| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Special Articles |
Department of Medicine and Pediatrics, Harbor-University of California Los Angeles Medical Center, Torrance, California 90509; and Lilly Laboratory for Clinical Research, Indianapolis, Indiana
Address all correspondence and requests for reprints to: Nancy Berman, Walter Martin Research Center, Harbor-University of California Los Angeles Medical Center, 1124 West Carson Street, Torrance, California 90502. E-mail: Berman{at}harbor6.humc.edu
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A possible explanation for erratic bursts of insulin release in patients with insulinomas lies in the phenomenon of insulin secretory oscillations (7, 8). High frequency regular oscillations of plasma insulin concentrations are a characteristic feature of the postabsorptive state in normal human subjects (8, 9). It is possible that exaggeration of normal physiological oscillations may occur in insulinoma patients and cause transient spontaneous hyperinsulinemia with hypoglycemia in the absence of any obvious secretagogue. This study was designed to test whether patients with insulinomas have plasma insulin oscillations, and whether abnormalities in this system would explain large insulin peaks in the fasting state.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Normal subjects. Nine healthy volunteers participated in this study (seven women, two men). Their mean age was 32.1 ± 2.0 yr, with a range of 25–41 yr. Mean body mass index was 21.5 ± 0.9 kg/m2, with a range of 17–25.7 kg/m2. Subjects were studied in the General Clinical Research Center at Harbor-UCLA Medical Center. None was using medications, and there was no personal or immediate family history of diabetes, hypoglycemia, or glucose intolerance.
Insulinoma subjects. Five patients with insulinomas (three women, two men) also participated. The patients ranged in age from 32–76 yr (mean = 47.6 ± 8.5 yr). Their mean body mass index was 25.8 ± 2.2 kg/m2 (range = 21.9–33.7 kg/m2). All patients were later found to have a surgically removable insulinoma. Control and insulinoma subjects consumed a weight-maintaining diet, including at least 200 g carbohydrate daily, for 3 days before study. Subjects gave informed consent for these studies, which were approved by the institutional review board.
Procedures
Studies were conducted in the morning after a 10-h overnight fast. Arterialized blood samples were drawn from an in-dwelling catheter in a heated hand (10). Blood samples were obtained at 2- or 3-min intervals for 60–100 min, for a total of 30–50 samples. The samples were maintained at 4 C until centrifugation which was performed within 60 min. Plasma was frozen at -70 C before analysis.
Analysis
Laboratory analysis. Plasma was assayed for glucose with an Abbott Biocromatic Analyzer (Abbott Diagnostics, Chicago, IL) using hexokinase methodology (11) and for insulin with a previously described RIA (12). The lower limit of sensitivity of the insulin assay was 7 pmol/L; proinsulin exhibited 60% cross-reactivity. Proinsulin was measured using a highly specific RIA with a sensitivity limit of 5 pmol/L (13) that does not cross-react (<1%) with insulin or C-peptide (14).
Data analysis. Spectral analysis was used to detect regular oscillations in time-series data (15). Briefly, this method models time series as a sum of sinusoids with different frequencies and phases, which are determined from the sampling interval and duration of the study. The amplitude computed at each frequency reflects the contribution of the sinusoid at that frequency. The squared amplitude (power) of each component is normalized to a percent of total power and shown plotted against period. Fisher’s test was used to determine whether the amplitude of the dominant frequency component is statistically different from noise (16). When spectral analysis indicated there was no regular pattern of oscillations, the DETECT Program (17) was used to define specific pulses. Student’s t test for unpaired data was used to evaluate the difference in means between groups. Data are presented as the mean ± SEM, unless stated otherwise.
Correction for proinsulin cross-reactivity. Insulinomas are known to hypersecrete proinsulin (1, 2, 3). Because proinsulin cross-reacts in the insulin assay, the pattern of plasma insulin concentrations in these patients might be influenced by the proinsulin component. To evaluate this possibility, the total insulin measured by insulin assay was corrected for the proinsulin component in each sample as follows: first, proinsulin was measured using a specific assay. The proinsulin component of each total insulin measurement was estimated using the concentration of proinsulin measured in each sample, and the known cross-reactivity of proinsulin in our insulin assay. The proinsulin component thus calculated was then subtracted to provide the estimated true insulin value for each measurement. Finally, the time series and spectral analysis were recomputed for each study using the adjusted true insulin values.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In normal lean subjects, mean plasma glucose and insulin
concentrations were 5.3 ± 0.1 mmol/L and 58 ± 9 pmol/L,
respectively (Table 1
). Plasma insulin
concentrations were stable over the duration of the study. However,
periodic, low-amplitude, high-frequency oscillations of plasma insulin
concentrations were found (Fig. 1a). In
all subjects, a dominant frequency component with a period of between
10 and 17 min was observed on spectral analysis, indicating a high
degree of regularity in the oscillations. In eight of nine subjects,
the amplitude of this component was significantly different (16) from
noise (P < 0.05, Table 1). Mean amplitude was 19
± 3 pmol/L and mean period was 12.5 ± 1 min. A representative
example using the results of subject 8 is demonstrated (Fig. 1a)
showing regular oscillations. Figure 1a shows the time series; Fig. 1b
shows the results of spectral analysis. The single large value at 17
min indicates regular oscillations with that period. Although most
controls in this study were lean and younger, similar results were
found in obese and elderly subjects (our unpublished observations).
