Observational Study of the Development of Diabetic Macular Edema Following Panretinal (Scatter) Photocoagulation Given in 1 or 4 Sittings FREE

The Diabetic Retinopathy Study demonstrated that panretinal (scatter) photocoagulation (PRP) reduced the risk of severe vision loss (<5/200 at 2 consecutive 4-month visits) due to complications of proliferative diabetic retinopathy (PDR) from 25% to 14% over 2 years.¹ The Early Treatment Diabetic Retinopathy Study (ETDRS) subsequently demonstrated that for patients with severe nonproliferative diabetic retinopathy or PDR, PRP and vitrectomy when necessary reduced the risk of severe vision loss to 4% over 5 years.² Consequently, the use of PRP has been accepted as the standard care for patients with PDR.³

Panretinal (scatter) photocoagulation has been associated with numerous complications that can result in decreased visual acuity, including macular edema.^{3 - 4} In the ETDRS, which was performed prior to the advent of optical coherence tomography (OCT), 18% of eyes that underwent full PRP (1200-1600 spots) were noted to have macular edema on stereoscopic fundus photographs by 4 months (F. L. Ferris, MD, unpublished data, June 17, 2008). There are multiple theories as to why the edema may occur, including oncotic fluid accumulation related to the tissue destruction or PRP-induced inflammation leading to cytokine release and increased permeability of the retinal capillaries.⁵

In the ETDRS, PRP generally was given over 2 or more sittings, usually within 4 weeks.⁶ Some clinicians, however, have recommended doing the PRP in a single sitting.⁷ In a 2004 survey of investigators participating in the Diabetic Retinopathy Clinical Research Network (DRCR.net), about a quarter of the responding investigators indicated that they routinely performed PRP in a single sitting while three-quarters used multiple sittings. While completion of PRP in 1 sitting might be more convenient with respect to number of office visits and compliance with completion of the treatment plan, there is at least a theoretical concern that this may increase the development of vision-disabling macular edema, although there are few data published to support this hypothesis. Other potential adverse effects include pain, sometimes requiring retrobulbar or peribulbar anesthesia; choroidal effusion or detachment; accommodative paresis; and acute elevation of intraocular pressure.^{8 - 14} Conversely, limiting the total number of spots at any 1 sitting by completing the scatter over multiple sittings may allow the edema to subside before the next sitting of PRP. As possible support for this hypothesis, Shimura et al¹⁵ found that macular edema was more likely when the PRP regimen was given in 4 sittings at weekly intervals compared with 4 sittings at biweekly intervals in a group of eyes with OCT evidence of preexisting macular edema.

There are more than 63 000 new cases of PDR each year in the United States.¹⁶ Panretinal (scatter) photocoagulation is indicated in the vast majority of these cases and in some of the 700 000 persons in the United States with PDR at any given time. Thus, determining if the morbidity from macular edema and accompanying visual acuity loss from PRP differs when the treatment is completed in a single sitting vs multiple sittings would be of interest to the many people receiving PRP each year. If there are no relevant clinical differences, then patient and physician convenience as well as possible economic consequences can be major factors in determining the number of PRP sittings. Potentially better compliance with the recommended treatment could also be realized.

The DRCR.net conducted a study to compare PRP in 1 sitting with PRP distributed over 4 sittings. Although a randomized trial was proposed, a majority of investigators had a strong bias for 1 vs more than 1 sitting, precluding a randomized trial design. Therefore, as a potential prelude to a randomized trial protocol, a prospective nonrandomized study was conducted. To minimize selection bias, investigators were required, prior to study initiation, to indicate whether they would perform PRP in 1 or 4 sittings for all study subjects they enrolled.

The study was conducted by the DRCR.net at 27 clinical sites in the United States. The protocol and Health Insurance Portability and Accountability Act–compliant informed consent forms were approved by multiple institutional review boards. Each subject gave written informed consent to participate in the study. Study oversight was provided by an independent data and safety monitoring committee. The study is listed at http://clinicaltrials.gov, under identifier NCT00687154, and the protocol is available on the DRCR.net Web site (http://drcr.net) and summarized later.

Eligible patients had to be at least 18 years of age with type 1 or type 2 diabetes mellitus. Patients were excluded if they had a history of chronic renal failure requiring dialysis, kidney transplant, pancreatic transplant, or intensive insulin therapy initiated within 4 months of randomization. The major eligibility criteria for a study eye included (1) early PDR or severe nonproliferative diabetic retinopathy for which the investigator intended to perform full scatter photocoagulation, (2) retinal thickness measured on OCT of less than 300 μm in the central subfield (for the first 5 months of the enrollment period, the eligibility criterion was center point thickness ≤200 μm and then the protocol was amended to increase the upper limit to 300 μm in the central subfield), and (3) visual acuity letter score of 73 or higher (20/32 or better) measured with the electronic ETDRS method.¹⁷ An eye was not eligible if it had received prior scatter photocoagulation, had high-risk (severe proliferative) retinopathy, had macular edema that had been treated in the prior 6 months or for which treatment was planned, had a history of major ocular surgery (including cataract extraction, vitrectomy, scleral buckle, or any intraocular surgery) within the prior 6 months or anticipated within the next 8 months, or had a history of YAG capsulotomy performed within the prior 2 months. A subject could have only 1 study eye.

Each investigator was required to declare, prior to enrollment, whether he or she was going to perform PRP in 1 or 4 sittings for all patients participating in this protocol under his or her care. The 1-sitting regimen consisted of the application of 1200 to 1600 burns. The 4-sitting regimen was spread over 12 weeks, with each sitting separated by 4 weeks (±4 days), and consisted of approximately 300 burns in each of the first 2 sittings and investigator judgment for number of burns for the third and fourth sittings as long as the total for the 4 sittings was between 1200 and 1600 burns. For both groups, treatment burns were to be spaced 1 burn width apart and were to extend from the vascular arcades to beyond the equator.

The burn characteristics were as follows: (1) size: argon laser using 200-μm spot size with Rodenstock lens (or equivalent) or 500-μm spot size with 3-mirror contact lens, (2) exposure: 0.1 second recommended, 0.05 to 0.2 second allowed, (3) intensity: standard mild white retinal burns (ie, 2+ to 3+ burns, no 4+ burns permitted),^{18 - 19} (4) distribution: edges at least 1 burn width apart, no closer than 1 row within the arcades, no closer than 2 disc diameters temporal to the fovea, (5) extent: arcades (approximately 3000 μm from the macular center) to at least the equator, and (6) wavelength: green or yellow (red could be used if vitreous hemorrhage was present precluding use of green or yellow). A retrobulbar, peribulbar, or sub-Tenon injection could be used at investigator discretion. An indirect laser delivery system could not be used. All treatments were performed with a laser that produced manually aimed burns, although a repeat mode could be used at investigator discretion.

Follow-up visits were performed in both groups after 3 days, 4 weeks, 17 weeks, and 34 weeks within prespecified time windows. At baseline and at each of these follow-up visits, best-corrected visual acuity was measured at 3 m. A certified visual acuity examiner completed a refraction following DRCR.net–specific protocol and tested visual acuity using the ETDRS electronic visual acuity system.¹⁷ Optical coherence tomography images were obtained through a dilated pupil by a certified operator using the Zeiss Stratus OCT (OCT3; Carl Zeiss Meditec, Jena, Germany). Seven-field fundus photographs were obtained at baseline and 3-field photographs, at 34 weeks. Both sets were sent to the DRCR.net Reading Center at the University of Wisconsin–Madison for grading.

If macular edema developed and resulted in a decrease of more than 10 letters in either treatment group, and this acuity decrease from macular edema still was present on a second examination at least 2 weeks later, focal/grid photocoagulation could be given at the discretion of the investigator and completion of PRP in the 4-sitting group could be deferred. If a vitreous hemorrhage occurred, additional PRP could be given.

