Journal of Neuroscience Research 27:50&504 (1990)
Effect of Gender on Glucose Utilization Rates in Healthy Humans: A Positron Emission Tomography Study S.A. Miura, M.B. Schapiro, C.L. Grady, A. Kumar, J.A. Salerno, W.E. Kozachuk, E. Wagner, S.I. Rapoport, and B. Horwitz Laboratory of Neurosciences, National Institute on Aging, National Institutes of Health, Bethesda, Maryland
Positron emission tomography (PET) was used with %uorodeoxyglucose to see if gender differences in resting cerebral glucose utilization could be detected. Thirty-two healthy subjects (15 women and 17 men; age range: 21-38 yr) were examined using a highresolution PET scanner to determine the regional cerebral metabolic rate for glucose (CMRglc) in 65 gray matter regions of interest. Whole brain CMRglc did not differ significantly between the two genders, nor did any of the regional CMRglc values. Only 1 of 65 ratios of regional-to-whole brain CMRglc differed significantly between men and women, which is consistent with chance. These results indicate that there are no differences in resting regional cerebral glucose utilization between young men and women. Key words: Positron emission tomography, deoxyglucose, brain, sex differences, brain metabolism
INTRODUCTION Gender differences have been observed in both functional and anatomic studies of the human brain. Ray and colleagues (1976) noted sex differences in EEGs, and differences in evoked potentials were observed by Gale et al. (1978). Recent investigations of cerebral blood flow (CBF) using xenon-133 inhalation and SPECT also have elicited gender differences, with women exhibiting a 15-20% higher rate of CBF, at rest or with cognitive activation. than men of the same age (Gur et al., 1982; Devous et al., 1986; Rodrigues et al., 1988). In addition, women also were observed to have a greater degree of Literalization of CBF than men on a spatial task that preferentially activated right hemispheric CBF (Gur et al., 1982). The relevance to metabolic studies is that rates of resting CBF are coupled to resting cerebral glucose metabolic rates (Baron et al., 1985; Raichle et al., 1976), though CBF may be influenced by the oxygen-carrying capacity in blood (which differs in the two sexes), as well as pC0, or neurogenic factors. 6 1990 Wiley-Liss, Inc.
Two previous studies of neocortical glucose metabolism using positron emission tomography observed that the women studied had 19% (Baxter et al., 1987) and 26% (Yoshii et al., 1988) higher cerebral metabolic rates for glucose (CMRglc) than men. This sex difference persisted with age (Yoshii et a]., 1988). Moreover, women exhibited a greater variance in CMRglc than did men (Baxter et al., 1987). Calculated brain volume and atrophy accounted for 21% of the variance observed in the study of Yoshii et al. (1988), with smaller brains exhibiting higher metabolic rates. The gender difference found by Baxter et al. (1 987) was global, rather than focal, with all neuroanatomical regions displaying essentially the same glucose metabolic difference between the sexes. Neuropsychologic 5tudies have elicited gender differences which are most prominent in tasks requiring spatial-perceptual skills and in those of verbal ability (Ray et al., 1976; Burstein et al., 1980; Piazza, 1980; Kimura, 1980; Butler, 1984; Kimura and Harshman, 1984). Several investigators have hypothesized that the gender differences observed in neocortical metabolism and neuropsychologic performance are a function of hormonal variation, i .e., estrogen concentration (Hampson and Kimura, 1988; Baxter et al., 1987; Nehlig et al., 1985). Baxter and colleagues speculated that the greater variation observed in cerebral glucose metabolism of women compared with men was a function of serum levels of estrogen. Similarly, Hampson and Kimura (1988) reported variations in cognitive performance by women with regard to the phase in the menstrual cycle during which they were tested. During the midluteal phase when both estrogen and progesterone levels are high, women performed better on complex manual tasks and worse on tasks requiring visual-spatial skills. The Received July 13, 1990; revised August 21, 1990; accepted August 22, 1990. Address reprint requests to Bany Horwitz, Ph.D., Laboratory of Neurosciences, Bldg. 10, Rm. 6C-414, NIA, NIH, Bethesda, MD 20892.
Cerebral Glucose Metabolism and Gender
performance of women taking oral contraceptives was more similar to performance of women in their midluteal phase than the women in the phase of their cycles during which estrogens and progesterone are low. Anatomic studies conducted by Diamond and colleagues support the hormonal theory of gender differences in brain function in that they noticed a higher concentration in cortical estrogen receptors in the right hemisphere of female rats in comparison to the higher left hemisphere receptor concentration in males (Diamond et a]., 1981). In the current study, 32 healthy volunteers were studied with high-resolution positron emission tomography to investigate whether gray matter glucose metabolism is influenced by gender.
METHODS Subjects 'Thirty-two healthy volunteers (15 women: 21-38 yr; 17 men: 21-38 yr) were selected for this study (Table I). Screening included medical and family history, physical and neurologic examinations, and laboratory tests as previously described (Duara et a]., 1984). Subjects with brain disease or illness that might affect the brain, including hypertension and other cardiovascular or cerebrovascular diseases, diabetes, excessive alcohol intake, toxic exposure, or head trauma, were excluded. All subjects were normotensivc. Subjects had an average of 15.3 (* 0.9) years of education for female participants, and 16.9 ( 2 1.9) years for male participants.
