Hypocalcemia is a documented electrolyte disturbance in people and animals with sepsis, but its mechanism is poorly understood.
Hypocalcemia is a documented electrolyte disturbance in people and animals with sepsis, but its mechanism is poorly understood.
To investigate mechanisms of hypocalcemia in dogs with experimentally induced endotoxemia.
Six healthy mixed breed dogs were included in this nonrandomized, placebo-controlled, crossover study.
Dogs initially were injected with placebo (0.9% NaCl; 1 mL, IV) and then lipopolysaccharide (LPS; 2 μg/kg, IV) after a 5-day washout period. Blood and urine samples were collected for measurement of serum total calcium (tCa), ionized calcium (iCa), total magnesium (tMg), ionized magnesium (iMg), parathyroid hormone (PTH), 25-hydroxyvitamin D (vitamin D), venous blood gases, and fractional excretion (FE) of calcium.
After LPS administration, body temperature increased and blood pressure decreased. Both iCa and tCa decreased (P < .01), but iMg was not significantly different between control and LPS treatments. PTH concentrations increased (P < .01) and vitamin D concentrations decreased (P < .01). Venous pH, bicarbonate, base excess, and blood glucose also decreased (P < .01). Urine tCa concentration was below the limit of detection for all dogs after LPS administration.
Hypocalcemia occurs during endotoxemia in dogs and is associated with hypovitaminosis D. Hypomagnesemia, hypoparathyroidism, alkalosis, and increased calciuresis are not associated with hypocalcemia in endotoxemic dogs.
diastolic arterial pressure
mean arterial pressure
partial pressure of carbon dioxide
systolic arterial pressure
Hypocalcemia is a common electrolyte abnormality in people and animals with sepsis.[1-11] The incidence of ionized hypocalcemia is 16% in critically ill dogs and 24% in dogs with sepsis. Hypocalcemia is associated with increased mortality in septic people and dogs.[1-3, 10] Recent retrospective studies in dogs and cats with naturally occurring sepsis indicate that the lowest recorded serum iCa concentration is associated with a longer duration of hospitalization.[10, 11] In addition, dogs that die in the hospital have more severe ionized hypocalcemia and cats that do not experience normalization of their serum iCa concentrations during hospitalization are less likely to survive.
The mechanism for hypocalcemia in septic patients is poorly understood and is probably multifactorial. Suggested causes include alkalosis,[3, 5] hypomagnesemia,[3, 5, 8] hypoparathyroidism,[1, 5, 8] hypovitaminosis D,[1, 5] increased calciuresis, chelation of calcium, accumulation of calcium in tissues or cells,[5, 7, 13-17] or increases in procalcitonin.[18, 19] Some of these mechanisms have been investigated with animal sepsis models, including experimentally induced endotoxemia, but the findings are variable depending on the species investigated and the sepsis model used.[7-9, 13-17] Little information exists on the mechanism for hypocalcemia in naturally occurring or experimental sepsis in dogs. Understanding the mechanism or mechanisms of hypocalcemia in dogs with sepsis might lead to novel treatment strategies and improved survival.
The objective of this study was to evaluate calcium homeostasis in dogs with experimentally induced endotoxemia by measuring serum and urine electrolyte concentrations, venous blood gases, and serum hormones that regulate iCa. We hypothesized that, similar to other species, dogs with endotoxemia would exhibit decreases in serum iCa and tCa concentrations and that the most likely associated causes would be hypomagnesemia, hypoparathyroidism, or hypovitaminosis D.
Six mixed-breed, sexually intact, purpose-bred male dogs aged 6–12 months old (median, 7.4 months) and weighing 18.7–22.4 kg (median, 19.8 kg) were used in this study. All dogs were considered to be healthy on the basis of a lack of clinically relevant abnormalities on physical examination and screening laboratory tests including a CBC, serum biochemistry, and urinalysis. Dogs were housed in animal facilities accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International and all care was provided in accordance with the principles outlined by the National Institutes of Health. All experimental procedures were reviewed and approved by the Animal Care and Use Committee at North Carolina State University.
