Diabetes, Obesity and Metabolism 12: 437–441, 2010.
© 2010 Blackwell Publishing Ltd
Adding insulin glargine vs. NPH insulin to metformin results
in a more efficient postprandial β-cell protection in
individuals with type 2 diabetes
T. Forst1,2, M. Larbig3, C. Hohberg1, S. Forst1, S. Diessel1, M. Borchert1, W. Roth1& A. Pf¨ utzner1
1Institute for Clinical Research and Development, Clinical Department, Mainz, Germany
2Johannes Gutenberg University Mainz, Department of Endocrinology, Mainz, Germany
3Sanofi-aventis, Medical Department, Berlin, Germany
Aim: Postprandial release of intact proinsulin (IP) is an independent marker for β-cell dysfunction in patients with type 2 diabetes. This
open-label, parallel-group, two-arm, pilot study compared the β-cell protective effect of adding insulin glargine (GLA) vs. NPH insulin to ongoing
Material and methods: Overall, 28 insulin-naive type 2 diabetes subjects (mean ± SD age, 61.5 ± 6.7 years; diabetes duration, 9.8 ± 6.5
years; HbA1c, 7.1 ± 0.5%; BMI, 30.7 ± 4.3 kg/m2) treated with metformin and sulfonylurea were randomized to add once-daily GLA or NPH
at bedtime. At baseline and after 3 months, subjects received a standardized breakfast, lunch and dinner, with pre- and postprandial blood
sampling to measure plasma IP, total insulin and blood glucose (BG).
Results: Insulin dose after 3 months was comparable in both groups (GLA vs. NPH: 23.6 ± 13.4 vs. 23.3 ± 12.7; p = NS). Both treatments
significantly reduced fasting BG levels (GLA: 158 ± 19 to 121 ± 23 mg/dl; NPH: 156 ± 34 to 119 ± 29 mg/dl; both p < 0.01 vs. baseline).
Fasting and postprandial BG levels did not differ between groups. IP levels decreased in both groups (p < 0.05 at all timepoints). Although
IP release after breakfast did not differ between treatments, GLA induced a greater reduction in IP release after lunch (p = 0.08) and dinner
(p = 0.04). Total plasma insulin levels did not differ between groups.
Conclusions: Adding basal insulin to metformin reduces postprandial β-cell load. While GLA and NPH had comparable effects at breakfast,
GLA reduces β-cell stress more effectively at dinner, and with a trend at lunch, most probably because of its longer lasting pharmacodynamic
Keywords: beta cell stress, insulin glargine, intact proinsulin, NPH insulin
Date submitted 23 November 2009; date of first decision 24 December 2009; date of final acceptance 19 January 2010
Increasing β-cell stress with subsequent failure to release suf-
ficient amounts of biologically active insulin is thought to
precede the deterioration of blood glucose control in individ-
uals with type 2 diabetes treated with oral agents [1,2]. Type
2 diabetes is classically monitored by measurement of labora-
tory markers, such as HbA1c, glucose, lipids, body mass index
and blood pressure. In addition to these traditional laboratory
markers, the measurement of intact proinsulin (IP) levels may
provide pursuing information with regard to β-cell function
and disease stage [3,4].
In addition to the much weaker glucose-lowering activity,
the increasing IP levels have been reported to be associ-
ated with a substantial increase in cardiovascular risk and
are now considered to be an independent cardiovascular
Correspondence to: Prof. Thomas Forst, Department of Endocrinology, Institute for Clinical
Research and Development, University Mainz, Parcusstrasse 8, D-55116 Mainz, Germany.
Re-use of this article is permitted in accordance with the Terms and Conditions set out at
risk marker in subjects, with and without disturbed glucose
metabolism [5–7]. In individuals with type 2 diabetes, the
increasing insulin requirements, owing to insulin resistance or
excess β-cell stimulation by sulfonylurea, result in β-cell
overload with impaired processing of IP into insulin and
C-peptide [8,9]. On the contrary,recent studies have indicated
that the early introduction of insulin treatment, even at low
doses that are insufficient to restore blood glucose control can
reduce β-cell stress, improve endogenous insulin processing
and mightcontribute to an improvement in the overallcardio-
vascular risk profile by reducing circulating IP levels [10–12].
