FORGOT CALCIUM? ADMISSION IONIZED-CALCIUM IN TWO CIVILIAN RANDOMIZED CONTROLLED TRIALS OF PRE-HOSPITAL PLASMA FOR TRAUMATIC HEMORRHAGIC SHOCK

Abstract

Background:

Randomized clinical trials(RCTs) support the use of pre-hospital plasma in traumatic hemorrhagic shock, especially in long transports. The citrate added to plasma binds with calcium, yet most pre-hospital trauma protocols have no guidelines for calcium replacement. We reviewed the experience of two recent pre-hospital plasma RCTs regarding admission ionized-calcium (i-Ca) blood levels and its impact on survival. We hypothesized that pre-hospital plasma is associated with hypocalcemia, which in turn is associated with lower survival.

Methods:

We studied patients enrolled in two institutions participating in pre-hospital plasma RCTs (Control=Standard-of-care; Experimental=Plasma), with i-Ca collected prior to calcium supplementation. Adults with traumatic hemorrhagic shock(SBP≤70 mmHg or 71–90mmHg+HR≥108bpm) were eligible. We use generalized linear mixed models with random intercepts and Cox proportional hazards models with robust standard errors to account for clustered data by institution. Hypocalcemia was defined as i-Ca≤1.0mmol/L.

Results:

Of 160 subjects(76% men), 48% received pre-hospital plasma, median age 40years(IQR:28–53), 71% suffered blunt trauma, median ISS=22(IQR:17–34). Pre-hospital plasma and control patients were similar regarding age, sex, ISS, blunt mechanism, and brain injury. Pre-hospital plasma recipients had significantly higher rates of hypocalcemia compared to controls (53% vs 36%, Adjusted Relative Risk, aRR=1.48; 95%CI: 1.03–2.12, p=0.03). Severe hypocalcemia was significantly associated with decreased survival(Adjusted Hazard Ratio:1.07;95%CI:1.02–1.13,p=0.01) and massive transfusion(aRR= 2.70;95%CI:1.13–6.46, p=0.03), after adjustment for confounders(randomization group, age, ISS, and shock index).

Conclusion:

Pre-hospital plasma in civilian trauma is associated with hypocalcemia, which in turn predicts lower survival and massive transfusion. These data underscore the need for explicit calcium supplementation guidelines in pre-hospital hemotherapy.

Introduction

Resuscitation techniques employing blood components have progressed from crystalloid alone to balanced blood components, and early plasma resuscitation in high-risk patients. These strategies were introduced for patients with liver injuries(1) and later expanded to all patients at risk of massive transfusion (2). In the last 15 years, observational studies in military settings showed reduced mortality with high ratios of plasma to red blood cells (RBC) during resuscitation (3). Subsequently, observational civilian trauma studies reproduced these outcomes (4). While a civilian randomized clinical trial(RCT) failed to confirm this benefit on 24-hour and 30-day mortality, it decreased hemorrhagic deaths at 24 hours (5).

An evolving concept of plasma-first resuscitation was postulated (6) and the Department of Defense funded two RCTs in civilian settings to test these effects. The Prehospital Plasma during Air Medical Transport in Trauma Patients at Risk of Hemorrhage (PAMPer) trial (7) demonstrated a significant mortality reduction in air-transported patients, who received pre-hospital plasma compared to normal saline. Conversely, the Control of Major Bleeding After Trauma (COMBAT) trial (8) did not demonstrate a survival advantage with early plasma in a short transport, ground-ambulance system. Differences in the proportion of blunt vs penetrating mechanism (COMBAT had more penetrating injuries than PAMPer) (9) and transport time (COMBAT had shorter transport time than PAMPer) (10) were invoked to explain the disparity in the outcomes of these two RCTs.

