Serum glucose and insulin response to mango and papaya
in type 2 diabetic subjects
Kaniz Fatema, MSc, Liaquat Ali, PhD*, Mohammad H Rahman, MSc,
Shahana Parvin, MBBS, MPhil, Zahid Hassan, MBBS
Biomedical Research Group, Bangladesh Institute of Research and Rehabilitation in Diabetes, Endocrine and
Metabolic Disorders (BIRDEM), Dhaka—1000, Bangladesh
Received 30 July 2002; received in revised form 16 September 2002; accepted 20 September 2002
To rank Bangladeshi Mango and Papaya in terms of their Glycemic Index (GI) and Insulinemic
Index (II), which are useful measures of glucose and insulin responses to a dietary component, thirteen
type 2 diabetic subjects consumed, under a cross-over design, equi-carbohydrate amounts of mango
(250 g), papaya (602 g) and white bread (the reference food, 63 g). Blood sample was drawn 5 times
between 0h and 3h. Serum C-peptide was measured to evaluate Insulinemic status. Mango and Papaya
showed higher serum glucose responses compared to that of bread. The similar glycemic responses of
Papaya and Mango were reﬂected in their GI values. Papaya showed higher insulin response compared
to both Mango and Bread (p⬍0.001). Papaya also showed signiﬁcantly higher C-peptide–glucose
ratio in comparison to that of Mango and Bread. The data suggest that equi-carbohydrate amount of
Papaya and Mango produce higher glycemic response as compared to bread, but the two fruits are
comparable regarding this property. The higher insulin response of papaya needs to be considered in
case of therapeutic management of diabetic patients and in assessing the risk of atherogenesis due to
hyperinsulinemia. © 2003 Elsevier Science Inc. All rights reserved.
Keywords: Diabetes Mellitus; Glycemic Index; Insulinemic Index; Mango; Papaya; Dietary management
A debate is going on for centuries regarding the amount of carbohydrate that individuals
with diabetes should consume. The latest consensus is towards liberalization of carbohydrate
* Corresponding author. Tel.: ⫹880-2-8613700, 8617130; fax: ⫹880-2-8613004; 8617130.
E-mail address: email@example.com (L. Ali).
Nutrition Research 23 (2003) 9–14
0271-5317/03/$ – see front matter © 2003 Elsevier Science Inc. All rights reserved.
intake with 50% to 60% of the total calories in the diet for this source . Special caution,
however, is required to deﬁne the nature of carbohydrate due to the central importance of
maintaining normoglycemia and due to the need for adjustment of drug regimens in
consideration of a particular diet.
Carbohydrate content varies from one food to another. So, a method of ranking the
carbohydrate containing food has been proposed on the basis of the incremental blood
glucose responses, glycemic index (GI), produced for a given amount of food . Glucose,
in turn, triggers insulin secretion from the pancreatic beta cells. So, in addition to GI, the
amount of insulin secreted in response to food is also important in the management of
diabetes and its vascular complications. Since insulin is now known to be atherogenic, a low
GI at the expense of hyperinsulinemia may not always be useful. Thus ranking of food based
on their insulinemic response, along with the glycemic response, is necessary. Measurement
of Insulinemic index (II) and C-peptide-glucose ratio may be two indicators of this response.
Due to the problem in differentiating endogenous and exogenous insulin, measurement of
C-peptide is preferred as a measure of serum insulin. Thus C-peptide Index and C-peptide-
glucose ratio are equated with insulin Index and insulin-glucose ratio.
Fruits form a major part of daily diet in both in healthy and diseased people and those have
a variety of pleasant and attractive ﬂavors. It has also laxative and base forming properties
inspite of equivalent different amount of carbohydrate . To prepare a diet plan for diabetic
and cardiovascular patients in the local perspective we need to grade these foods according
to their glycemic and insulinemic responses. However, no attempt has so far been made in
this direction for Bangladeshi foods. Very recently our group has undertaken a long-term
program to assess the GI and II of common dietary components of Bangladesh. The present
work is the ﬁrst one to investigate the fruits in this country under the series. Langra Mango
(Mangiﬁra Indica L) and Papaya (Carica Papaya L), being two of the most popular and
available fruits in the country, have been chosen for this study.
