Literature Thesis Master Forensic Science
Gestational age
estimation based
on the long bones.
A literature review of the current methods and an evaluation if they are
interchangeable.
Willemijn Lolcama
10-2-2020
Student ID: 10788948
Supervisor: Bernadette S de Bakker, MD PhD
Examiner: Dr. A. E. van der Merwe
1
Abstract
When fetal remains are discovered, one of the most important tasks of a forensic specialist is to estimate the gestational age of the fetus. This is highly relevant because from 24 gestational weeks a fetus is considered viable, in which case the fetal remains are handled like a human corpse and forensic research will have to be done. One characteristic of a fetus that is used to estimate the gestational age is the length of the long bones; including the femur, humerus, tibia, fibula, radius and ulna.
Gestational age estimation can be done in three different situations: for living, in-utero fetuses, for deceased, ex-utero fetuses and for skeletonized fetal remains. For living, in-utero fetuses sonographic measurements are obtained using an ultrasound. These measurements in combination with known gestational age are used to create reference sets for estimating gestational age of other fetuses. From deceased, ex-utero fetuses the length of the long bones are measured using radiographic techniques, such as X-ray and CT. These measurements are used to create reference sets and to calculate regression formulas. To estimate the gestational age of skeletonized remains a few reference sets exist, which are based on skeletonized material.
Although the described techniques are all used to estimate the gestational age, there are several challenges that stand in the way of making them interchangeable. Measurements obtained by a sonogram can not directly be compared to the length of skeletonized remains or radiographically measured bones. Sonographic measurements are subject to errors, due to the fetus moving inside the uterus and the error rate of the sonogram itself. Also research has shown that the reference sets and regression formulas created based on radiographic measurements are not reliable enough to be used on skeletonized remains. A final challenge is that bones can shrink post-mortem. The level of shrinkage of all the bones on average decreased from 10.09% ± 2,67% at 4 month of gestion to 1,28% ± 0,55% in new-borns
At this moment the methods for gestational age estimation are not directly interchangeable and coming up with specific correction factors is not possible without further research. To create more reliable and interchangeable reference sets more fetuses must be analysed, to obtain more data about the length of the long bones.
2
Content
Abstract ... 1
Introduction ... 3
Gestational age estimation in in-utero fetuses ... 5
Gestational age estimation in intact remains of ex-utero fetuses ... 9
Gestational age estimation in skeletonized remains ... 12
Influence of skeletonization on the long bones of fetuses ... 14
Interchangeable methods ... 16
Discussion ... 17
Conclusion and recommendations... 18
List of references ... 20
3
Introduction
When the remains of a fetus are found an important step in the forensic investigation is to estimate the gestational age before death. The gestational age is counted from the first day of the last menstruation before conception and is also known as the menstrual age. This should not be confused with fetal age, which is counted from the day of conception, approximately two weeks later. Estimation of the gestational age is important because when a fetus is older than 24 weeks of gestion it is considered viable, which means that it is assumed that the fetus can survive outside of the womb. So when the remains of a fetus with a gestational age above 24 weeks are discovered, they will be handled as a human corpse and forensic investigation will be done. This will not happen when the remains are younger than 24 gestational weeks, because at that point the fetus has never been viable. Also, terminating a pregnancy after 24 weeks of gestation without severe medical reasons is considered a criminal offense (1).
Gestational age can be estimated using various methods based on different body parts of the fetus. One way to estimate gestational age is by using long bone length, which includes the diaphyseal length of the femur, humerus, tibia, fibula, radius and ulna (2). The diaphyseal part of a long bone is the main shaft of the bone in between the epiphysis as shown in Figure 1 (3). Since in fetuses the epiphyses are not yet fused with the diaphysis, the term long bone length in this paper will refer to only the diaphyseal part of the bone.
An important point to consider when estimating the gestational age is that there are different situations when this might be necessary. Gestational age estimation is relevant for living, in-utero fetuses, in which case reference sets for the length of the long bones are based on ultrasound measurements (in-utero) (4). However, for deceased fetuses
FIGURE 1-AN OVERVIEW OF A FULL GROWN FEMUR, ONE OF THE LONG BONES OF THE HUMAN SKELETON.THE MIDDLE PART, OR THE SHAFT OF THE BONE, IS CALLED THE DIAPHYSIS.THE TWO HEADS OF THE LONG BONES ARE CALLED THE EPIPHYSES, WHICH ARE CONNECTED TO THE DIAPHYSIS BY THE METAPHYSIS. THE EPIPHYSES CONTAIN SPONGY BONE, WHILST THE DIAPHYSIS IS HALLOW AND CONTAINS BONE MARROW, THIS IS CALLED THE MEDULLARY CAVITY. THE OUTSIDE OF THE LONG BONES IS MADE OUT OF COMPACT BONE (3).
THE LENGTH OF THE DIAPHYSIS CAN BE USED FOR GESTATIONAL AGE ESTIMATION.
4 other measurement techniques are needed. The remains of deceased fetuses can be found in different taphonomic phases (5). The remains can still be intact and fresh, completely skeletonized and dry and of course everything in between. Since the long bones are subject to shrinkage when they dry (6), a problem is that the different age estimation standards for long bones lengths might not be useable for all fetal remains that are found.
