Approximately 2 million women in the UK have osteoporosis also called brittle bone disease. It causes painful and often life-threatening fractures after minor injuries such as falling while walking. Medicines like bisphosphonates make bones stronger and prevent fractures. However, their prolonged use prevents bone healing e.g. after dental surgery and causes bones to stiffen up and break on their own. Such breaks mostly occur in the thigh. To prevent these side-effects, treatment is often interrupted for 1-2 years after initial treatment for several years. However, such drug holidays can themselves cause fractures due to osteoporosis.
This study will examine the relation between osteoporotic fracture and preceding duration of pause in bisphosphonate prescription after an initial course of treatment for three and five years respectively. It will also explore the rate of thigh bone fracture and poor healing of jawbone after dental surgery during long-term bisphosphonate treatment.
Anonymised data originated from the NHS will be used. Data will be obtained from primary-care, hospitalisation, and death records. The effect of having a fracture with a one-year, two-years, or longer pause in bisphosphonate treatment after an initial treatment for three and five years will be calculated compared to continued treatment. Next, people treated with bisphosphonates for three or five years will be followed up electronically to find out the rate of thigh-bone fracture or poor healing after dental surgery.
These findings will aid shared decision making between patients and health professionals about duration of drug holiday after the initial treatment with bisphosphonates.
Background: Osteoporosis causes low-trauma fractures and is mostly treated with bisphosphonates. They suppress bone-turnover, increase bone mineral density, and prevent osteoporotic fractures. However, long-term suppression of bone-turnover is associated with atypical femoral fractures and jaw osteonecrosis. Pausing treatment can be associated with increased risk of osteoporotic fractures. There is conflicting guidance on the duration of initial bisphosphonate treatment and the length of the following treatment pause to minimise these risks.
Objectives:
Work-package-1: To examine the association between osteoporotic fractures and preceding duration of interruption in bisphosphonate prescription, after initial treatment for three-years and five-years respectively, compared to continued treatment.
Work-package-2: To examine the incidence of atypical femoral fractures and osteonecrosis of jaw after three-years and five-years of bisphosphonate treatment.
Methods: Data from the Clinical Practice Research Datalink (CPRD) Aurum linked to hospitalisation and mortality records will be used. The study will span from 01-Jan-2007 to 31-Dec-2022.
Work-package-1: Separate cohorts of patients aged ≥18 years, prescribed bisphosphonate for ≥3-years and ≥5-years will be ascertained and followed-up to the occurrence of osteoporotic fracture or being censored. Cases will be participants that experienced an osteoporotic fracture after completing at-least three years (study-1) and at-least five years (study-2) of bisphosphonate treatment. Four contemporaneous age and sex matched controls from the cohort without an osteoporotic fracture will be selected using risk-set matching. Participants will be defined as having had or not had a treatment pause in each of the preceding 12-months before the index date. Multivariable conditional logistic regression will be used to assess the association.
Work-package-2: The cohorts in work-package-1 will be followed-up from after three or five years of bisphosphonate prescription to the earliest of osteonecrosis of the jaw, atypical femoral fracture, drug-discontinuation, death, date of last data collection or study end. The incidence (95% confidence interval) of each outcome will be calculated.
Work-package-1: Nested case-control study.
Primary outcomes: Osteoporotic fracture after completing at-least three years (study-1) and at-least five years (study-2) of bisphosphonate treatment.
Osteoporotic fracture will be defined using a comprehensive list of fractures with their primary-care medical code (for CPRD) and ICD-10 terms and codes (for HES and ONS). For Snomed codes, this will include diagnostic codes (e.g. Fracture of thoracic vertebra) and procedure codes (e.g. Vertebroplasty of fracture of spine). Procedure codes will be used to identify the occurrence of a fracture in the absence of a diagnostic code for fracture.
The list will include an inventory of fragility fractures. As it is very difficult to determine the mechanism of fracture from the patients’ clinical records, as customary in epidemiological studies, osteoporotic fractures will be defined by the anatomical sites that are generally affected. Consistent with current definitions and previous studies [27, 28], major osteoporotic fractures will include those at hip, wrist, vertebra, and shoulder. Atypical femoral (sub-trochanteric) fractures will be excluded from this outcome.
Work-package-2: Outcomes: Atypical femoral fracture, osteonecrosis of jaw. They will be ascertained using ICD-10 codes and primary-care medical codes.
Abhishek Abhishek - Chief Investigator - University of Nottingham
Georgina Nakafero - Corresponding Applicant - University of Nottingham
Sara Muller - Collaborator - Keele University
Tricia McKeever - Collaborator - University of Nottingham
Victoria Welsh - Collaborator - Keele University
Zoe Paskins - Collaborator - Keele University
HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation