Symmetric dimethylarginine and renal function analysis in horses with dehydration

This study was a prospective investigation performed on clinically dehydrated horses. Serum SDMA, urea and creatinine concentrations, urine renal markers and short-term prognosis until discharge were included and analysed. reference range until T 48 while their creatinine concentrations into the normal range after T 12 postulate that SDMA might also reflect the tubular damage, while creatinine does not. A total of 18.5% (5/27) of horses had an increased GGT/creatinine ratio at T 0 . One study showed that the GGT/creatinine ratio was increased in all colic horses that underwent surgery. By contrast, only one of the 18 horses that colic had an increased GGT/creatinine ratio in the current study. The five horses with an increased GGT/creatinine ratio all had different primary complaints, ranging from colic orthopaedic problems. The for the significant correlation between the GGT/creatinine ratio and SDMA only at T 24 is unclear. The positive correlation between changes in SDMA and the GGT/creatinine ratio within the first 12 h could be related to the acute temporary disturbance of the renal tubule caused by renal during dehydration, but potential nephrotoxic aminoglycosides, gentamicin, our


Introduction
Symmetric dimethylarginine (SDMA) and its enantiomer asymmetric dimethylarginine are both amino acids derived from tissue endogenous proteins. While a major part of asymmetric dimethylarginine is metabolised by the enzyme dimethylarginine dimethylaminohydrolase, more than 90% of SDMA is excreted by the kidneys. [1][2][3] One meta-analysis paper including 18 studies in human medicine showed that plasma SDMA had a significant correlation with the glomerular filtration rate (GFR). 4 Studies in small animals found that SDMA has a better diagnostic value than creatinine for detecting a decrease of GFR since it can detect a decrease <30%, while creatinine only increases after a 75% loss of nephron function. [5][6][7] The plasma concentration of SDMA increased both in dogs with acute kidney injury (AKI) and chronic kidney disease (CKD).
Moreover, it was less effected by extra renal factors, such as lean body mass, age and gender, than creatinine in some studies, which makes it more suitable for detecting CKD patients with weight loss. 8,9 Primary kidney disease is believed to be comparably rare in horses. The prevalence of AKI in hospitalised horses was 14.8% in one study and the severity was lower than in other animal species. 10 However, the risk of developing AKI could be higher in diseases leading to dehydration and hypovolemia, such as colic or diarrhoea. 11 Early detection of renal injury and adequate therapy would be beneficial in these horses and drives researchers to search for a more sensitive biomarker. SDMA has not been widely studied in horses.
The aim of this study was to compare the concentration of SDMA with traditional renal biomarkers and establish its relationship with kidney function analysis in dehydrated horses. We hypothesised that SDMA concentrations would: 1) correlate significantly with dehydration and current renal markers, especially markers that are known to detect a decrease of GFR, and 2) provide a reliable value regarding short-term prognosis. The result should provide a prospective view as to whether SDMA is a potential renal marker to help diagnose early kidney injury in

Accepted Article
This article is protected by copyright. All rights reserved horses.

Study design and study population
This study was a prospective investigation performed on clinically dehydrated horses. Serum SDMA, urea and creatinine concentrations, urine renal markers and short-term prognosis until discharge were included and analysed.
Horses that were presented to the equine clinic, Free University of Berlin, between August 2018 and December 2019 with at least 6% dehydration without primary or history of kidney disease were included in this cohort study. Horses that had at least two or more abnormal criteria on admission were included. The criteria considered were: heart rate >60 beats/min, packed cell volume (PCV) >40%, total protein concentration (TP) >70 g/L, capillary refill time >2 s, lactate >0.9 mmol/L and clinical signs indicative of hypovolemic shock including cold extremities, pale mucous membranes or decreased jugular fill. The assessment of the grade of dehydration was based on the clinical examination, PCV and TP at admission (Table S1). 12 Foals younger than three months were excluded from the study in order to avoid spurious hypercreatininemia in foals. 13 The horses were divided into three groups: those with 1) mild dehydration (6-8% dehydration), 2) moderate dehydration (8-10% dehydration) and 3) severe dehydration (>10% or if horses were in hypovolemic shock).
Horses with 6-8% dehydration were rehydrated either with infusion therapy (Ringer-lactate or Ringer's solution [B Braun Melsungen AG, Melsungen, Germany]) for at least 24 h, as indicated by PCV and TP, or water by nasogastric tube; horses with moderate or severe dehydration were treated with infusion therapy in all cases. Other treatments were chosen based on the horses' main

