New insulins and insulin therapy


  • Disclosures: TD has received honoraria for speaking engagements or advisory board participation from several companies involved in the diabetes field. Furthermore he has received grant support from these companies (Abbott, Sanofi-Aventis, Bayer, Roche, Johnson&Johnson, Lilly, Medtronic, DexCom, NovoNordisk) for the conduct of studies or scientific meetings. JB is on the speaker's bureaux of Abbott, Medtronic and Sanofi-Aventis; and is a member of scientific advisory boards for AstraZeneca, Medtronic and Merck, Sharp & Dohme.

  • Endorsed by the International Conference on ATTD organized by Kenes International.

Thomas Danne,
Diabetes-Zentrum für Kinder and Jugendliche, Kinderkrankenhaus auf der Bult, Hannover, Germany
Tel.: +49-511-8115-340
Fax: +49-511-8115-344


The introduction of the so-called ‘designer’ insulins, the insulin analogues, has offered new opportunities in the clinical management of diabetes. Two additional new entities are close to reaching clinical practice. Linjeta™ (formally called VIAject) is not an analogue but rather a different formulation of human insulin which may give it a more rapid onset of action, potentially even faster than the currently available rapid-acting insulin analogues. Degludec™, on the other hand, is an insulin analogue molecule with an ultra-long clinical profile derived from the soluble multi-hexamer formation, resulting in a continuous slow and stable release of insulin degludec monomers which may last longer than currently available long-acting analogues. As with any new type of drug, the safety of the ‘designer’ insulins has to be closely scrutinised. Last year the increased cancer risk in diabetes entered the spotlight and the potential role of insulin analogues led to controversial discussions. In spite of recent new in vitro and observational data no new conclusive evidence became available. The need for multiple well-conducted and appropriately designed prospective observational studies to follow up the effectiveness and safety of the new insulins and the new insulin treatment regimens remains.

In this chapter it was our mission to chose articles published about “new insulins” over the last year that have the most important contribution for the on-going development of ultra-fast- and ultra-long-acting insulin analogs and preparations and their potential side-effects, particularly cancer. This has been done by means of PubMed searches as well as a review of abstracts of the recent large international diabetes meetings such as ADA, EASD and ISPAD.

Linjeta™– a new rapid acting insulin

Postprandial vascular effects of VIAject compared with insulin lispro and regular human insulin in patients with type 2 diabetes

T. Forst,1, A. Pfützner,1F. Flacke,2A. Krasner,2C. Hohberg,1E. Tarakci,1P. Pichotta,2S. Forst,1S. Steiner2

1Institute for Clinical Research and Development, Mainz, Germany, and2Biodel, Danbury, CT, USA

Diabetes Care 2010; 33: 116–20

Background: Prandial insulin delivery may influence the postprandial regulation of tissue blood flow. The effect of VIAject with human regular insulin and insulin lispro on postprandial oxidative stress and endothelial function was compared in patients with type 2 diabetes.

Methods: Fourteen patients (seven men) aged 61.5 ± 1.8 years, with mean diabetes duration of 6.6 ± 4.6 years and A1c 7.2% ± 0.5%, received a prandial injection of VIAject, human regular insulin and insulin lispro. The postprandial increases in asymmetric dimethylarginine (ADMA) and nitrotyrosine levels were checked at baseline and after a standardised liquid meal test (Ensure Plus), In addition, the postprandial effects on microvascular blood flow, skin oxygenation and vascular elasticity were measured.

Results: A significant reduction in the peak postprandial generation of ADMA was found with VIAject treatment compared with human insulin and insulin lispro (VIAject –27.3 ± 22.6 nmol/l, human insulin 97.7 ± 24.4 nmol/l and insulin lispro 66.9 ± 33.9 nmol/l; p < 0.05, respectively). The postprandial increases in nitrotyrosine levels were significantly less after VIAject than after human regular insulin (VIAject –0.22 ± 0.17 vs. human insulin 0.25 ± 0.15 μg/ml; p < 0.05), whereas nitrotyrosine after insulin lispro was in between (insulin lispro 0.09 ± 0.07 μg/ml). After VIAject earlier and more pronounced increases in microvascular blood flow and skin oxygenation were obtained compared with those after human insulin or insulin lispro (p < 0.05). All insulin formulations resulted in comparable improvements in central arterial elasticity.

Conclusions: Treatment with VIAject had an advantage compared with human regular insulin or insulin lispro in reduction of postprandial oxidative stress and improved endothelial function.