Hypocalcemia could be a factor contributing to dampen the benefit of pre-hospital citrated plasma, as studies in pre-hospital and hospital settings have detected it in association with transfusions (11–13). The discovery that citrate rendered whole blood incoagulable was a major advance in transfusion medicine (14). Although the effect of citrate on coagulation via binding calcium was known since the 1800’s, it was only in 1915 that several physicians around the globe used sodium citrate as an anticoagulant to transform the transfusion procedure from direct to indirect (14, 15). During World War I, Oswald Robertson, an American Army officer, considered the developer of the first blood bank, published a remarkable paper describing transfusions of stored red cells given near the front line in France (16). However, the residual iatrogenic effect of reduced circulating calcium due to binding with citrate in patients following transfusion, which had been appreciated for over 100 years, fueled caution (17). Adams et al. in 1944 (18) reported that “citrate intoxication” driving lethal hypocalcemia in animals was found to occur after greater than 60% blood loss was replaced with citrated whole blood, but was rapidly reversible with calcium gluconate. These authors warned that administration of plasma was even more concerning, as the amount of citrate in plasma was greater than in citrated blood. However, due to the perceived rarity of severe hypocalcemia from citrated blood product transfusion, its clinical significance was questioned (19); one study even warned against “the potential dangers of intravenous calcium salt therapy.” Guidelines in the 1960’s (20), however, cautioned for the risks of hypocalcemia associated with citrate intoxication due to rapid administration of blood, and recommended 10ml of 10% calcium gluconate be given for each liter of blood transfused.

Recent data suggest that severe hypocalcemia is common in patients undergoing massive transfusion(11), and severe hypocalcemia has been associated with an increased risk of death(21). A recent military analysis found that hypocalcemia was as common as 70% in patients treated with pre-hospital blood products(13). Yet the 2018 Advanced Trauma Life Support guidelines (22) do not recommend calcium supplementation for most transfusions.

The two pre-hospital RCTs (COMBAT and PAMPer) (7, 8) provided an opportune setting to address the effects of early citrated plasma resuscitation on calcium homeostasis. We hypothesize that patients who received pre-hospital plasma compared to crystalloid resuscitation would be at a higher risk of hypocalcemia upon presentation to the hospital, and those patients with hypocalcemia would be at increased risk of mortality and massive transfusion.

Methods

We included adults with traumatic hemorrhagic shock (systolic blood pressure,SBP≤70 mmHg or 71–90mmHg +Heart Rate, HR ≥108bpm) enrolled in two institutions participating in two pre-hospital plasma RCTs: 1)University of Pittsburgh Medical Center (UPMC), one of the trauma centers participating in the PAMPer trial (7); and 2)Ernest E Moore Shock Trauma Center at Denver Health (DH) in Denver, Colorado, the single center in the COMBAT trial (8). We were unable to obtain i-Ca levels from the other facilities participating in PAMPer. Both trials were approved by the Food and Drug Administration, the Human Research Protection Office of the Department of Defense, and the institutional review boards at the participating sites. Both trials were pragmatic and designed simultaneously with similar methods to be subsequently harmonized (7, 8). In brief, adult patients (COMBAT: age≥18 years; PAMPer: 18–90 years) with hemorrhagic shock (SBP≤70 mmHg or 71–90mmHg with HR ≥108bpm) were randomized in the field to receive either two units of universal donor (AB) thawed plasma (Experimental) or standard of care (Control: COMBAT=normal saline; PAMPer=normal saline with or without red blood cells if required). Exclusion criteria were prisoner status, known pregnancy, isolated penetrating injury to the head, asystole or cardiopulmonary resuscitation before randomization, known objection to blood products, opt-out bracelets or necklaces, or family objection to the patient’s enrolment. PAMPer also excluded documented cervical cord injury. The major difference between trials was the mode of transport: PAMPer was limited to patients transported by air, using nurse-paramedic teams over relatively long distances, compared to COMBAT patients, transported by DH affiliated ambulances, using paramedics over short distances in Denver City and County.

For this study, we selected patients with blunt or penetrating injuries, for whom ionized calcium (i-Ca) was collected prior to calcium supplementation. Although i-Ca measurement was not part of the studies’ protocols, it was often obtained as part of the resuscitation protocol of both institutions. The i-Ca was measured in arterial blood gas samples, thus reflecting actual free calcium in-vivo concentrations regardless of altered pH and plasma proteins. Notably, the arterial pH of these patients upon admission was similar in the two study groups (Table 1). Hypocalcemia was defined as in previous studies as i-Ca≤1.0mmol/L (11). Hypercalcemia was defined as i-Ca>1.25mmol/L (21). Traumatic brain injury(TBI) was defined as Abbreviated Injury Scale(AIS) score for head/neck>2. Injury severity was assessed by the Injury Severity Score(ISS). Massive transfusion was defined as >10 RBC units or death within 24 hours postinjury.

