This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ciric, J. Articles by Drezgic, M. Search for Related Content PubMed PubMed Citation Articles by Ciric, J. Articles by Drezgic, M.

Effect of Weight Loss on the Pulsatile Insulin Secretion

Milo arkovi, Jasmina iri, Zorana Penezi, Boo Trbojevi and Milka Drezgi

Institute of Endocrinology, School of Medicine, University of Belgrade, 11000 Belgrade, Yugoslavia

Address correspondence and requests for reprints to: Milo arkovi, Institute of Endocrinology, Dr Subotia 13, 11000 Belgrade, Yugoslavia. E-mail: mzarkov{at}eunet.yu

	Abstract

Top Abstract Introduction Subjects and Methods Results Discussion References

 
The aim of the study was to assess whether pulsatile insulin secretion is variable in the same individual and is related to changes in insulin sensitivity. Insulin sensitivity and pulsatility were measured before and after weight reduction in nine obese subjects. A pulsatility analysis was done using the PulsDetekt program. Blood was sampled every 2 min over a period of 90 min. The secretion randomness was quantified using approximate entropy (ApEn), and ApEn normalized by SD of the insulin time series (nApEn). Lower values indicate more regular secretion. Insulin sensitivity was measured using the homeostasis model assessment. Data are presented as median, minimum-maximum. After weight loss insulin sensitivity was increased (12.16, 7.60–76.70 vs. 38.96, 19.88–74.96%), the number of insulin pulses was reduced (11, 8–16 vs. 9, 6–12), and they were more regular (ApEn, 0.92, 0.53–133 vs. 0.69, 0,40–1.27; nApEn, 1.07, 0.74–1.33 vs. 0.97, 0.54–1.42). Before and after the weight loss there was a correlation between ApEn and nApEn and insulin sensitivity. Therefore, insulin secretion regularity is variable in the same individual and is related to insulin sensitivity.

	Introduction

Top Abstract Introduction Subjects and Methods Results Discussion References

 
REGULAR OSCILLATIONS are an inherent feature of the insulin secretion (1, 2). A decrease in the regularity of blood insulin fluctuations has been reported both in noninsulin-dependent diabetes mellitus patients and in obese subjects (3, 4, 5, 6, 7). It was also shown that the insulin sensitivity is correlated with the regularity of insulin oscillations in peripheral circulation, in such a way that less regular insulin oscillations are found in subjects with decreased insulin sensitivity (6). Reduced insulin sensitivity has also been associated with a higher number of detected insulin pulses and with an increase in the plasma proinsulin concentration (8, 9, 10). Troglitazone, an insulin sensitizing agent, was used to assess the influence of insulin sensitivity on insulin secretion in the obese subjects with impaired glucose tolerance. An increase in insulin sensitivity was concurrent with an improvement in the ability of the ß-cell to entrain to an exogenous oscillatory glucose infusion (11). Weight loss also improved ß-cell responsiveness to glucose and was associated with a reduction in proinsulin immunoreactivity (12). Accordingly, impairment of pulsatile insulin secretion would be present in conditions with reduced insulin sensitivity, and changes in the insulin sensitivity would influence the pulsatile secretion of insulin. So, the characteristics of blood insulin oscillations are not invariable in the same individual, but are related to insulin sensitivity, and they might represent the marker of the ß-cell strain. Therefore, the aim of the study was to assess whether changes in insulin sensitivity induced by weight loss would influence the characteristics of the blood insulin fluctuations, under basal, nonstimulated conditions. To assess the influence of the insulin sensitivity changes on pulsatile insulin secretion, obese subject with normal glucose tolerance were tested, before and after weight reduction. Therefore, we tested only the influence of the changes in insulin sensitivity and not the presence of diabetes or impaired glucose tolerance. The weight reduction program was conducted under close medical supervision, and all subjects consumed the same defined diet. The weight reduction period was followed by a 7-day refeeding period to obtain similar metabolic conditions during sampling for the pulsatility determination. Characteristic parameters of the insulin oscillations in the peripheral blood were determined before and after the weight reduction, using the deconvolution-based PulsDetekt program.¹

	Subjects and Methods

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Subjects

Studies were performed in nine obese subjects before and after weight reduction. Inclusion criteria were: 1) body mass index greater than 33 kg/m²; 2) normal renal and liver function tests; 3) no intercurrent disease in the last 3 weeks; 4) no signs of endocrine dysfunction; 5) normal oral glucose tolerance test; and 6) not taking any drugs. The characteristics of the study group are detailed in Table 1. A local ethical committee approved all the studies. All subjects gave informed consent.