In individuals with type 2 diabetes, a combination of oral
antidiabetic drugs (OAD) and basal insulin is commonly
initiated after OAD treatment has failed to achieve sufficient
metabolic control. In these subjects, basal insulin is usually
given in the evening to suppress hepatic gluconeogenesis
overnight, while OADs are continued to provide sufficient
blood glucose control during the day. Recent studies
comparing the long-acting insulin analogue glargine (GLA)
vs. NPH insulin in combination with metformin revealed the
advantages of insulin GLA, owing to its flat and long-lasting
DIABETES, OBESITY AND METABOLISM
The goal of the recent study was to investigate the effect of
basal insulin treatment,by adding insulin GLAor NPH insulin
to metformin, on postprandial β-cell function in patients
with type 2 diabetes using a combination of metformin and
Research Design and Methods
combination of metformin and sulfonylurea were randomized
to receive treatment with insulin GLA or NPH insulin at
bedtime along with 1000 mg metformin twice daily. The doses
of metformin and sulfonylurea were to be stable over the last 3
for study participation were an HbA1c level between 6.5 and
8.5%, and an intact fasting proinsulin level between 7 and
20 pmol/l.Subjects were excluded iftheyhad been treatedwith
glinides or glucosidase inhibitors within the last 4 weeks prior
to the screening visit. All other concomitant treatments were
therapy, both insulin treatments were titrated to reach a target
fasting glucose level of 100 mg/dl (5.6 mmol/l).
in the morning after an 8-h overnight fast for a test meal day
consisting of three standardized meals. For this test meal day,
the subjects were hospitalized in a comfortable environment
and an intravenous cannula for blood sampling was placed
standardized breakfast at 08:00 hours (434 kcal, 26.7 g protein,
15 g fat, 48 g carbohydrates), a standardized lunch at 12:00
hours (642 kcal, 48 g protein, 25 g fat, 53 g carbohydrates) and
a standardized dinner at 18:00 hours (427 kcal, 18 g protein,
23 g fat, 55 g carbohydrates). Blood samples were collected
before the test meals and 60 and 120 min after food intake to
measure plasma glucose, insulin and IP levels.
of Helsinki and was approved by the local ethical committee.
All subjects gave a written, informed consent.
All laboratory measurements were analysed at the Institute
for Clinical Research and Development (ikfe GmbH, Mainz,
Germany). Blood samples were centrifuged and maintained
at −20◦C until analysis. Plasma glucose concentrations were
determined by the glucose dehydrogenase method (Super GL,
RLT, M¨ ohnesee-Delecke, Germany). Insulin was measured
by a chemiluminescence assay (Invitron, Monmouth, UK),
which shows a high cross-reactivity with insulin GLA and
NPH Insulin. Therefore, the plasma insulin levels given in
the study represent the total insulin plasma level comprising
endogenous and exogenous insulin. Intact proinsulin was
measured using an enzyme-linked immunosorbent assay
(LincoResearch, St Charles, MO, USA) and HbA1c was
determined by high-performance liquid chromatography
(Menarini Diagnostics, Neuss, Germany).
Adverse events experienced by subjects during the study were
documented by the investigator at each visit.
Subjects Sample Size Considerations and Statistical
the effects of basal insulin supplementation on postprandial IP
secretion. Therefore, this study was designed as a pilot study,
without confirmatory sample size consideration. The number
of participating subjects was estimated based on a previous
study investigating the effect of low-dose prandial insulin
supplementation on postprandial IP levels . Enrolment of
Results are presented using descriptive summary statistics.
(SD). For the postprandial time course of IP levels, the
area under the curve (AUC) was calculated according to
the trapezoidal rule. Statistical comparison between baseline
and at 3 months of insulin treatment, and between the two
treatment groups were performed using the Student’s t-test
(paired and unpaired as appropriate). A two-tailed p < 0.05
was considered statistically significant.
Thirty subjects were randomized (15 individuals in each study
arm). One subject in the NPH insulin group terminated the
follow-up investigation. One subject in the insulin GLA group
was excluded owing to abnormal IP levels. In total, 28 subjects
(14 subjects in each group) were included in the per-protocol
analysis. Baseline demographic data and clinical characteristics
of the two groups are presented in Table 1.
Slight but non-significant differences between both study
As shown in Table 1, HbA1c at the start of the study
was comparable between the two groups. A slight, but
non-significant reduction in HbA1c was observed with both
treatments. At the end of the study, the HbA1c levels were
comparable between both treatment groups (insulin GLA: 6.9
Table 1. Clinical characteristics of the study subject.
