Aquarium-housed elasmobranchs were fed a variety of fish and shellfish (capelin, clam, shrimp, herring, and smelt) based on body weight. Animals were offered an elasmobranch micronutrient supplement (hereafter termed ‘supplement’) specifically made for Disney’s Animal Kingdom (Table 1) at the recommended dose [1 tablet (1.5 g) per ½ lb. of fish; Mazuri Vita-Zu Shark/Ray tabs 5M1Y, Mazuri Exotic Animal Nutrition, St. Louis, MO]. This supplement was altered from the commercial product (5MD8-2011 analysis shown) during the course of this data collection at Disney’s request unrelated to this study at the time (Table 1). Supplements were hidden inside target-fed fish, but consumption was variable across species and individuals, and not always able to be clearly tracked. SRs in the aquarium setting had vitamins withheld from 8/1/2015 to 9/28/2016 as a response to elevated levels of iodine.

Lagoon animals were housed in a netted lagoon pen in natural sea water (Castaway Cay, Bahamas). They were fed a daily diet consisting of shrimp with shells but no tails (20% total diet), squid without heads or pens (30% total diet), and an aquatic gel product fed as 50% of the total diet offered (Aquatic gel 57 W9, Mazuri Exotic Animal Nutrition, St Louis, MO).

Wild SR diets were unknown, but diet and stable isotope analysis has been described (24).

Serum for total (the sum of inorganic and organic/protein-bound) and inorganic iodine was submitted to the MSU VDL for analysis as per Wahlen et al. (32).

Serum was submitted to the MSU VDL for cobalt analysis as part of a trace mineral panel, using inductively coupled plasma-mass spectrometry (ICP-MS). Standards were from Inorganic Ventures. In-house serum pools were used as controls.

Thyroid ultrasounds were reviewed by one author (Mylniczenko). The purpose of using this tool in this study was to identify animals that had grossly different thyroid appearance from known normal animals, suggesting disease as identified by experience from the authors and in comparison to other species. Images were identified retrospectively from exams which were conducted by multiple individuals; they were not always done with uniform technique and quality, thus acquisition of consistent measurement parameters was not possible. There are limited references on ultrasound evaluation of thyroid glands in elasmobranchs (33). The shape and size of the thyroid varies among species and can be difficult to identify (34). A normal thyroid has a symmetrical, homogenous appearance (35, 36) and in elasmobranchs they are smaller or equal in size to adjacent coracohyal and coracomandibular musculature (34, 37). Any variation from this uniformity was documented.

For a basic or ‘high level’ assessment of thyroid disease, the following criteria were used:

Cysts and foci were not quantified, they were subjectively evaluated and listed as mild, moderate, and severe.

All analyses were performed using R statistical Software v4.3.2 (38) or SigmaPlot (v.14.0), or MedCalc software Ltd. v.23.0.9. All statistical tests were evaluated at an 𝛼< 0.05.

Quantitative measures (thyroid, cobalt, and iodine) were tested for normality using the Shapiro–Wilk normality test and visual inspection of the data normality using Q-Q plots (quantile-quantile plots) in the R ‘stats’ package (38). Since most of the data were left skewed and failed tests for normality even after standard transformations, comparisons between habitats and study cohorts were evaluated using non-parametric methods and corrections were applied for multiple post-hoc tests as appropriate. Log-transformation was employed on iodine measures during linear and ordinal regression model testing and evaluations. Single pairwise comparisons of serum thyroid, cobalt, or iodine measures evaluating differences between species and habitats used Mann–Whitney Rank Sum Tests. Study cohorts and habitat results were summarized as median (χ̃) and interquartile range [IQR; 25th and 75th percentile (%)], where appropriate.

Iodine values measured as high as 606,740 ng/mL [I (I); N = 151 χ̃=48.0 ng/mL, IQR = 25.00, 130.00] for inorganic iodine, and 493,529 ng/mL [I (T); N = 157, χ̃=150.0 ng/mL, IQR = 53.00, 495.00] for total iodine in WSBS with moderate to severe thyroid disease. Cobalt ranged from 4.24 to 2,818.91 ng/mL (N = 146, χ̃=112.33, IQR = 27.78, 316.17; see Figure 3). The highest iodine and cobalt values (>1,000 ng/mL) were observed in aquarium-housed animals in 2015 through 2020 (N = 15), and primarily in WSBS species (12 of 15 samples, N = 2 individuals, samples collected from 2013 to 2020). These high values were obtained when the WSBS were offered (and consumed) supplement “C” and “D” with approximately 100,000 ppm DM iodine (see Table 1). Highest cobalt values were observed in 2017 and 2018 when animals were offered supplement “D” (5M1Y; 5MD8) approximately 2 years after iodine content had been reduced by over 33% from 202,603 ppm DM down to 133,658 ppm DM, and more than 5 years after cobalt had been reduced from 151.4 ppm DM (supplement “A”) down to 4.2 ppm DM (supplement “B”) and 3.8 ppm DM (supplement “C”; see Table 1). Only 29 of 53 samples with values <300 ng/mL cobalt were evaluated as thyroid-normal. The remaining samples included 5/53 goiter; 9/53 Normal (LG); 5/53 mild; 5/53 moderate; and 8/53 moderate–severe. There was a mix of thyroid disease scores in the higher cobalt measures (>300 ng/mL), including 7/27 evaluated as normal, 6/27 Normal (LG), 0/27 goiter, 9/27 mild, 1/27 moderate, and 4/27 moderate–severe thyroid disease. The highest cobalt values observed (N = 13 > 700 ng/mL) primarily came from HCR, PR, and WSBS species. The lowest iodine and cobalt values were observed in wild and lagoon-housed SR (see Figures 2D–F), where all measured total iodine values were < 400 ng/mL, inorganic iodine values were < 50 ng/mL, and cobalt values were < 200 ng/mL. There was one exception in an aquarium-managed HCR with undetermined thyroid disease status, measuring a low of 14 ng/mL for both inorganic and total iodine.

Spearman’s rank correlations were computed to assess all available pairwise relationships between quantitative measures. Overall, there were strong, significant positive relationships between inorganic iodine and total iodine (⍴(149) = 0.85, p < 0.0001), total iodine and cobalt (⍴(131) = 0.54, p < 0.0001), TT4 and FT4 (⍴(112) = 0.69, p < 0.0001). Significant but weaker relationships were also observed between TT3 and TT4 (⍴(111) = 0.42, p < 0.0001), and TT3 and cobalt (⍴(45)=0.14, p < 0.0001; see Figure 4). All other correlations were of lower strengths of effect, but it was observed that in this correlation matrix, inorganic iodine had a weak, negative correlation with two of the thyroid measures (TT3 ⍴(46)= − 0.05, and FT4 ⍴(47)= − 0.14, see Figure 4, light red negative ellipses).

Ordinal regression models were determined to be the best fit to evaluate the ability of serum inorganic iodine, total iodine (or total iodine <400 or > 400 ng/mL), cobalt, TT3, TT4, and FT4 to predict thyroid disease evaluation scores in the aquarium cohort. In the first model, iodine values were log-transformed, and total iodine and cobalt were removed due to high VIF, collinearity and best fit. For Model 1 the factors determined to best predict thyroid disease severity (Y) was Y ~ log inorganic iodine + TT3 + TT4 + FT4 (Nagelkerke R² = 0.98; see Supplementary Table S3). Holding all other predictors constant, samples with 1 unit higher value of log inorganic iodine were 2.7 times more likely to be evaluated in a higher thyroid evaluation score (OR = 2.738, 95% CI = 1.280, 6.833).

Serum TT3 had a more complex association with iodine and thyroid disease score. Serum TT3 values were elevated in many of the samples from supplemented animals with the exception of the highest severity of thyroid disease (see Figure 6A). This negative relationship can also be seen in Figure 7B predictor effects plot as a negative association of TT3 with thyroid disease such that the highest probability of moderate–severe disease was associated with the lowest TT3 values. Holding all other predictor values constant, with every one nmol/L increase in TT3 we would expect a 1.125 decrease in the cumulative probability of thyroid disease evaluation level (log odds scale). Calculating for the OR (OR = 0.325, 95% CI = 0.091, 0.944) the reciprocal (inverse of the odds ratio = 3.077) can be more easily interpreted. In this model, where increased inorganic iodine was correlated with increasing disease severity, there was also a 3.1 times higher odds to be evaluated into a higher thyroid disease category for every 1 nmol/L decrease observed in serum TT3. The cumulative probabilities for each change in thyroid evaluation score are visualized as predictor effects plots in Figures 7A,B (log inorganic iodine, and TT3).