The weight reduction diet was conducted on an in-patient basis, as described previously (13). In short, during the diet period subjects consumed 1675 nonprotein kJ/day (400 kCal), supplemented with protein to 1 g/kg ideal body weight and vitamins and minerals as necessary to make up the recommended daily allowances. The duration of the weight reduction period was 3 weeks, and there was a 1-week refeeding period. During the refeeding period the caloric intake was gradually increased. Caloric intake was increased by 837 kJ (200 kCal) on the first day, 1255 kJ (300 kCal) on the second day, 2090 kJ (500 kCal) on the third day, and during the next 4 days in equal amounts up to calculated energy requirements.

Before and after the weight reduction, insulin sensitivity and pulsatility assessments were made. Over the 48 h before the study the subjects consumed a standardized isocaloric meal composed of 55% carbohydrate, 30% fat, and 15% protein. Energy requirements were calculated using Harris-Benedict equation increased for 30% (activity factor). After a 12-h fast, blood was sampled for insulin determination. An indwelling venous catheter was placed in the antecubital vein. Sampling started at 0800 h and lasted for 90 min at 2-min intervals. Blood was collected using a plastic syringe and transferred to plain glass tubes. Heparin was not used. Before each blood sample collection 0.2 mL (catheter volume) blood was taken with another syringe and discarded. The glucose samples were measured immediately after sampling, whereas insulin samples were frozen at -20 C until assayed.

Pulse analysis

Pulse analysis was done using the PulsDetekt program that is based on the deconvolution of secretory and metabolic events (6).

The program validation was done both on the simulated series and in vivo. To validate the program, 135 synthetic time series were generated by computer simulation with the following characteristics: 1) increase in hormone concentration, followed by monoexponential decay; 2) sampling length, 90 time units; 3) intersample interval, 0.1 time unit; 4) half-life of monoexponential decay, 8 time units; 5) pulse amplitude of 10, 20, and 40 units with a coefficient of variation of 20%; 6) interpulse interval of 6, 10, and 15 time units with a coefficient of variation of 20%; and 7) normally distributed noise, 5, 10, and 20% of series value at a given point in time. Each series was resampled to an intersample interval of 1, 2, and 3 time units, thus generating three new series. The PulsDetekt program correctly predicted synthetic series parameters (half-life, number of pulses, interpulse interval, and pulse amplitude). The program also proved to be robust to noise increase, but a reduction in the sampling rate from 2–3 units led to a significant increase in the false negative rate (false positive rate, 0.00% at noise level 5% and sample interval 1 unit to 0.49% at noise level 20% and sample interval 3 units; false negative rate, 0.00% at noise level 5% and sample interval 1 unit to 13.58% at noise level 20% and sample interval 3 units). For the simulated series set with an amplitude of 10 units, sampled at 2 time units at 10%, noise coefficient of variation was 1.9% for half-life, 2.0% for amplitude, and 0% for number of pulses and interpulse interval. The same synthetic series were also used to calculate approximate entropy (ApEn) coefficients of variation.

Pulse detection was also validated by the assessment of insulin concentration profiles during suppression of the endogenous insulin secretion. To simulate pulsatile secretion insulin was given in iv boluses. Five healthy subjects were included in the study. Suppression of endogenous insulin secretion was achieved using octreotide infusion (0.5 µg/min), preceded by 25 µg bolus. Blood was sampled for 90 min with an intersample interval of 2 min. The octreotide bolus was given at time 0, and the continuous infusion was given during the whole sampling period. At 30, 40, 50, 60, 70, and 80 min 72 mU insulin (Actrapid HM; Novo Nordisk, Bagsvaerd, Denmark) boluses were given into a cubital vein, contralateral to the sampling one. The insulin bolus was calculated to give a pulse amplitude of about 10 mU/L. Detected insulin pulses were assessed for temporal correlation with the pulses of exogenous insulin. From 30 min of sampling only exogenous pulses were detected. All but one pulse of exogenous insulin were detected (false negative, one (3.3%); false positive, zero). On visual inspection at the time point where the undetected pulse was expected only a small increase in the insulin concentration was found, which was not significant enough to be detected as a pulse. To assess reproducibility of the pulse detection parameters, coefficients of variation of the data obtained by the analysis of exogenous insulin were also calculated, as the same study was repeated in five different subjects (number of pulses, 7.7%; interpulse interval, 13.8%; amplitude, 34.8%; amplitude expressed as the percentage of maximal amplitude, 11.2%). The same series were used to calculate ApEn coefficients of variation.

Assays

Glucose was determined using a Beckman Coulter, Inc. (Fullerton, CA) Glucose Analyzer. The precision of glucose determination was 0.17 mmol/L. Insulin was determined using a RIA (INEP, Zemun, Yugoslavia). All samples from one subject were processed in the same batch. The intra-assay coefficient of variation was 5.8%. The minimal detectable concentration of insulin was 1 mU/L.

Data analysis

ApEn.ApEn analysis was used to quantify the degree of randomness of the insulin secretion (14). ApEn determines the likelihood that a certain secretory pattern will be the same throughout the sampling period, although the pulses contained in that pattern can be irregularly spaced. ApEn is a relative measure of system randomness, and a higher approximate entropy value represents higher system randomness. ApEn was computed according to the program described by Pincus et al. (15). We calculated two sets of ApEn values: 1) m = 1 and r = 1.5 mU/L; and 2) m = 1 and r = 20% of SD of the individual subject insulin time series. In choosing the r input parameter in ApEn as a fixed percentage of the SD of each data set, we normalize each epoch by its overall SD. By calculating ApEn and normalized ApEn (nApEn), we quantify amplitude (ApEn) vs. frequency (nApEn) time series irregularities from a separate point of view (16).

The ApEn analysis was done on the secretory data obtained by the deconvolution, because these data are considered to be noise free and they display no trend (the baseline of the secretory data is 0 if there is no significant tonic secretion).

Coefficients of variation for the synthetic series were 5.4% for ApEn and 7.0% for nApEn, and for the in vivo validation coefficients of variation were 19.5% for ApEn and 17.1% for nApEn.

Insulin sensitivity quantification.The homeostasis model assessment (HOMA) was used to quantify insulin sensitivity. HOMA is a mathematical model of insulin/glucose interactions that estimates a set of insulin sensitivity and ß-cell function that is expected to give the fasting glucose and insulin concentrations observed in the subject. Results are expressed as a percentage of the values found in a reference population (young, fit subjects with an ideal body weight) (17, 18). We used HOMA to obtain the insulin sensitivity from the same set of data that we used for the pulsatility analysis.

We used HOMA to obtain the insulin sensitivity from the same set of data that we used for the pulsatility analysis. The average of all 46 insulin and glucose samples was used to calculate insulin sensitivity. The coefficient of variation for the raw data (HOMA insulin sensitivity calculated at each time point, 46 HOMA measurements per subject) was 28.7 ± 3.7% and for the smoothed data [7 points, 14-min moving average, as suggested by Matthews et al. (17)] 15.6 ± 1.7%. To assess if there are any systematic oscillations in HOMA insulin sensitivity, the sampling period was divided in four or nine epochs that were compared using split-plot analysis of variance with subjects as random factor. No significant variations in insulin sensitivity were found between the epochs.

Statistical analysis.Statistical analyses were done using the Wilcoxon signed ranks test and the Spearman correlation. The results are expressed as median, minimum-maximum. P values less than 0.05 were considered to be statistically significant.

	Results

Top Abstract Introduction Subjects and Methods Results Discussion References

 
The weight reduction diet caused significant loss of weight in all of the subjects. The plasma insulin and glucose concentrations were decreased, and the insulin sensitivity was increased (Table 1).

Weight loss caused significant alterations in all parameters of pulsatile secretion of insulin, except the insulin half-life. The number of pulses and pulse amplitudes were reduced, and the interpulse interval was prolonged. After the weight reduction, the insulin secretion was more regular, and the values for ApEn and nApEn decreased (Fig. 1 and Table 2).

View larger version (67K): [in this window] [in a new window]   Figure 1. Insulin pulsatility profiles. All subjects are presented. Detected pulses are marked with vertical lines. The top line represents measured data, and the bottom line represents calculated data (deconvolution).

View larger version (21K): [in this window] [in a new window]   Figure 2. Relation between the change in the ApEn, the change in the nApEn, and the change in the HOMA insulin sensitivity.

There was a correlation between mean insulin concentration and the ApEn both before and after the weight reduction (before: ApEn, r = 0.750, P = 0.020; after: ApEn, r = 0.717, P = 0.030). There was a correlation between mean insulin concentration and the nApEn after but not before the weight reduction (before: nApEn, r = 0.650, P = 0.058; after: nApEn, r = 0.717, P = 0.030). Differences between ApEn and nApEn values before and after the weight loss did not correlate with the change in insulin concentration (ApEn, r = 0.100, P = 0.798; nApEn, r = -0.283, P = 0.460).

A change in the mean insulin concentration in the samples obtained for the pulsatility analysis did not correlate with a change in the number of detected insulin pulses or the interpulse interval (number of pulses, r = -0.560, P = 0.117; interpulse interval, r = 0.633, P = 0.067).

	Discussion

Top Abstract Introduction Subjects and Methods Results Discussion References

 
In this study, we wanted to assess whether the pulsatile insulin secretion is variable in the same individual and whether it is possible to induce changes in it by simple weight loss. We also wanted to evaluate the relation between variations in blood insulin oscillations and changes in insulin sensitivity. A deconvolution-based pulse detection program PulsDetekt was used to obtain a series of noise-free insulin pulses. The secretion randomness of those series was quantified using the ApEn analysis.

Weight reduction caused a significant increase in insulin sensitivity. The insulin half-life was not significantly altered and was comparable with previously published data (19, 20, 21).

The frequency of insulin pulses found in the present study was higher, and the interpulse interval was shorter (before: 11, 8–16 pulses/90 min; 7.60, 5.47–12.29 min; after: 9, 6–12 pulses/90 min; 9.25, 6.80–15.20 min) than some previous studies (interpulse interval 11–13 min) (4, 3, 22) and similar to more recent studies (number of insulin pulses 11.8 ± 0.9 per 90 minutes; interpulse interval 4.1 to 6.5 min) (8, 23). Probably the principal cause of this difference is the pulse detection methodology. Earlier studies used autocorrelation or spectral analysis, whereas more recent studies, including ours, used heuristic methods of pulse detection. Autocorrelation and spectral analysis are not suited for hormone series analysis due to the assumption of absolutely regular pulses and the lack of resolution and precision for short time series. The spectral density function of a noise-free series sampled at 1 min for 150 min would have a resolution of about 5 min at 95% confidence level (24). Heuristic methods make no assumptions on pulse regularity and are not dependent on sampling length.

After weight loss, the insulin pulse frequency was reduced, and the interpulse interval was prolonged. The decrease in the insulin pulse frequency at low insulin concentrations might be inherent to the analyzed system (insulin in the blood), because the lower limit of the assay sensitivity acts as a filter for low amplitude pulses. The change in the detected number of insulin pulses might also be caused by the properties of pulse detection, because the pulse detection is dependent on the pulse amplitude (because of the algorithm assumptions and the assay characteristics). Due to these characteristics of the system and the pulse detection process, the number of detected pulses would rise as the insulin sensitivity declines, due to the insulin concentration increase. In that case, we would expect that the change in the number of detected pulses would correlate with the change in the mean insulin concentration. In our study, neither changes in the number of detected insulin pulses nor the interpulse interval correlated with the change in the mean insulin. Therefore, the inherent properties of the pulse detection were not the cause of the insulin pulse frequency decrease.

