JCHR

 EXTENSION STUDY FOLLOW-UP

Background – Protocol T

Between August 2012 and August 2013 DRCR.net randomized 660 participants into Protocol T, a comparative effectiveness randomized trial of aflibercept, bevacizumab, or ranibizumab for eyes with decreased visual acuity from diabetic macular edema.  The trial reported primary 1-year results in 2015 and the final 2-year results in 2016. Participants were followed for a total of 2 years.  Treatment with the assigned anti-VEGF continued through 2 years based on a structured retreatment algorithm.  The two-year visit was completed by 578 (88%) participants.  Of the 82 participants who did not complete the 2 year visit, 24 had died.

The median numbers of injections were 5, 6, and 6 in year 2 and 15, 16, and 15 over 2 years in the a?ibercept, bevacizumab, and ranibizumab groups, respectively.  At the 2 year visit, eyes with worse baseline VA (20/50 to 20/320) (N=284) had a mean visual acuity improvement of 18.1, 13.3, and 16.1 letters, respectively (a?ibercept vs. bevacizumab, P = 0.02; a?ibercept vs. ranibizumab, P = 0.18; ranibizumab vs. bevacizumab, P = 0.18). In eyes with better baseline VA (20/32 to 20/40) (N=293), mean improvement at 2 years was 7.8, 6.8, and 8.6 letters, respectively (P > 0.10, for pairwise comparisons). In eyes with worse baseline visual acuity 25% of eyes in the aflibercept group, 54% in the bevacizumab group, and 34% in the ranibizumab group had center-involved DME on OCT at 2 years.  Among eyes with better baseline visual acuity 33%, 63%, and 36% of eyes, respectively had center-involved DME on OCT.  Anti-Platelet Trialists’ Collaboration (APTC) events occurred in 5% with a?ibercept, 8% with bevacizumab, and 12% with ranibizumab (P=0.047).

In conclusion, all 3 anti-VEGF groups showed VA improvement from baseline to 2 years with a decreased number of injections in year 2. Visual acuity outcomes were similar for eyes with better baseline VA. Among eyes with worse baseline VA, a?ibercept had superior 2-year VA outcomes compared with bevacizumab, but superiority of a?ibercept over ranibizumab, noted at 1 year, was no longer identi?ed at 2 years. Higher APTC event rates with ranibizumab over 2 years warrant continued evaluation in future trials. 

Available Data on Longer Term Outcomes

There have been several published trials on the effects of anti-VEGF therapy on DME beyond 2 years.  DRCR.net Protocol I demonstrated that on average eyes initially treated with ranibizumab maintained vision gains obtained by the first year through 5 years with minimal treatment after 3 years. At 5 years, the ranibizumab and deferred laser group gained an average of 9.8 letters from baseline even though the median number of injections in year 4 was 1 and year 5 was 0. Eyes in the ranibizumab and prompt laser group gained an average of 7.2 letters from baseline, even though the median number of injections in years 4 and 5 was 0.¹

RISE and RIDE showed visual acuity (VA) outcomes seen at month 24 in ranibizumab groups were consistent through month 36 with continued monthly ranibizumab dosing. In RIDE, the mean number of ETDRS letters change from baseline at month 24 versus change from baseline at month 36 in patients randomized to 0.3 mg ranibizumab was 10.9 versus 10.6. In RISE, 12.5 versus 13.2.^{2 2} Results from the RIDE/RISE open label extension study, during which participants who participated in the core studies were treated with ranibizumab on a pro re nata basis for DME, also demonstrated maintenance of visual gains from the first 3 years of treatment despite reduced frequency dosing.³

The VIVID and VISTA studies, which assessed efficacy of intravitreous aflibercept for DME treatment, demonstrated that participants’ 148 week and 100 week visual outcomes were consistent with the 52-week visual outcomes. . The mean BCVA gain from baseline to week 148 was 10.4 and 10.5 for the intravitreous aflibercept 2q4 regimen and 2q8, respectively in VISTA and 10.3 and 11.7 in VIVID.⁴  The mean BCVA gain from baseline to week 100 was 11.5 and 11.1 for the intravitreous aflibercept 2q4 regimen and 2q8, respectively in VISTA and 11.4 and 9.4 in VIVID.⁵ The mean BCVA gain from baseline to week 52 in VISTA was 12.5 and 10.7 and 10.5 and 10.7 in VIVID, in 2q4 and 2q8 groups respectively.⁶

Protocol T Beyond 2 Years

Bringing back Protocol T patients to complete one five year visit will provide information on treatment course, changes in visual acuity and macular edema after protocol specific treatment was stopped.  Patients who were randomized in Protocol T will be asked to return to one of the approximately 80 Protocol T clinical sites currently active in the Network for an eye exam, visual acuity, OCT, and fundus photography approximately 5 years from initial randomization date.  Medical history, diabetic retinopathy and DME treatment history will be collected.

Estimation of Number of Participants in Follow-up Study

83* of the original 88 clinical sites were surveyed to assess potential availability of participants.  Of the 588 participants at the survey sites who were not known to be deceased prior to 2 years, 364 (62%) participants were seen at the practice within the prior 6 months and 395 (67%) were seen within the prior 12 months.    The percentage of participants who died during the clinical trial was 2% in both the first and second year.  Assuming that trend increases slightly over years 3-5, we can roughly expect approximately 10% of the participants alive at the end of the 2-year randomized trial to have died between years 3-5.  Beginning with 588 participants alive at the end of the Protocol T clinical trial at active DRCR.net sites, applying a 10% death rate and 80% participation rate results in approximately 425 participants.  However, since some sites may decline participation in the extension study, the goal is to include at least 400 participants in the extension study.

