the glargine group even after adjusting for baseline HbA1c and oral therapy. FPG was also significantly lower and lipid parameters were significantly improved in the glargine group. There was no significant difference in hypoglycaemia, and the glargine group had a significantly greater weight increase. Level 1+

11.2.6. Health economic evidence statements

In the long-acting insulin TA193 there was an estimated cost-effectiveness ratio of £33,000 compared to NPH insulin, using the price of a vial of glargine. Using cartridges or pens gave higher cost-effectiveness ratios, £41,000 and £43,000 respectively. The results were most sensitive to the assumption on utility gained from reducing fear of hypoglycaemia. If it was assumed that there was no utility gain from this then the cost-effectiveness ratio rose to approximately £10 million per QALY.

The second study216 found a cost-effectiveness ratio of £13,000 per QALY gained compared to NPH insulin. But it did not take into account the disutility associated with the side effects of insulin glargine and no comparison was made with other third-line therapies.

The base-case results of the analysis of third-line therapy conducted for this guideline (see appendix C available at www.rcplondon.ac.uk/pubs/brochure.aspx?e=247) found that human insulin was as effective but less expensive than biphasic insulin, and more effective and less expensive than insulin glargine.

11.2.7. From evidence to recommendations

Pre-mix insulin

There was limited evidence for comparisons of pre-mix insulin with NPH insulin in people with diabetes. Because of the use of unselected populations of people with Type 2 diabetes taking little account of factors such as degree of insulin deficiency, high or low mealtime insulin requirement, diurnal patterns of blood glucose control, and sensitivity to hypoglycaemia, the studies did not help inform clinical decision making. These insulins, compared to basal insulins, target postprandial blood glucose control. The issue of whether postprandial blood glucose control was of any specific importance, rather than being important because glucose levels are highest at that time, is not being addressed in this guideline. There was confidence that no health outcome studies on the issue had been published. The GDG felt that it was inappropriate to make strong recommendations promoting pre-mix insulin over NPH or the opposite, except to observe that as insulin deficiency progressed mealtime insulin therapy would be more likely to be indicated.

There was limited evidence on the comparisons between insulin analogue pre-mixes and human insulin pre-mixes. There was definite evidence statistically of some reduction in postprandial blood glucose control in the period after injection when using an analogue rather than human insulin, as was to be expected from other data with rapid-acting insulin analogues. Equally there was some data on the reduction of hypoglycaemia, consistent with other analogue data. These effects were clinically quite small and therefore of questionable cost-effectiveness, a view supported by the health economic modelling.

Unfortunately all comparative trials had been performed using different recommendations of timing of insulin injection before meals for human and analogue insulins (in line with licences). The advantage of injecting immediately before meals (usually twice a day) in daily life to people with diabetes was felt to be a significant quality of life issue justifying the use of the analogues. Studies asking whether human insulin pre-mixes could be given immediately before meals without deterioration of blood glucose control (hyperglycaemia early and hypoglycaemia late) compared to analogues had not been performed.

Basal insulins including long-acting insulin analogues

The previous guidance for use of insulin glargine endorsed its use in people with Type 2 diabetes where the injections were given by a carer, where hypoglycaemia was a problem when using NPH insulin, and where insulin administration would otherwise require twice daily insulin injections. The studies performed since were a useful contribution not only to the understanding of insulin glargine, but more so, to the optimal use of insulin in people with Type 2 diabetes, in particular for people starting insulin therapy.

Very little useful information was found to assist in advising on the optimal insulin regimen once progression of islet B-cell failure had progressed further, for example in people 3–5 years or more after starting insulin therapy. The observational study from India was open to bias in patient and provider selection, and the subgroup analysis from A Trial comparing Lantus® Algorithms to achieve Normal blood glucose Targets in patients with Uncontrolled blood Sugar (AT.LANTUS) was similarly open to bias and in small numbers of people. The preferred view was that as islet B-cell deficiency progressed people tended to a state of insulin deficiency closer to those with Type 1 diabetes, suggesting that prior NICE guidelines advice for that group of patients could be applied.

