Estimating Exposed Pulp Lengths of Tusks in the African Elephant (loxodonta Africana Africana)

a INTRODUCTION Tusks are the only maxillary incisors present in African elephants (Loxodonta africana). These incisors are continuously growing teeth (elodont teeth) with an open apex 6. The growth rate of tusks varies between the sexes with males having large tusks that increases in size, while females have smaller tusks of which the growth rate decreases at the time of puberty 12. The outer mineralised layer of a tusk is ivory, which comprises dentin not covered by enamel. Ivory has a unique pattern on a cut section. Enamel is only present for a short period when the permanent tusk erupts 14. The ivory protects the pulp of the tusk which consists of numerous blood vessels, lymphatic and neural tissue, bound in a loosely arranged connective tissue stroma 2,23. The tusk anatomy and biology create several potential clinical and management conundrums. Although captive and wild elephants suffer tusk fractures 6,9,19,20 , the


INTRODUCTION
Tusks are the only maxillary incisors present in African elephants (Loxodonta africana).These incisors are continuously growing teeth (elodont teeth) with an open apex 6 .The growth rate of tusks varies between the sexes with males having large tusks that increases in size, while females have smaller tusks of which the growth rate decreases at the time of puberty 12 .The outer mineralised layer of a tusk is ivory, which comprises dentin not covered by enamel.Ivory has a unique pattern on a cut section.Enamel is only present for a short period when the permanent tusk erupts 14 .The ivory protects the pulp of the tusk which consists of numerous blood vessels, lymphatic and neural tissue, bound in a loosely arranged connective tissue stroma 2,23 .
The tusk anatomy and biology create several potential clinical and management conundrums.Although captive and wild elephants suffer tusk fractures 6,9,19,20 , the incidence in captive elephants is higher than that in free-ranging ones 20 .Tusk fractures in captivity result from trauma during the transport and offloading of elephants as well as individuals falling into moats or damaging their tusks against enclosure barriers 19 .In many cases, institutions shorten or blunt the tusks 25 to reduce fractures and injury.Pulp exposure through fracturing or coronal amputation leads to pain for the elephant 2,3,10 , for which various techniques are available as treatment 19 .Even so, tusk amputation is a regularly used option that aims to shorten a tusk without exposing the pulp.Assessment of coronal pulp length (exposed pulp length) relies on measuring the distance from the lip (labio-dental fold) surrounding the tusk to the eye 15 .It assumes a 1:1 ratio between this measure and exposed pulp length.This method is at best a crude guide, since it does not consider variation associated with age or sex.
The aim of this study was to find a complementary and alternative way of estimating the coronal pulp length in African elephants.This will allow veterinarians to evaluate if a tusk fracture has exposed the pulp, and to assess the need for general anesthesia to treat a fracture.In the case of institutions housing captive African elephants, estimating coronal pulp length from external tusk features may guide veterinarians when amputating long tusks without causing pulp exposure.