66.9 ± 6.2
30.0 ± 3.7
11.6 ± 7.5
7.1 ± 0.6
58.0 ± 8.5
31.5 ± 4.9
8.6 ± 4.7
7.1 ± 0.4
Body mass index (kg/m2)
Duration of diabetes (years)
438 T. Forst et al.Volume 12 No. 5 May 2010
DIABETES, OBESITY AND METABOLISM
Table 2. Blood glucose, insulin and intact proinsulin levels at baseline and 3 months after treatment with insulin GLA or NPH insulin.
NPH Baseline156 ± 34217 ± 44204 ± 41 141 ± 34 158 ± 41 140 ± 32100 ± 16 145 ± 33147 ± 37
3 months 119 ± 29∗
3 months 121 ± 23∗
197 ± 28
233 ± 27
200 ± 29
183 ± 38
211 ± 29
181 ± 28∗
64.0 ± 25.5
136 ± 38
131 ± 28
117 ± 24
39.6 ± 20.7
163 ± 36
142 ± 26
160 ± 23
52.3 ± 22.5
146 ± 28
131 ± 20
151 ± 23
46.6 ± 17.6
108 ± 15
102 ± 16
108 ± 19
24.0 ± 16.9
156 ± 21
147 ± 18
159 ± 21
35.8 ± 19.1
154 ± 21
148 ± 19
158 ± 26
39.5 ± 22.3
158 ± 19
18.6 ± 11.3 56.2 ± 22.2
3 months 20.0 ± 9.6
3 months 19.4 ± 7.8∗44.9 ± 16.7
NPH Baseline20.5 ± 12.7 33.3 ± 20.5
46.8 ± 20.7∗51.6 ± 27.5∗33.0 ± 16.7
55.5 ± 27.7
49.8 ± 24.8
48.3 ± 27.3
42.5 ± 15.6∗33.2 ± 13.7∗17.6 ± 9.3∗
44.3 ± 21.2
24.6 ± 11.1∗36.9 ± 15.7
47.6 ± 27.4
28.9 ± 11.9∗29.0 ± 10.8∗
28.0 ± 13.1
25.7 ± 9.3
41.1 ± 22.8
14.7 ± 8.366.7 ± 43.1 42.4 ± 23.238.8 ± 30.4
31.5 ± 13.7
57.5 ± 33.4
17.5 ± 8.8
16.2 ± 6.2
38.5 ± 24.7
31.2 ± 15.2
26.7 ± 8.9
45.2 ± 26.9 Proinsulin
54.4 ± 31.4
3 months 7.2 ± 6.3∗13.5 ± 9.5∗
16.5 ± 6.1
5.7 ± 2.8∗12.8 ± 5.8∗
23.1 ± 13.0∗22.3 ± 13.7∗25.5 ± 14.7∗24.4 ± 14.2∗15.0 ± 11.4∗17.8 ± 12.0∗20.2 ± 12.5∗
44.9 ± 15.6
23.1 ± 13.0∗17.7 ± 7.9∗
27.1 ± 7.5 48.9 ± 12.9 48.6 ± 14.7
20.1 ± 9.1∗
48.0 ± 16.6
20.7 ± 6.7
29.7 ± 10.4
10.6 ± 5.3∗
31.4 ± 11.0
12.5 ± 5.7∗
36.4 ± 11.9
14.5 ± 5.9∗
Mean ± standard deviation. GLA, glargine.
∗p < 0.05 at 3 months vs. baseline.
± 0.6%; NPH insulin: 6.8 ± 0.7%). Insulin doses at 3 months
were comparable with both treatment regimens (insulin GLA:
23.6 ± 13.4 IU; NPH insulin: 23.3 ± 12.7 IU; p = NS). The
postprandial plasma glucose, insulin and IP levels recorded
during the test meal days at baseline and after 3 months of
insulin treatment are presented in Table 2.
A significant reduction in fasting blood glucose levels from
baseline to endpoint was observed for both insulin treatments,
insulin concentrations increased significantly after adding
insulin GLA, but not with NPH insulin. No other consistent
difference in the time course of total (endogenous and exoge-
nous) blood insulin concentrations from baseline to endpoint
or between the two basal insulin treatments were observed.
