Results from this study are suggestive of mild transient non-azotaemic AKI in dogs envenomated by V. berus. These findings are supported by previous studies of snake-bitten dogs (5, 6).
Increased urinary AKI biomarkers have previously been reported in dogs envenomated by snakes compared to healthy controls (6). Although significant differences in uAlb/uCr between cases and controls existed in this study, the ratios are lower than reported in previous studies (6), possibly due to a milder kidney insult following V.berus envenomation compared to other snake species, although an effect of the extra freeze-thaw cycle these samples were subjected to, cannot be ruled out (34). Urinary albumin can reflect either glomerular or tubular injury (35) but lacks specificity, with increases also seen with extreme exercise (36), macroscopic hematuria and urinary tract infection (37). Hematuria was detected in several samples in this study but did not exceed 100 RBC/HPF and was not observed to be macroscopic in any sample. We therefore consider hematuria unlikely to have influenced uAlb/uCr ratios in these individuals.
One other study has specifically examined urinary markers for AKI in V.berus envenomated dogs, and reported higher uGGT/uCr and uALP/uCr compared to controls, suggestive of renal injury (5). Urinary GGT/uCr ratios in both cases and controls in our study were in accordance with the aforementioned study, but uALP/uCr was unexpectedly low. Given that ALP and GGT are both renal brush border enzymes, simultaneous leakage into urine upon tubular cell injury is expected. There are several possible explanations for a lack of increase in uALP in our study compared to others. Poor assay sensitivity might be considered most likely, but differences in timing and storage of samples are also possible explanations (38). An optimal uALP detection window of less than 12 hours after renal insult is reported in humans (39); thus, the detection window may have been missed in our study. Urinary ALP can also originate from the epididymis and prostatic fluid, thus, differences in proportions of intact male dogs between studies might also influence results (40).
Urinary CysC was not significantly different between envenomated dogs and controls in this study. Urinary CysC increases as a result of decreased reabsorption after proximal tubular injury (41), and an association between uCysC concentration and severity of AKI has been described in humans and dogs (42, 43). Mild AKI is therefore a likely explanation for the lack of increase in uCysC/uCr in envenomated dogs in the present study. A recent study indicates that the assay used in our study might measure lower concentrations of uCysC compared to other assays (44). Care should therefore be taken in comparing results from different assays.
Urinary KIM-1 is an early and highly sensitive and specific diagnostic biomarker for proximal tubular injury, approved by the US Food and Drug Administration (FDA) as a marker for drug induced AKI in rodent models (45, 46). Human studies describe upregulated uKIM-1 expression as early as 2 hours, and lasting up until 48 hours, after toxic or ischemic insult to the kidney (46). In our study, significant differences in uKIM-1/uCr between envenomated dogs and controls were not observed until T4. Species differences, sample size, and different mechanisms of renal toxicity are possible explanations for the seemingly later induction of uKIM-1 in our study. However, it has also been suggested that uKIM-1 may not be as sensitive and early a marker of AKI in dogs compared to humans (47, 48). Urinary uKIM-1/uCr and absolute uKIM-1 concentrations in healthy dogs in our study were lower than reported previously, also in one study using the same assay (33, 49). The same studies reported comparatively higher uKIM-1/uCr and absolute uKIM-1 in dogs with AKI than found in the envenomated dogs in our study. This indicates that although uKIM-1/uCr was significantly higher at T4 compared to both controls and T5, these ratios are overall low and likely represent mild AKI given previous findings of correlations between KIM-1 and extent of injury (50). All dogs in the present study had uKIM-1/uCr values below the cutoff of 0.75 ng/mg for diagnosing AKI proposed by one canine study (49), further supporting the theory that dogs bitten by V. berus might sustain a milder form of renal injury compared to AKI of other etiologies. A recent study suggests that uKIM-1 concentrations measured in the multiplex assay used in this study could be lower than those detected using other assays (44). An assay specific reference interval is needed and direct comparisons between concentrations detected in different studies should therefore be made with caution.
Neutrophil gelatinase-associated lipocalin is a ubiquitous LMW epithelial protein, subject to glomerular filtration and tubular reabsorption. Local NGAL production is induced in the distal tubule during renal injury (51). Urinary NGAL may therefore result from either proximal or distal tubular injury. Higher uNGAL/uCr ratios have previously been reported in snake-bitten dogs that developed AKI compared to controls (52), but this is the first study describing increases in uNGAL in dogs envenomated specifically by V.berus. In accordance with another canine AKI study (30), uNGAL/uCr ratios were higher already from 12 hours after bite in envenomated dogs, compared to controls, indicating its potential use as an early marker of AKI, as has previously been suggested (53). A recent study indicated that systemic inflammation has a significant impact on uNGAL/uCr ratios (30). Two forms of NGAL exist, of which the monomeric form is kidney specific, whereas increases in dimeric NGAL are seen in UTI and other inflammatory disease (54, 55). Assays for detection of monomeric NGAL in canine urine are not currently available, thus the possible lack of specificity should be taken into consideration when interpreting results.
