Treating Idiopathic Pulmonary Fibrosis with the Addition of Co-Trimoxazole: An Economic Evaluation Alongside a Randomised Controlled Trial

Abstract

Background

Idiopathic pulmonary fibrosis (IPF) is a fibrotic disease of the lungs of unknown origin with a poor prognosis. A small trial of co-trimoxazole demonstrated improvements in symptoms and functional parameters over a 3-month period. We therefore conducted a larger trial with a concurrent economic evaluation to investigate this antibiotic further.

Methods

We report an economic evaluation alongside a multi-centre, randomised, placebo-controlled, double-blind trial of 12 months therapy with 960 mg co-trimoxazole daily in 181 patients with fibrotic idiopathic interstitial pneumonia (IIP). Patients were recruited from 28 university and district hospitals in the UK and were aged over 40 years with fibrotic IIP. We report costs to the National Health Service (NHS) and society, change in forced vital capacity (primary endpoint) and quality-adjusted life-years (QALYs) gained, incremental cost effectiveness and cost utility ratios over 12 months.

Results

From the perspective of society, mean cost per patient in the co-trimoxazole arm was approximately £1177 higher than in the placebo arm, but mean QALYs were 0.053 higher yielding an incremental cost-effectiveness ratio of £22,012 per QALY gained with a 54.44 % probability of being below £30,000. The cost of IPF to UK society in 2011 is tentatively estimated at £124 million, of which 13 % is NHS costs, 1 % social services, 2 % patient out-of-pocket costs and 84 % lost productivity.

Conclusions

Given commonly employed thresholds in the UK NHS, on balance co-trimoxazole may be a cost-effective treatment for IPF, although there is substantial decision uncertainty. However, recent guidance on the use of immunosuppressive therapy in IPF patients should be taken into account prior to any policy decision.

Appendix

1.1 Details of Imputation

Missing data were imputed in STATA using the ‘ice’ command at the level of total NHS and societal costs, QALYs and change in FVC at 12 months. Explanatory variables were baseline NHS and societal cost, utility and FVC, whether the patient was prescribed prednisolone, age left education, marital status, employment status and intervention arm. Five datasets were generated, each based on 1000 markov chain monte carlo (MCMC) iterations. The code is provided below:

1.2 Details of Adjustment to Baseline Values

The imputed, adjusted increments were calculated as follows:

• The imputed datasets resulting from the ‘ice’ command above were resampled with replacement (thus five ‘new’ datasets were generated).

• OLS regressions were used to estimate NHS cost, societal cost, QALYs and change in FVC for each of the ‘new’ datasets, using treatment group and baseline NHS or societal cost, utility or FVC as explanatory variables respectively.

• The coefficients on treatment group were recorded as the adjusted incremental cost or outcome respectively.

• Steps 1-3 were repeated 1000 times, thus generating 5000 sets of adjusted incremental cost and outcomes (1000 from each imputed dataset).

• The percentile method was used to calculate 95 % confidence intervals around point estimate increments.