Hypertension and Diabetes Mellitus as Determinants of Multiple Lacunar Infarcts
Background and Purpose We investigated the relationship between hypertension, diabetes mellitus, and lacunes.
Methods From 1237 cases of ischemic stroke in the Stroke Data Bank of the National Institute of Neurological and Communicative Disorders and Stroke, data from 637 patients whose initial computed tomograms showed lacunar (n=184) or nonlacunar infarcts (n=453) were analyzed. The group with lacunar infarcts was further divided into subgroups according to whether the patients had multiple (n=40) or single (n=144) lacunar infarcts. The association of hypertension and diabetes mellitus with lacunar infarcts was investigated using logistic regression models that included age, sex, and cardiac disease. Similar models were used to analyze the effects of diastolic and systolic blood pressure.
Results Hypertension (odds ratio [OR], 2.5; 95% confidence interval [CI], 1.1 to 6.0) and diabetes (OR, 2.3; 95% CI, 1.1 to 4.5) were significantly related to multiple but not to single lacunes. Cardiac disease was inversely associated with both single and multiple lacunes. Diastolic blood pressure significantly affected the probability of multiple lacunar infarcts (OR, 1.4; 95% CI, 1.04 to 1.9), whereas systolic pressure did not.
Conclusions There may be etiologically distinct lacunar infarct subgroups, with multiple lacunes being strongly related to hypertension and diabetes mellitus. Other stroke risk factors may be more important in patients with single lacunes. Diastolic rather than systolic pressure seems to be a major determinant of multiple lacunes.
The role of hypertension in the etiology of lacunar infarcts is controversial. Studies by C.M. Fisher1 indicating that almost all lacunes are caused by hypertensive microangiopathy prompted some clinicians to regard further search for other possible stroke mechanisms in hypertensive patients with lacunes as unnecessary. Others have questioned this notion, arguing that hypertension is equally important for both lacunar and nonlacunar infarcts.2 Reported rates of hypertension in patients with lacunar infarct in fact vary widely, ranging from less than 50%3 to 97%.1
Diabetes mellitus is a well established risk factor for small-artery occlusive disease affecting the distal extremities, kidneys, retinas, peripheral and cranial nerves, and other tissues. However, its role in cerebral small-vessel disease is less well characterized. Autopsy studies have yielded ambiguous results, some of them suggesting an association between diabetes and lacunes4 and others no special relationship.1 Epidemiological studies have demonstrated an independent contribution of diabetes to the risk of stroke, but have not investigated the relationship between diabetes and lacunes.5
Little attention has been paid to the identification of subgroups of lacunar infarcts. On the basis of a study of 100 cases, Boiten et al6 proposed that single and multiple lacunes constitute morphologically and clinically distinct entities. They found an association between multiple lacunes and hypertension, but concluded that it was not independent. No synergistic effects of hypertension and diabetes in the etiology of lacunar infarcts have been found.7
This study investigates whether hypertension and diabetes are associated with lacunes.
The National Institute of Neurological and Communicative Disorders and Stroke (NINDS) instituted the Stroke Data Bank (SDB), a multicenter project, to prospectively collect data on the mechanisms, clinical course, and outcome of stroke. Between 1983 and 1986 a total of 1805 patients with intracranial hemorrhage or ischemic stroke were enrolled at four participating institutions (Boston University School of Medicine, The Neurological Institute of New York, Michael Reese Hospital and Medical Center [Chicago], and the University of Maryland School of Medicine [Baltimore]).
Variables identified on hospital admission and documented in the SDB included hypertension, diabetes, myocardial infarction, coronary artery disease, congestive heart failure, atrial fibrillation, other arrhythmias, and blood pressure. Hypertension was assumed to be present under the following conditions: (1) the patient or the patient’s family reported to SDB staff that they had been informed of the diagnosis by a physician before the SDB-classifying stroke event and antihypertensive medication had been recommended, or (2) the patient’s medical documents showed that hypertension requiring treatment had been diagnosed by a physician before the SDB-classifying stroke event. The patient’s treatment status (on medication or noncompliant with antihypertensive medication) was also documented. Diabetes was defined as chronic hyperglycemia requiring diet, oral medication, or insulin treatment, diagnosed before stroke onset.
Cranial computerized tomography (CT) scans were evaluated by the SDB investigators at the individual centers. The SDB protocol did not call for estimates of interrater variation or masked CT evaluation. Further details on the methods, design, and baseline characteristics of the SDB have been described in previous reports.8
Of the 1805 SDB cases, 1237 with complete cranial CT and risk factor documentation were diagnosed as cerebral infarctions. From these 1237 patients, 637 with ischemic lesions on their initial CT scans were judged eligible for the current study. Of the 600 cases excluded, 533 showed no ischemic lesion on first CT and 67 had both lacunar and nonlacunar infarcts. We chose to limit our analysis to the initial CT scan because the SDB protocol did not require a second scan. For the 637 patients included, 96% of the CT scans had been carried out more than 6 hours after the onset of symptoms and 66% more than 1 day after stroke onset; thus, only a few patients may later have developed an undiscovered lacune, just as others may have developed an undiscovered nonlacunar lesion.
The 637 patients were divided into the lacunar and the nonlacunar infarct groups. Lacunes were defined as small (diameter <2 cm) deep lucencies on CT in the penetrator territories. All other infarcts (pial-artery territorial type, large deep lesions, and border-zone infarcts) were considered nonlacunar, and these patients served as a comparison group.
The first statistical analysis consisted of three logistic regression models. The first was used to estimate the effect of hypertension and diabetes on all patients with lacunar infarcts. The second and third models were used to analyze data from two subgroups, one for patients with one lacune and the other for patients with multiple (two or more) lacunes. Age (years divided by 10 to create 10-year increments and facilitate interpretation of the results) and sex were included in each model as controls. To control for the effect of possible cardiac sources of embolism, the variable cardiac disease was also included. It was defined as the presence of one or more of the following: history of myocardial infarction, valvular heart disease with or without surgery, atrial fibrillation and other arrhythmias (including electrocardiogram findings), congestive heart failure, or angina pectoris. An interaction term for diabetes and hypertension was added to the regression models to allow for the possibility that these variables have a synergistic effect.
The second and third analyses estimated the impact of diastolic and systolic blood pressure on lacunes (one analysis for patients with the clinical diagnosis of hypertension only, and one analysis for all cases). Three regression models (for all patients with lacunes, those with single lacunes, and those with multiple lacunes) were again provided in each analysis. Diastolic and systolic blood pressure on admission to the hospital, diabetes, age, sex, and cardiac disease were included as independent variables. Blood pressure values were divided by 10 to create 10-mm Hg increments and facilitate interpretation of the results.
Using the definitions described above, 184 patients were assigned to the lacunar group and 453 were assigned to the nonlacunar infarct group (Table 1⇓). Patients with lacunes were further subdivided into those with one (n=144) or multiple (n=40) lacunes. In 19 patients with lacunes the lesion was in the brain stem; in the remaining 165 it was located supratentorially. Among the 40 patients with multiple lacunar infarcts, 33 patients had two lacunes, 3 had three, 3 had four, and 1 had six. The lacunar and nonlacunar groups did not differ by age, but there was a significantly higher proportion of women in the single–lacunar infarct group than in the nonlacunar-infarct group (Tables 1⇓ and 2⇓).
Hypertension and diabetes were more frequent in the lacunar group than in the nonlacunar group (Table 1⇑). The first regression model revealed a significant association of lacunes with hypertension and a weaker, nonsignificant association of lacunes with diabetes (Table 2⇑). Cardiac disease was strongly and inversely associated with lacunes. In the subgroup analyses (Table 2⇑), patients with multiple lacunes were significantly more likely to have hypertension and diabetes. They were also less likely to have cardiac disease. Single lacunes, however, were not significantly associated with either hypertension or diabetes, but were inversely related to cardiac disease. The interaction between hypertension and diabetes was not significant in any of the models.
Among patients with a clinical diagnosis of hypertension, admission blood pressures were higher for those with lacunes than for those with nonlacunar lesions (Table 3⇓). In the logistic regression model, the effect of diastolic pressure was significant for the total group of patients with lacunes and for the multiple–lacunar infarct subgroup, independent of systolic pressure and treatment status. Its effect was not significant for single lacunes. Systolic pressure had no significant effect for any lacunar group.
