We analysed integrated outcomes and follow-up data from two randomised, double-masked, placebo-controlled, multicentre, trials done at a total of 28 academic referral tertiary specialised centres offering joint thyroid eye clinics, or orbital clinics or practices, or both, in Europe and the USA. Participants were adult patients with a diagnosis of
Graves' disease and active moderate-to-severe
thyroid eye disease (clinical activity score [CAS] ≥4). Patients received eight
intravenous infusions of either
teprotumumab (10 mg/kg
body weight for the first infusion, 20 mg/kg for subsequent infusions) or placebo every 3 weeks. The final study visit was at week 24, 3 weeks after the final infusion. In our analysis, the prespecified primary outcome was the between-group difference from baseline to week 24 in the proportion of patients with a
proptosis response (≥2 mm reduction in the study eye without similar deterioration in the fellow eye at week 24) stratified by tobacco non-use and current use. Secondary endpoints at week 24 were the proportion of patients with improved
diplopia (≥1 Bahn-Gorman grade), an overall response (reduction of ≥2 mm in
proptosis and reduction of ≥2 points in CAS), mean change from baseline in
proptosis measurement in the study eye, mean change from baseline in
Graves' ophthalmopathy quality of life (GO-QOL) questionnaire scores (overall, visual functioning, and appearance), and the proportion of patients with disease inactivation (ie, a CAS score of 0 or 1). We also assessed data for the primary and secondary outcomes by patient subgroups (tobacco use; age <65 years or older; sex; time to diagnosis; CAS score 4 or 5, or 6 or 7; and
thyrotropin binding inhibiting
immunoglobulin [TBII] concentration <10 IU/L or ≥10 IU/L) versus placebo. Additional outcomes included short-term and long-term responses at 7 weeks and 51 weeks after the final dose, and post-hoc assessments of disease severity (more severe baseline disease defined as
proptosis ≥3 mm or constant or inconstant
diplopia, or both, as compared with all others), and an ophthalmic composite outcome (improvement in ≥1 eye from baseline without deterioration in either eye in ≥2 of the following: absence of eyelid swelling; CAS ≥2;
proptosis ≥2 mm; lid aperture ≥2 mm;
diplopia disappearance or grade change; or improvement of 8 degrees of globe motility). All outcome endpoint analyses were done by intention-to-treat (ITT) except where noted.
FINDINGS: The pooled ITT population consisted of 84 patients assigned
teprotumumab and 87 assigned placebo. More patients receiving
teprotumumab achieved a reduction of at least 2 mm in
proptosis at week 24 versus placebo (65 [77%] of 84 patients assigned
teprotumumab vs 13 [15%] assigned placebo; stratified treatment difference 63%, 95% CI 51-75; p<0·0001). Numbers-needed-to-treat (NNT) were 1·6 for
proptosis response, 2·5 for
diplopia response (treatment difference 39%, 95% CI 23-55), 1·7 for overall response (treatment difference 60%, 48-72), and 2·5 for disease inactivation (treatment difference 40%, 27-53); all p <0·0001. The post-hoc assessment of the composite outcome showed that it was reached by 68 (81%) patients in the
teprotumumab group and 38 (44%) in the placebo group (NNT 2·5, treatment difference 40%, 95% CI 26-53; p<0·0001). There were significantly more
proptosis responders with
teprotumumab in all subgroups at week 24; the number of
diplopia responders was also significantly higher with
teprotumumab for all subgroups except tobacco users and patients with TBII less than 10 IU/L at baseline. Integrated treatment differences for
proptosis ranged from 47% in tobacco users (95% CI 21-73, p=0·0015; NNT=2·1) to 83% in patients aged 65 years and older (60-100, p<0·0001; NNT=1·2), and for
diplopia ranged from 29% in tobacco users (95% CI -3 to 62, p=0·086; NNT=3·4) to 47% in those with baseline CAS of 6 or 7 (95% CI 23-71, p=0·0002; NNT=2·1). All other integrated subgroup results were p≤0·033. Integrated responses were observed at 7 weeks and 51 weeks after final dose for
proptosis in 62 (87%) of 71 patients and 38 (67%) of 57 patients respectively; for
diplopia in 38 (66%) of 58 and 33 (69%) of 48 respectively; and for the composite outcome in 66 (92%) of 72 and 48 (83%) of 58, respectively. During the 24-week study, compared with placebo, there were moderate-to-large improvements with
teprotumumab for GO-QOL total scores (19 vs 6, p<0·0001), visual scores (20 vs 7, p=0·0003), and appearance scores (18 vs 6, p=0·0003), respectively, which were maintained during follow-up. Of all adverse events during the treatment period, 63 (94%) of 67 patients with
teprotumumab and 59 (98%) of 60 patients with placebo were mild to moderate (grade 1 or 2), with three (4%) serious adverse events related or possibly related to
teprotumumab of diarrhoea, infusion reaction, and
Hashimoto's encephalopathy (co-incident with
confusion) leading to study discontinuation. Of the most commonly reported adverse events with
teprotumumab,
muscle spasm (18%, 95% CI 7·3-28·7),
hearing loss (10%), and hyperglycaemia (8%, 1·7-15·0) had the greatest risk difference from placebo.
INTERPRETATION: