The Renin-Angiotensin-Aldosterone System in Greyhounds and Non-Greyhound Dogs

Background

The renin‐angiotensin‐aldosterone system (RAAS) regulates blood pressure, electrolyte homeostasis, and renal function. Blood pressure, serum sodium concentrations, and urinary albumin excretion are higher in Greyhounds than other purebred and mixed‐breed dogs.

Hypothesis

Alterations in the RAAS in Greyhounds are associated with hemodynamic and clinicopathologic differences observed in the breed.

Animals

Clinically healthy Greyhound and non‐Greyhound dogs consecutively enrolled as blood donors (n = 20/group).

Methods

Prospective study. Standard chemical analysis was performed on serum and urine. Serum angiotensin‐converting enzyme (ACE) activity was determined by fluorometric assay. All other RAAS hormones were determined by radioimmunoassay. Symmetric dimethylarginine (SDMA) was measured by immunoassay. Measurements were compared to blood pressure and urine albumin concentration. Data are presented as mean ± SD or median, range.

Results

Serum creatinine (1.5 ± 0.2 vs 1.0 ± 0.1 mg/dL, P < .001), sodium (149, 147–152 vs 148, 146–150 mEq/L, P = .017), and SDMA (16.1 ± 2.9 vs 12.2 ± 1.8 μg/dL, P < .001) were significantly higher in Greyhounds versus non‐Greyhounds, respectively. Plasma renin activity (0.69, 0.10–1.93 vs 0.65, 0.27–2.93 ng/mL/h, P = .60) and ACE activity (4.5, 2.1–8.5 vs 4.6, 2.1–11.4 activity/mL; P = .77) were similar between groups and did not correlate with higher systolic pressures and albuminuria in Greyhounds. Plasma aldosterone concentration was significantly lower in Greyhounds versus non‐Greyhounds (11, 11–52 vs 15, 11–56 pg/mL, respectively, P = .002).

Conclusions and clinical importance

Basal RAAS activation did not differ between healthy Greyhounds and non‐Greyhounds. Lower aldosterone concentration in Greyhounds is an appropriate physiologic response to higher serum sodium concentration and blood pressure, suggesting that angiotensin II effects in the renal tubule predominate over those of aldosterone.

Hyperhomocysteinemia in Greyhounds and its Association with Hypofolatemia and Other Clinicopathologic Variables

Background

Folate and cobalamin are essential cofactors for homocysteine (HCY) metabolism. Hyperhomocysteinemia, a multifactorial condition, may reflect B vitamin deficiency and is associated with increased risk of cardiovascular disease, thrombosis, and neurodegenerative and chronic gastrointestinal diseases in humans. Hyperhomocysteinemia has been reported in Greyhounds with suspected chronic enteropathy.

Objectives

To evaluate the frequencies of and the association between hypofolatemia and hyperhomocysteinemia in Greyhounds.

Animals

Data and serum samples from 559 Greyhounds.

Methods

Nested case‐control study. The frequency of hypofolatemia in Greyhounds was determined by a laboratory database search. The relationship between hyperhomocysteinemia (measured by gas chromatography‐mass spectrometry) and hypocobalaminemia and hypofolatemia was evaluated, and its frequency compared between healthy Greyhounds and Greyhounds with thrombosis or chronic diarrhea.

Results

Hypofolatemia was identified in 172 of 423 (41%) Greyhounds and was more common in hypo‐ than in normocobalaminemic dogs (49% vs. 35%; P = .0064). Hyperhomocysteinemia was detected in 53 of 78 (68%) of Greyhounds, being more common in hypo‐ than in normofolatemic dogs (88% vs. 59%; P = .0175). All healthy Greyhounds, 21 of 30 (70%) of dogs with chronic diarrhea and 6 of 8 (75%) of those with thrombosis, were hyperhomocysteinemic. Serum HCY concentrations were inversely correlated with serum folate concentration (ρ = −0.28; P = .0386) and were positively associated with serum albumin concentration (ρ = 0.66; P = .0022).

Conclusions and Clinical Relevance

Hyperhomocysteinemia occurs frequently in the Greyhound population. Its association with hypofolatemia suggests decreased intracellular availability of B vitamins, but the functional implications warrant further investigation. Hyperhomocysteinemia in Greyhounds potentially may serve as a spontaneous canine model to further investigate hyperhomocysteinemia in humans.