The strongest of the new evidence for insulin starters appeared to relate to comparisons with NPH insulin, and of these the data on comparison with once daily (bedtime) human NPH insulin was the most novel. It was noted that these treat-to-target studies have the problem, given their limited duration, of driving control in the compared groups towards the same levels, and indeed pre-breakfast glucose levels and HbA1c were similar for insulin glargine and NPH, at similar insulin doses. The differences in nocturnal hypoglycaemia were convincing, if small in absolute terms. Despite post hoc analyses of the relationship between HbA1c and nocturnal hypoglycaemia showing convincing advantage of insulin glargine over NPH insulin, it was impossible to determine what the balance of advantage between the two measures would be in real clinical practice, where differences in hypoglycaemia tend to drive differences in insulin dosage and thus overall blood glucose control (which would be to the advantage of the long-acting analogue).

Although not the subject themselves of a randomised comparison, the approaches used in the treat-to-target studies of active dose titration in the context of appropriate education, self-monitoring and support were an important means of obtaining optimal blood glucose control whatever insulin was employed.

An issue relates to the choice of insulin preparation for starting insulin in people with Type 2 diabetes. As noted above, and provided that insulin was started reasonably early in the disease process before HbA1c had deteriorated too far, there was little justification for the use of more intensive mealtime plus basal insulin regimens in this situation. The studies comparing insulin glargine with pre-mix insulin regimens gave mixed results, with improved HbA1c apparently resulting from an ability to titrate twice daily insulin dosage faster (in total) than once daily injections, but at a cost of increased hypoglycaemia and weight gain. These results and the absence of longer term data on performance of the two regimens, together with complexities such as the possibility of using three injections of pre-mix, or of adding mealtime insulin to basal glargine, meant that the GDG was unable to identify overall advantage to one approach or the other.

The previous NICE guidance in relation to a single daily injection of insulin glargine not having to be given at any precise time was noted to be useful for those whose injections are given by others.

The GDG found the health economic modelling problematic in the area of insulin therapy. Major problems seem to relate to the difficulties of including fear of hypoglycaemia and its effect on everyday lifestyle, restrictions on lifestyle with insulin injections, and the present day educational costs associated with intensive insulin dose adjustment to achieve good target control. While some attempts had been made to incorporate some of these in sensitivity analyses, it was not possible to be sure of their validity, though the face value results all suggested that human insulin regimens were the only cost-effective approach.

RECOMMENDATIONS

R50.

When other measures no longer achieve adequate blood glucose control to HbA1c <7.5% or other higher level agreed with the individual, discuss the benefits and risks of insulin therapy. Start insulin therapy if the person agrees.

R51.

When starting insulin therapy, use a structured programme employing active insulin dose titration that encompasses:

  • structured education
  • continuing telephone support
  • frequent self-monitoring
  • dose titration to target
  • dietary understanding
  • management of hypoglycaemia
  • management of acute changes in plasma glucose control
  • support from an appropriately trained and experienced healthcare professional.
R52.

Insulin therapy should be initiated from a choice of a number of insulin types and regimens.

  • Preferably begin with human NPH insulin, taken at bedtime or twice daily according to need.
  • Consider, as an alternative, using a long-acting insulin analogue (insulin glargine) for a person who falls into one of the following categories:

    those who require assistance from a carer or healthcare professional to administer their insulin injections

    those whose lifestyle is significantly restricted by recurrent symptomatic hypoglycaemic episodes

    those who would otherwise need twice daily basal insulin injections in combination with oral glucose-lowering medications.

  • Consider twice-daily biphasic human insulin (pre-mix) regimens in particular where HbA1c is elevated above 9.0 %. A once-daily regimen may be an option when initiating this therapy.
  • Consider pre-mixed preparations of insulin analogues rather than pre-mixed human insulin preparations when:

    immediate injection before a meal is preferred, or

    hypoglycaemia is a problem, or

    there are marked postprandial blood glucose excursions.

  • R53.

    Offer a trial of insulin glargine if a person who has started with NPH insulin experiences significant nocturnal hypoglycaemia.

    R54.

    Monitor a person using a basal insulin regimen (NPH or a long-acting insulin analogue (insulin glargine) for the need for mealtime insulin (or a pre-mixed insulin preparation)). If blood glucose control remains inadequate (not to agreed target levels without problematic hypoglycaemia), move to a more intensive, mealtime plus basal insulin regimen based on the option of human or analogue insulins.

    R55.

    Monitor a person using pre-mixed insulin once or twice daily for the need for a further preprandial injection or for an eventual change to a mealtime plus basal insulin regimen, based on human or analogue insulins, if blood glucose control remains inadequate.

    11.3. Insulin detemir

    The GDG considered including sitagliptin and insulin detemir in this guideline; however, they were advised by NICE not to do so. NICE is undertaking a rapid update of recommendations in this guideline on second- and third-line drugs for managing blood glucose, which will cover these drugs. The updated guideline will be published early in 2009. For more information see www.nice.org.uk and search for ‘Type 2 diabetes newer agents’.

