Spina bifida, the normal, the pathological and the in-between: first evidence from a forensic osteological collection

Personal identification has always been an important issue in human societies, particularly the identification of human remains, which has gained even greater relevance in recent times [2]. In 1996, the universal right to be identified after death was recognized by the General Assembly of INTERPOL [3].

In the processes of identifying human remains there are a number of requirements, both biological and technical, that a scientific method must fulfill, to be considered valid, such as being based on characteristics that are “unique and immutable” (2, p. 401). At the same time, scientific methods must be testable, replicable, reliable, and scientifically valid [4, 5].

Forensic Anthropology (FA) contributes to the identification process by analyzing the biological profile, post mortem interval, traumatic injuries, aspects of the circumstances of death and individualization factors [4, 6]. These are characteristics that last during decomposition and are recognizable after death, representing permanent changes in the morphology of the skeleton and reflecting events in the individual’s life, such as cultural modifications, trauma, surgical interventions or pathological conditions [7, 8]. For a morphological feature to be considered an individualization factor, it has to occur with a low enough frequency (< 10%), and it has to be easily observable in imaging exams and dry bone [9, 10]. The value of these characteristics also depends on the existence of medical records, such as ante mortem radiographs, which allow comparison with post mortem observations. Distinguishing between simple anatomical variation and pathological change requires an in-depth knowledge of osteology and extensive experience in observing skeletal variation and common anomalies [9, 10].

The spectrum of skeletal manifestations associated with Spina bifida (SB) is a good example of the challenges when distinguishing between normal and pathological in dry bone. SB is a congenital disease that consists of the incomplete fusion of the vertebral arches, associated with a failure in the development of the neural canal [11]. It can occur in one or more vertebrae and in any region of the spine, more commonly in the lumbar and sacrum. Like other congenital diseases, the etiology of SB is multifactorial, involving the interaction of genetic and environmental factors [12]. Several studies have found evidence that SB is associated with a maternal folic acid deficiency during the embryonic period [13, 14].

The neural tube is the brain’s precursor to the spinal cord, which during the embryonic period develops through a process called neurulation, that is divided into a primary and secondary phase [12]. Defects that occur during primary neurulation result in open lesions, that is, they are not covered by normal skin [1]. Anomalies that may result from this stage of development are anencephaly (the absence of a brain) and meningomyelocele, or in rarer cases myelocele. In addition to the neurological deficiencies that arise in the spinal cord, meningomyelocele is accompanied by SB cystica that manifests in vertebrae with thin pedicles, deformed and/or absent laminae, underdeveloped spinous apophyses, and the edges of the bone anomaly are elevated by the protruding cystic mass [1].

When defects occur during the secondary neurulation, the resulting lesions are closed, meaning they are covered by normal skin, and are accompanied by SB occulta, as is the case with meningocele. This anomaly occurs most frequently in the lumbosacral region. The associated symptoms vary depending on the expression of the lesion [1].

The presence of cleft vertebral segments may not be associated with a neural tube defect (NTD), as the other conditions mentioned so far. This failure of the vertebral arch to fuse without the occurrence of neural tube defects (NTDs) is often considered a simple anatomical variant, which occurs mostly in the sacrum, without any clinical implications and is more common than cases involving NTDs [1].

It is important to emphasize that there is great divergence in the literature regarding how NTDs and SB are defined and classified. This discrepancy, both in the theoretical framework and in the practical context, is a problem recognized by several researchers [12, 15‐19]. Differences in the nomenclature used, classification systems and methods of analysis lead to incomparable results, as illustrated by the very high variation in the prevalence of SB occulta obtained in different population studies (between 1.2% and 50%) [15, 16]. An author who sought to establish a method to standardize the analysis of SB was Barnes [1] by separating the anomaly that occurs in the neural tube and the bone lesion that is observed in the spine, considering these two phenomena as distinct, but intrinsically associated. NTDs such as meningomyelocele or meningocele are congenital anomalies that occur in the nervous system while SB refers to a lesion that occurs in the skeletal system. The occurrence of SB may be a consequence of a NTD, which in Barnes’s view is the case of both SB cystica and occulta, it may also be an isolated phenomenon. In this case, the author suggests the name of “Cleft Neural Arch” (CNA). This condition usually affects only one or two vertebral segments in the transition regions between different types of vertebrae, and in certain cases it may extend to the entire dorsal sacral wall. Mulhern and Wilczak [18] also chose a similar terminology, “Complete Cleft Sacra” instead of SB, to specifically describe the condition they observed in skeletal remains from Native Americans.

The terminological disparity exposed here is reflected in a central point of discussion within the topic of SB, which is the debate on the clinical importance of SB occulta. Many authors do not distinguish between SB occulta and CNA using the first formulation to refer to both [12, 17, 19]. However, SB occulta has associated symptomatology and therefore clinical importance, whereas CNA does not [1].