What set of standards is used depends on the institute that does the examination and the status of the remains (e.g. intact remains or dry bones).
We aimed to review all methods used to estimate gestational age based on the long bones and to evaluate whether these methods are interchangeable between different settings (in-utero, ex-(in-utero, skeletonized). The main research question is: “Which methods based on
the long bones are used for estimation of gestational age and are they interchangeable between different settings?” To answers the main research question the different situations
in which gestational age estimation happens will be described as well as the corresponding techniques that are used. The influence of post-mortem bone shrinkage will be analysed and finally it will be assessed if the different techniques are interchangeable.
5
Gestational age estimation in in-utero fetuses
Sub question 1: What methods based on the long bones are used for gestational age estimation in in-utero fetuses?
Estimation of the gestational age happens during pregnancy to determine when the baby will come to full term. A normal pregnancy last on average 40 weeks, or 10 lunar months, counted since the first day of the last menstruation period. The first four gestational weeks are called the germinal phase, from five to tenth weeks is the embryonic phase and the final phase until birth, also called to term, is called the fetal phase (7). To determine in which phase, or more specifically which gestational week a fetus is the lengths of the long bones can be used. The long bones include the femur, humerus, tibia, fibula, radius and ulna. After 10 weeks of gestation it is possible to obtain sonographic measurements of the long bones from a fetus (4). This is when the femur starts to resemble the shape of the final adult bone (8).
Human bones are created by ossification, which is the process of bone formation. The process of bone formation starts around the 8th gestational week and continues until a human is fully grown (around 25 years). The flat bones of the skeleton, including the skull, clavicle and most of the cranial bones, are formed by intramembranous ossification. Intramembranous ossification happens when the mesenchymal stem cells are directly transformed into bone tissue. The long bones and the axial skeleton are formed by endochondral ossification. With this type of ossification the mesenchymal tissue is first transformed in a cartilaginous intermediate, which is later transformed into bone (9). As can be seen in Figure 2 the early cartilage model begins to thicken as a layer surrounds the cartilage, which start to mineralize and form periosteal bone around the middle area of the cartilage (7). Growth plates start to form and this allows the bone to grow. The bone turns into the final shape of the long bone from which it will keep on growing until the human is full grown.
6 FIGURE 2-THE EARLY STAGES OF LONG BONES DIAPHYSEAL OSSIFICATION.LONG BONES ARE FORMED BY ENDOCHONDRAL OSSIFICATION.THE PROCESS STARTS WITH A CARTILAGE MODEL THAT WILL START TO GROW IN LENGTH BY CELL DIVISION.A PERIOSTEAL COLLAR STARTS TO FORM AROUND THE CARTILAGE, WHILST THE MATRIX STARTS TO CALCIFY.THE PERIOSTEAL COLLAR TRANSFORMS INTO BONE SURROUNDING THE CARTILAGE AND GROWTH PLATES START T O FORM (7).
As mentioned before, it is possible to make sonographic measurements of the long bones after the 10th week of gestation. Ultrasound is a non-invasive method where sound waves
with a very high frequency are send into the skin. The waves then bounce back when they reach organs or the fetus. These reflected echoes are send back to an ultrasound scanner where an image of the internal situation will be created (10). By adjusting the settings of the ultrasound different parts of the fetus can be visualized, in Figure 3 the femur of a fetus in the first trimester is visible (11).
7 Using ultrasound imaging has advantages over other imaging techniques; besides being portable and relatively inexpensive, ultrasound does not give forth ionizing radiation (12). Because of this it is safe to use this technique on in-utero fetuses. Multiple papers are written where the authors analysed sonographic measurement of in-utero fetuses to create mathematical models that can predict the gestational age of other fetuses (13). For these studies the pregnant women of whom the fetus was investigated were all screened, and several groups were excluded. Women with maternal diseases that could affect normal fetal growth, women with multifetal pregnancy, pregnancies where dwarfism runs in the family and women who smoked were excluded (13, 14).
In 1980 Queenan, O’Brien and Campbell did a study with 41 pregnant women, with known gestational periods based on their last menstrual period. Every one to three weeks the long bones of the fetuses were measured using ultrasound. For every long bone the length of the bone was plotted against the weeks of gestation. The femur had the largest growing curve and measurements of this bone showed the smallest weekly standard deviation. They concluded that from 12 to 22 gestational weeks the long bones grow linear, so accurate age estimation could be done at this stage (15).
In 1981 O’Brien and Queenan did two other studies about using long bones for gestational age estimation. The first study was about assessing if the measurements of the femurs in a fetus could accurately and reproducible be done by using ultrasonic measurements, which was tested by doing 30 experiments. The mean standard deviation of these experiments was 0.8 mm, which was prove that the measurements were reproducible. Also they used 47 patients to study the accuracy of the method, which resulted in 95% confidence limits of ±6,7 days (16). They suggested that the use of the femur length could be added as an extra parameter in the estimation of gestational age, but that more research should be done before the femur length could be the only indicator of gestational age.
