VERTOS-II. Percutaneous Vertebroplasty Versus Conservative Therapy

Design:

This will be a multicenter (6 centers), prospective, randomized, intervention study. Eligible patients will be randomized for percutaneous vertebroplasty or analgesics.

Study population:Two hundred patients will be included.

Patient selection:

Patients will be referred from the GP to the radiology department for acute back pain. The GP will order an X-ray of the thoracic and lumbar spine to establish an osteoporotic vertebral fracture(s). Subsequently, the hospital radiologist will pick out the cases with vertebral fracture(s). Subsequently, the general physician/ geriatrician will confirm the diagnosis of 'osteoporotic' vertebral fracture, and exclude other causes of vertebral fracture (e.g. tumour, major trauma). Finally, the patient and their GP will be asked to participate in our study.

Eligible patients will be randomized for percutaneous vertebroplasty (and pain medication if necessary) or only analgesics.

Procedure:

The treatment consists of a transpedicular injection of polymethylmethacrylate bone cement into the collapsed vertebral body (maximal 2 vertebrae in 1 treatment session) using fluoroscopic guidance. The procedure is performed under the use of local anesthetics and the patient can be discharged the same day or after 1 night of hospital stay.

Conventional treatment (analgesics):

The internist optimizes the use of analgesics in ascending order: (1) Paracetamol, (2) Tramadol, (3) Tramadol and Paracetamol, (4) Morphine.

All patients receive osteoporosis medication.

Outcome parameters:

The primary outcome for the clinical study as well as the economic evaluation will be adequate relief of pain. Pain intensity will be measured by means of a visual analogue scale (VAS). Relief of up to 3,0 to 4,4 scale points (scale 0-10) is associated with a clinically relevant pain reduction (Beurskens et al 1995). The main goal of the economic evaluation is to assess the balance between costs and effects of percutaneous vertebroplasty and pain medication as compared to usual care (analgesics, physiotherapy, visits to general practitioner) after 4 weeks and after 1 year.

The clinical research objectives are to compare the effects (1 day,1 week, 1 month, 3 months, 6 months,12 months and 24 months) of vertebroplasty with conventional therapy on a) back pain, b) activities of daily living (ADL), c) mood, d) general health and e) quality of life in patients with painful osteoporotic vertebral fractures. Other objectives are to study the 1 and 2-year incidence of recurrent fractures, especially adjacent to treated levels of

Measurements:

a), b), c), d) and e) are measured by using the visual analogue scale (VAS) for pain intensity, the Roland Disability Questionnaire (RDQ) for functional status, while, the EQ6D and the Qualeffo-41 focus on quality of life, mood and general health. Data will be collected about units of resource utilization with an explicit clause asking whether it can be related to the vertebral fracture.

(New) osteoporotic vertebral fractures are assessed at baseline by conventional radiography and magnetic resonance (MR) imaging. The incidence of recurrent fractures will be determined by conventional X-rays at 1,3,12 and 24 months.

Economic evaluation:

The goal of the economic evaluation is to assess the balance between costs and effects of vertebroplasty as compared to usual care.

We will initially perform a cost-effectiveness analysis with a one month time horizon. Given immediate pain relief in a majority after vertebroplasty, already a cost-effectiveness ratio can be calculated.

Estimates of other relevant outcomes such as costs and quality of life will also be entered in the model to allow a full economic evaluation. We will estimate incremental costs per additional year freed of pain of immediate vertebroplasty as compared to vertebroplasty after one month and (synthetic scenario of) care as usual. As the latter does involve extrapolation using estimates not actually observed in the trial we can not use bootstrap simulation to assess uncertainty. Accordingly, multivariate probabilistic sensitivity analysis (Monte carlo simulations) will be used to evaluate uncertainty in the cost-effectiveness ratios. As we will take into account a time horizon up till one year costs nor effects will be discounted. Finally, a cost-utility analysis is foreseen based on utility scores obtained over the first month of observation after randomization.

The direct medical costs will be estimated from a societal perspective. This implies that resource use such as GP visits, analgesics, physiotherapy and consultation of orthopedic surgeons or general physicians will be recorded in the CRF and in patient diaries. In parallel, actual costs of the resources, i.e., unit costs, will be estimated. The majority will be based on the estimates gathered in the Dutch guidelines for economic evaluation published by CVZ. Subsequently, multiplication of units costs with resource use will yield cost estimates on an individual patient level. With regard to costs due to losses in productivity we pose that the majority of patients does not have a paid job. Time spent by spouses, family and friends taking over certain household activities or other unpaid activities will be accounted for using shadow prices.

The primary outcome for the economic evaluation will be adequate relief of pain. Pain will be measured by means of a visual analogue scale (VAS). Relief of up to 3,0 to 4,4 scale points is considered adequate. Time until this outcome is attained will explicitly be accounted for in the economic evaluation.

In addition to pain, health related quality of life will be measured. Subsequent summation over the period of follow up will yield overall QALYs up to a year for each arm of the trial.

In the cost effectiveness analysis we will take care of the differences in the Belgium and Dutch healthcare system.

Statistical analyzes All analyzes will be performed on an intention to treat principle. In addition, analyzes will on the different outcomes will be compared between the two groups with the Student's t-test for continuous normally distributed variables, with the Mann-Whitney test for variables that are not normally distributed and with the chi² test for categorical variables. In addition, subgroups (e.g. long/shot duration of complaints) will be examined in which vertebroplasty has a more (or less) pronounced effect by introducing interaction terms in linear regression models with change in pain and quality of life as outcome variables.

The one-year cumulative incidence of new fractures adjacent to the treated/old fracture levels will be compared between patients who were treated with vertebroplasty and those who were not with Cox-proportional hazards analyzes, adjusting for confounders.

Power calculations:

The planned number of subjects will be 200 in total, 100 in each intervention arm. This number is based on conventional assumptions of alfa=0.05 and ß=0.20, withdrawal from intervention of 20%. Based on pilot data and literature we expect a difference of 25% in significant pain relief. If we assume that 20% withdraws from intervention we need approximately 100 patients in each group.

Time schedule:

The total study will take 36 months.

Recruitment of patients (200) has been completed. In May 2009 the 1 year follow-up will be finished.