Part of this work was presented at the ACVIM Forum, Louisville, KY, in June 2006.
Corresponding author: Dan Rosenberg, DVM, PhD, Internal Medicine Unit of the National Veterinary School of Alfort, 7 avenue du Général de Gaulle, 94704. Maisons-Alfort, France; e-mail: email@example.com.
Background: Adrenal ultrasonography (US) in dogs with hyperadrenocorticism (HAC) is commonly used to distinguish adrenocorticotropic hormone (ACTH)-independent (AIHAC) and ACTH-dependent hyperadrenocorticism (ADHAC). To date, no cut-off values for defining adrenal atrophy in cases of adrenal asymmetry have been determined. Given that asymmetrical hyperplasia is sometimes observed in ADHAC, adrenal asymmetry without ultrasonographic proof of adrenocortical tumor such as vascular invasion or metastasis can be equivocal.
Objective: The purpose of this study was to compare adrenal US findings between cases of ADHAC and AIHAC in dogs with equivocal adrenal asymmetry (EAA), and to identify useful criteria for their distinction.
Animals: Forty dogs with EAA were included.
Methods: Ultrasound reports of HAC dogs with adrenal asymmetry without obvious vascular invasion or metastases were reviewed. Dogs were classified as cases of ADHAC (n = 28) or AIHAC (n = 19), determined by plasma ACTH concentration. The thickness, shape, and echogenicity of both adrenal glands and presence of adjacent vascular compression were compared between AIHAC and ADHAC groups.
Results: The maximal dorsoventral thickness of the smaller gland (SDV) ranged from 2.0 to 5.0 mm in AIHAC and from 5.0 to 15.0 mm in ADHAC. The 95% confidence intervals for estimated sensitivity and specificity of a SDV cut-off set at 5.0 mm in the diagnosis of AIHAC were 82–100 and 82–99%, respectively. Other tested US criteria were found to overlap extensively between the 2 groups, precluding their usefulness for distinction.
Conclusion and Clinical Importance: In EAA cases, an SDV ≤5.0 mm is an appropriate cut-off for AIHAC ultrasonographic diagnosis.
maximal dorsoventral thickness of the larger gland
receiver operating characteristic
maximal dorsoventral thickness of the smaller gland
Hyperadrenocorticism (HAC) or Cushing's syndrome is a commonly diagnosed endocrine disorder in dogs. Adrenocorticotropic hormone (ACTH)-dependent hyperadrenocorticism (ADHAC), mainly pituitary-dependent hyperadrenocorticism (PDH), is the most common form of HAC. Only 10–15% of HAC cases are ACTH-independent (AIHAC).1 The major form of AIHAC is triggered by an unilateral functional adrenocortical tumor (FAT) (adenoma or carcinoma). The distinction between AIHAC and ADHAC is crucial for prognosis and treatment.
Abdominal ultrasonography (US) currently is considered the “tool of choice” by many clinicians for determining the cause of HAC in dogs and is commonly used by veterinarians for such cases.2,3 Being common in veterinary practices, and not very time-consuming, US often is described as a reliable and practical method for identifying the origin of HAC.2,4 Ultrasound data obtained for the etiological characterization of HAC in dogs can be unequivocal. For instance, bilateral normal-sized or large adrenal glands with normal shape and echogenicity are considered strong evidence of adrenocortical hyperplasia secondary to ADHAC.4,5 Likewise, a large solitary and abnormally shaped adrenal mass invading adjacent structures or associated with distant nodules characterized by histopathology as adrenocortrical metastases is highly indicative of adrenocortical carcinoma in the context of HAC.6,7 However, apart from these typical ultrasonographic images, other images, referred to in this paper as “equivocal adrenal asymmetry” (EAA), can be more difficult to analyze for various reasons. First, adrenal asymmetry associated with an adrenal mass and contralateral atrophy is expected in unilateral FAT; however, to date, there is no established cut-off for defining adrenal atrophy. Cases of AIHAC caused by unilateral FAT have even been described with opposite gland thickness within reference values.8 Additionally, the appearance of FATs themselves can sometimes differ from the enlarged, rounded, or irregularly rounded shapes with small and homogeneous functional nodules described.7,8 Second, adrenocortical nodular hyperplasia involving one or both adrenal glands has been described in dogs with ADHAC generating adrenal asymmetry to a variable extent, changes in adrenal shape and, in some cases, heterogeneous echogenicity of the adrenal gland. In such situations, the images can be confused with those observed in cases of FAT.5
The purpose of this study was to analyze adrenal US data in HAC dogs with EAA and to retrospectively determine ultrasonographic criteria to differentiate AIHAC from ADHAC.
