ABSTRACT: Exhibited in the Egyptian Galleries in the British Museum (London, UK) is an exquisitely crafted artificial big toe for the right foot. It is made from cartonnage, a type of paper mache (layers of linen soaked with animal glue and coated with tinted plaster), and has been dated to before 600 BC. Distinct signs of wear prompted one researcher to call it “one of the earliest working prostheses to have been identified from the ancient world.” Housed in the Egyptian Museum (Cairo, Egypt) is a second example found strapped onto the right foot of a female mummy and dated to between 950 and 710 BC. Made in three pieces, it has a complex series of laces that joins the three sections together. To date, this artifact may be the oldest known intravital limb prosthesis in existence. The objective of this research was to assess the functionality of their design. Design replicas (P1 and P2, respectively) were made and assessed using two volunteers (V1 and V2) exhibiting complete disarticulation at the first metatarsophalangeal joint of the right foot. Computerized gait analysis evaluated selected gait parameters against their sound left side (control). Foot kinematic and kinetic data were recorded (average of 10 trials), plus a volunteer questionnaire was completed. Kinematic data from both replicas demonstrated a very satisfactory level of relative dorsiflexion, especially for V1, when worn with an Egyptian-style sandal (87.13% of the normal average range). Kinetic data showed no significant increase in plantar pressures between the affected and the sound side when worn with or without a sandal. However, when the volunteers wore sandals without the toe prosthesis, significant differences in peak pressures were recorded, indicating a potential benefit of P1 and P2 when wearing Egyptian sandals. Both volunteers indicated in the questionnaire that P2 was the more comfortable of the two. This research concluded that although both devices could have afforded at least limited ambulation, the performance and perceived comfort of the three-part wooden replacement infers that nascent prosthetic science may have been emerging in the Nile Valley as early as 950 to 710 BC.
Do two ancient Egyptian artificial restorations of the right hallux predate the earliest known prosthetic device, the Roman Capua leg of 300 BC? This study examines the evidence and concludes that possibly the earliest known prosthetics practitioner may indeed have been a craftsman living in Egypt&#x2019;s Nile Valley sometime between 950 and 710 BC.
JACQUELINE LOUISE FINCH, PhD, and ANN ROSALIE DAVID, OBE, PhD, FRSA are affiliated with the KNH Centre for Biomedical Egyptology, University of Manchester, Manchester, UK. GLYN HARVEY HEATH, PhD is affiliated with the School of Health, Sport and Rehabilitation Sciences, University of Salford, Salford, UK. JAI KULKARNI, MA, FRCP is affiliated with the University Hospitals of South Manchester, Manchester, UK.
Disclosure: The authors declare no conflict of interest.
Correspondence to: Jacqueline Louise Finch, PhD, 60 Sunningdale Drive, Bromborough, Wirral, Merseyside, United Kingdom, CH63 0JE; e-mail: firstname.lastname@example.org
Evidence for the restoration of missing body parts by ancient Egyptian embalmers is plentiful.1 Such practices were included in the burial preparations to reinstate the completeness of the physical body and, in so doing, enabled the deceased to be mystically reanimated in the afterlife. False eyes, noses, and genitals could be added and missing limbs were often reconstructed, but these were often crude imitations constructed from mud, reed, linen, and resin. Two artifacts in this assemblage of restorations are, however, much more sophisticated in both appearance and design. Exhibited in the Egyptian Galleries in the British Museum, UK, is an exquisitely crafted artificial big toe for the right foot. Made from cartonnage using layers of linen soaked with animal glue and coated with tinted plaster, it has been dated to before 600 BC.2 Housed in the Egyptian Museum (Cairo, Egypt) is a second artificial toe. This was found strapped onto the right foot of a female mummy and has been dated to between 950 and 710 BC. Made in three pieces, it has a complex series of laces that joins the three sections together.3 Examination of both artifacts has prompted some to question whether they were intended solely for the next world or whether these rare objects do indeed represent the first glimmers of the development of prosthetic devices.4
Currently, the oldest known “prosthetic” device is one rescued from a wealthy Roman burial in Santa Maria di Capua Vetere (the site of modern Capua) and securely dated to 300 BC. This bronze and wooden leg housed in The Royal College of Surgeons (London, England) was unfortunately destroyed in the bombing raid of May 10–11, 1941. Bliquez5 describes how the hollow wooden core may have been attached to a foot or metal rocking peg, with suspension achieved by leather straps attached to a bronze waist band. Fortunately, a detailed examination was undertaken in 1920 by Sudhoff, the eminent medical historian, and a paper was published by von Brunn of Rostock in 1926.5 These records enabled the reconstruction of a replica, which is now displayed in the Science Museum (London, England). With no gait analysis yet undertaken on the replica of the Capua leg, its functionality is still questionable, although many assume that the original was indeed one of the oldest known prosthetic devices.
Bliquez also reports on early examples of “prosthetic” feet found in two European burial sites. At Bonaduz, Switzerland, the artificial right foot dates to the 5th to 7th century AD. Analyzing the remnants and surrounding earth suggests that this was a leather pouch that had been stuffed with hay or moss. A wooden base had been attached by iron nails to assist stability. A Frankish grave in Griesheim (7th–8th century AD) yielded the second example made from wood and bronze. With the lower part of the left leg missing, the foot appears to have been fixed to a wooden extension as far as the knee. Czarnetzki et al.6 show how this device may have been strapped to the residual limb.
In 1881, the British Museum acquired for its Egyptian collection a cartonnage toe known as the Greville Chester great toe (Figure 1). Made from linen soaked with animal glue and coated with tinted plaster (gesso), it is made in the shape of the right big toe and a portion of the right foot. Its exact find spot is uncertain, but Thebes, near present-day Luxor, has been suggested. Beautifully crafted, it once sported a false nail, and clear signs of wear are evident. Eight holes on the medial side show distinct signs of rubbing, indicative of pull lines from laces. Four holes on the lateral side have been filled in and overlaid with paint, perhaps some form of refurbishment that had been carried out before burial. These two sets of holes were probably used to attach it onto the foot or perhaps to fasten it onto a sock or sandal. Areas where the top coat has cracked and fallen away may correspond to the position of the knot and thong of an ancient Egyptian sandal. During extensive scientific analysis of the object between 1989 and 1992 at the British Museum, textile characteristics of the linen dated it to before 600 BC. Reeves,2 who published a paper in 1999, described it as “far more than a simple cosmetic restoration applied by the embalmer to make the body whole for the hereafter” and as such represented “one of the earliest working prostheses to have been identified from the ancient world.”
In 2000, The Lancet published an article by Nerlich et al.3 on the discovery of the second example of an ancient Egyptian big toe restoration. Found in a cemetery site near Luxor, this example was fastened onto the right foot of its female owner, who lived between 950 and 710 BC. They concluded that she had lost her big toe some time before her death, with the wound completely healed and imaging revealing demineralization of the first metatarsal head (MTH) accompanied by mild osteopenia restricted to the leg bones. The artificial toe, now at the Egyptian Museum, has been made in three sections, two of wood and the third possibly of leather (Figure 2). Numerous holes have been drilled, through which a series of marlin hitches (more commonly used to lash sails) secures the sections together. It has a similar arrangement of lacing holes to the Greville Chester toe, eight along the medial edge, with four on the lateral but lacking any signs of rubbing. There has also been some attempt to incorporate a degree of flexibility into the design with a rudimentary hinge in the region of the first metatarsophalangeal (MTP) joint. The plantar surface, which has been planed flat to produce a stable contact with the ground, shows clear signs of abrasion, and the proximal edge is rounded and polished to limit rubbing in the navicular area on the dorsum of the foot during flexion. The inclusion of a hinge plus flattening of the plantar surface could be considered as unnecessary procedures if intended as purely a postmortem adornment.
THE EFFECT OF HALLUX AMPUTATION ON GAIT
According to Mann et al.,7 many physicians believe the hallux to be essential for locomotion. However Greer Richardson and Tooms8 dismiss the hallux as being insignificant for walking. With it bearing some 30% of the overall pressure applied to the forefoot region, a reduction in contact time (CT) of the toes due to injury or amputation would be expected to place more pressure on the MTHs.
Poppen et al.9 studied the effect of hallux amputation on patients who had undergone reimplantation to reconstruct a missing thumb. The resulting slight collapse of the medial longitudinal arch and incompetence of the first ray was thought to be due to a compromised windlass mechanism and consequent loss of power in this area by the plantar aponeurosis. An editorial comment by Jahss questioned their suggestion, proposing that the incompetence of the first ray was due to the loss of tendon and muscle attachments inserting into the base of the toe. Mann et al.7 changed the procedure for hallux amputation: 3 cm of extensor hallucis longus and flexor hallucis longus tendon was left in place. At the first MTH, the articular cartilage was removed in all but one patient; the intrinsic muscles, the sesamoids, and plantar aponeurosis were all left unattached. After 36 months, gait analysis was undertaken.
By removing the hallux at the first MTP joint, instability of the medial side of the foot appears to develop. With an intact hallux, the extrinsic and intrinsic muscle insertions into the proximal phalanx are able to create compression, stabilizing the joint and increasing its weight-bearing function. When the plantar aponeurosis and intrinsic muscles are intact, stabilization of the arch is transferred mainly to the second and third rays and permits the forward and medial progression of the center of pressure. The main problem in such amputees appears to be the instability created in the first MTP joint by the excision of some of the tendon insertions, with these normally assisting in maintaining the medial longitudinal arch.
The aim of our study was to determine whether the comments by both Reeves2 and Nerlich et al.3 concerning the significance of the two ancient Egyptian artificial big toes could be justified. Assessing the functionality of design replicas required both objective computerized gait analysis and subjective patient feedback.
ASSESSMENT OF WEAR ON THE ORIGINAL ARTIFACTS
Visual and photographic examination of both artifacts was used to determine the position and extent of surface anomalies in five areas: 1) the plantar aspect of the false hallux (for friction with the ground), 2) the distal region (often abraded during toe-off), 3) the medial and lateral holes (due to lacing), 4) the external surface finish ( cracks and degradation), and 5) the internal surface finish (rubbing against the foot).
Only participants presenting with complete disarticulation at the first MTP joint of the right foot could be included in the research. Finding suitable volunteers with a stable pathology and lacking complicating contradictions in the affected or contralateral limb (deformity, ulcers, or neuropathy) proved extremely difficult, and it became apparent that those with fragile pathologies rendered them unsuitable for the study. After collaboration with the subregional Manchester Disablement Service Centre, a press release was issued from the University of Manchester Media Office. From those who responded, two participants were sent letters of invitation. Research Participation Information Sheets were made available before medical screening.
Foot status and stability of the residual limb were assessed, as well as medical history recorded by consultants in rehabilitation medicine. Detailed neurocirculatory foot assessment was conducted, and cadence and step were observed. It was only when the clinicians, participant, and researcher were satisfied with the overall suitability of the patient that Research Participant Consent Forms were offered. Throughout the whole of the consultation process and subsequent trial period, it was stressed that the individual was free to withdraw from the research at any time. Volunteers 1 and 2 (V1 and V2) never met each other during the course of the research. Table 1 contains the demographic data of the volunteers.
PRODUCING THE REPLICAS
Negative impressions of both feet and of the left big toe were taken from each volunteer using plaster of Paris bandage. Three-dimensional plaster positive models were made from these and were subsequently vacuum molded to produce polypropylene negative impressions. Stable polyurethane foam positive models were then cast. These were used for the formation of the cartonnage replica (P1) and formed the template for the carving of the articulated wooden replica (P2). Replica leather Egyptian-style sandals were also made for each volunteer based on original artifacts. Preliminary fitting and attachment of both replicas using linen lacing under the foot plus strapping of the sandals were done before gait analysis so that adjustments could be made to optimize both comfort and fit.
COLLECTING THE DATA
During the trials, V1 and V2 wore comfortable clothing exposing the lower limbs. Before the collection of data, the volunteers were acquainted with the laboratory and the 10-m walkway. They were asked to walk at a self-selected cadence, either leading with the right or the left foot. They performed as many practice walks as was required for them to feel comfortable. Markers were placed at the start of the track to assist them in placing a clean strike on the mat. The 10 best walking trials were recorded for each foot in each variable with each volunteer. At the end of each trial, the subjects stopped and were asked to return to their starting position as indicated by a marker attached to the floor.
Kinematic data were obtained using 10 stationary integrated cameras using the Qualisys Proreflex MCU240 motion analysis system (Götenburg, Sweden) at the CRHPR Gait Laboratory, The University of Salford. Each volunteer wore four 10-mm reflective markers (R1–R4 on the right foot; L1–L4 on the left foot) carefully placed in the following positions: R1/L1 on the distal phalanx; R2/L2 on the proximal first phalanx, immediately distal to the MTP joint; R3/L3 on the sustentaculum tali, about 2.5 cm below the medial malleolus; and R4/L4 on the medial malleolus (Figures 3A, B). Kinematic data were collected for the following variables: wearing replica P1 with both feet unshod, wearing replica P1 with replica sandals on both feet, wearing replica P2 with both feet unshod, and wearing replica P2 with replica sandals on both feet. No kinematic data were collected with the volunteers unshod without wearing prosthetic toes. Position R1 on the distal phalanx and position R2 on the proximal first phalanx, distal to the MTP joint, were both absent because of the disarticulation. With the absence of these positions for the reflective markers on the affected side, it was deemed unsatisfactory to gather incomplete data for this particular variable. There were 107 frames in one complete gait cycle (GC). The data were exported to Visual3D™ (C-Motion, Inc, Germantown, MD, USA), and the sagittal plane motion was derived. This enabled the effect of each of the four chosen variables on the sagittal plane motion of the right (replica) hallux to be determined. This could then be compared with the sagittal plane motion of the left hallux, the control. The effect of each of the four variables on the total stance dorsiflexion (TSD) of the left (control) and right (replica) hallux, to determine the relative dorsiflexion (RD) of replicas P1 and P2 with and without a replica sandal (shod and unshod), was calculated. The relative effect of the sandal (RES) on the TSD produced by replicas P1 and P2 was used to show how the presence of a sandal could assist or hamper the overall dorsiflexion achieved in stance. The duration of the stance phase for both the left side (control) and the right affected side was noted. From this toe-off, as a percentage of the GC was calculated.
Kinetic data were collected using the Medilogic® Pressure Measuring System (T and T Medilogic Medizintechnik GmbH, Schönefeld, Germany) placed on the walkway. The peak plantar pressure, average plantar pressure, and CT for one GC were recorded from the best (maximum of 10) trials for both the left and right feet for the following seven variables: unshod, shod in volunteer’s own footwear, wearing replica sandals, wearing replica P1 on the right foot with both feet unshod, wearing replica P1 on the right foot with replica sandals on both feet, wearing replica P2 on the right foot with both feet unshod, and wearing replica P2 on the right foot with replica sandals on both feet.
All gait analysis data were plotted and tabulated. The kinematic data of each volunteer were plotted to show sagittal plane motion of the left hallux versus that of the replica on the right foot in each of the four variables. The kinetic data were tabulated. Using Windows Excel (2003), the mean and standard deviation for each of the seven kinetic data sets (right and left) were calculated. Pairwise comparisons using a two-sample t-test determined whether there was a statistically significant difference between the left and right sides for that variable (df =18 unless stated); α level of statistical significance was set at a P value<0.05, and critical value t stat was ±2.101.
The volunteers were asked to describe if the working environment had influenced their performance in any way and also whether they perceived any difference in performance and comfort when wearing the replicas with or without sandals. With the use of a maximum score of 10 to denote total comfort and ease when walking barefoot on the walkway and 1 as having no value as a prosthetic toe, they were asked to mark on a continuous line the perceived performance in the other six variables. This was then converted into a numerical equivalent and recorded.
EVALUATION OF WEAR ON THE ORIGINAL ARTIFACTS
For the Greville Chester artificial toe (Figure 1), the following observations were made: 1) On the plantar surface, the outer coating of brown gesso showed signs of thinning, revealing the darker underlying coating. 2) On the medial side of the distal region, at the tip, there were also distinct signs of thinning and surface abrasion. 3) The coating surrounding the eight holes on the medial border appeared eroded on their inferior aspect. The lateral holes had been infilled in antiquity and recoated with brown gesso; no wear was evident here. 4) On the surface, the underlying layers of linen were exposed in three regions: over 4 mm2 at the toe tip on the plantar surface, in the first web space, and 1.8 mm × 7.3 mm along the proximal edge. 5) There was random thinning of the dark red coating on the internal surface.
For the three-part wooden/leather artificial toe (Figure 2), the following observations were made: 1) On the plantar surface, there was extensive scuffing mainly on the medial side, consistent with that reported by Nerlich et al.3 in 2000. 2) There was no scuffing or wear on the toe tip. 3) There was no wear apparent in or around any of the medial or lateral lacing holes. 4) There was no cracking or degradation over the external or internal surfaces.