Study populations
Our samples included tusks collected from elephants in two regions.Firstly, the Namibian sample comprised stockpiled tusks originating mainly from the Etosha National Park, but also included individuals from the Caprivi and Khaudum Game Reserves.The elephants in the Etosha National Park in Namibia (22 270 km 2 , 19°S, 16°E), live in a dry savanna 16 (annual rainfall: 316-442 mm) 8 .This population is stable and comprised 1754-3080 elephants in 2002 1 .The population has a high incidence of tusk fractures but low incidence of tusklessness 20 .Authorities removed 316 female elephants from 1983 to 1985, water is provided artificially and the park is partially fenced 8 .
The Khaudum Game Reserve (3841 km 2 , 19°S, 20°E) is partially fenced and has several boreholes that provide water yearround (D.Alberts, Namibian Ministry of Environment and Tourism, pers.comm., 2005).The Reserve had 993-2939 elephants in 2000 1 and comprises dry savanna 16 (annual rainfall 487 ± 221 mm, data provided by the Ministry of Environment and Tourism, Tsumkwe, Namibia).Nearly 25 % of tusks in Khaudum are fractured, lower than recorded for Etosha 20 .
The 3353-5799 elephants living in the Caprivi Game Reserve (5715 km 2 , 18°S 22°E) during 1998 1 are part of a large regional population that spans several countries.It has no management and elephants are free to move in and out of the reserve.The vegetation is typically that of a dry savanna 16 .Tusklessness is not common in the regional population, but the incidence of fractures is about the same as that noted in the Khaudum Game Reserve 20 .
The second population consisted entirely of individuals from the Kruger National Park, SA (Kruger) (19 442 km 2 , 24°S, 31.5°E).The elephants in Kruger live in a transition savanna 16 (annual rainfall 500-700 mm) 22 .Here the population comprised 10 459 elephants in 2002 1 .The population has a low incidence of tusk fractures and tusklessness 20 .Authorities removed 17 219 elephants between 1967 and 1996 21 , water is provided artificially 13 and the park is mostly fenced 24 .

Data collection
Several measurements were made (Table 1) on 242 tusks (all natural deaths) from Namibia and collated data for 448 tusks collected during the culling programme in Kruger from 1991 to 1996 (A.Hall-Martin, African Parks, pers.comm., 2004).All tusk length measurements were measured along the outside curvature of the tusk.Total tusk length (lt) was the length from the base of the tusk to the incisal edge of the tusk (Fig. 1).Tusk circumference was measured at the base of each tusk (cb), as well as at the lip margin visible as a brown ring on each tusk (cl).Tusk diameter was also measured using calipers at the same sites as described for circumference (diameter at the base -db, diameter at the lip -dl).Pulp length (lp) was measured on the inside of the outer curvature of the tusks, using a wire that would follow the shape of the tusk.Pulp volume (vp) was determined using fine river sand and filling the pulp until the sand was level with the rim of the tusk.All lengths and heights are in millimetre, volume in millilitre and weight in kilogram.
Only 217 of the Namibian tusks were included since it was not possible to measure all defined parameters on broken or split tusks.For the Kruger group, 447 tusks were included in the study (Table 1).At the time of sampling, the Kruger observers sexed individuals and assigned ages to them using tooth eruption criteria 7 .The collated data for this sample did not include pulp length and tusks were no longer available for further measurements.

Data analysis
To accommodate the missing pulp lengths for the Kruger sample we first tested whether the two samples is one statistical population.We thus tested whether the relationships between the circumference at the base of the tusk (cb) and at the lip (c l ) and the circumference at the base of the lip (cl) and pulp volume (v p ) is the same for Namibia and Kruger.Least square regression analyses was used to find the linear relationships and tested whether the slopes differed using t-tests 18 .Intercepts were set at zero.Proving statistical unity would enable prediction of the pulp length (lp) for the Kruger tusks.For this purpose, the relationship between pulp length (lp) and circumference at the lip (cl) for the Namibian population through least square linear regression analyses.This relationship was used to predict the expected pulp lengths (lp) for tusks that had circumference at lip (cl) noted in the Kruger data set.
From measured and predicted vari-   ables, coronal pulp length was calculated i.e. the pulp that extends into the tusk crown beyond the labio-dental fold also termed the exposed pulp length (ep) (Fig. 1).To determine the exposed pulp length (ep) the length of the tusk was subtracted from its base to the labio-dental fold (ll) from the total pulp length (lp) (ep = lp -l l ,) using both measured and predicted data.With linear least square regression analysis, the relationship between the exposed pulp length (ep) and the circumferences at the lip (c l ) in both data sets was determined.The predicted pulp lengths together with the known sex and age data of the Kruger set also allowed a study of the influence of sex and age on exposed pulp length (ep).
No relationship was found between exposed pulp length (ep) and circumference at the lip (cl) (F1,216 = 1.352,P = 0.246) for the Namibian data set.However, a wide spread was noted of this derived variable ranging from 300 mm coronal of the labio-dental fold to 300 mm apical of it (Fig. 4).Some of this variation could in exposed pulp length could be explained when we the derived Kruger data were separated into age and sex categories (Fig. 5).For the observed Namibian data, the maximum exposed pulp length was 300 mm and the upper 95 % confidence interval for this group was 255.5 mm (Fig. 4).

