• 
• 

Three Part Question

For [paediatric trauma patients who have sustained mangled extreity type injuries], does the [use of published scoring tools attempting to predict success of attempted limb salvage, including PSI, LSI, NISSSA, GHSS or MESS], provide [statistically superior prognostic prediction of successful limb salvage]?

Clinical Scenario

A 10 year old child has been transferred to hospital after being involved in a pedestrian versus vehicle collision. They have sustained a lower limb injury which demonstrates significant soft tissue loss with a comminuted, open fracture of tibia/ fibula. Their pedal pulses are weak with dusky appearanes of the extremity of the affected lower limb

Search Strategy

The search-term strategy below was applied in PubMed, SCOPUS, Cochrane Library and Web of Science publication databases:

"(outcome OR assessment OR treatment outcome OR recovery of function[MeSH Terms] OR "functional status") AND limb salvage[MeSH Terms] AND ("upper extremity" OR "upper limb" OR "arm OR "forearm" OR "lower extremity" PR "lower limb" OR "Leg" OR "Foot" OR "Ankle") AND Trauma[MeSH Terms) AND ("Amputation"[Majr] OR "Amputee"[Majr]) AND ("children"[MeSH terms] OR "[aediatric"[MeSH terms] OR "adolescent"[MeSH terms]) AND ("Mangeled Extremity Severity Score" OR "Gangha Hospital Severity Score" OR "Limb salvage index" OR "Predictive salvage index" OR "Nerve Injury, ischaemia, soft-tissue injur, skeletal injury, shock and age score")"
Inclusion criteria: paediatric populations aged 0-18 years, upper and lower limb mangled extremities, polytrauma patients civilian and military studies. Minimum sample size of 10
Exclusion criteria: Studies not written in english, non-traumatic limb salvage/ amputation, adult studies, unclear documentation of management plan or patient outcome

Search Outcome

580 studie imported for screening, 115 duplicates removed. 465 studies screened, 362 studies found to be irrelevant. 103 full-text studies assessed for eligibility, 96 studies excluded. 7 studies selected. From the Selected Papers, 6 reported on MESS scores, 2 reported on GHSS, and only 1 reported on LSI/ PSI/ NISSSA.
Two Selected Papers were prospective cohort studies, with the remainder being retrospective, accessing data from local hospital or trauma network databases. Geographical locations conducting the studies included: 1 British, 1 Pakistani, 1 German, 1 Irani, 1 Australian, 2 North American and 1 Indian.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Behdad S; Rafiei MH; Taheri H; Mohammadzadeh M; Kiani G; Hosseinpour M December 2012 USA	Children with Grade IIIB and III C open fractures secondary to trauma, n= 200	Retrospective analysis comparing criteria of the MESS predictive index in mangled limb extremity trauma patients to attempt to validate use of MESS score, comparing two subsets within the cohort who underwent limb salvage versus amputation	Prospective cohort study with data collected from September 2009 to 2010, comprising 200 children who had sustained GAIIIB/IIIC femoral, tibial, or multiple injuries. comparing criteria of the MESS predictive index in mangled limb extremity trauma patients to attempt to validate use of MESS score, comparing two subsets within the cohort who underwent limb salvage versus amputation. Compared with degree of systemic shock, limb ischaemia, and degree of injury for predictability of limb salvage. MESS > or = 7 indicative of high risk of non-successful limb salvage. P Values 0.05 used for assessment of statistical significance in analysis between	The authors noted an amputation rate of 7.5% (n=15). Mean MESS value was 7.5 +/- 1.59 for the amputation group, versus 6.4 +/- 2.02(table 7). It was not possible to calculate positive or negative predictive values relating to a MESS > or < 7 with available data, however, it was possible to generate the ROC graph below with further calculation of AUC = 0.738	Although primary amputation excluded, limited detail on management protocol other than describing clinical judgement from paediatric surgeon. No description of infrastructure / equipment / transfer time / post-op protocol /rehab regimen / clinical decision pathway for delayed amputation. No report of patients who underwent multiple surgeries or how decision for delayed amputation was made. No clear description of management protocols. Moderate risk of bias
Fagelman MF Mar-Apr 2002 USA	36 paediatric patients who had undergone either primary amputation, or attempted limb. This assessed mangled extremity injuries of lower limbs. This group contained nine Gusto Anderson III B (“GAIIIB”) femur fractures, two Gusto Anderson III C (“GAIIIC”) femur fractures, nine GAIIIB tibia fractures and eight GAIIIC tibia fractures (total n=28). ). Similar data was not available for the amputation group (n=8)	To determine the efficacy of the use of MESS score in preduction of successful limb salvage in paediatric trauma patients.	Measured if calculated MESS score correctly predicted successful limb salvage and/ or if high scores were associated with limb amputation	amputation, MESS values predicted limb salvage in 37.5% of patients (n=3). For MESS score threshold <7 or >7, analysis found sensitivity of 92.9%, specificity 62.5%, a positive predictive value of 89.7% and a negative predictive value of 71.4%(table 8). Limb salvage failures were due to failed revascularisation (n=1) and post-operative infection (n=1). Incorrect predictions in the primary amputation group were attributed to transection of tibial nerve in a GAIIIC tibial injury (n=1), plus one patient who sustained bilateral GAIIIC tibial fractures whose simultaneous salvage and reconstructions were deemed non-survivable (n=2). Complication rates for patients who had undergone correctly predicted limb salvage or primary amputation were not provided.	No detailed surgical management plans for all patients, but circumstantial descriptions for cases where MESS score incorrectly predicted salvage. No detailed protocol. No description of hospital(s) infrastructure or resources available – specifically, no description of differences between two centres. No report for all patients who underwent multiple surgeries or how decision for delayed amputation was made. Exceptions made to describe clinical decision making for cases of incorrect MESS prediction towards salvage. Did compare initial and definitive limb salvage rates for MESS scores less than plus =/ greater than 7
Messner et al July 2020 Netherlands	data collected between 2013-2018, assessed 32 paediatric patients who sustained open fractures of the lower limb (29 tibia fractures, 1 femur fracture, 1 talus fracture, 1 open ankle fracture) of GAIIIB/C severity	examined the management and outcome of patients suffering complex paediatric lower limb injuries with bone and soft tissue loss	Measured GHSS score prognostic value, with further evaluation of fracture complexity, surgical techniques, time to surgery, QOL scores, union and complication rates. Comparator GHSS = / > 14, limb salvage vs amputation	GHSS values ranged from 6-13 which, in keeping with a threshold value of 14, correctly predicted limb salvage in all patients (no primary amputation or delayed amputations described). This paper therefore demonstrated GHSS had sensitivity and specificity of 100% for their patient cohort. The study did, however, provide data on complications sustained by patients. Whilst 0% developed compartment syndrome, 15.625% (n=5) of patients required revision of soft tissue flap closure for poor arterial in-flow (n=1), haematoma formation (n=1), partial necrosis (n=1) and microvascular complications (n=2(table 5). No patients developed ‘deep’ infections, however 27% of patients treated with external fixation (n=6) required oral antibiotics for pin site infections. Eight patients required corticotomy with planned distraction osteogenesis to correct leg length discrepancies. Four patients had injury-related physical growth arrest and had epiphysiodesis performed to avoid angular deformity.	No specific management of antimicrobials/ blood products/ specific physio regimen. No specific description of hospital infrastructure or nature of transfer from other units. Patient cohort exclusively comprised of successful limb salvage patients
