NORMAL VARIATIONS I N T H E CALIBER O F T H E HUMAN CEREBRAL AQUEDUCT G. FLYGER AND U. HJELMQUIST' Department of Histology, Earolinska Institutet and the Neurosurgical Clinic, S6dersjukhuset, Stockholm, Sweden ONE FIGURE About 10 years ago Lelcsell elaborated a new surgical procedure for the relief of non-communicating hydrocephalus due to obstruction at the aqueduct of Sylvius. Briefly, the method consists of, through the usual suboccipital exposure, easing a soft small rubber catheter up the aqueductal lumen until it is insured, by suction, that the tip has entered the third ventricle, when the catheter is removed. A spiral of metal wire is threaded onto the tip of the catheter, which is then reinserted. When the spiral is in correct position, the catheter is removed leaving the spiral in situ. The catheter used was a No. 7 or 8 CharriGre, i.e., with an outer area of 10 or 13 mm2. The chief operative risk appears to be the infliction of small lesions in this region. This is also pointed out by Bailey in the 1949 Yearbook of Neurosurgery in a review of Norlh's ('49) report of two cases in which the spiral was used for inoperable tumor causing obstruction. Instead, the methods of Torkildsen ('47) or Stookey and Scarff ('36) are recommended in these cases. Both of these are anastomotic operations. I n the former free communication is established between the posterior horn of one lateral ventricle and the cisterna magna by means of a rubber tube through a subgaleal tunnel. I n the latter, communication between the third ventricle and the subarachnoid space is established by two openings: one through the lamina terminalis, connecting the third 'This study mas made possible by a grant from the Research Fund at Sodersjukhuuet, Stockholm. 151 152 G . FLYGER AND U. HJELMQUIST ventricle with the cisterna chiasmatis, and one through the floor of the ventricle directly into the cisterna interpeduncularis. From the anatolmical standpoint, however, Leksell's method seems more attractive, since it follows the natural anatomical arrangement. Moreover, it has the advantage of permitting inspection of the fourth ventricle and the posterior surface of the aqueduct. Leksell's operative procedure provides the point of departure f o r the present study of the normal variations in the cross-sectional areas of the aqueduct at different ages. The standard textbooks of Anatomy described only the topography and approximate length of the aqueduct and state that the cross-sectional area of the aqueduct varies at different levels (Rauber-Kopsch, Gray, Villiger and others). These descriptions appear to be taken from Gerlach (1858). This author, however, carried out his investigations mainly on infant brains which were fixed in 1%potassium bichromate f o r a period of 5 to 6 weeks. A not inconsiderable amount of distortion is inevitable from hardening in bichrotmate. He made serial sections along the whole aqueduct. Each third section was traced under magnification, and he then found that the lumen of the aqueduct was not uniform throughout the extent of the aqueduct, but that it varied at different levels. No measurements of the aqueduct are given, but he published a number of drawings together with the scale. Turkewitsch ( '35, '36) has published anatomical studies on the internal structure of the aqueduct based on plastic reconstructions, but no figures are given to indicate the variations in caliber encountered. On the other hand, they give a good idea of variations in the shape of the aqueduct at different levels which, however, are a t variance with those of Gerlach. I n the normal encephalogram, in which the ventricular system is filled with air, the aqueduct is visualized as a curvilinear streak of air, the radius of which varies from case to case (Lindgren and di Chiro, '53). It is not possible, however, to detect any variations in contour which might serve as a 153 V B K I A T I O N S OF SYLVISN AQUEDUCT basis for a n anatomical division. Turkewitsch ( ’36) divided the aqueduct into five parts : Pars Anterior, Ampulla, P a r s Media, Gcnu and P a r s Posterior (fig. 1). Woollam and Millen ( ’ 5 3 ) coizsidcred this suhdivisiori rather elaborate, and suggested a division into three parts : Pars Anterior, Ampulla and Pars Posterior. They derive their. points of division from the two natural constrictions of the E’ig. 1 Normal appear:riice of the aqueduct of Rylvius as viewed p h n e (reproduced from Turkewitsch, ’36). ill a sagittal aqucductal luii1cii, one at the level of the rriiddle of the superior colliculus and the other at the level of the intercollicular sulcus. Russel wrote in 1949: “As matters stand, therefore, \ye do not Biiow either the normal range of variation in the caliber of the aqueduct . . . .” Since then oiily two studies on tlie dimensions of the aqueduct have becn published. Beckett, Netsky and Zirrimei*rnan ( ’SO), whose study was based on 11 cases of steriosis of the aqueduct a i d 50 non-hyciroceplialic cases, stated that the smallest cross-sectional area encountered in the latter was 0.09 ern2. S o further details of this series are 154 0. PLYGEIL A N D U. HJELMQUlST given. TVoollarri and lllillen ( ‘53) questioned the figure given by Beckett et al., and on measurement of tlic lumen of the appropriate aqueduct in their illustrations with tlie planimeter found tlie area was 0.009 em2. Woollam arid Rlilleri’s results were based on planimetric rneasureinent of the cross-sectional a r e a in 14 normal human adult brains. The mean crosssectional area of the aqueduct i n their series ranged from 0.6 mm2 to 2 mm2. They stated that the shapc of the aqueducts they examined approxiniated generally to the illustmtions published by G erlach. JIATERIAL The present series consists of 2 1 brains removed from individuals whose age at death ranged from 1 day to 75 years. I n tlic majority of older persons - those over 50 years of age -death was due to natural causes. Fourteen brains mere from females and 10 from males. I n no case was the cause of death related to disease of the central nervous system. Macroscopic sections of the brains revealed no abnorrnality a pa r t from moderate atrophy in the very oldest. Tlie sex, age, arid roported cause of death are listed in table 1. The irregular numhcring of the preparations is due to the fact that after the comriiencement of the investigation a number of brains were rejected on the grounds that the causc of death might have influenced the anatomical structure of the brain. MET HOD The brains were removed at autopsy in thc usual rnariner and fixed by suspension, by means of a string threaded under the basilar artery, in a bath of 10% forcmalin for a period of 2 to 4 weeks. Because of their softer consistency, tlie t,wo infant brains required the longer period of hardening in order to prevent distorting the enceplialon during removal. After fixation the mesencephalic region was removed, passed successively through increasing concentrations of alcohol, and The histcilogieal prcparations werc made by Mrs. Ulla Flyger, t o whoin n c cxprcss our hcarty thanks. 155 VARIATIONS O F SYLVIAN AQUEDUCT finally through methylbenzoate. Serial sections cut perpendicular to the long axis of the brain stem were made of the entire mesencephalon. Two contiguous sections were cut at intervals of 150 p, and one stained for cells and the other for glia. I n preparations I to I11 the sections were cut 2 0 p in thickness and in the remainder 7 p. It was found that sections TABLE 3 Present series PREPARATION NO. I I1 I11 IV VI VIII X XI XI1 XI11 xv XVI XVII XVIII XXI XXII XXIV xxv XXVII XXIX XXX XL XLI XLII AGE 76 yrs. 47 60 68 42 51 20 22 30 77 51 58 47 52 74 63 74 60 75 1 day 63 yrs. 69 62 1 day SEX F F M F F M F F F F M F M M M F F M M F M ?if F F CAUSE O F DEATH Chronic endocarditis Barbiturate poisoning Cancer of the tongue Gastric carcinoma Uterocervical carcinoma Gastric ulcer (operated) Cardiac defect Chronic hepatitis Chronic endocarditis Syphilitic aortitis Nephrosclerosis Hypertonia Gastric ulcer (operated] Addison’s disease Duodenal ulcer pneumonia Chronic endocarditis Ovarian carcinoma Hydronephrosis Laryngopharyngeal carcinoma Asphyxia Cardiac arteriosclerosis Cardiac infarction Cholangitis Asphyxia + cut at 20 ~1 were too thick for a study of the ependyma upon which the present authors are engaged. Ehrlich’s eosin was used for staining cells. All measurements are based on these sections. For the staining of glia were used Haggqvist ’s modification of Alzheimer-Mann ’s method in preparations I to X, and Ranke’s method in preparations XI to XIJII. 