These pages are best viewed with Netscape version 3.0 or higher or Internet Explorer version 3.0 or higher. When viewed with other browsers, some characters or attributes may not be rendered correctly.

Debate: Does Frequent, Severe Hypoglycemia

Cause Significant Cognitive Decrements?
On the Pro Side . . .
Hypoglycemia-Induced Cognitive Decrements in Adults With Type I: A Case to Answer?

Ian J. Deary, PhD, MRC Psych, FRCPE

A prima facie case
It should be stated at the outset that the hypothesis to be articulated and examined here is that repeated episodes of severe hypoglycemia might cause cognitive decrements in adults with type I diabetes. At times it will be referred to as "the hypothesis." By definition, an episode of severe hypoglycemia is one in which the person requires the aid of another person to effect recovery. It may or may not involve coma or seizure.

Type I diabetes is a multi-system disease. Tragically, the clinical evidence is all too obvious: visual problems, renal dysfunction, and neuropathy are borne by patients and confront the caregivers of people with type I diabetes day in, day out. However, psychological aspects of central nervous system dysfunction are not commonly numbered among the sequelae of type I diabetes and its treatment. For example, decrements in memory, judgment, attention, perception, thinking, and psychomotor function are not prominent. More specifically, despite hypoglycemia’s being the most feared side effect of insulin treatment, there has not arisen a groundswell of clinical and lay opinion to the effect that repeated bouts of hypoglycemia signal a decrement in mental function. Therefore, does this article needlessly raise an additional worry about type I diabetes? No: there are several good reasons why more thought should be given to the possibility that repeated episodes of hypoglycemia might compromise mental functions.

First, although it is rare, severe hypoglycemia may cause death.^1-3 Moreover, the neuropathological findings in humans after hypoglycemia-induced deaths reveal substantial brain changes.⁴

Second, there are well-documented cases of cerebral damage after single severe episodes of hypoglycemia. Again, these are rare. Nevertheless, problems such as profound memory disorder,⁵ cortical blindness,⁶ chronic pontine dysfunction,⁷ and other central nervous system syndromes⁸ have been reported after hypoglycemia, often supported by brain imaging evidence of structural brain changes.

Third, in the short term, mild to moderate hypoglycemia that does not involve a loss of consciousness compromises many cognitive functions, such as memory, information processing, psychomotor skills, and language.⁹ While these findings are based largely on the results of psychometric testing, more objective assessments of brain function, such as reaction times, psychophysical procedures, and "cognitive" evoked potentials also attest to the decrements of brain function during mild to moderate hypoglycemia.^9,10 Although these cognitive changes are reversible, hypoglycemia has the potential to cause at least temporary cerebral derangement. It is notable that normalization of the brain changes found during controlled hypoglycemia experiments often lags behind the return to euglycemia.⁹

Fourth, hypoglycemia might be conceived as just one of a number of individual small brain "insults" that accumulate over a lifetime to affect cognitive function permanently. Other factors that are known to affect brain function cumulatively include repeated head injuries, hypoxia (from chronic obstructive airways disease, sleep apnea, cardiac bypass surgery, and so forth), and excess alcohol.^11-14 These and other factors may be seen as agents that merely speed the process of cognitive aging, whereby "fluid" intelligence declines, slowly but measurably, from young adulthood to old age. There is some circumstantial evidence that unusually frequent and severe hypoglycemia may, rarely and perhaps in combination with other risk factors, affect cognitive function and personality permanently in a few people with type I diabetes.¹⁵

In summary, severe hypoglycemia can be fatal or can cause permanent brain damage. In the short term, it deranges many brain functions. It may be added to a list of other putative cerebral insults that may combine to produce cumulative brain decrements and accelerate the process of cognitive aging. Given these facts, it seems reasonable to hypothesize that repeated episodes of severe hypoglycemia in adulthood might affect cognitive function in proportion to the sum of any actual neural damage caused by individual episodes. Otherwise, one would have to hypothesize that the effect of severe hypoglycemia on cognitive function was all or none: either a major, clinically-apparent deficit occurred, or none at all. Prima facie, it seems more likely that this proven, potentially pathological phenomenon may cause a continuum of cerebral effects, which are mostly mild but possibly cumulative.