Materials and Methods
Medical records of dogs evaluated at the Internal Medicine Unit of the National Veterinary School of Alfort (France) between June 1, 2003, and November 1, 2008, were reviewed. We included all dogs with:
1History, physical examination, biochemistry, and hematological findings consistent with HAC.
2At least 1 endocrine test result, ACTH stimulation test or low-dose dexamethasone suppression test (LDDST), consistent with the diagnosis of HAC.
3An unequivocal characterization of HAC origin by plasma ACTH concentration measurement. Additional information for the characterization of HAC (feedback inhibition in response to dexamethasone suppression tests [DSTs], histopathology of adrenocortical tissue removed at surgery or necropsy) was collected when available.
4An ultrasonographic examination of the 2 adrenal glands allowing measurement of the maximal dorsoventral thickness and showing EAA without obvious vascular invasion or organ metastasis.
5No treatment of HAC started before the ultrasound examination.
Endocrine Tests and Hormone Assays
Serum cortisol concentrations for the dynamic endocrine tests were determined with a kita previously validated for dogs.9
The ACTH stimulation test, based on intravenous (IV) administration of 0.25 mg of tetracosactide,b the LDDST, based on IV administration of 0.01 mg/kg of body weight (BW) of dexamethasone,c and the high dose dexamethasone suppression test (HDDST), based on IV administration of 0.1 mg/kg BW of dexamethasone, were performed as described previously.10–14 Spontaneous HAC was confirmed in all dogs by either a marked increase in serum cortisol concentration (≥500 nmol/L) 1 hour after ACTH stimulation, an inadequate suppression of serum cortisol concentration (≥40 nmol/L) 8 hours after dexamethasone injection or both.
For the DSTs, feedback inhibition demonstrating ADHAC was defined as a 4 or 8-hour serum cortisol concentration <50% of the basal cortisol concentration, a 4-hour cortisol concentration ≤40 nmol/L or, for the HDDST only, an 8-hour cortisol concentration ≤40 nmol/L. In cases for which no decrease in cortisol was observed after DSTs, 4-hour and 8-hour plasma ACTH concentrations, additionally to basal ACTH, were measured to distinguish potential dexamethasone insensitivity of the ACTH-producing tissue from adrenocortical autonomy.
Blood samples for measurement of plasma ACTH concentration were collected from the jugular vein in ethylenediaminetetraacetic acid-coated tubes. Samples were centrifuged immediately at 4°C, at 500 ×g for 8 minutes, and plasma promptly stored in plastic tubes at −80°C until assay. Plasma ACTH concentrations were determined using Immulite ACTH,d an immunoluminometric assay, previously validated for dogs15 samples were run on the Immulite 2000 analyzer.e Dogs were considered to have AIHAC when ACTH concentration was below the limit of quantification of the assay (<5 pg/mL). ADHAC was identified by unsuppressed basal ACTH plasma concentrations (>6 pg/mL). Confidence intervals (CI) for estimated sensitivity and specificity of these criteria to detect ADHAC were previously demonstrated as ranging from 97 to 100% and from 85 to 100%, respectively.16
Ultrasonographic procedures were carried out by board-certified or equivalent experienced examiners, with a 5–8 MHz curved array transducer.f At the time of ultrasound examination, no information about the plasma ACTH concentration of the dogs was available to the examiner. The maximal dorsoventral thickness of the larger gland (LDV) and of the smaller gland (SDV) was measured from static images as described previously.17 Briefly, the maximal dorsoventral thickness was defined as the maximal dorsoventral dimension, which was perpendicular to the long axis of the adrenal gland. Measurements were obtained with the transducer in a subcostal rather than an intercostal position whenever possible. For each pair of measurements, the dorsoventral thickness difference ratio (DVTDR)—defined as the difference between LDV and SDV with respect to the mean [DVTDR = 2 × (LDV − SDV)/ (LDV + SDV)]—and the dorsoventral thickness ratio (DVTR)—defined as the ratio between LDV and SDV (DVTR = LDV/SDV)—were calculated. Adrenal asymmetry was arbitrarily defined before starting the study as DVTDR ≥ 0.2.