PET Positron emission tomographic (PET) scans were obtained using the Scanditronix PC 1024-7B (Scanditronix, Uppsala, Sweden) tomograph. In the center of the field of view, the transverse resolution is 6 mm for straight and cross slices, and the axial resolution is 1 1 and 8 mm for straight and cross slices, respectively. Subjects were placed in the scanner with their eyes covered and ears plugged. Head movement was minimized by using a thermoplastic mask which was individually fit. A line defining the inferior orbitomeatal (IOM) line was drawn on the face mask to align the subject. A 5 mCi bolus of ['8F]-2-fluoro-2-deoxy-D-glucose ("FDG) was injected via an intravenous catheter. Arterial blood was collected during the uptake period and scan for measurement of plasma radioactivity and plasma glucose concentration. Two multislice PET scans were obtained beginning 45 min after injection of tracer. Each scan consisted of 7 slices ranging from 10 to 100 mm above the IOM line. The sets of interleaved slices were offset by 6.9 mm. Each PET slice had a minimum of 2 million coincidence counts. Two 7-slice transmission scans to correct for attenuation were obtained at the same two lcvels as the emission scans.
501
The Brook's modification (1982) of the operational equation of Sokoloff et al. (1977) was used to calculate glucose metabolic rates (units: mg/100 g/min). Values of the kinetic constants were k , =O. 102, k , = O . 130, k3=0.062, and k,=0.0068 (Huang et al., 1980). The lumped constant in the operational equation was taken as 0.418 (Huang et al., 1980). PET slices were analyzed by dividing each brain into 64 lateral, 28 medial, and 7 subcortical regions of interest (ROIs) per hemisphere. Placement of these ROIs was aided by use of an atlas of neuroanatomy (Eycleshymer and Shoemaker, 191 1). Including three midline ROIs, a total of 201 individual ROIs were analyzed. These ROIs then were grouped into frontal, parietal, temporal, occipital, sensorimotor, paralimbic, and subcortical regions (a total of 31 bilateral and 3 midline regions). Men and women were compared using the absolute rates of CMRglc in each of the 65 ROIs, as well as the ratios of regional to whole brain CMRglc (whole brain CMRglc was evaluated as the mean CMRglc of 5 slices between 30 and 90 mm above the IOM line).
CT X-Ray computer-assisted tomography (CT) was performed on 7 of the female and 10 of the male subjects who were evaluated by PET. Noncontrast CT images were obtained using a General Electric 8800 CT/T scanner (General Electric Corp., Milwaukee, WI). Serial slices, each 10 mm thick and separated by 7 mm, were taken parallel to and from 0 to 110 mm above the IOM line. Intracranial structural volumes were derived by multiplying the summed number of pixels in each outlined ROI across all slices in which the region is visualized by the pixel area (0.0064 cm') and the interslice distance (0.7 cm). For the purposes of this study, the structures of interest are the 7-slice intracranial volume, which is defined as the 50 mm horizontal segment of the brain, commencing at the lowest slice that contained the body of the third vcntricle, and the total ventricular volume, which is defined as the sum of the lateral and third ventricular volumes. Details of the CT procedure can be found in Schwartz et al. (1985) and DeLeo et al. (1985).
RESULTS There was no significant difference between men and women for age (P > 0.05), but male subjects had a significantly ( P < 0.02) higher level of education than did the female subjects. Global CMRglc (the average of all gray matter ROIs) and whole brain CMRglc (mean of all whole slice values of CMRglc) did not differ between the men and women ( P > 0.05) (see Table I). In addition, not one of the 65 regional CMRglc (rCMRglc) values showed a
502
Miura et al.
TABLE I. PET and CT Results Men
Women
28.5 5 4.9" 16.9 f 1.9 6.74 k 1.07 8.63 i 1.31 1274 t 92 21 t 17
28.7 t 5.5 15.3 2 0.9h 6.78 2 0.89 8.67 ? 1.03 1130 L 12Sb 12 f 11
Male vs Female: Mean rCMRglc in 65 ROls 11
Age (yrj Education (yrj Whole-brain CMKglc (mgi100 gimin) Global CMRglc (mgi100 gimin) 7 - ~ 1 i c cranial e volume (mP) Total ventricular volume (m13)
,
"Mean t standard deviation. hMean for women significantly different than for men (p < 0.02).
significant difference between the two groups ( P > 0.05). This is illustrated in Figure l A , where mean rCMRglc in each ROI is plotted in women versus the corresponding value i n men. The closeness of the values in the two groups is expressed by the fact that the best-fit regression line through these data has a slope near I (0.89) and a y-intercept near zero (0.97 mg/100 gimin); the correlation coefficient for these data is 0.98. Only one ratio of rCMRglc to whole brain CMRglc (posterior cingulate) showed a significant ( P < 0.05) difference between the two sexes. This result is consistent with chance, given the large number of t tests that were performed. Figure 1B displays a plot of the mean value of the ratio of regional-to-whole brain CMRglc in each ROI in women versus the corresponding quantity in men. The best-fit regression line for these points has a slope of 0.89, y-intercept of 0.13 and correlation coefficient of 0.98, which shows how similar the mean ratios are for the two sexes. Total intracranial volume, as measured by CT, was significantly ( P < 0.02) larger in men than in women (see Table I). Total ventricular volume (sum of the lateral and third ventricles) did not differ significantly between the sexes ( P > 0.2).
