Cervical fractures can occur in patients of any age. However, in older adults the force required to cause these injuries may be much less than in younger patients.[[1]] Factors inherent to the older adult such as frailty, medical comorbidities and medications can all lead to an increased risk of falls. The risk of fracture itself is also increased due to osteoporosis[[2]] and degenerative changes in the cervical spine.[[3]] Studies have also shown that, despite a decreasing number of neck injuries in patients under the age of 65, there is an increasing incidence in patients aged over 65 years.[[4,5]]

In all age groups, C2 is the most frequent location for a fracture in the upper cervical spine. This is particularly so in patients aged over 80.[[6,8]] The two most common types of C2 fracture are through the dens (type II or III odontoid) and body of C2 (hangman) fractures, and all are unstable.[[7,8]] Although these injuries can be caused by high-energy impacts such as motor vehicle accidents, low-energy injuries such as falls are the most common mechanism of C2 fractures in older people.[[7,8]]

The management of C2 fractures is stabilisation to prevent neurological injury and promote bony union. The three main methods used are (1) rigid cervical collar immobilisation, (2) halo-vest immobilisation and (3) surgical fixation. The decision between these options depends on fracture and patient characteristics, but there is a lack of consensus, especially in older patients, as to the best treatment option.[[3,5,9,10]] Although use of surgical fixation may be increasing over time, rates of non-union, complications and mortality are variable and there are concerns about operating on comorbid and frailer older patients.[[3,11–15]] Complications of management include non-union (which can be associated with sudden death) and neurological complications from instability as well as postoperative and comorbid medical issues.[[3]] Complication rates reported from any management vary widely, from 21% to 44%,[[3,7,15–18]] and no treatment modality seems better.[[3,5,9,10]] In our centre, most of these patients are managed conservatively with external immobilisation, and very few are having surgical fixation.

Our clinical experience is that older patients with C2 fractures are very frail, have comorbidities and become functionally compromised by their immobilisation with a rigid cervical collar. The complications of rigid collar use appear to be infrequently reported.[[11,15,18,19]] Therefore, we chose to review the management, complications and functional outcomes of consecutive older patients with odontoid (C2) fractures.

Methods

The study was carried out at Christchurch Hospital, the sole acute hospital in Canterbury, New Zealand, with a catchment population of 520,000 people. It was a retrospective review of the management and outcomes of all patients aged 65 and over who were discharged with an ICD-10 discharge code of S12.1 (fracture of second cervical vertebra) from Christchurch hospital over five years (1 January 2009 and 31 December 2013 inclusive). All patients were initially managed and followed-up by the orthopaedic spinal team. The clinical records (both electronic and paper) were reviewed by a single clinician (KC), and data regarding mechanism of injury, comorbidities, complications, discharge location, requirement for rehabilitation and mobility and mortality outcomes were collected. Outpatient clinic notes were accessed via the electronic patient management system (Health Connect South (HCS)). Mortality data (out to two years post fracture) were obtained from the same electronic record (HCS), which is linked to the National Mortality Database using each patient’s unique National Health Index number.[[20]]

Low-impact injuries were defined as a fall from standing height or less. High-impact injuries were defined as a fall from greater than standing height or other trauma, such as injuries sustained in motor vehicle collisions.

A Charlson Comorbidity Index (CCI) score was calculated for each patient.[[21]]

Complications were recorded and categorised into one of six categories:

1. Complications directly related to the injury (pain, transfusions, other injuries)
2. Direct neurological damage from the injury (intracranial events, cord or nerve impingement)
3. Fracture fixation or collar problems (poor compliance, swallowing difficulties, pressure injuries from the collar)
4. Medical complications (infections, cardiac problems, incidental scan findings, malignancies, endocrine abnormalities),
5. Immobility-related complications, such as constipation, deep vein thrombosis/pulmonary embolism (DVT/PE) or pressure areas not from collar
6. Psychological problems (Delirium, depression, anxiety)

Data that were not normally distributed are presented as medians and interquartile ranges, and Chi-square analyses were used for comparisons of categorical data.

The New Zealand Health and Disability Ethics Committees (HDEC) advised that formal ethics approval was not required as the review was considered a retrospective audit. The Māori Health Team (Te Komiti Whakarite) at Christchurch Hospital, New Zealand, also reviewed the study plans and gave consent.