Plasma Vasoprotective Eicosanoid Concentrations in Healthy Greyhounds and Non-Greyhound Dogs

BACKGROUND

Hypertension and albuminuria often coexist in Greyhounds, suggesting generalized vascular dysfunction that could contribute to the development of a variety of diseases in this breed. Eicosanoid metabolites of arachidonic acid (AA) mediate endothelial function, vascular reactivity, and proteinuria in humans and in rodent models.

HYPOTHESIS

The eicosanoid profile of Greyhounds is shifted toward metabolites that promote vascular dysfunction, hypertension, and proteinuria.

ANIMALS

Healthy Greyhounds (n = 20) and non-Greyhound (n = 20) dogs that were consecutively enrolled in a blood donor program.

METHODS

Prospective study. Plasma eicosanoid metabolites were assayed by liquid chromatography/electrospray ionization mass spectrometry (LC/ESI/MS) and compared to systolic blood pressure (SP) measurements and urine albumin concentration.

RESULTS

Isomers of hydroxyeicosatetraenoic acid (HETE) were higher in Greyhounds than non-Greyhounds (median, range in pmol/mL: 5(S)HETE 19.82, 8.55-32.95 versus 13.54, 4.33-26.27, P = .033; 8(S)HETE 9.39, 3.28-19.84 versus 5.80, 2.25-17.66, P = .002; 9(S)HETE 9.46, 2.43-13.79 versus 5.82, 1.50-17.16, P = .026; 12(S)HETE 10.17, 3.81-40.06 versus 7.24, 2.9-16.16, P = .022). Dihydroxyeicosatrienoic acid (DHET) isomers also were higher in Greyhounds compared to non-Greyhounds (mean ± SD in pmol/mL: 8,9DHET 5.78 ± 2.13 versus 4.03 ± 1.36, P = .004; 11,12DHET 11.98 ± 2.86 versus 8.90 ± 3.48, P = .004; 14,15DHET 7.23 ± 2.19 versus 5.76 ± 1.87, P = .028). Albuminuria correlated with total DHET (rs = 0.46, P = .003). SP was positively correlated with 11,12EET (rs = 0.42, P = .006) and 20(S)HETE (rs = 0.38, P = .017). SP and 8,9EET were inversely correlated (rs = -0.49, P = .001).

CONCLUSIONS AND CLINICAL IMPORTANCE

Plasma eicosanoid profile in Greyhounds was consistent with activation of metabolic pathways known to promote vascular dysfunction and might contribute to higher blood pressures and albuminuria. Inhibition of these eicosanoid pathways should be evaluated as therapeutic targets in Greyhounds.

Markers of iron metabolism in retired racing Greyhounds with and without osteosarcoma

BACKGROUND

Greyhounds have well-described clinicopathologic idiosyncrasies, including a high prevalence of osteosarcoma (OSA). Hematocrit, HGB, and HGB oxygen affinity are higher than in other dogs, while haptoglobin concentration is lower, so we hypothesized that Greyhounds have a different iron metabolism. To our knowledge, there are no reports on serum iron profiles in Greyhounds.

OBJECTIVES

To elucidate iron metabolism in Greyhounds, we wanted to compare serum iron concentration, total iron-binding capacity (TIBC), and percent transferrin saturation (%SAT) in healthy retired racing Greyhounds (RRGs) with OSA (RRGs - OSA), and also with non-Greyhounds (NGs), without and with OSA (NGs - OSA).

METHODS

Serum iron concentration and unsaturated iron-binding capacity (UIBC) were measured by standard methods, and TIBC and %SAT were calculated in RRGs (n = 25), RRGs - OSA (n = 28), NGs (n = 30), and NGs - OSA (n = 32).

RESULTS

TIBC was lower in RRGs than in NGs (P < .0001), and in RRGs - OSA than in NGs - OSA (P < .0001). NGs - OSA had lower TIBC than healthy NGs (P = .003). Percent SAT was higher in RRGs than in NGs (P < .0001) and in RRGs - OSA (P = .008), and %SAT was also lower in NGs than in NGs - OSA (P = .004). Percent SAT was also higher in RRGs - OSA than in NGs - OSA (P = .001). Both RRGs - OSA (P = .02) and NGs - OSA (P < .0001) had lower serum iron concentrations than their healthy counterparts.