That same year O’Brien and Queenan used the previously described measuring technique to establish the average femur length based on ultrasonic measurements. They studied 411 pregnant women, from 14 to 40 weeks of gestation which was known based on their last menstrual period. The result of this study were graphs showing the ultrasound fetal femur length compared to the weeks of gestation (17). The authors claimed that their measurements of the femur are the first reproducible set of ultrasound measurements throughout pregnancy. The results of their measurements can be seen in Figure 4.
8 FIGURE 4 - GRAPH OF THE MEAN ULTRASOUND FEMUR LENGTH ± 2 SD FOR EACH WEEK OF GESTATION FROM 14 WEEKS TO TERM BASED ON THE DATA FROM O’BRIEN AND
QUEENAN (1981).ON THE X-AXIS THE NUMBER OF WEEKS OF GESTATION IS SHOWN, ON THE Y
-AXIS THE ULTRASOUND FETAL FEMUR LENGTH IN MILLIMETRES IS SHOWN (17).
Jeanty et al. (1981) measured the femoral length, the humeral length and the biparietal diameter (BPD) with an ultrasound and correlated these lengths with the gestational age in 450 normal pregnancies. The gestational age of these pregnancies was based on the crown-rump length. The correlation between femur length and gestational age as well as the correlation between the humerus and gestational age was 0,95 (18). This resulted in them doing another study, where the tried to calculate the gestational age based on the long bones.
In the second study, conducted in 1984, the authors referred to O’Brien et al. (1981) who suggested only using the femur length. So for this second study they used different long bone lengths as observed value and gestational age as estimated value. From all the long bones the estimations based on the femur length seemed to best the best fit, however the measurement of the biparietal diameter for the age estimation seemed more accurate (19). The authors suggested that using more than one long bone for the age estimation would result in a more reliable estimation of the gestational age (19).
In 1982 Hadlock et al. designed a study in which measurements of one femur of 338 fetuses were compared to the gestational age. The technique previously described by O’Brien was used to align the ultrasound (16). Part of this technique is taking multiple measurements and using the longest one, since tangential sections tend to shorten the
9 measured length. Based on the measurements several models were explored and the optimal model turned out to be a linear quadratic function. The variability in predicting the gestational age from the femur length was ± 9,5 days for fetuses between 12-23 weeks and ± 22 days for 23-40 weeks (13). One of the final conclusions from Hadlock et al. was that the measurements of the femur could give a good estimation of the gestational age for fetuses between 12 and 23 weeks. However for fetuses of 23 weeks and above the variability of the data is too large.
A problem with gestational age estimation based on ultrasound measurements is that the variability between 23-40 weeks can be up to several weeks (13). This results in a large age-interval which is unpreferable. Also, mathematical modelling showed that the femur growth curve is nonlinear after the 23rd week, which makes it harder to get a fitting model.
Another challenge is that there is a difference between the sonographic measurements of the long bones and the actual length of the bone when it is dissected (20). There can be a ± 2 mm error rate due to the resolution of the ultrasound technique (21). According to Cronk (1980) the dimensions that an ultrasound can measure depend on the gestational age of the fetus and the quality of the image that is created (22). Also the fact that a fetus can move around in the uterus can result in errors.
Gestational age estimation in intact remains of ex-utero
fetuses
Sub question 2: What methods based on the long bones are used for gestational age estimation in intact remains of ex-utero fetuses?
When intact fetal remains are discovered several reference tables and regression formulas, based on growth standards, are available to estimate the gestational age. Collecting data from deceased fetuses with a known gestational age helps to create a model for age estimation (2). For this examination radiographic measurements are taken using X-rays. This can be done by using conventional radiography (X-ray) or a computed tomography scan (e.g. CT-scan or Micro-CT).
10 X-rays are electromagnetic radiation waves with a higher energy level
than for example visible light. X-rays can travel through the body, where some of the energy will be absorbed by the tissue and bone it passes depending on the radiological density of the material. On the other side of the body an X-ray detector will show an image based on the contrast between the absorbed X-rays. Since bones have a high radiological density and thus absorb most of the X-rays, they will appear white against the black background of a radiograph. Using conventional radiography a clear two-dimensional image, see Figure 5 (23), of the bones can be created (24).
With a CT-scan a three-dimensional image of the bones can be created which will have a more accurate visual of the length of the bones. With the CT-scan a beam of X-rays is aimed at the subject and quickly rotates around it. A computer will process the signals that are produced to create cross-sectional images of the body. Digitally a three-dimensional image of the subject is formed which will contain more details about the subject than the conventional X-ray (25). Since the extremities in a fetus can have different positions it can be more reliable to make a CT-scan to get the most accurate measurements of the long bones. Adalian et al. described in 2001 how the long bones of a fetus should be measured using radiographic techniques. According to the authors the fetus should lay directly on the table, with its head and limbs fixated. The legs had to be half bent to prevent superposition, see Figure 6. The images were taken from a frontal and a lateral view. FIGURE 5-X-RAY IMAGE OF A FEMUR AT 16 GESTATIONAL WEEKS (23).
FIGURE 6-X-RAY TAKEN FROM THE FEMURS OF A FETUS.THE FETUS LAID DIRECTLY ON THE TABLE, WITH ITS HEAD AND LIMBS FIXATED. TO PREVENT SUPERPOSITION THE LEGS ARE HALF BENT.THIS X-RAY IS TAKEN ACCORDING TO THE PROTOCOL DESCRIBED BY ADALIAN ET AL.(20).
11 Using data about the length of the long bones in relation to the gestational obtained from radiographical imaging techniques is preferred over ultrasound measurements, since the data is more reliable. Also, the use of radiology is easier when the remains are still intact or contain soft tissue (2). Besides, the ultrasound data, as mentioned in the previous chapter, is not representative for the real anatomical lengths of the bones, due to the measuring errors. Another problem with ultrasound measurements is that they are taken when the fetus is alive inside the womb. The movement of the fetus can result in errors in the measurements which can result in incorrect reference sets (20).
In 2001 Adalian et al. published a study where they assessed a precise and easily reproducible method to create new fetal growth standards (20). The reason for this study was that most of the already existing reference tables were based on ultrasound measurements which might not be fitting for ex-utero fetuses. For the study 498 fetuses were selected, who died for various reasons. Using radiography they measured the diaphyseal length of both the femurs in situ. To test whether the measurements done by X-ray were accurate, the femurs of 30 fetuses were dissected and measured ex situ using the same technique. The intra- and inter- variability was also tested. The average difference in measurements in situ and ex situ was 0,58 mm, which after statistical testing lead to a very significant pairing; there was no significant difference between measurements by X-ray in situ and ex situ. They concluded that these measurements techniques could reliable be used to create new fetal growth standards.
Carneiro et al. published a study in 2016 considering the application of forensics anthropological methodologies to skeletonized and non-skeletonized human remains. For the study they used the method of Adalian et al. (2001) to update the fetal radiographic data from fetal long bones to estimate gestational age. The research took place in Portugal and thus Portuguese population-based equations were created. Radiographs from 257 fetuses of know gestational age were taken and regression formulas for each long bone were obtained. They concluded that all long bones showed a very strong positive correlation between length and gestational age. Reference tables with age intervals of approximately three to four weeks of gestation were published and tested on 30 different fetuses. The models appeared very accurate although there was a small over-estimation of the gestational age. The regression formulas were also compared to formulas developed by Adalian, Scheuer et al. (1980) and Fazekas & Kósa (1978). The new models seemed to have a higher accuracy and a smaller bias.
In 2016 another study was conducted, this time based on the Mexican population by Chávez-Martínez et al. (26). In their article they mention that previous studies (such as
12 Adalian, Scheuer et al. and Fazekas & Kósa) did not use the right regression formulas to fit the growth curve of the long bones. So the aim of their research was to create standards to estimate fetal age in Mexico by using reliable and reproducible systems. For the study 97 fetuses were used who were examined three times using X-ray to reduce intra-observer errors. For every fetus the fetal age and the average length of the long bones was presented. Using statistical methods it was analysed if there were any significant differences between the characteristics of the fetuses and the measurements of the bones. Finally, a quadratic regression formula was created for each long bone. An important note was that the estimation of gestational age was hard since there are many factors influencing the development and growth of a baby. They considered using multiple regression formulas combined to increase diagnostic precision.
Gestational age estimation in skeletonized remains
Sub question 3: What methods based on the long bones are used for gestational age estimation in skeletonized remains?
There are three main methods for age estimation of skeletonized fetal remains. Age estimation can be done by analysing dental development, ossification levels of the remains and skeletal growth. Although dental development can be the most reliable, since dental elements are less influenced by racial and environmental factors (27), chances are that the teeth cannot be identified in case of skeletonized remains (28). In that case using the diaphyseal length of the long bones is a good option for estimating gestational age. To measure the diaphyseal length of the long bones a sliding calliper or an osteometric board, see Figure 7, can be used, depending on the size of the bone. The measured lengths can then be compared to reference sets created by different scientist to estimate the gestational age of the remains. According to the Human Bone manual by White and Folkens the procedure of age assessment should always be done using a reference set based on a related set of skeletal material (29).
13 In 1978 Fazekas & Kósa did a study were they examined and measured most of the bones of deceased fetuses, with an approximate gestational age of 3 lunar month to full grown fetus (10 lunar months). For their study Fazekas & Kósa used 138 Hungarian samples and the measurements were done directly on the skeletonized bones (31). The data obtained from this study was for a long time the only available reference set for the forensic community. A problem with this method is that the reasoning behind the data is circular (7). Since the gestational age of the fetuses was not known, the crown-heel length was used to group the fetuses based on their length in a half-lunar month interval. This estimated gestational age interval was used next to create a reference set for the length of the long bones at these moments. This resulted in a dataset where the long bone lengths were used as independent variables, although this is not the case.