Accepted Article
This article is protected by copyright. All rights reserved complaint.
Survivors were defined as the horses which survived to discharge; non-survivors were those which died or were euthanised during the hospitalisation. Horses which were euthanised due to financial constraints were excluded from the analysis for the prognostic value of SDMA in order to reduce the statistical error.
AKI was defined as an increase of serum creatinine concentration ≧ 26.5 μmol/L within 48 h, according to the veterinary AKI staging system. 10

Blood sample collection
A total of 10 ml full blood was taken from the external jugular vein at time-point 0 (T 0 ) when the horses arrived at the clinic before infusion therapy. Further such samples were also taken at 12,24 or 48 ± 2 h (T 12 , T 24 and T 48 , respectively) after admission. Each sample was filled into a serum tube with a clot activator (Sarstedt AG & Co, Nümbrecht, Germany) and centrifuged at 3,800 g for 10 min. Two 1.2 ml samples were frozen at -80 for each time point for later corrections or ℃ estimation if necessary; one sample was kept at 4 until, sent to an external laboratory ℃ (SYNLAB. vet GmbH, Berlin, Germany) and analysed within 24 h. The concentrations of serum creatinine, urea nitrogen, glucose, TP, albumin and electrolytes were measured utilising an automated AU680 clinical chemistry analyser (Beckman Coulter GmbH, Krefeld, Germany).
Serum SDMA concentrations were measured with a DLD SDMA ELISA Kit (DLD Diagnostika GmbH, Hamburg, Germany). The latter has been validated in healthy horses and horses with AKI. 14

Urine sample collection
Urine samples were taken from mares at T 0 with a urine catheter before or within the first 30 min of infusion therapy; stallions and geldings' urine was collected during surgery, if the horse underwent surgery directly after admission or from the midstream of naturally voided urine in the

Accepted Article
This article is protected by copyright. All rights reserved stable within 30 min of admission. The urine samples at T 12 , T 24 and T 48 were taken in the stable during spontaneous urination after admission. All urine samples were analysed by a dipstick Combur9 Test (Roche Deutschland Holding GmbH, Freiburg, Germany) in the clinic. Urine (10 ml) collected in a sterile urine collection tube (Labor-und Medizintechnik Specht GmbH, Eresing, Germany), was sent to an external laboratory (SYNLAB. vet GmbH) for renal function analysis together with serum samples from the same time point. The urine specific gravity (SG), fractional excretion of electrolytes, urine TP (uTP) and the gamma-glutamyltransferase (GGT)/creatinine ratio were measured by a refractometer and the AU680 clinical chemistry analyser (Beckman Coulter GmbH) within 24 h. Sediment interpretation was performed by technicians with a microscope at an external laboratory (SYNLAB. vet GmbH).

Data analysis
Analysis was performed using IBM SPSS software (IBM Deutschland GmbH, Ehningen, Germany) for Windows, version 25. Serum concentrations of SDMA, creatinine and urea, urine SG, fractional excretion of sodium (FE Na+ ), uTP and GGT/creatinine ratio were analysed by Shapiro-Wilk tests to check the distribution of parameters. The Kendall Tau b coefficient test was used to test the correlations between concentrations of SDMA, creatinine and urea and parameters of renal function analysis, respectively, from T 0 to T 48 . The correlation between changes of SDMA concentrations and other parameters from T 0 to T 12 was performed by Kendall Tau b test in order to evaluate the reaction of SDMA and renal markers to the initial rehydration therapy. The Kruskal-Wallis test was used to analyse the differences of SDMA concentrations among the three dehydration groups. The distribution of serum creatinine and urea concentrations among three dehydration groups at T 0 were also analysed by Kruskal-Wallis test. The distribution of serum SDMA concentrations at T 0 in survivors/non-survivors groups were analysed by Mann-Whitney U test. Linear mixed regression models with repeat measurement were applied to access the

Accepted Article
This article is protected by copyright. All rights reserved association between the concentrations of SDMA at four time points and the three dehydration groups independently. Mauchly's test for sphericity was applied and the Huynh-Feldt correction was used to determine differences between the time points and interactions between time point and group. Model diagnostics included the visual inspection of normality and homoscedasticity of the residuals per time point. The level of significance was set at 5% for all analyses.