Methodological Problems
The possibility that repeated severe hypoglycemia permanently affects brain function is an empirical question; it should not be decided on principle. However, it is not an easy question to frame within the confines of a do-able experiment. Clinically, the question might be put as follows: "Is the mental ability of type I diabetes patients less than might otherwise be expected, and does the extent of the decrement correlate significantly with the amount of severe hypoglycemia they have suffered?" To highlight the problems of answering this seemingly simple question, some experimental issues will be considered.

It is necessary to decide which brain functions might be affected and to measure these in a valid way. That is, a practical way must be devised to measure the actual neurobehavioral change, if any, inflicted by hypoglycemia. Although there exist many objective measures of brain integrity, those most commonly used are psychometric tests of mental abilities. Therefore, among the thousands of available mental tests, a battery must be chosen that will give coverage of major cognitive functions such as intelligence, memory, language, attention and concentration, perception, motor and psychomotor function, and so forth.¹⁶Moreover, note that it is not necessarily the instantaneous cognitive level that is of interest; it is the difference between the measured level and that which would have prevailed had a person had no hypoglycemic damage. Because this might involve a longitudinal study, the same brain functions may have to be assessed on more than one occasion. This introduces the problems of practice effects due to repeated testing, a problem not met with, say, by biochemical or electrophysiological indices. The above considerations emphasize (a) the importance of the particular mental tests included in a study, and (b) the experimental design that is implemented. With regard to (a), the mental tests employed must assess the appropriate cognitive domains with sufficient sensitivity to detect what are probably small changes.

At the heart of the hypothesis being examined here is the assumption that a cumulative "dose" of hypoglycemia-induced brain insult expresses itself as less-than-expected performance on mental tests, and by implication, on the mental tasks that comprise everyday living at work and home. Therefore, having become armed with a valid battery of mental tests and an appropriate experimental design, a measure of the "dose" of hypoglycemia that has been delivered to a patient’s brain must be constructed. This is problematic and rarely done with any accuracy. It is quite common to count within the "dose" only those episodes of hypoglycemia that are defined as severe. Any contribution from mild episodes is, thereby, discounted. Of course, one can make no sharp division between episodes of mild and severe hypoglycemia in terms of their true brain impact as opposed to arbitrary clinical definition.

The upshot of imposing such a binary division on hypoglycemic events is that some essentially trivial episodes will be counted as severe and some significant episodes that have effected some neuronal damage will be dismissed as mild. Even if a study is performed prospectively, episodes of severe hypoglycemia tend to be counted as if they were equal in effect, though some studies have separated those episodes involving coma and seizure from others. It is obvious that the assumption of equal effect is false. In this type of episode-counting, the person who is given prompt help from other people to recover from several episodes of hypoglycemia may be deemed to have suffered many times more damage that another person who suffers a few untreated, prolonged, and profound attacks, perhaps involving coma and seizure. Thus, the differential "dose" entailed in each episode tends not to be discerned with accuracy, adding to the error of measurement. Studies that rely on retrospective accounts of hypoglycemia experience are likely to be even less accurate and more prone to errors.

An improved measure of "hypoglycemia dose" may prove possible if multiple measures of hypoglycemia were combined. In a different clinical domain, the advanced statistical technique of partial least squares has been used to combine multiple measures of pregnant mothers’ alcohol intake to provide a more accurate, latent measure of alcohol dose.¹⁷ This has been used to demonstrate dose-dependent effects of alcohol on their children from birth to ¹⁴ years of age.