DISCUSSION This study addressed the question of whether significant gender differences exist in resting regional cerebral glucose utilization. A comparison of rigorously screened healthy young men and women showed no significant differences in absolute global or regional glucose metabolism. Moreover, only 1 of 65 ratios of rcgionalto-global CMRglc differed significantly between sexes, which, given the large number of t tests performed, is most likely a chance finding. These results suggest that there is no difference in resting cerebral metabolism between young men and women. The results of this study differ from those of both Baxter et al. (1987) and Yoshii et al. (1988). This may bc attributable to differences in technique between the different investigations. The Scanditronix tomograph used
7
6
5
Mean
8
9
I
10
rCMRglc (mQ/lOOglmin) Males
Male vs Female: Ratios of Regional-to-Whole Brain CMRglc
0.8
1 .o
1.2
1.4
1.6
Mean Replonal-to-Whole Brain CMRglc Ratio Males
(A) Plot of the mean regional cerebral metabolic rate for glucose (rCMRglc) in each of 65 regions of interest (ROI) in women versus the corresponding mean in men (each point represents the mean values in a specific ROI). The line represents the best-fit, least-squares regression line through the 65 points. Slope and y-intercept for the line are given in the equation; R represents the correlation coefficient for the 65 points. (€5) Plot of the mean ratio of regional-to-whole brain CMRglc in each of 65 ROIs in women versus the corresponding mean in men. The least-squares regression line is shown, as are the slope, y-intercept, and correlation coefficient for the 65 points.
in this study provides better sensitivity and a higher degree of spatial resolution (6 mm within plane) than do the scanners used in the other studies; Baxter et al. (1987) used a NeuroEcat (spatial resolution is 1 1 mm in plane), whereas Yoshii et al. (1988) used a PETT V (spatial resolution is 15 mm in plane). The higher degree of spatial resolution exhibited by the Scanditronix PC 1024
Cerebral Glucose Metabolism and Gender
decreases partial volume effects of CSF and white matter (Mazziotta et al., 1981), allowing for a more accurate comparison of cerebral metabolic rates between men and women. However, this effect is unlikely to account for the gender differences found by the other research groups, because the partial volume effect could reduce the distinction between groups, rather than increase it. As pointed out by Baxter et al. (1987), because women, as a group, have smaller brains than men, the partial volume effect should decrease rCMRglc for a b‘ riven structure in women compared with the same structure in men, since a smaller brain is more likely to include more white matter and/or sulcal CSF for a given size ROT than a larger brain. The most likely explanation for the difference between our results and the other investigations is associated with how attenuation of photons by brain and skull is corrected. Both Baxter et al. (1987) and Yoshii et al. (1 988) used a calculated value to correct for attenuation, whereas we used a measured attenuation correction in this study. The measured method provides a greater degree of accuracy since it takes into account scx differences in brain size, skull thickness, and does not introduce the operator bias that is present in determining the size and placement of the ellipse used in the calculated method. The female subjects studied in this investigation had significantly smaller intracranial compartments than did the male subjects, as determined by CT. The effect of smaller brain size on calculated rCMRglc values also has been noted in another study from this laboratory. Young adult subjects with Down syndrome were found to have had higher calculated mean glucose metabolism than controls, when scanned using an ECAT I1 tomograph (spatial resolution 17 mm) and a calculated attenuation correction (Schwartz et al., 1983; Horwitz et al., 1990). In a subsequent study from our laboratory of 14 young adults with Down syndrome on the Scanditronix 10247B tomograph using a measured attenuation correction, Schapiro et al. ( I 990) found no significant difference in global or regional glucose metabolic rates compared to controls. Subjects with Down syndrome generally have smaller brain sizes than do controls (Schapiro et a]., 1987). These empirical results suggest that differences in brain size, skull thickness, and operator bias in determining the size and placement of the ellipse used in the calculated method for correcting attenuation may be sufficient to explain the differences between our results and those of other investigators. The discrepancy between sex differences detected using neuropsychologic measures and the absence of gender differences in cerebral metabolism might be explained by the fact that the PET studies were conducted with participants in the resting state rather than during tasks requiring cognitive activation. No gender differ-
503
ences in resting cerebral blood flow or metabolism have been related to neuropsychological measures in the previous studies of cerebral blood flow or metabolism (Devous et al., 1986; Rodrigues et al., 1988; Baxter et al., 1987; Yoshii et al.. 1988). It is perhaps not surprising that no gender differences in resting rCMRglc were found by us using PET. In a [‘4C]deoxyglucose study by Nehlig et al. (1985) in rat, significant gender differences in rCMRglc were found only in some hypothalamic nuclei. Given the much finer spatial resolution of the quantitativc autoradiographic technique compared with PET, one might have predicted that few regions in the human brain would display significant gender differences in PET-measured rCMRglc. This is not to say that differences between the sexes will not be found when cognilive activation studies using PET are performed. One limitation of this study is that we did not control for variations in serum estrogen, by studying young women in different phases of their menstrual cycles. Baxter et al. (1987) scanned their female subjects between days 5 and 15 (day 1 was the first day of menstruation) of the menstrual cycle. However, there have been no human cerebral metabolic studies to show that brain metabolism varies with estrogen levels in females. The best available support for the influence of estrogens on cerebral metabolism is provided by the studies conducted by Nehlig et al. (1985) who showed higher rCMRglc in a few brain regions of female rats than in male rats (i.e., some hypothalamic areas). Female rats showed a nonsignificant cyclic variation in global CMRglc during the estrous cycle. In summary, we find no significant differences in resting global or regional glucose utilization, as measured by PET and FDG, in healthy young women compared to heal thy young men.
REFERENCES Baron JC, Rougemonl D, Collard P, Bustany P, Bousser MG, Comar D (1985): Coupling between cerebral blood flow, oxygen consumption and glucose utilization: Its study with positron tomography. In: Reivich M., Alavi A (eds): “Positron Emission Tomography.” New York: Alan R. IAs, pp 203-218. Baxter LR, Jr., Mazziotta JC. Phelps ME, Seline CE, Guze BH, Fairbanks L (1987): Cerebral glucose metabolic rates in normal human females versus normal males. Psychiat Res 21 :237245. Brooks RA (1982): Alternative forinula for glucose utilization using labeled deoxyglucose. J Nucl Med 23538-539. Butler S (1984): Sex differences in human cerebral function. In: De Vries GJ, De Bruin JPC, Uylings HBM, Corner M A (eds): “Sex Differences in the Brain: The Relation Between Structure and Function. Progress in Brain Research, Vol 61.’. Amsterdam: Elscvier Science Publishers, pp 443-455. Burstein B, Bank L, Jarvik LF (1980): Sex differences in cognitive functioning: Evidence, determinants, implications. Hum Dev 23: 289-313.
504
Miura et al.
DeLeo JM, Schwartz M, Crea5ey H, Cutler N . Rapport S1 (1985): Computer-assisted categorization of brain computerizcd tomography pixels into cerebrospinal fluid, white matter, and gray matter. Comput Biomed Res l8:79-88. Devous ME, Sr, Stokely EM, Chehabi HH, Bonte FJ (1986): Normal distribution of regional cerebral blood flow measured by dynamic single-photon emission tomography. J Cereb Blood Flow Metabol 3:95-104. Diamond MC, Dowling GA, Johnson RE (1981’): Morphologic cerebral cortical asymmetry in male and female rats. Exp Neurol 71:261-268. Duara R, Grady C. Haxby J , Ingvar D, Sokoloff L. Margolin RA, Manning RG, Cutler NR, Rapoport SI (1984): Human brain glucose utilization and cognitive function in relation to age. Ann Neurol 16:702-713. Eycleshytner AC, Schoemaker DM (191 I ) : “A Cross-Section Anatomy.” New York: Appleton. Gale A, Brown A, Osborne K. Smallbone A (1978): Further evidence of sex differences in brain organisation. Biol Psycho1 6203208. Gur RC, Cur RE, Obrist WD, Hungerbuhler JP, Younkin D,Rosen AD: Skolnick BE, Reivich M (1982): Sex and handedness differences in cerebral blood flow during rest and cognitive activity. Science 217559-661. Hampson E, Kimura D (1988): Reciprocal effects of hormonal fluctuations on human motor and perceptual-spatial skills. Behav Neurosci 1O2:456-459. Horwitz R , Schapiro MB, Grady CL, Rapoport SI (1990): Cerebral metabolic pattern in young adult Down’s syndrome subjects: Altered intercorrelations between regional rates of glucose utilization. J Mental Defic Res 34:237-252. Huang SC, Phelps ME, Hoffman EJ, Sideris K, Selin CJ. Kuhl DE ( 1980): Non-invasivc determination of local cerebral metabolic rate of glucose in man. Am J Physiol 238:E69-E82. Kimura D ( 1980): Sex differences in intrahemispheric organization of speech. Behav Brain Sci 3:240-241. Kimura D, Harshman RA (1984): Sex differences in brain organization for verbal and non-verbal functions. In: De Vries GJ, De Bruin JPC. Uylings HBM, Corner MA (eds): “Sex Differences in the Brain: The Relation Between Structure and Function. Progress in Brain Research, Vol 61 . ” Amsterdam: Elsevicr Science Publishers, pp 423-441. Mazxiotta JC, Phelps ME, Plummer D, Kuhl DE (1981): Quantitation
in positron emission tomography: 5 . Physical-anatomical effects. J Corriput Assist Tomogr 5:734-743. Nehlig A, Porrino LJ, Crane AM, Sokoloff 1. (1985): Local cerebral glucose utilization in normal female rats: Variations during the estrous cycle and comparison with males. J Cereb Blood Flow Metabol 5:393-400. Piazza DM (1980): The influence of sex and handcdncss in the hemispheric specialization of verbal and nonverbal tasks. Neuropsychologia 18:163-1 76. Raichle ME, Grubb RL, Gad0 MH, Eichling JO, Ter-Pogosaian MH ( I 976): Correlation between regional cerebral blood flow and oxidative metabolism. Arch Neurol 33523-526. Ray WJ, Morel1 M, Frediani AW (1976): Sex differences and latcral specialization of hemispheric functioning. Ncuropsychologia 14:391-394. Rodrigucs G , Warkentin S , Risberg J, Rosadini C (1988): Sex difference? in rcgional cerebral blood flow. J Cereh Blood Flow Metabol 8:783-789. Schapiro MB, Haxby JV, Grady CL, Rapoport S1 (1987): Quantitative C T analysis of brain morphometry in adult Down’s syndrome at different ages. Neurology 37: 1424-1427. Schapiro MB, Grady CL, Kumar A , Herscovitch P, Haxby JV. Moore A M , White B, Friedland RP. Rapoport SI (1990): Regional cerebral glucose metabolism is normal in young adults with Down syndrome. J Cereb Blood Flow Metabol 10:199-206. Schwartz M, Duara R , Haxby J, Grady C , White BJ, Kay AD, Cutler NR, Rapoport S1 (1983): Down’s syndrome in adults: Brain metabolism. Science 221:781-783. Schwartz M, Creasey €1. Grady CL, DeLeo JM, Fredcrickson HA, Cutler NR, Rapoport S1 (1985): CT analysis of brain morphometrics in 30 hcalthy men, aged 21 to 81 years. Ann Neurol 17:146-157. Sokoloff L, Keivich M, Kennedy C , DesRosiers MH, Patlak CS, I’ettigrcw KD, Sakurada 0, Shinohara M (1977): The [“Cldeoxy glucose method for the measurement of local cerebral glucose utilization: Theory, procedure and normal values in the conscious and anesthetized albino rat. J Neurochem 28:897916. Yoshii F, Barker WW, Chang JY, Loewenstein L), Apicella A, Smith D, Boothe T, Ginbberg MD, Pascal S, Duara R (1988): Scnsitivity of cerebral glucose metabolism to age, gender, brain volume, brain atrophy and cerebrovascular risk factors. J Cereh Blood Flow Metabol 8:654-661.