Results

A total of 69 patients were identified as having the discharge S12.1 code. Five patients were excluded from final analysis: two patients because they did not have a C2 fracture, and three because the relevant paper notes could not be obtained. This left a total of 64 patients whose medical notes were analysed in full. Full patient characteristics are shown in Table 1. Patients had a mean age of 80.6 years and were predominantly of European ethnicity (95%). Forty percent had high-impact injuries and 60% of patients had low-impact falls.  All patients were managed conservatively with a non-operative method of fixation. The majority (72%) were managed in a rigid cervical collar. Of those in the collar and brace system, one was managed in a halo-vest immobiliser and traction for a short duration, but this was stopped due to delirium and they were converted to a rigid collar, whereas the others were in the Aspen®CTO system only. Thirty-seven patients (58%) had inpatient rehabilitation associated with this injury (13 orthopaedic geriatric rehabilitation unit, nine spinal unit and nine older persons health rehabilitation unit, six in other district health boards (DHBs)).

Table 1: Patient characteristics. View Table 1.

Table 2 demonstrates the outcomes in this group of patients. The most frequent complications were medical (60%) or complications directly arising from the use of a collar (58%).

The most frequent medical difficulties were lower-respiratory infections (14 (22%)), and four (6%) had respiratory distress/needed intubation. Urinary retention (7 (11%)) and falls (6 (9%)), with two (3%) patients sustaining additional fractures, were next most frequent issues.

Collar complications were predominantly pressure related: early reddening or irritation (n=8 (13%)), established pressure areas under chin (n=4 (6%)) or on chest (n=3 (5%)), difficulty eating (n=9 (14%), with one requiring nasogastric tube) and discomfort/pain (n=21 (33%)). Pain from the collar was severe enough that six abandoned the hard collar, and a further two patients attempted intermittent self-removal. Difficulty walking or participating in rehabilitation was reported in four  patients (6%) with a hard collar.

Constipation (11 (17%)) was the most frequent immobility-related problem. No cases of venous thromboembolism were recorded, and there were two pressure injuries (not related to the collar).

Seven patients sustained direct neurological injuries, all of which (except one) were as a result of high-energy injuries. Five of these seven patients died as a direct result of the severity of their intracranial injuries. Delirium (psychological complication) occurred in at least 12 patients (19% of all patients).

Table 2: Patient outcomes. View Table 2.

The majority (75%) of patients who were living at home prior to their fracture were able to discharge to their own homes from either their acute inpatient stay or their time in rehabilitation. Of those patients who were originally admitted from residential care, only one patient died, and the remainder were able to return to their residential care facility. There were six new admissions to residential care, although one of these patients only stayed for short-term care until their collar was removed and then they were able to return home.

Six patients (9%) died during their inpatient stay: five of these sustained significant neurological injuries as a result of high-energy impacts, and one had multiple medical comorbidities and recurrent falls. These patients all died within 10 days. Fourteen patients had died at one year (22%) and 21 by two years (33%).

Significantly more patients required a mobility aid on discharge than on admission (Chi square=43, p<0.0001). Sixty percent of patients required a walking frame on discharge and only 14% were able to walk independently (without an aid) on discharge, whereas 67% were documented to be independent walkers on admission. Four patients were discharged home from hospitals outside of the Canterbury area, so their final discharge mobility is unknown.

All patients alive at discharge (and in region) were followed in outpatients until their fractures were stable. At six months, three patients had bony non-union recorded, but all three were thought to have fibrous union and were stable.

Discussion

The key findings in this observational study were that all patients were managed conservatively, mortality at one year was 21.9%, which is similar or better than other studies,[[3,8,11,15,17]] and complications were very common.[[3,7,15,18]] Most older people needed further inpatient rehabilitation, in part because of the rigid collar fixation and in part because of comorbidities.  Sixty-seven percent walked independently without an aid prior to their injury, but the majority required some walking aid after discharge.

A rigid collar is an important treatment modality in the management of C2 fractures, but poor compliance can lead to non-union, which is associated with sudden death.[[3]] Previous studies have suggested that halo fixation is poorly tolerated in older people and concluded that rigid cervical collars are better tolerated.[[3,6,14]] Halo fixation was not tolerated by one patient in our cohort. Most of our patients were managed conservatively in a rigid collar, which is thought to be better tolerated. Despite this, nearly two-thirds (58%) of our cohort suffered a complication related to their collar, including difficulty eating and pressure areas as well as intolerance leading to poor compliance. The collars also created physical limitations on patients, such as being unable to look down to see their food or feet. This appeared to impact their ability to carry out activities of daily living (ADLs) such as walking, toileting and washing. Many older adults rely on visual cues for walking, and so rigid immobilisation of the neck can impact their ability and their confidence to walk. This may be one reason why there is an increased need for a walking aid on discharge, as was seen in our cohort of patients. The longer length of stay (mean 23 days, range 1–99 days) demonstrates the time it takes for patients to overcome their acute injury and learn how to manage their ADLs, injuries, collar cares and pain sufficiently to manage in the community.