156 G. FLYGER AND U. HJELMQUIST The only definition of the boundaries of the aqueduct available in the literature is that given by Woollam and Millen. Like those of these workers, our sections begin first caudal to the posterior commissure and continual to a point immediately caudal to the inferior colliculus. It may justifiably be argued that the cross-sectional appearance of the aqueduct in our investigations does not correspond exactly with that encountered in vivo, since the fixatives which we used might have caused shrinkage or other distortion of the brains. Casts of the ventricles made with the brain in situ might be expected to produce more satisfactory results. Such casts have been made by, among others, Torkildsen ('33) using melted wax, and Woollam ('52) who employed Neoprene, a plastic mass for the purpose. These methods, however, are also subject to errors, since the amount of pressure used in the injection can affect the final size of parts distended by the injected material and thus produce erroneous values. From a practical standpoint, therefore, it seemed 'more valuable to fix the brains by a well-recognized pathologic-anatomic technique. The sections were magnified X 50 and the area of each section of the lumen of the aqueduct was measured with a planimeter. Each eosin-stained section was measured 3 to 5 times and the mean determined. The small number of measurements on each section did not warrant the calculation of the mean error. The differences encountered, however, never exceeded 5 units, and hence the figures are exceedingly exact. According to Romeis ( '48) considerably more reliable values are obtained by planimetric measurelment than, e.g., by excision and weighing. If correctly carried out, the exactitude attained is between 1,400 and 1,600 of the area measured. The statistical calculations of the variations in the crosssectional areas of the aqueduct are based on the formula: ds2:-PF, n-1 where n = the total values, M = the mean and x = the variable. The statistical tabulations were made by Mr. Torkel Westling. 157 VARIATIONS O F SYLVIAN AQUEDUCT RESULTS The cross-sectional areas of the aqueduct in the present series are shown in table 2. All measurements are expressed in square millimeters. The beginning and the end of the aqueduct indicate the cross-sectional areas immediately caudal TABLE 2 T h e cross-sectional area of the normal cerebral aqueduct ( i n mmz) at various levels at different ages PREP. NO. I I1 I11 IV VI VIII x XI XI1 XI11 XV XVI XVII XVIII XXI XXII XXIV XXV XXVII XXIX XXX XL XLI XLII BEGINNING AGE 76 yrs. 47 60 68 42 51 20 22 30 77 51 58 47 52 74 63 74 60 75 1 day 63 yrs. 69 62 1 day SEX F F M F F M F F F F M F M M M F F M M F M M F F END o~ n p U ~ ~ U c TAQUEDUCT 2.53 2.91 5.50 2.94 1.76 1.48 2.28 0.71 1.68 3.12 2.27 1.86 5.22 1.53 2.40 0.98 1.24 1.24 2.69 1.12 7.22 3.34 0.75 1.55 6.98 3.40 3.28 2.84 2.02 5.20 4.26 1.50 1.38 2.74 1.92 5.15 5.88 4.07 3.59 1.16 4.45 3.87 2.29 1.46 9.84 5.92 1.10 0.83 MAXIMUM AREA MINIMUM AREA DIFFERENCE BETWEEN MAX.AND MIN. 6.98 3.40 3.50 2.94 2.02 5.20 4.68 2.29 1.38 3.12 2.37 5.19 5.88 4.07 3.59 1.25 4.86 3.87 3.25 1.46 9.84 7.06 2.09 1.76 1.70 1.08 2.60 1.78 1.15 1.18 2.12 0.71 0.69 1.49 1.26 0.95 2.07 0.71 1.62 0.40 0.44 1.21 1.74 0.84 7.22 2.67 0.52 0.54 5.28 1.32 2.90 1.16 0.87 4.02 2.56 2.58 0.69 1.63 1.11 4.24 3.81 3.36 1.97 . 0.85 4.22 2.66 1.51 0.62 2.62 4.59 1.57 1.22 to the posterior comrnissure and immediately caudal to the inferior colliculus, respectively. I n the next two columns are given the maximum and minimum areas encountered, and in the last column the difference between the maximum and minimum values. 158 G. FLYGER A N D U. HJELMQUIST What. strikes one most is the wide range of variation in area, from 0.4mm2 to 9.84mm2. It might be expected that comparable levels from preparations of approximately the same age would show some agreement, but such is not the case. For example, in preparations I and XI11 the maximum area of cross-section was 6.98 mm2 and 3,12 mm2, respectively, and in preparations XXI and XXIV the difference between the maximum and minimum areas was 1.97 mm2 and 4.22 mm2, respectively. Neither does sex appear to play any part. Accordingly, as the table shows, the size of the aqueductal lumen is highly individual. I n 17 of the 24 preparations the lufmen of the aqueduct was larger at the caudal end than at the cranial end, in 8 of them as much as 2 mm2, a very high figure, when one considers the minute size of the aqueductal lumen. In 8 preparations the cranial and caudal lumina were approximately the same size ; in none of these 8 cases did the difference exceed 0.4 mm2. I n three preparations the lumen was more than 0.4 mm2 larger at the