A nonrandomized, placebo-controlled, crossover study was performed. All dogs were initially treated with a placebo (saline solution; 0.9% NaCl; 1 mL, IV) and then undiluted lipopolysaccharide1 (LPS; 2 μg/kg, IV) after a 5-day washout period. Food was withheld from the dogs before each day of the study on which saline or LPS was administered, but water was available throughout the study. Baseline vital signs (ie, rectal temperature, heart and respiratory rates, blood pressure) were measured after each dog was obtained from its housing area. A 20-gauge catheter was placed in a cephalic vein of each dog 1 hour before initiation of the study. Dexmedetomidine2 (2 μg/kg, IV, once) then was administered to facilitate jugular venous catheter3 placement and indwelling urinary catheter4 placement. Urinary catheters were capped and maintained without a collection system. After instrumentation, atipamezole5 (volume equal to that of dexmedetomidine, IV, once) was administered via the cephalic catheter to reverse the sedative drug. After vital signs returned to baseline values, either saline or LPS was administered via the cephalic catheter. At the completion of the intensive study period (12 hours after baseline), the jugular, cephalic, and urinary catheters were removed, and the dogs were rehoused and offered food. Subsequent blood samples were collected by direct venipuncture and urine samples were collected by catheterization.
An oscillometric blood pressure device6 was used to record a systolic (SAP), diastolic (DAP), and mean arterial pressure (MAP) with the dog positioned in sternal or lateral recumbency. The cuff was placed in the same location (above the carpus or below the hock) for each recording. Heart rate (HR) was measured by palpation of the femoral or dorsal pedal pulse or by cardiac auscultation. Vital signs were recorded at time 0 (time of injection) and every 30 minutes thereafter for 6 hours and then at 8, 10, 12, and 24 hours. Dogs that became hemodynamically unstable (SAP < 90 mmHg, MAP < 70 mmHg, or HR > 180 bpm) after LPS administration were given lactated Ringer's solution7 (10 mL/kg, IV, bolus), which was repeated as necessary to maintain hemodynamic stability.
Blood samples (6 mL) were collected from the jugular catheter by means of a 3-syringe technique followed by a catheter flush (NaCl 0.9%, 1 mL, IV) or by means of direct venipuncture at times 0, 1, 2, 3, 4, 5, 6, 8, 10, 12, and 24 hours. Venous blood gas8 analysis was performed immediately with a cage-side analyzer9 and the remaining blood was deposited into clot activator tubes10 that were centrifuged (2000 rpm, 15 minutes, 15°C). Serum was separated and immediately stored at −80°C until batch analysis. Urine samples (6 mL) were collected from the urinary catheter at times 0, 2, 4, 6, 8, 10, 12, and 24 hours and immediately stored at −80°C until analysis. The urinary bladder was emptied after each urine sample collection.
Frozen serum and urine samples from times 0, 2, 4, 6, 8, 10, 12, and 24 hours were shipped on dry ice to the Diagnostic Center for Population and Animal Health at Michigan State University for measurement of serum iCa, iMg, PTH, and 25-hydroxyvitamin D (vitamin D) concentrations, as well as serum and urine tCa, total magnesium (tMg), and creatinine concentrations. Ionized calcium and magnesium concentrations were measured with an ion-sensitive electrode.11 PTH was measured by a 2-site immuno-radiometric assay12 as previously described. The vitamin D assay consisted of a 2-step procedure involving rapid extraction of 25-hydroxyvitamin D from serum with acetonitrile followed with an assay13 by an equilibrium radioimmunoassay procedure. A chemistry analyzer14 was used to measure serum and urine electrolyte concentrations with indirect potentiometry and serum and urine creatinine concentrations with a kinetic modification of the Jaffe procedure.
A Shapiro–Wilk test was used to assess the assumption of normality. LPS treatment resulted in nonnormal distribution of data; therefore, logarithmic transformation of data was done before analysis. The influence of time and treatment on variables was assessed between saline- and LPS-treated dogs by analysis of variance with repeated measures comparing the factors “time” and “treatment.” All parameters displayed a significant interaction between time and treatment (P < .001), allowing for differences between saline- and LPS-treated dogs to be assessed at different time points with paired t-tests. For the t-tests, the level of significance was set at alpha ≥ 0.0063 after Bonferroni's correction. Commercially available software was used for the statistical analyses15 and graph generation.16
Baseline values for all vital signs and laboratory parameters were not statistically different between dogs before administration of saline or LPS. All dogs developed lethargy and gastrointestinal signs within 30 minutes of LPS administration. All dogs vomited 2–6 times (median, 3.5 times) and 5 of 6 dogs developed diarrhea. All dogs developed fever by 2 hours after LPS administration as well as a decrease in MAP by 1 hour after LPS administration (Supporting Information, Fig S1). Five of 6 dogs required IV fluid resuscitation 1–2 hours after LPS administration because of hemodynamic instability. Lactated Ringer's solution was administered IV at doses of 10–50 mL/kg per dog (median, 20 mL/kg). By 6 hours after LPS administration, lethargy was improved and vital signs returned to normal. Activity level also was normal by 24 hours after LPS administration. Other signs noted included blepharospasm and periorbital swelling in 1 dog, icteric serum and an episode of regurgitation in 1 dog, and intermittent coughing in 1 dog. All these signs resolved within 6–24 hours after LPS administration and none of the dogs required additional therapeutic intervention. No long-term clinical sequelae or laboratory abnormalities were detected in any of the dogs.
Serum iCa and tCa concentrations decreased in dogs after LPS administration; serum iCa concentration was decreased between 4 (P = .0031) and 12 (P = .0011) hours and serum tCa concentration was decreased between 2 (P < .001) and 24 (P = .0039) hours (Fig 1). Total Mg concentration was decreased in dogs only at 2 hours after LPS administration (P < .001) and iMg concentration was not significantly changed after LPS administration at any time point (Fig 2). Total calcium concentration decreased in concordance with iCa concentration in dogs after LPS administration, but tMg concentration did not (Fig 3).
Parathyroid hormone concentrations were increased in dogs between times 4 (P = .0038) and 24 (P = .0042) hours after LPS administration (Fig 4). Vitamin D concentrations were decreased at times 2 (P < .001), 6 (P < .001), and 12 (P = .0033) hours after LPS administration (Fig 4). Thus, dogs exhibited hyperparathyroidism as an appropriate response to the decrease in iCa concentration, but hypovitaminosis D also was noted in association with the decrease in iCa concentration (Fig 5).
Venous blood gas results indicated that the dogs developed metabolic acidosis after LPS administration characterized by decreases in venous pH, bicarbonate concentration, total carbon dioxide (CO2), and base excess at various time points throughout the study (Supporting Information, Fig S2). Venous partial pressure of CO2 (PCO2) was not significantly changed at any time point after LPS administration (data not shown). Urine tCa concentration was below the limit of detection of the laboratory for all dogs after LPS administration (data not shown).
In the present study, the administration of LPS (endotoxin) to healthy dogs resulted in fever, tachycardia, tachypnea, and hypotension, as well as vomiting and diarrhea. Both serum iCa and tCa concentrations decreased in dogs after LPS administration. This finding could be attributed to hypovitaminosis D, but is unlikely the result of hypoparathyroidism, alkalosis, or increased calciuresis. The potential role of hypomagnesemia in the development of hypocalcemia is unclear because there was only a transient decrease in tMg concentration with no changes in iMg concentration, the biologically active form.
Ionized hypocalcemia occurs in several species during naturally occurring and experimental sepsis. Specifically, ionized hypocalcemia occurs during naturally occurring sepsis in people,[1-6] cats, and dogs,[10, 12] as well as in experimental sepsis in pigs, horses, and rats. Recent retrospective studies in dogs and cats with naturally occurring sepsis document incidences of ionized hypocalcemia of 24% and 93%, respectively. In dogs, the severity of ionized hypocalcemia is associated with a longer duration of hospitalization and increased risk of cardiopulmonary arrest.