  • Comment: Linjeta™ (formally called VIAject) is a more rapid-acting insulin formulation that is currently under review by the US Food and Drug Administration (1). Data presented so far only in abstract form at the American Diabetes Association and European Association for the Study of Diabetes 2010 congresses indicate that it has a more rapid onset of action than the insulin analogues (2,3). The paper by Forst et al. focuses on the potential benefits of such an insulin formulation on postprandial metabolism in patients with type 2 diabetes. Linjeta™ is not an insulin analogue but rather a different formulation of human insulin which may promote a more rapid absorption. Early studies were done with insulin at pH 4 and a concentration of 25 U/ml. The marketable formulation of pH 7 and 100 U/ml apparently retains these characteristics of a more rapid onset but this has to be seen in larger trials (see figure). This insulin would be of particular benefit in type 1 diabetes, especially in insulin pump therapy. Early data suggest that it is biocompatible with insulin infusion sets (3). In current clinical practice, many patients claim that even when using rapid-acting insulin analogues the onset of action is still too slow to effectively provide postprandial glucose control. A large analysis of 1041 paediatric patients on continuous subcutaneous insulin infusion revealed the close correlation between the number of daily boluses and HbA1c. Only the subgroup of patients taking more than 12 boluses a day had an average HbA1c in the target range (= 87, HbA1c 7.3 ± 0.9) (4). Such frequent bolussing is likely to predispose to insulin stacking. Thus, any insulin with a more rapid onset of action and a shorter duration theoretically would be beneficial. This need for a more rapid insulin is also echoed from researchers involved with the artificial pancreas.
  • image(1)

[  A novel pH-neutral formulation of the monomeric insulin VIAject® has a faster onset of action than lispro. Leszek Nosek, Tim Heise, Frank Flack2, Alan Krasner2, Philip Pichotta2, Lutz Heinemann, Solomon Steiner2, Profil Institute for Metabolic Research, Neuss, Germany, 2Biodel, Danbury, CT, USA. Adapted from the EASD 2010 oral presentation, with permission. ]

Degludec™, a long-acting insulin analogue

Insulin degludec, a new generation ultra-long-acting insulin, used once daily or three times weekly in people with type 2 diabetes: comparison to insulin glargine (Abstract)

C. Mathieu,1 G. Fulcher,2P. V. Rao,3N. Thomas,4L. Endahl,5T. Johansen,5A. J. Lewin,6J. Rosenstock,7M. Pinget,8B. Zinmann9

1UZ Gasthuisberg K.U. Leuven, Belgium,2Royal North Shore Hospital, University of Sydney, Australia,3Nizam’s Institute of Medical Sciences University, Hyderabad, India,4Christian Medical College, Vellore, India,5Novo Nordisk A/S, Søborg,, Denmark,6National Research Institute, Los Angeles, CA, USA,7Dallas Diabetes and Endocrine Center at Medical City, Dallas, TX, USA,8University Hospital Strasbourg, Strasbourg, France, and9Samuel Lunenfeld Research Institute, Mount Sinai Hospital, University of Toronto, Canada

Diabetologia 2010; 53 (Suppl 1): S8

Background: Degludec (IDeg) is a neutral, soluble ultra-long-acting new generation insulin that forms multi-hexamer after subcutaneous injection. This randomised, four-arm, parallel group, treat-to-target trial examined the efficacy and safety of IDeg administered once daily vs. three times weekly in patients with type 2 diabetes who did not achieve good glycaemic control with oral antidiabetic medications.

Methods: Patients with a mean HbA1c level of 8.7% were randomised to treatment with subcutaneous insulin injected in the evening as follows: once daily IDeg (= 60), three times weekly IDeg (= 62), once daily alternative IDeg formulation (= 61) or insulin glargine once daily (= 62), all in combination with metformin during a 16-week period.

Results: At the end of the trial (16 weeks), the mean reduction from baseline in HbA1c and the final mean value of HbA1c were similar across the four treatment arms. Body weight remained stable and the rate of confirmed hypoglycaemia was low during the study in the four arms. No significant differences in adverse events (mainly mild or moderate in severity) were documented across the treatment arms.

Conclusions: IDeg used once daily or three times weekly was safe and well tolerated with similar glycaemic control to glargine.