As shown in Table 2, adding once-daily basal insulin to
metformin resulted in a remarkable decrease in fasting and
postprandial IP levels at all test meals. Figure 1 shows the AUC
While the total release of IP after breakfast was comparable
between the two insulin treatments, the postprandial release
of IP tend to be lower after lunch, and was significantly lower
after dinner with insulin GLA when compared with NPH.
Overall, 12 individuals (six people in each treatment group)
reported a total of 16 adverse events (6 and 10 events in
the insulin GLA and NPH insulin group, respectively). The
most frequent type of adverse events in the NPH insulin
group was musculoskeletal and connective tissue disorders
(three patients reporting four events) and in the insulin GLA
group was infections (two patients reporting one event each).
Hypoglycaemiawas reported by two patients (one event each),
both occurring in the NPH insulin arm. On the contrary,
the only adverse event that was considered serious (a case of
Meniere’s disease) occurred in the insulin GLA group.
In individuals with type 2 diabetes, OAD treatment over time
is often followed by a deterioration in β-cell function, as
indicated by a sustained loss of blood glucose control [1,2].
Early β-cell dysfunction may have important physiological
implications and may serve as a target for novel treatment
strategies for diabetes mellitus . Even before β-cell failure
becomes apparent with a clinically detectable deterioration
in metabolic control, there is a marked increase in release
of IP, which was reported to be an early marker for β-cell
dysfunction [3,15,16]. Furthermore, stimulation of the β cell
by sulfonylureas might impair the conversion rate from IP to
insulin and C-peptide [8,9]. On the contrary, the initiation of
insulin in individuals with type 2 diabetes, even if it is not
titrated to reduce blood glucose levels, was found to reduce
β-cell stress and improve the conversion rate of IP [11,12].
Once OADs are unable to maintain glycaemic control, insulin
treatment in type 2 diabetes is often initiated by basal insulin
supplementation in the evening, while maintaining OADs
with inadequate control with sulfonylurea and metformin
treatment, the initiation of basal insulin treatment not only
offers a reduction in fasting glucose levels but also improves
β-cell function during the day. Although a reduction in
postprandial β-cell stress was observed in our study, there
was no significant improvement in postprandial blood glucose
levels after breakfast, lunch and dinner. Therefore, this finding
suggests that the initiation of insulin supplementation in
individuals with type 2 diabetes might evolve β-cell protective
effects which go beyond the regulation of blood glucose.
In addition to its use as a marker forβ-cell damage, elevated
IP levels were shown to stimulate adipocytes and decrease
to increase plasminogen activator inhibitor-1 levels, which is
Volume 12 No. 5 May 2010 doi:10.1111/j.1463-1326.2010.01209.x 439
DIABETES, OBESITY AND METABOLISM
1 h2 h
AUC pp intact proinsulin
1 h 2 h
p = 0.08
AUC pp intact proinsulin
p = 0.05
p = 0.04
AUC pp intact proinsulin
Figure 1. Area under the curve (AUC) for 1- and 2-hour postprandial
intact proinsulin levels after 3 months of treatment with adding evening
insulin glargine vs. NPH insulin to metformin 1000 mg twice daily at
breakfast (A), lunch (B), dinner (C) (white: NPH Insulin, black: Insulin
In addition, elevated IP levels have been shown to predict
coronary atherosclerosis  and the risk of cardiovascular
events in subjects with and without diabetes [5,6,19,20].
Furthermore, we recently reported a postprandial increase in
IPlevels in non-diabeticsubjects with increased cardiovascular
events . Therefore, our findings of a marked overall
of basal insulin treatment might have important implications
not only for metabolic control, but also for cardiovascular risk
reduction in individuals with type 2 diabetes.
Insulin GLA is a long-acting human insulin analogue with
a longer time–action profile and no pronounced peak of
action when compared with NPH insulin . In a recent
study, treatment with insulin GLA in combination with OADs
achieved better postprandial metabolic control when com-
pared with NPH insulin in combination with OADs . In
our study population, no significant differences in glucose
control were observed between the two insulin formulations.
On the contrary, treatment with insulin GLA in combination
prandial release of IP after lunch and dinner when compared
with NPH insulin. Owing to the shorter time–action profile of
NPH insulin, the rate of insulin release from the subcutaneous
tissue depot is more rapid, which exhausts the supply more
quickly and, thus, requires earlier endogenous insulin release.