Identifying the presence or absence of a NTD and therefore SB through the observation of dry bone is very difficult or even impossible. Both NTDs and the CNA frequently occur in the lumbosacral region, and both can result in a completely open sacrum [1]. In NTDs the medullary canal is widened, pushing the edges of the bony cleft outwards, comparatively, when there is no NTD the medullary canal remains open, but it is not enlarged and the edges are not raised [1].

Paleopathological studies mainly analyze SB occulta in the sacrum, because, in general, in an archaeological context, when SB is found in an adult skeleton, it can be assumed that it was not a case of SB cystica, due to the low life expectancy associated with it [17]. Still, there are reported paleopathological cases of SB cystica. At the Windover archaeological site (Florida, USA), an individual aged 15 years old was found with pathological changes possibly associated with SB cystica [20]. Garralda et al. [21] identified an opening in the L5 vertebral arches and sacral vertebrae in a child aged 8–9 years old, from a necropolis in Cathedral of El Burgo de Osma cloister (Soria, Spain). This lesion, together with the presence of hydrocephalus led to the diagnosis of SB cystica with meningomyelocele. These cases demonstrate the importance of a detailed analysis of the entire skeleton and differential diagnosis as essential tools to final conclusions. The presence of other skeletal anomalies beyond those occurring at the sacrum can contribute to the most correct classification of the sacral lesions observed [17].

The relevance of SB to FA and its value as an individualizing factor cannot really be ascertained within the current theoretical framework, as no method of classifying this pathology can fit the aforementioned criteria, to be considered valid, while there isn’t an agreement amongst researchers, in the clinical and paleopathological domains, on its definition and classification. The following cases exemplify this well. In a study on the relevance of discrete traits for FA, Verna and co-authors [10] found eight traits having a prevalence lower than 5%, including SB occulta. The authors consider SB occulta to be an asymptomatic form resulting from an ossification failure during the posterior arch fusion. In another similar research, which investigates whether certain morphological characteristics are rare enough to constitute individualization factors, Komar and Lathrop [8] consider that SB occulta has minimal value as an individualizing characteristic, since it would not be detectable during life, by the individual or his family. The authors do not specify what they consider to be SB occulta, but refer to it as an anomaly, which suggests that they define it similarly to Verna et al. [10]. Therefore, it is important to establish a classification system/method approved and applied by the entire scientific community, followed by a review of past studies to standardize the existing data.

The main goals of this research were to highlight the existing divergences in the literature regarding SB, propose a system that allows the analysis of SB in an objective and replicable manner and apply it in the 21^st Century Identified Skeletal Collection of the University of Coimbra, arguing the relevance of this condition in FA. To our knowledge, this is the first systematic work of this kind, in an identified collection with forensic relevance.

A sample derived from the 21^st Century Identified Skeletal Collection from the University of Coimbra (CEI/XXI) was studied [22]. Housed at the Laboratory of Forensic Anthropology (LFA) of the Department of Life Sciences at the University of Coimbra, Portugal, the skeletons from CEI/XXI originate from the Capuchos Cemetery in Santarém, Portugal [23, 24]. It is currently composed of 302 adult individuals of both sexes (162 females and 140 males), who died between 1982 and 2012 and were exhumed between 1999 and 2016. The ages at death range from 25 to 101 years old [24].

The studied sample comprised 209 skeletons. The proportion of females (57.9%; n = 121) is significantly higher (χ² = 5.211; d.f. = 1; p = 0.022) than that of males (42.1%; n = 88). The 207 individuals with age at death information were divided into six age groups (Table 1), which ranged between 44–99 years (mean = 80.73; median = 82; SD = 10.059).

In addition, a set of 31 sacra (training sample) from the Unidentified Osteological Collections of the Department of Life Sciences of the University of Coimbra (see Bárrios, [25]) were randomly selected for training purposes. From this sample, the parameters to be applied in the CEI/XXI were developed and the inter- and intra-observer errors were calculated using Cohen’s Kappa. 

The sacra were macroscopically analyzed considering the state of preservation according to Buikstra and Ubelaker [26] (>75%, complete; 25% – 75%, partially complete; <25%, incomplete), the number of foramina and the opening of the median sacral crest, both proximally and distally. The remaining vertebrae were inspected for gross morphological changes, especially those located at the posterior vertebral (neural) arch.