The regularity of insulin secretion, quantified using ApEn and nApEn, was increased after weight loss. Insulin sensitivity correlated to ApEn and nApEn, both before and after weight reduction. Changes in measures of insulin secretion randomness (ApEn and nApEn) correlated to the change in insulin sensitivity, but not with the change in weight.

In conclusion, the present study indicates that the randomness of the insulin secretion is related to insulin sensitivity and that insulin secretion regularity is variable in the same subject. There are two possible explanations for this finding. One is that the change in the insulin secretion pattern induced by a change in insulin sensitivity is an artefact of the pulse detection process, due to reduced insulin concentration and consequent failure to detect low amplitude pulses. Another explanation is that a less regular and higher number of insulin pulses is the consequence of a reduction in insulin sensitivity. Therefore, because less regular insulin secretion is associated with insulin resistance, it can be a marker of an over worked ß-cell.

	Footnotes

 
¹ The PulsDetekt program can be downloaded at http://solair.eunet.yu/mzarkov/.

Received March 1, 2000.

Revised May 24, 2000.

Accepted July 12, 2000.

	References

Top Abstract Introduction Subjects and Methods Results Discussion References

 

1. DeFronzo RA, Bonadonna RC, Ferrannini E. 1992 Pathogenesis of NIDDM. A balanced overview. Diabetes Care. 15:318–368.[Abstract]
2. Stagner JI, Samols E, Weir GC. 1980 Sustained oscillations of insulin, glucagon, and somatostatin from the isolated canine pancreas during exposure to a constant glucose concentration. J Clin Invest. 65:939–942.
3. Lang DA, Matthews DR, Burnet M, Turner RC. 1981 Brief, irregular oscillations of basal plasma insulin and glucose concentrations in diabetic man. Diabetes. 30:435–439.[Abstract]
4. O’Rahilly S, Turner RC, Matthews DR. 1988 Impaired pulsatile secretion of insulin in relatives of patients with non-insulin dependent diabetes. N Engl J Med. 348:1225–1230.
5. Schmitz O, Porksen N, Nyholm B, et al. 1997 Disorderly and nonstationary insulin secretion in relatives of patients with NIDDM. Am J Physiol. 272:E218–E226.
6. Zarkovic M, Ciric J, Stojanovic M, et al. 1999 Effect of insulin sensitivity on pulsatile insulin secretion. Eur J Endocrinol. 141:494–501.[Abstract]
7. Sonnenberg GE, Hoffman RG, Mueller RA, Kissebah AH. 1994 Splanchnic insulin dynamics and secretion pulsatilities in abdominal obesity. Diabetes. 43:468–477.[Abstract]
8. Hunter SJ, Atkinson AB, Ennis CN, Sheridan B, Bell PM. 1996 Association between insulin secretory pulse frequency and peripheral insulin action in NIDDM and normal subjects. Diabetes. 45:683–686.[Abstract]
9. Peiris AN, Stagner JI, Vogel RL, Nakagawa A, Samols E. 1992 Body fat distribution and peripheral insulin sensitivity in healthy men: role of insulin pulsatility. J Clin Endocrinol Metab. 75:290–294.[Abstract]
10. Reaven GM, Chen YD, Hollenbeck CB, Sheu WH, Ostrega D, Polonsky KS. 1993 Plasma insulin, C-peptide, and proinsulin concentrations in obese and nonobese individuals with varying degrees of glucose tolerance. J Clin Endocrinol Metab. 76:44–48.[Abstract]
11. Cavaghan MK, Ehrmann DA, Byrne MM, Polonsky KS. 1997 Treatment with the oral antidiabetic agent troglitazone improves ß cell responses to glucose in subjects with impaired glucose tolerance. J Clin Invest. 100:530–537.[Medline]
12. Polonsky KS, Gumbiner B, Ostrega D, Griver K, Tager H, Henry RR. 1994 Alterations in immunoreactive proinsulin and insulin clearance induced by weight loss in NIDDM. Diabetes. 43:871–877.[Abstract]
13. Drezgic M, Penezic Z, Zarkovic M, et al. 1996 Influence of three-week fasting on gonadotrophin pulsatility in obese menstruating women. Int J Obes. 20:608–612.
14. Pincus SM, Keefee DL. 1992 Quantification of hormone pulsatility via an approximate entropy algorithm. Am J Physiol. 262:E741–E754.
15. Pincus SM, Gladstone IM, Ehrenkranz RA. 1991 A regularity statistics for medical data analysis. J Clin Monit. 7:335–345.[CrossRef][Medline]
16. Siragy HM, Vieweg WVR, Pincus S, Veldhuis JD. 1995 Increased disorderliness and amplified basal and pulsatile aldosterone secretion in patients with primary aldosteronism. J Clin Endocrinol Metab. 80:28–33.[Abstract]
17. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. 1985 Homeostasis model assessment: insulin resistance and ß-cell function from fasting plasma glucose and insulin concentration in man. Diabetologia. 28:412–419.[CrossRef][Medline]
18. Turner RC, Rudenski AS, Holman RR, Matthews DR, O’Rahilly SP. 1987 Quantitative modelling of endocrine diseases as exemplified by diabetes. Clin Endocrinol (Oxf.). 26:107–116.
19. Dreyer M, Matthaei S, Kuhnau J, Rudiger HW. 1986 Prolonged plasma half-life of insulin in patients with a genetic defect of high affinity binding sites. Horm Metab Res. 18:247–249.[Medline]
20. Kuzuya T, Matsuda A. 1976 Disappearance rate of endogenous human C-peptide from blood. Diabetologia. 12:519–521.[CrossRef][Medline]
21. Lanng S, Thorsteinsson B, Roder ME, Nerup J, Koch C. 1994 Insulin sensitivity and insulin clearance in cystic fibrosis patients with normal and diabetic glucose tolerance. Clin Endocrinol (Oxf.) 41:217–223.
22. Lang DA, Matthews DR, Peto J, Turner RC. 1979 Cyclic oscillations of basal plasma glucose and insulin concentrations in human beings. N Engl J Med. 301:1023–1027.[Abstract]
23. Storch MJ, Rossle M, Kerp L. 1993 Pulsatile Insulinsekretion in die Pfortader bei Leberzirrhose. Dtsch Med Wochenschr. 118:134–138.[Medline]
24. Priestley MB. 1987 Spectral analysis in practice. In: Priestley MB, ed. Spectral analysis and time series, ed 1. London: Academic Press; 502–612.

