Outcomes of hip fracture in centenarians: a systematic review and meta-analysis

Key summary points Aim We aimed to determine outcomes of hip fracture in centenarians. Finding One-year mortality following a hip fracture was 53.8% (95% CI 47.2 to 60.3%). Dementia (26.2%, 95% CI 15.7 to 38.2%) and hypertension (15.6%, 95% CI 3.4 to 33.1%) were the most common comorbidities. Message Effective cross-discipline communication and intervention is suggested to promote treatment outcomes.


Introduction
The number of oldest-old is increasing in high income countries [1], with declining mortality in the 80 years and older age group and improved treatments contributing to an increasing number of centenarians in the last decades.The number of centenarians is projected to increase to between 13 and 50 million individuals worldwide during the twentyfirst century [2].Centenarians experience a higher mortality than other older age groups, such as those aged 80-99 years [1].
Older people are susceptible to trauma owing to an increased risk of delirium and dementia [3], incontinence, frailty, impaired vision and drug interactions [4].In addition, side effects of commonly used drugs with central nervous system effects (such as benzodiazepines and antipsychotics) can increase risk of falls, as well as orthostatic hypotension.Resulting fractures are the third most common cause of hospitalisation for this population [5].
Fractures in older people can lead to significant disruptions in daily activities and loss of independence.A review of outcomes following hip fracture estimated between 40 and 60% of patients do not recover their pre-fracture level of mobility, while 30 to 60% do not regain their level of independence for basic activities of daily living [6].
Outcomes of hip fracture in centenarians remain comparatively under-explored owing to the small number of patients reaching 100 years of age [7].This group of patients are at particular risk of poor outcomes following hip fracture.Centenarians have a high comorbidity burden and polypharmacy, resulting in increased surgical risk [8].Progressively poorer bone quality than younger individuals predisposes the oldest of people to more complex fractures [9].
To the authors' knowledge, no systematic reviews have been undertaken assessing the outcomes or characterisation of centenarians who have sustained a hip fracture.Understanding hip fracture outcomes of people 100 years and older is required considering the number of centenarians is projected to increase [2].Though negative outcomes may be expected due to a high comorbidity burden, these are yet to be formally synthesised.Accordingly, this review aimed to determine outcomes of hip fracture in centenarians and to identify the most common comorbidities among centenarians with hip fractures to better characterise this population.

Methods
The PRISMA 2020 checklist was satisfied in the reporting of this systematic review [10].The protocol for this systematic review was prospectively registered (PROSPERO Registration: CRD42022377719).

Study eligibility
Studies were eligible if they reported outcomes of treatment for hip fractures in centenarians (both conservative and surgical), and/or comorbidities in centenarians who sustain a hip fracture.Both full-texts and abstracts were included.Eligible study designs were case series, case-control, cross-sectional and cohort studies, as well as randomised controlled trials.Both retrospective and prospective studies were eligible.Papers not reporting original data such as literature or systematic reviews were excluded, along with case reports and letters to the editor.There was no constraint based on language, publication status or patient demographics.Eligibility assessment was performed independently by two reviewers (DAAL, CKB).Disagreements regarding study eligibility were solved through discussion.

Search strategy and data extraction
We searched the following electronic databases via OVID: MEDLINE, Global Health, and Embase.Currently registered studies were reviewed using the databases: ISRCTN registry, the National Institute for Health Research Portfolio, the UK National Research Register Archive, the WHO International Clinical Trials Registry Platform, and OpenSI-GLE (system for information on grey literature in Europe).Conference proceedings from the European Federation of National Associations of Orthopaedics and Traumatology (EFORT), British Orthopaedic Association (BOA) and British Trauma Society (BTS) were searched.The reference lists of included studies were also searched (backwards-searching).Finally, papers citing the studies included were also reviewed for eligibility (forward-searching).
Database search and data extraction were conducted independently by two reviewers (DAAL, CKB).Searches were conducted twice for quality assurance.The final search was completed on the 25 th of January 2023.The search strategy is presented in Appendix 1 and modified for each respective database.
Data were extracted onto a data extraction template.Data extracted included: baseline characteristics including number of patients, treatment received, patient sex, age, study location, fracture type, fracture management, comorbid diseases, mortality and complications.

Outcomes
All-cause mortality was calculated, including in-hospital mortality, mortality at 1, 3, 6 months, and 1-year post-fracture.At this extreme of age, mortality beyond 1 year was excluded, as it may be unrelated to the hip fracture.Other outcomes included prevalence of comorbidities in centenarians and non-centenarians, and complications (occurrence and rate) over follow-up periods.The latter were not restricted to a specific time after fracture, since no study defined the duration of the observation period for which complications were monitored.Therefore, we assumed these took place before the final follow-up period post-operatively reported in the studies.