Effect of Gender on Glucose Utilization Rates in Healthy Humans: A Positron Emission Tomography Study S.A. Miura, M.B. Schapiro, C.L. Grady, A. Kumar, J.A. Salerno, W.E. Kozachuk, E. Wagner, S.I. Rapoport, and B. Horwitz Laboratory of Neurosciences, National Institute on Aging, National Institutes of Health, Bethesda, Maryland
Positron emission tomography (PET) was used with %uorodeoxyglucose to see if gender differences in resting cerebral glucose utilization could be detected. Thirty-two healthy subjects (15 women and 17 men; age range: 21-38 yr) were examined using a highresolution PET scanner to determine the regional cerebral metabolic rate for glucose (CMRglc) in 65 gray matter regions of interest. Whole brain CMRglc did not differ significantly between the two genders, nor did any of the regional CMRglc values. Only 1 of 65 ratios of regional-to-whole brain CMRglc differed significantly between men and women, which is consistent with chance. These results indicate that there are no differences in resting regional cerebral glucose utilization between young men and women. Key words: Positron emission tomography, deoxyglucose, brain, sex differences, brain metabolism
INTRODUCTION Gender differences have been observed in both functional and anatomic studies of the human brain. Ray and colleagues (1976) noted sex differences in EEGs, and differences in evoked potentials were observed by Gale et al. (1978). Recent investigations of cerebral blood flow (CBF) using xenon-133 inhalation and SPECT also have elicited gender differences, with women exhibiting a 15-20% higher rate of CBF, at rest or with cognitive activation. than men of the same age (Gur et al., 1982; Devous et al., 1986; Rodrigues et al., 1988). In addition, women also were observed to have a greater degree of Literalization of CBF than men on a spatial task that preferentially activated right hemispheric CBF (Gur et al., 1982). The relevance to metabolic studies is that rates of resting CBF are coupled to resting cerebral glucose metabolic rates (Baron et al., 1985; Raichle et al., 1976), though CBF may be influenced by the oxygen-carrying capacity in blood (which differs in the two sexes), as well as pC0, or neurogenic factors. 6 1990 Wiley-Liss, Inc.
Two previous studies of neocortical glucose metabolism using positron emission tomography observed that the women studied had 19% (Baxter et al., 1987) and 26% (Yoshii et al., 1988) higher cerebral metabolic rates for glucose (CMRglc) than men. This sex difference persisted with age (Yoshii et a]., 1988). Moreover, women exhibited a greater variance in CMRglc than did men (Baxter et al., 1987). Calculated brain volume and atrophy accounted for 21% of the variance observed in the study of Yoshii et al. (1988), with smaller brains exhibiting higher metabolic rates. The gender difference found by Baxter et al. (1 987) was global, rather than focal, with all neuroanatomical regions displaying essentially the same glucose metabolic difference between the sexes. Neuropsychologic 5tudies have elicited gender differences which are most prominent in tasks requiring spatial-perceptual skills and in those of verbal ability (Ray et al., 1976; Burstein et al., 1980; Piazza, 1980; Kimura, 1980; Butler, 1984; Kimura and Harshman, 1984). Several investigators have hypothesized that the gender differences observed in neocortical metabolism and neuropsychologic performance are a function of hormonal variation, i .e., estrogen concentration (Hampson and Kimura, 1988; Baxter et al., 1987; Nehlig et al., 1985). Baxter and colleagues speculated that the greater variation observed in cerebral glucose metabolism of women compared with men was a function of serum levels of estrogen. Similarly, Hampson and Kimura (1988) reported variations in cognitive performance by women with regard to the phase in the menstrual cycle during which they were tested. During the midluteal phase when both estrogen and progesterone levels are high, women performed better on complex manual tasks and worse on tasks requiring visual-spatial skills. The Received July 13, 1990; revised August 21, 1990; accepted August 22, 1990. Address reprint requests to Bany Horwitz, Ph.D., Laboratory of Neurosciences, Bldg. 10, Rm. 6C-414, NIA, NIH, Bethesda, MD 20892.
Cerebral Glucose Metabolism and Gender
performance of women taking oral contraceptives was more similar to performance of women in their midluteal phase than the women in the phase of their cycles during which estrogens and progesterone are low. Anatomic studies conducted by Diamond and colleagues support the hormonal theory of gender differences in brain function in that they noticed a higher concentration in cortical estrogen receptors in the right hemisphere of female rats in comparison to the higher left hemisphere receptor concentration in males (Diamond et a]., 1981). In the current study, 32 healthy volunteers were studied with high-resolution positron emission tomography to investigate whether gray matter glucose metabolism is influenced by gender.
METHODS Subjects 'Thirty-two healthy volunteers (15 women: 21-38 yr; 17 men: 21-38 yr) were selected for this study (Table I). Screening included medical and family history, physical and neurologic examinations, and laboratory tests as previously described (Duara et a]., 1984). Subjects with brain disease or illness that might affect the brain, including hypertension and other cardiovascular or cerebrovascular diseases, diabetes, excessive alcohol intake, toxic exposure, or head trauma, were excluded. All subjects were normotensivc. Subjects had an average of 15.3 (* 0.9) years of education for female participants, and 16.9 ( 2 1.9) years for male participants.