This article has been cited by other articles:

				P. Jahanshahi, R. Wu, J. D. Carter, and C. S. Nunemaker Evidence of Diminished Glucose Stimulation and Endoplasmic Reticulum Function in Nonoscillatory Pancreatic Islets Endocrinology, February 1, 2009; 150(2): 607 - 615. [Abstract] [Full Text] [PDF]

				C. S. Nunemaker, M. Zhang, D. H. Wasserman, O. P. McGuinness, A. C. Powers, R. Bertram, A. Sherman, and L. S. Satin Individual Mice Can Be Distinguished by the Period of Their Islet Calcium Oscillations: Is There an Intrinsic Islet Period That Is Imprinted In Vivo? Diabetes, December 1, 2005; 54(12): 3517 - 3522. [Abstract] [Full Text] [PDF]

				M. O. Larsen, C. B. Juhl, N. Porksen, C. F. Gotfredsen, R. D. Carr, U. Ribel, M. Wilken, and B. Rolin {beta}-Cell function and islet morphology in normal, obese, and obese {beta}-cell mass-reduced Gottingen minipigs Am J Physiol Endocrinol Metab, February 1, 2005; 288(2): E412 - E421. [Abstract] [Full Text] [PDF]

				R. Feneberg, M. Sparber, J. D. Veldhuis, O. Mehls, E. Ritz, and F. Schaefer Altered Temporal Organization of Plasma Insulin Oscillations in Chronic Renal Failure J. Clin. Endocrinol. Metab., May 1, 2002; 87(5): 1965 - 1973. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ciric, J. Articles by Drezgic, M. Search for Related Content PubMed PubMed Citation Articles by Ciric, J. Articles by Drezgic, M.