cranial end than at the caudal end. Table 3 shows the mean cross-sectional areas of the aqueduct and the standard deviation. The minute size of the aqueductal lumen is remarkable, and it is difficult to believe that this canal provides the sole means of transit for the cerebrospinal fluid from the third to the fourth ventricle. This has also been pointed out by Woollam and Millen ('53) and Hassin ( '48). The possibility of other means of exit must be taken into consideration. DISCUSSION As Broman ( '27), Bickers and Adams ( '49) and others noted, an absolute as well as a relative decrease in the size of the aqueductal lumen occurs progressively from the second foetal nionth to the time of birth. This is attributed to the influence of the nuclear masses and fiber tracts surrounding the aqueduct. This explanation would lead one to expect a progressive decrease in the aqueductal lumen up to time of 159 VARIATIONS O F SYLVIAN AQUEDUCT onset of those degenerative changes which usually accompany aging and result in atrophy of adjacent parts of the brain. The normal variations with increase in age are not known with certainty. Spiller ('16) introduced the hypothesis that the aqueductal lumen decreased with age much like the central canal of the spinal cord. To test this theory, the material was grouped according to age (table 4). TABLE 3 Mean cross-sectional area of the aqueduct ( i n mmZ) b t different levels and standard deviation MEAN STANDARD DEVIATION Beginning of aqueduct 2.4 & 1.6 End of aqueduct 3.5 2 2.2 Maximum area 3.9 2 2.1 Minimum area 1.5 zk 1.4 Difference between maximum and minimum 2.4 r+ 1.4 ~ TABLE 4 Mean cross-sectional area of the aqueduct [in mmz) in different age groups AGE GROUP BEGINNING OF AOUXDUCT END O F AQUEDUCT MAXIMUM AREA MINIMUM AREA 0-30 31 - 50 50 1.6 3.3 2.6 1.9 3.7 4.0 2.3 3.7 4.1 1.0 1.4 1.7 The table shows the mean cross-sectional areas of the aqueduct in each age group. Except at the cranial end of the aqueduct, there is a definite although not, perhaps, a statistically significant increase in area with increase in age. Whether this is due to general old age atrophy or to a reduction in the size of the surrounding nuclear masses and fiber tracts has not been studied in this investigation. It may seem peculiar that the size of the aqueductal lumen does not, like the central canal, diminish with age. The chief 160 G. FLYGER AND U. HJELMQUIST reason is doubtless the cerebrospinal fluid pressure, which does not permit any shrinkage. Now, how do our results compare with those of others? As f a r as we are aware, the only similar investigations are those of Beckett and co-workers ( '50) and Woollam and Millen ( '53). Our technique was almost identical with theirs. Beckett and his co-workers, however, reported only the smallest crosssectional area they encountered in their non-hydrocephalic series, 0.9 mmZ (corrected by Woollam and Millen), which is at variance with that encountered in this investigation 0.4mm2. It is impossible to be sure that the difference between Becliett's findings and ours are as they appear, since his paper does not describe the exact method of fixation used or the details of the technique employed in measuring the cross-sectional areas. I n this procedure, the slightest deviation in the transverse plane from the perpendicular to the long axis of the brain stem is sufficient to produce an appreciable difference from the true value. A comparison of our results with those of Woollam and Millcn must be limited to only the cross-sectional area of the aqueduchl lumen at the cranial and caudal ends and the maximum and minimum area of cross-section. All our values are considerably higher (see table 3 ) . I n their series the mean area of cross-section at the crainal end was 1.7mm2, and at the caudal end 2.5mm2. What this diffcrence is due to is difficult to say. As stated earlier, differences in the crosssection technique are a factor to be taken into consideration. Another not unessential difference lies in the fact that Woollam and Millen employed 5% formalin t o fix the brains while we used 10%. The latter concentration is stated by Hansen ('07) to be that which produces a minimum of distortion. There rema.ins, however, the fact that the size of the aqueductal lumen is so minute that one is justified in questioning whether this canal is actually the sole outlet f o r the cerebrospinal fluid. Sjoqvist ( '36) estimated that cerebrospinal