Headache prevalence is greater in patients with diabetes. Split and Szydlowska found that 95 of 154 patients with noninsulin-dependent diabetes mellitus had migraine and 32 had tension headache, almost twice the frequency of the control group without diabetes.1 The basis for this increased prevalence is uncertain. As early as 1933, Critchley and Ferguson referred to a dietetic form of migraine caused by fasting.2 Fasting is commonly reported as a headache trigger by patients with and without recurrent headaches (Yom Kippur headache).3 In contrast, patients with migraine often report cravings for sweets prior to the onset of their headache; migraine develops in those patients only after ingestion of sweet food items, ie, chocolate (personal observation). It is commonly believed that hypoglycemia is the mechanism underlying fasting-induced headache, in particular, in the diabetic population. An insulin-dependent unstable diabetic patient had chronic, isolated migraine episodes, developing only after correction of recurrent profound hypoglycemia.
Objective.—To describe a patient with specific hypoglycemia rebound migraine.
Background.—There is an increased prevalence of headache in persons with diabetes. Although hypoglycemia may precipitate headache in some diabetic (and nondiabetic) patients, it is not a universal pathogenetic mechanism responsible for headache in those individuals or in normal fasting subjects.
Methods.—Clinical history, review of past medical records, neurologic examination, follow-up evaluation, electroencephalogram, and computerized tomography of the head.
Results.—A 56-year-old man with unstable diabetes mellitus had recurrent monthly episodes of profound hypoglycemia for 40 years. These episodes were followed by severe, global, pulsatile headache after glycemia became normal subsequent to intravenous infusion of glucose and the patient was no longer confused and lethargic. He had no headache preceding or throughout the actual hypoglycemic phase. His neurologic examination was normal when asymptomatic. His baseline electroencephalogram was normal, but showed mild slowing of the background in the immediate posthypoglycemic state. Computerized tomography of the head demonstrated mild atrophic changes. His severe bouts of hypoglycemia and migraine were ameliorated by prophylactic treatment with valproic acid.
Conclusion.—Posthypoglycemic migraine may occur exceptionally in patients with unstable diabetes as a rebound phenomenon, caused by an unidentified mechanism.
A 56-year-old man of Italian descent had unstable insulin-dependent diabetes mellitus for 40 years, treated with both intermediate and short-acting insulin. He had history of recurrent episodes of profound hypoglycemia with documented blood sugars in the range of 20 to 30 mg/dL, with an average frequency of one a month. He was usually treated in the emergency department with intravenous infusions of glucose. Confusion, disorientation, amnesia, and occasional agitation manifested his hypoglycemic episodes. During the recovery phase when he was no longer confused, he experienced severe bilateral pounding headaches for several hours, associated with nausea and vomiting and only partially alleviated by opiates. At times, he was able to prevent the attacks by promptly drinking orange juice or by intramuscular injections of glucagon. He had more frequent milder episodes of hypoglycemia without consequence.
Past medical history was significant for remote use of recreational drugs. He smoked a half package of cigarettes a day. He had moderate chronic obstructive lung disease, peripheral vascular insufficiency, and chronic depression. He had previously undergone aortofemoral bypass surgery. There was a strong family history of diabetes but no family history of migraine, epilepsy, or psychiatric illness. He did not suffer from headaches other than when recovering from profound hypoglycemia. He complained of occasional numbness and tingling of his feet. He had been a carpenter and homebuilder in the past, before becoming incapacitated because of diabetes and depression. He received both NPH and regular insulin twice a day at variable doses. He took lorazepam for anxiety, doxepin for depression, albuterol for wheezing, and occasional analgesics for leg pains.
Magnetic resonance imaging of the brain in the past and electroencephalograms (EEGs) while asymptomatic were normal. An EEG in the immediate recovery phase obtained after one of his severe hypoglycemic episodes, showed mild generalized background slowing but no epileptic discharges. Computed tomography of the head showed mild atrophic changes. His electrocardiogram revealed sinus tachycardia when he was agitated. A complete blood count, chemistries (other than blood sugars), and liver function tests were normal.
Valproic acid (VPA) was prescribed at a dose of 500 mg twice a day, with subsequent control of the episodes of profound hypoglycemia and complicating migraines. He experienced no side effects from VPA and continues improved after 20 months of follow-up.