Unfortunately, the hormones and substances that regulate magnesium concentration in the body are incompletely understood. Although PTH is thought to increase serum iMg concentrations by decreasing renal excretion of magnesium, the true effect of both PTH and vitamin D on magnesium homeostasis is not clear. Hypomagnesemia is a suggested cause of ionized hypocalcemia during sepsis, because of its involvement in calcium homeostasis. Ionized magnesium, comprising 63% of serum tMg concentration in dogs, is the biologically active form and is responsible for stimulating the release of PTH and acting as a cofactor for PTH activity in the skeletal system and kidneys, as well as hydroxylation (activation) of 25-hydroxyvitamin D.[24-28] Serum tMg concentration represents < 1% of total body magnesium; therefore, serum iMg concentration is considered a more accurate reflection of magnesium status.[29, 30] Hypomagnesemia has been associated with hypoparathyroidism and hypovitaminosis D in people.[24-27] However, it is unlikely that hypomagnesemia is a cause of hypocalcemia during endotoxemia in dogs, because iMg concentration was unchanged during the present study.
The parathyroid response to hypocalcemia in patients with sepsis is variable, probably because parathyroid hormone release is stimulated by both decreases in serum iCa concentration and increases in circulating catecholamine concentrations. Previous studies identified relative hypoparathyroidism characterized by normal to low serum PTH concentrations in the face of concurrent ionized hypocalcemia in septic people and horses.[1, 8] Relative hypoparathyroidism might be secondary to multiple organ dysfunction syndrome or cytokine-mediated suppression of the parathyroid glands.[31, 32] Given the increase in serum PTH concentration beginning 2 hours after LPS administration and lasting the duration of the present study, hypoparathyroidism is a highly unlikely cause of hypocalcemia during endotoxemia in dogs. Decreased serum iCa concentrations overlapped almost completely with the increase in serum PTH concentration. It is unclear, however, whether the increase in serum PTH concentration was an appropriate response to the hypocalcemia or simply a result of catecholamine release after LPS injection. Increased serum PTH concentrations also are documented in critically ill human patients and are associated with increased mortality.[33, 34] This could be because more severely ill patients with sepsis have higher increases in PTH secretion caused by β-adrenergic stimulation of the chief cells in the parathyroid gland by catecholamines.
Serum vitamin D concentrations were significantly decreased in dogs after LPS administration throughout almost the entire duration of the present study, in concordance with decreased serum iCa concentrations. Therefore, hypovitaminosis D is a possible cause of ionized hypocalcemia in dogs with endotoxemia. Serum vitamin D concentrations measured in this study were 25-hydroxyvitamin D (calcidiol), which has 500 times less biological activity than 1,25-dihydroxyvitamin D (calcitriol), the active form of the hormone. 25-hydroxyvitamin D is produced in the liver and is the major circulating form of vitamin D, which is available for further activation by 1α-hydroxylation in the kidney. Impaired 1α-hydroxylase activity and subsequent 1,25-dihydroxyvitamin D deficiency occurs in septic human patients and is associated with decreased total calcium concentrations. Recently, vitamin D deficiency has been documented in healthy and critically ill people, and its role in neoplastic, cardiovascular, autoimmune, and infectious diseases is under investigation. Interestingly, LPS and infectious agents stimulate toll-like receptors on macrophages and monocytes, which signal the expression of the vitamin D receptor and 1α-hydroxylase. Activated vitamin D increases the expression of peptides that promote innate immunity and destroy infectious agents or LPS. Therefore, vitamin D deficiency might reduce the natural response to infection and predispose to increased morbidity.
Decreases in 25-hydroxyvitamin D, in association with decreases in vitamin D- (Gc-) binding protein, have been documented in critically ill people with and without sepsis. Decreased vitamin D-binding protein also has been documented in experimental endotoxemia in rats.Vitamin D-binding protein is the major carrier for both forms of circulating vitamin D in humans, mice, rats, cows, horses, and dogs. It is required to recover filtered 25-hydroxyvitamin D from the urine, and thus decreased vitamin D-binding protein might lead to renal wasting of 25-hydroxyvitamin D. Vitamin D-binding protein might also transport LPS and bind to circulating neutrophils to simulate chemotaxis to specific tissue sites. Reductions in vitamin D-binding protein are associated with decreased survival in people with septic shock as well as the development of organ dysfunction and sepsis after trauma.[41-43] Supplementation with vitamin D improved survival 24–48 hours after LPS injection in endoxotemic mice. Thus, hypovitaminosis D in dogs in the present study could be related to an LPS-induced reduction in vitamin D-binding protein, with a subsequent loss of 25-hydroxyvitamin D in the urine. Urine 25-hydroxyvitamin D, serum 1,25-dihydroxyvitamin D, and vitamin D-binding protein concentrations were not measured in the dogs during the present study.
