https://orcid.org/0000-0002-3839-1874
Figures
Abstract
Objectives
In adult major trauma patients admission hypocalcaemia occurs in approximately half of cases and is associated with increased mortality. However, data amongst paediatric patients are limited. The objectives of this review were to determine the incidence of admission ionised hypocalcaemia in paediatric major trauma patients and to explore whether hypocalcaemia is associated with adverse outcomes.
Methods
A systematic review was conducted following PRISMA guidelines. All studies including major trauma patients <18 years old, with an ionised calcium concentration obtained in the Emergency Department (ED) prior to the receipt of blood products in the ED were included. The primary outcome was incidence of ionised hypocalcaemia. Random-effects Sidik-Jonkman modelling was executed for meta-analysis of mortality and pH difference between hypo- and normocalcaemia, Odds ratio (OR) was the reporting metric for mortality. The reporting metric for the continuous variable of pH difference was Glass’ D (a standardized difference). Results are reported with 95% confidence intervals (CIs) and significance was defined as p <0.05.
Results
Three retrospective cohort studies were included. Admission ionised hypocalcaemia definitions ranged from <1.00 mmol/l to <1.16 mmol/l with an overall incidence of 112/710 (15.8%). For mortality, modelling with low heterogeneity (I2 39%, Cochrane’s Q p = 0.294) identified a non-significant (p = 0.122) estimate of hypocalcaemia increasing mortality (pooled OR 2.26, 95% CI 0.80–6.39). For the pH difference, meta-analysis supported generation of a pooled effect estimate (I2 57%, Cochrane’s Q p = 0.100). The effect estimate of the mean pH difference was not significantly different from null (p = 0.657), with the estimated pH slightly lower in hypocalcaemia (Glass D standardized mean difference -0.08, 95% CI -0.43 to 0.27).
Citation: Hibberd O, Price J, Thomas SH, Harris T, Barnard EBG (2024) The incidence of admission ionised hypocalcaemia in paediatric major trauma—A systematic review and meta-analysis. PLoS ONE 19(5): e0303109. https://doi.org/10.1371/journal.pone.0303109
Editor: Jean Baptiste Lascarrou, CHU Nantes, FRANCE
Received: January 16, 2024; Accepted: April 18, 2024; Published: May 28, 2024
Copyright: © 2024 Hibberd et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its supporting information files.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Major trauma is a leading cause of death in children [1, 2]. Haemorrhage is a key aetiology of potentially survivable death following trauma, and it is well recognised that the ‘lethal triad’ of coagulopathy, hypothermia, and acidosis can further potentiate haemorrhage [3, 4]. Recently it has been recognised that ionised hypocalcaemia (iHypoCa) also contributes to this ‘lethal triad’, and as such this may now be considered the ‘diamond of death’ [5–7]. An iHypoCa contributes to coagulopathy and cardiovascular decompensation due to its role in clot formation, vascular tone, and cardiac contractility [6–8]. Consequently, current guidelines recommend maintaining normal calcium levels in the bleeding trauma patient [9].
There is good agreement between arterial and venous calcium measurement on blood gases, which are often obtained shortly after hospital arrival in severely injured paediatric patients [10]. Blood gases report ionised calcium (iCa), which is the free-form and physiologically active calcium in the blood [11]. Transfusion of citrated blood products leads to iHypoCa due to calcium chelation with citrate [12]. However, in trauma, early iHypoCa may occur in the absence of exogenous citrate and the aetiology is complex, with multiple potential mechanisms [12, 13]. Pathophysiological mechanisms include dilution by crystalloid fluid administration, cellular calcium shifts secondary to ischaemia and reperfusion injury, calcium-lactate binding, and impaired calcium homeostasis [5–7]. In adult major trauma patients iHypoCa occurs in approximately half of cases and is associated with coagulopathy, increased blood transfusion requirements, and increased mortality [14–19] However, in paediatric major trauma patients data are limited. In critically ill paediatric patients on the Paediatric Intensive Care Unit (PICU) iHypoCa has been observed to be independently associated with more severe organ dysfunction and more frequent amongst non-survivors compared to survivors [20, 21]. In trauma, children may be more sensitive to iHypoCa due to different injury kinematics and developmentally dynamic physiological responses to injury [22, 23]. Therefore, it may not be appropriate to extrapolate results from studies involving adult major trauma patients to the paediatric major trauma patient.
The objective of this systematic review and meta-analysis was to determine the incidence of admission iHypoCa in paediatric major trauma patients and explore whether this is associated with adverse clinical outcomes when compared to normocalcaemia (iNormoCa).
Materials and methods
This study adheres to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist (S1 Checklist) [24].
Eligibility criteria
This systematic review explores the incidence of iHypoCa in paediatric major trauma patients and whether admission iHypoCa (iCa <1.16 mmol/l), compared to iNormoCa (iCa ≥1.16 mmol/l) is associated with a greater incidence of adverse outcomes. Paediatric (<18 years old) major trauma patients (Injury Severity Score >15, or trauma team activation) with a documented iCa concentration on admission to the emergency department (ED) were included. Studies involving patients who had an iCa concentration obtained after the administration of blood products in the ED were excluded, in keeping with the methodology of a systematic review of adult studies [17]. The full Population, Intervention, Comparison, Outcomes, and Study Design (PICOS) eligibility criteria are available in the previously published protocol [25].