DISCUSSION
Elephants fracture their tusks and do more so in captivity 20 .When fractures expose pulp, clinicians may need general anesthesia to treat the tusk.However, general anesthesia in elephants carries the risk of animals having adverse reactions to drugs 4 , falling into moats surrounding their enclosures 19 or damaging the enclosures from falling on anesthetic induction and/or recovery 5 .Clinicians can minimise these risks by estimating the length of exposed pulp and applying general anaesthesia only to those individuals that need treatment.This study thus focused on finding an externally visible tusk measure from which to estimate the coronal pulp length and complement the existing guidelines 15 .
Application of these findings extends further to inform clinical decisions regarding tusk amputations that can prevent complications associated with pulp exposure.Some challenges were encountered though since the majority of environment-independent measurements of elephant tusks are within the bony alveolus of the maxilla and obscured when observing a live animal.However, several techniques (e.g.digital photogrammetry) 17 easily allow measures of tusk circumference (cl) and diameter (d l ) at the lip.We therefore focused on such variables and in particular on circumference at the lip (cl).
Initially we encountered some analytical constraints due to partially available information for some data sets.Namibia had all the tusk measurements, but little data on the sex and age of the individuals from which the tusks came.Kruger had the sex and age data, but missed some of the morphological tusk measurements required.It was thus first necessary to evaluate whether these samples could be pooled into one statistical populationthat allowed prediction of the missing data for Kruger.Two morphological relationships were found that did not differ between the geographically separated Namibian and Kruger samples irrespective of sex or age.These two samples were therefore treated as one statistical population for the purpose of this study.
To find a stable, clinically visible measure for exposed pulp length, focus was placed on the total pulp length (lp), a crucial variable that was not recorded for Kruger.However, accepting the samples from Kruger and Namibia as a single statistical population allowed prediction of the total pulp length (lp) for the Kruger samples.The relationship between total pulp length (lp) was used and the circumference at the lip (cl) constructed from the Namibian samples to estimate total pulp lengths (lp) for the Kruger samples from their circumference at the lip (cl) measurements.That allowed calculation of a measure of exposed pulp length for both Kruger and Namibia.
It was expected that older and thus larger elephants would have larger tusks and presumably more exposed pulp beyond the lip margin.However, no relationship was found between exposed pulp (ep) and the circumference at the lip (cl) of a tusk.Several reasons for this finding were considered.The surrounding soft tissue and associated alveolar bone enlarge with the growing tooth that may result in constant exposed pulp lengths -when elephants increase in size the exposed pulp lengths do not increase at the same rate.As elephants age, they become larger and the surrounding tissue increases owing to growth-enhancing factors while bone mass increases to accommodate the heavier tooth.Rates of increase in pulp length, tissue and bone volume need to be the same to keep exposed pulp constant when elephant size increases.No measures of how tissue and bone volume change with age were available.However, the large variation in exposed pulp length that we have noted suggest that age related rates of change in pulp length, tissue and bone volume are not the same.Other factors thus most likely influence allometric relationships of exposed pulp length.
Alternatively, the analytical approach to accommodate the shortcomings of the data sets may have masked the results, particularly in the case of Kruger.It is conceded that the pooling of sex and age may also influence our results.Sex and age differences in exposed pulp length, (ep) as evident for the values predicted for Kruger, explained some of the variations noted.Males tended to have a larger variance in exposed pulp length (ep) irrespective of age than that of females.However, interpretation of these sex-and agespecific patterns are cautious, as much of the variation will be removed by the statistical relationship used to predict total pulp length (lp) from which exposed pulp length (ep) is calculated.Furthermore, geographical differences may skew the conclusions.The relationships for Namibia, from which the exposed pulp lengths for Kruger were predicted, came from tusks collected from three widely separated areas spanning a large rainfall gradient.It was not possible to separate these tusks.Since the incidence of tusk fractures seems to decrease with rainfall 20 , it is not inconceivable that tusk growth and development may differ across a rainfall gradient.
Finally, unknown biological processes may be a more compelling reason for the lack of a relationship between exposed pulp (e p ) and the circumference at the lip (cl).The maximum exposed pulp (ep) calculated for tusks from Namibia were 300 mm (Fig. 4), a relatively small value compared with the total tusk lengths measured in this sample.This was surprising as it was expected that the total pulp length and therefore exposed pulp length to elongates as the tusk elongates.Given that the volume of the pulp is conical, both the diameter at the base of the tusk (db) and the total pulp length (l p ) can influence pulp volume (vp).Tusks do not only increase in length with age, but the circumference of the tusk also increases.It is likely that the increased pulp volume of older male elephants, for instance, may be a result of an increase in circumference more than an increase in length.This conceivably influenced our results.In addition, secondary dentin formation is a well-recognised entity 11 .It is a life-long continuous, physiological deposition of dentin that eventually results in reduced pulp volume.Tertiary dentin, by contrast, is deposited specifically in response to injury of the tooth 11 .It was therefore also postulated that continuous deposition of dentin, either as a normal physiological process or as a result of trauma/injury or a combination thereof to be responsible for this reduction in pulp volume that affects both the diameter at the base of the tusk (db) and the total pulp length (lp).For instance, the process could result in reduction of pulp length even while tusk length and circumference increases.This may explain why a tendency was noted for the range of exposed pulp (ep) values to reduce, as females get older.
The biological and statistical reasons highlighted above may explain why a linear correlation between exposed pulp length (ep) and the circumference at the lip (cl) was not found.What does this mean clinically?The variance noted in exposed pulp defined an upper 95 % confidence level of 255.5 mm that served as a guideline.It is therefore proposed that no tusk should be amputated less than 300 mm from the labio-dental fold, regardless of sex or age of the individual animal.It is suggested that this guideline is a compliement to the existing index 15 .

