Chloride:Sodium Ratio May Accurately Predict Corrected Chloride Disorders and the Presence of Unmeasured Anions in Dogs and Cats

Disorders of chloride and mixed acid–base disturbances are common in veterinary emergency medicine. Rapid identification of these alterations and the presence of unmeasured anions aid prompt patient assessment and management. This study aimed to determine in dogs and cats if site-specific reference values for [Cl−]:[Na+] ratio and [Na+] − [Cl−] difference accurately identify corrected chloride abnormalities and to evaluate the predictive ability of the [Cl−]:[Na+] ratio for the identification of unmeasured anions. A database containing 33,117 canine, and 7,604 feline blood gas and electrolyte profiles was generated. Institution reference intervals were used to calculate site-specific reference values for the [Cl−]:[Na+] ratio and the [Na+] − [Cl−] difference. Contingency tables were used to assess the ability of these values to correctly identify corrected chloride disorders. Unmeasured anions were estimated by calculating strong ion gap (SIG). Continuous variables were compared using the Mann–Whitney U test. Correlations between continuous variables were assessed using Spearman’s rho (rs). In dogs, site-specific reference values for the [Cl−]:[Na+] ratio correctly identified 94.6% of profiles as hyper-, normo-, or hypochloremic. For dogs with normal sodium concentrations, site-specific reference values for the [Na+] − [Cl−] difference correctly identified 97.0% of profiles. In dogs with metabolic acidosis (base deficit > 4.0), [Cl−]:[Na+] ratio and SIG were moderately but significantly negatively correlated (rs −0.592, P < 0.0001). SIG was significantly greater in dogs with metabolic acidosis and hypochloremia compared to those without hypochloremia (P < 0.0001). In cats, site-specific reference values for the [Cl−]:[Na+] ratio correctly identified 93.3% of profiles as hyper-, normo-, or hypochloremic, while site-specific reference values for [Na+] − [Cl−] difference correctly identified 95.1% of profiles. In cats with metabolic acidosis [Cl−]:[Na+] ratio and SIG were moderately significantly negatively correlated (rs −0.730, P < 0.0001). SIG was significantly greater in cats with metabolic acidosis and hypochloremia compared to those without hypochloremia (P < 0.0001). Site-specific values for [Cl−]:[Na+] ratio and [Na+] − [Cl−] difference accurately identify corrected chloride disorders in both dogs and cats and may aid identification of the presence of unmeasured anions.

inTrODUcTiOn Chloride is the principal anion in the extracellular fluid (ECF) and thus its regulation is crucial for maintenance of osmolality and for acid-base balance (1). Metabolic acid-base disorders are common in veterinary emergency and critical care medicine and point of care measurement of electrolytes and acid-base status are integral to the diagnosis and management of emergency patients (2,3). The acid-base status of plasma (and hence the ECF) is affected by alterations in alveolar ventilation changing PaCO2, through manipulation of the plasma strong ion difference (SID) by the kidneys and secondary to alterations in the concentrations of weak acids, as indicated by Atot (4,5). The kidneys regulate SID by differential reabsorption of sodium and chloride ions in the renal tubules. Disorders of plasma chloride concentration are strongly associated with acid-base disturbances. A decrease in plasma chloride increases SID, causing a hypochloremic alkalosis, while an increase in plasma chloride decreases SID, causing a hyperchloremic acidosis (1). Identification of such alterations can be clinically valuable to increase the index of suspicion for gastrointestinal obstruction (6), or the presence of unmeasured anions.
In addition to disorders that cause gain or loss of chloride, changes in plasma chloride concentration can also result from changes in water balance. When changes in water balance affect measured chloride concentrations, the measured sodium concentration also changes (7). Correct identification of a true chloride change occurring independent of changes in water balance requires calculation of the corrected chloride, by taking into account the corresponding changes in sodium concentration: (1) Typically, the mid-point of the sodium reference interval (RI) is used to approximate normal sodium for this calculation. The corrected chloride value is then compared against the institution RIs for chloride to determine if an independent chloride disorder exists. An alternative method for rapid identification of the presence of a corrected chloride disorder is the calculation of the chloride:sodium ratio ([Cl − ]:[Na + ] ratio) (8).
A high [Cl − ]:[Na + ] ratio suggests a gain of chloride relative to sodium and a hyperchloremic metabolic acidosis independent of free-water changes, while a low chloride to sodium ratio suggests a loss of chloride relative to sodium and a hypochloremic metabolic alkalosis. Reference values for the chloride to sodium ratio have not been established for dogs and cats. Anecdotally, it has been suggested that values >0.78 in dogs and >0.80 in cats are associated with hyperchloremic metabolic acidosis, while values <0.72 in dogs and <0.74 in cats are associated with hypochloremic alkalosis (9). Since the corrected chloride calculation and the chloride to sodium ratio calculation incorporate the same values, the two calculations are mathematically linked. Thus, it is possible for every institution to calculate site-specific reference values for the [ When the sodium concentration is normal, the difference between the sodium and chloride concentrations ( can also be used to identify the presence of a corrected chloride disorder. If sodium is abnormal then using a simple difference will either underestimate or overestimate the influence of free water on chloride levels. Normal values for the sodium-chloride difference have also been previously reported, wherein values >40 mmol/L indicate hypochloremic alkalosis, while values <32 mEq/L are associated with hyperchloremic acidosis, but these values have not been validated to date (9). Following similar logic to the sitespecific chloride to sodium ratio calculation, a site-specific RI for the sodium chloride difference may be calculated as:

Na
Cl difference lower bound Na RI midpoint

Na
Cl difference upper bound Na RI midpoint  (9); that corrected hypochloremia in the presence of a BD correlates with the presence of unmeasured anions and that corrected normochloremia in the presence of a BD correlates with the presence of mixed acid-base disturbances.

MaTerials anD MeThODs electrolyte and Metabolite analyses
Blood gas and electrolyte analyses were conducted with pointof-care analyzers (RapidPoint 405; Siemens, Malvern, PA, USA) equipped with ion-selective electrodes using blood samples heparinized with dry balanced lithium/zinc heparin (Arterial Blood Gas Sampler, Westmed Inc., Tucson, AZ, USA). The point-of-care instrument employed in this study uses an external cartridgebased system to provide quality control (QC). This cartridge contains three separate levels of control material that span clinically relevant ranges. The QC procedure on the analyzer is automated and runs at a predetermined frequency (three times per day). Analysis channels that fail repeated QC are turned off and do not produce patient test results until the problem is rectified by an operator. Day-to-day and month-on-month performance based on tabulated QC data and Levey-Jennings graphs is assessed as required. Local RIs for the blood-gas analyzer were previously generated using heparinized blood samples collected from 20 healthy dogs and 20 healthy cats that were not part of the study population. These animals were considered healthy on the basis of history, physical examination and the results of complete blood count and serum chemistry profiles. Serum chemistry analyses were conducted using an automated chemistry analyzer (Cobas ModP, Roche-Hitachi, Indianapolis, IN, USA). Blood lactate concentrations were measured with a handheld point-of-care lactate meter (Lactate Pro, Arkray, Minneapolis, MN, USA) using heparinized whole blood samples.

case selection and Database compilation
An electronic database of blood gas and electrolyte analyses conducted at Cornell University Hospital for Animals between May 31, 2007, and January 03, 2015, on patients presented to the emergency room or intensive care unit was searched for results from dogs and cats. The database was visually inspected and manually curated to remove samples from species other than dogs and cats, samples with missing, erroneous or untraceable case numbers, analyses from sample types other than blood (e.g. abdominal fluid) and analyses with missing data. Medical records from dogs with very high measured chloride concentrations were manually checked to identify patients receiving potassium bromide as an anticonvulsant (11). These cases were removed from the dataset to limit the impact of a known interfering substance (12). Provided the datasets were complete, multiple samples from a single animal were kept in the database. Institution computerized medical record systems were then searched for data on patient signalment, presenting complaint, final diagnosis, outcome, hospitalization dates and for data from serum biochemistry analyses. Four separate databases were thereby created containing the electrolyte data, point-of-care analyses, biochemistry analyses and case demographics. A custom script (Visual Basic, Microsoft Visual 328 Studio for Windows; Microsoft, Redmond, WA, USA) was written to search each database via the unique patient identifier and created a final composite database containing the relevant data from each of the separate databases corresponding to the time and date stamp from the blood-gas and electrolyte analyses. The final database was then manually checked for accuracy by crossreferencing the database entries with the parent data sources for a randomized selection of cases, spanning the entire range of case numbers and representing 0.1% of the total case entries. From the parent database, canine and feline cases were segregated for further analyses.