Plasma glucose was also stable, but oscillations were infrequently
observed (data not shown).
|
|
The insulinoma patients had lower mean plasma glucose
concentrations after an overnight fast: 3.6 ± 0.3 mmol/L
(P = 0.01 vs. normal subjects; Table 2). Mean plasma insulin was significantly
higher than controls at 150 ± 42 pmol/L (P =
0.01). Overall variability in insulin concentrations was much greater
in the insulinoma patients (Fig. 1c
). Plasma insulin pulses were also
observed but showed a lack of regularity (Fig. 1c). Spectral analysis
did not identify a statistically significant dominant frequency
component in four of five patients. This is evident from the multiple
low amplitude peaks without a dominant frequency, identified on
spectral analysis (e.g. Fig. 1d). Data were
therefore analyzed with the DETECT program; the number of pulses and
interval between pulse maxima varied considerably within and between
individuals (Table 2). Pulses occurred at intervals ranging from 4–54
min (Table 2). The average overall interpulse interval, 19.0 min, was
slightly longer than the range of the period of regular oscillations
found in the normal subjects.
|
, patient 4). Plasma insulin
concentrations doubled from approximately 300 pmol/L at 46 min to 565
pmol/L in a period of a few minutes. This is equivalent to a
postprandial increase in insulin concentrations, yet it clearly
occurred in an apparently unprovoked fashion while the subject was in
the fasting state. The result of this pulse was a decrement in plasma
glucose concentrations, which had stabilized at 2.5 mmol/L in the 20
min before the appearance of this insulin pulse, then dipped to 1.9
mmol/L over the next 20 min (Fig. 2). Similar large insulin pulses were
noted in another two time series. Overall, variability in the amplitude
of insulin pulses is also evident as shown in Table 2. Mean pulse
amplitudes showed considerable intersubject variability (range 19–134
pmol/L). Large intrasubject variation was also observed in some
subjects. The range of pulse amplitudes was wide in patients 3, 4, and
5; in patients 1 and 2 this range was small, reflecting a level of
consistency closer to normal subjects. Also in insulinoma subjects, no
regular oscillations in plasma glucose concentrations were observed,
though levels tended to drift lower as the study progressed (see ranges
of glucose concentrations in Table 2).
|
Mean plasma proinsulin concentrations were elevated in all patients, ranging from 34–164 pmol/L (upper limit of reference range = 20 pmol/L). This represents 12–70% of the mean total insulin concentrations. The proinsulin component of each measurement of total insulin (described in Subjects and Methods) was subtracted, and the resulting time series reevaluated. Mean insulin concentrations and amplitude of oscillations were significantly lower in each case, but there was no change in degree of regularity of oscillations when analyzed by spectral analysis (data not shown).
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Specifically, we tested whether insulinoma patients have insulin oscillations, and whether abnormal pulsatility might explain the observation of sudden, erratic plasma insulin peaks (5, 6). Using high-frequency serial sampling, this study confirmed the presence of consistent insulin pulsatility in these patients, but with patterns that are erratic compared with normal subjects. In normal controls, regular plasma insulin oscillations with a mean period of 12.5 min were observed, similar to the observations of others (7, 8). Spectral analysis evaluates oscillatory data using a sinusoid wave model, thus the finding of a single dominant high-frequency peak in the power spectrum in normal subjects implies a degree of regularity approximating a sine wave statistically (15). The degree of consistency in both frequency and amplitude implied by sine wave-type oscillations, reflects a physiological system that is highly regulated. In contrast, the insulin pulses in the insulinoma patients did not demonstrate a regular pattern in either amplitude or frequency. The amplitude of the pulses in some patients was much larger than in normal subjects and varied within and between subjects. The periodicity observed in insulin oscillations of normal subjects was also lacking. Spectral analysis failed to demonstrate regularity of insulin pulses in most insulinoma subjects, and the interpulse interval estimated by DETECT demonstrated great variability. This high degree of variability in both frequency and amplitude in insulinoma patients is evidence for loss of the highly organized physiological regulation of oscillations seen in normal subjects.
