Comorbidities and outcomes in trauma admissions of all severities: a retrospective study

Traumatic injury remains a major contributor to morbidity and mortality around the world.¹ In New Zealand, trauma continues to pose a heavy burden on health services and communities, particularly as demographic ageing and rising chronic-disease prevalence intersect with injury risk.² The Te Manawa Taki (TMT) regional trauma registry,³ which covers a regional population of a million people, has recently published a 10‑year descriptive analysis (2013–2022) of trauma admissions in its catchment, noting that 92% of admissions were low severity and that the proportion of higher-severity trauma increased over the period.⁴ This evolving case mix highlights the importance of understanding not only injury patterns but also how patient factors, such as comorbidities, influence in-hospital courses and outcomes. 

Comorbid conditions are increasingly recognised as crucial modifiers of trauma outcomes. The recent scoping review by Glynn et al.⁵ emphasised that major-trauma patients with pre-existing comorbidities consistently face complications, longer stays and higher mortality. Globally, studies have linked chronic diseases such as diabetes, chronic kidney disease, cardiovascular disease and chronic pulmonary disease to elevated risks of acute complications, including sepsis, acute kidney injury (AKI) and hospital-acquired pneumonia.^6,7 Moreover, registry-based prediction work has begun to integrate comorbidity metrics to improve the risk adjustments that reflect a broader trend in trauma quality and safety—acknowledging that “host factors” such as pre-existing conditions are important in determining outcomes. Still, that evidence remains patchy, especially for their role in in-hospital complications, including patients with low-severity injuries, who make up approximately 90% of all trauma admissions.⁴

A recent publication,⁸ drawing on the TMT trauma registry of patients with all injury severities, found that the incidence of in-hospital complications can be meaningfully described for the TMT Region and are also generalisable New Zealand–wide, underscoring the feasibility and relevance of the approach. The TMT Region’s population is broadly representative of New Zealand as a whole^9,10—with a population of approximately 1 million, representing 20% of the New Zealand population—and displays demographic characteristics reflective of the country for median age (35 vs 38 years) and sex (48 vs 50% male),¹¹ but a higher proportion of Māori (26.5 vs 17.1%).¹² That work motivates further enquiry into which comorbidities are most strongly associated with key complications among trauma patients, not only in major injuries but in all injury severities.

Accordingly, this study aims to extend the current evidence by examining the relative prevalence of pre-existing comorbidities and their association with key in-hospital complications in all trauma patients recorded in the TMT trauma registry between 2014 and 2023. The study includes all ages, sexes, ethnicities and injury severities to capture the full spectrum of risk, with the expectation of providing evidence on the most significant comorbidities associated with complications in trauma patients during their journey in the hospital. The findings will support future work to develop risk stratification models for clinical use, helping clinicians anticipate and mitigate preventable complications.

Methods

A retrospective review of data from the TMT trauma registry was conducted. The data include patients of all injury severities, ages, sexes and ethnicities admitted to any TMT hospital over 10 years from 1 January 2014 to 31 December 2023.

All trauma patient admission records entered into the TMT trauma registry include a patient’s unique identifier, or National Health Index (NHI) number, as well as facility-specific arrival and discharge dates per admission. Data were extracted from the TMT trauma registry using DI Writer/Collector^TM.¹³ Consistent with trauma registries internationally, patients were excluded if they sustained periprosthetic or insufficiency fractures; exertional injuries; hanging, drowning or asphyxiation without evidence of external force; poisoning; ingestion of a foreign body; injury as a direct result of pre-existing medical conditions or late effects of injury; or the injury occurred more than 7 days before admission.^14,15 Demographic and injury-event information for the trauma registry was gathered from prehospital records, hospital systems and, when necessary, directly from patients. The Injury Severity Score (ISS) is calculated from the sum of the squares of the highest Abbreviated Injury Scale (AIS) grade score in each of the three most severely injured body regions, producing scores ranging from one to a maximum of 75. An ISS of 75 is considered to have the worst prognosis.¹⁶ The TMT trauma registry employs AIS version 2008. In this study, high severity or major trauma were defined as any patient with an ISS of >12, or who died in the hospital, including the emergency department; low severity or non-major trauma were defined as any patient with an ISS less than or equal to 12.¹⁷

The New Zealand National Minimum Dataset (NMDS)¹⁸ is a national collection of public and private hospital admission and discharge information. Both individual patient-level information and facility-specific admission information are collected and stored in the NMDS. All patient-level records include a NHI number, in common with the TMT trauma registry, and any clinical diagnoses information recorded in the NMDS is also specific to a named hospital admission.³ This allows any pre-existing comorbidities of the patient at the time of admission and hospital-acquired-complications diagnoses coding to be linked to a specific facility and admission-date range for any given patient. For this study, a subset of the NMDS containing all diagnoses for the hospital episodes in the TMT region was obtained from Data Services, Ministry of Health – Manatū Hauora, National Collections, NMDS (hospital events), allowing a complete mapping of TMT trauma registry records with NMDS data and classifying pre-existing comorbidities and complications.

The NMDS records were subsequently merged with trauma patient admission data from the TMT trauma registry across multiple fields, including the patient’s unique NHI identifier, facility name, hospital admission date-time, trauma event admission date-time, pre-existing comorbidities based on the comorbidity onset date–field and complication episode date–field elements. All patients were de-identified before analysis. International Classification of Diseases (Tenth Revision, Australian Modification; ICD-10-AM) entries, reflected in both the NMDS dataset provided and those initially entered by hospital-based clinical coders, were used to confirm the accuracy of dataset merging. A select subset of ICD-10-AM codes and groupings, provided by the Australia and New Zealand Trauma Registry (ANZTR), was used for classifying comorbidities and complications (see Appendix Table 1 and Appendix Table 2)¹⁹ and pre-existing comorbidities at the time of first presentation to a TMT hospital with injury for trauma care. Additionally, the classification used in the Charlson Comorbidity Index (CCI) was applied to the comorbidity data to produce 17 categories.²⁰ The CCI score was calculated by applying weights to each comorbidity category with hierarchy conditions and age-weights (see Appendix Table 3 for more details on the weight scores applied).^21–23

Variables examined include patient-demographic characteristics; age, grouped into life stages of 0–14, 15–64 and 65 years and older; sex, classified into female and male; and ethnicity, classified into Māori and non-Māori. To verify differences between in-hospital comorbidities and categorical variables such as sex and ethnicity a Chi-squared test was used, and the Student’s t-Test was used for continuous variables such as age and CCI. Prevalence and incidence rate ratios (IRRs) corrected for demographic characteristics such as ethnicity, sex and age group were used to produce adjusted rate ratios (aRRs). The aRR were calculated with a negative log-binomial regression²⁴ using 10-year mean events in the TMT trauma cohort as offset. To examine the association between the comorbidity categories and complications, correlation statistics were produced using a generalised linear model of logistic regression.²⁵ In all tests, the significance level was set at p<0.05. All statistical analyses were performed using RStudio version 2025.05.1,²⁶ using packages tidyverse²⁷ and epitools²⁸ with corrplot²⁹ for plotting the correlogram.

As administrative de-identified data were used, ethical approval was deemed out of scope by the New Zealand Health and Disability Ethics Committee, and research approval was provided under the locality authorisation process at Health New Zealand – Te Whatu Ora Waikato (RD024081).

Results

Over the period 2014–2023, there were a total of 56,113 trauma events and 7,818 cases with a pre-existing comorbidity, accounting for an average of 13.9% of all trauma admissions in the TMT Region over this period.