The kinematic data for V1 are displayed in Figures 4A–D. When wearing replicas P1 or P2 without sandals (Figures 4A, C) the dorsiflexion in the stance phase is initiated early, is prolonged, and is less effective than that produced by the left (control) side. With the addition of a sandal, replica P1 produced strong dorsiflexion in stance, although toe-off was delayed. On the right side, the plantarflexion in swing was also strong. With the addition of a sandal, the dorsiflexion achieved with replica P2 was still prolonged (Figure 4D). The TSD on the right side in all four variables was less than that on the left (control) side, although this was greater when replica sandals were worn (Table 2). This can be clearly seen in the % RD, with P1 + sandals recording 87.13% and P2 + sandals recording 77.62% of the TSD achieved by the left control side. The relative effect of wearing sandals with the replicas was determined by calculating the RES values: when wearing replica P1, the RES on the right side was +43.36% (25.79 − 17.99/17.99 × 100), whereas when wearing P2, the RES was +28% (22.31 − 17.43/17.43 × 100). With P1 laced onto the right foot, the RES for the left side decreased and was −5.19% (29.60 − 31.22/31.22 × 100), whereas when wearing P2, the RES was −9.62% (28.74 − 31.80/28.74 × 100). Toe-off was most delayed (>60% of the GC) when wearing sandals with replica P1 (66.36%).
The kinematic data for V2 are displayed in Figures 5A–D. The sagittal plane motion in all four variables was similar. Dorsiflexion on the right side (replica) was tending to be initiated slightly earlier and was less strong than on the left. In all four variables, the TSD on the right side was again less than on the left control side (Table 2). Wearing sandals with either P1 or P2 caused a reduction in the TSD achieved by both the right and left sides. There was little difference between the % RD in each of the four variables, with P2 with the sandal giving the greatest value (63.20%). When wearing P1, the RES on the TSD of the right side was −18.31% (16.19 − 19.82/19.82 × 100), whereas when wearing replica P2, it was −13.81% (18.29 − 21.22/21.22 × 100). With P1 laced onto the right foot, the TSD on the left side decreased and the RES was −21.33% (25.96 − 33.00/33.00 × 100), whereas when wearing P2, the RES was −15.77% (28.94 − 34.36/34.36 × 100). Toe-off on the right side was consistently earlier (<60% of the GC), the earliest when replica P2 was worn with sandals (54%, GC).
A summary of the mean peak plantar pressure for V1 is displayed in Table 3. Wearing replica P1 unshod, the right foot recorded the highest value, 56.31 N/cm2. When wearing replica sandals only, there was great disparity in mean peak pressure between the left and right feet (12.67 N/cm2 [54.18 N/cm2 − 41.51 N/cm2]). A reduction in mean peak pressure on the right side compared with the corresponding left occurred when wearing replica P1 + sandals (4.81 N/cm2 [54.77 N/cm2 − 49.96 N/cm2]), replica P2 unshod (1.78 N/cm2 [50.99 N/cm2 − 49.21 N/cm2]), and replica P2 + sandals (4.78 N/cm2 [52.22 N/cm2 − 47.44 N/cm2]). Table 4 shows a summary of the position of peak pressure in all variables. Only on the left and when unshod was this consistently experienced under the intact hallux (7). When the volunteer’s own footwear was worn, it was placed under the heel (left 8, right 9); otherwise, it tended to spread laterally across the MTHs. On the affected side, it was experienced under the heel but mainly under the second to fourth MTHs. For V1, pairwise comparisons showed no statistically significant effect on the mean peak pressure except when wearing the replica sandals (p value L:R <0.05; see Table 4).
A summary of the mean average plantar pressures for V1 appears in Table 3. Under the right foot, wearing replica P1 unshod produced the highest mean average pressure, 34.43 N/cm2. Wearing replica sandals only produced the greatest disparity in mean average plantar pressures between the right and left sides (5.48 N/cm2 [32.62 N/cm2 − 27.14 N/cm2]). Table 4 shows a summary of the position of average plantar pressure in all variables. When unshod on the left side, average plantar pressure was experienced mainly under the intact hallux (6). On the right side, this occurred most often either under the heel (8, 10, 5, 5) or across the second to fourth MTHs (5, 7, 6, 5). Pairwise comparisons showed no statistically significant effect on the mean average plantar pressures in any of the variables (see Table 3).
A summary of the mean CTs for V1 appear in Table 3. When unshod, there was a disparity in mean CTs between the right and left sides (left, 0.79 seconds; right, 0.74 seconds). The longest mean CTs were recorded both on the left side, wearing replica P1 unshod and when wearing replica P2 with sandals (0.84 seconds). The mean CTs for both left and right sides when wearing replica P1 with sandals and replica P2 unshod were similar to that when the subject’s own footwear was worn. For V1, pairwise comparisons showed statistical significance in only one variable, and this was marginal, wearing replica P2 with sandals (p value L:R = 0.050; see Table 3).
