A systematic autopsy survey of human infant bridging veins

In early childhood, subdural haemorrhage (SDH) can occur through accidental and non-accidental mechanisms [1‐4] as well as from birth-related trauma [5]. One of the proposed non-accidental mechanisms of injury involves a baby being subjected to repetitive rotational and acceleration/deceleration forces within the head due to violent shaking, with or without an additional impact force. The term abusive head trauma (AHT) is often used to include a pattern of injuries that are thought to occur through the inertial forces associated with a shaking event. Of these injuries, SDH is the most commonly observed feature and has been demonstrated in up to 92% of AHT cases [6].

The origin of infantile SDH has been debated throughout the scientific and medical literature [7‐9]. Historically, damage to the bridging veins has been the most commonly suggested source of subdural bleeding in AHT [10‐12]. However, previous experimental research on head injuries has centred around animal and biomechanical models [13, 14]. These studies have been used to estimate the amount of force required to tear bridging veins (which some authors state are large blood vessels [15, 16]) to magnitudes that would be far greater than that induced by inertial forces, such as those that can be produced by vigorous shaking. Direct evidence of ruptured bridging veins is also limited, demonstrated in only a few imaging studies by the extravasation of contrast agent [17] and during autopsy examination [18].

There are several detailed descriptions from neuroimaging and cadaver dissection studies on the locations, size and anatomical characteristics of adult bridging veins [19‐21]. However, there is scanty published research on infant bridging veins, which would be expected to be proportionally smaller than adult vessels. The most comprehensive studies on the anatomical locations of adult bridging veins can be found in neurosurgical books and papers as it is crucial for neurosurgeons to be aware of these veins during various surgical approaches, as damage to a bridging vein can result in post-operative brain damage [22‐24].

In this study, we map the number and anatomical locations of bridging veins from 43 human neonatal, infant and young child post-mortem examinations onto a 3D model of the surface of a representative infant brain. We also describe our observations of the drainage routes of these vessels, either directly into the dural venous sinuses or indirectly through venous pathways within the dural membrane. The mean diameter of 79 bridging veins from five additional cases within these age groups is also reported.

Post-mortem studies of the infant bridging veins are technically challenging due to their location below the dural membrane, as the infant dura is still attached to the calvarium at the sutures, rendering removal of the calvarium without damage to the bridging veins extremely difficult by conventional methods. When inspected at autopsy, bridging veins are fragile and easily disrupted by applying excessive stretching force during manipulation and under the force of gravity. Paediatric research is further complicated by the sensitivity of such studies, together with limited access to autopsy cases.

Light microscopy, X-ray, digital subtraction angiography and cadaveric anatomical dissections with vascular casting have previously been used to determine the diameter, number and anatomical locations of bridging veins [28‐30]. Variation exists in the diameter measurements reported in these studies, with values ranging between 1 and 7 mm, except when using anatomical dissections where veins as small as 0.1 mm have been infrequently recorded [19]. In a radiological study, the diameters were shown to vary considerably between individuals but not in any one individual [28]. To our knowledge, no studies have specifically detailed the physical characteristics of infant bridging veins, with the exception of one study in which outer diameter measurements of between 0.14 and 0.72 mm were recorded for six vessels [31].

Owing to the delicate nature of the bridging veins, we were only able to measure the diameters of the larger bridging veins in the more accessible areas (i.e. the frontoparietal parasagittal regions and superior cerebellum). Due to the collapse of the blood vessels and lack of blood flow, post-mortem measurements can only be considered to represent approximations of the true in vivo vessel diameters. However, it is likely that the stated values for diameter measurements are over-estimations due to the flattened appearance of the veins from collapse and also due to the necessity of photographically documenting the veins lying against a very thin wire or probe to ascertain the bridging section for subsequent measurement of vein diameters. Although more bridging veins were measured from the right hemisphere than the left, it is possible that this is due to our standard method for opening the dura overlying the right hemisphere first, rather than reflecting anatomical asymmetry in these cases. Despite these limitations, the reported values were still considerably smaller than those documented in the majority of adult studies [29, 30] but broadly correlate with the limited measurements recorded by Morison for infant bridging veins [31].