2. Methods and materials
Thirteen (13) Type 2 diabetic subjects (male 7 and female 6) were included in this study.
The participants were selected from the BIRDEM Outpatient Department. Diabetes was
diagnosed and classiﬁed by the WHO criteria . Patients with acute or chronic complica-
tions of diabetes mellitus and those using insulin, oral contraceptives or steroid were
excluded. Pregnancy was also an exclusion criteria. All participants gave their written
consent after being fully informed about the nature of the study.
2.2. Study design
Detailed family history was taken, and clinical and biochemical ﬁndings were recorded in
a predesigned case record form. The body weight, in kilogram, was measured by using
appropriate scales (Detect-Medic, Detect Scales INC, USA). Body Mass Index and waist-hip
10 K. Fatema et al. / Nutrition Research 23 (2003) 9–14
ratio of the subjects were calculated by appropriate formula. The subjects were requested to
fast overnight (8-10 hr) and suggested not to take medicine. An intravenous cannula was
inserted into a superﬁcial vein in the forearm, fasting blood samples were collected and meal
was served. The meal was consumed within 15 min along with 200 ml water. Blood samples
were then collected at 30, 60, 120 and 180 minutes. Separated serum was preserved at
⫺40°C for future biochemical analysis.
2.3. Test meals
Langra Mango (Mangiﬁra Indica L) (250 g) and Papaya (Carica Papaya L) (602 g),
having 50 g available carbohydrate in each fruit (calculated from a food table) , were
given to the subjects. White bread (63 g calculated by chemical analysis) was given as a
2.4. Laboratory method
Serum glucose was measured by glucose-oxidase method (Sera Pak, Bayer, USA),
C-peptide was measured by an ELISA technique (DRG International Inc, Germany) and
Fructosamine was measured by colorimetric method (Boaheringer-Mannheim, Germany).
2.5. Statistical analysis
All analyses were done using the SPSS Windows Package. Experimental values were
expressed as mean ⫾SD (standard deviation) or median (range) as appropriate. To compare
results between two groups students t-test were performed where appropriate. ANOVA
(Duncan) was performed to test the difference between means and Nonparametric (Mann-
Whitney) test was done to calculate the difference of different test meals and Bonferroni ‘t’
test for ratio of different parameters. Integrated serum glucose values over the 3 hour time
period was calculated by area under the curve using a mathematical model .
The study was conducted on non-obese subjects (BMI 24.57 ⫾2.53 and waist-hip ratio
1.07 ⫾0.01, M ⫾SD) with mild to moderate hyperglycemia (serum fructosamine 391 ⫾93,
mol/l, M ⫾SD) (Table I). The absolute value of serum glucose in fasting as well as in
postprandial states showed no signiﬁcant difference among the Bread, Papaya and Mango
groups. The integrated serum value during the 3-hour time period also showed no difference
(Table II). Glycemic Index (GI) of Papaya (124 ⫾14; M ⫾SD) was close to that of Mango
(122 ⫾10, Table II). Fasting serum C-peptide (ng/ml, M ⫾SD) levels in cases of Bread,
Papaya and Mango did not differ among themselves. The respective 3-hour postprandial
serum C-peptide levels after feeding Bread, Papaya and Mango were as follows: 3.99 ⫾1.39,
5.54 ⫾1.67 and 4.15 ⫾1.70 respectively (p ⬍0.01 in Papaya vs Bread; and p ⬍0.003 in
Papaya vs Mango, Table III). Absolute increment of C-peptide values [median (range),
11K. Fatema et al. / Nutrition Research 23 (2003) 9–14
ng/ml] was calculated in response to three test meals. Secretory response to Mango 1.05
(0.21-3.95) was signiﬁcantly low compared to both Bread 2.73 (0.65-5.0) and Papaya 2.72
(0.65-5.0) (Table III). Fasting C-peptide-Glucose ratio of Bread (0.34 ⫾0.16) showed lower
value compared to Papaya (0.44 ⫾0.21) but the difference did not reach level of signiﬁcance
(p ⫽0.058). Postprandial C-peptide-glucose ratio of Papaya was signiﬁcantly higher com-
pared to Bread (p ⫽0.01) and Mango (p ⫽0.003). C-peptide Index (CI) of Papaya (129 ⫾
33) did not show statistically signiﬁcant difference compared to mango (112 ⫾45, t ⫽
⫺1.746, p ⫽0.106).