For Model 2 we were specifically interested in the predictive ability of serum total iodine measures <400 or > 400 ng/mL as categorical variables. Using a final model of Y ~ total iodine >400 + TT3 + TT4 + FT4, samples with total iodine >400 ng/mL were 3.6 times more likely to be evaluated as a higher thyroid disease score (OR = 3.591, 95% CI = 1.280, 6.833, p = 0.012). Serum TT3 was not significant in this more generalized model (p = 0.104), but had similar trends as in Model 1, with lower TT3 values occurring in the highest disease categories. The differences between the highest thyroid disease evaluations for mild/moderate (p < 0.01) and moderate/moderate–severe (p < 0.001) can be readily distinguished between the <400 and > 400 ng/mL total iodine categories (Figure 7C). In the <400 ng/mL total iodine category, the probability of a normal thyroid evaluation was 61% (vs. 31% for >400), and only 19% probability of scoring a 4 or 5 (moderate to moderate–severe) in the <400 ng/mL total iodine (vs. 45% probability for >400).

Commercial laboratory analysis of blood thyroid hormone results consistently showed values in elasmobranchs were markedly below those routinely measured in other non-wildlife, terrestrial vertebrate species at MSU VDL (e.g., Felis catus or Canis familiaris). This was especially unusual given that clinical cases of thyroid disease (not goiter) still resulted in near undetectable blood thyroid hormone levels. Working with the laboratory (Bousianny), extant RIA methodology was reviewed. First, a direct measurement for known thyroid hormone was necessary. As a biological validation, HPLC performed on homogenized and extracted thyroid tissue showed high concentrations of immunoreactive TT3, TT4 and FT4 (Figure 1), while values from a matched cardiac blood sample were low. These results suggest that thyroid hormones are indeed highly conserved in elasmobranchs (55), and exhibited expected cross reactivity with established commercial RIA antisera. Secondly, serum was evaluated from a subset of ‘thyroid-normal’ SR and compared to thyroid abnormal SR. Normal values were uniformly quite low, most at or near zero, and in fact, below the readable range of the kit standard curve. Increased sample volume (2x for TT3 and 8x for TT4) and additional standards (down to the assay lower limit of sensitivity) allowed for improved sample measurement (Tables 2, 3). While both T3 and T4 are found in the thyroid itself, presumably large amounts are not routinely released into circulation resulting in relatively low levels which may not correlate with hypothyroidism.

RIA is generally considered the “gold standard” for thyroid hormone assays. Developed decades before the more current EIAs and CLIAs, RIAs generally had the advantage of improved discrimination and greater sensitivity and could easily be adapted “in-house” to meet the needs of species-specific lower ranges by adjusting working dilutions of the radio-labled tracer or antisera titer (56, 57). In contrast, current commercial labs typically follow a standardized RIA kit protocol using unextracted serum that had been optimized for species with higher thyroid hormone ranges (e.g., canine, feline (21)). Previously published data as well as older research were focused on RIA and likely varied methodology. Many labs may switch to non-RIA assays because of ease and safety (no need for radiation) but since they are often validated for targeted species, they might not be sufficiently sensitive to work well on elasmobranchs. Direct comparison of this study to previously published elasmobranch values is not a simple task. Current commercial labs and assays cannot easily perform historical sample processing or match assay methods (49, 51, 56–61). A further question was which of these values have biologic relevance. Thyroid hormone physiology is reviewed elsewhere (62) but salient content notes that in many species T3 is the biologically active form with T4 being a prohormone that requires conversion (62). Additionally T3 may be produced peripherally in multiple organs, like the liver (in dogfish (Squalus acanthias)) and possibly the kidney. Overall, the literature continues to be data-poor for thyroid function in elasmobranchs. What exists is challenging to interpret due to a variety of sample processing and assay methodology, and a potential confound of increased TT3 and TT4 in supplemented animals that may also vary in rates of consumption.