Table 1:

Characteristics and outcomes stratified by study group

	Control group N=84 (52.5%)	Experimental group N=76 (47.5%)	P-value
	N (%) or Median (IQR)	N(%) or Median (IQR)
Trial
COMBAT	33 (39.3%)	30 (39.5%)	1.00
PAMPer	51 (60.7%)	46 (60.5%)
Age (years)	39.5 (26–52)	41 (29–54)	0.72
Female sex	18 (21.4%)	20 (26.3%)	0.58
Injury Severity Score	23 (17–33)	22 (15–35)	0.56
Field Heart Rate (bpm)	115 (108–128)	119 (109–127)	0.58
Field SBP (mmHg)	73 (61–83)	72 (63–81)	0.83
Field Shock Index	1.53 (1.37–1.87)	1.60 (1.38–1.85)	0.49
Field Glasgow Coma Score	12 (3–15)	14 (3–15)	0.32
Scene to ED time (min)	35 (20–54)	39 (21–55)	0.62
Blunt mechanism	64 (76.2%)	50 (65.8%)	0.16
TBI (AIS Head>2)	29 (34.5%)	24 (31.6%)	0.74
Calcium
Time to first i-CA (min)	41 (9–110)	47 (22–110)	0.63
First i-Ca (mmol/L)	1.03 (0.97–1.09)	0.99 (0.94–1.06)	0.02
Hypocalcemia (i-Ca<1.0mmol/L)	30 (35.7%)	40 (52.6%)	0.03
Ca supplementation received	58 (69.1%)	45 (59.2%)	0.11
Minutes to Ca supplementation	126 (61–469)	99 (43–228)
Physiology/Coagulation
Admission pH	7.24 (7.15–7.33)	7.25 (7.18–7.32)	0.66
Admission Hemoglobin (g/dL)	12.7 (11.2–14.1)	12.3 (10.3–14.3)	0.29
Admission Prothrombin time/INR	1.26 (1.10–1.50)	1.2 (1.1–1.4)	0.39
R-TEG R (sec)	0.8 (0.7–1.0)	0.8 (0.8–0.9)	0.72
R-TEG MA (mm)	56.5 (49.0–60.6)	56.5 (51.9–62.9)	0.40
R-TEG Angle (degrees)	69.4(61.1–73.1)	70.5 (64.5–75.3)	0.42
R-TEG LY30 (%)	0.7 (0.1–2.1)	0.2 (0–1.7)	0.13
Transfusions/Fluids/Tranexamic acid
RBC units/24hrs	3 (0–8)	2 (0–8)	0.96
Pre-hospital RBC received (PAMPer only)	6 (15.8%)	4 (12.1%)	0.74
Massive transfusion	15 (17.9%)	16(21.1%)	0.69
Plasma units/24hrs	0 (0–4)	2 (2–7)	<0.001*
Platelet units/24hrs	0 (0–1)	0 (0–1)	0.95
Cryoprecipitate units/24hrs	0 (0–0)	0 (0–0)	0.48
Crystalloids (ml)	350 (150–800)	250 (0–750)	0.17
Tranexamic acid administered	2 (2.4%)	0	0.50
Outcomes
ICU Days	5 (2–14)	7 (2–14)	0.52
Mortality	15 (17.9%)	9 (11.8%)	0.38

^*by design. R-TEG: rapid thrombelastogram; MA: maximum amplitude; RBC: red blood cells; ICU: intensive care unit; INR: prothrombin time international normalized ratio; TBI: traumatic brain injury (Abbreviated Injury Scale Head/neck >2); i-Ca: ionized calcium; Ca: calcium

Statistical analysis:

we expressed numerical variables as median and interquartile range(IQR) and categorical variables as number and percent. We performed univariate comparisons with Chi-square or Fisher exact tests for categorical variables, and with t-test or Mann-Whitney U-test for numerical variables. We use generalized linear mixed models(GeLM) with random intercepts and Cox proportional hazards models with robust standard errors to account for clustered data by institution/trial. For analysis of survival and massive transfusion, variables fitting the definition of a confounder (i.e., associated with both the outcome and the exposure), and univariately associated with the outcome at p<0.25 were included in the models. Violations of the proportionality assumption were checked and, when present, remedied by entering an interaction of the offending variable with time. Kaplan-Meier curves illustrated the survival by i-Ca strata. All tests were two-tailed with significance established at p<0.05. The analyses were performed with SAS vs 9.4 (SAS Institute, Cary, NC).