	44 paediatric patients treated for traumatic extremity arterial lesions in level 1 Trauma Centre between 1971 and 2006	Determination of prognostic value of MESS, review of epidemiology/diagnotics/treatment	Determination of prognostic value of MESS, review of epidemiology/diagnotics/treatment, comparator MESS equal to or greater than 7	This assessed paediatric patients who had sustained arterial injuries of the upper and/or lower limbs. 6.8% of patients had undergone primary amputation (n=3), with 11.4% of all patients undergoing delayed amputation (n=5). All patients undergoing amputation reported a MESS value > 7. 9.1% of patients (n=4), however, had undergone successful limb salvage despite having MESS values >7. The authors found that all patients who had undergone amputation had sustained lower limb injuries, plus all patients who had MESS values <7 had undergone successful limb salvage. MESS values differed significantly between upper and lower limb injuries (upper limb 3.3+/-1.4 versus lower limb 6.1+/-2.6), with 38.7% of observed arterial injuries occurring in the upper limb versus 61.3% in the lower limb. MESS values were significantly greater for injuries resulting from blunt trauma and “multiple” trauma (5.5+/-2.3 and 7.0+/-2.4 respectively) versus penetration lesions or supracondylar distal humeral fractures (3.5+/-2.0 and 3.7+/-1.6 respectively). The mortality rate was 6.8% (n=3), all of whom had sustained polytraumatic injuries. The authors reported a 18.2% post-operative complications rate, including a 13.6% rate of vascular thrombosis or stenosis and a 4.5% rate of secondary peripheral ulceration. A MESS prognostic threshold value of 7 provides 100% specificity & sensitivity for upper limb injuries however 79% sensitivity plus 100% sensitivity for lower limb injuries. This would provide, for lower limb injuries, a Positive Predictive Value (“PPV”) of 100% plus a Negative Predictive Value (“NPV”) of 66%	Implicit use of energy of trauma for MESS scores however no clear descriptions. Further description of pattern and type of vascular injuries secondary to penetrating and blunt. No clear sub-analysis to describe if certain types of repair or additional procedures such as fasciotomy were more associated with limb amputation. Infrastructure not described / equipment described / transfer time not described / post-op protocol not described /rehab regimen, clinical decision pathway for delayed amputation both not described. No report of patients who underwent multiple surgeries or how decision for delayed amputation was made. Did compare initial and definitive limb salvage rates for MESS scores less than plus =/ greater than 7. treatment protocols were performed individually in every patient depending on age, severity of vascular lesions and accompanying injuries – no clear standardization, described lack of guidelines.
ZURehman et al June 2020 Pakistan	75 (67 males and 8 females) with peripheral arterial injuries 2008-2018	Observational study including analysis of NESS and revised trauma score	Measured probability of successful prediciton of limb salvage (successful vascular repair) relative to revised trauma score and MESS scores (= / > than 7)	75 paediatric patients who had sustained peripheral arterial injuries of the upper and/or lower limb (56% and 44% respectively). 85.3% of patients had sustained associated injuries, with 60% having sustained “complex” long bone fractures and/or nerve injuries. The authors reported a 4% primary amputation rate from their cohort, whose mean MESS value was 7.7. For patients whose MESS value was <7 (n=64) there was a secondary amputation rate of 6.2%(table 7), whereas patients whose MESS value was >7 (n=4) demonstrated a secondary amputation rate of 25%(table 6). The authors concluded there was no statistical significance to indicate that greater MESS value was associated with greater risk of limb loss (p value 0.163). For MESS threshold value 7, the data generates a sensitivity of 95%, a specificity of 50%, a PPV of 94% and a NPV of 57%. The authors reported a cohort mortality rate of 2.67%, where both patients had sustained multiple injuries including severe head injuries. There was no description of specific complications rates post-operatively for this patient cohort.	Despite detailed management protocol/ plan, no specific description of course of care for amputees beyond decision for primary amputation and “Failed revascularisation”. Described as occurring in developing country. Description of general template for clinical examination and Angiography with further description of surgical technique. No specific description of hospital infrastructure
Stewart et al Oct 2010 Australia	26 children selected, 24 studied (-2 insufficient data).	Retrospective cohort study. Single centre study in Australia. Study focused on Successful prediction of successful limb salvage vs amputation. data collected between 2000-2010, 24 paediatric trauma patients who sustained mangled lower extremity injuries (GAIIIA/ GAIIIB/ GAIIIC plus dysvascular injuries included	Comparison between MESS, LSI, PSI, NISSSA, HFS-98 in sensitivity & specificity. MESS 7, NISSSA 11, LSI 6, HFS-98 11, PSI 8	The authors found a 12.5% primary amputation rate, with no reported delayed amputations. For MESS threshold value 7, a sensitivity of 67%, Specificity rate= 86%, PPV= 40% and a NPV= 94.7%. For LSI threshold value 6, the paper found a sensitivity of 67%, specificity 81%, PPV 33.3% and NPV 94.4%. For PSI threshold value 8, the paper found a sensitivity of 100%, specificity 90%, PPV 60% and NPV 100%. For NISSSA threshold value 11, the paper found a sensitivity value of 67%, specificity 81%, PPV 33.3% and NPV 100%. With sub-analysis for tibial trauma patient only, all scores demonstrated 100% sensitivity with MESS & PSI demonstrating 87% specificity and LSI & NISSSA demonstrating 81% specificity(table 8). There were no documented mortalities or co-morbidities	No specific details of management plan or protocols, types of surgery performed or necessitated. Paired descriptions of mechanism plus site of injury with ages with outcome. No description of infrastructure, diagnostics, theatre teams. No specific description of teams involved, however authors include members of orthopaedics and plastics/ max-fax. No specific description of how interventions may have varied between these two groups (amputation/ limb salvage)
Venkatadass et al	data collected between January 2008 – March 2015, 52 paediatric patients who had sustained mangled lower extremity traumatic injuries	Retrospective cohort study. hildren (0-18 years) who were admitted with Open type IIIB injuries of lower limbs between January 2008 and March 2015 were included. MESS and GHOISS were calculated for all the patients. There were 50 children with 52 type IIIB Open injuries of which 39 had open tibial fractures and 13 had open femur fractures.	Direct comparision of predictive capabilities of MESS and GHSS criteria in paediatric trauma patients (MESS cut-off 6 or 7, GHSS cut-off 14)	The authors reported two primary amputations, plus two delayed amputations occurring due to post-operative distal limb ischaemia and/or muscle necrosis. Both primary amputation patients had MESS values =6, with the secondary amputation patients having MESS values of 6 & 7. Each primary amputation patient had GHSS values of >17, whereas the secondary amputation patients had GHSS values of 16 (corresponding MESS value 6) and 17 (corresponding MESS value 7) respectively(table 6). There were no reported cases of mortality. The independent rates of primary and secondary amputation were 3.85% for both types. The paper did not comment on complication rates. The authors reported, for GHSS threshold value 17, a sensitivity= 75%, specificity= 94%, PPV= 66% and NPV= 95%. For the same patients, the authors demonstrated for MESS threshold value 7 a sensitivity= 25%, specificity= 89%, PPV 17% and NPV 94%	Low sample size (however similar weakness across all selected papers). s, it becomes difficult to determine the presence of a grey zone for amputation and the chances of type II error becomes greater. A multicenter prospective analytical study is needed to further prove the applicability of GHOISS in children