Testing methods for TT4 were changed during the timeframe of this research from RIA to CLIA due to assay kit availability. To examine the usefulness of this new test in comparison to the validated RIA, comparisons were conducted on the same samples, where available (N = 54). No correlation was found between the two methods (r² = −0.015). Additionally, CLIA values did not repeat well with the same samples when resubmitted. It is unclear why the repeat samples were not consistent but antisera differences or interference factors within the sample matrix are suspected (6, 63, 64).

These findings overall suggest that only TT4 and TT3 of the current RIA assay have potential value with the assay modifications as described. Using the assay ‘as is’, TT4 and TT3 (without increased sample volume) low values near zero may not mean hypothyroidism, and values that are readable (e.g., elevated TT4 and FT4), may indicate a problem. When disease progressed to a severe category, TT3 concentrations dropped to near baseline. TT4 CLIA results were questionable, and further research should be conducted to validate it.

Considering that iodide deficiency is the most likely cause of thyroid enlargement and at the time when this study was initiated, with near zero thyroid hormone results in nearly all samples, serum iodine seemed a plausible tool to explore for deficiency and as a proxy potentially for goiters. There are well established connections between acute iodide loads and thyroid inhibition known as the Wolff-Chaikoff effect (65). Total iodine (including unbound and protein-bound forms) and inorganic iodine (unbound only) were run on serum to establish some baselines and also to see if there were changes related to disease. Wild SRs (considered normal) were compared to semi-wild lagoon-housed SR (natural sea water but nutritionally managed) and closed system aquarium-housed SR, that was overrepresented with reproductive disease. In general, serum iodine levels of <20 ng/mL are abnormally low and 20–40 ng/mL questionably low, while 40–400 ng/mL seems to be a range consistent with a healthy animal. The wild SR stood out with iodine values in the 300 ng/mL range and this is likely because of a diet rich in natural iodine [crustaceans, seaweed, etc. (66, 67)]. Conversely, lagoon-housed SR did not have access to high volumes of natural foods and were fed a managed diet while living in natural sea water. Their diet however differed from aquarium-housed SR in that they did not receive a supplement and only a gel diet offered with low levels of iodine and cobalt.

High iodine values (>400 ng/mL) were frequently identified and associated with a negative health status and abnormal thyroid appearance. In practice, the authors recommend any serum values above 200 ng/mL prompt an evaluation of the animal and its history, thyroid appearance, and any comorbidities. A possible outcome might be a reduction of supplement in a mg/kg of fish or a reduction of frequency per week. Extant supplements that a majority of aquaria use have a moderate amount of iodide in their formulations although this has decreased over the last several years. Others have suggested that the typical level of iodide offered to elasmobranchs is more than required to balance a deficient diet (37, 68).

Serum total and inorganic iodine values in wild SR ranged from 177 to 307 ng/mL, and 32 to 48 ng/mL (respectively). In the lagoon-housed SR, values were lower than expected. These animals were exposed to natural iodide in natural sea water and minor amounts in the diet, but they were not supplemented. Neither of these groups had any thyroid abnormalities identified on physical exam or ultrasound. In the overall dataset including aquarium held animals, there were some very high values of iodine (Figure 2). Levels of supplement correlated with serum levels: if there were higher levels of dietary iodine, this tended to correlate with greater serum iodine levels. Animals that reliably accepted and consumed supplements were documented with elevated iodine levels compared to individuals that did not (pers experience Mylniczenko). One cohort had reproductive disease and abnormally high levels of iodine (>2000 ng/mL, unpublished) (22). Those individuals had all supplementation removed (oral and water sources) for more than a year to observe serum levels of iodine over time as well as thyroid size and shape. Iodine levels dropped but took several weeks to months, suggesting accumulation and slow excretion rate (69), yet the thyroid never enlarged although showed abnormal architecture. This prompted a smaller investigation in healthy animals given one dose of supplement (70) that showed iodine peaked at 24 h and declined over 72 h. This suggests that repeated dosing at an interval less than 72 h may result in accumulation, helping at least in part to explain higher levels of iodine.