Study Population
A total of 57 horses met the inclusion criteria. Sixteen were excluded due to lack of obvious laboratory changes which made grading of the accurate dehydration status impossible. The remaining 41 horses were included in the analyses. Patient data and final diagnosis can be seen in Table 1.
Most horses were admitted as emergency cases. Thirteen horses were assigned to the mild dehydration group with 6-8% dehydration according to PCV/TP and clinical characterisations.
Sixteen horses were in the moderately dehydrated group with 8-10% dehydration. Twelve horses were in hypovolemic shock on admission and belonged to the severely dehydrated group. A total of 46.3% (19/41) of horses in the current study underwent surgery because of the primary disease: 18 had colic surgery and one had orthopaedic surgery. Seventeen horses were treated with gentamicin during the sampling period and 33 horses received non-steroidal anti-inflammatory drugs.  Table 2.

Renal parameters and renal function analysis
Urine samples were collected from 28 horses at T 0 ; samples in the other 13 horses could not be obtained at this time point. A total of 17.9% (5/28) of horses were sampled before the beginning of infusion therapy at T 0 . The urine SG and uTP were increased in 42.9% (12/28) and 17.9% (5/28) of horses, respectively. One patient's urine was too concentrated to carry through the whole renal function analysis. Unfortunately, even after dilution of the sample, a homogeneous solution was not formed and could not be analysed by the laboratory equipment. Therefore, the FE Na+ and GGT/creatinine ratio were measured in 27 horses and found to be increased in 18.5% ( There were no correlations between SDMA and creatinine or urea concentrations from T 12 to T 48

Accepted Article
This article is protected by copyright. All rights reserved (Table S2).
No significant correlations at T 0 were identified between SDMA concentrations and the parameters of renal function analysis: urine SG, FE Na+ , uTP and GGT/creatinine ratio. The urine TP had a moderate correlation with SDMA concentrations at T 12 (r = 0.394, P = 0.04) and T 48 (r = 0.565, P = 0.01). The GGT/creatinine ratio at T 24 correlated significantly with SDMA concentrations (r = 0.547, P = 0.02). Neither urine SG nor FE Na+ correlated with SDMA concentrations from T 0 to T 48 (Table S2).
In order to compare the response of each marker to the rehydrated therapy, changes in serum SDMA concentrations and other renal markers from T 0 to T 12  One horse had creatinine concentrations above the reference range persistently until T 48 , which fit the criteria of AKI. The SDMA concentrations above the cut-off value over 48 h were observed in two horses, one of them was the horse with AKI, while another one had increased creatinine concentrations only until T 12 . Two horses that did not have increased SDMA concentrations at T 0 developed increased SDMA concentrations above the cut-off value at T 24 and T 48, respectively, meanwhile, they both had normal creatinine concentrations persistently throughout the study period.
A total of 25% (7/28) of horses had no urine casts according to the sediment examination. The rest of the horses (19/28) had calcium-carbonate, -oxalate and struvite within the physiological amount. Since the sediment examinations were all carried out in an external laboratory, although

Accepted Article
This article is protected by copyright. All rights reserved examined within 24 h, the rapid degeneration of the cast in alkaline urine could not be totally avoided in this study. In addition, the estimation of sediment amounts and types with numerical SDMA data is difficult and imprecise. Regardless of the type of the urine cast, there were no significant correlations between the SDMA concentrations and amounts of cast from T 0 to T 48 .
Findings of the other indicators including erythrocytes and leucocytes made the analysis between these indicators and SDMA concentrations impossible: pathologically increased erythrocytes and leucocytes were only found in one and two horses respectively throughout 48 h, meanwhile, the rest of the patients had no or acceptable normal amounts of erythrocytes and leucocytes in their urine. Without convincing statistical estimation, the results of the sediment examinations are not discussed further.

Relationship between SDMA and dehydration groups within 48 hours
The Kruskal-Wallis test revealed that there were significant differences of SDMA concentrations at T 0 among dehydration groups (P = 0.03; Figure 2, Table 2

Accepted Article
This article is protected by copyright. All rights reserved the other time points. Although higher mean concentrations of SDMA could be observed in the moderately and severely dehydrated group from T 0 to T 48 , the distribution of SDMA concentrations in the three groups overlapped easily with each other. The intraclass correlation coefficient was calculated as 81.2%. This means that 81.2% of the variance was due to the variance between horses whereas values did not differ much between the individual horses.
Besides SDMA, there were also significant differences of the serum creatinine and urea concentrations among three dehydration groups at T 0 (P<0.001 and P = 0.04, respectively, Table   2).