Fruits can be excellent components of the diet in diabetic patients. This is for their high
nutritional values (particularly in respect of micronutrients) and for their palatability which
increases patient compliance. However, starchy and juicy fruits are generally avoided by
diabetic patients and those are advised by the physicians and dieticians with substantial
restriction. To rationalize the advice of fruits in diabetic (as well as cardiovascular) patients,
it is important to know their chemical composition and above all their biological responses.
The substitution of calories and other nutrients may then be done on the basis of patient
choice, socioeconomic capability and availability of fruits.
Glycemic Index (GI) is useful in monitoring the biological response of a food in relation
to glycemic status. The present study evaluates the GI of Mango and Papaya in type 2
Clinical and sociodemographic characteristics of the study subjects
Age (years, mean ⫾SD) 39.46 ⫾6.21
BMI (mean ⫾SD) 24.57 ⫾2.53
Waist-hip ratio (mean ⫾SD) 1.07 ⫾0.01
Male : Female 27:23
Rural : Urban 4:9
Annual income (median-range) in US Dollars 2316 (1050–4210)
Serum fructosamine (mean ⫾SD,
mol/l) 391 ⫾93
Effect on glycemic status of the study subjects (n ⫽13) after feeding the different test meals
foods Serum glucose (mmol/l) AUC (mmol/
l/ 180 min) Glycemic
0 min 30 min 60 min 120 min 180 min
Bread 8.7 ⫾3.5 9.8 ⫾2.8 10.9 ⫾2.5 9.8 ⫾2.5 8.3 ⫾2.7 1745 ⫾450
Papaya 9.2 ⫾3.4 13.0 ⫾3.1 14.3 ⫾3.0 11.6 ⫾3.5 9.5 ⫾3.4 2138 ⫾639 124 ⫾14
Mango 9.7 ⫾3.2 13.4 ⫾3.3 13.7 ⫾3.5 11.2 ⫾4.1 9.8 ⫾3.5 2153 ⫾566 122 ⫾10
Results expressed as mean ⫾SD; ANOVA (Duncan) test was performed to test the difference between means.
p⬍0.05 was taken as the level of signiﬁcance; n, Number of subjects; AUC, Area under the curve.
12 K. Fatema et al. / Nutrition Research 23 (2003) 9–14
diabetic subjects. In line with the popular impression, the results suggest that, Mango and
Papaya produces higher glycemic response compared to Bread (the reference diet) (Table II
& III). These properties of Papaya and Mango need to be considered when dietary advice is
given. It is interesting to note that Mango and Papaya have equivalent GI. It means that, on
calculating the exact calorie requirement, the dietician may freely criss-cross Papaya and
Mango so far as the blood glucose level is concerned. This fact may provide substantial
ﬂexibility to the clinicians and dieticians regarding the advice on foods and it may improve
the satisfaction of the patients and increase their compliance.
Insulin is the central hormone in maintaining blood glucose homeostasis. In type 2
diabetes mellitus the absolute level of insulin may be low, normal or higher in the blood
although there is always a relative insulin deﬁciency. In spite of the life-saving role of the
hormone the level of higher insulin in blood (hyperinsulinaemia) has been shown to be
associated with increased atherosclerosis leading to cardiovascular disorders . Due to this
atherogenic role of insulin it is desirable to control the blood glucose of patients keeping the
insulin levels as low as possible. In this context, serum C-peptide (a marker of insulin)
measurement in the present study has an important implications. The results show that
although GI is same for Papaya and Mango, the insulin response is substantially higher
[reaching signiﬁcant levels in case of absolute values (Table II) and C-peptide Index (Table
III)] in case of papaya. This insulin secretogogue property may have some advantages in
cases of patients with low insulin levels, but in patients with normal or higher insulin levels
this may not be desirable. In the absence of detailed biochemical investigations for insulin
secretory capacity (or at least for insulin/C-peptide level measurements) it is difﬁcult to
assess the Insulinemic status of a subject. In these circumstances it seems that a cautious
approach is necessary regarding the quantity of papaya advised. The physicians should also
be aware of this property of papaya when an oral drug or insulin is prescribed. If a free choice
is given, mango should be the preferred food for a diabetic patient whose insulin secretory
status is unknown.