acid-Base and strong ion calculations
Unmeasured anions were estimated by calculating the strong ion gap (SIG) as follows (2,13,14): where Alb. contribution = Albumin g dL 2.  Only profiles with normal sodium concentrations were used to assess the performance of the Na-Cl difference RIs. There were 16,662 profiles with normal sodium concentrations. Based on calculations using equation 1, of the 16,662 profiles that were normonatremic, 11,131 profiles had normal chloride, 3,113 profiles were hypochloremic, and 2,418 profiles showed evidence of hyperchloremia. The previously reported RIs for the Na-Cl difference (32-40) correctly identified 8,355 of the normal profiles, 1,467 of the hypochloremic profiles, and 2,418 of the hyperchloremic profiles for a total accuracy percentage of 73.46%. Using the newly proposed normal ranges for Na-Cl difference evaluation (29.0-38.0), 10,863 of the normal profiles were correctly identified, 2,946 of the hypochloremic profiles were correctly identified, and 2,352 of the hypochloremic profiles were correctly identified, for a total accuracy percentage of 96.99%. Overall a net improvement of 23.53% in accuracy was attained over the previously published RI.
After removal of cases for which requisite data were missing, there were 1,106 records available for acid-base analysis using the quantitative strong ion approach. Using the whole dataset, there was a mild, but significant negative correlation (rs = −0.428, P < 0.0001) between the [Cl − ]:[Na + ] ratio and SIG (Figure 1A). In dogs with a metabolic acidosis (n = 739) (defined as BD greater than 4.0) the correlation between the [Cl − ]:[Na + ] ratio and the SIG was greater (rs = −0.592, P < 0.0001) (Figure 1C). The SIG was significantly greater in dogs with a metabolic acidosis in which hypochloremia (defined using the site-specific [Cl − ]:[Na + ] ratio < 0.74) was present than in which it was absent ([Cl − ]:[Na + ]  ratio > 0.74) 11.5 (34.9 to −1.6) with hypochloremia vs. 5.5 (23.3 to −6.5) without hypochloremia (P < 0.0001) (Figure 2A) ,042 of the normal profiles were correctly identified, 2,858 of the hypochloremic profiles were correctly identified, and 191 of the hypochloremic profiles were correctly identified. Using the site-specific RI led to a net improvement in accuracy of 32.51% over the previously published RI.
As with dogs, only profiles with normal sodium were evaluated when assessing the performance of the Na-Cl difference RIs. There were 4,472 feline profiles with a normal sodium concentration. Based on the established corrected chloride calculation, of the 4,472 profiles that had normal sodium concentrations, 2,468 profiles had normal chloride, 1,919 profiles were hypochloremic, and 85 profiles showed evidence of hyperchloremia. The currently published normal ranges for the Na-Cl difference (32-40) correctly identified 1,656 of the normal profiles, 618 of the hypochloremic profiles, and 85 of the hyperchloremic profiles for a total accuracy percentage of 52.75%. Using the site-specific RI for Na-Cl difference (26.0-36.0), correctly identified 2,350 of the normal profiles, 1,823 of the hypochloremic profiles, and 78 of the hypochloremic profiles, for a total accuracy percentage of 95.06%. This yielded an overall improvement of 42.31% in accuracy over the previously published RI.
After removal of cases for which requisite data were missing, there were 671 records available for acid-base analysis using the quantitative strong ion approach. Using the whole dataset, there was a moderate, but significant negative correlation (rs = −0.548, P < 0.0001) between the [Cl − ]:[Na + ] ratio and SIG (Figure 1B). In cats with a metabolic acidosis (n = 490) (defined as BD greater than 4.0) the correlation between the [Cl − ]:[Na + ] ratio and the SIG was greater (rs = −0.730, P < 0.0001) (Figure 1D). The SIG was significantly greater in cats with a metabolic acidosis in which hypochloremia (defined using the site-specific [Cl − ]:[Na + ] ratio < 0.76) was present (P < 0.0001) than in which it was absent ([Cl − ]:[Na + ] ratio > 0.76) ( Figure 2B). The anion gap calculation incorporates the calculated [HCO3 − ] that depends upon the measured CO2 partial pressure. Thus, in patients with a marked respiratory acid-base abnormality, the AG calculation may be inaccurate. In contrast, the BD is a CO2 independent measure of acid-base status. Additionally, if only the [Cl − ] and [Na + ] values are available, such as on a chemistry panel, the ratio may be still be calculated and used to suggest the presence of unmeasured anions and by using both the low and high reference values the ratio may also help exclude the presence of unmeasured anions.

DiscUssiOn
Attempts to improve the identification of the nature, cause, and severity of metabolic acidosis in clinical patients has driven the development of acid-base analysis methods, including the Stewart approach (4), the strong ion concept (15,16), and the quantitative acid-base analysis systems (17,18). In contrast to the Henderson-Hasselbalch approach, the strong ion concept posits that disorders of sodium and chloride are two of the major determinants of acid-base disturbances. Hyperchloremia can be readily identified with the [Cl − ]:[Na + ] ratio and is associated with a metabolic acidosis. Hypochloremia as a primary disorder causes a metabolic alkalosis. In patients with increased concentrations of negatively charged ion species such as lactate, ketoacids and ethylene glycol metabolites, the presence of hypochloremia may mask or diminish the resulting metabolic acidosis. Measuring these other ion species may be costly and time-consuming. While validated point-of-care assays for lactate and ketones now exist (19)(20)(21), they are not universally available and their measurement is unnecessary in every case. Data from the present study suggest that evaluating the BD and the [Cl − ]:[Na + ] ratio in combination will identify those patients in which a compensating hypochloremia (10)

eThics sTaTeMenT
The study was exempt from ethics committee approval because it presents a retrospective analysis of electrolyte and acid-base data collected as part of clinician-driven care provided to patients treated at the institution hospital. No client or patient identifying information is presented.
aUThOr cOnTriBUTiOns RG conceived the study, analyzed data, and cowrote the manuscript; MM analyzed data and cowrote the manuscript; EZ analyzed data and edited the manuscript; SR and DF collected and analyzed data and edited the manuscript.