CONCLUSION

Lower TIBC and higher %SAT may constitute another Greyhound idiosyncrasy compared with other dogs. In this study, all dogs with OSA had higher serum iron concentrations and %SAT than healthy dogs.

Effects of racing on reticulocyte concentrations in Greyhounds

BACKGROUND

Greyhounds have several hematologic variables that are outside of the respective reference intervals of other dog breeds. In addition, increases in HCT, total protein and HGB concentration, and RBC and WBC counts occur immediately after exercise; these values return to resting values within a few hour after racing.

OBJECTIVE

This study evaluated the effects of exercise on the concentration of reticulocytes in circulating blood in racing Greyhounds. We hypothesized that reticulocyte numbers are significantly increased immediately after a race, and return to baseline within one to 2 h postrace.

METHODS

Fifty actively racing Greyhounds at the Wheeling Island Racetrack and Casino were included in the study. Samples were collected by jugular venipuncture one day prior to racing at the kennel (resting), immediately after racing, and one to 2 h after the race (recovery). Reticulocyte counts were determined with an IDEXX ProCyte Dx Hematology Analyzer (IDEXX Laboratories, Inc., Westbrook, ME, USA). Due to a nonparametric distribution, the results were statistically compared using the Friedman test.

RESULTS

Reticulocyte concentrations were significantly different among the 3 sample collection times (P < .0001). There was a significant increase in reticulocyte concentration immediately after racing (P < .001); one to 2 h after racing, the reticulocyte numbers decreased significantly (P < .001) to counts comparable to resting samples.

CONCLUSION

The increase in reticulocyte concentration is probably related to splenic contraction secondary to the release of catecholamines, although premature bone marrow release could also account for these changes. Thus, it is important to consider a Greyhound's activity and degree of excitement when interpreting selected hematologic data in a clinical setting.

Haematological and biochemical values in North American Scottish deerhounds

OBJECTIVE

Sighthounds, including deerhounds, have unique physiological traits that result in laboratory test results that may lie outside reference intervals for the general dog population. Although reference intervals for most analytes are thought to be similar among sighthounds, breed-specific reference intervals are available mainly for greyhounds. The aim of this study was to establish reference intervals for haematology and serum biochemical profiles in deerhounds.

METHODS

Venous blood samples were collected from healthy deerhounds. Haematological and biochemical analytes were examined and reference intervals were established using the 5th and 95th percentiles.

RESULTS

The reference intervals obtained from 96 dogs for platelets, reticulocytes, total thyroxine, chloride, gamma glutamyl transferase, bilirubin and glucose were lower than the general dog population. Reference intervals for mean cell volume, potassium, urea, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase and cholesterol were higher than the general dog population. Reference intervals for eosinophils and globulin were wider than that of the general population.

CLINICAL SIGNIFICANCE

These results confirm that differences in haematological and biochemical values exist in the deerhound. Some appear to be shared by all sighthounds but others may be unique to this breed.

Clinical pathology of Greyhounds and other sighthounds

Owing to the development of Greyhounds as racing sighthounds, these dogs have acquired unique physiologic adaptations that distinguish them from other breeds. Reference intervals for many analytes in retired racing Greyhounds (RRGs) differ from those of other breeds; most of the hematologic differences have also been described in other sighthounds. In this review, we provide a survey of the literature on clinical pathology of Greyhounds and other sighthounds and results of laboratory testing, including analysis of CBCs, biochemical profiles, coagulation tests, and blood gases, in RRGs at The Ohio State University. Major clinicopathologic differences in this breed include higher RBC mass, creatinine concentration, glomerular filtration rate, activities of hepatic enzymes, and concentration of cardiac troponin, as well as lower WBC, neutrophil, and platelet counts, thromboelastographic values, and concentrations of serum haptoglobin, total globulins, and T4.

White-coat effect on systemic blood pressure in retired racing Greyhounds

BACKGROUND

Greyhounds are known to have a higher systemic arterial blood pressure (BP) than non-Greyhound dogs.

OBJECTIVES

The purpose of this study was to determine whether the high systemic BP was because of the white-coat effect.

ANIMALS

Twenty-two healthy retired racing Greyhounds (RRG) enrolled in a blood donation program.

