The administration of intermittent parathyroid hormone affects functional recovery from trochanteric fractured neck of femur

Patients and Methods

The design was an open label, prospective feasibility and pilot study with randomised treatment allocation, comparator controlled with blinded outcome assessment and a nested qualitative sub-study as previously published by the study group.⁸ The study was approved by South West 2 Research Ethics Committee (10/H0206/34) and registered with EudraCT (2010-020081-22) and ISRCTN (03362357). Funding was provided by a National Institute for Health Research (NIHR) Research for Patient Benefit grant (PB-PG-0408-16292) and sponsored by North Bristol NHS Trust (R& I 2185).

Recruitment was undertaken from a continuous sample of patients screened to meet the inclusion criteria from admissions to the trauma and orthopaedic services at six National Health Service teaching hospitals in the United Kingdom. Patients aged 60 years or older admitted with a trochanteric hip fracture were included. Exclusion criteria included contraindications to the study medication, lack of capacity to give informed consent or the inability to comply with the study protocol. Consenting participants were randomised between the intervention group and standard care using an internet-based, computer-generated random allocation provided by Sealed Envelope (London, United Kingdom) on a 1:1 ratio in blocks of ten. Randomisation was performed by the research associate at the recruiting site following the consent of the participant.⁸ The intervention group was prescribed a daily subcutaneous injection of PTH 1-34 (Teriparatide) for 42 days, commencing within ten days of surgery. Other medications related to bone-health (except vitamin D and calcium), including bisphosphonates and denosumab, were withheld for the duration of the intervention in this group. Participants were taught by the study team to self-administer the injections, with assistance from relatives or carers if required. The standard care group received treatment according to National Institute for Health and Clinical Excellence (NICE) guidance.⁹ Adherence to this guidance was monitored as part of the study.

Standardised data collection was undertaken at screening, on recruitment to the study, on discharge from the acute hospital stay and at six and 12 weeks, and six months after surgery. Hospital systems were checked for occurrence of any further fractures within 12 months of surgery. Outcome measures including the Short Physical Performance Battery (SPPB),^10-12 the Short Form (SF)-36,¹³ EuroQol 5-D (EQ-5D)¹⁴ general health questionnaires and visual analogue scales (VAS)¹⁵ for pain on bearing weight (0 to 10 scale where 10 was the worst pain imaginable and 0 was no pain) were completed with participants by an assessor blinded to treatment allocation, at outpatient appointments at six and 12 weeks post-operatively. The general health questionnaires were repeated in a telephone consultation with participants at six months. Pre- and post-operative radiographs were used to classify the fracture according to AO/OTA system,¹⁶ ascertain the type of operative intervention, measure the amount of collapse as measured on the anteroposterior final post-operative radiograph, assess union and identify any failure of fixation.

As this was a pilot study to determine the appropriateness of the proposed methodology for a main study with adequate power, no formal sample size calculations were performed and groups of 20 participants were deemed suitable.

On this basis, statistical comparison of outcomes in the intervention and comparison group were not carried out. Instead, we report measures assessing the feasibility and acceptability of the study (such as eligibility, retention, compliance and data completeness).

Results

Of the 724 patients screened to meet the inclusion criteria, 143 (20%) were eligible for the study; of these, 123 were approached, and 29 (20%) of those eligible consented to participate (4% of the total patients screened) (Fig. 1). These numbers were significantly lower than the 50% of the population originally predicted to be eligible. Lack of capacity precluded 314 of the 581 (54%) of those screened as eligible from giving informed consent. Recruitment took place between March 2012 and March 2013.

Fig. 1

Consolidated Standards of Reporting Trials (CONSORT) diagram.

In total, 192 patients (33%) of the population were excluded due to contraindications to the intervention, such as renal failure, cancer or pre-existing metabolic bone disease. Inability to comply with the study schedule excluded 272 patients (47%) but 129 of these were already excluded for other reasons. The demographics of those screened and consented are shown in Table I.

A total of seven patients (24%) failed to complete the study. Of these, two patients died, contact was lost with one and four withdrew quoting “stress of comorbidities,” being “fed up”, “not wanting to participate further” and the burden of recovery as reasons. Only two of these withdrawals were by patients in the intervention group. The timing of withdrawal varied from two to 214 days after enrolment. A total of 26 participants (90%) completed the six week follow-up visit and 22 (76%) the 12-week and six-month visits (Fig. 1).

All patients remaining in the study at the relevant point completed the SPPB at six (n = 26) and 12 weeks (n = 22). Full completion of the EQ-5D was performed by all patients remaining at six weeks, by 21 of 26 (96%) at 12 weeks and 20 of 21 (91%) at six months. The SF-36 was fully completed by 20 of 26 (77%) patients at the six week follow-up and 17 of 22 (77%) at 12 weeks, which may reflect the more complicated nature and greater length of the questionnaire. This improved, however, to 18 of 22 patients (92%) at the six month follow-up, when this was generally completed via telephone consultation. The results at 12 weeks are reported in Table II.

Of the 15 patients randomised to the intervention group, 11 completed the course of injections. The injections were predominantly self-administered - two patients required assistance from family to deliver the injection and others reported reminders or assistance in accessing the medication, provided by their family. Of those not compliant in the intervention group, one withdrew following randomisation, one stopped with potentially related side effects (night sweats and shivering) but continued in the trial, one withdrew from the study entirely on discharge from hospital and one started the injections late and therefore received a reduced treatment duration, due to misunderstanding the instructions.

Adverse events were recorded for patients in both groups. In all 15 events were reported as being serious according to ICH-GCP definitions.¹⁷ In all, eight of these were in the intervention group and seven in the comparison group. None of the serious adverse events were related to the study intervention and all may be expected in this patient population.

A total of six patients sustained AO/OTA A1 fractures, 15 A2 fractures and four A3 fractures. The implants used were 25 sliding hip screws and four cephalomedullary nails. There were no follow-up radiographs for four patients. Radiographs were available for the other 25 patients at six weeks and 21 had further images at 12 weeks. All fractures had united at 12 weeks with ten patients demonstrating no collapse of the fracture, nine up to 5 mm collapse and six up to 10 mm collapse. No fracture had collapsed more than 10 mm. There were no complications reported regarding fracture fixation.

A total of 13 of the 14 patients in the comparison group were managed in line with NICE guidelines,⁹ while patient 14 died one month into the study. Five of these participants were already taking osteoporosis medications on admission. At six weeks, 17 of 26 (65%) of patients reported taking their calcium and vitamin D supplements as prescribed. Six of eight patients (75%) under 75 years of age on entry to the study were referred for a Dual-energy X-ray absorptiometry scan in line with NICE guidance. One of the remaining two had undergone a recent scan and the other died a month into the study.

Discussion

PTH has been shown to improve bone mineral density (BMD) and reduce subsequent fractures in those with osteoporosis.¹⁸ Despite evidence of benefit from animal studies in the treatment of fractures^19-23 the dose, duration and cost effectiveness of treatment remain in question.²⁴ Subcutaneous injection remains the only licensed route of administration²³ and although other delivery systems are under investigation, they are not yet licensed.²⁵ While it is increasingly frequently prescribed, evidence from human studies is lacking.⁶

Patients with trochanteric fractures were chosen, as hip fractures represent the most costly fracture with large socio-economic costs.²⁶ Assessing union and measuring an impact on the recovery of function in lower limb fractures is very difficult²⁷ and so this study tested the option of a surrogate measure such as the SPPB.^10-12 This requires patient contact and can be influenced by other physical comorbidities, which are common in this frail group of patients.

This study has highlighted important challenges that should be considered when designing research studies in patients with hip fractures. These challenges could be extrapolated to any frail, elderly trauma patient population where the combination of age, comorbidities and an acute injury requiring surgery present a unique situation in which to conduct research. The greatest challenge of delivering this study was that of identifying a sufficient number of eligible patients and gaining informed consent in this elderly population, with only 4% of those 724 patients who met the inclusion criteria sustaining trochanteric fractures being both eligible and willing to consent.

If a patient was recruited, the protocol has been shown to be deliverable across multiple sites. Once enrolled and taught the injection technique, the compliance and adherence to the injection intervention was good. Data completion to six weeks was good but dropped at the later time points.

As a non-blinded study it was important to define the care received by the comparison group. Since the design of the study, the implementation of a national tariff through the National Hip Fracture Database has standardised the care of this patient group.²⁶ Surrogate markers of the outcome of rehabilitation, such as discharge destination and length of stay were successfully recorded. It was difficult, however, to obtain measures of other aspects of rehabilitation such as the intensity of therapy and social support required, due to lack of access to and detail of recorded data.

Limitations of the study included its pilot nature, and the small groups which precluded objective assessment of the value of PTH in this injury. Recruitment was only open to those with no cognitive impairment, no contraindications to PTH and the ability to comply with the study protocol, which excluded 80% (581) of the patients screened. Follow-up of these patients was difficult and seven (24%) failed to complete the study. This makes the results difficult to generalise to the whole population.

With these difficulties in mind, expansion into a large multicentre trial should be undertaken with caution. Any definitive study recruiting a representative cohort would need to include the population unable to give informed consent. Sample size calculations using the data from this study based on detecting a one-point change in the SPPB at 12 weeks as the primary outcome, assuming an 80% completion rate, would require 405 patients. Any design should, therefore, include strategies to maximise retention and completion such as follow-up in the community or by telephone.

If PTH is given to aid bone healing in trochanteric fractures, it should still be considered experimental.