Among those who were injured, significant differences were observed with respect to the presence of pre-existing comorbidities across injury severity and demographic groups (Table 1). As more than 90% of the injuries were non-major, this is reflected in the sub-group with any comorbidities. However, pre-existing comorbidities were more common among those with major injuries than non-major injuries. Specifically, 791 of 4,405 major-injury patients (18.0%) had pre-existing comorbidities compared with 7,027 of 51,708 non-major-injury patients (13.6%). Pre-existing comorbidities were more prevalent among non-Māori. Specifically, 5,910 of 38,665 non-Māori (15.3%) had at least one comorbidity compared with 1,908 of 17,448 Māori (10.9%). Among those with any comorbidities, 75.6% were non-Māori and 24.4% were Māori. There were more females with a pre-existing comorbidity (4,117 out of 17,448 [19.4%] of the injured females and 52.7% of those with any comorbidity) compared with males (5,910 out of 38,665 [10.6%] of the injured males and 47.3% of those with any comorbidity). Most trauma admissions involving comorbidities occurred in the older age group, 65+ years (59.5% of those with a pre-existing comorbidity), and 4,649 out of 11,197 accounted for 41.5% of the age group 65+.

View Table 1–3, Figure 1–2.

Among individuals with pre-existing comorbidities, the mean of CCI was 4.7 (SD=2.4). The mean CCI was higher among non-Māori compared with Māori (5.1 [SD=2.3] versus 3.5 [SD=2.4], p<0.001), and among females compared with males (5.0 [SD=2.3] versus 4.4 [SD=2.5], p<0.001). Mean CCI increased with age, with the highest scores observed in those aged 65+ years (5.9 [SD=1.9]), followed by 2.8 (SD=1.7) for those aged 15–64 and 1.2 (SD=0.6) and lowest for 0–14 year olds (F=2,687, p<0.001).

The incidence of pre-existing comorbidities in injured patients (median age of 35 years [IQR 17–59]) has steadily increased since 2014 (Figure 1). Similarly, the incidence of complications among trauma admissions has gradually increased over this period.

The aRR show that patients of non-Māori ethnicities, females and those 65 years and older are at higher risk (p<0.001) of having a comorbidity and a complication during their admission (Table 2). For those injured patients with pre-existing comorbidities, people 65 years and above have a higher risk of complications.

Among the sub-group with any comorbidities, diabetes (including type 1 and 2) is the most commonly observed (27.4% without or with complications), followed by cerebrovascular diseases (12.5%) and chronic pulmonary diseases (12.4%) (Table 3). The distribution of specific comorbidity conditions among major-injury and non-major-injury severities were not significantly different for most conditions. Significant differences were seen only in diabetes with chronic complications (p<0.05) and chronic pulmonary diseases (p<0.01), with a higher proportion among the non-major-injury group.

Several comorbidities show significant correlations with in-hospital complications in trauma patients (Figure 2). Diabetes, with or without chronic complications, is associated with AKI (p<0.05). Diabetes with chronic complications, however, is linked to other complications such as cardiac arrest, catheter-associated urinary tract infection (UTI), central line–associated bloodstream infections, coagulopathy and surgical-site infections (p<0.05).

Cerebrovascular disease is associated with several complications, including AKI, catheter-associated UTI, coagulopathy and post-haemorrhagic anaemia (p<0.05). Chronic pulmonary diseases are linked to acute respiratory distress syndrome and central line–associated bloodstream infections (p<0.05). Chronic kidney disease is associated with various in-hospital complications, such as AKI, catheter-associated UTI, central line–associated bloodstream infections, coagulopathy, sepsis, pneumonia and flap/graft failure (p<0.05). Similarly, peripheral vascular disease is associated with several complications, including AKI, central line–associated bloodstream infections, coagulopathy, pneumonia and graft failure (p<0.05). The generalised linear model statistics with a log link function indicate a significant relationship between age-weighted CCI score and the occurrence of complications (F[428]=0.22, p<0.001).