Each examiner subjectively described the shape, the echogenicity, and the relation with adjacent structures of the 2 adrenal glands. A regular bilobed shape for the left adrenal gland long axis and a smooth wedge to elongated ovoid shape for the right adrenal gland long axis were considered as “normal.”17“Nodular enlargement” referred to a round, well-defined focal bulge of the adrenal gland without loss of its global shape. Adrenal glands with close maximal dorsoventral thickness and maximal length measurements, resulting in a short ovoid to spherical appearance and complete loss of normal shape, were subjectively described as “rounded.” For both glands, echogenicity was considered “normal” when adrenal gland parenchyma appeared homogeneously hypoechoic to surrounding fat, or when a layered appearance delineating cortex and medulla was observed.17,18 Mixed echogenicities with focal areas of hyperechogenicity of adrenal parenchyma, or of the adrenal cortex when corticomedullary differentiation was possible, were designed as “heterogeneous.”5 Narrowing of vessels in contact to left or right adrenal glands without obvious vascular invasion was considered to be a sign of “vascular compression.”
For histologic examination, the adrenocortical tissue removed surgically or during necropsy was fixed in 10% neutral buffered formalin and routinely embedded in paraffin. Sections were stained with hematoxylin and eosin. Histopathological analyses were performed by board-certified pathologists. Adrenal tumors, when identified, were classified as benign or malignant. Tumor architecture, cell pleomorphism, mitosis frequency, and invasive behavior (ie, capsule or vascular invasion) were recorded. Malignancy was considered when at least evidence of invasive behavior was observed. Adrenal tumor functional status was considered when atrophy of the adjacent nontumoral adrenocortical tissue was noted.
Data were analyzed with commercially available software.g,h Data were not normally distributed and are reported as medians, ranges, and number of subjects. Continuous values obtained from the 2 groups were compared with the Mann-Whitney rank-sum test. Discontinuous values were compared with the Fisher's exact test. A Spearman correlation coefficient (r) was used to assess the correlation between variables. P < .05 was considered significant. A threshold for SDV was determined from the receiver operating characteristic (ROC) curve. The 95% CI for estimated sensitivity and specificity for SDV were determined by free software.i
Abdominal US was performed on 81 HAC dogs over the study period. US examinations were considered unequivocal in 34 dogs. These dogs were classified as having ADHAC on the basis of DVTDR < 0.2, confirmed by an unsuppressed plasma ACTH concentration. The remaining 47 dogs were classified as having EAA and were included in the study. Of these, 19 dogs had AIHAC and 28 dogs had ADHAC. Dogs in the ADHAC group were between 4 and 15 years old (median, 12 years) at the time of diagnosis. Thirteen were male and 15 were female. Their BW ranged from 2 to 36 kg (median, 9 kg). Dogs in the AIHAC group were between 6 and 16 years old (median, 10 years). Thirteen were female and 6 were male. Their BW ranged from 5 to 47 kg (median, 14 kg).
Basal plasma ACTH concentrations of dogs in the AIHAC group were undetectable (<5 pg/mL). In dogs with ADHAC, basal plasma ACTH concentrations ranged from 10 to 286 pg/mL (median, 32 pg/mL).
In the ADHAC group, 18 dogs were subjected to a DST, of which 16 showed feedback inhibition: cortisol suppression was observed in 5 dogs after LDDST and in 11 dogs after HDDST. In the remaining 2 dogs, the absence of a decrease in serum cortisol concentration was associated with an absence of suppression of plasma ACTH concentrations (detectable at baseline, 4 and 8 hours).
In the AIHAC group, 14 adrenal glands were removed by surgery (n = 9) or necropsy (n = 5) and examined by histopathology. No medical treatment for HAC was administered to dogs before removal. Functional tumors were identified for all 14 adrenal glands, 10 of which were carcinomas and 4 adenomas.