Methodological appraisal
Level of evidence and risk of bias of each included study were evaluated independently by two reviewers (DAAL, CKB).The level of evidence of the studies presented was determined with the March 2009 Oxford Centre for Evidence-Based Medicine: Levels of Evidence (5 = lowest level of evidence, corresponding to case reports; 1a = highest level of evidence, corresponding systematic reviews of randomised controlled trials) [11].Each study was appraised with an appraisal tool reflecting the study design.Accordingly, tools used included the Institute of Health Economics case series studies quality appraisal checklist [12] and the Downes and Black Tool for cross-sectional studies [13].

Data analysis
Where sufficient (at least two) and homogeneous studies reported on the same outcome, a random-effects metaanalysis was performed using MetaXL version 5.3 software (EpiGear International Pty Ltd, Wilston, Queensland, Australia).Data on prevalence was presented as weighted prevalence percentage and 95% confidence intervals (CI).Data on continuous outcomes, i.e.Charlson Comorbidities Index (CCI) was presented as weighed mean difference and 95% CI.Statistical heterogeneity was assessed using Cochran's Q value and Higgins I 2 statistic for each pooled analysis.This was interpreted in accordance with Higgins and Green [14].Subgroup analysis of surgical management of hip fractures was undertaken, comparing mortality following surgical intervention with the overall cohort mortality (irrespective of treatment received).

Search results
In total, 4671 records were screened, of which 23 studies were eligible and evaluated 6970 centenarians (Fig. 1, Table 1).Mean patient age was 101.5 (range 100-111) years.Twenty-two studies reported patient sex (816 males, 11.7% and 6133 females, 88.3%).No study reported whether their patients had concomitant injuries, patient osteoporotic status, nor commented on the severity of hip fracture.

Study quality assessments
All studies identified were case series or cross-sectional studies, with a corresponding level of evidence being four.Whether relevant outcome measures were established a-[ priori or if outcome assessors were blinded to intervention administered were unclear for all studies (Table 2).Overall, the majority of studies included exhibited methodological limitations pertaining to low level of evidence and concerns regarding risk of bias.