Discussion

This study profiles comorbidities in trauma admissions from a population of approximately 1 million people in New Zealand. The pre-existing morbidity patterns of injured patients reflect broader morbidity and public health trends in New Zealand, with diabetes, chronic pulmonary disease and cerebrovascular disease being the most common pre-existing conditions.³⁰ This pattern aligns with national and international evidence. The age-standardised prevalence of diabetes rose from 36.6 per 1,000 in 2013 to 47.0 per 1,000 in 2024.³¹ Further projections suggest that the number of people living with diabetes in New Zealand will nearly double by 2044.³² These findings help to explain why “diabetes without chronic complications” often appears as the leading comorbidity in injury populations.⁵

The demographic variations observed, such as higher rates of comorbidities among older age groups and females, reflect known health inequities. Age-group patterns in this study align with New Zealand health statistics: in children (0–14 years), chronic pulmonary disease predominantly influences morbidity; among older adults (65+), both cerebrovascular disease and malignancies are leading causes of morbidity, consistent with international literature.^33–35 However, the differences by ethnicity in the trauma cohort are not consistent with existing studies and need further exploration. In their “inequity of morbidity” study, Gurney et al.³⁶ showed that Māori and Pacific peoples carry a disproportionately high burden of many common chronic diseases, often at younger ages, compared with non-Māori/non-Pacific groups.

This study reinforces and expands on growing national and international evidence that some of the pre-existing comorbidities are strong predictors of in-hospital complications in trauma patients. Within the TMT cohort from 2014 to 2023, patients with pre-existing comorbidities at admission had markedly higher risks of complications such as AKI, catheter-associated UTI and pneumonia, compared with those without comorbidities, consistent with prior New Zealand and global evidence.^5,8 Complications such as AKI, coagulopathy, pneumonia and central line–associated bloodstream infections are notably linked with several comorbidities. In addition to diabetes, chronic pulmonary diseases and cerebrovascular diseases, chronic kidney disease and peripheral vascular diseases are also associated with complications. Because we included the entire trauma population across all severities and ages, our findings emphasise that the influence of comorbidity is not limited to major-trauma cases but extends across the full spectrum of injury. The presence of multiple comorbidities, indicated by a higher CCI score, correlates with an increased likelihood of several complications in trauma patients.

Importantly, our results dovetail with recent region-specific work: the analysis of TMT in-hospital complications⁸ indicates that complication incidence is measurable at scale in New Zealand trauma cohorts. This study further showcases the possibilities for other regions of New Zealand to integrate their trauma data with other hospital data systems to identify risk predictors for trauma patients. By linking specific comorbidities (e.g., diabetes, chronic pulmonary disease, cerebrovascular disease, renal disease) to discrete complication types (such as central line–associated bloodstream infections, coagulopathy, sepsis, pneumonia, AKI), the regional evidence base can move beyond simple comorbidity counts toward complication-specific risk profiling. However, future studies are needed to achieve this, as this study has limitations. This study did not examine the mediating and moderating role of age in comorbidities and duration of the pre-existing comorbidity, which could have affected the findings on ethnicity and needs further study. Further, the study used ICD-10 coding to classify complications, which is done post-discharge and limits its applicability in a clinical setting. Further examination of complications using clinical and biochemical variables is needed to enable risk profiling of trauma patients.

From a clinical and systems perspective, these findings suggest that early incorporation of comorbidity assessment in trauma pathways is essential, as is early referral to appropriate speciality services in trauma centres. Deploying comorbidity and frailty indices at admission, combined with injury severity and demographic data, can enhance the precision of risk stratification.^37,38 Given the integrated registry infrastructure in the TMT Region and the work being done towards digitising the trauma registry data collection, it is feasible to embed predictive models and real-time decision support into existing workflows. Doing so will link the registry to clinical decision support systems, allowing clinicians to flag high-risk patients immediately, allocate preventive resources (e.g., renal protective strategies, early catheter removal, antimicrobial prophylaxis) and monitor performance against complication rates over time.

Conclusion

By integrating newer evidence, including regional registry-based complication analyses and international scoping reviews, this study both confirms known associations and advances locally calibrated risk insight for the New Zealand trauma system. The next steps should include validating model performance, assessing prospective implementation and iteratively refining algorithms with incoming registry data. Over time, embedding this predictive capacity into the trauma care pathway may reduce preventable complication burden and improve patient outcomes.

View Appendix.