Evidence Based Medicine
June 5, 2009
Christine Williams
CASE: Mr. Jones, a 45 yo WM with history of GERD underwent a screening EGD and path demonstrated high-grade dysplasia. He talked to GI and they recommend a relatively new procedure for treatment: RFA rather than esophagectomy or photodynamic therapy. He wants to know if you agree. Luckily, you find a great article just published on this very topic while perusing NEJM over your morning coffee.
Radiofrequency Ablation in Barrett’s Esophagus with Dysplasia
NJ Shaheen, P Sharma, et al. NEJM. 28 May 2009, Vol 360, No 22, p 2277-2288
Background: About 10% of patient’s with chronic reflux have Barrett’s esophagus (intestinal metaplasia) with an overall prevalence of ~1.6%. Most cases do not progress beyond nondysplastic metaplasia or low-grade dysplasia. For those with high-grade dysplasia, the risk of esophageal cancer may be >10% per year. RFA is a newer technology for the treatment of Barrett’s for which outcomes are still being assessed.
Study Design/Participants: Multicenter, randomized, sham-controlled trial with concealment of study-group assignments during data analysis. 127 participants from 19 tertiary care centers with dysplastic Barrett’s esophagus
Inclusion Criteria: Age 18-80 y with endoscopically evident, non-nodular dysplastic Barrett’s esophagus <8cm in length. Patients with high-grade dysplasia had no lymphadenopathy or esophageal wall abnormalities on endoscopic ultrasound within the last 12 months
Exclusion Criteria: Age <18 or >80, pregnancy, active esophagitis, esophageal stricture precluding passage of the endoscope, history of esophageal cancer, varices, coagulopathy or life expectancy <2 yrs (judged by site investigator)
Methods: Randomization was stratified to four groups: low- and high-grade dysplasia and length <4cm or between 4-8cm of metaplasia. All patients were placed on esomeprazole 40mg bid, underwent EGD with measurement of the esophageal inner diameter and were randomized by a computer-generated block-randomization sequence to obtain a 2:1 intervention vs control group assignment.
Sham group: study catheter was removed
Ablation group: Radiofrequency ablation at baseline then +/- 2, 4, and 9 months
- 3cm long bipolar electrode array (12J or 40W/cm2) applied in 3cm long increments initially
- Focal RFA for areas of residual metaplasia at subsequent visits
All patients underwent biopsies based on grade of dysplasia which were all assessed by a study pathologist at a central laboratory (assumed to be blinded to study-group assignment)
- Low-grade at 6 and 12 months
- High grade at 3, 6, 9, and 12 months
Outcomes:
Primary:
1. Proportion of pts w/low-grade dysplasia w/complete eradication of dysplasia at 12 mos
2. Proportion of pts w/high-grade dysplasia w/complete eradication of dysplasia at 12 mo
3. Proportion of all pts with complete eradication of intestinal metaplasia at 12 mos
Secondary:
1. Proportion of pts with progression of dysplasia
2. Proportion of biopsy samples free of intestinal metaplasia at 12 mos
3. Discomfort level via diary of daily symptoms
4. Adverse events
Statistical Analysis:
- Power of 80% to detect a difference of 50% in eradication of dysplasia and 45% in eradication of metaplasia based on a two-sided test w/a significance level of 0.05
- Intention to treat analysis with logistic regression used to control for risk factors and assess for predictors in response to therapy
- Secondary per-protocol analysis in those not lost to follow-up
- Fisher’s exact test and Student’s t-test for baseline variables
- Fisher’s exact test for differenced in eradication of dysplasia/metaplasia
- Wilcox on rank-sum test for chest-pain scores (non-normal distribution)
Results: (see figure 2 and table 1. pg 2281/2 for enrollment/baseline characteristics)
Primary outcomes: summarized in Table 2 (p 2284)
- RFA was highly effective in eradication of dysplasia both in the low-grade (38/42) and high-grade (34/42) groups and only 5/22 and 4/21, respectively in the control groups.
- Complete eradication of intestinal metaplasia (all pts) occurred in 65/84 with RFA and 1/43 in control group
Secondary outcomes:
- Eradication of intestinal metaplasia occurred at a high rate in low grade (34/42) and high-grade (31/42) ablation groups and 1/22 and 0/21 respectively for controls.
- Eradication of dysplasia (all) was 72/84 in RFA and 9/43 in controls
- Progression of dysplasia was less likely in RFA group 3/84 and 7/43 in controls. This effect seemed higher in the high-grade to cancer subgroup
- Biopsy specimens free of metaplasia were higher in both groups (low-grade 1228/1260 and 1442/1464 and 313/550 and 360/614 for RFA and control
- Chest pain occurred in in many of the RFA group but resolved within 8 days (median pain score report of 23 on 100 point scale on day1, 0 in controls)
Validity:
-Was assignment of patients randomized: yes
-Was the randomization concealed: yes, to investigators when reviewing data set and presumably to pathologists. Procedure obviously can’t be blinded.
-Was follow-up of patients sufficiently long: Debatable—would argue NO
-Were all patients analyzed in groups to which they were randomized? Yes (int. to treat)
-Were the groups treated equally other than experimental rx? Yes
-Were the groups similar at the start of the trial: Higher BMI in pts w/high grade dysplasia among controls
Strengths:
- randomized, sham-controlled intervention
- intention to treat analysis
- most patient’s did complete all follow-up
- likely translatable to our patient care population (tertiary care center)
Weaknesses:
- overall number of patients low
- race and sex limited (almost all patients white males)
- follow-up limited to 1 year which may predispose to seeing a more robust response
- use of surrogate markers for effect on morbidity and mortality
- unable to blind completely given nature of intervention
- higher BMI among patients with high-grade dysplasia in control group
- the sponsor managed the database
- many possible conflicts of interest with investigators
- not powered to evaluate disease progression/morbidity/mortality
Discussion:
Low NNT for RFA to eradicate dysplasia and intestinal metaplasia at 12 months based on these results with a low rate of adverse effects but with a very limited time span to evaluate for other concerns down the road like QOL or esophageal strictures. It is unclear if such an aggressive treatment and monitoring would be truly beneficial for morbility and mortality particularly for those with low-grade dysplasia which is much less likely to progress to cancer over 12 months than high-grade dysplasia (<1% vs 10%). Overall, this is a very well planned and executed study which is highly favorable to RFA for Barrett’s esophagus with the multiple above mentioned caveats.
So, what should you tell Mr. Jones after your recommendation that he quit smoking and drinking?