    11.4. Insulin delivery devices

    Insulin pumps are not considered here; they have been the subject of a recent NICE TA, and are not widely used in people with Type 2 diabetes.217

    11.4.1. Clinical introduction

    Insulin was previously normally delivered from syringes, necessitating accurate measuring of insulin doses drawn up from insulin vials under suitably hygienic conditions. Modern pen-injector devices obviate most of the problems of measuring up doses while avoiding most of the hygiene problems, and offer a convenient and safe means of carrying around injection equipment. However, several models of injector are available, including some designed for those with visual and physical impairments.

    The clinical question addressed here was whether any particular pen-injector had an evidence-based advantage over any other, including groups of people with difficulty using such devices.

    11.4.2. Methodological introduction

    Six crossover RCTs were identified which compared insulin pens or other delivery systems with conventional syringes.219–224 One study was excluded for methodological reasons.224 Two crossover RCTs were also identified which compared different types of insulin pens.220,225

    This area was not covered in detail by the previous guideline, and studies were only searched for from 1995 onwards to prevent the inclusion of obsolete devices.

    None of these studies were of a particularly high methodological quality with few reporting any details of randomisation, concealment or a power analysis. Few studies took into account the insulin delivery method that patients had used previously. Most studies assessed patient preference by use of their own specifically developed for purpose questionnaires; it was notable that some of these contained ‘leading’ questions.

    11.4.3. Health economic methodological introduction

    No health economic papers were identified for this question.

    11.4.4. Evidence statements: syringes vs other insulin delivery systems

    Glycaemic control

    One study found pre-lunch blood glucose values were lower during pen treatment (p<0.01) but no other significant differences were found between pens and syringes for blood glucose profiles or in terms of HbA1c.219 Three other studies found no differences between syringes and other delivery devices in terms of glycaemic control.221–223 Level 1+

    Hypoglycaemic episodes and adverse events

    Two studies noted no significant difference in the incidence of hypoglycaemic episodes between pens and syringe treatments.219,221 In other studies no AEs were considered by the investigator to be related to study treatment223 or the safety profiles for pen and the vial/syringe appeared similar.222 Level 1+

    Main patient acceptance outcomes

    Operational use

    In one study patients starting insulin using a pen found the insulin injections easy (63%) or very easy (33%) at the end of 12 weeks, whilst those who commenced insulin with conventional syringes found it more difficult with only 24% finding it very easy by the end of 12 weeks and 51% finding it easy (p=0.0005).221 Level 1+

    Other studies (which did not report significance) found that the operations needed for insulin administration with a pen compared to a syringe were faster (88%)219 and that the pen device was found easier to use overall compared to the syringe (74% vs 21% respectively).222 Level 1+

    In a study of patients with motor dysfunction and/or visual problems, an insulin injection device with a large easy-to-read dial, large push button for injection and audible clicks for each unit injected, was found to be easier to use compared to a vial and syringe by 82% of patients with the practical aspects of the injection device (dosing and injecting) rated as very easy or easy by 86%.223 Level 1+

    A study of visually impaired patients found that 80% were able to set and dispense three insulin doses after written instructions when using the insulin injection device with easy-to-read dial, large button for injection and audible clicks for units injected. This was significantly more than those using a syringe (27%, p<0.001) or a pen device (61%, p<0.001).220 Level 1+

    Pre-selection of dose

    A study comparing a pen with a conventional syringe and vial found that setting and drawing up the dose of insulin was significantly easier for patients using the pen (p=0.0490).221 Level 1+

    Other studies (which did not report significance) reported that 86% of participants found that pre-selection of insulin dose with a pen was easier than insulin withdrawal from a vial with a conventional syringe219 and that 85% of patients reported that they found it easier to read the insulin dose scale with the pen than the vial/syringe (10% found reading the insulin dose scale easier using the vial/syringe).222 Level 1+

    Pain

    A study found that injection pain was significantly lower with a pen than with syringes and vials (p=0.0018). Patients commencing on syringes reported a significantly lower level of injection pain after the switch to using the pen (p=0.0003).221 Another study reported participants found insulin injections with the pen, compared to the conventional syringe, were 55% less painful, although 43% did not notice any difference.219 Level 1+