fluid is formed at the rate of about 500cm3 per day. VARIATIONS O F SYLVIAN AQUEDUCT 161 A further object of this investigation has been, on the basis of the normal variations in the size of the aqueductal lumen, to try to decide what attitude should be adopted towards Leksell’s operative method for the relief of noncolmmunicating hydrocephalus. Our results would seem t o indicate that catheterization of an aqueduct which normally has such a small lumen cannot be performed without considerable risk. Furtherlmore, the aqueduct is not a straight tube, but bow-shaped, and surrounded by very soft tissues which may easily be damaged by the passage of a catheter. It is to be noted that the operation comes into question only in cases of stenosis of the aqueduct where, accordingly, the lumen is smaller than normal. From the anatomical standpoint, therefore, this operative procedure seems to us to be extremely dangerous. No extensive postoperakive follow-up study has been published, but we hope to be able to do so at a future date. SUMMARY The normal variations in the size of the lumen of the cerebral aqueduct have been studied by means of sections in a series of 24 normal human brains. The cross-sectional areas of the aqueduct were measured with the planimeter. Variations in area ranged from 0.40 mm2 to 9.84 mm2. The large individual variations are pointed out. The advisability of catheterization in cases of noncommucating hydrocephalus is discussed. LITERATURE CITED BAILEY, P. 1949 I n : Yearbook of Pieurosurgery: 454-455. The Year-Book Publishers, Inc., Chicago, Ill. BECITETT,R., M. NETSKP AND H. M. ZIMNERNAN 1950 Developmental stenosis of the aqueduct of Sylvius. Am. J. Path., 26: 755-787. BICKERS, D., AND R. ADANS 1949 Hereditary stenosis of the aqueduct of Sylvius. J. Neuropath. Exp. Neuro., 8 ; 104-105. 1949 Hereditary stenosis of the aqueduct of SyIvius as a cause of congenital hydrocephalus. Brain, 78 ; 246-262. BROMAN, I. 1927 Manniskans utveckling fore fodelsen. C. W. K. Gleerups forlag, Lund. GERLACH,J. 1858 Mikroskopische Studien aus dem Gebiete der menschlischen Morphologie. Verlag von Ferd. Enke, Leipzig. 162 GRAY,H. G . FLYGER AND U. HJELMQUIST 1954 Textbook of Anatomy (ed. 3 1 ) : 988. Longmans, Green Company, London. HANSEN,F. C. C. 1907 Om efterfixering af formolpreparater. Saertryk ur Hospitalstidende, 21. 3-10. Kobenhavn. HASSIN,Q. B. 1948 Cerebrospinal fluid ( i t s origin, nature and iorrnation). J. Neuropath. Exp. Neurol., 7 : 172-187. HAGGQVIS’P, G. 1936 Analyse der Faserverteilung i n eineni Riickenmarkquerschnitt (Th. 3 ) . Ztschr. f. mikr. anat. Forsch., 39: 1-34. LEKSELL, L. 1949 A surgical procedure f o r atresia of the aqueduct of Sylvius. Acta Psych. e t Neur., 2 4 : 559-568. LINDGREN,E., A N D G. DI CHIRO 1953 The roentgenologic appearance of the aqueduct of Sylvius. Acta Radiologica, X X I X : 117-125. N O R L ~ NG. , 1949 Surgical treatment of iiioperablc tumors causing obstruction of Sylvius aqueduct. Yearbook of Neurosurgery: 454-456. RAUBER-KOPSCH 1955 Lehrbuch uiid Atlas der Anatomie des Menschen (ed. 1 9 ) , I I : 342-343. Verlag von Georg Thieme, Leipzig. ROMEIS,I). 1948 Mikroskopische Technik (ed. 15). Leibnitz Verlag, Miinchen. RUSSEL, D. 1949 Observations on the pathology of hydrocephalus. Medical Research Council, no. 265. SJOQVIST, 0. 1937 Beobachtungen iiber die Liqvorsekrption beim Menschen. Zbl. Neurochir., 1: 8-18. SPILLER, W. G. 1916 Syringomyelia. Syringoencephalomyelia. J. Nerv. Ment. Dis., 4 4 : 395-414. STOOKEY, B., AND J. E. S C A R F F 1936 Occhsion of aqueduct of sylvius by neoplastic and nou-neoplastic processes with rational surgical treatment f o r relief of resulting obstructive hydrocephalus, €3~11.Neuro. Iiist. h’ew Pork, 5 : 348-377. TURKETYITSCH, N. 1935 Die Entwicklung des Aqueductus cerebri bei Menschen. Morph. Jahrhuch, 76: 421-477. - 1936 La constitution anatomique d e I’aqucductus cerebri cle l’honime. Arch. Anat. Strasbourg, 21 : 323-358. TORXILDSE:N, A. 1933 Cit. from Woollam. - 1947 Ventriculoristernostomy. A palliative operation i n different types of non-communicating hydrocephalus. Johan Grundt Tanuni Forlag, Oslo. VILLIGER,E. 1922 Gehirn und Ruckenmark. Verlag von Wilhelm Engelmann, Leipzig. WOOLLAM, D. 1952 Casts of the ventricles of the brain. Brain, 7 5 : 259-267. WOOLLAM, D., AND J. MILLEN 1953 Anatomical considerations i n the pathology of the stenosis of the cerebral aqueduct. Brain, 7’6: 104-112.