The hypothesis is made even more difficult to test because severe hypoglycemia might not be the only cause of any, probably mild, cognitive decrements in adults with type I diabetes. Prolonged hyperglycemia and neuropathy have also been implicated as causal factors in cognitive decrements in type I diabetes, and they might interact with severe hypoglycemia in producing neurobehavioral effects.^18-19 When there is more research evidence available, it will be interesting to review whether different metabolic effects have different toxic effects on different neurobehavioral systems. For example, it has been suggested that chronic hyperglycemia may affect psychomotor efficiency rather than learning and memory.¹⁸ The fact that there is more than one metabolic effect of diabetes that can affect cognitive function has implications for any control groups against which type I diabetes patients with repeated severe hypoglycemia are compared.²⁰

In summary, a simply formulated research question is not straightforwardly answerable. Measuring the appropriate mental function levels, their difference from expected levels, and the actual "dose" of hypoglycemia incurred by individuals, and isolating any effect of hypoglycemia from other effects on cognitive function related to type I diabetes may prove problematic, if not impossible. Different research designs, all imperfect, have been implemented to address the question, as will become evident when the results of empirical studies are examined. It is important to note that these problems will tend to make studies more error-prone and more likely to produce Type 2 statistical errors. A Type 2 statistical error is said to exist when a study indicates no effect although a real effect, in truth, exists; that is, results that suggest no apparent association between repeated episodes of severe hypoglycemia and cognitive decrements may be "unsafe."

The Evidence
Let us now assess the evidence for and against the hypothesis that repeated episodes of severe hypoglycemia might cause cognitive decrements in adults with type I diabetes. Those studies that compared the mental abilities of people with type I diabetes with controls but that did not measure hypoglycemia overtly will be discounted.e.g., ^21-23 This is not intended to dismiss alternative potential causes of cerebral under-performance among people with diabetes. Indeed, there is growing evidence that chronic hyperglycemia may result in some lasting cognitive decrement, especially in combination with a history of hypoglycemia.^19,20 However, the present report will concentrate on the connection, if any, between severe hypoglycemia and cognitive decrements in adults with type I diabetes. More detailed discussions of many of the studies introduced below may be found elsewhere.^24-26

1. Evidence for the hypothesis
All of the studies that support the hypothesis that repeated severe hypoglycemia might be associated with cerebral decrements are cross-sectional and involve retrospective accounts of hypoglycemia. More than 20 years ago, Bale reported that, out of 100 people with type I diabetes and 100 matched controls, 17 people with diabetes and no controls scored at "brain damaged" levels on a word learning test.²⁷ Proportionately, type I diabetes patients with a past history of severe hypoglycemia were more likely to appear in the low-scoring group. Though this study was the first to offer citable evidence relevant to the hypothesis, the use of the terms "brain damaged" and "dementia" (in the title and body of the article, respectively) is unwarranted; there is no evidence that cognitive decrement remotely approaching the severity that might justify these emotive terms exists among people with type I diabetes other than in some rare cases such as those mentioned earlier.

Skenazy and Bigler tested 19 people with type I diabetes and found that the Wechsler Performance IQ (PIQ) and trail-making scores of people with type I diabetes were lower than those of ill and healthy controls.28 Moreover, the number of "insulin reactions" (assessed by interview) had a significant negative correlation with Performance IQ (r = -0.44, P < 0.04). There were no memory decrements in the diabetic group.

Wredling and associates compared cognitive abilities in two groups each comprising 17 people with type I diabetes: those with and those without repeated episodes of severe hypoglycemia.²⁹ The latter group, though well-matched to the former on demographic, educational, and medical factors, performed more poorly on some cognitive tasks. However, the detailed results were not consistent with other studies, as their subjects were not impaired on trail-making and reaction-time tests but were affected on tests of simple immediate memory and finger tapping. The digit symbol task, which is a subtest of the Wechsler Performance IQ battery, was performed more poorly by people with hypoglycemia, offering some indication that Performance IQ is likely to be affected by repeated severe hypoglycemia.

Sachon and colleagues studied three subject groups—type I diabetes patients with and without warning symptoms of hypoglycemia who had suffered many or no episodes of severe hypoglycemia, respectively (n’s = 30 and 25), and 25 age- and social class-matched controls—and found evidence for a detrimental effect of repeated severe hypoglycemia on word-list recall and verbal fluency and a possible effect of diabetes per se on trail-making.³⁰ This suggestion that severe hypoglycemia might affect memory and verbal fluency functions is not replicated in the studies to be discussed next.