    Preference for a device

    In the study of patients with motor dysfunction and/or visual problems, the insulin injection device with the easy-to-read dial, large button for injection and audible clicks for units injected, was significantly preferred to the vial and syringe (82% vs 10%, p<0.001).223 Level 1+

    In all studies comparing pens with conventional syringes more patients stated a preference for the pens over the conventional syringe and vial.219–222 Level 1+

    Insulin delivery devices vs other insulin delivery devices

    NovoPen® 3 vs HumaPen Ergo® vs Humalog Pen® vs InnoLet® vs FlexPen®

    Auditory confirmation of dose setting was heard by 100% of study participants for NovoPen® 3, 98% for FlexPen®, 90% for InnoLet®, 75% for HumaPen Ergo® and 63% for the Humalog Pen®. This was significantly different between the NovoPen® 3 and the Humalog Pen® (p<0.001), the HumaPen Ergo® (p<0.001), and InnoLet® (p<0.01), and the FlexPen® and the Humalog Pen® (p<0.001), and HumaPen Ergo® (p<0.01).225 Level 1+

    For tactile feedback, (the proportion of patients physically sensing they had dialled a correct dose) this was 100% for the FlexPen®, 92% for the NovoPen® 3, 81% InnoLet®, 67% HumaPen Ergo® and 50% for the Humalog Pen®. Significantly more patients reported that they had dialled the correct dose for the FlexPen® compared with the Humalog Pen® (p<0.001), HumaPen Ergo® (p<0.001) and InnoLet® (p<0.01). Significant differences were also noted between the NovoPen® 3 and Humalog Pen® (p<0.001) and the HumaPen Ergo® (p<0.01).225 Level 1+

    Patients reported most confidence in setting the correct dose when rating the NovoPen® 3 and FlexPen®. Scores for the NovoPen® 3 were significantly higher than those for the InnoLet® (p<0.001), HumaPen Ergo® (p<0.001) and Humalog Pen® (p<0.001), whereas the FlexPen® scored significantly higher than the Humalog Pen® (p<0.01).225 Level 1+

    InnoLet® vs Humulin Pen®

    In a group of visually impaired patients, the InnoLet® insulin device (easy-to-read dial, large button for injection and audible clicks for units injected) was found to be significantly more effective than the Humulin Pen® in terms of visual accuracy when reading the dose scale (92% vs 45%, p<0.001). Additionally, significantly more patients using InnoLet® were able to intuitively set and dispense a 20U insulin dose (84% vs 41%, p<0.001) and InnoLet® was significantly preferred to the Humulin Pen® (87% vs 13%, p<0.001).220

    11.4.5. From evidence to recommendations

    There was no strong published evidence that insulin pen injectors were a preferred option for insulin injection, but in clinical practice this was not questionable. The studies comparing devices did not compare all devices, were inevitably unblinded, and were manufacturer sponsored in single centres for the most part. The issue of bias was real. It was considered that some devices preformed better than others, but also that this was generally known to regular prescribers. Prescribers should be fully familiar with the devices they were recommending; this would be difficult for all the devices available.

    One injection device, the InnoLet®, was not a pen injector, but was aimed more at people with physical disabilities in manipulating injection systems. The studies were consistent with clinical experience in suggesting that this device was successful in enabling self-injection in some people who could not otherwise do it easily or reliably.

    Please refer to the Diabetes UK guidance for the issue of disposal of devices/sharps.

    RECOMMENDATIONS

    R56.

    Offer education to a person who requires insulin about using an injection device (usually a pen injector and cartridge or a disposable pen) that they and/or their carer find easy to use.

    R57.

    Appropriate local arrangements should be in place for the disposal of sharps.

    R58.

    If a person has a manual or visual disability and requires insulin, offer a device or adaptation that:

    • takes into account his or her individual needs
    • he or she can use successfully.
    Copyright © 2008, Royal College of Physicians of London.

    All rights reserved. No part of this publication may be reproduced in any form (including photocopying or storing it in any medium by electronic means and whether or not transiently or incidentally to some other use of this publication) without the written permission of the copyright owner. Applications for the copyright owner’s written permission to reproduce any part of this publication should be addressed to the publisher.

    Cover of Type 2 Diabetes
    Type 2 Diabetes: National Clinical Guideline for Management in Primary and Secondary Care (Update).
    NICE Clinical Guidelines, No. 66.
    National Collaborating Centre for Chronic Conditions (UK).

    PubMed Health Blog...

    read all...

    Recent Activity

    Your browsing activity is empty.

    Activity recording is turned off.

    Turn recording back on

    See more...