Comment(s)

Despite the global burden of paediatric trauma, the Selected Papers demonstrated a paucity of data evaluating the prognostic efficacy of the Scoring Tools. The MESS score was the most frequently reviewed Scoring Tool. However, mixed values for sensitivity and specificity were apparent, with none of the Selected Papers demonstrating reliable statistical power regarding its prognostic value. Whilst there was evidence that the use of GHSS in lower limb trauma provided high sensitivity (and reliable statistical power), this was apparent in only 2 of the Selected Papers, with just 1 evaluation of its use in paediatric amputation patients. The remaining Scoring Tools, NISSSA, PSI and LSI, were applied in only one Selected Paper, which illustrated that patients may present with multiple injuries, plus prolonged ischaemic time and higher energy levels of injury mechanism being associated with greater risk of limb loss and mortality. The Selected Papers highlighted that patients may undergo limb salvage or amputation with a spectrum of Scoring Tool values, providing contradicting advice regarding likelihood of successful salvage. This illustrates the clinically challenging and multi-factorial decision-making associated with savage prediction of mangled limb injuries. Application of Scoring Tool metrics modelled on adults with little or no adaptation indicates further research is required to develop fit-for-purpose Scoring Tools for children. A dearth of evidence for salvage in upper limb trauma creates a difficulty in assessing if outcomes of limb salvage are “better predicted” by a particular Scoring Tool. The Selected Papers did, however, show discrepancy in threshold values in the Scoring Tools applied to upper and lower limb trauma, indicating that separate Scoring Tools for each may be required to increase accuracy in predication of successful salvage. Furthermore, this may assist with clinical decision making as to avoid unsuccessful limb salvage, ultimately leading to delayed amputation. Lastly, the Selected Papers applied varying criteria to selection of underlying data components applied in the use of Scoring Tools themselves, including, inter alios, age intervals, geographical location, and collection period. Absence of consistently controlled variables in the Selected Papers has corresponding implications on the consistency of Scoring Tools. Therefore, their sensitivity and specificity further skews the prospect in selecting a fit-for-purpose, or indeed a preferred Scoring Tool.

Clinical Bottom Line

There is insufficient evidence to recommend any prognostic scoring tool as adequately powered to correctly predict limb salvage.

References

1. Behdad S Evaluation of Mangled Extremity Severity Score (MESS) as a predictor of lower limb amputation in children with trauma European Journal of Pediatric Surgery Dec 2012;22(6):465-9
2. Mitchell F Fagelman, Howard R Epps, Mercer Rang Mangled extremity severity score in children Journal of pediatric orthopedics 2002 Mar-Apr;22(2):182-4
3. J Messner, P Harwood, L Johnson, V Itte, G Bourke, P Foster Lower limb paediatric trauma with bone and soft tissue loss: Ortho-plastic management and outcome in a major trauma centre Injury July 2020;51(7):1576-1583
4. Mommsen P; Zeckey C; Hildebrand F; Frink M; Khaladj N; Lange N; Krettek C; Probst C Traumatic extremity arterial injury in children: epidemiology, diagnostics, treatment and prognostic value of Mangled Extremity Severity Score Journal of orthopaedic surgery and research April 2010; 15:5:25
5. Zia Ur Rehman , Amna Riaz , Zafar Nazir Peripheral Arterial Injuries in Children: An Audit at a University Hospital in Developing Country Annals of vacular diseases Jun 2020; 13(2):158-162
6. David A Stewart, Christopher J Coombs, H Kerr Graham Application of lower extremity injury severity scores in children Journal of childrens orthopedics 2012 Oct;6(5):427-31
7. K Venkatadass, Tarani Sai Prasanth Grandhi, S Rajasekaran Use of Ganga Hospital Open Injury Severity Scoring for determination of salvage versus amputation in open type IIIB injuries of lower limbs in children-An analysis of 52 type IIIB open fractures Injury 2017 Nov;48(11):2509-2514