Scheuer et al. (1979) published a paper in which they provided evidence that the long bones could be used to estimate the age of a fetus. In the paper they mention three methods for age estimation, two based on the long bones and one based on dental development. Since the method based on dental development is hard to use in the field they do not elaborate more on this topic. The long bones can be used in two different ways: directly asses the age based on the lengths, or by using the length of the long bones to estimate the crown-heel. The authors criticize the latter, since age estimation based on CHL is a two-step method which will not result in an accurate estimation because CHL is an estimations itself.
For their own study Scheuer et al. used material from two locations in the United Kingdom. From the Bristol Royal Hospital for Sick Children the data was used from approximately 17 fetuses, that had already been measured based on X-rays taken. This data was used to check the reliability of the second set. From the second location, The London University Institute of Child Health, radiographs were taken from premature live fetuses, of which the gestational age was known. Based on the data from the University age was regressed based on the length of single long bones and based on a combination of multiple long bones. Multiple linear regression, single linear regression as well as logarithmic regression formulas were published. The multiple regression formula containing all long bones had the lowest standard error of the estimation (1,88 mm) (32).
According to Scheuer et al. the number of subjects they used for their research, 17, was limited, which could have resulted in less accurate formulas. Also the fact that the measurements came from radiographic images in utero give rise to some concern, since it is hard to find the exact right position of the long bones in the uterus to take precise radiographs (32).
14 A more recent study by Thornton et al. (2019) assed if the femur is an useful bone to estimate the age of an immature individual in the African population (33). For this study 74 remains were used, of which the femurs were dissected and dried before they were measured. The measurements were done by a digital calliper. The measurements of the femur were done over the maximal length of the bone, so this is different than only the diaphyseal length that is normally measured. The results of the study included that since there are significant changes in the morphology of the bone, these measurement could be of use in the estimation of the gestational age. Thornton and colleagues were pleased that the measurements could be used as a reference set for age estimation in the African population instead of the European standard that were used before.
Influence of skeletonization on the long bones of fetuses
Sub question 4: What is the influence of skeletonization on the long bones of fetuses? The estimation of gestational age is relevant for all types of fetal remains that are found. There can be only partial remains present, decomposed remains, burned remains, mummified or skeletonized remains and so on. Since the remains are subject to post-mortem change, the standards created for age estimation based on intact remains might not be applicable for skeletonized remains. To investigate this is it is interesting to know if the length of the long bones may change post-mortem. According to Clark et al. (1997) bone will be broken down over the years, but will remain the same shape but become more brittle (5). Not many other studies have been conducted about this topic, however Angie Huxley in 1998 designed a study to investigate post-mortem shrinkage of the bones.According to Huxley the shrinkage can happen due to the dehydration of the bone matrix (6). For the study data from Petersohn and Köhler (1965) was used. They measured diaphyseal lengths from 490 fetuses, in their fresh state and when they were dry, calcinated bones. The long bones are sorted into groups of one lunar month based on the gestational age of when the fetus died. Although the original paper of Petersohn and Köhler is not available, their data is shown in the paper of Huxley. Based on this data Huxley calculated the shrinkage rate per long bone per lunar month. The analysis of the data from Peterson and Köhler resulted in the conclusion that the rate of shrinkage is different per long bone and per age. In the paper of Huxley the average shrinkage per lunar month is shown for each long bone; in Figure 8 the shrinkage rate for every bone on average shown. It appeared that the level of shrinkage of all the bones on average decreased from 10.09% ± 2,67% at 4 lunar months to 1,28% ± 0,55% in new-borns. The decrease of shrinkage rate is shown in Figure 9.
15 FIGURE 5-FEMORAL SHRINKAGE RATES FOR FETUSES BETWEEN 4-10 LUNAR MONTHS AND NEW-BORNS (6).
Based on the analysis of the data from Petersohn and Köhler it seems that the shrinkage rate decreases the older the fetus was at the moment of decease. This is due to the fact that the shrinkage happens because of loss of water and organic material. When the fetus starts growing more bone due to ossification, there is less water present and thus the bone can shrink less post-mortem. Finally Huxley advised to use a correction in the age estimation in the early lunar months, since the rate of shrinkage can be up to 13,85%, however in the older lunar months it might be less relevant to use a correction.
Another post-mortem modification that can occur is that the remains get burned. Cunha (2009) reviewed multiple papers where authors discussed the matter of burned remains (27). It was stated by Bradtmiller and Buikstra (1984) that when bones shrink due to burning it does not have a significant influence on the age estimation (34), however this study was not specifically focussed on fetuses. A later study from 2005 by Thompson claimed that the statement by Bradtmiller and Buikstra was premature and that more research was necessary (35).
FIGURE 4-COMPARISON OF SHRINKAGE RATES BY SKELETAL ELEMENT FOR FETUSES RANGING BETWEEN 4-10 LUNAR MONTHS (LM) AND NEW-BORNS (7).
16
Interchangeable methods
Sub question 5: Are the methods for gestational age estimation of in-utero fetuses interchangeable with the methods for estimation of intact or skeletonized remains?
As described in the previous chapters, there are many studies done focussing on the use of the lengths of the long bones for gestational age estimation. For in-utero fetuses mainly sonographic measurements are taken, whilst ex-utero radiographic methods are used to measure the long bones. The third way of measuring the long bone is when they are dissected and an osteometric board is used to directly measure the bones. All these methods have the same goal: assessing the age of a fetus. It would be very useful if all these standards obtained from these studies would be interchangeable between different settings. As previously mentioned, sonographic measurements can have a slight error compared to the real anatomical size, due to an error rate in the ultrasonic machine as well as the movement from the fetus in the uterus. Also ultrasound measurements are subject to inter- and intra- observability, which can result in a higher error rate. This causes the standards obtained from sonographic measurements of the long bones, especially the femur, to be slightly smaller than the actual anatomical length of the bone and thus these standards cannot be used on fresh bones nor dry bones (36).
In 2016 Carneiro et al. conducted a study were they created equations for gestational age based on the long bones. The equations were created using radiographic measurements obtained post-mortem from a known hospital sample in Portugal (2). In the final remarks of the article the authors claimed that the technique would be appropriate to use on dry bones as well as semi decomposed bones. This would suggest that the standards obtained from radiographic measurements of intact fetal remains would be interchangeable with the standards for skeletonized remains.
However, in 2019 Carneiro et al. conducted another study where they tested whether these equations could also be used on skeletonized material (37). The equations were assessed on a known set of fetal dry bones from the Granada collection in Spain. Since the equations are only suitable for fetuses, only 17 samples could be used for the assessment. These fetuses had a gestational age between five to nine gestational months. The Granada collection consists of all skeletonized material, so all the relevant bones were measured using a digital calliper, with an instrument error rate of 0.01 mm. These measurements were directly used in the relevant equations to predict the gestational age at death of the individuals.
17 After analysing all the results of the calculations Carneiro et al. had to conclude that although the study provided some effective results, with a smaller mean absolute error rate and bias than previous studies (Scheuer et al. (1980) and Fazekas & Kósa (1978)), the overall accuracy and unbiasedness compared to the reference set was not sufficient enough. They proposed that a new method based on solely dry bones should be applied because it will be more accurate, less biased and more easy to use in the forensic context (37). Another challenge with making the standards interchangeable is that research showed that there is a difference in the length of the long bones among ethnical groups (28, 38).
Based on the prior literature it seems that the different methods for age estimation are not directly interchangeable. The sonographic measurements are not accurate enough compared to the real anatomical measurements, and based on the study of Carneiro et al. (2019) the formulas established based on radiographic measurements are not reliable enough to use on dry bones.
Since there is a possibility of error in many stages during the measurements of the long bones it will be hard to come up with an standard correction rate that can solve the accuracy problem. Based on the shrinkage rate described by Huxley, a correction rate could be useful for the early gestational weeks, since there is a significant shrinkage rate. However this rate is not constant during the entire gestational period so the correction rate should be adjust based on the gestational age of the fetus. A clear problem with that is that it is impossible to adjust the correction rate based on the gestational age whilst that rate must be used to estimate the right gestational age, that seems like a paradox.
Discussion
Using the length of the long bones of a fetus to estimate the gestational age is a useful and promising technique. However, a few problems and challenges arise when assessing whether the correct techniques are interchangeable or even useful at all.
The first problem is that for every population or ethical group different reference sets or regression formulas are needed, since growth is population specific. In the ideal situation every population would have a specific standard, based on the exact same method. In that way all data from different populations could easily be compared and maybe even correction factors based on population could be calculated. Another solution for the difference in population would be to use a large sample set with subjects from different racial groups. In 2017 researchers claimed that a racial-neutral regression formula gave similar results when estimating the gestational age as a racial- specific formula (39).
18 Something else that came to light when reviewing all relevant literature is that the methods used by various researchers does not seem ideal. There is not an unlimited provision of deceased fetuses that can be used for research, which is why some of the studies had a rather small number of subjects for their research. This is of course also due to the fact that a significant percentage of the deceased fetuses had underlying medical reasons and could therefore not be used for the study.
What also seems problematic is the circular reasoning used by Fazekas & Kósa. The measurements and standard created by them are cited in almost every article concerning anthropological fetal age estimation, however it seems that the data they used is not very accurate. By using inaccurate data to support new data the accuracy might decrease more. The biases created by Fazekas & Kósa should be taken into account when creating new standard sets for age estimation.
Using ultrasonic measurements for dry bone analysis will be hard to arrange. Ultrasonic measurements are subject to a few variables that influence the accuracy which makes it hard to come up with one correction method to account for all those variables. Comparing ultrasonic in-utero measurements related to the gestational age with other in-utero ultrasonic measurements is a well described topic. Many articles are written about this technique and in hospitals the femur length can be used in-utero to estimate gestational age. But to use this data also on archaeological remains, more research is necessary. First the exact rate of error from sonographic measurements compared to the anatomical length must be determined and then also the level of post-mortem shrinkage should be taken into account.
The rate of post-mortem shrinkage is a field where possibly more research could be useful. The only researcher that wrote an original article about post-mortem shrinkage is Huxley. Every other mention of post-mortem shrinkage of bones is as a reference to the article by Huxley. This could however be because shrinkage does not occur in post-natal humans. In the study of Huxley shrinkage significantly occurs in the early gestational weeks, due to the high level of hydration in the bone.
