Discussion of a Post- Marketing Approval Study for 30-day Continuous Wear Contact Lenses

Gene Hilmantel, O.D., M.S. Bernard P. Lepri, O.D., M.S., M.Ed. James F. Saviola, O.D. Rosalie A. Bright, Ph.D.( OSB/DPS )

New contact lens materials with much higher oxygen transmission may have the potential for safer continuous wear for longer periods of time.

The incidence of corneal ulcers is the main concern.

Although a serious problem, the incidence of ulcers is too low to reliably determine the risk in a reasonable PMA study.

The FDA’s position is that the best way to address this concern is to require a “Post-Marketing Approval Study” of the risk posed by 30-day “continuous wear.”

What is an unacceptable ulcer rate?

Relative Risk (Risk of EW ulcer/Risk of DW ulcer) from Schein O, et al. The relative risk of ulcerative keratitis among users of daily-wear and extended-wear soft contact lenses. New Eng J Med 1989;321: (continuous wear)

From the 1989 Schein data, it seems that a “relative risk” of about (compared to Daily Wear) was considered unacceptable.

Poggio’s 1989 study* found that the incidence of ulcers was: 4 per 10,000 in DW patients, 20 per 10,000 in EW patients. Poggio’s 1989 study* found that the incidence of ulcers was: 4 per 10,000 in DW patients, 20 per 10,000 in EW patients. *Poggio EC, et al. The incidence of ulcerative keratitis among users of daily-wear and extended-wear soft contact lenses. New Eng J Med 1989;321:

Assuming that 15x the risk of daily wear is unacceptable, this means that : 15x(4/10,000) = 60/10,000 is too much risk. 60/10,000 is about times the risk of 7-day Extended Wear. Assuming that 15x the risk of daily wear is unacceptable, this means that : 15x(4/10,000) = 60/10,000 is too much risk. 60/10,000 is about times the risk of 7-day Extended Wear.

Lifetime ulcer risk Vs Number of Years Wearing Contacts Number of Years Wearing Contacts Lifetime ulcer risk One year ulcer risk

Question: Does the panel feel that using 60/10,000 ulcers/patient-year (about 3x the 7-day EW rate) as an upper limit is reasonable? Question: Does the panel feel that using 60/10,000 ulcers/patient-year (about 3x the 7-day EW rate) as an upper limit is reasonable?

What Type of Study Should Be Recommended? What Type of Study Should Be Recommended? IncidenceRelative Risk Case-Control Study Follow Cohort

 Can assess relative risk of different actual wearing schedules.  Good for the study of rare diseases.  Relatively inexpensive.  Can assess relative risk of different hygiene practices.  “Real World” environment. Patients and practitioners NOT self or other selected.  Can assess relative risk of different actual wearing schedules.  Good for the study of rare diseases.  Relatively inexpensive.  Can assess relative risk of different hygiene practices.  “Real World” environment. Patients and practitioners NOT self or other selected. Advantages of a Case-Control Study:

 Requires a waiting period (3 - 5 years?), until 30-day lenses have sufficient market share.  Only assesses relative risk, not actual incidence of ulcers. Ulcer rate for 7-day lenses may have decreased since 1989 (since 30-day wear was eliminated).  Generally produces large confidence intervals, or else need very large number of ulcers. E.g., in Schein(1989): RR of overnight EW to DW was 10.2, with a CI of  Requires a waiting period (3 - 5 years?), until 30-day lenses have sufficient market share.  Only assesses relative risk, not actual incidence of ulcers. Ulcer rate for 7-day lenses may have decreased since 1989 (since 30-day wear was eliminated).  Generally produces large confidence intervals, or else need very large number of ulcers. E.g., in Schein(1989): RR of overnight EW to DW was 10.2, with a CI of Disadvantages of a Case-Control study:

Question: Will there be difficulty in getting enough EW ulcer cases? Schein (1989) had 86 in study (52 were EW). Half of all CL-related ulcers are DW. Change in pattern of care for ulcers since 1980s. Question: Will there be difficulty in getting enough EW ulcer cases? Schein (1989) had 86 in study (52 were EW). Half of all CL-related ulcers are DW. Change in pattern of care for ulcers since 1980s.

*Table is for testing H 0 : Relative Risk= 1. (30-Day/7-Day Wear) ONE-SIDED TEST,  =.05. Cases ( EW Ulcers) Required for 80% Power to Detect Relative Risk* Relationship between Sensitivity of test and required Sample Size.

Statistical POWER is a key measure of our confidence in product safety. Power and sample size are strongly related. Statistical POWER is a key measure of our confidence in product safety. Power and sample size are strongly related. Cases (Ulcers) Required to Detect Relative Risk* *Table is for testing H 0 : Relative Risk= 1. (30-Day/7-Day Wear) ONE-SIDED TEST,  =.05.

There is an interplay between Market Penetration and Sensitivity & Power of the test. Table is for testing H 0 : Relative Risk= 1. ONE-SIDED TEST,  =.05 Cases (Ulcers) Required for 90% Power to Detect Relative Risk*

Questions: What statistical POWER would the panel recommend to ensure confidence in the result? Should we wait for greater penetration of the market, in order to achieve greater sensitivity and power? Questions: What statistical POWER would the panel recommend to ensure confidence in the result? Should we wait for greater penetration of the market, in order to achieve greater sensitivity and power?

Following a cohort of 30-day wearers is an alternative way to assess risk. Could be done by requiring a large number of practitioners to fill out a small follow-up questionnaire after 1 year of experience with the lens. Following a cohort of 30-day wearers is an alternative way to assess risk. Could be done by requiring a large number of practitioners to fill out a small follow-up questionnaire after 1 year of experience with the lens.

Disadvantages of a Cohort study: Selected patients. Selected practitioners. Relatively controlled follow-up environment. Cost? Advantages of a Cohort Study: May yield results within ~ 2 years(?) of FDA approval. Can assess incidence of ulcers. May assess incidence of other complications.

*For ONE-SIDED TEST; alpha=0.05 Significant increases in ulcer rate are detectable with a large sample size. (Here, Testing the null hypothesis: ulcer rate  )*

Question: What type of clinical setting does the panel recommend for implementation of a post- approval study? private practitioners commercial chains HMOs all of the above sponsor’s discretion Question: What type of clinical setting does the panel recommend for implementation of a post- approval study? private practitioners commercial chains HMOs all of the above sponsor’s discretion

Would the panel recommend: a case-control study? following a large cohort? both? Would the panel recommend: a case-control study? following a large cohort? both?

Question: How would the panel define the endpoints that we are interested in for the study? Endpoints used could be:  Patients with presumed infectious ulcers;  Patients with any type of ulcers;  Patients with central or para-central scars (not present at pre-fit);  Patients with reduction in acuity associated with scaring;  Other? Endpoints used could be:  Patients with presumed infectious ulcers;  Patients with any type of ulcers;  Patients with central or para-central scars (not present at pre-fit);  Patients with reduction in acuity associated with scaring;  Other?

Questions? Comments? Questions? Comments?