A summary of the mean peak pressure recorded for V2 appears in Table 3. All values were higher under the right foot except when V2 was wearing replica P1 unshod (1.43 N/cm2 [59.48 N/cm2 − 58.05 N/cm2]). Note that 64.00 N/cm2 was the highest value recorded during the trials and the threshold of the Medilogic pressure mat. When wearing replica sandals only, there was again great disparity in mean peak pressure between the left and right feet (10.79 N/cm2 [64.00 N/cm2 − 53.21 N/cm2]). Table 4 shows a summary of the position of peak pressure in all variables. On the left side, there is consistent load bearing onto the heel in each variable (10, 7, 7, 7, 8, 8, and 9, respectively). On the right side, the heel or the first MTH was preferred. When the subject’s footwear was worn, the peak pressure tended to move laterally across the second to fourth MTHs. Pairwise comparisons showed a statistically significant effect on the mean peak pressure when the subject’s own footwear was worn and when wearing only the replica sandals (only five repeats were successfully achieved for this variable) (p values L:R <0.05, see Table 3).
A summary of the mean average plantar pressures recorded by V2 appears in Table 3. The highest mean average pressure recorded was under the right foot unshod, 44.15 N/cm2. Wearing replica sandals only produced the greatest disparity in mean average plantar pressures between the right and left sides (8.68 N/cm2 [41.78 N/cm2 − 33.10 N/cm2]). In all variables except when wearing replica P1 unshod, the higher value was experienced under the right foot. Table 4 shows a summary of the position of average pressure in all variables. On the left side, this was consistently placed under the heel (9, 7, 6, 7, 8, 6, 9), whereas on the right, the heel is used in combination with the first MTH. Pairwise comparisons showed a statistically significant effect on the mean average pressure for the following three variables: unshod, wearing own footwear, and wearing sandals (only five repeats were successfully achieved for this variable) (p values L:R <0.05; see Table 3). Volunteer 2 recorded pressure under the right hallux of replica P1 when unshod (peak pressure, 3; average pressure, 2) and also when wearing P1 with sandals (peak pressure, 4; average pressure, 1) (Table 4).
A summary of the mean CTs for V2 appears in Table 3. Examination of mean CTs in all the variables showed a maximum difference of 0.02 seconds only. Pairwise comparisons showed no statistically significant effect on the mean CTs in any of the variables (see Table 3).
EVALUATION OF VOLUNTEER QUESTIONNAIRE
Both volunteers felt that neither the walkway, the presence of the 10 integrated cameras, nor the technicians had compromised their performance during the trials. Neither of the volunteers felt that P1 or P2 was heavy to wear.
Volunteer 1 noted that P1 “felt very inflexible,” “it didn’t bend,” and “I seem to be thrown over to the right.” With the sandal, it was a little more comfortable, but this problem persisted. Volunteer 1 was unsure as to the robustness of P1 for prolonged use. For comfort and for ease of walking, V1 gave P1 a score of 2 (without sandal) and 3 (with sandal). Describing the performance of P2, V1 commented “I felt in time I could get accustomed to walking in it” and “this one (P2) was very good with the sandal.” For comfort and ease of walking, V1 gave P2 a score of 9 (for performance with or without a sandal).
The comments of V2 regarding P1 included “I found it difficult to walk in as it didn’t bend” but “cosmetically it is very good.” With or without the addition of a sandal made no difference to this volunteer; both variables being scored 1 for comfort and ease of walking. Comments regarding P2 included “this one did not rub at all; it was very comfortable” and “was softer to walk on but still moved inwards.” Overall, replica P2 scored 8 for comfort and for ease of walking (with or without a sandal).
Analysis of the kinematic and kinetic data showed that each volunteer responded differently to the variables, and thus, the results were highly subjective. It was, however, very encouraging that both volunteers were able to walk wearing the replicas, albeit within the confines of the gait laboratory.
THE PERFORMANCE OF P1
Despite the adverse comments of V1, our kinematic results showed that replica P1 performed well especially when supported by a replica Egyptian sandal. With the stance phase being prolonged and with a much delayed toe-off, an impressive amount of dorsiflexion relative to the unaffected left side was achieved. Our results for V2 showed that although P1 produced a slightly better degree of dorsiflexion relative to the left side when worn alone, the addition of the sandal was problematic with an anxiety to initiate an earlier toe-off.
Discomfort with possible tissue damage from excessive overloading was determined by the magnitude of peak pressure. Our kinetic results for both volunteers showed that there were especially large differences in mean peak pressure between the right and left feet when wearing only Egyptian-style sandals without the “prosthesis.” They may have felt unstable wearing just the sandals and this caused them to offload the left foot. The position of the sandal thong is traditionally located into the first web space. Because of the lack of a hallux on the right, it was necessary to reposition this thong into the second web space for the trials in this variable. In this way, the sandal on the right may have felt more secure than that on the left, possibly contributing to this offloading effect. An ancient Egyptian with complete disarticulation at the MTP joint trying to wear a sandal would have had similar difficulty holding it onto the foot and would also have been forced to adjust the thong.
Trials using transfemoral amputees have shown that more time is spent standing on the normal leg; consequently, this bears a larger vertical load.10 Prolonged loading asymmetry is known to lead to joint problems in the contralateral limb, such as weight-bearing osteoarthritis, degeneration,11 and the development of osteophytes and cysts.12 This loading asymmetry would outwardly manifest as a limp. It should be noted that these differences in mean peak pressures between the two sides decreased for both volunteers when P1 was worn on its own and to a lesser degree when the sandals were added, with V2 having greater problems with the sandals. The high incidence of peak pressure located at position 3 (second–fourth MTHs) when P1 was worn alone or with a sandal justified the perception of V1 “being thrown over to the right.” Volunteer 2 felt uncomfortable in P1, consistently placing peak pressure under the heel. Overall, however, these results suggest that the design of P1 was producing a beneficial effect on the symmetry of peak pressure loading.
THE PERFORMANCE OF P2
In terms of kinematic performance, replica P2 performed well when V1 wore it with a replica sandal, although the sandal adversely affected the performance of the left foot more than when wearing replica P1. Volunteer 2 was able to produce only a slightly enhanced level of function with this “prosthesis,” the sandals appearing detrimental to any improvement. When wearing P2 without a sandal, both volunteers again experienced a reduction in the peak pressure; thus, the effect of P2, like P1, reduced loading asymmetry. When both volunteers wore P2 with a sandal, loading asymmetry was again reduced but to a lesser degree. Comments from V1 regarding the comfort of the material were most encouraging. Despite still throwing the wearer over to the right onto the second to fourth MTHs and with a sandal marginally causing anxiety to move off earlier from right foot, P2 was perceived as the more convincing of the two devices.
Any contemporary partial foot prosthesis (PFP) must provide stability, control, and protection from shock while affording meticulous attention to alignment and fit at the socket-residual limb interface.13–16 Competent construction and an appreciation of the demands of the wearer are paramount. Today, clinical experience in patients with first MTP amputation indicates that functionality is of more importance than cosmesis, and any PFP that decreases plantar foot pressures without obviating comfort is clinically efficacious. Toe prostheses are not in common use, and the normal optimal biomechanical method is the shoe filler and rigid sole insert.
The purpose of our research was to evaluate the functionality of the design of two ancient Egyptian artificial restorations of the right big toe. Those results, coupled with examination of the original artifacts for any signs of wear and feedback from the volunteers, form the basis for our conclusions regarding their significance to the history of prosthetics.
After loss and during healing, ambulation for the ancient amputee would have been limited and gait dysfunctional. Walking without any kind of footwear would have put the foot at risk of further damage. Although wearing a sandal would have been regarded as the norm, holding it in place on the affected side would have been problematic. A sock or “boot” may have given both protection and support.
The quality of execution of both original artifacts is indisputable. Fashioning these replacement toes would have required skills and expertise afforded only to the elite in the society. There are distinctive signs of wear on the Greville Chester toe, which may of course be due to the nature of the material alone. The wood/leather three-part example is by its very nature more durable; however, the carver has made a deliberate attempt to satisfy both comfort and functionality. How well the originals fitted their owners will never be known, but undoubtedly, some consultation with the craftsman would have taken place, inferring some form of prescriptive and iterative process. Wearing either device, the owner could certainly have gained some clinical benefit with better distribution of plantar pressure and a more symmetrical heel-to-toe gait pattern. Finally, we can only speculate if they were worn mainly for special occasions (in temples/on ceremonial days) or more frequently on a daily basis.
Constructing and testing design copies using volunteers were deemed the most appropriate method of resolving the question of functionality and, thus, their possible significance to the history of prosthetics. The limitations of such research are numerous and challenging. These include difficulties in the selection of suitable volunteers, the sourcing of materials and manufacture of replicas, the method of attachment to the foot, the problems of biomechanical assessment confined under laboratory conditions, and the positioning of reflective markers plus data overload restricting the number of volunteers.