Upregulation of the inflammatory cytokines MCP-1 and IL-8 is described during renal injury in dogs (27, 33) and in human snake envenomation (56). Few studies have measured uMCP-1 in dogs; hence, less is known regarding its kinetics. The finding of increased uMCP/uCr 12–36 hours after envenomation compared to controls in our study, corresponds with previous nephrotoxicity studies (28, 33), indicating that MCP-1 might be useful as an early marker of AKI. An interesting finding in the present study was the inverse relationship between bodyweight and uMCP-1/uCr in envenomated dogs. Although a lower relative venom concentration in larger dogs is a possible explanation, this was considered more likely to be an incidental finding, since a similar effect was not observed for the other biomarkers. A high number of samples in this study had uIL-8 concentrations under the LLOQ. This was also described in another canine AKI model, despite elevations in other AKI markers (33), raising questions as to the sensitivity of uIL-8 as a marker for canine AKI.
The chemokine OPN has rarely been quantified in canine urine, thereby limiting comparisons with our study. Urinary OPN is a sensitive marker for renal tubular injury in rodent drug induced AKI models (57). Absolute uOPN concentrations in the present study were comparable to those in a canine hemorrhagic shock model (33). In the aforementioned study, absolute uOPN values did not increase significantly from baseline after induction and treatment of shock, whereas in our study, uOPN/uCr was significantly higher in envenomated dogs 24 and 36 hours after bite, compared to controls. The difference in findings might be explained by difference in sample size, mechanism of AKI, or a lack or normalization to creatinine in the other study. A lack of specificity in the presence of inflammation is also possible and thus a systemic contribution to the uOPN measured, cannot be ruled out.
Serum concentrations of MCP-1, IL-8 and OPN can increase in various inflammatory states, including muscle injury and snake bite (57–60). Given their LMW, we cannot rule out a systemic contribution via glomerular filtration during systemic inflammation induced by snake envenomation (61), and further studies are therefore needed to ascertain the specificity of these markers.
Except for uCysC and uAlb, all urinary biomarker/Cr ratios were higher 36 hours after bite compared to controls. Whilst clearance kinetics of the various biomarkers may differ, our findings suggest that AKI might occur until at least 36 hours after V. berus bite. A resolution of AKI by 14 days is suggested by a lack of significant difference between cases and controls at T5, but further studies are needed to clarify whether AKI is present beyond 36 hours after bite. Although our study design does not allow us to establish a direct benefit of IVFT in the treatment of AKI in these dogs, hospitalization with monitoring and targeted IVFT is a sensible recommendation for dogs envenomated by V.berus, and results from our study indicate that this treatment should be implemented for a minimum of 36 hours.
In accordance with previous studies (9, 62), sCr concentrations were within the reference interval for all dogs in this study, likely due to a lack of renal dysfunction, although an effect of hemodilution due to IVFT cannot be ruled out. According to veterinary AKI grading guidelines (16), two of the envenomated dogs in this study would be classified as having AKI grade I due to a non-azotaemic increase in sCr of 26.4 µmol/L within 48 hours, and eight dogs due to microalbuminuria at one time point or more. A grading system incorporating novel renal injury biomarkers such as those measured in this study, to identify non-azotaemic AKI biomarker-positive individuals, would be of benefit by allowing early identification and treatment. More work is needed to establish the specificity of many of these biomarkers as well as to generate reference intervals for their clinical use.
In accordance with one other study (5), a positive correlation was found between severity of clinical signs at presentation and uGGT/uCr in our study. Although SSS at presentation might thus be an indicator of peak uGGT/uCr after snake bite, further studies are needed to fully assess its use in this setting, especially since the relationship between uGGT/uCr and extent of injury is unknown. The lack of correlation of SSS with the other biomarkers in this study, suggests that severity of clinical signs after envenomation is of overall limited use in predicting which dogs are most likely to develop AKI.
There are limitations to this study. As previously mentioned, the specificity of uOPN, uNGAL, uIL-8, and uMCP-1 in the face of systemic inflammation needs clarification. However, the parallel increases in the urinary biomarkers evaluated in this study are suggestive of renal tubular injury in dogs envenomated by V. berus. Sample size limits the statistical inferences that can be made. Differences in assay sensitivity and imputation methods used for values under LLOQ hinder direct comparisons between our study and others for some of the biomarkers. A high number of samples had values under the LLOQ for uALP, uIL-8 and uAlb, thereby limiting their interpretation. Evidence suggests that uALb/uCr and uNGAL/uCr values are unlikely to be significantly affected by microscopic haematuria and haemoglobinuria, respectively (52). Likewise, haematuria and haemoglobinuria are unlikely to significantly affect uGGT/uCr (37, 38), but it is unknown to what extent results for the other biomarkers in this study might be affected. Although the extent to which myoglobinuria might interfere with biomarker measurement is unknown, it is thought unlikely to influence the overall conclusions of this study since this was observed only in one dog. Urine samples were not subjected to bacterial culture, and thus subclinical bacteriuria may have been missed, although the extent to which this might influence results is unknown. Treatment with antivenom is a confounding factor, although unavoidable due to the non-interventional nature of this study. Renal histopathology would have been a useful addition to this study but was not included due to ethical considerations.