A series of studies by our research team in Edinburgh, Scotland, has examined cognitive functions and brain integrity on an initial cohort of 100 people with type I diabetes.^31-35 It was hypothesized that the estimated frequency of hypoglycemia within a group of people with type I diabetes would correlate significantly with the difference between expected and actual Performance IQ scores. That is, the "dose-response" relationship between repeated severe hypoglycemia and IQ decrement was examined.³¹ IQ decrement was estimated by subtracting current Performance IQ (using the Wechsler Adult Intelligence Scale Revised [WAIS-R]) from estimated premorbid IQ (using the National Adult Reading Test [NART]). Frequency of severe hypoglycemia was estimated at interview. Both of these measures were examined for stability by re-assessing them 18 months later in the majority of the cohort. Stability coefficients were greater than 0.7.³² Validity of hypoglycemia histories was checked by comprehensively examining the medical notes of those people who claimed either to have had five or more or no episodes of hypoglycemia. Relatives were also interviewed. The histories appeared accurate.

There was a significant association between estimated IQ decrement and frequency of severe hypoglycemia (r = 0.33, P < 0.001, after removing potential confounding effects of age, duration of diabetes, and blood glucose level at the time of testing).³¹ This is almost certainly an underestimate because of the imperfect reliability of the measures being correlated. Frequency of severe hypoglycemia was also found to correlate significantly with reaction time, reaction-time variability, and inspection time, suggesting that following severe hypoglycemia, speed of information processing is slowed slightly and attention is affected. This suggestion was confirmed in detail in a follow-up of this group.³²

For descriptive purposes, subgroups within the cohort were identified who had had five or more (n = 24) or no (n = 23) episodes of severe hypoglycemia. Whereas the estimated IQ decrement was 7.9 IQ points in the former, it was a nonsignificant 2.1 points in the latter.³¹ Therefore, within the most severely affected quartile of the cohort (with respect to severe hypoglycemia) the IQ decrement was statistically significant, but not large. Those with the most frequent severe hypoglycemia also had longer and more variable reaction times than those with none. Lincoln and associates in Nottingham, England, studied 70 patients with type I diabetes in an attempt to replicate the Edinburgh design.³⁶ The researchers found a correlation of -0.22 (P < 0.05) between the estimated decline in intelligence in their patients and the reported frequency of major hypoglycemic attacks.

We shall now discuss some follow-up studies of the Edinburgh cohort of 100 people with type I diabetes. A comparison with 100 closely matched healthy subjects showed that the type I diabetes group was significantly lower on both Verbal IQ and Performance IQ, but that only the Performance IQ difference disappeared when the effects of severe hypoglycemia were removed statistically.³³ This means that adults with type I diabetes score slightly lower than people without diabetes on tests measuring two broad types of mental ability. It appears that severe hypoglycemia might be implicated in affecting mental performance that involves novel materials and speeded performance. Scores on mental tests involving knowledge, reasoning, and vocabulary are also performed more poorly by people with diabetes, but the cause is unlikely to be severe hypoglycemia. It might be that some other metabolic derangement is involved, or that the education of people with diabetes suffers due to the illness. A further study of the Edinburgh cohort also raised the possibility of a link between hypoglycemia unawareness and IQ decrement.³⁴

From the cohort, the regional cerebral blood flow (estimated by single photon emission computed tomography [SPECT]) of 10 patients with and without repeated severe hypoglycemia was compared with healthy controls.³⁵ Patients with type I diabetes had greater tracer uptake in the superior pre-frontal cortex, and this was especially marked in those with previous severe hypoglycemia. These latter patients alone had a relative reduction in tracer uptake compared with controls in the calcarine cortex. As discussed in the published paper, such cerebral blood flow differences are consistent with neuropsychological and electroencephalographic studies of hypoglycemia.