PET Positron emission tomographic (PET) scans were obtained using the Scanditronix PC 1024-7B (Scanditronix, Uppsala, Sweden) tomograph. In the center of the field of view, the transverse resolution is 6 mm for straight and cross slices, and the axial resolution is 1 1 and 8 mm for straight and cross slices, respectively. Subjects were placed in the scanner with their eyes covered and ears plugged. Head movement was minimized by using a thermoplastic mask which was individually fit. A line defining the inferior orbitomeatal (IOM) line was drawn on the face mask to align the subject. A 5 mCi bolus of ['8F]-2-fluoro-2-deoxy-D-glucose ("FDG) was injected via an intravenous catheter. Arterial blood was collected during the uptake period and scan for measurement of plasma radioactivity and plasma glucose concentration. Two multislice PET scans were obtained beginning 45 min after injection of tracer. Each scan consisted of 7 slices ranging from 10 to 100 mm above the IOM line. The sets of interleaved slices were offset by 6.9 mm. Each PET slice had a minimum of 2 million coincidence counts. Two 7-slice transmission scans to correct for attenuation were obtained at the same two lcvels as the emission scans.
501
The Brook's modification (1982) of the operational equation of Sokoloff et al. (1977) was used to calculate glucose metabolic rates (units: mg/100 g/min). Values of the kinetic constants were k , =O. 102, k , = O . 130, k3=0.062, and k,=0.0068 (Huang et al., 1980). The lumped constant in the operational equation was taken as 0.418 (Huang et al., 1980). PET slices were analyzed by dividing each brain into 64 lateral, 28 medial, and 7 subcortical regions of interest (ROIs) per hemisphere. Placement of these ROIs was aided by use of an atlas of neuroanatomy (Eycleshymer and Shoemaker, 191 1). Including three midline ROIs, a total of 201 individual ROIs were analyzed. These ROIs then were grouped into frontal, parietal, temporal, occipital, sensorimotor, paralimbic, and subcortical regions (a total of 31 bilateral and 3 midline regions). Men and women were compared using the absolute rates of CMRglc in each of the 65 ROIs, as well as the ratios of regional to whole brain CMRglc (whole brain CMRglc was evaluated as the mean CMRglc of 5 slices between 30 and 90 mm above the IOM line).
CT X-Ray computer-assisted tomography (CT) was performed on 7 of the female and 10 of the male subjects who were evaluated by PET. Noncontrast CT images were obtained using a General Electric 8800 CT/T scanner (General Electric Corp., Milwaukee, WI). Serial slices, each 10 mm thick and separated by 7 mm, were taken parallel to and from 0 to 110 mm above the IOM line. Intracranial structural volumes were derived by multiplying the summed number of pixels in each outlined ROI across all slices in which the region is visualized by the pixel area (0.0064 cm') and the interslice distance (0.7 cm). For the purposes of this study, the structures of interest are the 7-slice intracranial volume, which is defined as the 50 mm horizontal segment of the brain, commencing at the lowest slice that contained the body of the third vcntricle, and the total ventricular volume, which is defined as the sum of the lateral and third ventricular volumes. Details of the CT procedure can be found in Schwartz et al. (1985) and DeLeo et al. (1985).
RESULTS There was no significant difference between men and women for age (P > 0.05), but male subjects had a significantly ( P < 0.02) higher level of education than did the female subjects. Global CMRglc (the average of all gray matter ROIs) and whole brain CMRglc (mean of all whole slice values of CMRglc) did not differ between the men and women ( P > 0.05) (see Table I). In addition, not one of the 65 regional CMRglc (rCMRglc) values showed a
502
Miura et al.
TABLE I. PET and CT Results Men
Women
28.5 5 4.9" 16.9 f 1.9 6.74 k 1.07 8.63 i 1.31 1274 t 92 21 t 17
28.7 t 5.5 15.3 2 0.9h 6.78 2 0.89 8.67 ? 1.03 1130 L 12Sb 12 f 11
Male vs Female: Mean rCMRglc in 65 ROls 11
Age (yrj Education (yrj Whole-brain CMKglc (mgi100 gimin) Global CMRglc (mgi100 gimin) 7 - ~ 1 i c cranial e volume (mP) Total ventricular volume (m13)
,
"Mean t standard deviation. hMean for women significantly different than for men (p < 0.02).
significant difference between the two groups ( P > 0.05). This is illustrated in Figure l A , where mean rCMRglc in each ROI is plotted in women versus the corresponding value i n men. The closeness of the values in the two groups is expressed by the fact that the best-fit regression line through these data has a slope near I (0.89) and a y-intercept near zero (0.97 mg/100 gimin); the correlation coefficient for these data is 0.98. Only one ratio of rCMRglc to whole brain CMRglc (posterior cingulate) showed a significant ( P < 0.05) difference between the two sexes. This result is consistent with chance, given the large number of t tests that were performed. Figure 1B displays a plot of the mean value of the ratio of regional-to-whole brain CMRglc in each ROI in women versus the corresponding quantity in men. The best-fit regression line for these points has a slope of 0.89, y-intercept of 0.13 and correlation coefficient of 0.98, which shows how similar the mean ratios are for the two sexes. Total intracranial volume, as measured by CT, was significantly ( P < 0.02) larger in men than in women (see Table I). Total ventricular volume (sum of the lateral and third ventricles) did not differ significantly between the sexes ( P > 0.2).