In a further analysis (data not shown) the three regression models were extended to include all patients; ie, data from those with no clinical diagnosis of hypertension were added. Again, the effect of diastolic blood pressure was significantly greater in patients with multiple lacunar infarcts than in those with nonlacunar infarcts (P=.03), but for the total group of patients with lacunes and the single-lacune subgroup neither diastolic nor systolic pressure had a significant effect.
Our findings suggest that hypertension is an independent determinant of multiple lacunar infarcts. The overall association between hypertension and lacunar infarcts primarily reflects this. Further analyses with a larger sample would be needed to establish whether there is a linear relationship between hypertension and number of lacunes.
In contrast to our result, a meta-analysis of data from 21 studies by Millikan and Futrell2 reported no significant difference in the proportion of hypertension in patients with lacunar and nonlacunar infarcts. However, that analysis may have been flawed by a misinterpretation of the frequency of hypertension in the Lausanne Stroke Registry.3 In that registry 47% of the patients with lacunar infarct were hypertensive, not 24% (the figure used by Millikan and Futrell). Our recalculation of the meta-analysis indicated that hypertension was present in 64% of 1018 patients with lacunar infarct and 56% of 2536 patients with nonlacunar infarct (P<.05). In other studies, like ours, in which the proportions of patients with lacunar and nonlacunar infarcts who had hypertension were directly compared, both significant9 14 and nonsignificant10 11 associations were found between lacunes and hypertension. These mixed results may reflect small numbers of cases and the fact that the association between hypertension and lacunes is marginal when single and multiple lacunes are not differentiated. Finally, an analysis of a small (n=100) sample of patients with lacunar infarct that lacked a nonlacunar infarct control group6 found more hypertension in patients with multiple (71%) than with single (43%) lacunes, but this difference was linked to coexisting leukoaraiosis, which was not measured in our data.
Obviously, the prevalence of hypertension in patients with lacunes and nonlacunar infarcts varies across studies, and the proportion of patients with lacunar infarct in our study who had hypertension−73%−was higher than the 60% to 64% proportion in meta-analyses of numerous investigations.2 12 The extent to which different definitions of hypertension or sample differences may explain these discrepancies has not yet been determined. Despite this, our findings seem to support the hypothesis of distinct lacunar infarct entities6 and to strengthen the notion of a strong relationship between hypertension-induced microangiopathy and multiple lacunar infarcts.
Unresolved issues remain. Given the high prevalence of hypertension in our study (64% overall) and in other studies assessing patients with nonlacunar infarcts, the question arises as to the determining factors for small-vessel disease among hypertensive stroke patients in general. Hypertension was present in 82% of our patients with multiple lacunar infarcts and in 64% of those with nonlacunar infarcts, a difference in prevalence of less than 20%. On this basis alone, in only one of five cases of multiple lacunar infarct would the lesions be potentially attributable to hypertension. On the other hand, the strong inverse association of lacunes with cardiac disease seen in both lacunar infarct subgroups is an argument against an important cardioembolic contribution to infarct mechanisms in lacunes. The proportion of stenosing large-vessel arteriosclerosis in our patients with lacunar infarct was also too low (9%) to account for the remaining etiologically unexplained cases.
Several elements of hypertension that might predispose a person to small-vessel disease rather than to other mechanisms of stroke have been investigated. In one clinical study,10 neither systolic and diastolic blood pressure measured at various stages nor cardiac hypertrophy was found to have a specific effect. However, in an autopsy-based investigation13 the number of lacunes was significantly higher in subjects with diastolic or combined diastolic and systolic hypertension than in subjects with isolated systolic hypertension. Our results are consistent with the findings of the autopsy study. We found an association of diastolic blood pressure with lacunes and little or no effect of systolic pressure.
Like hypertension, diabetes was independently associated with multiple lacunes. This suggests that diabetes, when present, plays an important role in the etiology of disseminated cerebral small-vessel disease. A synergistic effect of hypertension and diabetes was not found; the effects of the two risk factors were apparently additive rather than multiplicative. Previous studies comparing the prevalence of diabetes in patients with lacunar and nonlacunar infarct9 10 11 or investigating a synergism between hypertension and diabetes7 revealed no significant association and no synergistic effect of the two risk factors, but did not present results for a subgroup of patients with multiple lacunes. The question of which characteristics of diabetes may render a diabetic patient more vulnerable to lacunar infarct than to nonlacunar infarct remains unanswered. Although our sample was larger than most others, it was still too small to address this question.