In summary, there is sufficient evidence for the hypothesis to suggest that there is a case to answer. Repeated severe hypoglycemia might especially affect Performance IQ, though the average effect of several episodes in young (between age 25 and 50 years) adults is modest.

The studies described above tend to have positive results, but a few cautions are warranted. The specific cognitive functions found to be affected differ across studies; the effects on memory are a particular point of disagreement. There are many different types of memory, involving different brain regions, and a definitive examination of memory function has yet to be conducted. Most studies test many mental functions but find effects of severe hypoglycemia on few; thus, Type 1 statistical errors are possible. (A Type 1 error means that so many comparisons have been made that some apparently significant findings occurred due to chance.) It is well known that, of studies which have been conducted, those most likely to be written up and published by researchers are the ones producing positive results. Due caution is urged, therefore, before accepting the above studies as a fair representation of the total set of those conducted to examine the hypothesis. Of course, all of the above studies are cross-sectional. This tends to attract blanket criticism with respect to (a) their inability to determine causation, and (b) possible selection biases.³⁷ With regard to (a) it is said that any association between hypoglycemia and mental function could be explained by people of lower ability controlling their illness less well and experiencing more episodes of hypoglycemia. However, the use of pre-morbid IQ and current IQ measures has shown that severe hypoglycemia is associated with IQ decrement but not original IQ levels.³¹With regard to (b), no specific bias effects have been identified that could have produced spurious positive results in the above studies.

2. Evidence against the hypothesis
Two longitudinal studies failed to find a link between severe hypoglycemia and cognitive decrements. The Stockholm Diabetes Intervention Study (SDIS) allocated adults with type I diabetes to either intensive glycemic management (n = 44) or normal management (n = 53). The groups were similar with respect to demographic and medical factors. Subjects kept records of every episode of severe hypoglycemia over 7.5 years. Over the study epoch, 80% of the intensively treated and 58% of the standard treated patients experienced severe hypoglycemia, with a mean number of episodes per patient per year of 1.1 and 0.4, respectively. The two groups were compared on a battery of neuropsychological tests at various time points, including the 7.5- year follow-up, with no evidence of cognitive decrement in the intensively treated group.^20,38,39

It has already been stated that this study had limited ability to address the key hypothesis being considered here.40 The groups were not separated efficiently on experience of hypoglycemia, the neuropsychological test battery was limited, especially with respect to Performance IQ-type tests, the power of the study was low, and the subjects may not have been studied for a sufficient period. In addition, the authors have hypothesized that the intensive group may have been spared any cognitive dysfunction caused by hyperglycemia and polyneuropathy but suffered some due to severe hypoglycemia, though such a balancing of effects would apply to so few patients in their study that this does not appear likely.

The Diabetes Control and Complications Trial (DCCT) was a multi-center, randomized study in which people between 13 and 39 years were allocated to intensified (n = 711) or conventional (n = 730) insulin therapy.41 Patients were followed for a mean of 6.5 years across 29 centers, and hypoglycemic events were noted and reported prospectively. A very extensive and appropriate battery of neuropsychological/psychometric mental tests was administered at 2, 5, and 7 years. No significant association was found between severe hypoglycemia and cognitive decrements.

This seems to be a compelling negative result. However, as has been stated elsewhere, there are several reasons for caution in accepting this influential null result.⁴²The subjects were not those in whom one might expect cognitive decrements: they were young (average 27 years), had high mean intelligence, their diabetes was of rather recent onset, and they had little experience of hypoglycemia before or during the study (only 23 out of 1,441 experienced more than five episodes of hypoglycemic coma or seizure during the trial).⁴³ The follow-up period was brief with respect to the time frame over which cognitive decrements might be expected to appear. The cognitive tests were sorted into eight domains of function on an impressionistic basis, with no formal data reduction techniques being employed. A statistically-derived Performance IQ measure should form the target of a reanalysis. Moreover, the grouping of tests under the banners of "attention" and "memory" was problematic, because each of these functions is multifaceted, with different aspects of each being subserved by different brain regions. Some of the analyses controlled for educational level, a problematic maneuver, since it is known that there is an association between intellectual function and cognitive level. The mixture of adults and adolescents necessitated some subjects’ being given different versions of test batteries, and changes over time must have involved, therefore, different amounts of practice and developmental effects in different subjects in addition to any effects of severe hypoglycemia. Also, in follow-up tests, some adolescents may have been switched from childhood to adult versions of tests. Most difficult of all for this researcher to admit, and with unknowable (perhaps no) implications for the veracity of the results, the analyses were presented in a terse and abstruse way, making it difficult to fully evaluate their adequacy.