Discussion
Knowledge of the mortality associated with hip fracture provides further evidence of the importance of health professionals to hold informed discussions with patients and relatives regarding prognosis following this injury type.However, the majority of studies included in this review exhibited methodological limitations pertaining to low level of evidence and concerns regarding risk of bias.Caution should, therefore, be placed when interpreting these findings.
Due to increasing mortality with age, it is important to ascertain whether hip fractures lead to increased mortality in centenarians.A single study found centenarians with a hip fracture had a lower mean survival time and hospitalization-free survival time than those without a hip fracture [22].Further research should perform such analysis to more reliably establish the disease-specific mortality associated with hip fracture.Alvarez et al. [35] stratified mortality in centenarians according to degree of frailty."Robust" and "intermediate" centenarians (higher cognitive and physical capabilities than frail centenarians) had a mortality of less than 10% within a year.A pooled 1-year mortality of 51% was calculated in this meta-analysis.Therefore, centenarians with hip fracture may experience a higher mortality than centenarians without hip fracture.Further research matching centenarians with hip fracture and autonomous centenarians without a hip fracture is required to validate this hypothesis.
Current evidence suggests centenarians suffering from a hip fracture have a higher pre-injury functional status [30], lower medication burden [26,30], and a lower number of comorbidities [26,29,31] than their younger peers (with the   latter demonstrated by meta-analysis).Individuals suffering from hip fractures are more likely to be ambulatory [7].This may render centenarians suffering from hip fractures a selfselecting group, with a lower comorbidity burden allowing them to ambulate, and subsequently fall.This may explain the lower prevalence of comorbidities when compared to a random selection of older people [33].However, we can only draw conclusions based on our study population (centenarians with hip fractures).Our findings are not applicable to centenarians without hip fractures.In addition, they are hindered by the low level of evidence of the studies included in this review.Dementia was the most commonly reported comorbidity, being present in over a quarter of patients suffering from a hip fracture (26.2%).This is lower than the estimated 40% prevalence in patients over 100 years of age [36].This may be explained by the fact that centenarians with moderate-severe dementia are mostly dependent on a wheelchair for ambulation, and are hence less susceptible to falls and hip fractures.Hypertension was the second most common comorbidity, with a pooled prevalence of 15.6%.
Antihypertensive medication use may predispose fall injuries among older people [37].Mazzola et al. [22] matched centenarians with and without a hip fracture, and found the former were more likely to be on an antihypertensive agent.However, whether hypertension itself or taking antihypertensive medications are the culprits remains unclear.A lower prevalence of hypertension was noted in this meta-analysis than is usually reported in other population-based studies [38].Centenarians with multiple morbidities, requiring multiple medications like anti-hypertensives, are more likely to be less ambulatory.Therefore, hypertension as a comorbidity may be under-represented in our study population.Other medications which may increase the risk of falls in the elderly include diuretics, sedatives, hypnotics, antidepressants, benzodiazepines and nonsteroidal anti-inflammatory drugs [39].Though medication reviews are commonly performed in clinical practice regardless of patient age, these are particularly relevant for centenarians given the detrimental outcomes resulting from fractures.
Pneumonia was the most common complication following treatment of hip fracture in centenarians, occurring in 10.0% of patients.This was followed by urinary tract infection (3.7%).Infective presentations could be attributed to centenarians with hip fractures having a long hospital stay (calculated at 28.95 days on average).Considering infection were the two most common complications, antibiotic prophylaxis in centenarians with a hip fracture may be warranted.Other common complications included arrhythmia (3.2%), heart failure (2.9%), and anaemia (2.8%).These may prompt cardiovascular input.Cases of hip fracture in centenarians are typically complex, with potentially multiple comorbidities and complications.A multidisciplinary approach is required for optimal treatment, with orthogeriatricians' involvement advocated by European and UK guidelines [40,41].Sub-group analysis comparing overall mortality irrespective of treatment, and mortality in patients undergoing surgical repair indicated similar survival and overlapping confidence intervals up to 1 year after the hip fracture.Therefore, surgical repair following a hip fracture may not provide an additional benefit in terms of mortality reduction.A fixed fracture may lead to improved pain management, quality of life, ease of nursing and personal care, which would justify operative intervention.This meta-analysis cannot determine whether surgical repair yields better functional outcomes or lower complication rates than conservative management due to the lack of studies evaluating these parameters in operative and conservative interventions separately.Further study should explore these parameters to ascertain whether non-operative treatment leads to increased functional recovery.
Time-to-surgery was reported in 11 studies.These ranged from 1.6 to 4 days.This is in contrast with current treatment guidelines, which recommend surgery be performed within 24 h of admission [40,41].Such delay could be explained by the need for medical optimisation before orthopaedic surgery can be performed on centenarians [42].However, lack of capacity in the operating theatre and/or ageism may be contributing factors [43].Only one study reported that, of nine patients, seven were operated within 48 h, whereas the remaining two were delayed due to medical stabilisation.Further work on hip fractures in centenarians should report reasons for delays in time to surgery, given this is a negative prognostic indicator and must be addressed [44].
Our review was limited by the inclusion of studies with varying dates and duration of retrospective periods.As surgical procedures may have changed over the course of long time periods, the mortality data may be affected.In addition, to calculate mortality regardless of treatment, patients who did not undergo surgical repair were included.These patients may have had severe comorbidities or have been at the end of life, which could have increased the mortality figures calculated.Current evidence has limitations which must be improved to increase the understanding of outcomes of fractures in centenarians.Firstly, the vast majority of studies carried a low level of evidence with major concerns regarding their risk of bias.All studies were case series, which is the most feasible study design to evaluate outcomes of fractures in centenarians given the small number of patients reaching this age.To conduct large clinical trials of this population, a national or international network of trial centres would be warranted.Secondly, only two studies reported patient ethnicity [23,27].
Studies also poorly reported the characteristics of their cohort in relation to concomitant injuries and osteoporotic status.Given that both factors significantly impact on prognosis, this is an important limitation which should be addressed in future study reporting on this population.Thirdly, no studies clarified the time-frame following fracture within which complications were identified.Therefore, it is not possible to determine when these occur following hip fractures in centenarians.Fourthly, 12 studies did not report time-to-surgery.This is a crucial parameter for successful operation and survival, and is particularly important for prognosis in centenarians [44].Further research should report on this parameter given its impact on prognosis.Finally, there was a very low number of studies reporting outcomes for surgical and conservative management separately.The comparison of these approaches, investigating both mortality but also health-related quality of life, is required to improve understanding on what is the best care for people aged 100 years and older who experience hip fracture.

Fig 3 )
Fig. 1 PRISMA diagram depicting the study collection process

Fig. 2
Fig. 2 Pooled overall mortality following a hip fracture

Fig. 3
Fig. 3 Pooled mortality in patients undergoing surgical intervention

Table 2
Results of risk of bias assessment

Table 4
Pooled mortality in patients undergoing surgical interventionLCI lower 95% confidence interval, HCI higher 95% confidence interval, N number of patients included in analysis

Table 6
Pooled comorbidities in centenarians with a hip fracture LCI lower 95% confidence interval, HCI higher 95% confidence interval