This finding highlighted that accumulated iodine levels take a long period of time to reduce and even with seemingly low levels, not all species develop goiters. In contrast, eagle rays not receiving supplements developed a goiter by 6 months post birth. Similarly, bamboo sharks developed goiters 2 months after removal of supplements (52). The water systems they were in were ozonated, making iodide absorption and use from water improbable (50). This is likely due to the assumption that any iodide that comes in contact with ozonated water, potentially even within the digestive tract, would react to become biologically unavailable in the form of iodate (50, 71).

It is tempting to recommend normal ranges or exact cut-offs for low serum iodine to facilitate diagnosis of goiters. However, lagoon-housed SR had both low serum values and low dietary iodine, with only the natural sea water providing biologically available iodine and they did not develop goiters. Iodine can be both absorbed and excreted through renal clearance at a variable rate dependent on species and the presence of other nutrients such as chloride, so the ocean’s bioavailable iodide may be processed differently than in an aquarium setting with faster clearance (12, 72, 73). In the few animals with confirmed goiters, those individuals were not on any supplements prior to diagnosis (pups under observation). Once a goiter was identified, either by gross appearance or ultrasound, supplementation was started and the goiter was resolved.

Hyperiodinism, as shown here, has not really been considered a problem of elasmobranchs. It has been suggested that ‘excess’ waterborne iodide did not change serum T3 and T4 (RIA) concentrations in white tip reef sharks relative to TT3 and TT4 (EIA) when compared to ranges reported in wild-caught Atlantic stingrays (H. sabina) (49, 57, 58) and trout (74). This direct comparison is challenged by the variability in delivery method, sample processing, antisera, assay methodology and sensitivity (free vs. total T3 or T4: RIA vs. EIA). The authors suggested no negative consequence from high iodide and concluded that they are insensitive to excess iodide. This is contrary to literature across species which does suggest iodine in excess can cause problems. Excess iodine can lead to thyroid dysfunction resulting in either hyperthyroidism as well as hypothyroidism (65, 75), thyroiditis, and goiter development (76). In some mammals, iodide can cause toxicity resulting in mortality of young animals, prolonged parturition, and other signs (75, 77). In birds, excessive iodine can cause reduced egg production by directly inhibiting ovulation (78, 79). Additionally, at least in fish, the ‘insensitivity’ to iodide was tested in short term high exposure, vs. chronic exposure. In the data presented here, iodine is shown to correlate with consumption, length of time on supplements, and disease state (reproductive disease and abnormal thyroid tissue findings). Some of the levels identified in this paper suggest that iodine may be stored in the body outside of the thyroid gland, allowing for later release, as has been documented in other species (80, 81). Iodine may vary in non-thyroid storage sites across species. It accumulates in bile (80) as well as ova (82) and in some of the animals reported here, follicular iodine was interpreted as high (~9,000–10,000 ng/mL compared to same day serum of >8,000 ng/mL; unpublished) in animals with reproductive disease (characterized by follicular stasis or retention). Normal comparisons are still necessary. In a recent paper on iodinated contrast in catshark (Scyliorhinus torazame) computed tomography scans, iodine was seen to accumulate in the gallbladder peaking at day 6 and ovary showing high values at day 260 post administration (83). This storage capacity may help explain some of the high values seen in the current study. It is unclear in males if there is a different level of accumulation as abnormal animals in this study were predominantly female (females comprised 81.7% of individuals & 87.2% of study samples; see Supplementary Table S1). Iodide, in excess, has been labeled as an endocrine disruptor with the thyroid as the main target (84). Additionally, excess iodine has been linked with hyperthyroidism and arrested ovulation which fits with several of the cases here (85). Clearly this is an area of much needed research for this particular disease process (6).

Serum iodine in normal thyroid SR showed a lot of variability with the wild group having higher inorganic iodine relative to the lagoon- and aquarium-housed SR conspecifics. This is likely because of access to iodine rich prey and plant material. In this study, there were not many true goiters to derive any ranges in order to offer diagnostic values for identifying risk of goiter. In fact, the lagoon-housed SR, which were in similar habitats to wild SR (natural sea water environment) had moderately lower iodine values and also showed no evidence of goiter formation evidenced by ultrasonography. These animals do not have access to the diversity of prey species as they are managed in an enclosed lagoon and are fed a prepared diet of shrimp, squid and gel that has a minor amount of iodine present. It is possible that the clearance rate of iodine in the lagoon population is high, but access to biologically available water iodide offers a physiologically appropriate level to prevent goiters.