Conclusions
What should be the verdict for the hypothesis that repeated severe hypoglycemia might cause cumulative cognitive decrements among adults with type I diabetes? Apart from innocent or guilty, Scots Law provides, uniquely, a third verdict: "Not Proven." This is the preferred conclusion here, just as another reviewer has called it "unproven."⁴⁴

In Scotland, this verdict is often taken to mean "We ken ye did it, but wi cannae proov it" ("We know you did it, but we cannot prove it"), but that is not what is meant here. Neither the null hypothesis nor the experimental hypothesis has sufficiently weighty evidence. If the retrospective studies are to be trusted, the effect sizes they suggest are quite modest; type I diabetes patients suffering five or more episodes of severe hypoglycemia can, on average, expect to lose about a third of a standard deviation in IQ during young to middle adulthood.

Whether type I diabetes and severe hypoglycemia affect cognitive function more dramatically later in life is another, largely unexplored, matter. It is often stated that cross-sectional studies may not be used to decide the direction of causation. Could it be, then, that the reverse hypothesis is true: that low intelligence leads to more severe hypoglycemia? That seems unlikely, because severe hypoglycemia has tended to correlate with estimated IQ decrement and not with premorbid IQ estimates.

On the other hand, the two negative prospective studies indicate that there is no very large effect of severe hypoglycemia on cognitive function for the great majority of young adults with type I diabetes. However, neither of these studies was specifically designed to test the hypothesis, and both, as we have seen, had insufficient power when measured against this criterion. Therefore, under the apparent contradictions between the cross-sectional and longitudinal studies may lie agreement that the experimental effect size is modest.

People value their thinking power, and patients with type I diabetes are right to ask whether severe hypoglycemia might compromise their mental capacities. This question is particularly important in the light of knowing that intensive glycemic control undoubtedly reduces retinopathy, nephropathy, and neuropathy, but at the same time increases the likelihood of severe hypoglycemia two- to threefold.⁴⁵ The definitive study to test the hypothesis has not yet been done, despite the widely recognized need for more research on this issue.^{42, 44, 46} Patients and their caregivers will make better-informed choices when they know for certain what risks they are running if they opt for tight glycemic management with its known sequela of increased episodes of severe hypoglycemia.

References

¹Tunbridge WMG: Factors contributing to deaths of diabetics under 50 years of age. Lancet 2:569-72, 1981.

²Kalimo H, Olsson Y: Effects of severe hypoglycaemia on the human brain: neuropathological case reports. Acta Neurol Scand 62:345-56, 1980.

³Tattersall RB, Gale EAM: Mortality. In Hypoglycaemia and Diabetes: Clinical and Physiological Aspects. Frier BM, Fisher BM, Eds. London, Edward Arnold, 1993, p. 190-98.

⁴Patrick AW, Campbell IW: Fatal hypoglycaemia in insulin-treated diabetes mellitus: clinical features and neuropathological changes. Diabetic Med 7:349-54, 1990.

⁵Chalmers J, Risk MTA, Kean DM, Grant R, Ashworth B, Campbell IW: Severe amnesia after hypoglycemia. Diabetes Care 10:922-25, 1991.

⁶Gold AE, Marshall SM: Cortical blindness and cerebral infarction associated with severe hypoglycemia. Diabetes Care 19:1001-1003, 1996.