Insulin degludec: less pharmacodynamic variability than insulin glargine under steady-state conditions (Abstract)

T. Heise,1L. Hermanski,1L. Nosek,1A. Feldmann,1S. Rasmussen,2T. K. Stryhn,2H. Haahr2

1Profil Institut für Stoffwechselforschung, Neuss, Germany, and2Novo Nordisk A/S, Søborg, Denmark

Diabetologia 2010; 53 (Suppl 1): S387

Background: Degludec (IDeg) is a new basal insulin analogue with action duration of more than 24 h. The aim of this randomised, double-blind, parallel group study was to compare the pharmacodynamic variability of IDeg with that of insulin glargine under steady-state conditions.

Methods: Patients with type 1 diabetes (= 54), aged 38 ± 10 years with a mean HbA1c of 7.7% ± 0.9%, were treated with 0.4 U/kg of IDeg or glargine once daily for 12 days.

Pharmacodynamic profiles were investigated over 24 h with the euglycaemic glucose clamp technique. Within-subject variability was estimated on log-transformed pharmacodynamic end-points, which were derived from the glucose infusion rate profiles during the clamps.

Results: IDeg produced significantly less pharmacodynamic variability than glargine on all protocol pharmacodynamic variability parameters, including total metabolic effect (p < 0.0001). The individual within-subject variability was consistently lower for IDeg compared with glargine. IDeg’s metabolic effect was exactly evenly distributed between the first and the second 12 h, and this distribution was less variable than with glargine (p < 0.001).

Both insulin types were well tolerated, without serious adverse events and without severe hypoglycaemic episodes. The rate of hypoglycaemic episodes was 166 (20 nocturnal) in the IDeg group compared with 182 (37 nocturnal) in the glargine group.

Conclusions: The results demonstrated that, under steady-state conditions, the IDeg administrated once daily is significantly less variable and more stable in glucose-lowering effect than glargine.

  • Comment: Degludec (NN1250) is a neutral, soluble ultra-long-acting new generation insulin with a duration of action of more than 24 h. So far no full publications are available, but these two abstracts were presented at this year’s diabetes meetings. Degludec has a very flat and smooth action profile. The product is intended for basal insulin treatment of diabetes and has the potential to be used less than once daily. It will be marketed by Novo Nordisk as Degludec (insulin degludec) and Degludec Plus (70% Degludec, 30% insulin aspart) and shows the potential to control blood glucose levels in type 1 and type 2 diabetes when injected once daily or thrice a week. The ultra-long clinical profile is derived from a soluble multi-hexamer formation that is believed to lead to a continuous slow and stable release of insulin degludec monomers. Clinical studies published so far only in abstract form showed that, after 16 weeks of treatment with insulin degludec, mean HbA1c reductions were similar across the once-daily and three-times-weekly insulin degludec groups and comparable to insulin glargine. The potential advantage may be in a particularly low risk of hypoglycaemia. In the preliminary data presented, 77% of patients treated with insulin degludec three times weekly did not experience any confirmed hypoglycaemia (defined as low blood glucose levels or episodes that required assistance) and, of those patients using insulin degludec once daily, 92% did not experience any confirmed hypoglycaemia. However, these rates were not significantly different from insulin glargine.

New insulins and the risk of cancer

Doses of insulin and its analogues and cancer occurrence in insulin-treated type 2 diabetic patients

E. Mannucci,1M. Monami,2D. Balzi,3B. Cresci,4L. Pala,4C. Melani,3C. Lamanna,1I. Bracali,2M. Bigiarini,4A. Barchielli,3N. Marchionni,2C. M. Rotella4

1Diabetes Agency, Careggi Teaching Hospital, Florence, Italy,2Section of Geriatric Cardiology and Medicine, Department of Cardiovascular Medicine, University of Florence, and Careggi Teaching Hospital, Florence, Italy,3Epidemiology Unit, Local Health Unit 10, Florence, Italy, and4Section of Endocrinology, Department of Clinical Pathophysiology, University of Florence, and Careggi Teaching Hospital, Florence, Italy

Diabetes Care 2010; 33: 1997–2003

Background: Epidemiological studies published in the last few years suggested that some insulin analogues could be associated with increased risk of cancer. The aim of this study was to assess the long-term association of different insulin analogues with cancer incidence.

Methods: A nested case–control study dataset was generated from the cohort study dataset (= 1340 insulin-treated diabetic outpatients) by sampling control subjects from the risk sets. For each case subject, the control subjects (up to five) were chosen randomly from those members of the cohort who were at risk for the same follow-up time as the case subject. Matching variables were based on 5-year age classes, sex and body mass index classes (< 18.5, 18.5–24.9, 25–29.9 and ≥ 30 kg/m2).