Results

Of 150 PAMPer subjects admitted to UPMC, 97 (64.7%) had i-CA measured before calcium supplementation, while in COMBAT, 63 of 125 (50.4%) subjects had this measurement (Figure 1). This resulted in a study sample of 160 subjects, of whom 48% received pre-hospital plasma. Table 1 shows the characteristics of these patients, stratified by experimental (pre-hospital plasma) and control (pre-hospital normal saline) groups. The groups were well balanced regarding baseline characteristics. Of note, the groups had similar median time to first i-Ca measurement before calcium supplementation: 41minutes(9–110) in the control group and 47minutes(22–110) in the pre-hospital plasma group(p=0.63). The differences in mortality and massive transfusions between the experimental and control groups were not statistically significant.

Figure 1:

CONSORT diagram

Overall, 44% had hypocalcemia, but the experimental group had a higher incidence of hypocalcemia than the control group (35.7% vs 52.6%, p=0.03), which remained significant after accounting for clustered data by institution/trial (GeLM, p=0.03, Adjusted Relative Risk, aRR=1.48; 95% Confidence Interval, CI: 1.03–2.12). Only two individuals had mild hypercalcemia (1.42 and 1.55mmol/L, one in each study site, none received pre-hospital plasma, both survived).

Table 2 depicts the characteristics of patients with and without hypocalcemia. Mortality in the hypocalcemia group was 18.6% compared to 12.2% in the normocalcemia group. Massive transfusion was required in 28.6% of patients with hypocalcemia, and in 12.2% of patients with normocalcemia. Hypocalcemia was associated with significantly lower angle and MA.

Table 2:

Patients characteristics stratified by normocalcemia (ionized calcium> 1.0 mmol/L) and hypocalcemia (ionized calcium ≤1.0mmol/L)

	Normocalcemia i-Ca>1.0mmol/L (n=90) N(%) or Median (IQR)	Hypocalcemia i-Ca≤1.0mmol/L (n=70) N(%) or Median (IQR)	p-value
Randomized to Plasma group	36 (40.0%)	40 (57.1%)	0.03
Trial=COMBAT	32 (35.5%)	31 (44.3%)	0.33
Trial=PAMPer	58 (64.4%)	39 (55.7%)
Age (years)	43 (26–58)	36 (28–51)	0.17
Female sex	22 (24.4%)	16 (22.9%)	0.81
Injury Severity Score	22 (13–20)	27(17–34)	0.05
Field Systolic Blood Pressure (mmHg)	79 (68–86)	70	<0.001
Field Heart Rate (bpm)	112 (108–122)	120	0.04
Field Shock Index	1.44 (1.33–1.63)	1.72	<0.001
Minutes to ED	39 (22–54)	34 (20–55)	0.42
Blunt trauma	67 (74.4)%	47 (67.1%)	0.31
TBI (AIS Head/neck>2)	28 (31.1)%	25 (35.7%)	0.54
Calcium
i-Ca (mmol/L)	1.08 (1.04–1.11)	0.95 (0.86–0.97)	<0.0001
Minutes to i-Ca measurement	33 (18–218)	37 (20–196)	0.16
Coagulation
Prothrombin time/INR	1.20 (1.10–1.40)	1.31 (1.16–1.50)	0.11
R-TEG ACT (seconds)	121 (105–136)	121 (105–128)	0.43
R-TEG ANGLE (degrees)	70.8 (65.3–75.8)	67 (60.6–72.2)	0.02
R-TEG MA (mm)	57.5 (53.8–64.4)	54 (38.7–59.9)	0.005
R-TEG LY30 (%)	0.70 (0–2.0)	0.5 (0–2.0)	0.69
Hyperfibrinolysis (LY30>7.6%)	3 (6.3%)	2 (5.0%)	0.82
Physiologic lysis (LY30 0.6–7.6%)	23 (47.9%)	17 (42.5%)
Lysis shutdown (LY30 <0.6%)	22 (45.8%)	21 (52.5%)
Transfusions /Fluids/Tranexamic acid
RBC units/24 hours	1 (0–5)	5 (2–10)	0.0002
Massive transfusion (>10 red blood cell units or death/24 hours)	11 (12.2%)	20 (28.6%)	0.009
Plasma units/24hrs	2 (0–4)	2 (1–7)	0.007
Platelet units/24hrs	0 (0–0)	0 (0–1)	0.30
Cryoprecipitate units/24hrs	0 (0–0)	0 (0–0)	0.0003
Crystalloids (ml)	250 (100–600)	400 (0–800)	0.43
Tranexamic acid administered	1 (1.1%)	1 (1.4%)	1.00
Outcomes
Length of ICU stay (days)	6 (2.0–11.0)	6 (2–17)	0.59
Mortality	11 (12.2%)	13 (18.6%)	0.26

R-TEG: rapid thrombelastogram; MA: maximum amplitude; RBC: red blood cells; ICU: intensive care unit; INR: prothrombin time international normalized ratio; TBI: traumatic brain injury (Abbreviated Injury Scale Head/neck >2); i-Ca: ionized calcium; Ca: calcium

Kaplan-Meyer curves (Figure 2) showed that the control group with hypocalcemia had the lowest survival, followed by the experimental group with hypocalcemia. Confounder candidates for the association of hypocalcemia with death and with massive transfusion were: age, ISS as well as systolic blood pressure (SBP), heart rate (HR) and shock index upon admission as well as randomization group(Table 2). Because SBP and HR are components of the shock index (thus highly correlated), we entered shock index in the Cox proportional hazards model as a confounder of the association of hypocalcemia with survival. Hypocalcemia was independently associated with decreased survival (Table 3). A quadratic term testing a U-shaped association of i-Ca levels with survival was tested; because it was not significant, it was dropped from the model. In addition, using the GeLM to adjust for the above confounders and account for clustered data by institution/trial, hypocalcemia was significantly associated with massive transfusion (aRR= 2.70; 95%CI: 1.13–6.46, p=0.03).

Figure 2:

Kaplan-Meier survival curves by study group (Pre-hospital plasma and Control) and hypocalcemia (Ionized calcium, i-Ca≤1.0mmol/L)

Table 3:

Cox proportional hazards model testing the effect of hypocalcemia on survival adjusted for age, injury severity score (ISS), and shock index

Variable	Hazard Ratio	95% Hazard Ratio Confidence Limits		P-value
Hypocalcemia	1.07	1.02	1.13	0.01
Experimental vs Control group	0.92	0.75	1.12	0.40
Age (10 years)	1.02	1.02	1.03	<.0001
Injury Severity Score (10 points)	1.04	1.01	1.07	0.004
Shock Index (1.00 unit)	1.02	1.01	1.03	0.002

Discussion

Combining two RCTs that tested plasma-first resuscitation in the pre-hospital setting identified a significant risk of hypocalcemia after administration of citrated blood products. Hypocalcemia was present in roughly one third of patients who did not receive plasma, compared to over half in patients who received plasma. Hypocalcemia was an independent predictor of mortality and of massive transfusion after adjusting for confounders (age, injury severity and shock index). While hypocalcemia appeared to dampen the survival benefit of pre-hospital plasma (Figure 2), the current study was not designed nor powered to assess the effect of the pre-hospital plasma on survival or massive transfusion

Optimal coagulation requires calcium as well as close-to-normal acid-base balance, hematocrit, and temperature; if these essential elements are missing, hemostatic therapy may be in vain as stable clotting cannot occur (23). Calcium plays an essential role in the formation and stabilization of fibrin polymerization sites and, consequently, it has an impact on all platelet-dependent functions. Yet laboratorial coagulation tests may mask the negative impact of hypocalcemia on coagulation, as blood samples are re-calcified prior to being assayed. In addition, calcium is essential in myocardial and vascular smooth muscle contraction.

Hypocalcemia’s association with blood product use in these two RCTs confirms the recent military and civilian experience (11–13, 21). Our study, however, advances this knowledge by examining the specific role of pre-hospital plasma, and demonstrating that even a relatively small volume of a blood component may decrease i-Ca levels.

The cause of hypocalcemia following hemorrhagic shock and injury remains unclear, yet there is an alarming paucity of recently published primary data addressing causes and interventions (24). Indeed, in the 34^th William Fitts Jr Oration, presented at the 67^th Annual Meeting of the American Association for the Surgery of Trauma in 2008, Charles E. Lucas (25) summarized his group’s decades-long work on this topic, and alerted for the little scientific attention given to calcium. Moreover, gaps in knowledge about the citrate content in blood products and its impact on i-Ca seem to exist. A recent survey of a large, urban, academic Level I trauma center in the northeast United States revealed that the majority of the respondents erroneously believed that red blood cells were the blood component with the largest amount of citrate, and ignored that plasma and platelets contained 60% of the total citrate (26, 27).

