This is the first study in Uganda and Africa to use the GHOISS to predict the need for amputation and 14-day treatment outcomes for Gustilo type III A and B tibia fractures. We hope that these results will help improve the management of these injuries and in turn reduce the socioeconomic burden from associated morbidity and mortality.
For a long time, the management of Gustilo type III tibia injuries has been intricately important in regard to limb salvage or removal (Rajasekaran et al. 2015). This dilemma has often resulted in incorrect decisions, i.e., unnecessary or delayed amputations, associated with substantial morbidity and mortality (O’Hara et al. 2018; Rajasekaran et al. 2006; Schirò et al. 2015; Ssebuufu et al. 2017). Unsuccessful limb salvage can result in long-term suffering, which can include pain, long hospitalizations, an increased number of operations and financial depletion (Sivagnanam et al. 2021).
The Ganga Hospital Open Injury Severity Score (GHOISS) was developed in 1994 and validated in 2006 by Rajasekaran and colleagues in an attempt to answer the question of limb salvage or amputation in OTF type III patients (Mehmood et al. 2019; Nicolaides, Pafitanis, and Vris 2021; S. Rajasekaran et al. 2015). In earlier studies, the GHOISS was demonstrated to effectively predict the need for amputation or limb preservation in Gustilo type IIIB and IIIA injuries (Elniel and Giannoudis 2018). The GHOISS has equal specificity and sensitivity for limb salvage or amputation in both adults and children (Nicolaides et al. 2021; Venkatadass et al. 2017). A GHOISS of 17 or more had the highest sensitivity and specificity for limb loss, whereas a GHOISS of 14 or less predicted successful salvage (Nicolaides et al., 2021; Rajasekaran et al., 2015).
This study included both children and adults, and the injury severity distributions according to the GHOISS were as follows: Group I (5 fractures), Group II (33 fractures), Group III (12 fractures), and Group IV (3 fractures). We found that the Group II GHOISS injury category was the most predominant (33/53; 62.3%), followed by Group III (12; 22.6%). This was consistent with the findings of (Rajasekaran et al. 2006), who reported group II injuries as most predominant (46.8%), followed by group III (26.6%), (Basu et al. 2022) (48% as group II and 44% as group III), and (Madhuchandra et al. 2015), who found 50% of injuries to be group II and 20% as group III. However, (Venkatadass et al. 2017) reported different findings, with group III injuries being the most predominant at 46.2% and 36.5%, respectively, for group II. The difference in the study design and inclusion criteria could be responsible for the differences observed. The retrospective study included only children with open tibia or femur fractures.
Males were the most common (i.e., 44/53 (83.0%)) and were likely to have a greater grade of injuries than females were, in agreement with the 90%, 95% and 98.2% observed in the cohorts of (Basu et al. 2022; Madhuchandra et al. 2015; Rajasekaran et al. 2006), respectively. This can be explained by the high-risk nature of the activity and lifestyle of men in these communities. Twenty-nine of the 53 patients (55%) had at least one or more comorbidities according to the Ganga score, which was close to the 50%, 51.4% and 57.5% observed in the (Basu et al. 2022), (Rajasekaran et al. 2006) and (Madhuchandra et al. 2015) cohorts, respectively, but lower than the 78.8% observed by (Venkatadass et al. 2017). This discrepancy may be due to the retrospective design and inclusion criteria of the Venkatadass cohort.
The mean age was 34 years, which is the age at which individuals are most active and engaged in productive work, increasing their risk of injury. This was similar to the 34.97 years observed by (Rajasekaran et al. 2006), probably due to the same age criterion consideration of all age groups but different from the 41.32 years in (Basu et al. 2022) and 38.15 years in (Madhuchandra et al. 2015) due to the different age considerations in these two studies that excluded children.
Forty-seven of the 53 (89%) Gustilo type III tibia injuries were successfully salvaged, while 6/53 (11%) resulted in amputation. Four of the six (4/6) amputations were performed primarily, while 2/6 amputations were performed after unsuccessful salvage. Secondary amputations were performed due to severe soft tissue infection. The mean GHOISS score for amputation was 14.8. Five of the six (83%) of the amputated limbs belonged to a high injury group, i.e., either group III or IV. All OTFs that resulted in amputation had a comorbid factor as described by the GHOISS system.
These findings are in keeping with the findings of (Basu et al. 2022; Madhuchandra et al. 2015; Rajasekaran et al. 2006; Venkatadass et al. 2017), who reported a successful salvage rate greater than 86%. They also found that the outcome was closely associated with the GHOISS category, with the majority of patients in the amputation group being in group III or IV, with a GHOISS score greater than 14. Madhuchandra et al., 2015, observed a very high salvage rate of 97.5% because his cohort was dominated by patients with mild form of injury and only one limb with a GHOISS score greater than 15. Secondary amputation was due to persistent infections, similar to the findings of other studies (Rajasekaran et al., 2006) (Basu et al. 2022; Venkatadass et al. 2017).
Binary logistic regression with amputation as the dependent variable and GHOISS score as the independent variable showed that GHOISS score significantly predicted amputation (p = 0.000). A GHOISS of 13 was found to be the threshold score above which amputation should be considered. In our cohort, only one patient (GHOISS = 10) underwent an amputation, which was performed secondarily after the GHOISS had predicted a salvage. The remaining 5 OTFs that resulted in amputation had (GHOISS) score of 13 or above. On plotting the receiver operator curve (ROC), the area under the curve (AUC) for GHOISS prediction for amputation was 0.927 (95% CI 0.830-1.000), with a sensitivity, specificity, and accuracy of 83.3%, 91.5% and 90.6%, respectively, for GHOISS 13.
This finding is in agreement with that of (Basu et al. 2022; Madhuchandra et al. 2015; Rajasekaran et al. 2006; Venkatadass et al. 2017), who also reported that the GHOISS score is a significant predictor of amputation in patients with Gustilo type III A and B tibia injuries. However, our threshold score of 13 is slightly lower than the 14 reported by these previous studies. The reason for this difference could be that the previous studies were conducted in highly specialized centres with improved health care systems as well as via the orthoplastic approach, where the orthopaedic surgeon worked with the plastic surgeon straight from the initial debridement through the stages of reconstruction.