Conclusion and recommendations
Gestational age estimation is a relevant forensic topic. After 24 gestational weeks a fetus is viable and should be treated accordingly as is stated in the law. The long bones of the fetus grow linear in between gestational week 14 - 22, and models have been created to also account for week 23 to term. Since fetal remains can be found in different taphonomic
19 states, different techniques and standards have been developed to fit the situation to as precise as possible estimate the gestational age.
For living, in-utero fetuses ultrasonic measurements are used since they are non-invasive and can give a clear image of the fetus. Based on these measurements compared with other ultrasonic standards the gestational age of a living fetus can be estimated. The ultrasonic measurements however, do not give accurate information about the actual anatomical length of the diaphysis bones, since an error of ± 2mm has been reported.
The use of radiographic techniques, like X-ray and CT, are used to measure the length of the long bones in intact, ex-utero, fetuses. Using X-ray a 2D image can be created and using CT a 3D image can be created. Although some researchers have said that the regression formulas created based on the radiographic measurements compared to the (known) gestational age are also applicable on dry, skeletonized remains, other studies show a significant difference between the results of the formulas and the known gestational age.
For skeletonized archaeological remains some standards have been created, where Fazekas & Kósa is one that keeps on being referred to. Most of these standard are quite old, and could be revised, also since the methods on which the standards are based are biased and created with circular reasoning.
The post-mortem shrinkage of bone is also a factor that should be taken into account when assessing if the methods for the different states of remains are interchangeable. In the early months of gestation bones can shrink up to 18,3 percent, whilst the bones of new-borns shrink only up to 1,48 percent according to Huxley.
So to conclude, the methods for gestational age estimation are at this moment not directly interchangeable and coming up with specific correction factors is not possible without further research. It will be useful to analyse as many fetuses, both radiographic and anthropological, to obtain more data about the length of the long bones. By studying the radiographic images and measurements in comparison with the anthropological data a correction factor for those techniques can be calculated. Finally statistical methods should be applied to analyse whether the shrinkage rate of the long bones has a significant influence on the age estimation especially in the case of early gestational age.
20
List of references
1. Artikel 82a Wetboek van Strafrecht, (2016).
2. Carneiro C, Curate F, Cunha E. A method for estimating gestational age of fetal remains based on long bone lengths. Int J Legal Med. 2016;130(5):1333-41.
3. Blaus B. Structure of a Long Bone Medical gallery of Blausen Medical 20142013 [ 4. Exacoustos C, Rosati P, Rizzo G, Arduini D. Ultrasound measurements of fetal limb bones. Ultrasound Obstet Gynecol. 1991;1(5):325-30.
5. Clark M, Worrell M, Pless J. Postmortal changes in soft tissues. Forensic
taphonomy: the postmortal fate of human remains. : CRC Press, Boca Raton, FL; 1997. p. 151–70
6. Huxley AK. Analysis of shrinkage in human fetal diaphyseal lengths from fresh to dry bone using Petersohn and Köhler's data. J Forensic Sci. 1998;43(2):423-6.
7. Scheuer L, Black S. The Juvenile Skeleton. London: Elsevier Academic Press; 2004.
8. Gardner E, Gray DJ. The prenatal development of the human femur. Am J Anat. 1970;129(2):121-40.
9. Breeland G, Sinkler M, Menezes R. Embryology, Bone Ossification. . In:
StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2020 Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK5397182020.
10. Ultrasound
https://www.nibib.nih.gov/science-education/science-topics/ultrasound: National Institute of Biomedical Imaging and Bioengineering (NIBIB); 11. Watson W, Seeds J. Diagnostic Obstetric Ultrasound. Glob. libr. women's
med2008.
12. Abuhamad A, Minton KK, Benson CB, Chudleigh T, Crites L, Doubilet PM, et al. Obstetric and Gynecologic Ultrasound Curriculum and Competency Assessment in Residency Training Programs: Consensus Report. J Ultrasound Med. 2018;37(1):19-50. 13. Hadlock FP, Harrist RB, Deter RL, Park SK. Fetal femur length as a predictor of menstrual age: sonographically measured. AJR Am J Roentgenol. 1982;138(5):875-8. 14. Chaithra R, Sunkeswari S, Kalghatgi R. The Study of Relation between the Gestational Age of Human Fetuses and the Diaphyseal Length of Femur Using Ultrasonography. International Journal of Anatomy and Research. 2017;5:3342-9. 15. Queenan JT, O'Brien GD, Campbell S. Ultrasound measurement of fetal limb bones. Am J Obstet Gynecol. 1980;138(3):297-302.
16. O'Brien GD, Queenan JT, Campbell S. Assessment of gestational age in the second trimester by real-time ultrasound measurement of the femur length. Am J Obstet
Gynecol. 1981;139(5):540-5.
17. O'Brien G, Queenan J. Growth of the ultrasound fetal femur length during normal pregnancy: Part I. American Journal of Obstetrics and Gynecology. 1981;141(7):833 - 7. 18. Jeanty P, Kirkpatrick C, Dramaix-Wilmet M, Struyven J. Ultrasonic evaluation of fetal limb growth. Radiology. 1981;140(1):165-8.
