• 
• 

Three Part Question

In [blunt cardiac trauma patients], can [any factors] predict [prognosis]?

Clinical Scenario

A 40-year-old man attends the emergency department after a road traffic accident. Although haemodynamically stable, he has sustained a blunt chest injury and has bruising across his anterior chest. You are uncertain about its significance. You wonder if you can perform any tests to help guide your management.

Search Strategy

Medline/PubMed (1966 to 31/01/2019 (date of search)):
((blunt[Title/Abstract]) AND (cardiac[Title] OR myocardial[Title]) AND ((((injur*[Title]) OR trauma[Title] OR contusion[Title]) OR blunt cardiac injury[MeSH Terms]) OR blunt cardiac trauma[MeSH Terms])) AND ( ( Clinical Trial[ptyp] OR Meta-Analysis[ptyp] OR Review[ptyp] OR Clinical Study[ptyp] OR systematic[sb] ))

Embase (1974 to 2019 Week 04):
(blunt.ab. or blunt.ti.) and (trauma or injur* or contusion).ti. and (cardiac or myocardial).ti.
Manually add these limits for the 1st search:
(Clinical Trial OR Randomized Controlled Trial OR Controlled Clinical Trial OR Multicenter Study)
Manually add these limits for the 2nd search:
(Evidence Based Medicine OR Consensus Development OR Meta Analysis OR Outcomes Research OR Systematic Review)

Cochrane (1966 to 31/01/2019 (date of search)):
(blunt):ti,ab,kw, (cardiac):ti OR (myocardial):ti, (trauma):ti OR (injur*):ti OR (contusion):ti

Search Outcome

104 papers found, 5 of which directly relevant and of sufficient quality for inclusion. A further 6 papers were found by scanning the references of all relevant papers.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Burrell et al. 2017 Australia	n=42 major trauma patients; of these: n=21 had evidence of chest injury and elevated troponin I (cTnI), n=21 had major trauma but no evidence of chest injury (controls). All patients had an ECG taken on admission and underwent CMR within 7 days. Serum troponin was taken on admission, then repeated daily for 3 days or until stabilised. For comparison, transthoracic echocardiography was also performed on n=31/42. Abnormal tests defined as below: CMR: any regional wall motion abnormality, elevated regional T2 STIR or LGE if also associated with focal hypokinesis, mild or greater pericardial effusion, intramural haematoma, or new valvular lesion; Echo: no definition given; ECG: evidence of acute myocardial injury or conduction defects. Major adverse cardiac events (MACE) defined as any of these during hospital stay: ventricular arrhythmia, unexplained hypotension requiring inotropes, or requirement for cardiac surgery	Prospective cohort study (Level 2b)	Sensitivity and specificity of predictors for inpatient MACE, defined by abnormal test (see patient group column) (5/21 with MACE vs. 16/21 without MACE)	CMR: 60%, 81%; Echo: 25%, 77%; CTnI peak > 1000ng/L: 40%, 57%; ECG: 80%, 69%	Small cohort size. Extended time of testing (up to 7 days) - impractical in the initial triage assessment. Definition of abnormal ECG non-specific; no definition for abnormal echo
			Outcomes of patients with chest injury vs. no chest injury (cases vs controls)	Chest injury: 47% had sternal fractures, 29% had an abnormal CMR, 22% had an acute echo change; No chest injury: none suffered any of the above outcomes; Chest injury vs. no chest injury: ~3-fold stay in ICU
			Characteristics of patients with abnormal CMR vs. those with normal CMR	Abnormal ECG: 67% of those with abnormal CMR vs. 33% of those with normal CMR; Abnormal echo: 50% of those with abnormal CMR vs. 8% of those with normal CMR; Mean peak troponin 4400ng/L (abnormal CMR) vs 2100ng/L (normal CMR)
Kalbitz et al. 2017 Germany	n=173 severely injured patients with injury severity score (ISS) of 25+. Serum cTnT/cTnI determined on admission. cTnT ≥ 15ng/L and cTnI ≥ 70ng/L defined as elevated by clinical routine diagnostic lab (Roche assay)	Retrospective observational study (Level 2b)	cTnT as an indicator of thoracic injury	cTnT levels in patients with severe thoracic injury (AIS3 and 4) were significantly increased (p = 0.008) compared to levels in patients without thoracic injury	cTnT 'cut-off' concentraion (9ng/L) not useful - far below the precisional limit of the point-of-care assay (which had a cut-off for abnormal results of 15ng/L). No correction was made for the significant difference (p = 0.004) found in cTnT levels, in patients who received norepinephrine on admission vs. patients who were not exposed to exogenous catecholamines
			Prognostic relevance of cTnT for mortality after trauma	Patients with ISS ≥25 who survived (n = 125) an observation period of 25 days after trauma had significantly lower (p = 0.02) cTnT levels compared to patients who died (n = 48). At a cut-off concentration of 9 ng/L, sensitivity for mortality was 88% and specificity was 50%
Amino et al. 2009 Japan	n=11 blunt chest trauma patients who were diagnosed with cardiac injury by ECG/myocardial enzyme/transthoracic echo (TTE), agreed to have regular examinations for 1+ year post-hospital admission. Patients assessed every 3 months by ECG, cardiac enzyme tests, TTE, and 24hr-Holter. Nuclear medicine study of the heart (NMSH) performed in the acute (<1 month) and chronic (>12 months) stages of the injury. Positive finding on NMSH defined as: reduced coronary blood perfusion (201Tl imaging), myocardial fatty acid metabolism (123I-BMIPP), or cardiac sympathetic nerve distribution (123I-MIBG)	Prospective cohort study (Level 2b)	Pericardial effusion on initial presentation (CT-diagnosed) as a predictor of long-term prognosis	The extent of effusion (seen in 4 patients) did not reflect the presence of (NMSH-diagnosed) damage	Small cohort size. No predictors for acute outcome. No standardised diagnosis of BCI. No definition for positive troponin. No inclusion of blunt chest trauma patients who were not diagnosed with cardiac injury, who are still at risk of post-admission cardiac events
			Proportion of patients with abnormal results after 12 months	ECG: 4/11 (RBBB, ventricular premature contraction, depolarisation abnormality, repolarisation abnormality) Holter ECG: 6/12 (supraventricular premature contraction, ventricular premature contraction, nonsustained VT) cTnI: 0/11 TTE: 1/11 (ventricular aneurysm - required treatment)
			Proportion of patients with NMSH positive findings in acute period	6/11
			Of the above, the number showing persistent abnormality on NMSH at 12 months	6/6
			Mortality rate after 12 months	0/11
Edouard et al. 2004 France	n=728. cTnI taken on admission, 6h, 12h. cTnI cut-off: 400ng/L (limit of Bayer assay detection = 40ng/L) for significant myocardial injury. According to the time course of cTnI, 3 groups of patients were defined: very transient (< 12hr) and limited release (cTnI < 2000ng/L), transient (< 36hr) and significant release (cTnI > 2000ng/L), and sustained (> 36hr) and significant release (cTnI >2000ng/L). Myocardial contusion diagnosed by significant ECG signs (ST segment deviation, abnormal T wave, tachyarrhythmias, and cardiac conduction defects)	Prospective cohort study (Level 2b)	Diagnostic yield of elevated cTnI for myocardial contusion (by ECG criteria)	Sensitivity: 63%, Specificity: 98%, PPV: 40%, NPV: 98%	Abnormal cTnI cut-off high (old assay), may explain false negatives and lack of perceived benefit. By using ECG criteria to define myocardial contusion, there is a risk of underdiagnosis - the standard for diagnosis of myocardial contusion (based on ECG criteria) were shown to be less sensitive than cTnI in the paper itself, questioning the validity of ECG diagnosis (32/37 (86.4%) of patients with transient rise and fall in Tn had normal ECGs)
			Correlation between abnormal cTnI and ECG abnormalities in early survivors	Of 37 patients with transient cTnI release, only 13% had ECG abnormalities, compared to 100% of those with prolonged cTnI rise
			Proportion of early survivors (n=54, alive at 48hrs post-trauma) diagnosed with myocardial contusion during their hospital admission, stratified by cTnI release	Very transient and limited cTnI release: 13% (3/22), Transient and significant cTnI release: 13% (2/15), Prolonged and significant cTnI release: 100% (17/17)*; * indicates p<0.05 vs. other 2 groups of patients
			cTnI as a predictor of mortality	Mortality was similar in early survivors with or without cTnI release (15% vs. 12%). Mortality was similar in early survivors with or without cTnI release (15% vs. 12%). Odds ratio for late mortality (>48hr after trauma) was 1.32 in patients with cTnI release
Rajan and Zellweger 2004 Australia	n=187 with blunt chest trauma. Serial measurements of cTnI, CK, CK-MB: admission, 6h, 12h, 18h, 24h, daily for 3d. Patients underwent sequential ECGs and echos. cTnI assay cut-off value was 350ng/L. After hospital discharge, patients were observed for 3.3 months (median)	Prospective cohort study (Level 2b)	Prognostic yield of cTnI	Patients with cTnI <1050 ng/L at admission and during the first 6hrs showed no cardiac abnormalities throughout entire study period based on ECG and echo. Increasing cTnI levels correlated with an increasing risk of arrhythmia (alpha = 1.791; p = 0.008) and a temporary decrease in LV function (alpha = 0.523, p = 0.005)	Excluded patients with pre-existing cardiac disease. Time elapsed from event causing injury to admission is unclear and this will have a bearing on blood tests done on admission
			Cut-off time for peak cTnI level	No positive conversion in cTnI levels occurred after 6hrs
Velmahos et al. 2003 USA	n=333 consecutive patients with significant blunt thoracic trauma. Serial ECG and TnI tests performed: at admission, and at 4 and 8 hours post-admission. Definitions: Clinically significant BCI (SigBCI) - presence of cardiogenic shock, arrhythmias requiring treatment, or posttraumatic structural defects (as seen on echo); Elevated troponin taken as >1500ng/L	Prospective cohort study (Level 2b)	Positive and negative predictive values for SigBCI (respectively)	ECG: 29%, 98%; TnI: 21%, 94%; Combination of ECG and TnI: 34%, 100%. The admission ECG or TnI was abnormal in 43 of 44 patients with SigBCI	Abnormal troponin cut-off high, may explain false negatives. By protocol, SigBCI was defined before Tn or ECG results were available
Mori et al. 2001 Italy	n=32 with clinical and/or radiological signs of blunt chest trauma. Abnormal TEE defined as wall motion abnormalities (i.e. cardiac contusion) with full resolution at follow-up; patients imaged within 24hrs of injury. Serial cTnI measurements at admission, 6hr, 12hr, 24hr, 48hr, 96hr. Normal cTnI taken as <400ng/L; minimal cTnI concentration detectable by assay = 350ng/L	Prospective cohort study (Level 2b)	Sensitivity of raised cTnI and CK-MB	Mean cTnI levels significantly higher in those with abnormal echo (2600+/-1600 vs. 600+/-1400 ng/L, p < 0.001), whereas difference in CK-MB levels was not statistically significant (p=0.77)	Small cohort size. Excluded patients with pre-existing cardiac disease. Disproportionately high number in the study had TEE, which does not apply to routine ED practice. The finding of normal echo with abnormal Tn raises the question of diagnosing contusion with echo
			Cut-off cTnI level for myocardial contusion	15/15 patients with normal cTnI had a normal echo. 7/7 patients with 400ng/L < cTnI levels < 1000ng/L had a normal echo. 6/10 patients with cTnI levels > 1000ng/L had an echo positive for myocardial contusion
			Mean cTnI levels over time	Mean cTnI levels peaked after 12hrs, whereas the increase disappeared after 48hrs
			Feasibility of echo	TEE feasibility (97%) significantly higher than that of TTE (67%) (p<0.002), and TEE is more accurate than TTE in identifying wall motion abnormalities (in n=4, it was detected only by TEE but not TTE)
			Outcomes	Of the myocardial contusion patients, echo abnormalities resolved in 3/6, 2/6 died during hospital stay (cerebral lesions), and LV wall motion abnormalities remained unchanged in 1/6
Collins et al. 2001 USA	n=66 with suspected blunt cardiac injury (not including patients who were hypotensive (SBP<90mmHg) or intubated). Underwent ECG on admission, and serum cTnI, CPK, and isoenzymes on admission and 4-6hrs post-injury. Clinically significant BCI (SigBCI) defined as arrhythmias requiring treatment, depressed cardiac function, or structural abnormality noted on ECG. Patients discharged on physician's clinical judgement	Prospective cohort study (Level 2b)	Prognostic value of normal cTnI and ECG on admission	n=40 had normal cTnI and ECG on admission: all 40/40 remained haemodynamically stable w/o any cardiac complications during admission, and were discharged the next day	Small cohort size, low prevalence of SigBCI. No explanation of 'structural abnormality on ECG', and no mention of whether this finding was compared to a previous ECG as a reference (if available). No definition of 'cardiac complications'
			Prognostic value of 4-6hr post-injury cTnI	n=16 had abnormal ECGs on admission, but normal cTnI 4-6hr post-injury: all 16/16 remained haemodynamically stable, with no arrhythmias requiring treatment, and no cardiac complications during admission. n=10 patients had elevated cTnI 4-6hr post-injury: 2/10 suffered a SigBCI (inc. 1 mortality)
			Sensitivity, PPV, and NPV (respectively) of 4-6hr post-injury cTnI	100%, 20%, 100%
			Correlation between cTnI rise and ISS scores	Patients with cTn rise vs. without had mean ISS scores of 15 vs. 4
Fulda et al. 1997 USA	n=71 with chest wall injuries not requiring ICU admission. Serial cTnT and CK-MB tested at admission, 6h, 12h, 18h, 24h, 48h. Signal-averaged ECGs obtained on admission, then daily for min. 2 days after admission. Definitions: Elevated cTnT: >200ng/L; Elevated CK-MB: >4%; Abnormal echo: wall motion abnormalities; Significant ECG: signal-averaged QRS complex > 114ms, RMS40 < 20mV, or LAS40 > 38ms with 40Hz high-pass filter; Clinically significant cardiac abnormalities = 1+ of the following: SVT, junctional tachycardia, AFib, AFlutter, sinus arrest, VT, VF, BBB, or acute ischaemic change	Prospective cohort study (Level 2b)	Significant predictors for patients at risk of clinically significant abnormalities	Significantly abnormal initial ECG and elevated initial cTnT (p<0.05)	Excluded patients with pre-existing cardiac disease. Excluded patients who were discharged within 48hrs, from the study (with no apparent reason). All patients chosen to have an echo were considered by protocol or attending physician to have blunt myocardial injury, which will have skewed the predictive factor statistics. The definition of clinically significant cardiac abnormalities was limited to arrhythmias and ischaemic ECG changes, rather than prognostic outcome. The objective of the study was to predict rhythm disturbances rather than BCI
			Sensitivity, specificity, PPV, and NPV of tests in predicting clinically significant abnormalities	Abnormal echo: 0.12, 0.98; 0.75, 0.66; Significant ECG: 0.38, 0.93, 0.77, 0.72; Abnormal CPK-MB: 0.12, 0.96, 0.6, 0.65; Abnormal cTnT: 0.27, 0.91, 0.64, 0.68
Maenza et al. 1996 USA	Patients with suspected blunt cardiac trauma; n=2,210 in 25 prospective studies, n=2,471 in 16 retrospective studies. Definitions for abnormal tests: ECG - new ECG abnormality, or any abnormality in patients with no previous ECG for comparison; CPK-MB - any elevation of MB fraction greater than the level considered normal by the authors of the article (range: 0-7%); Scan - any radionuclide scan reported as abnormal by authors; Echo - any echo reported as abnormal by authors. Definition of complication: any new cardiac problem that required treatment	Meta-analysis of 41 studies (Level 2a)	Odds ratio of developing a complication with abnormal test result. * = significant	Prospective: ECG: 3.23* (1.71-6.10); CPK-MB: 21.2* (5.89-76.9); Scan: 1.38 (0.16-3.25); Echo: not significant (CI contains 1). Retrospective: ECG: 26.0* (18.51-36.52); CPK-MB: 7.7* (5.21-11.38); Scan: 0.79 (0.07-23.8); Echo: 84.0* (32.77-215.34*). Combined: ECG: 19.19* (1.92-25.77); CPK-MB: 6.6* (2.14-9.27); Scan: 1.08 (0.88-1.38); Echo: 31.11* (1.50-62.50)	Inconsistent definitions of 'complications' between studies (with inclusion of 'soft complications' in some studies, where therapy may not be necessary) leads to conservative OR calculations. Inconsistent definitions of 'abnormal test' between studies, no mention of transience/persistence of abnormality
			Negative predictive value of tests	Normal ECG and CPK-MB correlated with the lack of clinically significant complications
Bu'Lock et al. 1994 United Kingdom	n=60 patients with blunt central chest trauma or multiple injuries involving the torso. Patients grouped by Chest Injury Score (CIS) and ISS. Group A1 (n=36): CIS 2-3, ISS<10, seatbelt-related trauma; Group A2 (n=11): CIS 2-3, ISS<10, direct sternal injuries; Group B (n=13): CIS 4-5, ISS>10. ECG, echo, and cardiac enzyme estimation performed within 12hr of admission. ECG and enzyme estimations repeated daily for 3 days, echo repeated at end of the 3 days. Myocardial injury on echo suggested by abnormal myocardial echogenicity. 8/11 patients with pericardial effusion were reassessed clinically and by ECG and echo after 1 year (2 died in hospital, 1 lost to follow-up)	Prospective cohort study (Level 2b)	Hospital mortality rate	Group A1: 1/36, Group A2: 1/11, Group B: 4/13	Small cohort size. Complete cardiac enzyme data only available for 36/60 patients. No indication as to how echo results (i.e. outcome) relate to ECG/cardiac enzyme data (due to different-sized, but overlapping groups) - so no correlation between test results and outcome can be calculated. This is a study to specifically evaluate the incidence and outcome of pericardial effusions, and not other types of cardiac injury
			Proportion of group who developed pericardial effusion (as shown on echo)	Group A1: 25% (but none required drainage, and none had evidence of haemodynamic compromise/evidence of myocardial injury), Group A2: 0%, Group B: 30% (2/3rds of these had evidence of significant cardiac injury)
			Outcome of 8/9 surviving patients with pericardial effusion	All 8 patients were well, with no residual effusion or evidence of an echo abnormality, nor signs of heart failure/constrictive pericarditis

Comment(s)

We have reviewed 10 original research studies and 1 meta-analysis in this BestBETs review. The role of cardiac biomarkers: Most studies concur that normal cardiac troponin (cTn) serum concentrations (particularly cTnI) strongly predict good clinical outcomes. Many indicate that cTnI >1000 ng/L is required to predict clinically significant blunt cardiac injury (BCI). However this observation should be interpreted with caution considering that the threshold value for an elevated cTn in the majority of these studies is more than one order of magnitude higher than current assays, which have detection thresholds ranging from 5-40ng/L. Some studies suggest that if cTn remains below the assay detection limit 6 hours post-admission, then the likelihood of adverse outcome is very low. In current practice, additional modalities such as bedside imaging may be useful adjuncts within this 6 hour time period. It is also important to note that the definition of an adverse cardiac event varies in these studies and include arrhythmias, mechanical complications and transient depression of cardiac function (this may include stress cardiomyopathy although the studies reviewed do no specify this). Due to this variability in definition, ECG and cTn levels demonstrated poor predictive value. A low injury severity score (ISS) appears to be associated with normal cTn. Scores in the moderate or higher range were more likely to be associated with a high cTn level. This is useful in the acute assessment of blunt chest trauma. Sternal fractures did not show a consistent relationship with elevated cTn levels. As with any other acute event associated with myocardial injury (defined as cTn levels above the 99th percentile of a normal population), an appropriately timed cTn assay remains a useful predictive tool for adverse outcome in BCI. The role of the resting 12-lead ECG: Any new abnormalities on ECG (including those of ischaemic or arrhythmic in nature) would be an indication for extended cardiac monitoring. However, a normal ECG in isolation does not exclude clinically significant BCI. Serial ECGs are helpful during the monitoring period. The role of the echocardiogram: Some studies indicate that if cTn 6 hours post-admission is normal and stable, echo does not provide any additional benefit. However, echo is extremely helpful in acutely unstable patients for diagnosis, risk-stratification and to identify patient who need immediate assessment by cardiothoracic surgeons. The role of other imaging modalities: Myocardial perfusion imaging by radioisotope tracer uptake is restricted by radiation exposure and availability in an acute setting. However, the role of cardiac MRI is likely to expand and may specifically have a role in diagnosing contained myocardial rupture and late sequelae of BCI.

Editor Comment

JB

Clinical Bottom Line

A low injury severity score appears to correlate with a low risk of BCI. A normal troponin level is an excellent predictor for good prognosis and when combined with a normal ECG, the two reach near 100% negative predictive value for clinically significant BCI. Timing and serial assays are important for both. Any patient who presents within 6 hours of BCI should not be sent home without serial Tn assays and ECGs. If both are abnormal, the likelihood of an adverse cardiac event increases, although there is no definitive protocol for risk-stratification due to inconsistent criteria for diagnosis of clinically significant BCI and the variability in practice in the reported studies.

References

1. Burrell, A., Kaye, D., Fitzgerald, M., Cooper, D., Hare, J., Costello, B. and Taylor, A Cardiac magnetic resonance imaging in suspected blunt cardiac injury: A prospective, pilot, cohort study Injury 2017, 48(5), pp.1013-1019
2. Kalbitz, M., Pressmar, J., Stecher, J., Weber, B., Weiss, M., Schwarz, S., Miltner, E., Gebhard, F. and Huber-Lang, M The Role of Troponin in Blunt Cardiac Injury After Multiple Trauma in Humans World Journal of Surgery 2017, 41(1), pp.162-169
3. Amino, M., Yoshioka, K., Morita, S., Iizuka, S., Otsuka, H., Yamamoto, R., Aoki, H., Aizawa, T., Ikari, Y., Nasu, S., Hatakeyama, K., Iino, M., Kodama, I., Inokuchi, S. and Tanabe, T One-Year Follow-Up and Convalescence Evaluated by Nuclear Medicine Studies and 24-Hour Holter Electrocardiogram in 11 Patients With Myocardial Injury Due to a Blunt Chest Trauma The Journal of Trauma: Injury, Infection, and Critical Care 2009, 66(5), pp.1308-1310
4. Edouard, A., Felten, M., Hebert, J., Cosson, C., Martin, L. and Benhamou, D Incidence and Significance of Cardiac Troponin I Release in Severe Trauma Patients Anesthesiology 2004, 101, pp.1262-1268
5. Rajan, G. and Zellweger, R. Cardiac Troponin I as a Predictor of Arrhythmia and Ventricular Dysfunction in Trauma Patients With Myocardial Contusion The Journal of Trauma: Injury, Infection, and Critical Care 2004, 57(4), pp.801-808
6. Velmahos, G., Karaiskakis, M., Salim, A., Toutouzas, K., Murray, J., Asensio, J. and Demetriades, D Normal Electrocardiography and Serum Troponin I Levels Preclude the Presence of Clinically Significant Blunt Cardiac Injury The Journal of Trauma: Injury, Infection, and Critical Care 2003, 54(1), pp.45-51
7. Mori, F., Zuppiroli, A., Ognibene, A., Favilli, S., Galeota, G., Peris, A., De Saint Pierre, G. and Dolara, A. Cardiac contusion in blunt chest trauma: a combined study of transesophageal echocardiography and cardiac troponin I determination Italian Heart Journal 2001, 2(3), pp.222-227
8. Collins, J., Cole, F., Weireter, L., Riblet, J. and Britt, L The usefulness of serum troponin levels in evaluating cardiac injury The American Surgeon 2001, 67(9), pp.821-5
9. Fulda, G., Giberson, F., Hailstone, D., Law, A. and Stillabower, M An Evaluation of Serum Troponin T and Signal-Averaged Electrocardiography in Predicting Electrocardiographic Abnormalities after Blunt Chest Trauma The Journal of Trauma: Injury, Infection, and Critical Care 1997, 43(2), pp.304-312
10. Maenza, R., Seaberg, D. and D'Amico, F A meta-analysis of blunt cardiac trauma: Ending myocardial confusion The American Journal of Emergency Medicine 1996, 14(3), pp.237-241
11. Bu Lock, F., Prothero, A., Shaw, C., Parry, A., Dodds, C., Keenan, J. and Forfar, J Cardiac involvement in seatbelt-related and direct sternal trauma: a prospective study and management implications European Heart Journal 1994, 15(12), pp.1621-1627