A small proportion (six) of our patients were managed with a soft collar only. Although this is not recognised as a standard management strategy, three of these patients were still alive at two years despite having a soft collar as their only treatment. Soft collars may be a management option, but only in very select older adults who cannot tolerate more rigid immobilisation. Soft collars do not limit swallowing or mobility as much as rigid collars and so might reduce complications.[[3]]

Nearly 60% of our patients sustained their injury as a result of a low-energy impact, which is consistent with previous studies.[[3,4,14]] Although more patients sustained their injuries from a low-impact accident, very few of these patients (2%) had any neurological complications compared with 23% who sustained their injuries from high-energy accidents. Four out of these six patients sustained their injuries from motor vehicle accidents.

There is a lack of consensus over the management of C2 fractures in the elderly population.[[1,3,5,10,15,22]] While there is a trend for increased surgery for selected patients,[[10]] this trend is not universal[[5]] and most studies show the majority of older people are still treated non- surgically. Surgical approaches may have benefits of better immobilisation and higher bony union rates, but there can also be an increased risk of ICU admission and need for a feeding tube.  There are concerns that an operative approach increases the early mortality rate compared with non-operative management, but this has not been demonstrated consistently.[[3,14]] For these reasons, surgery has tended to be for carefully selected patients—often those who are younger and with multi-trauma or less stable fractures.[[8]] In older patients, the risks of surgery are higher, in part due to comorbidities, and reports suggest conservative management may give similar longer-term outcomes.[[3,12,23]] This current non-comparative study does not address the surgery versus no surgery debate, but it does describe functional outcomes of conservative management of C2 fractures, including mobility and domicile, as well as highlighting some issues with rigid collar immobilisation.

The complication rate was high in this cohort as all patients had at least one complication during their inpatient stay. Documented complication rates vary enormously, from 0% to 91%, for non-operative management strategies.[[3]] Most recent studies report complication rates in the range of 35% to 46% of patients,[[7,15,17,18]] rates that are lower than those found in our study. There is no consistent approach for assessing complications, which makes it difficult to compare rates between studies. However, complications are clearly common in older patients who have a C2 fracture, regardless of treatment approach (surgical or conservative).

The mortality rates were 9% at 30 days and 22% at one year for our cohort of patients, which are comparable or better than other reported mortality rates.[[3,8,11,15,16,23–25]] Functional dependency pre-fracture was the main predictor of mortality in one study,[[16]] whereas increasing age and comorbidities were predictors in others.[[11,17,24]] The unadjusted median CCI score of 2.0 found in our study predicts a one year mortality rate of 21%, which is similar to our one year mortality rate.[[21]] These data, together with the New Zealand paper by Chan et al,[[11]] suggest that late mortality (at one year) for older C2 fracture patients may be predicted more by their comorbidities and less by the C2 fracture itself. This is similar for other frailty fractures, such as femoral neck fractures, where frailty and comorbidities influence longer-term survival.[[26]]

As with other retrospective studies, there are limitations. It was a single-centre study and all patients were managed non-surgically. Inpatient rehabilitation occurred in three different units on two geographic sites, both of which are separate from the acute admitting hospital. There was less-frequent orthopaedic input in the older persons health unit at time of the study. Since then, all three units have moved to the same campus, with easy access to both orthopaedic and spinal input. Retrospective review of notes may underestimate complications, such as falls or delirium, if they were not accurately recorded at the time. Also, if discharge coding was different to S12.1, then some C2 fractures may have been missed. Three patients’ notes could not be accessed, which resulted in a slightly smaller cohort than planned. However, in mitigation of these shortcomings, this is a comprehensive review of a five-year cohort of older patients with a C2 fracture that included a longer follow-up period (two years) and recorded complications and functional outcomes such as mobility and domicile, as well as use of rehabilitation services. Bony union was not consistently recorded in outpatient records and so could not be formally assessed, but all surviving patients were followed until their fractures were stable.

This study has shown that conservative management of C2 fractures in older patients has a significant impact on patients. All patients suffered from at least one complication. Nearly three in five patients required inpatient rehabilitation, and a similar number had collar-related complications. Functional limitations, such as increased need for walking aids, were very common post fracture and rehabilitation.