⁷Perros P, Sellar RJ, Frier BM: Chronic pontine dysfunction following insulin-induced hypoglycemia in an IDDM patient. Diabetes Care 17:725-27, 1994.

⁸Lins P-E, Adamson U: Neurological manifestations of hypoglycaemia. In Hypoglycaemia and Diabetes: Clinical and Physiological Aspects. Frier BM, Fisher BM, Eds. London, Edward Arnold, 1993, p. 347-54.

⁹Deary IJ: Effects of hypoglycaemia on cognitive function. In Hypoglycaemia and Diabetes: Clinical and Physiological Aspects. Frier BM, Fisher BM, Eds. London, Edward Arnold, 1993, p. 81-92.

¹⁰Bendtson I: Neurophysiological changes of hypoglycaemia. In Hypoglycaemia and Diabetes: Clinical and Physiological Aspects. Frier BM, Fisher BM, Eds. London, Edward Arnold, 1993, p. 72-79.

¹¹Richardson JTE: Clinical and Neuropsychological Aspects of Closed Head Injury. London, Taylor and Francis, 1990.

¹²Murkin JM, Newman SP, Stump DA, Blumenthal JA: Statement of consensus on assessment of neurobehavioral outcomes after cardiac surgery. Ann Thorac Surg 59:1289-95, 1995.

¹³Cheshire K, Engleman H, Deary I, Shapiro C, Douglas N: Factors impairing daytime performance in patients with sleep apnea/hypopnea syndrome. Arch Int Med 152:538-41, 1992.

¹⁴Ryan C, Butters N: Learning and memory impairments in young and old alcoholics: evidence for the premature-aging hypothesis. Alcohol Clin Exp Res 4:288-93, 1980.

¹⁵Gold AE, Deary IJ, Jones RW, O’Hare JP, Reckless JPD, Frier BM: Severe deterioration in cognitive function and personality in five patients with long-standing diabetes: a complication of diabetes or a consequence of treatment? Diabetic Med 11:499-505, 1994.

¹⁶Lezak M: Neuropsychological Assessment. Oxford, Oxford University Press, 1995.

¹⁷Streissguth AP, Barra HM, Sampson PD, Bookstein FL: Prenatal alcohol and offspring development: the first fourteen years. Drug and Alcohol Dependence 36: 89-99, 1994.

¹⁸Ryan CM, Williams TM: Effects of insulin-dependent diabetes on learning and memory efficiency in adults. J Clin Expl Neuropsychol 15: 685-700, 1993.

¹⁹Ryan CM, Williams TM, Finegold DN, Orchard TJ: Cognitive dysfunction in adults with type 1 (insulin-dependent) diabetes mellitus of long duration: effects of recurrent hypoglycaemia and other chronic complications. Diabetologia 36:329-34, 1993.

²⁰Reichard P, Pihl M: Mortality and treatment side-effects during long-term intensified conventional insulin treatment in the Stockholm Diabetes Intervention Study. Diabetes 43:313-17, 1994.

²¹Franceschi M, Cecchetto R, Minicucci F, Smizne S, Baio G, Canal N: Cognitive processes in insulin-dependent diabetes. Diabetes Care 7:228-31, 1984.

²²Dejgaard A, Gade A, Larsson H, Balle V, Parving A, Parving H-H: Evidence for diabetic encephalopathy. Diabetic Med 8:162-67, 1991.

²³Prescott JH, Richardson JTE, Gillespie CR: Cognitive function in diabetes mellitus: the effects of duration of illness and glycaemic control. Brit J Clin Psychol 29:167-75, 1990.

²⁴Deary IJ: Neuropsychological manifestations. In Hypoglycaemia and Diabetes: Clinical and Physiological Aspects. Frier BM, Fisher BM, Eds. London, Edward Arnold, 1993, p. 337-46.

²⁵Deary IJ: Intelligence, personality and severe hypoglycaemia in diabetes. In International Handbook of Personality and Intelligence. Saklofske DH, Zeidner M, Eds. New York, Plenum, 1995, p. 711-26.

²⁶Gold AE, Deary IJ, Frier BM: Recurrent severe hypoglycaemia and cognitive function in type 1 diabetes. Diabetic Med 10:503-508, 1993.

²⁷Bale RN: Brain damage in diabetes mellitus. Brit J Psychiatry 122:337-41, 1973.

²⁸Skenazy JA, Bigler ED: Neuropsychological findings in diabetes mellitus. J Clin Psychol 40:246-58, 1984.

²⁹Wredling RAM, Levander S, Adamson U, Lins P-E: Permanent neuropsychological impairment after recurrent episodes of severe hypoglycaemia in man. Diabetologia 33:152-57, 1990.

³⁰Sachon C, Grimaldi A, Digy JP, Pillon B. Dubois B. Thervet F: Cognitive function, insulin-dependent diabetes and hypoglycaemia. J Int Med 231:471-75, 1992.

³¹Langan SJ, Deary IJ, Hephurn DA, Frier BM: Cumulative cognitive impairment following recurrent severe hypoglycaemia in adult patients with insulin-treated diabetes mellitus. Diabetologia 34:337-44, 1991.

³²Weary IJ, Langan SJ, Graham, KS, Hephurn DA, Frier BM: Recurrent severe hypoglycemia, intelligence and speed of information processing. Intelligence 16:337-59, 1992.

³³Deary U. Crawford JR, Hephurn DA, Langan SJ, Blackmore LM, Frier BM: Severe hypoglycemia and intelligence in adult patients with insulin-treated diabetes. Diabetes 42:341-44, 1993.

³⁴MacLeod KM, Deary IJ, Graham KS, Hepburn DA, Frier BM: Hypoglycaemia unawareness in adult patients with type 1 diabetes: relationship to severe hypoglycaemia and cognitive impairment. Diabetes Nutr Metab 7:205-12, 1994.

³⁵MacLeod KM, Hepburn, DA, Deary IJ, Goodwin GM, Dougall N. Ebmeier KP, Frier BM: Regional cerebral blood flow in IDDM patients: effects of diabetes and recurrent severe hypoglycaemia. Diabetologia 37:257-63, 1994.

³⁶Lincoln NB, Faleiro RM, Kelly C, Kirk BA, Jeffcoate WJ Effect of long-term glycemic control on cognitive function. Diabetes Care 19:656-58, 1996.

³⁷Reichard P: Intensified conventional insulin treatment and Neuropsychological impairment. Brit Med J 304:447,1992,

³⁸Reichard P. Berglund B. Britz A, Levander S. Rosenqvist U: Hypoglycemic episodes during intensified insulin treatment: increased frequency but no effect on cognitive function. J Int Med 229:9-16, 1991.

³⁹Reichard P. Britz A, Rosenqvist U: Intensified conventional insulin treatment and neuropsychological impairment. Brit Med J 303:1439-42, 1991.

⁴⁰Dreary IJ, Frier BM: Intensified conventional insulin treatment and Neuropsychological impairment. Brit Med J 304:447, 1992.

⁴¹The DCCT Research Group: Effects of intensive diabetes therapy on Neuropsychological function in adults in the Diabetes Control and Complications Trial. Ann Int Med 124:379-88, 1996.

⁴²Deary IJ, Frier BM: Severe hypoglycaemia and cognitive impairment in diabetes: link not proven. Brit Med J 313:767-68, 1996.

⁴³Shillitoe R: Psychological aspects of the Diabetes Control and Complications Trial. Practical Diabetes 11:93, 1994.

⁴⁴Amiel S: Diabetes and dementia: a causal association? Diabetic Med 11:430-31, 1994.

⁴⁵The DCCT Research Group: The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulindependent diabetes mellitus. N Engl J Med 329:977-86, 1993.

⁴⁶Ryan CM: Neurobehavioral complications of type 1 diabetes: examination of possible risk factors. Diabetes Care 11:86-93, 1988.

Ian J. Deary, PhD, MRC Psych, FRCPE, is a professor in the Department of Psychology at the University of Edinburgh, in Scotland.