Mechanisms other than citrated blood products are involved in the development of hypocalcemia. Indeed, Vivien et al. (28) compared observed i-Ca and expected i-Ca (corrected for hemodilution, pH, lactate and colloid binding of calcium) upon admission in trauma patients prior to receiving any blood component, and noted a significant difference between them in severely injured patients, presumably due to shock and/or ischemia-reperfusion. Another potential causal mechanism for hypocalcemia could be low hepatic clearance of citrate due to defective hepatic perfusion caused by the hemorrhagic shock (29). Lucas et al. (30) in the early 1980s demonstrated that low total calcium and i-CA in 41 injured patients who required blood transfusions were associated with shock time and hypoalbuminemia, as well as blood products and crystalloid administration during resuscitation. Further investigation in animals by this group suggested a direct association between shock and hypocalcemia (31).

Our findings regarding the association of hypocalcemia with death and the need for massive transfusion confirm the observations of other investigators (32). Our study was also consistent with the previous association of increased mortality in patients with hypocalcemia who were critically ill due to reasons other than trauma(33) and with a large retrospective study of 15,000 critically ill patients, in which hypocalcemia remained a predictor of mortality (34).

Limitations of this study include the lack of i-Ca measurements in all patients enrolled in the RCTs, although a very good balance in baseline risk factors was retained in the subgroup with these measurements, strongly suggesting selection bias was not at play. This preserved the internal validity of the assessment of the association between pre-hospital plasma with hypocalcemia. In contrast, survivor bias and the modifying effects of pre-existing disease severity may have confounded the impact of hypocalcemia on survival. It is important to note that mode of transport (air transport in PAMPer and ground in COMBAT) is also a surrogate for different levels of care (a critical care team nurse + medic vs a single medic in the back of the ambulance), potentially introducing some survival bias (patients have to survive long enough for helicopter to arrive on-site) and selection bias (air transport is only activated if patient care warrants resources unavailable locally, and for more severely injured patients). These biases can potentially limit the generalizability of our results (i.e., less severely injured patients may have lower incidence of hypocalcemia).

We did not conduct analyses on the impact of calcium replacement on survival, as these would be severely flawed due to survivor bias (patients must be alive to receive it), intervention bias (patients with more severe injuries and/or in-extremis are more likely to receive it) and incomplete risk adjustment due to small sample size. Of note, although calcium replacement is usually part of in-hospital massive transfusion protocols, this is not the case in pre-hospital plasma-first resuscitation, or in patients who require less blood products than those for whom massive transfusion protocols are activated. Collectively, these deficiencies underscore the need for large prospective trials testing the signal detected in these solid analyses of the association of citrated-blood resuscitation with hypocalcemia.

In summary, trauma patients resuscitated with pre-hospital plasma often present to the hospital with hypocalcemia, which place them at increased risk of mortality. Citrate in the plasma contributes to hypocalcemia but other causes of low i-Ca remain unclear as some patients who did not receive plasma also had hypocalcemia. Thus, further research into the mechanisms of postinjury hypocalcemia, and associated mortality are needed. A randomized controlled trial will be required to provide definitive answers regarding the optimal therapeutic interventions. Yet while such evidence is unavailable, pre-hospital administration of calcium is a simple, relatively innocuous approach to attenuate injury- and resuscitation-induced hypocalcemia, particularly in patients who received blood products. The recently published European guidelines(35) favored 10% calcium chloride (270 mg of elemental calcium/10 mL) over 10% calcium gluconate (90 mg of elemental calcium/10 mL), especially in the presence of abnormal liver function, because decreased citrate metabolism results in slower release of i-Ca. It is important to note that both calcium salts should be given slowly to avoid cardiac effects, and that extravasation into soft tissue can cause severe lesions (36). Due to the low risk of iatrogenic effects of calcium gluconate, we advocate that 1gram(10ml) be given for every 1 to 2 units of blood products (plasma inclusive), and that admission i-Ca levels be monitored early and often in patients with hemorrhagic shock.