Department of Endocrinology, Mayo Clinic, Rochester, Minnesota 55905
Address correspondence and requests for reprints to: F. J. Service, Mayo Clinic, Department of Endocrinology-West 18, 200 First Street SW, Rochester, Minnesota 55905.
![]() |
Introduction |
---|
![]() ![]() ![]() |
---|
In the September 2000 issue of JCEM, Hirshberg et al. (4) question the necessity for extending the period of fasting beyond 48 h in patients being evaluated for a hypoglycemic disorder. They have confirmed previous reports that most persons with insulinoma have a positive fast within 48 h. The Mayo Clinic experience of positive fasts within 48 h in large series of patients with insulinoma has been reported as: 94% (n = 108, 19271965) (5) and 92% (n = 51, 19641975) (6). In the subsequent period, 1976 to the present, we have histological confirmation in 205 patients operated on for insulinoma. Among these, 170 underwent prolonged fasts according to a standard protocol. The fast was terminated within 12 h in 33%, 24 h in 65%, 36 h in 84%, 48 h in 93%, and 72 h in 99%. Two patients had negative fasts at 72 h; in one the fast remained negative through 96 h. Although a very rare event, negative 72-h fasts in insulinoma patients have been observed by others.
Could any of these prolonged fasts have been terminated earlier without sacrificing diagnostic accuracy, as suggested by Hirshberg et al. (4)? The answer to this question is predicated partly on the purpose for which the prolonged fast is conducted. When the prolonged fast is done to determine the role of ß-cell polypeptides in the genesis of previously documented hypoglycemia, the fast may be terminated when the plasma glucose is 55 mg/dL or less (7) because ß-cell polypeptide levels are suppressed below that glucose level. Assuming that plasma glucose is being monitored on a frequent basis, there is no good reason for a fast done for that reason to be unduly prolonged. However, recognition of this purpose for the prolonged fast is relatively new (7). The classic end point of the prolonged fast has been the demonstration of Whipples triad. The latter requires knowledge of the current plasma glucose concentration and identification of concurrent neuroglycopenic symptoms. The decision to end a prolonged fast requires considerable judgement by the physician in attendance and is not amenable to retrospective review, except in the rare instances of egregious inattentiveness by the responsible physician.
Could the prolonged fast be arbitrarily terminated at 48 h in the asymptomatic patient and an interpretation made later based on the plasma glucose and insulin concentrations, as suggested by the NIH investigators? To do so would preclude demonstration of Whipples triad, an important factor in the diagnosis of a hypoglycemic disorder. In addition, although a low plasma glucose is necessary, it is not sufficient to make a diagnosis of a hypoglycemic disorder because of the overlap between normal and hypoglycemic patients (7). Furthermore, even with the use of carefully generated criteria for hyperinsulinism, plasma insulin levels fall below the criterion for hyperinsulinemia in a small number of insulinoma patients in our experience and that of the NIH group. Although the criteria for hyperinsulinemia were generated from the 72-h fast (7, 8) and apply when the plasma glucose is low regardless of the duration of food deprivation the obverse, the demonstration of ß-cell suppression in the nonhypoglycemic patient probably requires 72 h of fasting to generate conditions conducive to complete ß-cell suppression. It is not known whether current criteria would apply at 48 h of fasting.
Since an accurate diagnosis is so important because of implications for radiologic studies (some of which may be invasive) and major surgery, we take the position of conducting evaluations with great care and measuring not only all three ß-cell polypeptides (insulin, C-peptide, and proinsulin) but also insulin surrogates such as ß-hydroxybutyrate and plasma glucose response to iv glucagon.
We are pleased that the NIH has followed our lead in measuring plasma proinsulin directly. However, due to limitations in their analysis, which failed to relate the proinsulin concentrations to glucose concentrations in insulinoma patients vs. controls, a diagnostic criterion could not be generated by them.
Recommendations regarding optimal duration of the prolonged fast based solely on experience with insulinoma fails to take into consideration that when a prolonged fast is undertaken in a patient suspected to have a hypoglycemic disorder the presence of and, indeed, the nature of that disorder is not known. For instance, in recent years we have subjected noninsulinoma hypoglycemic patients to the prolonged fast who subsequently had the following diagnoses: insulin factitious hypoglycemia, sulfonylurea factitious hypoglycemia, end-stage renal disease, nonislet cell tumor hypoglycemia, insulin autoimmune hypoglycemia, drug-induced hypoglycemia, and noninsulinoma pancreatogenous hypoglycemia syndrome. In 14% of these patients with positive fasts, Whipples triad developed after 48 h of fasting.
The primary reason to extend the prolonged fast to 72 h in the as yet asymptomatic patient is both to minimize failure to detect a hypoglycemic disorder and to maximize certainty that a person does not have a fasting state hypoglycemic disorder. Truncating the fast at 48 h or less increases the risk for misdiagnosis. However, this is not to deny that changes cannot be made in the conduct of the prolonged fast. On occasion, we have stopped the prolonged fast in persons who experienced nausea and vomiting as a result of ketosis; the latter rules out hyperinsulinemia. ß-Hydroxybutyrate concentrations probably should be measured regularly during the prolonged fast and used as a possible end point.
Although the 72-h fast has classically been considered an inpatient procedure, we have been initiating the fast in recent years in our outpatient endocrine testing center after an overnight fast in virtually all of our patients who have high likelihood of a hypoglycemic disorder, including insulinoma. Forty percent of patients (insulinoma and other causes of hypoglycemia) studied since 1995 have had positive fasts in the outpatient setting. Those patients whose fast is not completed by the end of the business day are admitted to the hospital to complete the prolonged fast.
In conclusion, we acknowledge that the Mayo Clinic may have a proprietary interest in insulinoma since the first case was identified at this institution. A similar charge may be laid regarding the 72-h fast. It is an undisputed fact that almost all patients with a fasting state hypoglycemic disorder will have symptomatic hypoglycemia within a few hours of food deprivation. However, to capture those few patients who become hypoglycemic in the 3rd day of fasting and to be certain that a fasted patient is normal or at least does not have a fasting state hypoglycemic disorder we recommend continuation of the prolonged fast up to 72 h.
Received July 24, 2000.
Accepted August 4, 2000.
![]() |
References |
---|
![]() ![]() ![]() |
---|