Information sources and search strategy
A literature search for articles in the English language was completed using MEDLINE on the EBSCO platform, CINAHL on the EBSCO platform, and Embase on the Ovid platform. Articles were searched from database inception to the search date (03/07/2023). The reference lists of all included studies, and the grey literature were also searched. The search strategy can be found in the online supplementary tables of the published protocol [25].
Selection process
The search strategy was undertaken by a trained librarian from the UK Defence Medical Services Burnett Library (S1 File). Duplications were manually removed and the combined abstracts from the search strategy were then independently screened by two reviewers (OH and JP) to identify studies that met the inclusion criteria. For abstracts meeting inclusion criteria, full texts were retrieved and independently screened against the eligibility criteria by two reviewers and an adjudicating third reviewer (EB).
Data collection process
A standardised data sheet (Microsoft ® Excel for Mac, Version 16.72, 2023) was used to extract data from included studies to facilitate data synthesis and assessment of quality and risk of bias (S2 File). Extracted data were independently verified by the second reviewer, and any discrepancies were adjudicated by the third reviewer. For incomplete data, or different reporting of central tendency (e.g., mean, and median) the corresponding authors of the articles were contacted directly for additional information.
Data items
The full protocol, detailing data items extracted, is published and available online [25]. The Microsoft Excel® data sheet containing full extracted data can be found in S2 File.
Study risk of bias assessment
The risk of bias was assessed for all included studies using the ROBINS-I tool [26]. The risk of publication bias was assessed with funnel plots. The quality of evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach [27]. The quality of evidence for each study was also assessed independently by two authors, with consensus agreement or adjudication by a third author for cases of disagreement.
Effect measures
The primary outcome of this systematic review was the incidence of admission iHypoCa. Secondary outcomes are the associations with adverse outcomes and physiological abnormalities. The endpoint for adverse outcomes was classified dichotomously as the odds of in-hospital mortality. Treatment requirements (such as need for transfusion) and the LOS (hospital and PICU) were also compared between studies. The end point for physiological abnormalities was the mean difference in pH, this was driven by available data and selected due to universal reporting across studies. Other physiological abnormalities compared between studies were markers of coagulopathy, lactate, and the odds of haemodynamic instability defined as an elevated Shock Index Paediatric Age-Adjusted (SIPA) (0-6yrs >1.22, 7-12yrs >1.00, 13-16yrs >0.90) [28, 29].
Ratio measures (odds ratio) and mean differences are used for measures of effect. Where possible, the estimate of effect was presented along with a confidence interval and a p-value.
Synthesis methods
Data were synthesised as per the PRISMA guidelines [24]. Studies were assessed clinically (PICO) and methodologically (study design, comparability, outcome ascertainment, and risk of bias). A narrative synthesis and summary of effect measures were conducted when heterogeneity was deemed too substantial for meaningful meta-analysis. All analysis and plotting were executed with Stata (Stata, version 18MP, 2024, Stata Corp, College Station TX, USA; www.stata.com). To assess the endpoints, we executed random effects modelling. After the number of studies (N) was determined to be small, we employed a Sidik-Jonkman (SJ) approach recommended for small-N meta-analysis [30]. The SJ approach was chosen due to its superior performance for random effects meta-analysis, especially in small studies [31]. Sensitivity to model selection was assessed by re-executing meta-analysis using a second random effects approach, the DerSimonian-Laird model. A third modelling approach, fixed-effect modelling, was executed solely as an indicator of study heterogeneity (i.e., to determine whether results were markedly different from random effects results). The assessment method used for comparing continuous outcomes (e.g., pH differences) was the Glass D metric standardized by control group. This approach was chosen conservatively, since the standard deviations between iHypoCa and iNormoCa groups differed substantially in one study (Epstein). We also report an effect estimate of the non-standardized mean between pH in iHypoCa- vs. iNormoCa groups in S3 File.
Formal heterogeneity and small-study bias (including publication bias) were evaluated using I2 (with 30–60% representing “potentially moderate” heterogeneity), Galbraith plotting, and funnel plots with trim-and-fill (imputed study) analysis [32]. Sensitivity analyses included omitted-study plotting and cumulative meta-analysis. Meta-regression and prediction intervals were planned in the event the analysis accrued the minimum requirement of ten studies [29].
Results
Study selection
A total of 22 abstracts were screened through a database search, reference review, and the grey literature (Fig 1).
Five unique full-text studies were examined, two of which were excluded [33–37]. One study was excluded as it reported total calcium rather than ionised calcium [33]; the other as blood products in the Emergency Department (ED) were received prior to calcium measurement [34]. This resulted in three studies being included in the review [35–37]. All three studies were retrospective cohorts (Table 1) undertaken at a paediatric trauma centre and the quality of evidence was either low or moderate, owing to confounders, risk of bias, and sample size [35–37]. Studies differed in sample size (n = 111–457), inclusion criteria related to trauma team activation, and prevalence of penetrating injury (4%-24%) [35–37].