Although the replica of the three-part wooden device appeared to perform marginally less well in the assessment, it was this device that was perceived by the volunteers as being by far the more comfortable and convincing “prosthesis.” Assuming that the original artifact did indeed fit its intended owner, it now seems quite reasonable to suggest that this rare and beautiful object represents the earliest tangible example of a prosthetic appliance. Predating the Roman Capua leg by possibly up to 400 years, the earliest practitioner (although unwittingly) of this branch of medical science may have indeed been a craftsman living in the Nile Valley between 950 and 710 BC.
The authors thank the Egyptian Supreme Council for Antiquities, the Egyptian Museum, Cairo and the British Museum, London, for allowing examination and photography of the artifacts. Thanks also go to Andrey Aksenov, University of Salford, who assisted with the gait analysis data collection for this study, and Mr Geoffrey Riley, state registered prosthetist/orthotist and departmental technician, who offered technical skills.
1. Finch JL. The Significance of Two Ancient Egyptian Hallux Restorations on the History of Prosthetic Medicine: Evaluation of the Original Artifacts and the Biomechanical Assessment of Replicas
[dissertation]. Manchester, UK: Faculty of Life Science, University of Manchester; 2009;46–67.
2. Reeves N. New light on ancient Egyptian prosthetic medicine. In: Davies V, ed. Studies in Honour of Egyptian Antiquities: A Tribute to TGH James
. London, England: British Museum; 1999; Occasional Paper 123: 73–77.
3. Nerlich AG, Zink A, Szeimies U, Hagedorn HG. Ancient Egyptian prosthesis of the big toe. Lancet
2000; 356: 2176–2179.
4. Finch JL. The ancient origins of prosthetic medicine. Lancet
2011; 377: 548–549.
5. Bliquez LJ. Prosthetics in classical antiquity: Greek, Etruscan and Roman prosthetics. In: Haase W, Temporini H, eds. Aufstieg und Niedergang der Römischen welt. Rise and decline of the Roman World
. Berlin, Germany: Walter de Gruyter; 1996; 37.3: 2640–2676.
6. Czarnetzki A, Uhlig C, Wolf R. Menschen des Frühen Mittelalters in Spiegel der Anthropologie und Medizin
. Stuttgart, Germany: Wurttembergisches Landesmuseum Stuttgart; 1983: 91–95.
7. Mann RA, Poppen NK, O’Konski M. Amputation of the great toe: a clinical and biomedical study. Clin Orthop Relat Res
1988; 226: 192–205.
8. Greer Richardson E, Tooms RE. Amputations about foot. In: Canale ST, ed. Campbell’s Operative Orthopedics Vol 2
. St Louis, MO: Mosby; 1998: 1973–1990.
9. Poppen NK, Mann RA., O’Konski M, Buncke HJ. Amputation of the great toe. Foot Ankle
1981; 1.6: 333–336.
10. Eberhart HD, Elftman H, Inman VT. The locomotor mechanism of the amputee. In Klopsteg P E, Wilson PD, eds. Human Limbs and Their Substitutes
. New York, NY: McGraw-Hill; 1954: 472–480.
11. Norvell DE, Czerniecki JM, Reiber GE, et al.. The prevalence of knee pain and symptomatic knee osteoarthritis among veteran traumatic amputees and non-amputees. Arch Phys Med Rehabil
2004; 86 (3): 487–493.
12. Melzer I, Yekutiel M, Sukenik S. Comparative study of osteoarthritis of the contralateral knee joint of male amputees who do and do not play volleyball. J Rheum
2001; 28 (1): 169–172.
13. Engelmeier RL. Technique for prosthetic replacement of missing toes. J Am Podiatr Assoc
1983; 73: 36–38.
14. Kulkarni J, Curran B, Ebdon-Parry M, Harrison D. Total contact silicone partial foot prostheses for partial foot amputations. Foot
1995; 5: 32–35.
15. Dillon MP, Barker TM. Can partial foot prostheses effectively restore foot length? Prosthet Orthot Int
2006; 30.1: 17–23.
16. Stills ML. Partial foot prostheses/orthoses. Clin Prosthet Orthot
1987; 12 (1): 14–18.
KEY INDEXING TERMS: prosthetic toe; ancient Egypt; computerized gait analysis; prosthetic science
© 2012 American Academy of Orthotists & Prosthetists