Results: During a median follow-up of 75.9 months (interquartile range 27.4–133.7), case subjects of incident cancer (= 112) were compared with matched control subjects (= 370). In case subjects, the mean daily dose of glargine was significantly higher than in control subjects (0.24 vs.0.16 IU/kg/day, p = 0.036). Glargine treatment with a dose ≥ 0.3 IU/kg/day was associated with incident cancer even after adjusting for Charlson comorbidity score, other types of insulin administration, and metformin exposure [odds ratio 5.43, 95% confidence interval (CI) 2.18–13.53, p < 0.001]. No association between incident cancer and insulin doses was found for human insulin or other analogues.

Conclusions: In light of the possible association between higher glargine doses and cancer, the dosage of glargine should always be considered in patients treated with this product.

Association of hyperglycaemia and insulin usage with the risk of cancer in type 2 diabetes: the Hong Kong diabetes registry

X. Yang,1G. T. Ko,2W. Y. So,1R. C. W. Ma,1L. W. Yu,1A. P. Kong,1H. Zhao,1C. C. Chow,1P. C. Tong,1,2J. C. Chan1–3

1Department of Medicine and Therapeutics, Chinese University of Hong Kong, Hong Kong, People’s Republic of China,2Hong Kong Institute of Diabetes and Obesity, Chinese University of Hong Kong, Hong Kong, People’s Republic of China, and3Li Ka Shing Institute of Health Sciences, Chinese University of Hong Kong, Hong Kong, People’s Republic of China

Diabetes 2010; 59: 1254–60

Background: Previous data have demonstrated that insulin has mitogenic effects and the presence of prolonged hyperglycaemia may be a risk factor for cancer in type 2 diabetes. However, it is unknown whether use of insulin with its effect on cell growth and proliferation increases cancer risk or its glucose-lowering effect decreases cancer risk.

Methods: From a cohort of Chinese patients with type 2 diabetes (= 4623), free of cancer and naïve to insulin at enrolment, a 1:2-matched new insulin user cohort was selected based on age (± 3 years), smoking status, and likelihood of initiating insulin therapy (± 0.05). Stratified Cox regression analysis on the matched pairs was used to obtain hazard ratios (HRs) of insulin therapy and A1c for cancer risk.

Results: Of 973 new insulin users, 971 had matched non-users (= 1935). The cancer incidence was significantly higher in insulin non-users than in insulin users (49.2 vs. 10.2, per 1000 person-years, p < 0.0001). After further adjustment, A1c was associated with an increased cancer risk (HR per percentage 1.26, 95% CI 1.03–1.55), whereas use of insulin was associated with a decreased cancer risk (HR of insulin users vs. non-users 0.17, 95% CI 0.09–0.32). Similar results were found when including all 973 insulin users and 3650 non-users in the analyses.

Conclusions: This study demonstrated that prolonged hyperglycaemia with elevated A1c levels was related to increased cancer risk, whereas insulin usage was associated with a reduced cancer risk among Chinese patients with type 2 diabetes.

  • Comment: We reviewed the issue of the risk of malignancy in patients treated with insulin analogues extensively in the 2009 Yearbook (5). Several additional position statements and commentaries and some new studies (6–11) have been published since then. Clearly, in vitro and pre-clinical studies themselves cannot be considered conclusive and should primarily inform design considerations for observational studies. The 2009 conclusion, that a relationship between insulin analogues, particularly glargin, and cancer cannot be confirmed or excluded on the basis of the currently available data, remains true also in 2010.

In the paper by Mannucci et al. cited above, a monocentric case–control study from a small region in Florence with a total of 482 cases and a median follow-up of 75.9 months appears to support the 2009 paper of Hemkens et al. (12) as incident cancer was associated with a dose of glargin ≥ 0.3 IU/kg/day but no association was found for human insulin or other analogues. The primary objective of the Mannucci study was the question of an increased cancer rate in Lantus-treated patients. This was not found. As in the previous controversial studies, post hoc analyses have shown the association of Lantus dose and cancer. However, as in the previous studies several biases (no matching for body mass index etc.) have been introduced and the observations are based on 29 cases with Lantus, 13 above the arbitrary threshold of 0.3 IU/kg and 16 below. Again significant differences in the overall type of insulin therapy were notable between Lantus and human insulin users, as approximately 85% of the Lantus users also had prandial insulin, compared with less than 60% of those on NPH. Also the cancer incidence in this region is about four times higher than usual with low rates of breast and colorectal cancer and a high rate of other gastrointestinal cancers.