Assessment of risk of bias
Overall, although all studies were well conducted, a serious to critical risk of bias was identified (Table 2). Bias was largely due to confounding, with differential determinants being unrecorded, for example, the interval time from injury to calcium measurement and pre-hospital treatments were not reported in two studies [36, 37]. Moreover, the selection of patients who received trauma team activation also risked missing major trauma patients who self-present or those with occult injuries identified on subsequent imaging.
Incidence of admission ionised hypocalcaemia
Admission iHypoCa was defined differently by each study (Table 3). Definitions ranged from <1.00 mmol/l to <1.16 mmol/l [35–37]. The overall incidence of iHypoCa was 112/710 (15.8%) in the emergency department for paediatric major trauma patients.
Patient demographics
Two studies observed a statistically significant difference in age between iNormoCa and iHypoCa; patients with iHypoCa were older (Table 1) [35, 36]. However, none of the included studies adjusted for age as a variable in regression modelling. None of the studies demonstrated a statistically significant difference in the Injury Severity Score (ISS) between iHypoCa and iNormoCa patients [35–37]. Across studies, the majority of injuries were from a blunt mechanism (91.9%) with the largest proportion of penetrating injuries being observed in the study by Ciaraglia et al. (38/142 (26.7%)) [35].
Treatments prior to calcium measurement
One study reported the association between prehospital variables and the incidence of iHypoCa [35]. Ciaraglia et al. did not observe any significant difference between groups with respect to prehospital (weight-adjusted) crystalloid volume (iHypoCa 8.04 ml/kg (SD 12.6) compared with iNormoCa 9.46 ml/kg (SD 17.91), p = 0.650) [35]. The mean interval between injury and hospital arrival was not different between iHypoCa and iNormoCa paediatric patients: iHypoCa– 189.6 minutes (standard deviation (SD) 282.1) compared with iNormoCa– 169.4 minutes (SD 173.9), p = 0.744.
Association with adverse outcomes
The requirement for blood transfusion was reported differently in each study. Ciaraglia et al. recorded weight-adjusted transfusion volumes at 4 and 24 hours, and observed at both time intervals that iHypoCa was associated with significantly higher weight-adjusted transfusion volumes (p = 0.002 and p = 0.040 respectively) [35]. Epstein et al. dichotomously recorded the need for blood transfusion within the ED and found this requirement to be significantly different between iHypoCa (7/24 (29.2%)) and iNormoCa (28/433 (6.5%)) (p<0.001) [36]. Gimelraikh et al. recorded the requirement for blood transfusion during the first 48 hours and did not observe this requirement to be significantly different for iHypoCa (6/22 (27.3%)) and iNormoCa (18/89 (20.2%) (p = 0.67) [37]. No studies reported whether paediatric patients received exogenous calcium replacement.
For all studies, hospital and PICU length of stay were reported in all studies and were not observed to be different between iHypoCa and iNormoCa patients (Table 4). The length of stay in PICU and the hospital in general, was observed to be greater for the cohort reported by Ciaraglia et al. suggesting a greater burden of injury [35].
Meta-analysis results for the association of ionised hypocalcaemia with mortality
The I2 value (39%) suggested low heterogeneity. Concerning heterogeneity was also judged unlikely based on non-significant Cochrane’s Q (p = 0.294), a reassuring Galbraith graph (S3 File), and on the fact that fixed-effect modelling (S3 File) generated similar effect estimates to those calculated using random-effects SJ modelling.
The overall calculation (Fig 2) was judged acceptable given the conclusion that heterogeneity did not preclude generation of a pooled effect estimate for the mortality OR. There was a non-significant (p = 0.122) mortality association: iHypoCa mortality OR (with >1 corresponding to higher mortality with iHypoCa) was 2.26 (95%CI 0.80–6.39); full results details are reported in the S3 File. Trim-fill analysis identified possibility of small-study bias, but two imputed studies fell within the region of non-significant p values. With the new study N of five studies the mortality OR 95% CI continued to overlap the null. Full results including contoured funnel plot are presented in the Supplement. Cumulative meta-analysis (S3 File) demonstrated that with increasing study N the effect estimate appeared to stabilize around the pooled effect estimate of approximately 2.2. At all points in the cumulative meta-analysis the effect estimate’s 95%CI overlapped zero. Similar findings were calculated when meta-analysis was repeated using omitted-study assessment (S3 File); omission of Ciaraglia et al. resulted in the pooled effect estimate (of the two remaining studies) approaching significance but leave-one-out analysis did not result in any p estimates dropping below 0.05 and all 95%CIs overlapped the null.
Association with physiological abnormalities
The presence of haemodynamic instability (an elevated SIPA on arrival) was reported in two studies [35, 36]. Ciaraglia et al. observed 28/66 (42.4%) of hypocalcaemic patients to be haemodynamically unstable, and that this was associated with an increased odds of haemodynamic instability when compared to iNormoCa patients OR 3.57 (95% CI 1.65–7.72) [35]. In contrast, Epstein et al. observed 6/24 (25.0%) of iHypoCa patients to be haemodynamically unstable and that this was not associated with an increased odds of haemodynamic instability when compared to iNormoCa patients–OR 0.88 (95%CI 0.34–2.27) [36]. Gimelrakih et al. did not undertake analysis for haemodynamic instability [37]. No studies reported the requirement for vasoactive medications or the requirement for invasive interventions (e.g., surgical intervention or interventional radiology).
Markers of coagulation were reported in two studies (Table 5) [36, 37]. In these studies, only PTT was recorded in both studies, and neither study observed a significant difference between iHypoCa and iNormoCa paediatric patients.
Ciaraglia et al. did not undertake analyses for any markers of clotting, and across studies, PTT was the only clotting test that was recorded in more than one study [35]. Epstein et al. measured PTT, INR, and platelets, using INR >1.5 to dichotomise the presence of coagulopathy, observing no significant difference between iHypoCa and iNormoCa paediatric patients [36]. Gimelraikh et al. measured PT, PTT and platelets, and also observed no significant difference between the iHypoCa and iNormoCa paediatric patients [37].
Meta-analysis results for the mean difference in pH between hypocalcaemic and normocalcaemic patients
The I2 value (57%) indicated potential for moderate heterogeneity (Fig 2). Concerning heterogeneity was judged unlikely based on non-significant Cochrane’s Q (p = 0.100), a reassuring Galbraith graph (S3 File), and on the fact that fixed-effect modelling (S3 File) generated similar effect estimates to those calculated using random-effects SJ modelling.
The overall calculation (Fig 3) was judged acceptable given the conclusion that marked heterogeneity did not preclude generation of a pooled effect estimate for the mean pH difference. There was a non-significant (p = 0.657) pH difference: pH’s standardized mean was calculated to be 0.08 lower in iHypoCa (95%CI -0.43–0.27). The non-standardized mean difference was calculated as pH 0.01 lower in iHypoCa ((95%CI -0.056–0.035, p = 0.657), full results details reported in S3 File).
Trim-fill analysis identified possibility of small-study bias. Two studies were imputed, but with the new study N of five studies the Glass D 95% CI continued to overlap the null. Full results including contoured funnel plot are presented in the S3 File.
Cumulative meta-analysis (details reported in S3 File) demonstrated that with increasing study N the effect estimate approached the null value. At all points in the cumulative meta-analysis the effect estimate’s 95%CI overlapped zero. Similar findings were calculated when meta-analysis was repeated using omitted-study assessment (S3 File); omission of any individual study did not result in a finding of statistically significant pH difference in iHypoCa vs iNormoCa.
Discussion
This review identified three retrospective cohort studies, two studies from Israel and one study from the USA. An iHypoCa was observed in approximately one in six paediatric major trauma patients, and this was reported to be more prevalent in older children. Although there was no significant difference in pH, hospital or PICU LOS, or mortality, there may be an increase in haemodynamic instability and an increased requirement for blood transfusions within the first 24 hours of admission for patients with iHypoCa compared to iNormoCa patients. The included articles do not provide evidence of a causal relationship between iHypoCa and adverse outcomes.
Incidence of ionised hypocalcaemia
The overall incidence of admission iHypoCa in 112/710 (15.8%) paediatric major trauma patients is lower than that reported in adult major trauma patients (13.0–56.2%) [14–19]. This lower incidence in the paediatric population has also been observed for transfusion-related iHypoCa [34, 38]. Similarly to this review, in a systematic review involving adult patients the definitions and incidence of iHypoCa varied between studies and may reflect a lack of consensus data on what represents clinically significant iHypoCa [17]. Overall, the studies included in the adult systematic review used lower cut-offs for iHypoCa than two out of the three studies included in this review, with one study in this review defining iHypoCa as <1.16mmol/l [37]. The inclusion of these higher cut-off values is important as this reflects a common laboratory reference value and may also inform subsequent work related to mild or moderate hypocalcaemia [39, 40].
In this review two out of three studies observed that older age was associated with iHypoCa [35, 36]. Moreover, the three studies included in this review included patients up to the age of 18 years of age, where physiology is more likely to be similar to adults [35–37]. As such, it remains unclear whether calcium imbalance in paediatric trauma should be considered different to the adult trauma population, and further studies that are powered to detect age-related differences in a younger cohort would be useful to clarify relevance.
The incidence of iHypoCa in each of the studies included in this review varied widely (5.3% to 46.5%) [35–37]. The higher incidence of iHypoCa in the study by Ciaraglia et al. (66/142 (46.5%)) may reflect the exclusion of isolated head injuries, the greater proportion of penetrating injuries in this group (38/142 (26.7%)) and differences in institutional trauma team activation criteria (S4 File) [35]. Penetrating injuries are associated with a higher incidence of iHypoCa [18] and trauma-induced coagulopathy [4]; however, markers of coagulopathy were not reported in this study [32]. Head injuries are recognised to increase the fibrinolytic system and are prone to coagulopathy; the true effect of excluding these isolated injuries in Ciaraglia et al.’s study is therefore unclear [41]. As such, the true incidence of iHypoCa and relevant cut-off values in paediatric major trauma remains unclear; however, with 15.8% of the population being identified as having iHypoCa in the ED, this does not appear to be a rare phenomenon.
Treatment requirements
Although the association between iHypoCa and the requirement for blood transfusion was explored differently in each study, this was observed to be significant in the first 24 hours, particularly in the first four hours [35, 36]. Among adult major trauma patients multiple studies have observed iHypoCa to be associated with increased transfusion requirements and increased requirement for massive transfusion [13, 17, 42]. This is particularly important when considering early Trauma Induced Coagulopathy [4, 8].
Additional treatments, such as vasoactive medications, or invasive interventions (interventional radiology or surgery) were not recorded in any of the included studies [35–37]. Moreover, none of the studies included in our review explored the effect of exogenous calcium administration. At present, there is not evidence to support calcium supplementation. Moreover, a recent systematic review of empirical calcium administration for adult patients in cardiac arrest has demonstrated harm [43].
Association with adverse outcomes
This review observed iHypoCa to increase the odds of admission haemodynamic instability two-fold but did not demonstrate any association between hospital or PICU LOS, nor mortality. Calcium concentration play an integral role in both vascular smooth muscle and cardiac contractility [44, 45]. Previous work involving adults has observed both hypotension and a significantly worse shock index amongst hypocalcaemic (iCa <1.0 mmol/l) trauma patients [13, 17–19]. However, these studies used a low cut-off for iHypoCa, with a similar cut-off only used in one out of the three studies explored in this review [33]. Epstein et al. used a cut-off of iCa <1.1mmol/l and did not observe a significantly increased difference, however, an overall meta-analysis of both Ciagliaria et al. and Epstein et al.’s studies did show an overall increased odds of haemodynamic instability [35, 36]. Gimelraikh et al. did not undertake analysis for haemodynamic instability and used a cut-off of iCa <1.16 mmol/l, as such, the association between milder iHypoCa and haemodynamic instability is unknown and further work is required to explore this [37].
This review demonstrated a trend towards increased hospital and PICU length of stay but this was not statistically significant. In a large retrospective cohort study involving 1305 injured paediatric patients (median age 11.2 years old, mean ISS 12.14), Cornelius et al. observed length of stay to be longer in paediatric patients with iHypoCa (serum calcium <9 mg/dl [2.25 mmol/l] compared to those with iNormoCa [33]. This study also observed an increased requirement for transfusion and for interventional radiology or operative management amongst patients with iHypoCa [33]. However, we excluded this study from this analysis as the authors used total calcium concentration, which may not reflect the physiologically active ionised calcium concentration [11]. Moreover, total calcium concentration have less utility in the ED, with values not being as quickly available as blood gas measurements [10]. The study also included low acuity isolated orthopaedic injuries, so the discrepancy between unwell paediatric patients with iHypoCa may be more pronounced [33]. The lack of difference between groups may reflect the practice of undertaking a conservative and observational approach to paediatric trauma [46, 47].
Although there was a trend towards increased odds of mortality for iHypoCa patients, this review did not observe any significant difference in mortality rates between paediatric patients with iHypoCa and iNormoCa. In contrast, studies in adult trauma patients looking at calcium concentration prior to blood transfusion have demonstrated an association between iHypoCa and mortality [14–16]. Additionally, hypercalcaemia (iCa ≥ 1.30 mmol/l) has also been associated with mortality and adverse outcomes in adult trauma patients, suggesting that there is a parabolic association between calcium levels and adverse outcomes [18].
Association with physiological abnormalities
This review did not demonstrate an association between iHypoCa and coagulopathy, pH, or hyperlactatemia. The presence of coagulopathy was heterogeneously explored between studies, and no studies used thromboelastography (TEG) or rotational thromboelastometry (ROTEM), thus limiting conclusions on the association between iHypoCa and coagulopathy [35–37]. In contrast, studies involving adult major trauma patients have observed an independent association between iHypoCa and coagulopathy (defined as a PTT ≥40 s or an INR ≥1.4) on arrival to the ED [17–19].
Acidosis is well-known to impair coagulation, and a positive correlation between iHypoCa and acidosis has previously been demonstrated post-injury [17, 48]. Mechanistically this seems paradoxical, as alkalosis leads to intracellular calcium shifts and iHypoCa, however, acidosis in trauma is largely due to ischaemia and reperfusion injury causing a raised lactate, which chelates calcium and causes pH-dependent calcium binding [6, 13, 17]. There is the potential for both acidosis and alkalosis to confound iCa levels; although a pH calcium correction formula can be used (with around 5% calcium level increase for each 0.1 pH unit decrease), this may risk underestimating iHypoCa and may be less relevant for interpretation in the clinical situation [11, 42]. In the studies included in this review, pH was similar amongst the iHypoCa and iNormoCa paediatric patients, with small standard deviations in individual studies and narrow confidence intervals in the meta-analysis. This suggests that few paediatric patients had severe acidosis. In contrast, acidosis is more prevalent amongst adult major trauma patients and has been associated with iHypoCa in this cohort [16, 18, 19].
It is also notable that none of the studies included in this review reported serum potassium concentrations, with raised concentrations being recognised as deleterious in traumatic coagulopathy it would be beneficial to quantify this in future studies [49].
Limitations
This review was primarily limited by the small number of available studies, with each study being a single-centre and retrospective design. Moreover, the pooling of data when different definitions of iHypoCa were used limits the true assessment of the incidence and implications of iHypoCa in paediatric trauma. However, the review has the advantage of being the first study to explore this area, and the methodology appropriately reflects the lack of a standardised definition of iHypoCa in current clinical practice.
The inclusion of Ciaraglia et al’s study which had a minority of patients who received citrated blood products prehospital prior to iCa measurement is potentially a significant confounder [35]. However, the prehospital transfusion volumes were small and similar with no statistically significant difference between the iHypoCa (34/66 (51.5%)) and iNormoCa (36/76 (47.4%)) groups, suggesting this did not impact upon admission iHypoCa rates or outcomes [35]. Moreover, these patients did not receive any exogenous calcium administration prior to calcium measurement. Therefore, the inclusion of these patients in meta-analysis is unlikely to be a significant confounder and is appropriate given the small number of studies available.
Additional limitations relate to the end points for meta-analysis. Given the low overall incidence of mortality, the study is likely underpowered to detect a mortality difference. Moreover, studies only report inpatient mortality, and as such, this may miss later morality in paediatric trauma [50]. Despite this, the study is novel and represents the largest sample size exploring iHypoCa outcomes in the paediatric population. There are also limitations to using pH as a marker of physiological derangement. pH is confounded by unmeasured metabolic and respiratory variables, and pH itself alters iCa levels [11, 51]. However, this was selected due to the availability of data, and has clinical relevance for paediatric trauma given that acidosis may further potentiate trauma-induced coagulopathy [6].
There is also a risk of selection bias, with all studies using trauma team activation as part of the inclusion criteria, since a reasonable percentage of injured paediatric patients are recognised to self-present (with their caregiver) to the ED this is likely to have excluded several seriously injured paediatric patients [52, 53].
Conclusion
Admission iHypoCa was present in at least one in six paediatric major trauma patients and may be associated with haemodynamic instability and increased blood transfusion requirements. An iHypoCa was not identified to have a statistically significant association with mortality.
Supporting information
S2 File. Excel spreadsheet featuring extracted data.
https://doi.org/10.1371/journal.pone.0303109.s003
(XLSX)
S4 File. Trauma team activation criteria for included studies.
https://doi.org/10.1371/journal.pone.0303109.s005
(PDF)
Acknowledgments
The authors would like to acknowledge and thank Catherine Hancox (Academic Librarian) and the Defence Medical Services Library Team for their assistance with the search strategy. Dr Peter Brooke is acknowledged for his review of the draft of this review, and Dr Mark Lyttle for his review of the protocol. The authors would also like to acknowledge and thank the authors of the included studies, in particular Dr Angelo Ciaraglia, Dr Danny Epstein, and Dr Nir Samuel, for their support and provision of data.
PROSPERO registration number: CRD42023425172
Protocol published in BMJ Open (doi:10.1136/bmjopen-2023-077429)
References
- 1. Wolfe I, MacFarlane A, Donkin A, et al. Why children die: death in infants, children and young people in the UK. Royal College of Paediatrics and Child Health. 2014. Accessed: October 23rd 2023: https://www.rcpch.ac.uk/sites/default/files/2017-06/Why%20children%20die%20-%20Death%20in%20infants%2C%20children%20and%20young%20people%20in%20the%20UK%20-%20Part%20A%202014-05.pdf.
- 2. Elbourne C, Cole E, Marsh S, et al. At risk child: a contemporary analysis of injured children in London and the South East of England: a prospective, multicentre cohort study. BMJ Paediatr Open. 2021;5(1):e001114. Published 2021 Nov 5. pmid:34796283
- 3. Davis JS, Satahoo SS, Butler FK, et al. An analysis of prehospital deaths. J Trauma Acute Care. 2014;77(2):213–8.
- 4. Moore EE, Moore HB, Kornblith LZ, et al. Trauma-induced coagulopathy [published correction appears in Nat Rev Dis Primers. 2022 Apr 22;8(1):25]. Nat Rev Dis Primers. 2021;7(1):30. Published 2021 Apr 29. pmid:33927200
- 5. Wray JP, Bridwell RE, Schauer SG, et al. The diamond of death: Hypocalcemia in trauma and resuscitation. Am J Emerg Med. 2021;41:104–109. pmid:33421674
- 6. DeBot M, Sauaia A, Schaid T, Moore EE. Trauma-induced hypocalcemia. Transfusion. 2022;62 Suppl 1:S274–S280. pmid:35748689
- 7. Ditzel RM Jr, Anderson JL, Eisenhart WJ, et al. A review of transfusion- and trauma-induced hypocalcemia: Is it time to change the lethal triad to the lethal diamond? J Trauma Acute Care Surg. 2020;88(3):434–439. pmid:31876689
- 8. Brohi K, Gruen RL, Holcomb JB. Why are bleeding trauma patients still dying? Intensive Care Med. 2019;45(5):709–711. pmid:30741331
- 9. Rossaint R, Afshari A, Bouillon B, et al. The European guideline on management of major bleeding and coagulopathy following trauma: sixth edition. Crit Care. 2023;27(1):80. Published 2023 Mar 1. pmid:36859355
- 10. Bilkovski RN, Cannon CM, Adhikari I, et al. Arterial and venous ionized calcium measurements: Is there a difference? Ann Emerg Med. 2004; 44 (4): S56.
- 11. Hamroun A, Pekar JD, Lionet A, et al. Ionized calcium: analytical challenges and clinical relevance. J Lab Precis Med. 2020; 5: 22.
- 12. Byerly S., Inaba K., Biswas S, et al. Transfusion-Related Hypocalcemia After Trauma. World journal of surgery. 2020;44(11): 3743–3750. pmid:32734451
- 13. Kronstedt S, Roberts N, Ditzel R, et al. Hypocalcemia as a predictor of mortality and transfusion. A scoping review of hypocalcemia in trauma and hemostatic resuscitation. Transfusion. 2022;62 Suppl 1:S158–S166. pmid:35748676
- 14. Cherry RA, Bradburn E, Carney DE, et al. Do early ionized calcium levels really matter in trauma patients? J Trauma. 2006;61(4):774–779. pmid:17033540
- 15. Magnotti LJ, Bradburn EH, Webb DL, et al. Admission ionized calcium levels predict the need for multiple transfusions: a prospective study of 591 critically ill trauma patients. J Trauma. 2011;70(2):391–397. pmid:21307739
- 16. Vasudeva M, Mathew JK, Fitzgerald MC, et al. Hypocalcaemia and traumatic coagulopathy: an observational analysis. Vox Sang. 2020;115(2):189–195. pmid:31845341
- 17. Vasudeva M, Mathew JK, Groombridge C, et al. Hypocalcemia in trauma patients: A systematic review. J Trauma Acute Care Surg. 2021;90(2):396–402. pmid:33196630
- 18. Helsloot D, Fitzgerald M, Lefering R, et al; TraumaRegister DGU. Trauma-induced disturbances in ionized calcium levels correlate parabolically with coagulopathy, transfusion, and mortality: a multicentre cohort analysis from the TraumaRegister DGU®. Crit Care. 2023;27(1):267. Published 2023 Jul 6. pmid:37415194
- 19. Vettorello M, Altomare M, Spota A, et al. Early Hypocalcemia in Severe Trauma: An Independent Risk Factor for Coagulopathy and Massive Transfusion. J Pers Med. 2022;13(1):63. Published 2022 Dec 28. pmid:36675724
- 20. Dias CR, Leite HP, Nogueira PC, Brunow de Carvalho W. Ionized hypocalcemia is an early event and is associated with organ dysfunction in children admitted to the intensive care unit. J Crit Care. 2013;28(5):810–815. pmid:23683566
- 21. Broner CW, Stidham GL, Westenkirchner DF, Tolley EA. Hypermagnesemia and hypocalcemia as predictors of high mortality in critically ill pediatric patients. Crit Care Med. 1990;18(9):921–928. pmid:2394115
- 22. Attard C, van der Straaten T, Karlaftis V, Monagle P, Ignjatovic V. Developmental hemostasis: age-specific differences in the levels of hemostatic proteins [published correction appears in J Thromb Haemost. 2019 Nov;17(11):1995]. J Thromb Haemost. 2013;11(10):1850–1854. pmid:23927605
- 23. Mullen S, Tolson A, Bouamra O, et al. Comparison of injury patterns and interventions between adolescent, adult and paediatric trauma cases: a cross-sectional review of TARN data. BMJ Open. 2023;13(5):e064101. Published 2023 May 9. pmid:37160391
- 24. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. Published 2021 Mar 29. pmid:33782057
- 25. Hibberd O, Price J, Harris T, et al. Incidence of admission ionised hypocalcaemia in paediatric major trauma: protocol for a systematic review and meta-analysis. BMJ Open. 2023 Nov 10;13(11):e077429. pmid:37949627
- 26. Sterne JA, Hernán MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355:i4919. Published 2016 Oct 12. pmid:27733354
- 27.
Schünemann HJ, Higgins JPT, Vist GE, et al. Chapter 14: Completing ‘Summary of findings’ tables and grading the certainty of the evidence. In: Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions version 6.4 (updated August 2023). Cochrane, 2023. Available from www.training.cochrane.org/handbook.
- 28. Linnaus ME, Notrica DM, Langlais CS, et al. Prospective validation of the shock index pediatric-adjusted (SIPA) in blunt liver and spleen trauma: An ATOMAC+ study. J Pediatr Surg. 2017;52(2):340–344. pmid:27717564
- 29. Nordin A, Coleman A, Shi J, et al. Validation of the age-adjusted shock index using pediatric trauma quality improvement program data [published online ahead of print, 2017 Oct 12]. J Pediatr Surg. 2017;S0022–3468(17)30645-0. pmid:29108845
- 30. Jackson D, Law M, Rucker G, Schwarzer G. The Hartung-Knapp modification for random-effects meta-analysis: A useful refinement but are there any residual concerns? Stat Med 2017;36(25):3923–3934. pmid:28748567
- 31. IntHout J, Ioannidis JP, Borm GF. The Hartung-Knapp-Sidik-Jonkman method for random effects meta-analysis is straightforward and considerably outperforms the standard DerSimonian-Laird method. BMC Med Res Methodol. 2014;14:25. Published 2014 Feb 18. pmid:24548571
- 32. Deeks JJ, Higgins JP, Altman DG. Analysing data and undertaking meta-analysis. In: Higgins J, Thomas J, Chandler J, et al., eds. Cochrane Handbook for Systematic Reviews of Interventions (version 6.3) Cochrane; 2022.