In two autopsy studies more than 50% of the subjects with lacunes had multiple small, deep lesions.4 13 This indicates that imaging-based analyses, including our own, underestimate the proportion of multiple lacunes. Such a systematic error tends to understate any difference found in CT-based studies. This suggests that the positive association of hypertension and diabetes with multiple lacunes is stronger than we have reported.
No transient ischemic attacks were included in our analysis, and the timing of most scans was appropriate. Nevertheless, because no matching of clinical stroke type and localization with the lesion seen on CT was attempted under the SDB protocol, some lacunar and nonlacunar infarct cases in our study may have been misclassified. The most likely error is the misclassification of true lacunar infarcts, which are more easily missed on CT scans than larger infarcts. The result of this would be that our findings underestimate the strength of the relationships that we have identified. Similarly, random error in the clinical diagnosis of hypertension and misclassification of reabsorbed small hemorrhages as single lacunar infarcts (which we consider a remote possibility) would tend to cause the positive association of hypertension with lacunes to be underestimated.
Why a higher proportion of patients with single lacunes were female remains unexplained. Other risk factors that were not analyzed as part of the NINDS SDB, such as hypercholesterolemia, smoking, and alcohol use, could be determinants of small-vessel disease, particularly single lacunes.
In summary, hypertension and diabetes are strongly and independently associated with multiple lacunar infarcts. Single lacunar infarcts appear to constitute a subgroup that is less closely linked to these risk factors. Diastolic rather than systolic blood pressure seems to determine the association of hypertension with multiple lacunes.
- Received July 7, 1994.
- Revision received September 14, 1994.
- Accepted October 5, 1994.
- Copyright © 1995 by American Heart Association
Fisher CM. Lacunes: small, deep cerebral infarcts. Neurology. 1965;15:774-784.
Millikan C, Futrell N. The fallacy of the lacune hypothesis. Stroke. 1990;21:1251-1257.
Bogousslavsky J, van Melle G, Regli F. The Lausanne Stroke Registry: analysis of 1,000 consecutive patients with first stroke. Stroke. 1988;19:1083-1092.
Tuszynski MH, Petito CK, Levy DE. Risk factors and clinical manifestations of pathologically verified lacunar infarctions. Stroke. 1989;20:990-999.
Boiten J, Lodder J, Kessels F. Two clinically distinct lacunar infarct entities? a hypothesis. Stroke. 1993;24:652-656.
Pullicino P. Small deep infarcts in diabetes. Neurology. 1990;40:249. Abstract.
Foulkes MA, Wolf PA, Price TR, Mohr JP, Hier DB. The Stroke Data Bank: design, methods, and baseline characteristics. Stroke. 1988;19:547-554.
Loeb C, Gandolfo C, Mancardi GL, Primavera A, Tassinari T. The lacunar syndromes: a review with personal contribution. In: Lechner H, Meyer JS, Ott E, eds. Cerebrovascular Disease: Research and Clinical Management. Amsterdam, The Netherlands: Elsevier Science Publishing Co; 1986; 1:107-156.
Lodder J, Bamford JM, Sandercock PAG, Jones LN, Warlow CP. Are hypertension or cardiac embolism likely causes of lacunar infarction? Stroke. 1990;21:375-381.
Boiten J, Lodder J. Lacunar infarcts: pathogenesis and validity of the clinical syndromes. Stroke. 1991;22:1374-1378.
Lodder J, Boiten J. Incidence, natural history, and risk factors in lacunar infarction. In: Pullicino PM, Caplan LR, Hommel M, eds. Advances in Neurology: Cerebral Small Artery Disease. New York, NY: Raven Press Publishers; 1993; 62:213-227.
Dozono K, Ishii N, Nishihara Y, Horic A. An autopsy study of the incidence of lacunes in relation to age, hypertension, and arteriosclerosis. Stroke. 1991;22:993-996.
Chamorro A, Sacco RL, Mohr JP, Foulkes MA, Kase CS, Tatemichi TK, Wolf PA, Price TR, Hier DB. Clinical-computed tomographic correlations of lacunar infarction in the Stroke Data Bank. Stroke. 1991;22:175-181.