There is considerable variation in the number of bridging veins recorded between studies in the literature [19, 20, 30], with average values ranging from 29 veins in one study [19] to only four or five in another study [20]. However, these studies often refer only to the larger veins draining directly into the SSS. Differences in reported numbers may also result from the various methods of investigation, cohort numbers and ages of the participants. A further area of study for bridging vein locations is underneath the tentorium, where vessels often enter the left and right transverse sinuses [32‐35]. Neurosurgical papers also provide descriptions of bridging veins when determining the safety of several surgical approaches. However, these publications are usually focused on small, specific areas, often focused on the drainage of blood into a specific dural venous sinus [32, 36]. To our knowledge, this paper is the first systematic paediatric autopsy study to map all the bridging veins visible with the naked eye, in any one individual, over the entire surface of the brain. For this reason, our reported average number of infant bridging veins is considerably higher than values reported previously in studies in adult populations. It is probable that our average bridging vein count may nevertheless represent an underestimation owing to the technical difficulties of examining very small veins, documenting vessels located in cranial recesses with limited visibility and the necessity to incise the dura to gain access. However, we consider that our counts represent a best estimate of bridging vein numbers that could be identifiable by the naked eye under otherwise ordinary autopsy conditions. We have also demonstrated the considerable variation in number of bridging veins between individuals. Although our counts of bridging veins were lower in the AHT cases than the non-head injured cases, we cannot at this point determine whether this was due to the elastic recoil of small calibre broken veins, decreasing the likelihood of detection with the naked eye, or by high numbers of veins being obscured from view by the presence of SDH.

Our technique of using neurosurgical equipment and meticulous dissection methods have allowed us to demonstrate the anatomical locations of infant bridging veins, including vessels which ‘bridge’ to the dura distant from the sinuses and therefore demonstrating indirect routes of drainage into the larger dural venous sinuses. The locations of bridging veins recorded within this paper are comparable with previous publications [19, 21, 37], and although variations in the nomenclature used to describe these vessels exist in the published literature, they have previously been grouped as: a superior sagittal group, a sphenoidal group, a tentorial group and a falcine group, depending on which dural venous sinus they drain into [38].

Tentorial sinuses (sometimes described as meningeal veins or lakes) have been well described in the literature as constant venous channels [35, 37, 39]. Bridging veins from both the cerebellum and cerebrum may enter the tentorial sinuses which then drain into the transverse and straight sinuses including at the point of the transverse-sigmoid junction [40]. However, we observe that bridging veins also appeared to join venous channels either loosely attached to the dura or within the membrane itself, often in the parasagittal regions. Although not often alluded to in the literature, this indirect drainage route to the sinuses has been noted previously by some authors who describe ‘adhesions’ of the bridging veins to the dura [19] and venous sinuses (sometimes referred to as meningeal veins) [30, 38, 41] extending out by more than a centimetre from the SSS. In several more historical papers, as well as being described as either attached or within the dura, bridging veins are suggested to be free in the subdural space for up to 3 cm [42, 43], which one author describes as being particularly noticeable in the infant [42]. Through microscopic examination, the subdural space has been shown to be a potential space as the dura and arachnoid are continuous with one another [44]. A space is suggested to be created by the collection of fluid or air which cleaves through a weak cell layer of the dura [45] known as the dural border cell layer [46]. During our observations of the bridging veins, upon reflection of the dura, we create a subdural space; therefore, we cannot say how much of the free bridging sections of the veins were or were not within the dura. It is possible that the veins which appeared loosely attached to the dura may have occupied the dural border cell layer, disrupted by reflection of the membrane.

As bridging veins are the most commonly suggested source of infantile SDH, detailed descriptions of these blood vessels in the young are highly important for medical professionals involved in infantile head injury cases. We have shown that not all bridging veins are large calibre vessels and that some vessels are extremely small. They also appear to be delicate as they are easily disrupted on reflection of the dura.

As shown within this paper and others [20, 32, 37], the cerebral drainage system is extremely variable from one individual to the next. Whether this variation in vasculature could result in one infant being more susceptible to traumatic head injury and subdural bleeding than another cannot be answered here, but this point is worth future consideration, especially given the wide range of clinical outcomes seen in AHT cases. Anatomical information, including the data within this paper, may aid future understanding of the potential mechanisms which result in SDH and confirm whether or not bridging veins can be traumatically damaged by inertial/rotational forces to produce this type of bleeding. These data could also provide biomechanical engineers with the opportunity to create more detailed computational models of human infantile head trauma, accidental or otherwise.