The greater demand on β cell will become evident, particularly
after a meal, when the requirements for insulin are high. In
individuals with type 2 diabetes who have barely compensated
β-cell function, this will lead to a greater release of IP from the
exhausted β cells [23–25]. Despite comparable glucose con-
trol, the prolonged pharmacodynamic profile of insulin GLA
results in stronger β-cell protection, lasting over 24 h.
The comparable total plasma insulin levels found in both
enhanced endogenous insulin release was able to compensate
for the shorter time–action profile of NPH insulin, thereby
keeping blood glucose levels comparable between the study
insulin-treated subjects will be followed by an increase in the
release of IP, particularly after meals, as observed in our study.
A potential limitation of our findings is that this was an
exploratory pilot study to evaluate the protective effects on
the β cell by initiating basal insulin therapy with metformin
in individuals with type 2 diabetes, pretreated with OADs
are needed to investigate if the effect of basal insulin
supplementation will translate into longer β-cell survival or
a reduction in cardiovascular risk. In addition, our study only
compared once-daily application of insulin GLA, with once-
daily injection of NPH insulin. Most probably, NPH insulin
In conclusion, the initiation of basal insulin in combination
with metformin results in an overall improvement of β-
cell function, as indicated by a reduction in the fasting and
postprandial release of IP from the β cells. Because of the
protracted pharmacokinetic profile of insulin GLA compared
with NPH insulin, treatment with insulin GLA may offer
insulin applied once daily.
This studywassupported byan unrestricted fund fromSanofi-
aventis, Berlin, Germany. Editorial support for this article
was provided by the Global Publications group of Sanofi-
aventis. The Clinical trial registry number (ClinicalTrials.gov)
Conflicts of Interest
Prof. Dr. Andreas Pf¨ utzner and Prof. Dr. Thomas Forst
received unrestricted research grants and speaker fees from
Sanofi-aventis. Dr. Martin Larbig is an employee of Sanofi-
aventis. Dr. Cloth Hohberg, Senait Forst, Dr. Stepahn Diessel,
440 T. Forst et al. Volume 12 No. 5 May 2010
DIABETES, OBESITY AND METABOLISM Download full-text
Marcus Borchert, and Werner Roth have declared no conflict
1. UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose
control with sulphonylureas or insulin compared with conventional
treatment and risk of complications in patients with type 2 diabetes
(UKPDS 33). Lancet 1998; 352: 837–853.
2. Kahn SE, Haffner SM, Heise MA et al. Glycemic durability of rosiglitazone,
metformin, or glyburide monotherapy. N Engl J Med 2006; 355:
3. Pfutzner A, Kunt T, Hohberg C et al. Fasting intact proinsulin is a highly
specific predictor of insulin resistance in type 2 diabetes. Diabetes Care
2004; 27: 682–687.
4. Kahn SE. The importance of the beta-cell in the pathogenesis of type 2
diabetes mellitus. Am J Med 2000; 108(Suppl. 6a): 2S–8S.
5. Zethelius B, Lithell H, Hales CN, Berne C. Insulin sensitivity, proinsulin and
insulin as predictors of coronary heart disease. A population-based 10-
year, follow-up study in 70-year old men using the euglycaemic insulin
clamp. Int J Obes Relat Metab Disord 2005; 48: 862–867.
6. Alssema M, Dekker JM, Nijpels G, Stehouwer CD, Bouter LM, Heine RJ.
Proinsulin concentration is an independent predictor of all-cause and
cardiovascular mortality: an 11-year follow-up of the Hoorn Study.
Diabetes Care 2005; 28: 860–865.
7. Forst T, Pfutzner A, Lubben G et al. Effect of simvastatin and/or
pioglitazone on insulin resistance, insulin secretion, adiponectin, and
proinsulin levels in nondiabetic patients at cardiovascular risk–the PIOSTAT
Study. Metabolism 2007; 56: 491–496.
8. Dworacka M,
Szczawinska K. Disproportionately elevated proinsulin levels in type 2
diabetic patients treated with sulfonylurea. Int J Clin Pharmacol Ther 2006;
Abramczyk M,Winiarska H, Kuczynski S,Borowska M,
9. Inoguchi T, Umeda F, Kakimoto M et al. Chronic sulfonylurea treatment
and hyperglycemia aggravate disproportionately elevated plasma
proinsulin levels in patients with type 2 diabetes. Endocr J 2000; 47:
10. Forst T, Hohberg C, Pfutzner A. Cardiovascular effects of disturbed insulin
activity in metabolic syndrome and in type 2 diabetic patients. Horm
Metab Res 2009; 41: 123–131.
11. Jain SK, Nagi DK, Slavin BM, Lumb PJ, Yudkin JS. Insulin therapy in type
2 diabetic subjects suppresses plasminogen activator inhibitor (PAI-1)
activity and proinsulin-like molecules independently of glycaemic control.
Diabet Med 1993; 10: 27–32.
12. Pfutzner A, Lorra B, Abdollahnia MR et al. The Switch from Sulfonylurea
to Preprandial Short- Acting Insulin Analog Substitution Has an Immediate
and Comprehensive beta-Cell Protective Effect in Patients with Type 2
Diabetes Mellitus. Diabetes Technol Ther 2006; 8: 375–384.
13. De Mattia G, Laurenti O, Moretti A. Comparison of glycaemic control in
patients with Type 2 diabetes on basal insulin and fixed combination oral
antidiabetic treatment: results of a pilot study. Acta Diabetol 2009; 46:
14. Wallander M, Bartnik M, Efendic S et al. Beta cell dysfunction in patients
with acute myocardial infarction but without previously known type 2
diabetes: a report from the GAMI study. Int J Obes Relat Metab Disord
2005; 48: 2229–2235.
15. Kirchhoff K, Machicao F, Haupt A et al. Polymorphisms in the TCF7L2,
CDKAL1 and SLC30A8 genes are associated with impaired proinsulin
conversion. Int J Obes Relat Metab Disord 2008; 51: 597–601.
16. Fritsche A, Madaus A, Stefan N et al. Relationships among age, proinsulin
conversion, and beta-cell function in nondiabetic humans. Diabetes 2002;
51(Suppl. 1): S234–S239.
17. Haffner SM, D’Agostino R, Mykkanen L et al. Proinsulin and insulin
concentrations in relation to carotid wall thickness: Insulin Resistance
Atherosclerosis Study. Stroke 1998; 29: 1498–1503.
18. Jia EZ, Yang ZJ, Zhu TB et al. Proinsulin is an independent predictor of
the angiographical characteristics of coronary atherosclerosis. Cardiology
2007; 110: 106–111.
19. Yudkin JS, May M, Elwood P, Yarnell JW, Greenwood R, Davey SG. Con-
centrations of proinsulin like molecules predict coronary heart disease
risk independently of insulin: prospective data from the Caerphilly Study.
Diabetologia 2002; 45: 327–336.
20. Lindahl B, Dinesen B, Eliasson M, Roder M, Hallmans G, Stegmayr B. High
proinsulin levels precede first-ever stroke in a nondiabetic population.
Stroke 2000; 31: 2936–2941.
21. Galloway JA, Hooper SA, Spradlin CT et al. Biosynthetic human proinsulin.
Review of chemistry, in vitro and in vivo receptor binding, animal and
human pharmacology studies, and clinical trial experience. Diabetes Care
1992; 15: 666–692.
22. Lepore M, Pampanelli S, Fanelli C et al. Pharmacokinetics and pharmaco-
dynamics of subcutaneous injection of long-acting human insulin analog
glargine, NPH insulin, and ultralente human insulin and continuous sub-
cutaneous infusion of insulin lispro. Diabetes 2000; 49: 2142–2148.
23. Pfutzner A, Pfutzner AH, Larbig M, Forst T. Role of intact proinsulin in
diagnosis and treatment of type 2 diabetes mellitus. Diabetes Technol
Ther 2004; 6: 405–412.
24. Pfutzner A, Kann PH, Pfutzner AH et al. Intact and total proinsulin: new
aspects for diagnosis and treatment of type 2 diabetes mellitus and insulin
resistance. Clin Lab 2004; 50: 567–573.
25. Haffner SM, Bowsher RR, Mykkanen L et al. Proinsulin and specific insulin
concentration in high- and low-risk populations for NIDDM. Diabetes 1994;
Volume 12 No. 5 May 2010 doi:10.1111/j.1463-1326.2010.01209.x 441