Fig. 1 :
Fig. 1: Diagrammatic illustration of a tusk sectioned longitudinally showing all the length, height and circumference measurements taken on the tusks.A, Total tusk length; B, length of tusk to lip.C, pulp length; D, height and circumference at base; E, height and circumference at lip; cb = circumference at tusk base; cl = circumference at lip margin; db = diameter at tusk base; dl = diameter at lip margin; ep = exposed pulp length; ll = length base of tusk to labial fold; lp = total pulp length; lt = total tusk length; vp = pulp volume;

Fig. 2 :
Fig. 2: Relationships between measures of the tusk circumference at base (cb) and values of tusk circumference at the lip (c l) as well as the third root of pulp volume (vp) recorded for samples from Namibia and Kruger.We found no difference between Namibia and Kruger in the respective relationships (see text) and thus pooled the data.

Fig. 3 :Fig. 4 :
Fig. 3: Relationship between pulp length (lp) and the circumference at the lip (c l) established for tusk collected in Namibia.We did not know the sexes from which these tusks came.The relationship allowed us to predict pulp length for tusks collected in Kruger since the tusks from Namibia and Kruger come from the same statistical population (see text and Fig. 2).

Fig. 5 :
Fig. 5: Sex differences in estimated exposed pulp (ep) are apparent, while age differences are not well defined for the values we predicted for tusks from Kruger.Males usually have a longer exposed pulp than females, while the variability in females decreases with age.