Prognostic value of SDMA
Twenty-eight of the 41 horses included in the analyses were alive until T 12 , 25 horses at T 24 and 21 horses survived to T 48 . Two horses that were euthanised due to financial constraints were excluded from the statistical estimation. A total of 53.8% (21/39) horses were euthanised in accordance with animal welfare and poor prognosis or deceased during hospitalisation, and 46.2% (18/39) of horses were discharged.
With the Mann-Whitney U test, there was no statistical significance in the association between SDMA concentrations at T 0 and survival (P = 0.1). The median concentration of SDMA at T 0 in the survivor group was 0.58 (IQR: 0.40-0.69,) µmol/L, while the median in the non-survivor group was 0.67 (IQR: 0.54-0.84) µmol/L. A total of 63.6% (7/11) of horses with increased SDMA concentrations at T 0 were euthanised or died, while 50% (14/28) of horses did not survive to discharge despite normal SDMA concentrations.

Relationship between SDMA and current renal biomarkers
This study aimed to examine the association of SDMA concentrations with other markers of renal

Accepted Article
This article is protected by copyright. All rights reserved function in dehydrated horses to test its value as a potential marker of early kidney injury.
We found a moderate correlation between SDMA and creatinine concentrations at T 0 , while there was no significant correlation between SDMA and serum urea concentrations from T 0 to T 48 . The moderate correlation between SDMA and creatinine concentrations was similar to a study in dogs with AKI. 8 Furthermore, the changes in the SDMA and creatinine concentrations after rehydration therapy measured at 12 h were positively correlated, indicating that both SDMA and creatinine might have the similar ability to detect the decrease of GFR. There were inconsistencies between SDMA, creatinine and urea concentrations in some patients, results which may relate to extra-renal factors: 3 of 9 horses with increased creatinine concentrations at T 0 had SDMA concentrations within the normal range. Serum creatinine concentrations can increase due to not only the kidney injury but also the dehydrated status of the patients. Although the authors in one study of dehydrated dogs concluded that SDMA might be influenced by the prerenal volume status in dogs with azotaemia, 15 we postulate that hydration status may not impact SDMA as much as creatinine. We observed clearer differentiation in creatinine concentrations than SDMA concentrations among the three dehydration groups, suggesting that creatinine might be affected by dehydration more easily than SDMA. However, since 2 of these 3 horses died shortly after T 0 , it remains unknown whether the hydrated status led to any effect on either biomarker within 12 h in the current study. These 3 horses were of middle age and with normal to obese body condition score. On the other hand, 5 of 11 horses with increased SDMA concentrations had creatinine concentrations within the normal reference range. Four of these horses were older than 20 years and one was estimated to have a body condition score of 2 out of 9. Decreased liver function, older age and less muscle mass, might contribute towards the differences in our observations between creatinine and SDMA and may explain why which might cause the concentration of creatinine to remain within the normal range with a potential kidney injury or decrease of GFR. 9,16 Serum urea

Accepted Article
This article is protected by copyright. All rights reserved nitrogen is not a sensitive marker of GFR and is also affected by different extrarenal variations. 17 Serum urea and creatinine are impacted by similar extrarenal factors and were found to have moderate correlation with each other.
Two horses developed increased SDMA concentrations above the cut-off value at T 24 and T 48 with persisting normal creatinine concentrations. Although neither horse was defined as having AKI according to traditional creatinine criteria, it could still indicate that SDMA might detect an early kidney injury prior to creatinine.
Most studies in small animals and humans with AKI and CKD focused on the relationship between SDMA and creatinine or GFR. In the current study, we also compared SDMA concentrations with parameters of renal function measured in urine. No significant correlations of SDMA concentrations with urine SG, FE Na+ , uTP and the GGT/creatinine ratio were found in the present study from T 0 to T 48 , except for the correlation with uTP at T 12 and T 48 and the GGT/creatinine ratio at T 24 . A lack of complete urine sampling throughout the whole study period might have been a factor affecting these results. In addition, different external factors, such as infusion therapy with and without electrolytes, or medications which have an impact on the renal parameters.
Urine SG has been used in the estimation of dehydration for a long time. In our study, urine SG was the parameter which showed the highest proportional increase at T 0 in 42.9% (12/28) of dehydrated horses. Although urine SG is sensitive to acute hypertonic dehydration, it could still lead to misclassified results. The urine in the bladder in horses with acute dehydration could still be physiologically diluted and then mixed with urine produced in the dehydrated state. 18,19 Furthermore, the SG is also affected by the infusion therapy, medications such as alpha-2-agonists from T 12 to T 48 , thus interpretation of our data was only possible at T 0 . The FE Na+ indicates the function and damage of the proximal tubule. However, it can be affected by breed, age, exercise,