This current data supports the need to provide supplementation to prevent classic colloidal goiters in elasmobranchs. It also highlights the need to reevaluate the current levels and frequency of recommended supplementation as the physiologic levels of some key elements are elevated over what seems required. While species specific data would be ideal for iodine and trace mineral panels, it seems unlikely that the proposed ranges vary significantly from species to species. Additionally, given the association with both thyroid pathology and reproductive disease in many species further investigation is truly warranted.

Iodine may serve as an indicator for thyroid health, but only for elevated thyroid hormones (‘hyperthyroidism’). Iodine levels in managed animals over 400 ng/mL (the upper limit of what wild animals showed) indicate an abnormal status. We recommend levels above 200 ng/mL prompt a recheck and comparison to thyroid ultrasound, often resulting in a reduction of supplement dose and/or increased interval in between dosing. Perhaps mature females require different levels of supplement. The data here suggests that hypothyroidism cannot be conclusively connected to low thyroid levels or low iodine, without corresponding clinical information (overt swelling or ultrasound). Further attention should be considered for long term effects in aquarium populations.

Serum cobalt was markedly elevated in aquarium-housed animals compared to wild and lagoon-housed SR counterparts; and higher when total iodine was greater than 400 ng/mL. Comparable studies note values of 1.3–2.6 ng/mL in wild sand tiger sharks (Carcharias taurus) (86) and 24.9–180.7 ng/mL in aquarium smooth dogfish (Mustelus canis) (87). In the current dataset values ranged from 4.2–2,818.9 ng/mL with the highest belonging to aquarium animals.

Cobalt is known to be stored in the thyroid (88, 89) and in excess has been reported as both causing hypothyroidism (with histologic changes in the thyroid) and having adverse reproductive effects (testicular atrophy and alteration of ovarian function) (90, 91). Cobalt serves as a goitrogen with excessive intake in humans, inhibiting normal iodine uptake by the thyroid, which in turn creates excess iodine, inhibiting thyroid activity (37, 92). Cobalt functions as a systemic radical when unbound by albumin, by generating reactive oxygen species involved with lipid peroxidation (93). In the elasmobranch that does not have albumin, there is potentially increased exposure to the ‘radical’ form (45, 93). We found a range of thyroid disease scores associated with the <300 and > 300 ng/mL cobalt values, including normal thyroid with high serum cobalt. Mild, moderate, and severe thyroid evaluations were almost two times more frequent when cobalt was >300 ng/mL. Chickens will alter deposition of iodine in yolk in response to feeding of cobalt supplementation (46), supporting cobalt’s potential role in reproductive disease.

Cobalt is primarily used in the folate cycle in vitamin B12, with excess excreted in mammals. Orally, it has low toxicity, with 100-fold margin of safety from minimum requirements for feeding domestic species (47). Cobalt can bind iodine in systems of high iodine saturation (94, 95). Pharmacological use of cobalt impacts race horses, due to its chemical interaction with iodine, inhibiting the thyroid (96). Excess cobalt intake is documented to alter thyroid metabolism, and impact goiter development despite iodine sufficient diets in humans (97).

Cobalt was correlated with total iodine and TT3, and moderately so, with inorganic iodine, TT4 and FT4 (Figure 4), though not statistically significant. Elevated cobalt levels, as with iodine, appear to negatively impact the health of elasmobranchs. Theoretically, abnormally elevated levels of either or both iodine and/or cobalt appear to result in thyroid dysfunction. This in turn would result in a slow progression of ovarian dysfunction culminating in reproductive disease. With a higher incidence of both iodine and cobalt in animals with reproductive disease and thyroid pathology, further investigation should be explored.

Ultrasound is a powerful tool to evaluate an organ like the thyroid where it can provide antemortem data on thyroid size and appearance. Subtle lesions can be seen as well as an enlarging gland prior to seeing outward signs of a ventral swelling. Here, ultrasound was used to tease out differences of thyroid disease, as traditional goiter was seen in only a few cases, while other thyroid pathology was seen in many cases. This allowed for some broad categorization to help understand some of the dynamics of thyroid hormone, iodine and cobalt. An in-depth study on ultrasound size, echogenic character and descriptions of lesions would be ideal but not feasible in this study. Still, it is clear that ultrasonographic changes to the thyroid correlated with disease.