MATERIALS/METHODS

We prospectively measured systemic BP in 3 environments: in the hospital by the investigator (Hosp), in the home by the investigator (H/I), and in the home by the owner (H/O). Five serial measurements of systolic, diastolic, and mean arterial pressures (SAP, DAP, MAP) as well as heart rate (HR) were measured by an oscillometric method on the distal forelimb and distal hind limb in all 3 environments.

RESULTS

There was a significant difference for SAP, MAP, and HR between the Hosp and both H/I and H/O (P < .001); there were no significant differences for any of the parameters between the H/I and H/O environments. HR, but not SAP, MAP, or DAP (P < .05) decreased in RRG with multiple hospital visits for blood donation before this study. The hind limb SAP was significantly higher than the forelimb SAP (P < .05).

CONCLUSIONS AND CLINICAL IMPORTANCE

We conclude that the high SAP, MAP, and HR seen in the hospital setting are likely because of a white-coat effect. Furthermore, consideration should be given to defining the parameters of normal BP in RRG according to the environment in which they are obtained.

Serum cardiac troponin I concentration in retired racing greyhounds

BACKGROUND

Cardiac troponin I (cTnI) is a polypeptide found specifically in cardiac muscle tissue that has been used as a diagnostic and prognostic indicator of cardiomyopathy. Increases in cTnI are associated with myocardial pathologic processes. However, high serum cTnI concentrations have been observed in normal Greyhounds.

HYPOTHESIS

We hypothesized that Greyhounds have cTnI concentrations higher than non-Greyhound dogs, and that a separate reference range should be established for Greyhounds.

ANIMALS

Blood samples were collected from the jugular vein from a group of 20 healthy Greyhound blood donors.

METHODS

Analysis of serum cTnI was performed with an immunoassay system with a detection level of 0.01 ng/mL, as described previously. The Greyhound values were compared with 2 groups of Boxers with and without arrhythmogenic right ventricular cardiomyopathy (ARVC), and to a group of non-Boxer control dogs from a previous study.

RESULTS

The mean cTnI concentration in Greyhounds was significantly higher (P < .0001) than that in non-Greyhound control dogs, although not significantly different from normal Boxers (P= .50), or Boxers with ARVC (P= .58). Greyhound serum cTnI concentrations were in the range found in Boxers with ARVC. The proposed reference range for cTnI in Greyhounds is 0.05-0.16 ng/mL.

CONCLUSIONS AND CLINICAL IMPORTANCE

Greyhounds have a reference range for serum cTnI concentrations that differs from that of other previously published reference ranges for dogs of other breeds. Until a broader database and more precise reference range can be established, caution should be exercised in interpreting serum cTnI concentrations in Greyhounds with suspected cardiac disease.

Effects of time of collection and sample processing on von Willebrand factor concentration in retired racing greyhounds

BACKGROUND

Concentrations of von Willebrand factor (vWF) are less than 30% in approximately 10% of Greyhounds.

HYPOTHESIS

That sample collection, processing, and storage can affect the concentration of vWF in plasma of Greyhounds.

ANIMALS

Nineteen healthy former racing Greyhounds.

METHODS

Prospective study: Blood samples were obtained from the jugular vein of dogs at 4 times during the day. Samples were divided and processed in each of 3 ways. The 1st tube was centrifuged immediately, the 2nd tube was kept in a household refrigerator, and the 3rd tube was kept at room temperature for 3 hours before centrifugation and processing.

RESULTS

There were no significant differences in the vWF concentration between different sample storage and processing (P = 5.31). There was no statistically significant diurnal variation in vWF concentration in the samples evaluated (P = .37).

CONCLUSIONS AND CLINICAL IMPORTANCE

Time of sample collection and short-term storage temperature do not influence the vWF concentration in retired racing Greyhounds.

Effects of 2 concentrations of sodium citrate on coagulation test results, von Willebrand factor concentration, and platelet function in dogs

BACKGROUND

Blood collection tubes containing 3.2% (0.109 M) sodium citrate, instead of 3.8% (0.129 M) sodium citrate, have recently become available in the United States. These tubes are visually indistinguishable from the traditional 3.8% sodium citrate tubes, except for wording on the label. Consequently, samples for hemostatic evaluation are frequently collected in tubes containing the lower concentration of sodium citrate.