Limitations of the present study include its nonrandomized design and the young adult age of the dogs. Some dogs were given large amounts of IV fluids during the LPS phase of the study, which could have affected their ionized calcium concentrations. The volume of IV fluids was not controlled among individual dogs and IV fluids were not administered to dogs during the placebo phase of the study. In addition, given that the majority of dogs were < 1 year old, it is likely that they were still growing and thus their calcium regulation in response to endotoxemia might be different from that of adult dogs. However, each dog served as its own control, and thus the results of the present study are still valid. Calcium concentrations vary dramatically in juvenile dogs depending on whether or not they are experiencing a growth spurt and thus a reference range for serum iCa concentration in growing dogs has not been established (P.S., personal communication). It is possible that growing dogs will not exhibit the same response to LPS with regard to calcium homeostasis in comparison to adult dogs.
Another limitation of the present study is that additional information pertaining to other possible mechanisms that could have contributed to ionized hypocalcemia was not measured such as serum lactate, calcitonin, vitamin D-binding protein, or extravascular calcium concentrations. Therefore, chelation with lactate, increases in calcitonin or vitamin D-binding protein, and redistribution of calcium cannot be ruled out as causes of hypocalcemia in dogs with endotoxemia. Unfortunately, assays for serum or urine calcitonin or vitamin D-binding protein were not available at the time of the study. Lastly, extrapolating the results of the present study to naturally occurring sepsis in dogs is limited by the use of endotoxemia as a model for sepsis. Although the clinical signs and inflammatory response to endotoxin are similar to the changes that occur during naturally occurring sepsis,[45-47] naturally occurring sepsis is substantially different from endotoxemia and other experimental models of sepsis. Therefore, findings from the present study might not coincide with what occurs during naturally occurring sepsis in dogs.
The present study showed that dogs with endotoxemia exhibit decreases in serum tCa and iCa concentrations, which are associated with hypovitaminosis D. These disturbances might be because of an LPS-mediated decrease in vitamin D-binding protein and subsequent loss of vitamin D in the urine. Additional studies are needed to examine the kinetics of calcium, vitamin D, vitamin D-binding protein, and lactate during sepsis in dogs and clinical studies are needed to determine whether vitamin D supplementation ameliorates ionized hypocalcemia or improves outcome during naturally occurring sepsis in dogs. Hypomagnesemia, hypoparathyroidism, alkalosis, and increased calciuresis were not associated with hypocalcemia in dogs with endotoxemia in this study.
The authors thank Drs Amy DeClue, Ashley Cruse, and Kevin Coleman for their assistance with the planning and completion of this study, as well as Samantha Berner and Maria Stone for their assistance with animal handling and sample collection. Funding for this study was provided in part by the Ontario Veterinary College Department of Clinical Studies.
Escherichia coli serotype 0127:B8, Sigma-Aldrich, St. Louis, MO
Dexdomitor, Pfizer Animal Health, New York, NY
BD Intracath, Becton Dickinson Infusion Therapy Systems, Sandy, UT
8F 22” Kendall Urethral Catheter, Tyco Healthcare Group, Mansfield, MA
Antisedan, Pfizer Animal Health
Cardell 9401 Blood Pressure Monitor, Sharn Veterinary, Tampa, FL
Lactated Ringers, Baxter Healthcare Corporation/Hospira, Lake Forest, IL
CG8+ cartridge, Abbott Laboratories, Abbott Park, IL
iSTAT handheld, Abbott Laboratories
BD Vacutainer No Additive, Beckton, Dickinson, and Company, Franklin Lakes, NJ
Novo 8 + electrolyte analyzer, Nova Biomedical, Waltham, MA
Intact PTH assay, Diagnostic Systems Laboratories, Webster, TX
25-hydroxyvitamin D total assay, Diasorin, Stillwater, MN
Olympus AU640e, Olympus America, Inc, Center Valley, PA
SAS v.9.1, SAS Institute, Cary, NC
Prism 5, GraphPad Software, La Jolla, CA