A recent Chinese study of 4623 patients with type 2 diabetes with a 5-year follow-up, of whom 973 were newly treated with insulin, found a higher incidence of cancer in non-insulin users (49.2 per 100 patient years) compared with insulin-treated patients (10.2 per 100 patient years) resulting in a decreased risk of cancer (hazard ratio 0.17) when using insulin. They found a relationship of HbA1c with cancer incidence, a factor that was not analysed in most of the previous studies, adding new complexity to the issue. A consensus report of US diabetologists and oncologists (13) concludes that there is an urgent need of multiple well-conducted and appropriately designed prospective observational studies.

Insulin therapy in type 2 diabetes

Optimal insulin regimens in type 2 diabetes mellitus: systematic review and meta-analyses

D. S. Lasserson,1P. Glasziou,1R. Perera,1R. R. Holman,2A. J. Farmer1

1Division of Public Health and Primary Health Care, University of Oxford, Oxford, UK, and2Diabetes Trials Unit, Oxford Centre of Diabetes, Epidemiology and Metabolism, University of Oxford, Oxford, UK

Diabetologia 2009; 52: 1990–2000

Background: To compare the effects of different insulin regimens (biphasic, basal or prandial insulin supplementation) on glucose control, clinical outcomes and adverse events in patients with type 2 diabetes.

Methods: A systematic review and meta-analyses of randomised controlled trials reported up to October 2008 were carried out.

Results: Twenty-two trials with altogether 4379 randomised patients were included. Pooled data from insulin-naïve patients were used in the meta-analyses. Using basal insulin supplementation as reference, greater HbA1c reductions were registered with biphasic insulin (0.45%; 95% CI 0.19–0.70, p = 0.0006) and prandial insulin (0.45%; 95% CI 0.21–1.65, p = 0.002), but with smaller decreases in fasting glucose of 0.93 mmol/l (95% CI 0.21–1.65, p = 0.01) and 2.20 mmol/l (95% CI 1.70–2.70, p < 0.00001), respectively. Larger insulin doses at study end were observed with biphasic and prandial insulin regimens. There were no differences between the three insulin regimens with regard to major hypoglycaemic events, whereas minor hypoglycaemic episodes with biphasic or prandial insulin were variably observed as higher than or similar to basal insulin therapy. Compared with basal insulin, prandial insulin therapy resulted in a larger increase in body weight (1.86 kg; 95% CI 0.80–2.92, p = 0.0006).

Conclusions: A more favourable effect on glycaemic control may be achieved by initiating insulin therapy in patients with type 2 diabetes with biphasic or prandial insulin rather than with basal insulin supplementation, but the risk of hypoglycaemia may be increased with the former regimens.

Three-year efficacy of complex insulin regimens in type 2 diabetes

R. R. Holman,1,2A. J. Farmer,2,3M. J. Davies,4J. C. Levy,2J. L. Darbyshire,1,2J. F. Keenan,1,2S. K. Paul,1,2for the 4-T Study Group

1Diabetes Trials Unit,2Oxford Centre for Diabetes, Endocrinology and Metabolism, and3Department of Primary Health Care and National Institute for Health Research, School of Primary Care Research, University of Oxford, Oxford, UK, and4Department of Cardiovascular Sciences, University of Leicester, Leicester, UK

N Engl J Med 2009; 361: 1736–47

Background: To investigate the optimal insulin regimen in patients with type 2 diabetes.

Methods: The three-year, open-label, multicentre trial included patients with type 2 diabetes (= 708) with suboptimal glycaemic control despite on-going combination therapy with metformin and sulphonylurea. The patients were randomised to insulin therapy based on biphasic insulin aspart twice daily, prandial insulin aspart three times daily, or basal supplementation with insulin detemir once or twice daily. During the first year, sulphonylurea therapy was substituted by an additional type of insulin in the case of unacceptable hyperglycaemia (HbA1c > 10% or ≥ 8% on two consecutive occasions at or after 24 weeks of therapy), and thereafter if HbA1c was greater than 6.5%. Effects on HbA1c, proportion of patients with HbA1c ≤ 6.5%, hypoglycaemia frequency and body weight were evaluated.