Ultrasonographic examination of the larger of the 2 adrenal glands disclosed nodular enlargement in 20 dogs in the ADHAC group and in 8 dogs with AIHAC (Table 1, Fig 1). A rounded shape of the whole adrenal gland was identified in 2 dogs with ADHAC and in 11 dogs with AIHAC (Fig 2). The basal ACTH concentrations of the 2 dogs with ADHAC and a rounded-shape adrenal gland were 21 and 74 pg/mL, respectively. The frequency of adrenal glands with rounded shape was significantly different between the 2 groups. Heterogeneous echogenicity was observed for 10 adrenal glands in the ADHAC group and for 10 adrenal glands in the AIHAC group (Fig 3). Echogenicity homogeneity did not differ between the 2 groups. Compression of adjacent vessels was identified in 3 adrenal glands from the ADHAC group (compressing aorta, caudal vena cava, and left renal vein, respectively) and in 6 adrenal glands from the AIHAC group (compressing caudal vena cava in 5 cases and left renal artery in 1 case). DST were performed on the 3 ADHAC cases and feedback inhibition was demonstrated after LDDST in 1 dog and HDDST in 2 dogs. Compression of adjacent vessels was not significantly different between the 2 groups.
Table 1. Comparison of the ultrasound appearance of the smallest and largest adrenal gland in AIHAC and ADHAC dogs with EAA.
Ultrasonographic examination of the smaller adrenal gland disclosed nodular enlargement in 2 dogs with ADHAC and heterogeneous echogenicity in 1 dog with ADHAC. Smaller adrenal glands had normal shape with normal echogenicity in all remaining dogs.
LDV measurements were statistically different between the ADHAC and AIHAC groups, with medians (range) of 12.7 (7.0–37.0) and 27.0 (9.8–88.0) mm, respectively (Fig 4). Three dogs of 19 in the AIHAC group had LDV measurements above the range observed in dogs with ADHAC. There was no correlation between LDV and BW in the 2 groups. SDV measurements were statistically different between ADHAC and AIHAC groups, with medians (range) of 7.5 (5.0–14.0) and 3.0 (2.0–5.0) mm, respectively (Fig 5). Eighteen dogs of 19 in the AIHAC group had SDV below the range observed for dogs with ADHAC. There was no correlation between SDV and BW in the 2 groups. DVTDR were statistically different between the ADHAC and AIHAC groups, with medians (range) of 47% (20–126%) and 157% (93–181%), respectively (Fig 6). Sixteen dogs of 19 in the AIHAC group had DVTDR above the range observed for dogs in the ADHAC group. There was no correlation between DVTDR and BW in the 2 groups. DVTR were statistically different for dogs in the ADHAC and AIHAC groups, with medians (range) of 1.6 (1.2–4.3) and 8.3 (2.7–19.7) mm, respectively (Fig 7). Sixteen dogs of 19 in the AIHAC group had DVTR above the range observed for the ADHAC group. There was no correlation between DVTR and BW in the 2 groups.
The ability to distinguish between AIHAC from ADHAC in EAA using SDV measurements was assessed with a ROC curve analysis (Fig 8). The area under the curve of the ROC curve was 0.999. The estimated sensitivity and specificity of SDV measurement delineating AIHAC from ADHAC were 95% (95% CI: 74–99%) and 100% (95% CI: 88–100%), respectively, for a threshold set at 4.5 mm, and 100% (95% CI: 82–100%) and 96% (95% CI: 82–99%), respectively, for a threshold set at 5.0 mm.
This study compares the US appearance of the adrenal glands in 47 dogs with HAC and EAA. Abdominal US is considered to be a reliable tool for distinction between ADHAC and AIHAC in dogs.2 Several criteria generally are applied, including homogeneity or heterogeneity of the adrenal glands, shape, compression of adjacent vessels, and symmetry.1 Symmetry of the adrenal glands is considered to be a crucial factor for this distinction, symmetry being indicative of ADHAC and asymmetry (reflecting an adrenal mass and contralateral adrenal atrophy) indicative of AIHAC.1 However, to the best of our knowledge, rigorous analyses of the accuracy of these criteria are still lacking. Furthermore, the concept of asymmetry itself is, to a certain extent, ill-defined: criteria for atrophy have not been defined, and a degree of asymmetry has been reported in reference papers describing adrenal ultrasound appearance in ADHAC.5,19
In this study, dogs were classified as having ADHAC or AIHAC based on basal plasma ACTH concentration. In dogs, Cushing's syndrome has long been thought to exist as only 2 forms—PDH or adrenal-dependent hyperadrenocorticism (ADH). Other forms of the disease (HAC because of ectopic ACTH secretion and food-dependent HAC) are well described in humans but have only been recently identified in dogs.20–22 Thus, we classified HAC according to criteria generally used in studies of humans, distinguishing between ACTH-dependent and ACTH-independent forms of HAC.23,24 Indeed, dogs with ADHAC could have either an ectopic or a pituitary origin of the syndrome, although a pituitary origin is the most common.
