Number of flare events per month during Period III (calculated as number of flares divided by number of months participant was enrolled during Period III). Flare was determined using pre-defined criteria, using an interactive voice response system.

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication

Time from first dose of double blind medication (start of Period III) to occurrence of first OA flare. Flare was determined using pre-defined criteria, using an interactive voice response system

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication.

Number of days subject was free from OA flare divided by number of days on study medication in Period III. Flare was determined using pre-defined criteria, using an interactive voice response system.

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication

Number of days subject was in OA flare divided by number of days on study medication in Period III. Subjects may have more than one flare. Flare was determined using pre-defined criteria, using an interactive voice response system.

Participant rated intensity of osteoarthritis pain on categorical scale from 0 (no pain) to 10 (worst pain). Scores analyzed as area under the curve (AUC) of participant's scores from each assessment in Period III.

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication.

These data are presented by the weeks from the start of Period II, 2 weeks before randomization. The weeks post-randomization, Period III, are different from the study weeks i.e. includes 2 weeks from Period II.

Participant's response to question "Considering all the ways the osteoarthritis in your hip or knee affects you, how are you doing today?" on scale from 1 (very good) to 5 (very poor). Scores analyzed as area under the curve (AUC) of participant's scores from each assessment in Period III.

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication.

These data are presented by the weeks from the start of Period II, 2 weeks before randomization. The weeks post-randomization, Period III, are different from the study weeks i.e. includes 2 weeks from Period II.

Physician assessed each participant's disease symptoms on a categorical scale from 1 (very good) to 5 (very poor).

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication

Total amount of rescue medication (acetaminophen in milligrams [mg]) taken per month per participant

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication.

Subjects who did not take rescue medication were assumed to have taken 0mg and were included in the analysis.

Number of subjects taking rescue medication: continuous use n=220; intermittent use n=239.

Days on rescue medication divided by number of days on study medication in Period III

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication. Number of subjects taking rescue medication: continuous use n=220; intermittent use n=239.

Subjects who did not take rescue medication were calculated as 0 and included in the analysis.

Number of days on flare medication per month per subject calculated as number of days on flare medication divided by the number of days on study medication in Period III

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication.

Subjects who did not take flare medication were calculated as 0 and included in the analysis.

Number of subjects taking flare medication: continuous use n=282; intermittent use n=339.

Score at end of Period III minus score at start of Period III. WOMAC assesses subject responses to 24 components regarding subscales of pain, stiffness and physical function (score range: 0=none to 4= extreme). Total score is sum of the 3 subscale scores. Negative change indicates improvement.

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication. Number of subjects evaluable: continuous use Period III start n=428, end n=427; intermittent use Period III start n=424, end n=424

Subject assessment on 7 sleep associated categories. Raw scores are transformed to a 0-100 scale. Higher score indicates more of the outcome (e.g. more snoring, more adequate sleep). Score at end of Period III minus score at start of Period III.

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication. Subjects assessed per scale n=continuous use (cont); n=intermittent use (inter)

WOMAC assesses subject responses to 24 components regarding subscales of pain, stiffness and physical function (score range: 0=none to 4= extreme). Total score is sum of the 3 subscale scores. Scores analyzed as area under the curve (AUC) of participant's WOMAC scores from each assessment in Period III.

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication.

Number of subjects evaluable: continuous use Period III start n=428, end n=427; intermittent use Period III start n=424, end n=424

Transformed score scale: 1=optimal; 0=not optimal; mixed = both optimal and non-optimal sleep during Period III

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication

SF-12v2 is a 12 item health survey covering 7 topics. Raw scores are transformed to a 0 to 100 scale. Higher scores indicate better state of health. Score at end of Period III minus score at start of Period III.

Intent to treat (ITT) population included subjects who were randomized and received at least one dose of double blind study medication.

Subjects assessed per scale n=continuous use (cont); n=intermittent use (inter)

Serious adverse events occuring during the 2 week run-in period (Period II) when all participants were dosed with celecoxib 200 mg daily

1197 participants entered the open-label run-in (period II) to allow observation of successful treatment of an osteoarthritis flare. 875 participants were randomized to double blind treatment (period III). 322 participants were not randomized.