DISCUSSION This study addressed the question of whether significant gender differences exist in resting regional cerebral glucose utilization. A comparison of rigorously screened healthy young men and women showed no significant differences in absolute global or regional glucose metabolism. Moreover, only 1 of 65 ratios of rcgionalto-global CMRglc differed significantly between sexes, which, given the large number of t tests performed, is most likely a chance finding. These results suggest that there is no difference in resting cerebral metabolism between young men and women. The results of this study differ from those of both Baxter et al. (1987) and Yoshii et al. (1988). This may bc attributable to differences in technique between the different investigations. The Scanditronix tomograph used
7
6
5
Mean
8
9
I
10
rCMRglc (mQ/lOOglmin) Males
Male vs Female: Ratios of Regional-to-Whole Brain CMRglc
0.8
1 .o
1.2
1.4
1.6
Mean Replonal-to-Whole Brain CMRglc Ratio Males
(A) Plot of the mean regional cerebral metabolic rate for glucose (rCMRglc) in each of 65 regions of interest (ROI) in women versus the corresponding mean in men (each point represents the mean values in a specific ROI). The line represents the best-fit, least-squares regression line through the 65 points. Slope and y-intercept for the line are given in the equation; R represents the correlation coefficient for the 65 points. (€5) Plot of the mean ratio of regional-to-whole brain CMRglc in each of 65 ROIs in women versus the corresponding mean in men. The least-squares regression line is shown, as are the slope, y-intercept, and correlation coefficient for the 65 points.
in this study provides better sensitivity and a higher degree of spatial resolution (6 mm within plane) than do the scanners used in the other studies; Baxter et al. (1987) used a NeuroEcat (spatial resolution is 1 1 mm in plane), whereas Yoshii et al. (1988) used a PETT V (spatial resolution is 15 mm in plane). The higher degree of spatial resolution exhibited by the Scanditronix PC 1024
Cerebral Glucose Metabolism and Gender
decreases partial volume effects of CSF and white matter (Mazziotta et al., 1981), allowing for a more accurate comparison of cerebral metabolic rates between men and women. However, this effect is unlikely to account for the gender differences found by the other research groups, because the partial volume effect could reduce the distinction between groups, rather than increase it. As pointed out by Baxter et al. (1987), because women, as a group, have smaller brains than men, the partial volume effect should decrease rCMRglc for a b‘ riven structure in women compared with the same structure in men, since a smaller brain is more likely to include more white matter and/or sulcal CSF for a given size ROT than a larger brain. The most likely explanation for the difference between our results and the other investigations is associated with how attenuation of photons by brain and skull is corrected. Both Baxter et al. (1987) and Yoshii et al. (1 988) used a calculated value to correct for attenuation, whereas we used a measured attenuation correction in this study. The measured method provides a greater degree of accuracy since it takes into account scx differences in brain size, skull thickness, and does not introduce the operator bias that is present in determining the size and placement of the ellipse used in the calculated method. The female subjects studied in this investigation had significantly smaller intracranial compartments than did the male subjects, as determined by CT. The effect of smaller brain size on calculated rCMRglc values also has been noted in another study from this laboratory. Young adult subjects with Down syndrome were found to have had higher calculated mean glucose metabolism than controls, when scanned using an ECAT I1 tomograph (spatial resolution 17 mm) and a calculated attenuation correction (Schwartz et al., 1983; Horwitz et al., 1990). In a subsequent study from our laboratory of 14 young adults with Down syndrome on the Scanditronix 10247B tomograph using a measured attenuation correction, Schapiro et al. ( I 990) found no significant difference in global or regional glucose metabolic rates compared to controls. Subjects with Down syndrome generally have smaller brain sizes than do controls (Schapiro et a]., 1987). These empirical results suggest that differences in brain size, skull thickness, and operator bias in determining the size and placement of the ellipse used in the calculated method for correcting attenuation may be sufficient to explain the differences between our results and those of other investigators. The discrepancy between sex differences detected using neuropsychologic measures and the absence of gender differences in cerebral metabolism might be explained by the fact that the PET studies were conducted with participants in the resting state rather than during tasks requiring cognitive activation. No gender differ-
503
ences in resting cerebral blood flow or metabolism have been related to neuropsychological measures in the previous studies of cerebral blood flow or metabolism (Devous et al., 1986; Rodrigues et al., 1988; Baxter et al., 1987; Yoshii et al.. 1988). It is perhaps not surprising that no gender differences in resting rCMRglc were found by us using PET. In a [‘4C]deoxyglucose study by Nehlig et al. (1985) in rat, significant gender differences in rCMRglc were found only in some hypothalamic nuclei. Given the much finer spatial resolution of the quantitativc autoradiographic technique compared with PET, one might have predicted that few regions in the human brain would display significant gender differences in PET-measured rCMRglc. This is not to say that differences between the sexes will not be found when cognilive activation studies using PET are performed. One limitation of this study is that we did not control for variations in serum estrogen, by studying young women in different phases of their menstrual cycles. Baxter et al. (1987) scanned their female subjects between days 5 and 15 (day 1 was the first day of menstruation) of the menstrual cycle. However, there have been no human cerebral metabolic studies to show that brain metabolism varies with estrogen levels in females. The best available support for the influence of estrogens on cerebral metabolism is provided by the studies conducted by Nehlig et al. (1985) who showed higher rCMRglc in a few brain regions of female rats than in male rats (i.e., some hypothalamic areas). Female rats showed a nonsignificant cyclic variation in global CMRglc during the estrous cycle. In summary, we find no significant differences in resting global or regional glucose utilization, as measured by PET and FDG, in healthy young women compared to heal thy young men.
REFERENCES Baron JC, Rougemonl D, Collard P, Bustany P, Bousser MG, Comar D (1985): Coupling between cerebral blood flow, oxygen consumption and glucose utilization: Its study with positron tomography. In: Reivich M., Alavi A (eds): “Positron Emission Tomography.” New York: Alan R. IAs, pp 203-218. Baxter LR, Jr., Mazziotta JC. Phelps ME, Seline CE, Guze BH, Fairbanks L (1987): Cerebral glucose metabolic rates in normal human females versus normal males. Psychiat Res 21 :237245. Brooks RA (1982): Alternative forinula for glucose utilization using labeled deoxyglucose. J Nucl Med 23538-539. Butler S (1984): Sex differences in human cerebral function. In: De Vries GJ, De Bruin JPC, Uylings HBM, Corner M A (eds): “Sex Differences in the Brain: The Relation Between Structure and Function. Progress in Brain Research, Vol 61.’. Amsterdam: Elscvier Science Publishers, pp 443-455. Burstein B, Bank L, Jarvik LF (1980): Sex differences in cognitive functioning: Evidence, determinants, implications. Hum Dev 23: 289-313.