The period of enrollment of this study overlaps with the enrollment period of a previously published study evaluating ACTH accuracy for ADHAC and AIHAC discrimination in 109 dogs at the same center.16 The 2 studies are characterized by distinct inclusion criteria limiting the risks of inclusion of the same dogs. However, 4/28 ADHAC dogs in the present survey were also enrolled in the previous study evaluating ACTH accuracy.16 The 4 dogs exhibited a cortisol inhibition after DST demonstrating pituitary dependency. Likewise, 5/19 AIHAC dogs in the present study also were already described.16 The presence of a FAT (ie, for tumors removed at surgery or necropsy, atrophy of adjacent nontumoral adrenocortical tissue was observed; for tumors removed at necropsy, atrophy of the contralateral adrenal gland was observed) was demonstrated by histopathology for all of them.
In the “ACTH accuracy study,” published previously, no overlap was found between ACTH concentration in 91 dogs with ADHAC and 18 dogs with AIHAC, measured with the same technique.16 Given the narrowness of the 95% CI for estimated sensitivity and specificity of plasma ACTH concentration for the discrimination of HAC cause, it is unlikely in the current study that dogs were wrongly classified by ACTH measurement. Indeed, in the present study, the histological examination of adrenal glands from 14/19 AIHAC dogs was consistent with a FAT. Additionally, 18/28 dogs with ADHAC had a DST, 16 of which showed feedback inhibition demonstrating the pituitary origin of the syndrome.13 The absence of feedback inhibition in 2/18 dogs tested with DST could be attributed to FATs—highly improbable because of the unsuppressed basal ACTH concentration—to PDH characterized by resistance to dexamethasone suppression, to ectopic Cushing's syndrome, or to HAC caused by concurrent pituitary and adrenocortical tumors. The presence of both pituitary and adrenocortical tumors in HAC seems to be rare.25 In the largest study of such cases, it was found in only 17 of approximately 1,500 HAC dogs. Moreover, as suggested later by the authors of the study, confusion of concurrent pituitary and adrenocortical tumors with ADHAC with adrenocortical nodular hyperplasia may have resulted in over-inclusion (cortisol suppression in 5 dogs, and, in the case of absence of cortisol suppression, no measurement at the same time of both ACTH and cortisol concentrations over DST in order to analyze the involvement of ACTH-producing tissue and adrenal cortex).j In our study, the absence of ACTH suppression in the 2 dogs does not demonstrate or exclude adrenal autonomy. However, given the frequency of ADHAC compared with that of concurrent pituitary and adrenocortical tumors, a single ACTH-producing origin—pituitary gland probably—remains the most likely explanation for these 2 cases.
We only included dogs for which adrenal US was carried out before the start of HAC treatment. Indeed, both trilostane and mitotane are associated with changes in adrenal gland size and appearance26–29; thus such treatment could generate bias in the observations.
The maximal dorsoventral thickness of the adrenal glands has been found to be the only dimension that could be measured in a reliable way with US.17,19,30 Measurement by US of other dimensions such as width and length has been shown not to be correlated with gross measurements.17 The relevance of other quantitative determinations that might be reliable indicators of adrenal gland size such as adrenal volume has still to be evaluated. In the context of this retrospective study, we used the maximal dorsoventral thickness measure for both glands, focusing on EAA. We arbitrarily excluded dogs with <20% difference between the 2 glands and those showing local invasion or organ metastasis consistent with an adrenocortical tumor. Such cases generally are considered unequivocal by clinicians, being indicative of ADHAC and adrenocortical carcinoma, respectively. Among the 81 HAC dogs that were examined by US over the study period, 47 dogs were included. Even if, to our knowledge, the frequency of EAA in dogs has never been studied so far, this frequency probably reflects a selection bias possibly linked to the particularity of the referral population recruited in our center.
In this study, we identified a significant difference in LDV between ADHAC and AIHAC groups. However, a major overlap between the measurements was observed for the 2 groups. A previous study indicated that a dorsoventral thickness >20.0 mm is more likely to reflect the presence of an adrenal tumor than hyperplasia.7 However, there are exceptions; ADHAC cases have been reported with adrenal gland diameter measuring ≥20.0.31 Our study confirms that such a criterion could lead to some cases being wrongly classified, with a LDV between 20.0 and 35.0 mm observed in 5/28 dogs in the ADHAC group. The ADHAC dog with a LDV of 35.0 mm elicited a feedback inhibition response after HDDST. A few days after diagnosis, the dog was euthanized, for reasons unrelated to HAC. The histological examination of the pituitary and adrenal glands showed an adenoma of the pars distalis, homogeneous hyperplasia of the smaller adrenal gland (SDV 5.0 mm), and nodular hyperplasia of the enlarged adrenal gland.
The ACTH-secreting tissue and adrenal cortex were not available for histological analysis in many other cases in the ADHAC group because many of them were still alive, and under medical treatment. In the absence of histological analysis, potential extra-adrenocortical pathological features (eg, pheochromocytomas) contributing to adrenal enlargement cannot be completely ruled out, although they appear unlikely because of their scarcity.32 Functional tumors with LDV < 20.0 mm have been described relatively often.4,7,8 This study is consistent with previous findings, with 5/19 AIHAC dogs showing LDV < 20.0 mm. The overlap in LDV between the 2 groups shows that use of LDV as the only tool for distinguishing AIHAC and ADHAC would lead to a high rate of misclassification.
In this study, the range of SDV measurements overlapped between the ADHAC and the AIHAC groups by only 1 measurement: 1 dog from each group was found to have an SDV of 5.0 mm. Indeed, in ADHAC, 27/28 SDV measurements were >5.0 mm, ranging from 5.4 to 14.0 mm. The results are in agreement with those of a previously published study in which both adrenal gland widths were measured in 28 dogs with HAC.4 In 19/21 dogs classified as having ADHAC, the maximal width of the smallest adrenal gland ranged between 5.2 and 14.3 mm. However, in the same study, 5 dogs classified as having AIHAC had both adrenal gland widths measured, and only 1 had a maximal width of the smallest adrenal gland <5.0 mm (4.3 mm) whereas in the present study, 18/19 SDV of dogs with AIHAC measured between 2.0 and 4.5 mm. In the previous study, allocation of the dogs to AIHAC group was based on indirect criteria (eg, resistance to mitotane treatment, radiographic signs compatible with an adrenal tumor without histological proof of its adrenocortical nature or visualization of images compatible with metastasis without histological proof of their adrenocortical nature) allowing doubts about correct classification.4 However, undetectable basal plasma ACTH concentration was identified in the 5 dogs as in the dogs of the present study limiting the risks of misclassification, and the reasons for these differences between the 2 studies remain obscure.
In addition, the observation of a SDV < 5.0 mm in most of the dogs with AIHAC does not agree with another study focusing on the ultrasonographic appearance of both adrenal glands in 15 dogs with AIHAC caused by FATs.8 Unilateral FAT was identified in 12/15 dogs, of which only 7 had a SDV < 5.0 mm. However, classification of AIHAC in this previous study was based on the absence of inhibition feedback after LDDST. Because this absence also is observed in approximately 40% of ADHAC cases,13 the inclusion of only dogs with AIHAC was not guaranteed without histological confirmation of FAT or ACTH measurement. In the same study, 3/15 dogs were considered as having 2 FATs, a condition only documented in approximately 10 cases in the literature.8,32,33 In the absence of histological analysis, confusion with cases of ADHAC with nodular adrenocortical hyperplasia remains possible. In our study, the absence of double adrenocortical tumors in the AIHAC group probably is linked to the low prevalence of the condition. It could also be explained by the inclusion criteria including only DVTDR ≥ 20%, as asymmetry is less likely in bilateral than in unilateral FAT even if cases of bilateral tumors characterized by a certain extent of asymmetry have been described.8,32
This study suggests that SDV measurement is reliable for ADHAC and AIHAC distinction in dogs with EAA: a SDV < 5.0 mm supports AIHAC. The autonomous secretion of cortisol by FAT suppresses pituitary secretion of ACTH. This suppression leads to atrophy of the opposite adrenal cortex.1,34 It highlights the importance of evaluating both adrenal glands for the characterization of HAC.1 Over a long time, visualization of nonenlarged adrenal glands was not guaranteed as many of the studies related to FAT did not give a description of some of the contralateral glands.31,33,35 This study has benefited from improvements in ultrasound equipment and skills, allowing the systematic evaluation of the 2 glands for the diagnosis of HAC origin.1,2 This study provides, for the first time, cut-offs allowing the distinction of AIHAC and ADHAC in cases of EAA. Using the threshold maximizing AIHAC sensitivity (SDV ≤ 5 mm) the 95% CIs for estimated sensitivity and specificity for AIHAC determination were 82–100 and 82–99%, respectively. Despite the narrow range of specificity and sensitivity values, the observation of adrenal asymmetry with SDV around the cut-off remains equivocal. We therefore used ratios depicting the extent of adrenal asymmetry as a further measurement in such cases. Although statistically different between the 2 groups, DVTR and DVTDR values overlapped between the 2 groups. The use of these 2 ratios didn't allow adequate classification of the 2 cases with SDV = 5.0 mm (data not shown). Other ratios—(LDV / BW, SDV / BW, LDV − SDV) / SDV and (LDV − SDV) / LDV—also were tested without benefit as well (data not shown).
Qualitative criteria related to shape, parenchymal echogenicity, and the effect on surrounding structures also were tested. Because these data were analyzed retrospectively from US reports, bias may have been introduced by under-counting and inaccuracies in the assessment of such subjective criteria. As a matter of fact, although being commonly used in the literature, many of these criteria (eg, nodular enlargement, rounded shape, heterogeneity of the echogenicity of adrenal glands) have not been precisely defined.4,5,8,18 Without demonstration of the reproducibility of these qualitative criteria, risk of disparity of their descriptions exists for each examiner as well as among the different examiners of the present study.
Despite these limitations, this study shows that these data provide inadequate levels of specificity in the AIHAC and ADHAC groups. Heterogeneous parenchyma and focal areas of increased echogenicity previously have been described in adrenal glands of various shapes (normal, nodule, mass), harboring nodular adrenocortical hyperplasia.5,7 This study confirms these observations and suggests that these criteria should not be considered as accurate indicators of tumoral origin for HAC. Moreover, this study describes for the first time compression by nodules of adjacent vessels in ADHAC (with ACTH dependency confirmed by plasma ACTH concentration measurement and by cortisol decrease after DST), and shows its limited specificity for determining the presence of tumors.1
In conclusion, this study highlights the importance of visualization of the atrophied contralateral gland for accurate diagnosis of AIHAC in cases of adrenal asymmetry. It provides, for the first time, a cut-off for SDV of 5.0 mm, allowing the distinction between AIHAC and ADHAC with high sensitivity and specificity in dogs with EAA. Although statistically different between ADHAC and AIHAC, the use of other criteria such as the LDV, the extent of the asymmetry between the 2 glands, changes in echogenicity, and the compression of adjacent vessels is hampered by overlap between the 2 forms of HAC.
aEnzymun-Test cortisol, Roche Diagnostics, Meylan, France
bSynacthène injectable, tetracosactide, 0.25 mg/mL, Novartis Pharma SA, Rueil-Malmaison, France
cDexadreson, dexamethasone phosphate, 2 mg/mL, Intervet—Shering Plough, Beaucouzé, France
dImmulite ACTH, Siemens Medical Solutions Diagnostics, Los Angeles, CA
eImmulite 2000, Siemens Medical Solutions Diagnostics
fPhilips Advanced Technology Laboratories HDI 3500, Philips, Eindhoven, The Netherlands
gMicrosoft Office Excel 2003 software, Microsoft Corporation, Redmond, WA
hSAS/STAT software, version 5.0, SAS Institute Inc, Cary, NC
iMedCalc software, version 10.0.1, Mariakerke, Belgium
jPeterson ME. Concurrent pituitary and adrenal tumors in dogs with Cushing's syndrome. Proceedings 19th Annual Forum of the ACVIM, Denver, CO 2001 (abstract)
This work was supported by postdoctoral grants from Intervet Pharma R & D SA, Angers, France, awarded to Ghita Benchekroun, and from the Consellería de Innovación e Industria (Xunta de Galica) awarded to María Isabel Rodríguez Piñeiro and by grants from the Direction Générale de l'Enseignement et de la Recherche (Ministère de l'Agriculture et de la Pêche).