HYPOTHESIS

Results of hemostasis assays are different in samples collected in 3.2% versus 3.8% sodium citrate.

ANIMALS

Twenty healthy dogs.

METHODS

This study aimed at determining whether results of standard coagulation tests, von Willebrand factor concentration (vWF:Ag), and platelet function with the platelet function analyzer PFA-100a were affected by the different concentrations of sodium citrate. Blood samples were collected in tubes containing either 3.2% or 3.8% sodium citrate concentrations and processed routinely for coagulation assays (one-stage prothrombin time [OSPT], activated partial thromboplastin time [aPTT], fibrinogen concentration, and platelet count), vWF:Ag, and platelet function assays with a PFA-100.

RESULTS

There was no significant difference between samples collected in 3.2% versus those collected in 3.8% sodium citrate for OSPT, aPTT, fibrinogen concentration, platelet count, or vWF:Ag. The closure times with collagen/adenosine diphosphate were significantly shorter (66 +/- 8.1 versus 74.8 +/- 9.7 seconds; P < .0001) with the 3.2% than with 3.8% sodium citrate concentration, and the hematocrit was significantly higher (47.9 +/- 5.6 versus 46.0 +/- 4.7 seconds; P = .03) in samples collected in 3.2% than in those collected in 3.8% sodium citrate.

CONCLUSIONS AND CLINICAL IMPORTANCE

There is no clinically relevant effect of collection of blood into 3.2% or 3.8% sodium citrate.

Evaluation of platelet aggregation using a point-of-care instrument in retired racing Greyhounds

Veterinarians involved in Greyhound rescue have anecdotally observed that 10-15% of Greyhounds bleed profusely after simple surgical procedures. In most patients, platelet counts and hemostasis profiles are normal; therefore, it is possible that these dogs have platelet dysfunction. The PFA-100 is a novel point-of-care platelet function analyzer that has recently been evaluated as a rapid method to assess platelet function in dogs. The objectives of this study were to characterize platelet function in a group of healthy Greyhounds by means of the PFA-100. Blood samples were collected from the jugular vein from 30 healthy Greyhounds. CBC, biochemical profile, PFA-100 assay with collagen/epinephrine (COL-EPI) and collagen/ adenosindiphosphate (COL-ADP), plasma von Willebrand factor antigen concentration (vWF:Ag), and vWF collagen-binding assay (vWF:CBA) were performed. PFA-100 closure times (CTs) with COL/ADP ranged from 63 to 92 seconds (mean +/- SD, 74.7 +/- 7.9 seconds) and with COL/EPI from 87 to 238 seconds (138 +/- 41 seconds); vWF: Ag ranged from 22 to 120% (87.52 +/- 25.5%) and vWF: CBA ranged from 36 to 102% (77.4 +/- 17.3%); and platelet counts ranged from 147 to 265 x 10(9)/L (194.6 +/- 31.64 x 10(9)/L). Greyhound CTs were significantly shorter than CTs in a mixed population of 50 healthy non-Greyhound dogs, in which the COL/ADP CTs ranged from 61 to 172 seconds (mean +/- SD, 87 +/- 21.6 seconds), and the COL/ EPI CTs ranged from 81 to 300 seconds (mean +/- SD, 183 +/- 67.6 seconds; P = 0.005 for COL/ADP CT; P = 0.001 for COL/ EPI CT). Also, platelet counts were significantly lower (P = 0.001) and packed cell volume was significantly higher (P = 0.001) in the Greyhound than in the non-Greyhound group. The PFA-100 is a reproducible method that can be used in the clinical setting to assess platelet function in Greyhounds; however, normal CTs in healthy Greyhounds are shorter than in other breeds. The results obtained in this study will be used to screen for abnormal platelet function in Greyhounds with postoperative bleeding.