This impaired regulation of plasma insulin oscillations in the insulinomas was accompanied by unusually high pulses of insulin concentrations in three of five subjects with insulinomas. Sudden changes of this type are caused by an increase in insulin secretion (plasma C-peptide at peak insulin levels was also elevated; data not shown). Occurrence of pulses in the postabsorptive state, when fuel concentrations are stable, suggest that the stimulus for secretion is an intrinsic one. This is consistent with findings in vitro that demonstrate that the isolated pancreas and isolated islets secrete insulin in a pulsatile fashion, even when fuel concentrations are held constant in perifusing media (18, 19). This is more specifically supported by the findings of Chou et al. (13) who demonstrated that insulinomas, when studied in vitro, secrete insulin in oscillatory fashion. Thus, in vivo, unusually high plasma insulin pulses appear to represent insulin oscillations that are intermittently exaggerated when secreted by an adenoma. This represents another feature of autonomy from physiological control that characterize endocrine tumors.
Unusually high insulin pulses were not observed in all the time series, despite ongoing oscillations. It is therefore possible that some adenomas secrete insulin in more stable fashion. This is not unusual in endocrine tumors, which may display different degrees of autonomy. However, in none of the control subjects were there any pulses of the magnitude observed in patients with insulinomas. This finding thus corroborates long-held clinical suspicions of erratic secretory behavior in some insulinomas (2, 3, 4). Demonstration of unexpected pulses in the postabsorptive state thus may be an additional tool that can raise the index of suspicion for the presence of an insulinoma. However, this determination requires more than just the finding of a high concentration of insulin; rather a series of at least three measurements to represent both nadir and pulse peak should be identified.
Possible mechanisms for abnormal regulation of insulin oscillations deserve some comment, although this study cannot distinguish between them. 1) Irregularity at the insulin secretory level. 2) Two independent sources of insulin secretion, each responding to a different pacemaker; i.e. oscillations of insulin originating from the insulinoma acting as an autonomous secretor and separately from normal pancreatic islets. If these two sources are not coordinated, the resulting oscillatory pattern in plasma are likely to be irregular. 3) The effect of changes in blood glucose concentrations. In vitro studies have demonstrated that high-frequency insulin oscillations are entrainable by fairly small changes in glucose concentrations (19). Glucose responses to erratic insulin pulses might alter the regularity of insulin oscillations by entrainment. 4) A role for proinsulin. Although cosecreted with insulin in pulsatile fashion (13), the plasma half-life of proinsulin is longer than insulin, and might interfere with the pattern of oscillations if both molecules are detected in the insulin assay. This possibility was excluded.
In summary, by using frequent serial blood sampling in the postabsorptive state, this study has demonstrated high-frequency insulin pulsatility in patients with insulinomas. However, the pulses are irregular in amplitude and frequency in contrast to normal controls. These data provide strong evidence for a loss of regulatory control over insulin secretory oscillations in these patients. The results form a basis for understanding the occasional unexpected episodes of hypoglycemia in insulinoma patients by demonstrating the presence of unusually large and erratic insulin pulsatility in some individuals. This finding provides additional evidence for the autonomous, unregulated nature of insulin secretion in patients with insulinomas. Finally, unexpectedly high pulses of plasma insulin concentrations found in fasting patients during evaluation of hypoglycemia, should be considered highly suggestive of insulinoma.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received February 6, 1997.
Accepted May 23, 1997.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  P. Wiesli, B. Schwegler, B. Schmid, G. A Spinas, and C. Schmid Mini-Mental State Examination is superior to plasma glucose concentrations in monitoring patients with suspected hypoglycaemic disorders during the 72-hour fast Eur. J. Endocrinol., April 1, 2005; 152(4): 605 - 610. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. S. Mao, N. Berman, and E. Ipp Loss of entrainment of high-frequency plasma insulin oscillations in type 2 diabetes is likely a glucose-specific {beta}-cell defect Am J Physiol Endocrinol Metab, July 1, 2004; 287(1): E50 - E54. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Vig, M. Lewis, K. J. Foster, and A. Stacey-Clear Lessons to be learned: a case study approach Insulinoma presenting as a change in personality The Journal of the Royal Society for the Promotion of Health, March 1, 2001; 121(1): 56 - 61. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