The degree of disease was only generally categorized. Further research and characterization may be worthwhile. Hormone and iodine levels were correlated with subjective evaluation of disease. Causality is difficult, but (as evidenced in the range of values observed in the aquarium-housed animals with normal thyroid) it is suspected that high levels of iodine preceded thyroid pathology, and high values for an extended period of time would likely be a factor. However, given the location and appearance of the thyroid, successful ultrasonography can be challenging. Additionally, a high resolution transducer is recommended as lesions (cysts and subtle nodules) can easily be missed if the correct technique and transducer are not used.

What is unknown, is what degree of percent damage must occur to the thyroid tissue before functionality is altered. This is true in other species where it has been stated that “there is no simple correlation between morphologic lesions and resultant clinical manifestations. A multinodular goiter, for example, in one instance may be associated with normal thyroid function, in another with hyperfunction, and yet another with hypofunction” (98). Overall, ultrasound proved to be either a good adjunct to characterization of severity of disease or an indicator of disease.

In the normal shark or ray, thyroid hormone values are low (TT3 < 3 nmol/L, TT4 < 5 nmol/L, and FT4 < 5 pmol/L; Figure 3). If animals were over supplemented or over consume (as indicated by elevated iodine and cobalt levels), all three thyroid parameters will be elevated. Paradoxically, in the most severe disease cases, TT3 dropped while TT4 and FT4 stayed elevated, which speculatively implies loss of functionality.

When looking at all the data together, high values in TT3, TT4, and FT4 positively correlated with degree of morphologic disease (either in size or thyroid lesions) and iodine with a proportional increase in cobalt until the disease became severe and then TT3 and FT4 dropped. The inorganic iodine and TT3 levels, when high, could ‘predict’ or specifically, be able to distinguish the moderate and moderate–severe disease categories. In the normal animals, uniformly, the iodines were low <300 with low TT4 and a TT3 of 1–3 nmol/L. An important factor is that the amounts of iodine and cobalt in the supplement decreased (by internal request), over several years. Concomitant to that is a general decrease in serum iodine and cobalt in the aquarium population. Serum cobalt was elevated when serum iodine was elevated, and over a short period of time, cobalt was proportionately higher than iodines when the supplement was initially lowered in iodine. Cobalt may take longer to reduce in the body than the iodine in this circumstance. Iodine and cobalt were also highest with females with reproductive disease, it is presumed this is due to iodine and cobalt accumulation in the ova (46, 80, 82). Low cobalt levels were found in animals with traditional goiters and in natural sea water habitats.

Goiters are a well known symptom of thyroid disease in elasmobranchs. Here it is shown that thyroid disease can present not only in enlargement of the gland, but also with lesions similar to that described in humans (8, 9). Thyroid disease was closely associated with reproductive disease in the aquarium study population. High iodine and cobalt levels were also associated with reproductive disease. Anecdotally, in the SR in this study, elevated reproductive steroid hormones concomitantly dropped when iodine levels also decreased (unpublished data). Additionally, it is worth noting that not all elasmobranchs get goiters and there are potentially species-specific susceptibilities (37).

Some changes are reversible. In the cases of classical goiters (e.g., eagle rays, Aetobatus spp.,), iodine supplementation indeed reversed the enlargement of the thyroid. In others, mineralizations and large cysts were not resolved over time, likely indicating permanent tissue alteration. Ultrasound would provide an excellent data set if collected over time by a uniform sonographer with high resolution linear transducers.

Limitations of this data set include a limited number of animals within species and disease categories and an inability to monitor and qualify consumption of supplementation directly in all cases. Additionally, there was a lack of a robust series of abnormal biological controls in terms of classic goiter to establish some low ranges of iodine. A future direction can include validation of TT4 CLIA especially at the test’s lower detection limits. Additionally, further research can be considered into the interplay of iodine, cobalt and thyroid hormone in elasmobranchs.