The Mango investigated in this study is Langra and the Papaya was only of one variety.
It is now necessary to study other common varieties of mango and papaya. Study should also
be undertaken to explore the mechanism of papaya induced increase in serum insulin levels
of the patients.
Fasting and postprandial serum insulinemic response, insulinemic index and c-peptide : glucose ratio of the
study subjects (n ⫽13) after feeding the different test meals
foods Serum C-peptide (ng/ml) AICP (ng/ml) Insulinemic
Index C-peptide : Glucose ratio
0 min 180 min 0 min 180 min
Bread 2.67 ⫾1.04 3.99 ⫾1.39
—0.34 ⫾0.16 0.44 ⫾0.21
Papaya 2.86 ⫾0.94 5.45 ⫾1.67
129 ⫾33 0.44 ⫾0.21 0.52 ⫾0.25
Mango 2.92 ⫾1.10 4.15 ⫾1.70
112 ⫾45 0.31 ⫾0.14 0.39 ⫾0.21
Results expressed as mean ⫾SD or median (range); Paired students ‘t’, Bonferroni ‘t’and nonparametric
(Mann-Whitney) test were performed where appropriate to calculate statistical difference between groups.
Different superscripts in each column indicate signiﬁcant statistical difference at p ⬍0.05. AICP, Absolute
increment of C-peptide over basal values.
13K. Fatema et al. / Nutrition Research 23 (2003) 9–14
Thanks are due to the participants in the study. We are grateful to Diabetic Association of
Bangladesh (DAB), ENRECA Project of Danish International Development Assistance
(DANIDA) and International Program in the Chemical Sciences (IPICS), Uppsala Univer-
sity, Sweden, for their ﬁnancial support.
As one of the study in a series conducted by his Group LAliclosely supervised and
monitored all aspects of this study from conception of the idea to submission of the paper;
MH Rahman was involved in the biochemical analysis of serum glucose and insulin; S
Parvin took part in project design and data interpretation; Z Hassan was involved in
 Vinik A, Wing RR, Lauterio TJ. Nuritional management of the person with diabetes. In: Porte D Jr, Sherwin
RS, editors. Ellenberg & Rifkin’s Diabetes Mellitus. 5th ed. Stamford Connecticut: Appleton & Large, 1996.
 Wolever TMS, Jenkins DJA, Jenkins AL, Josse KG. The glycemic index: methodology and clinical
implications. Am J Clin Nutr 1991;54:846–54.
 Englyst HN, Kingman SM. Carbohydrate. In: Garrow JS, James WPT, editors. Human Nutrition, and
Dietetics. 9th ed. Edinburgh London: Churchill Livingstone, 1994. p. 38–55.
 Expert Committee on the Diagnosis and Classiﬁcation of Diabetes Mellitus. Report of the expert committee
on the diagnosis and classiﬁcation of diabetes mellitus. Diabetes Care 1997;20:1183–97.
 Desio Khadder Pustiman. In: Sastha Rakhay Pusti. Ali SMK (editor), Bangle Printers, Dhaka. 1995. p. 314.
 Tai MM. A mathematical model for the determination of total area under glucose tolerance, and other
metabolic curve. Diabetes Care 1994;17:152–4.
 Kaplan NM. Hypertension and diabetes. In: Porte D Jr, Sherwin RS, editors. Ellenberg & Rifkin’s Diabetes
Mellitus. 5th ed. Stamford Connecticut: Appleton & Large, 1996. p. 1097–1104.
14 K. Fatema et al. / Nutrition Research 23 (2003) 9–14