*5 sites have been dropped from the Network since Protocol T completed

 Study Objectives

 The primary objective is to perform descriptive analyses for the following:

• Visual acuity outcomes at 5 years
• DME outcomes at 5 years
• Types of DME treatments used since 2 year study visit
• Frequency of DME agents used since 2 year study visit
• Treatments for diabetic retinopathy since 2 year study visit
• Diabetic retinopathy outcomes at 5 years
• APTC events occurring in participants since 2 year study visit

 Secondary analyses will include original treatment group comparisons for the following:

• Visual acuity outcomes at 5 years
• DME outcomes at 5 years
• Diabetic retinopathy outcomes at 5 years

Study Design

• Cohort study

 Eligibility Criteria

•  Study eyes of previously randomized participants in Protocol T

 Visits and Procedures

Participants would return for one follow-up visit 5 years (± 6 months) from original randomization for the following procedures to be performed:

• Assessment of DME and diabetic retinopathy treatments since 2 year visit
• Ocular and medical history since 2 year visit
• Best-corrected E-ETDRS visual acuity
• Comprehensive, dilated eye exam
• Optical coherence tomography
• HbA1c
• Color fundus photographs
• Assessment of APTC events since the 2 year visit

Sample Size

Sample size is expected to be between 400 and 425 participants.

 Eligibility

All participants randomized in Protocol T will be eligible for this extension follow-up study.

Patient Recruitment

All patients who were randomized in the Protocol T clinical trial who are not known to be deceased will be targeted for recruitment, including those who did not complete their 1 or 2 year visit in the clinical trial.  Currently active DRCR.net clinical sites that participated in Protocol T will be asked to participate in this extension study.  Participants from sites who decline to participate will not be included.  Clinical coordinators from participating sites will contact the patients who were randomized at their site to invite participation either in-person while the patient is undergoing routine care at the DRCR.net clinical site, or by letter or telephone if the patient is not receiving eye care at the clinical center.  Coordinators will make use of additional contact information provided by the patient at entry into Protocol T to attempt to reach the patient.

Follow-up Visit

Visit Schedule

Each willing participant will have one follow-up study visit, approximately 5 years (±6 months) after original randomization.

Testing Procedures

The testing procedures will be the same as in the Protocol T clinical trial and are detailed in the DRCR.net Procedures Manuals.  Visual acuity testing, ocular exam, fundus photography, and OCT will be performed by DRCR.net certified personnel.  When feasible, visual acuity, OCT, and fundus photography should be performed by personnel masked to the participant’s original treatment group assignment.

The following procedures will be performed.

• Assessment of DME and diabetic retinopathy treatments since 2 year (or last) visit for the study eye
• Ocular and medical history since 2 year visit
• Best corrected E-ETDRS visual acuity testing (including protocol refraction) in each eye
• Ocular examination on each eye including slit lamp, measurement of intraocular pressure, lens assessment, and dilated ophthalmoscopy
• Spectral domain OCT on study eye using Zeiss Cirrus or Heidelberg Spectralis

·         The same OCT machine used at the 2 year visit should be used if possible

Digital fundus photographs of the study eye

• 4 wide field and 7 modified field images are preferred

·         Ultrawide field images are accepted only if 4 wide and 7 modified are not available

• Hemoglobin A1c

·         HbA1c does not need to be repeated if available in the prior 12 months.

 Main Analyses

  The primary objective is to perform descriptive analyses for the full cohort for the following:

• Types of DME treatments used since 2 year study visit
  • Types of anti-VEGF injections
  • Focal/grid laser treatments
  • Types of other DME treatments
• Frequency of DME agents used since 2 year study visit
• Treatments for diabetic retinopathy since 2 year study visit
• Visual acuity outcomes at 5 years
  • Mean change in visual acuity from baseline and from 2 year (or last) visit (primary VA outcome)
  • Proportion of eyes with 2 and 3 or more line gains or losses in visual acuity from baseline and from 2 year (or last) visit. 
  • Distribution of visual acuity levels at 5 years
• DME outcomes at 5 years
  • Mean change in OCT from baseline and from 2 year (or last) visit (primary OCT outcome)
  • Proportion of eyes with OCT central subfield thickness < 250 µm on Zeiss Stratus or the equivalent on spectral domain OCT based on gender specific cutoffs at 5 years
• Diabetic retinopathy outcomes at 5 years
  • Diabetic retinopathy severity on fundus photos at 5 years
  • Improvement in diabetic retinopathy from baseline and from 2 years
  • Worsening in diabetic retinopathy from baseline and from 2 years  

 Secondary analyses will include original treatment group comparisons for the following:

• Visual acuity outcomes at 5 years (see outcomes above)
• DME outcomes at 5 years (see outcomes above)
• Diabetic retinopathy outcomes at 5 years (see outcomes above)

 

Note: If there are a substantial number of participants in all three treatment groups who were solely treated with their randomized treatment during the 5 years following randomization, a sensitivity analysis will be performed including only those participants. 

Analyses will consist of three two-group comparisons.  Within each outcome, the Hochberg approach will be used to control the Type 1 error.  Binary outcomes will be analyzed using logistic regression models adjusting for baseline factors where appropriate.  Continuous outcomes will be analyzed using an analysis of covariance model adjusting for baseline measures where appropriate.  All linear model assumptions will be verified including linearity, normality of residuals, and homoscedasticity.  If model assumptions are not met data transformation or a nonparametric analysis will be considered.

Subgroup analyses mirroring all analyses described above will be performed for participants with baseline visual acuity 20/32 to 20/40 and participants with baseline visual acuity 20/50 or worse. 

Missing data will be excluded. 

Since treatment for DME and DR between 2 and 5 years was at investigator discretion and may or may not have been the same treatment as initially randomly assigned, treatment during this stage will be considered when interpreting the results.  

 

Back