- 33. Cornelius BG, Clark D, Williams B, et al. A retrospective analysis of calcium levels in pediatric trauma patients. Int J Burns Trauma. 2021;11(3):267–274. Published 2021 Jun 15. pmid:34336394
- 34. Livingston MH, Singh S, Merritt NH. Massive transfusion in paediatric and adolescent trauma patients: incidence, patient profile, and outcomes prior to a massive transfusion protocol. Injury. 2014;45(9):1301–1306. pmid:24950797
- 35. Ciaraglia A, Lumbard D, Deschner B, et al. The Effects of Hypocalcemia in Severely Injured Pediatric Trauma Patients [published online ahead of print, 2023 Feb 13]. J Trauma Acute Care Surg. 2023;10.1097/TA.0000000000003902. pmid:36787433
- 36. Epstein D, Ben Lulu H, Raz A, Bahouth H. Admission hypocalcemia in pediatric major trauma patients-An uncommon phenomenon associated with an increased need for urgent blood transfusion. Transfusion. 2022;62(7):1341–1346. pmid:35638746
- 37. Gimelraikh Y, Berant R, Stein M, et al. Early Hypocalcemia in Pediatric Major Trauma: A Retrospective Cohort Study [published online ahead of print, 2022 Apr 13]. Pediatr Emerg Care. 2022;10.1097/PEC.0000000000002719. pmid:35413033
- 38. Giancarelli A, Birrer KL, Alban RF, Hobbs BP, Liu-DeRyke X. Hypocalcemia in trauma patients receiving massive transfusion. J Surg Res. 2016;202(1):182–187. pmid:27083965
- 39.
Goyal A, Anastasopoulou C, Ngu M, Singh S. Hypocalcemia. In: StatPearls. Treasure Island (FL): StatPearls Publishing; October 15, 2023.
- 40. Bloom BM, Grundlingh J, Bestwick JP, Harris T. The role of venous blood gas in the emergency department: a systematic review and meta-analysis. Eur J Emerg Med. 2014;21(2):81–88. pmid:23903783
- 41. Peng S, Lv K. The role of fibrinogen in traumatic brain injury: from molecular pathological mechanisms to clinical management. Eur J Trauma Emerg Surg. 2023;49(4):1665–1672. pmid:35972516
- 42. Imamoto T, Sawano M. Effect of ionized calcium level on short-term prognosis in severe multiple trauma patients: a clinical study. Trauma Surg Acute Care Open. 2023;8(1):e001083. Published 2023 Jun 30. pmid:37396952
- 43. Hsu CH, Couper K, Nix T, et al. Calcium during cardiac arrest: A systematic review. Resusc Plus. 2023;14:100379. Published 2023 Mar 27. pmid:37025978
- 44. Eisner DA, Caldwell JL, Kistamás K, Trafford AW. Calcium and Excitation-Contraction Coupling in the Heart. Circ Res. 2017;121(2):181–195. pmid:28684623
- 45. Amberg GC, Navedo MF. Calcium dynamics in vascular smooth muscle. Microcirculation. 2013;20(4):281–289. pmid:23384444
- 46. Henry S. ATLS 10th edition offers new insights into managing trauma patients. Bulletin of the American College of Surgeons. Am Coll Surg. 2018: 56. https://bulletin.facs.org/2018/06/atls-10th-edition-offers-new-insights-into-managing-traumapatients/#Chapter_3_Shock.
- 47. Committee on Pediatric Emergency Medicine, Council on Injury. Management of Pediatric Trauma. Pediatrics. 2016;138(2):e20161569. pmid:27456509
- 48. Vivien B, Langeron O, Morell E, et al. Early hypocalcemia in severe trauma. Crit Care Med. 2005;33(9):1946–1952. pmid:16148464
- 49. Wolf J, Geneen LJ, Meli A, et al. Hyperkalaemia Following Blood Transfusion-a Systematic Review Assessing Evidence and Risks. Transfus Med Rev. 2022;36(3):133–142. pmid:35534372
- 50. McLaughlin C, Zagory JA, Fenlon M, et al. Timing of mortality in pediatric trauma patients: A National Trauma Data Bank analysis. J Pediatr Surg. 2018;53(2):344–351. pmid:29111081
- 51. Wang S, McDonnell EH, Sedor FA, Toffaletti JG. pH effects on measurements of ionized calcium and ionized magnesium in blood. Arch Pathol Lab Med. 2002;126(8):947–950. pmid:12171493
- 52. Hibberd O, Price J, Laurent A, Agrawal S, Barnard EBG. Imaging appropriateness in paediatric trauma patients conveyed to a trauma unit compared to a major trauma centre following the introduction of trauma networks. Eur J Emerg Med. 2023;30(4):301–303. pmid:37387636
- 53. Hibberd O, Price J, Laurent A, Agrawal S, Barnard E. Paediatric Major Trauma: A Retrospective Observational Comparison of Mortality in Prehospital Bypass and Secondary Transfer in the East of England. Cureus. 2023;15(3):e36808. Published 2023 Mar 28. pmid:37123802