Accepted Article
This article is protected by copyright. All rights reserved medication or crystalloid fluid therapy in horses. 20 Only 5 horses were sampled before the beginning of sodium-containing infusions at T 0 . Two horses had increased GGT/creatinine ratios and FE Na+ until T 24 and T 48 respectively, indicating advanced tubular damage. These two patients also had increased SDMA and creatinine concentrations at T 0 and a relatively high concentration of SDMA within the reference range until T 48 , while their creatinine concentrations decreased continuously into the normal range after T 12 . We postulate that SDMA might also reflect the tubular damage, while creatinine does not. A total of 18.5% (5/27) of horses had an increased GGT/creatinine ratio at T 0 . One study showed that the GGT/creatinine ratio was increased in all colic horses that underwent surgery. 21 By contrast, only one of the 18 horses that underwent colic surgery had an increased GGT/creatinine ratio in the current study. The five horses with an increased GGT/creatinine ratio all had different primary complaints, ranging from colic to orthopaedic problems. The reason for the significant correlation between the GGT/creatinine ratio and SDMA concentrations only at T 24 is unclear. The positive correlation between changes in SDMA concentrations and the GGT/creatinine ratio within the first 12 h could be related to the acute temporary disturbance of the renal tubule caused by renal ischaemia during dehydration, but administration of potential nephrotoxic aminoglycosides, such as gentamicin, may also have influenced our results. 22 Proteinuria occurs in glomerular disease, bacteriuria or pyuria, and it may increase in equine urine after exercise. 12 Only one of the five horses that showed increased uTP at T 0 had an increased uTP at T 12 as well. Other than that, increased urine protein seemed to be a coincidental and transient result in each patient at different time points and the moderate correlation between uTP and SDMA concentrations at T 12 and T 48 might be an accidental result.
Several studies showed that SDMA might be eliminated by the liver and other non-renal enzymatic degradation in humans. [23][24][25] Furthermore, SDMA showed neither an advantage in predicting CKD in dogs with Leishmaniosis, nor the ability to detect CKD in cats with diabetes

Accepted Article
This article is protected by copyright. All rights reserved mellitus in several studies. 26,27 These results indicate that, in addition to being a potential marker of renal function, SDMA might also be involved in other physiological or pathological processes in human beings and small animals. Similarly, the elimination process of SDMA might not only be limited to the kidneys in horses, which might have influenced the SDMA's correlation with renal function in the current study.
Although SDMA concentrations varied significantly among three dehydration groups at T 0 , most differences were observed between the mild and moderate dehydration groups. Since the subgrouping of the patients depended only on their PCV/TP and clinical characteristics but not plasma osmolality, the wrong assignment could not be totally avoided which might have contributed to these statistical observations.

Prognostic value
No significant association was identified in this study between the SDMA concentrations and outcome whereas SDMA has been shown to be an independent prognostic indicator for long-term mortality in critical human patients and was associated with adverse clinical outcome 30 days after an ischaemic stroke. 28,29 In critically ill dogs, no significant difference in serum SDMA concentrations between survivors or non-survivors was found. 30

Conclusion
We observed moderate correlation between SDMA and serum creatinine concentrations but no persistently significant associations between renal function parameters and SDMA concentrations in dehydrated horses. SDMA concentrations were different between groups with different hydration status but SDMA was not different between survivors and non-survivors. Extra-renal factors are likely to have influenced our results and further studies of SDMA including serial monitoring will help clarify the role of this biomarker in equine renal disease.

Authors' declaration of interests
Author Hsiao-Chien, Lo received support for this work from SYNLAB.vet GmbH, Berlin, including part of study design, measurement of Symmetric dimethylarginine (SDMA; object Biomarker) and renal functional analysis with the coordination of co-author Judith C. Winter. The

Accepted Article
This article is protected by copyright. All rights reserved author has full access to the study data and take complete responsibility for the integrity of the data and accuracy of data analysis. Other co-authors have declared no competing interests.

Ethical animal research
This study was approved by the Ethics Committee of Free University Berlin.

Informed consent:
Owners consented for their horses to take part in this study.

Date accessibility statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Source of funding
The measurement of Symmetric dimethylarginine (SDMA, object Biomarker in this study), renal functional analysis were provided and executed by SYNLAB.vet GmbH, Berlin. Abbreviations: BCS = body condition score; N = number of subset. a Included one 4-month-old stallion that was excluded from the calculation of median age.
b Survivors were defined as horses which survived to discharge; non-survivors died or were euthanised during hospitalisation. Two horses in the mild dehydration group that were euthanised due to financial constraints were excluded.   This article is protected by copyright. All rights reserved