The exact anatomical proximal and distal limits of the median sacral crest were determined having as topographical references the associated foramina pairs and the respective sacral vertebrae involved. In those individuals who presented complete opening of the sacral canal, a more detailed morphological analysis was done, in order to differentiate between SB cystica, SB occulta and CNA, based on criteria proposed by Barnes [1]. According to this author, the manifestation of SB cystica associated with meningomyelocele usually involves several vertebrae and occurs mostly in the lumbosacral region. In the unfused vertebral arches, the pedicles are thin, the laminae are deformed or absent, the spinous process does not develop, and the edges of the skeletal defect are raised by the cyst. Similarly, SB occulta associated with meningocele occurs predominantly in the lumbosacral region. It results in pedicles and vertebrae with a wider appearance and is distinguished from SB cystica by manifesting in a smaller and narrower lesion that may be fusiform when several vertebrae are affected. Finally, CNA is characterized by not being associated with a NTD, unlike the conditions mentioned above. Thus, the edges of the bony defect are not raised, and the spinal canal remains normal. It usually involves one or two vertebral segments, in the transition regions between different types of vertebrae, but it can also cause fusion failure of the sacral dorsal wall [1]. As SBC and SBO occur in association with a NTD, these are accompanied by other neurological manifestations in the rest of the body, namely different levels of paralysis depending on the site and extent of the neurological defect, gait disturbances and foot deformities, among others [1]. Therefore, the complete skeleton of these individuals was analyzed in order to check for evidence of additional bone changes that could be associated with the observed defect in the sacrum towards a differential diagnosis, if needed. The data collected was systematically recorded (Excel, Microsoft) and subjected to a statistical study (IBM/SPSS Statistics version 21.0). Expected and observed frequencies in the categories under analysis were evaluated using Chi-square test (χ²), including Yates corrected Chi-square in the case of 2 × 2 contingency tables and Fisher’s exact test when Chi-square basic assumptions were not fulfilled. Correlation analyses were performed using Spearman’s coefficient. Binary logistic regression was used, with p values ​​determined by Wald’s chi-square test [27]. Statistical significance was considered for p < 0.05.

The inter- and intraobserver errors were calculated using Cohen’s Kappa (Table S1). In general, there were no noticeable discrepancies, either in the inter- and intraobserver errors, except for the variable proximal point of the sacral crest (Table S1). In this variable, the differences between observations arise because one observer (MTM) considered the proximal origin of the sacral crest between two vertebrae, while observer two (VMJM) had the sacral vertebra as reference and not the intervertebral spaces.

The 209 observed sacra were complete (52.2%, n = 109), partially complete (27.3%, n = 57) or incomplete (20.6%, n = 43). Figure 1 illustrates the different preservation states found in the CEI/XXI.

The observation of the most proximal point of the sacral crest, which corresponds to the location where it starts, was possible in 181 of the 209 sacra under study. This point varied between S1 (71.3%[129/181]) – or L5 (5.5%[10/181]) in those cases presenting sacralization of the last lumbar vertebra – and S3 (0.6%[1/181]). The sacral crest started at the S2 level in six sacra (3.3% of 181). This wide amplitude demonstrates the great variability of this sacral anatomical element.

Distally, the opening of the sacral canal comprised the S4 and S5 vertebrae in most individuals (71.2%[111/156]). This is usually known as the sacral hiatus, which in some individuals also included the S3 (9.0%[14/156]) and in others was limited to the S5 (17.3%[27/156]).

The complete opening of the sacral canal was detected in only 4 individuals (2.6%[4/156]), and in one of these (CEI/XXI_106) the opening of the posterior arches extends to L5 (see Fig. 2). Of these 4 individuals three are men (75%[3/4]) (Table 2). Although the prevalence recorded in males (4.4%[3/68]) is higher than in females (1.1%[1/88]) (Table 2), no significant differences between sexes were found (Fisher exact test: p = 0.318). Each individual, of those with complete opening of the sacral canal, belongs to a different age group, as Table 3 shows with the lowest age at death being 65 years old (CEI/XXI_185) and the highest 90 years old (CEI/XXI_106).

The applied binary logistic regression model (see Table S2) shows that the independent variables, namely sex (χ2 Wald = 1.494; p = 0.222) and age at death (χ2 Wald = 0.050; p = 0.823), are not significant predictive factors for the complete opening of the sacral canal (dependent variable).

When crossing the data obtained in the analysis of the most proximal point of the sacral crest with those of the opening of the sacral canal, it was verified that, in addition to the four individuals mentioned above, two others have very reduced sacral crests, which extend for only one or two vertebral segments (Table 4). In particular, individuals CEI/XXI_126 (male; 78 years old) and CEI/XXI_265 (female; 79 years old), whose sacral crests cover only vertebrae S3 and S4 and S2 and S3, respectively. In these individuals, there is, in fact, a fusion of the vertebral arches and, at the same time, the lack of it, both proximally and distally.

This is the first systematic study of SB and CNA in an identified collection relevant to FA. Complete sacral cleft was found in four individuals (2.6%) from the CEI/XXI. Although there were no significant differences between sexes, the statistical study showed that men are four times more likely to have complete sacral cleft. The presence of SB was not found in any of the individuals, and the bony defect observed in each of them was considered to be a CNA. Altogether in the sample, 11 individuals had CNA (7.05%).

Therefore, instead of considering a range of variation in the prevalence of “spina bifida occulta” between 1.2% and 50%, as reported in the literature, we sought to compare the prevalence of complete sacral cleft with studies whose results were compatible, leading to a variation between 1.2% and 2.7%. SB itself is even rarer. Thus, both complete sacral cleft and the SB can be considered as individualizing factors in the context of personal identification and FA.

The continuation of the research of SB within the context of FA and paleopathology should focus on the standardization of methodologies and classification systems applied, mainly with regard to the distinction between the presence of SB occulta or CNA, in order to develop compatible and realistic data about their prevalence.