Results: Median HbA1c was similar with the three insulin therapies: biphasic insulin 7.1%, prandial insulin 6.8%, and basal insulin supplementation 6.9%, respectively. Of the patients receiving biphasic insulin 31.9% attained HbA1c ≤ 6.5%, which was less than in the prandial group (44.7%, p = 0.006) and in the basal group (43.2%, p = 0.03). In total, 67.7% in the biphasic, 73.6% in the prandial and 81.6% in the basal insulin group, respectively, received an additional type of insulin (p = 0.002). Median rate of hypoglycaemia was highest in the prandial group (5.7 hypoglycaemic events per patient per year), intermediate in the biphasic group (3.0) and lowest in the basal group (1.7) (p < 0.001). The increase in body weight was higher in the prandial insulin group than in the biphasic and basal insulin groups. No differences in other adverse events were registered.

Conclusions: In patients with type 2 diabetes with inadequate glycaemic control while on combined metformin and sulphonylurea therapy, addition of basal or prandial insulin resulted in better HbA1c control than in those who were given a biphasic insulin regimen. Basal insulin supplementation was associated with fewer hypoglycaemic events and lesser increase in body weight.

  • Comment: Owing to the progressive nature of the disease, many patients with type 2 diabetes will eventually require exogenous insulin supplementation to maintain adequate glucose control. While several types of insulin preparations (human, analogues, various forms of premixed) and a number of different insulin regimens are accessible, the optimal mode of insulin therapy in type 2 diabetes is a matter of uncertainty. Moreover, feasible combination therapies with other glucose-lowering agents as well as unwanted side-effects of the therapy regimens, such as risk of hypoglycaemia and weight gain, need to be considered.

Seemingly, the two referenced studies arrived at contradictory conclusions: the meta-analyses by Lasserson et al. favoured biphasic (premixed) insulin supplementation when combined effects on glycaemic control, risk of hypoglycaemia and weight gain were considered, whereas the prospective trial by Holman and co-workers indicated superior outcomes when insulin therapy was initiated with basal insulin. These dissimilarities may be more apparent than real, however.

The meta-analyses were based on a mixture of short-term studies (3–12 months) in insulin-naïve type 2 diabetes patients who had initiated insulin therapy with conventional or analogue insulins alone or in various combinations with oral hypyglycaemic agents (metformin, sulphonylureas and/or thiazolidinediones). Complete information about titration algorithms, glycaemic targets and insulin doses at study end were not available. Although these variations were accounted for in the statistical evaluation of the data, some of the included trials may have been too short to allow sufficient up-titration of insulin.

The referenced paper by Holman et al. is an extension of the 4-T study, in which patients with inadequate glycaemic control despite maximum tolerable doses of metformin and sulphonylurea were randomised to add-on therapy with biphasic (premixed insulin aspart), prandial insulin aspart or basal insulin supplementation with insulin detemir, using standardised insulin dosing algorithms. The 1-year results of the 4-T study published in 2007 (14) showed that prandial and biphasic insulin therapy resulted in greater reductions in HbA1c levels than basal insulin. The risk of hypoglycaemia and weight gain, on the other hand, were highest in the prandial group, intermediate in the biphasic group, and lowest in patients randomised to basal insulin therapy. Notably, these results were included in the meta-analyses by Lasserson et al., and most probably had a major influence on the outcome of these assessments.

In contrast, in the 3-year evaluation of the 4-T study – which was not included in the meta-analyses – the median HbA1c levels were comparable with the three insulin regimens, albeit with a greater proportion of patients attaining HbA1c ≤ 6.5% in the prandial and basal insulin groups than in the biphasic insulin arm. It should be emphasised that substitution of sulphonylurea therapy with the addition of a second type of insulin supplementation because of insufficient glycaemic control (defined as HbA1c ≤ 6.5%) was necessary in the majority of patients in all three groups; this was done by adding midday prandial aspart insulin in the biphasic group, bedtime basal insulin in the prandial group, and prandial aspart insulin three times daily in the basal insulin group, respectively. This occurred most frequently in the basal insulin group (82%), with intermediate frequency in the prandial group (74%) and least often in the biphasic insulin group (68%). In other words, approximately half the patients in each insulin arm who received additional insulin therapy attained the treatment goal of HbA1c ≤ 6.5%. Hence, within a relatively short period of time, more complex insulin regimens seem to be needed in most patients to maintain the glycaemic target, irrespective of the initial mode of insulin therapy. Considering also the risk of hypoglycaemic events and weight gain, initiating add-on insulin therapy with basal insulin supplementation appears to be a feasible strategy in most patients who fail on oral therapy, with subsequent escalation to basal prandial insulin regimens when needed. Whether this should be carried out with human insulins or with the use of insulin analogues is another disputed issue. However, taking treatment convenience and quality of life aspects into account, less complex biphasic-based insulin regimens may still be appropriate in many cases.