504
Miura et al.
DeLeo JM, Schwartz M, Crea5ey H, Cutler N . Rapport S1 (1985): Computer-assisted categorization of brain computerizcd tomography pixels into cerebrospinal fluid, white matter, and gray matter. Comput Biomed Res l8:79-88. Devous ME, Sr, Stokely EM, Chehabi HH, Bonte FJ (1986): Normal distribution of regional cerebral blood flow measured by dynamic single-photon emission tomography. J Cereb Blood Flow Metabol 3:95-104. Diamond MC, Dowling GA, Johnson RE (1981’): Morphologic cerebral cortical asymmetry in male and female rats. Exp Neurol 71:261-268. Duara R, Grady C. Haxby J , Ingvar D, Sokoloff L. Margolin RA, Manning RG, Cutler NR, Rapoport SI (1984): Human brain glucose utilization and cognitive function in relation to age. Ann Neurol 16:702-713. Eycleshytner AC, Schoemaker DM (191 I ) : “A Cross-Section Anatomy.” New York: Appleton. Gale A, Brown A, Osborne K. Smallbone A (1978): Further evidence of sex differences in brain organisation. Biol Psycho1 6203208. Gur RC, Cur RE, Obrist WD, Hungerbuhler JP, Younkin D,Rosen AD: Skolnick BE, Reivich M (1982): Sex and handedness differences in cerebral blood flow during rest and cognitive activity. Science 217559-661. Hampson E, Kimura D (1988): Reciprocal effects of hormonal fluctuations on human motor and perceptual-spatial skills. Behav Neurosci 1O2:456-459. Horwitz R , Schapiro MB, Grady CL, Rapoport SI (1990): Cerebral metabolic pattern in young adult Down’s syndrome subjects: Altered intercorrelations between regional rates of glucose utilization. J Mental Defic Res 34:237-252. Huang SC, Phelps ME, Hoffman EJ, Sideris K, Selin CJ. Kuhl DE ( 1980): Non-invasivc determination of local cerebral metabolic rate of glucose in man. Am J Physiol 238:E69-E82. Kimura D ( 1980): Sex differences in intrahemispheric organization of speech. Behav Brain Sci 3:240-241. Kimura D, Harshman RA (1984): Sex differences in brain organization for verbal and non-verbal functions. In: De Vries GJ, De Bruin JPC. Uylings HBM, Corner MA (eds): “Sex Differences in the Brain: The Relation Between Structure and Function. Progress in Brain Research, Vol 61 . ” Amsterdam: Elsevicr Science Publishers, pp 423-441. Mazxiotta JC, Phelps ME, Plummer D, Kuhl DE (1981): Quantitation
in positron emission tomography: 5 . Physical-anatomical effects. J Corriput Assist Tomogr 5:734-743. Nehlig A, Porrino LJ, Crane AM, Sokoloff 1. (1985): Local cerebral glucose utilization in normal female rats: Variations during the estrous cycle and comparison with males. J Cereb Blood Flow Metabol 5:393-400. Piazza DM (1980): The influence of sex and handcdncss in the hemispheric specialization of verbal and nonverbal tasks. Neuropsychologia 18:163-1 76. Raichle ME, Grubb RL, Gad0 MH, Eichling JO, Ter-Pogosaian MH ( I 976): Correlation between regional cerebral blood flow and oxidative metabolism. Arch Neurol 33523-526. Ray WJ, Morel1 M, Frediani AW (1976): Sex differences and latcral specialization of hemispheric functioning. Ncuropsychologia 14:391-394. Rodrigucs G , Warkentin S , Risberg J, Rosadini C (1988): Sex difference? in rcgional cerebral blood flow. J Cereh Blood Flow Metabol 8:783-789. Schapiro MB, Haxby JV, Grady CL, Rapoport S1 (1987): Quantitative C T analysis of brain morphometry in adult Down’s syndrome at different ages. Neurology 37: 1424-1427. Schapiro MB, Grady CL, Kumar A , Herscovitch P, Haxby JV. Moore A M , White B, Friedland RP. Rapoport SI (1990): Regional cerebral glucose metabolism is normal in young adults with Down syndrome. J Cereb Blood Flow Metabol 10:199-206. Schwartz M, Duara R , Haxby J, Grady C , White BJ, Kay AD, Cutler NR, Rapoport S1 (1983): Down’s syndrome in adults: Brain metabolism. Science 221:781-783. Schwartz M, Creasey €1. Grady CL, DeLeo JM, Fredcrickson HA, Cutler NR, Rapoport S1 (1985): CT analysis of brain morphometrics in 30 hcalthy men, aged 21 to 81 years. Ann Neurol 17:146-157. Sokoloff L, Keivich M, Kennedy C , DesRosiers MH, Patlak CS, I’ettigrcw KD, Sakurada 0, Shinohara M (1977): The [“Cldeoxy glucose method for the measurement of local cerebral glucose utilization: Theory, procedure and normal values in the conscious and anesthetized albino rat. J Neurochem 28:897916. Yoshii F, Barker WW, Chang JY, Loewenstein L), Apicella A, Smith D, Boothe T, Ginbberg MD, Pascal S, Duara R (1988): Scnsitivity of cerebral glucose metabolism to age, gender, brain volume, brain atrophy and cerebrovascular risk factors. J Cereh Blood Flow Metabol 8:654-661.