19. Jeanty P, Rodesch F, Delbeke D, Dumont JE. Estimation of gestational age from measurements of fetal long bones. J Ultrasound Med. 1984;3(2):75-9.
20. Adalian P, Piercecchi-Marti MD, Bourliere-Najean B, Panuel M, Fredouille C, Dutour O, et al. Postmortem Assessment of Fetal Diaphyseal Femoral Length: Validation of a Radiographic Methodology. Journal of forensic sciences. 2001;46(2):215-9.
21 21. Issel EP. Ultrasonic measurement of the growth of fetal limb bones in normal pregnancy. J Perinat Med. 1985;13(6):305-13.
22. Cronk CE. Fetal growth as measured by ultrasound. American Journal of Physical Anthropology. 1983;26(S1):65-89.
23. Pazzaglia UaDCaICaBMaGGaBM. Thanatophoric dysplasia. Correlation among bone X-ray morphometry, histopathology, and gene analysis. Skeletal radiology. 2014;43. 24. X-rays https://www.nibib.nih.gov/science-education/science-topics/x-rays: National Institute of Biomedical Imaging and Bioengineering; [
25. Computed Tomography (CT)
https://www.nibib.nih.gov/science-education/science-topics/computed-tomography-ct: National Institute of Biomedical Imaging and Bioengineering; [
26. Chávez-Martínez P, Ortega-Palma A, Castrejón-Caballero JL, Arteaga-Martínez M. Equations to estimate fetal age at the moment of death in the Mexican population. Forensic Sci Int. 2016;266:587.e1-.e10.
27. Cunha E, Baccino E, Martrille L, Ramsthaler F, Prieto J, Schuliar Y, et al. The problem of aging human remains and living individuals: a review. Forensic Sci Int. 2009;193(1-3):1-13.
28. Cardoso HF, Abrantes J, Humphrey LT. Age estimation of immature human
skeletal remains from the diaphyseal length of the long bones in the postnatal period. Int J Legal Med. 2014;128(5):809-24.
29. White TD, Folkens PA. The human bone manual.: Elsevier Academic Press; 2005. 464 p.
30. Lynch JJ, Maijanen H, Prescher A. Analysis of Three Commonly Used Tibia Length Measurement Techniques. Journal of Forensic Sciences. 2019;64(1):181-5.
31. Ubelaker D. Estimating age at death. In: Rich J, Dean D, Power R, editors. Forensic Medicine of the lower extremity: Human identification and trauma analysis of the tigh, leg, and foot2005. p. 99-112.
32. Scheuer JL, Musgrave JH, Evans SP. The estimation of late fetal and perinatal age from limb bone length by linear and logarithmic regression. Ann Hum Biol.
1980;7(3):257-65.
33. Thornton R, Edkins AL, Hutchinson EF. Contributions of the pars lateralis, pars basilaris and femur to age estimations of the immature skeleton within a South African forensic setting. Int J Legal Med. 2020;134(3):1185-93.
34. Bradtmiller B, Buikstra JE. Effects of burning on human bone microstructure: a preliminary study. J Forensic Sci. 1984;29(2):535-40.
35. Thompson TJ. Heat-induced dimensional changes in bone and their consequences for forensic anthropology. J Forensic Sci. 2005;50(5):1008-15.
36. Nagaoka T, Kawakubo Y. Using the petrous part of the temporal bone to estimate fetal age at death. Forensic Science International. 2015;248:188.e1-.e7.
37. Carneiro C, Cunha E, Botella M, Alemán I, Curate F. Fetal age at death estimation on dry bone: Testing the applicability of equations developed on a radiographic sample. Revista Argentina de Antropología Biológica. 2019;21.
38. Kiserud T, Piaggio G, Carroli G, Widmer M, Carvalho J, Neerup Jensen L, et al. The World Health Organization Fetal Growth Charts: A Multinational Longitudinal Study of Ultrasound Biometric Measurements and Estimated Fetal Weight. PLoS Med.
2017;14(1):e1002220.
39. Skupski DW, Owen J, Kim S, Fuchs KM, Albert PS, Grantz KL, et al. Estimating Gestational Age From Ultrasound Fetal Biometrics. Obstet Gynecol. 2017;130(2):433-41.
22
Appendix A: Search strategy
First PubMed was used to search for articles, and secondly Google Scholar was used to find additional information. Two books were available: The Juvenile Skeleton by Scheuer and Black and The Human Bone Manual by White and Folkens, of these books the relevant chapters were included in this review.
Per sub question articles were found and collected using the following search criteria: SQ1: ((gestational age) AND (ultrasound)) AND (long bones) → 154 results
SQ2: ((gestational age) AND (radiographic)) AND (long bones) → 36 results SQ3: ((forensic anthropology) AND (long bones)) AND (age) → 81 results SQ4: ((shrinkage)) AND (bone)) AND (fetal) → 13 results
Snowballing through the relevant papers resulted in approximately an extra 20 relevant articles.