The pathogenesis of headache in diabetic patients is unknown. Hypoglycemia is a suspected mechanism since fasting may trigger headaches, especially in those with headache.3 Furthermore, headache is a recognized “malaise” symptom of hypoglycemia and diabetic patients with nocturnal hypoglycemia secondary to iatrogenic evening insulin overdose complain of headache on arousal.4,5 Blau and Pyke determined, however, that hypoglycemia was a precipitant of migraine only in some persons with diabetes; in the majority of their patients, diabetes had no influence on their migraine.6 In a similar fashion, Pearce found that insulin-induced hypoglycemia produced migranous symptoms in only 2 of 20 migraineurs in his series, confirming that simple or uncomplicated hypoglycemia is not the universal precipitant of headache in fasting controls, migraineurs, and diabetic patients.7 Alternative explanations for the occurrence of glycemia-sensitive headaches (“hypoglycemic” headache) include rapidly falling blood sugar levels even if within the reference range, flat or low glucose tolerance curves between 2 consecutive hourly readings, and blunted response to glucagon-induced hyperglycemia.8,9 In support of the first mechanism, a high-protein, low-carbohydrate diet with frequent meals (the so-called hypoglycemic diet) may improve the severity and frequency of headaches in those patients exhibiting transient or relative hypoglycemia on the glucose tolerance test.8 Finally, in discussing the possible role of glycemia in precipitating headaches in diabetic and nondiabetic patients, and migraineurs, it must be remembered that peripheral venous hypoglycemia may occur even with normal brain glucose levels and that diabetic patients have regional differences in their cerebral blood flow and glucose utilization.10,11
A different perspective to this problem is that, perhaps, parallel endocrine abnormalities are responsible for headaches in migraineurs. For instance, patients with migraine (and by extension persons with diabetes) may have an enhanced sensitivity to endogenous insulin with a slow hypoglycemic recovery phase, even in the presence of a normal hypothalamic-pituitary-adrenal axis. Patients with cluster headache may also exhibit blunted ACTH and cortisol responses to the insulin tolerance test.12,13 A traditionally held view is that plasma free fatty acid concentrations are increased in migrainous patients following a period of fasting. This increase perhaps plays a role in the appearance of their headache.14 Changes in the concentration of neurotransmitters, receptor density, and affinity occur in rodents with streptozotocin-induced diabetes. Lim et al reported changes in the central dopaminergic systems, and Sandrini et al changes in brain serotonin concentration, 5-HT1A and 5-HT2 receptors in this experimental animal model of diabetes.15,16 It is conceivable that similar changes occur in humans with diabetes. These changes could explain the increased prevalence of migraine in diabetic patients since significant alterations in the metabolism of serotonin and dopamine are known to occur in migraine.17,18
Of interest was this patient's response to VPA. It is unclear if his headaches improved indirectly as a result of the amelioration of his hypoglycemic events or if it was the result of the specific pharmacodynamic antimigraine effect of VPA. Either explanation will be difficult to ascertain because VPA possesses a highly complex mechanism of action. At the cellular level, VPA has different effects on the concentration of dopamine and serotonin depending on the area of the brain being analyzed, ie, hippocampus or basal ganglia.19 At the cortical level, VPA reduces catecholamine synthesis, leaving unchanged the synthesis of serotonin.20 Finally, VPA has an insulin-dependent neuroprotective anti-apoptotic effect.21
The metabolic-endocrine effects of VPA administration in humans is equally complex; VPA reduces plasma glucose and probably alters the beta-oxidation of fatty acids, while VPA abolishes oxytocin release following insulin-induced hypoglycemia.22 At the clinical level, VPA is known to cause centripetal obesity and hyperandrogenism in women, associated with elevated leptin and insulin plasma levels and unfavorable serum lipid profiles.23,24 Of parallel significance is the propensity of VPA to lower plasma carnitine in persons with diabetes, which, in the view of some investigators, constitutes the basis for VPA encephalopathy and VPA-induced neonatal hypoglycemia subsequent to exposure in utero.25,26 It is unknown if carnitine plays a role in the pathogenesis of migraine. However, one patient with a subtle defect in mitochondrial beta-oxidation, abnormal carnitine metabolism, migraine, and VPA-induced encephalopathy reversed by riboflavin was reported by Triggs et al.27
This patient is unique because of the following clinical aspects: (1) he has a long-standing history of recurrent profound hypoglycemia, (2) migraine developed solely during the recovery (rebound) phase of his hypoglycemia, and (3) he had an excellent therapeutic response to the administration of VPA.