Left basilar systolic murmur in retired racing greyhounds

Nineteen of 28 (67%) Greyhounds enrolled in the Blood Donor Program at The Veterinary Teaching Hospital, The Ohio State University (Columbus, OH), had a left basilar systolic murmur. Ten Greyhounds with murmurs and 9 without murmurs were evaluated to gain knowledge about the pathogenesis of this murmur. Echocardiograms were performed without sedation by means of a GE Vivid 7 Echocardiographic System with a continuous ECG; systolic arterial blood pressure (SABP) was measured with an Ultrasonic Doppler Flow detector model 811-B. The mean peak aortic velocity in the Greyhounds with murmurs (2.15 m/s; range, 1.8-2.2 m/s) was significantly higher than in the Greyhounds without murmurs (1.89 m/s; range, 1.6-2.0 m/s) (P < .001); there were no significant differences between groups for aortic valve or annulus diameter, fractional shortening, pulmonic velocity, SABP, hematocrit, serum protein concentration, or red blood cell counts. In this study, Greyhounds with soft, left basilar systolic murmurs had mildly (but significantly) higher mean peak aortic velocities than similar dogs without murmurs. In the dogs with murmurs (and higher velocities), we could not identify structural abnormalities, such as valvular lesions or other congenital defects. There was no inverse correlation between the systolic murmur and the higher hematocrit and red blood cell counts observed in this breed. This 1-2/6 basilar systolic murmur is common in Greyhounds, and it does not appear to be of any clinical consequence.

Effects of blood donation on arterial blood pressure in retired racing Greyhounds

The purpose of this study was to evaluate changes in systolic arterial blood pressure (SABP) immediately after collection of blood for transfusion in retired racing Greyhounds. We prospectively evaluated 19 blood donor Greyhounds before and after the collection of a unit (450 mL) of blood. The SABP was measured with Doppler in the right forearm after the dogs had been in the blood collection room for a few minutes (PRE-FLOOR) and again 5-10 minutes after the dogs were placed on the table where they would be bled (PRE-TABLE). A total of 3-5 minutes after completing the blood collection, the SABP was measured again while the dogs were still in lateral recumbency on the table (POST-TABLE) and once more 60-90 minutes later, when the dogs were on the floor after completing the donation (POST-FLOOR). All dogs were monitored for clinical signs of hypotension, including depression, weakness, collapse, and pallor, for a minimum of 2 hours after donation. There was a significant difference in SABP for the group between PRE-FLOOR and POST-TABLE (P = .02) and between PRE-TABLE and POST-TABLE determinations (P = .01). There were no significant differences for any of the other time points; there were no adverse events. Therefore, we conclude that the collection of 450 mL of blood from normal Greyhounds results in a short-lived yet significant decrease in SABP, but the likelihood of adverse events is negligible.

Morphologic characterization of specific granules in Greyhound eosinophils

BACKGROUND

"Vacuolated" eosinophils (ie, eosinophils with empty, nonstaining granules) have been described previously in normal Greyhounds. However, to our knowledge, detailed studies of granules in vacuolated and normal eosinophils in this breed have not been performed.

OBJECTIVE

The objective of this prospective study was to characterize some of the morphologic, ultrastructural, and cytochemical staining features of specific (primary) granules in both normal and vacuolated eosinophils in Greyhound blood.

METHODS

Morphologic features of eosinophils in Wright's- and Diff-Quik-stained peripheral blood smears from 49 Greyhounds were compared with 200 blood smears from non-Greyhound dogs. Transmission electron microscopy was done on blood from 3 Greyhounds with vacuolated eosinophils and 3 with normal eosinophil granules. Blood smears from 4 of these dogs also were stained cytochemically with alkaline phosphatase (AP), chloracetate esterase (CAE), and alpha naphthyl butyrate esterase (ANBE). The morphologic features and tinctorial properties of vacuolated and normal eosinophils were compared.

RESULTS

Twenty-six Greyhounds (53%) had vacuolated eosinophils and 23 (47%) had normal granulated eosinophils in smears stained with Wright's stain. Only 1% of eosinophils were vacuolated in non-Greyhound dogs. Twenty of the 23 (85%) Greyhounds with normal granulated eosinophils on Wright's-stained smears had vacuolated eosinophils in smears stained with Diff-Quik. Ultrastructurally, no morphologic differences were observed between granules of vacuolated and normal eosinophils. Both vacuolated and normal eosinophils in Greyhounds were positive for AP and negative for CAE and ANBE, as expected for normal dogs.

CONCLUSION

Vacuolated eosinophils in Greyhounds likely reflect, at least in part, differential staining properties of the specific granules with different hematologic stains. Ultrastuctural and cytochemical features of eosinophil granules were similar in normal and vacuolated eosinophils from Greyhounds.