Background/Purpose:

To calculate the sample sizes needed in future randomized controlled trials (RCTs) testing a therapeutic intervention in pSS, methodologists usually rely on data obtained from previous studies. To evaluate the reverse influence of each usual criteria on the number of patients needed to demonstrate the efficacy of rituximab in pSS, we have done a post hoc analysis on the multicenter, randomized, double-blind, placebo-controlled trial TEARS.

Methods:

122 Patients were assigned to receive either RTX infusions (1g) or placebo (P) at weeks 0 and 2. All patients fulfilled the new American-European Consensus Group criteria for pSS, had an active disease as assessed by mean values of the 2 highest visual analog scales (VAS) ≥50 evaluating dryness, pain, fatigue and global, and had either a recent and a biologically active pSS or at least one extra-glandular manifestation. Results concerning the primary end point (improvement of at least a 30 mm on 2 of 4 VAS) and secondary end points have shown a statistically improvement of the disease on global VAS score (continuous data) but not on the primary criteria and discrete variables, excepting VAS sicca). We also evaluated in a single centre (Brest, France) the echostructural changes (parotid and submandibular) using a semi-quantitative scoring (0 to 4). Here, we report the impact of various end points on the sample size calculated using EpiInfo. Analysis was made only when the proportion of patient improved in the RTX group was higher to the one obtained in the placebo group.

Results: 

The table shows the proportion of patients improved depending on the chosen cut-off of four VAS evaluating dryness, pain, fatigue and global disease or ultrasound and the number of patients needed to demonstrate a significant difference considering that the same proportions should be observed in a future study. At week 24, the more important clinical differences between the P group and the RTX group were observed with sicca VAS; the cut-off chosen between 20 or 30 mm gave quite similar results. ESSDAI improvement ≥was observed in 9/54 (17%) in the P group vs 12/61 (20%) in the R group. Ultrasonographic changes seemed to be the most sensitive end point in terms of sample size needed to demonstrate an effect of RTX. Nevertheless, the US changes were not associated with the mean VAS improvement registered between W0 and W24 and the number of patient evaluated was low.

VAS improvement	≥10 mm P vs R;  sample needed	≥20 mm P vs R -sample needed	≥30 mm P vs R  sample needed
Disease	18/53 (34%) vs 31/60 (52%)  N : 258	13/53 (24%) vs 21/60 (35%) N :574	12/53 (23%) vs 9/60 (15%)
Pain	21/54 (39%)  vs 26/60 (43%) N : 4844	18/54 (33%) vs 15/60 (25%)	13/53 (24%)  vs 8/60 (13%)
Fatigue	15/54 (28%) vs 29/60 (48%) N :204	9/54 (17%) vs 17/60 (28%) N :486	5/53 (9%) vs 11/60 (18%) N : 494
Sicca	15/53 (28%) vs 32/60 (53%)  N:136	11/53 (21%)  vs 24/60 (40%) N :204	6/53 (11%) vs 16/60 (27%) N:212
Number of VAS improved ≥30 mm	≥1	≥2	≥3
	26/59 (44%) vs 30/63 (48%) N:4972	16/59 (27%) vs 18/63 (29%) N: 16024	13/59 (22%) vs 12/63 (19%)
Improvement of the ultrasonographic grade	≥1 grade
	1/14 (7%) vs 7/14 (50%)  N : 42

Conclusion:

In future RCTs in pSS, ultrasound salivary gland changes could represent the most sensitive end point for pSS study, followed by sicca and fatigue. Nevertheless, ultrasound improvement was not associated with any clinical improvement. Change in a single VAS can be used but assesses only one domain of the patients’s complaints. Change in ESSDAI could be the most logical end-point reflecting all the domains of activity of the disease but its change appeared low in the TEARS study.

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/impact-of-different-end-points-on-the-patient-cohort-size-needed-to-demonstrate-the-efficacy-of-a-therapeutic-intervention-in-pss-a-post-hoc-analysis-of-the-tears-study-tolerance-and-efficacy-of-rit/