Optical coherence tomography scans were 6 mm in length and included the 6–radial line pattern (fast macular scan option with OCT3) for quantitative measures and the crosshair pattern (6-12 to 9-3 o’clock) for qualitative assessment of retinal morphology. The OCT scans were sent to the DRCR.net Reading Center at the University of Wisconsin–Madison for grading. Based on review of these scans by an expert grader at the Reading Center, a total of 11% of the 155 baseline scans and 11% of the 567 follow-up scans were judged by the Reading Center to have inaccurate automated thickness measurements because the automated placement of lines by the OCT software was judged morphologically not to be at the inner aspect of the retina or outer aspect of the retina or both. In these cases, center point thickness was measured manually and the resultant value used to impute a value for the central subfield (based on a correlation of the 2 measures of 0.98), as previously published.²⁰

A sample size was planned to be approximately 150 eyes (only 1 eye per subject was entered into the study), with approximately equal numbers receiving each of the 2 treatment regimens. If both eyes were eligible, the investigator at his or her discretion selected 1 of the eyes to be the study eye. As an observational study, this protocol aimed to determine if any trends exist and if the trends were strong enough to warrant a phase 3 trial.

The primary study outcome was OCT-measured central subfield thickening at 34 weeks; the main secondary outcome was visual acuity at 34 weeks. Another secondary outcome was the presence of retinal thickness in follow-up of at least 250 μm in the central subfield combined with an increase in OCT-measured retinal thickness from baseline of at least 25 μm (representing the half width of the 95% confidence interval on the difference between replicate OCT measurements in eyes with levels of macular edema similar to this cohort²¹ ). Outcomes related to PRP safety and efficacy included the need for additional PRP or the development of vitreous hemorrhage.

Burn number and average power were compared between 1-sitting and 4-sitting groups using Wilcoxon rank sum tests. Signed rank tests were performed on changes in OCT central subfield thickness from baseline at each follow-up visit for the 1-sitting and 4-sitting groups, respectively. Statistical comparisons of the 1-sitting group vs the 4-sitting group for the continuous outcomes were performed using nonparametric analysis based on ranks (van der Waerden scores) adjusted for baseline values. The models were also adjusted for race and baseline retinopathy severity to account for potential confounding due to the slight baseline imbalances between treatment groups. Logistic regression models adjusted for baseline values, race, and baseline retinopathy severity were used to assess dichotomous outcomes. Subgroup analyses were also performed stratified by baseline retinopathy severity (nonproliferative vs proliferative retinopathy). Missed visits were excluded from the analysis. Analyses using the last observation carried forward imputation method gave similar results (data not shown).

All Pvalues are 2-tailed. SAS version 9.1 (SAS Institute Inc, Cary, North Carolina) was used for all analyses.

Between 2005 and 2007, 155 subjects (mean [SD] age, 55 [11] years; 46% women) were enrolled at 27 sites. Mean visual acuity was 20/25 (mean [SD] letter score, 83 [5]) and mean (SD) OCT central subfield retinal thickness was 202 (26) μm, with the central subfield being less than 250 μm in 151 eyes (97%). Among the 155 subjects, 144 had fundus photographs that could be graded by the Reading Center; 73 eyes (51%) were classified as having nonproliferative retinopathy, 50 eyes (35%) had non–high-risk PDR, and 21 eyes (15%) had high-risk PDR (even though these latter 21 eyes were not considered to have high-risk PDR by the enrolling ophthalmologist). Macular edema was considered to be present on baseline photographs in 27 eyes (18%). Eighty-four eyes were enrolled in the 1-sitting group and 71, in the 4-sitting group. The baseline characteristics of the 2 groups are presented in Table 1.

In the 1-sitting group, 84 eyes received a median of 1274 burns (interquartile range, 1220-1406) compared with 1260 burns (interquartile range, 1200-1456) in the 71 eyes in the 4-sitting group (P = .73). Eleven eyes in the 4-sitting group had fewer than 4 sittings because of missed visits, with the median number of burns in these eyes being 618. Median average power was 280 mW and 250 mW in the 2 groups, respectively (P = .89). Thirty-nine (46%) of the eyes in the 1-sitting group received a retrobulbar anesthetic injection prior to treatment while an injection was received for at least 1 sitting in 10 eyes (14%) in the 4-sitting group (P < .001). Additional scatter photocoagulation after completion of the initial regimen, and before 34 weeks, was given to 6 eyes in the 1-stting group.

Visit completion was 98% at the 3-day visit, 94% at the 4-week visit, 92% at the 17-week visit, and 88% at the 34-week visit in the 1-sitting group and 94%, 94%, 89%, and 82%, respectively, in the 4-sitting group (Figure). One subject died (in the 4-sitting group). Among the 11 subjects in the 4-sitting group with an incomplete laser regimen, 2 completed both the 17-week and 34-week visit, 2 completed just the 17-week visit, and 7 dropped out prior to the 17-week visit.

Median OCT central subfield thickness at baseline was 207 μm in the 1-sitting group and 198 μm in the 4-sitting group. During follow-up, thickness tended to increase slightly from baseline levels in both the 1-sitting and 4-sitting groups (1-sitting group: P < .001 at each visit compared with baseline; 4-sitting group: P = .004 at 3 days and <.001 at 4 weeks, 17 weeks, and 34 weeks) (Table 2). However, in the 1-sitting group, the median changes appeared stable across all follow-up visits, while in the 4-sitting group, the median changes appeared to increase throughout the study.

At the 3-day visit (after completion of the 1-sitting regimen and the first sitting of the 4-sitting regimen), the median change from baseline in central subfield thickness was +9 μm in the 1-sitting group compared with +5 μm in the 4-sitting group (P = .01), with 9% and 5% of eyes in the 2 groups, respectively, having increased by 25 μm or more (P = .32). At 4 weeks, the 1-sitting group had a greater median change in retinal thickness from baseline (+13 μm) than the 4-sitting group (+5 μm; P = .003); at this time, there was also an increase from baseline retinal thickness of 25 μm or more in 18% of the 1-sitting group compared with 8% (P = .04) in the 4-sitting group.

By 17 weeks (approximately 17 weeks after the 1-sitting group regimen had been completed and approximately 5 weeks after the 4-sitting group regimen had been completed), median central subfield thickness in the 4-sitting group had increased to a level similar to that in the 1-sitting group (median change from baseline, +14 μm vs +15 μm, respectively; P = .08). Seventeen percent of eyes in the 1-sitting group and 41% in the 4-sitting group had an increase of at least 25 μm (P = .001), and 6% and 17%, respectively, had an increase of at least 25 μm from baseline, with a thickness of 250 μm or greater (P = .003).

At 34 weeks (approximately 34 weeks after the 1-sitting group regimen had been completed and approximately 22 weeks after the 4-sitting group regimen had been completed), the median change in the central subfield thickening in the 1-sitting group was stable at +14 μm but there was a further increase in the 4-sitting group to +22 μm (P = .06). Twenty-five percent of the eyes in the 1-sitting group and 45% of eyes in the 4-sitting group had an increase in the central subfield thickening of at least 25 μm (P = .005), including 13% and 24%, respectively, having an increase of at least 25 μm to a thickness of 250 μm or greater (P = .02). Results for OCT-measured retinal volume were similar to the central subfield results, with a greater increase seen in the 1-sitting group compared with the 4-sitting group at 3 days and 4 weeks but a greater increase in the 4-sitting group compared with the 1-sitting group at 17 weeks and 34 weeks (Table 2).

Subgroup analysis of eyes with nonproliferative and proliferative retinopathy at baseline gave similar results to the pooled analyses for both central subfield thickening and retinal volume.

Prior to the 34-week visit, macular edema was treated with focal/grid photocoagulation in 1 eye in the 1-sitting group and 1 eye in the 4-sitting group. On the 34-week fundus photographs, macular edema was present in 17 eyes (23%) in the 1-sitting group and 15 eyes (29%) in the 4-sitting group.

The visual acuity results generally paralleled the retinal thickening results (Table 3). The visual acuity was slightly worse in the 1-sitting group compared with the 4-sitting group at the 3-day visit, with a median change from baseline in letter score of −3 and −1, respectively (P = .005). At weeks 4 and 17, the visual acuity change from baseline was similar, with a median change from baseline in letter score of −1 in each group at both visits (P = .37 and .66, respectively). By 34 weeks, the visual acuity was slightly worse in the 4-sitting group, with a median change from baseline in letter score of 0 vs −2 (P = .006). Although few eyes experienced a worsening of 10 or more letters from baseline, this change was greater at the 3-day visit in the 1-sitting group (13% vs 2%; P = .004); thereafter, little difference for this outcome was observed between groups. At 34 weeks, 7% in the 1-sitting group and 9% in the 4-sitting group had visual acuity 10 or more letters worse than baseline (P = .75). There was no obvious correlation of an increase in retinal thickening by 25 μm or more with visual acuity loss (≥5 letters or ≥10 letters).

Similar results were observed when analyzed separately for eyes with nonproliferative vs proliferative retinopathy at baseline or with respect to the anesthesia used to apply PRP.

Between the 17-week and 34-week visits, a vitreous hemorrhage reducing acuity by 10 or more letters from baseline occurred in 2 eyes in each group. These eyes had mild and moderate PDR in the 1-sitting group and one had high-risk PDR in the 4-sitting group (the Reading Center was not able to grade the retinopathy level of the other eye) at baseline.

This study evaluated 155 individuals undergoing PRP for diabetic retinopathy with relatively good visual acuity and no or minimal macular edema within 2 nonrandomized groups, completion of PRP in 1 or 4 sittings. While some eyes had an increase in OCT-measured central retinal thickening following treatment in either treatment group, most eyes maintained a central retinal thickness within a normal range. Retinal thickening was greater in the first few days after completion of the entire treatment regimen of 1200 to 1600 burns in a single sitting compared with the initial 300 burns from the first of 4 sittings. However, this treatment group difference was transient, and the results at 34 weeks of follow-up suggested that retinal thickening might be slightly greater with the 4-sitting regimen. At 34 weeks, the eyes in the 4-sitting group had a shorter time from the last laser sitting than in the 1-sitting group. The eyes within the 4-sitting group might have improved further on OCT and visual acuity measurements if more follow-up was obtained.

The number of burns administered in the 2 groups was similar. Not unexpectedly, retrobulbar/peribulbar anesthesia was used more commonly in the 1-sitting group, although without any apparent serious adverse events noted nor any differences in OCT or visual acuity outcomes. Although there was a slight increase in OCT-measured central subfield thickness in both groups overall, only 1 eye in the 1-sitting group and 1 eye in the 4-sitting group developed macular edema sufficient to be treated with focal/grid photocoagulation. The only difference judged clinically relevant between groups at any point was the increased chance for visual acuity loss in the 1-sitting group at the 3-day visit, which no longer was apparent between 4 weeks and 17 weeks. By the 34-week visit, it is possible that the course of the disease, independent of the number of sittings used to complete the PRP, could have a greater influence on the OCT and visual acuity measurements than any differential impact of the number of sittings on these outcomes.

Fundus photographs were not obtained to evaluate level of retinopathy at the end of the 34 weeks of follow-up; therefore, direct comparison of the efficacy of the scatter treatment in the 2 groups cannot be made. Few eyes required additional scatter treatment after completion of the initial regimen and few developed a vitreous hemorrhage.

In one study published in 1982, investigators reported a single-center randomized study evaluating the beneficial or adverse effects of argon laser photocoagulation for PDR depending on whether treatment was administered in a single sitting or multiple sittings. Results showed no major differences between groups in the effect of treatment on visual acuity, visual field scores, or retinopathy risk factors; OCT was not available at the time. According to Doft and Blankenship,²³ exudative retinal detachment, choroidal detachment, and angle closure occurred more commonly in the single-sitting regimen, but these adverse effects were transient, and no long-term difference between treatment groups was found. These adverse effects also have been reported by others^{8 - 11} ; however, none of these adverse effects were identified in our study, although we did not perform ultrasonography to look for shallowing of the anterior chamber. We are not aware of any other prospective studies that have compared PRP administered in 1 sitting and multiple sittings. However, other prior studies have evaluated the development of macular edema following a multiple-sitting PRP regimen. Most of the literature consists of case reports and case series.^{4 ,15 ,24 - 26} Shimura et al¹⁵ conducted a prospective study in which 36 patients with type 2 diabetes who had bilateral symmetric severe nonproliferative or early proliferative retinopathy without clinically significant macular edema received PRP in 4 sittings. Each subject received the treatment once a week for 4 weeks in one eye and once every other week over 8 weeks in the other eye. They reported that there was a greater increase in central retinal thickening in the eyes treated weekly than in the eyes treated biweekly and the resolution of the edema was slower in the eyes treated weekly. Their results suggest a somewhat greater increase in retinal thickening after scatter treatment than was reported in this study. However, since their results excluded 7 eyes that developed macular edema and 39% of eyes received additional scatter treatment after completion of the initial regimen (whereas only 4% of eyes in our cohort did), it is difficult to compare their results with the results presented in this study. In the ETDRS, which was performed prior to OCT availability, 18% of eyes that had been treated with full scatter photocoagulation in 2 or more sittings were noted to have macular edema graded on fundus photographs at 4 months. While this value is similar to what was reported in this study within the 4-sitting group based on OCT criteria for macular edema, the correlation between presence of edema on fundus photographs and OCT thickening is not strong.²⁰

The strengths of this study were that it was prospective and followed a standardized protocol. However, the results of this study must be viewed in the context that treatment group assignment was not determined using randomization and the number of subjects enrolled was not large enough to derive definitive conclusions. The majority of DRCR.net investigators had a strong bias for 1 vs more than 1 sitting, precluding a randomized trial design. This was primarily because of concern of investigators who routinely performed PRP in multiple sittings who feared that a single-sitting regimen would produce clinically relevant macular edema in their patients. It was thought that if results from a nonrandomized trial showed no large difference between 1 vs 4 PRP sessions, randomization might be possible in a future trial. Some control of selection bias was obtained by requiring participating investigators to indicate prior to the study whether they would perform scatter treatment in 1 sitting or 4 sittings for all of their subjects enrolled in the study. In an attempt to have treatment groups with similar levels of retinopathy severity, eyes with high-risk proliferative disease were excluded because scatter treatment generally would be given over a shorter period than the 12-week period mandated in the 4-sitting group. Baseline characteristics of the 2 treatment groups were fairly well balanced despite the lack of randomization, although, as expected from previous DRCR.net studies,²⁷ the level of retinopathy as determined by the investigator did not always match the level as determined by the Reading Center. Nevertheless, selection bias cannot be excluded in this study since investigators might not offer study participation to a patient for whom they did not believe that the study treatment regimen was appropriate. In addition, only 85% of subjects completed the 34-week follow-up period, a percentage lower than the goal of 90% or greater. Also, there were 11 eyes in the 4-sitting group in which the PRP was not completed because of missed visits and in which the median number of laser spots was only 618. Though unlikely, this study does not allow one to determine if the outcomes in the 4-sitting group would have been different if the PRP was completed in these eyes and complete follow-up was obtained.

Results of the current study suggest that PRP for diabetic retinopathy can be safely administered in a single sitting in patients with relatively good visual acuity and no or mild preexisting center-involved macular edema. Because of its nonrandomized design and small sample size, this study is not sufficient to determine unequivocally that applying PRP in 1 sitting is either noninferior or superior to applying PRP in more than 1 sitting. More definitive results would require a large randomized trial. The implications of this study, if true, have important public health implications, as the costs to the patient in terms of travel and lost productivity as well as to the eye care provider could be reduced.

Correspondence: Alexander J. Brucker, MD, c/o Jaeb Center for Health Research, 15310 Amberly Dr, Ste 350, Tampa, FL 33647 (drcrnetF1@jaeb.org).

Submitted for Publication: July 22, 2008; final revision received September 22, 2008; accepted October 10, 2008.

Financial Disclosure: Dr Brucker has been a consultant and has stock ownership or options for Acuity. Dr Antoszyk has been a consultant for Genentech, Allergan, and Alimera Sciences; received honoraria from/been on the speakers' bureau for Genentech and Bausch & Lomb; and received research grants from Genentech, Allergan, Regeneron, Alcon, Acuity, Alimera Sciences, Bausch & Lomb, Eli Lilly, Genaera, National Eye Institute, NeoVista, OSI Eyetech, Othera, Sirion Therapeutics, and VisionCare Ophthalmic Technologies. Dr Bressler has received research grants from Acucela, Allergan, Bausch & Lomb, Carl Zeiss, Emmes, Fovea Pharmaceuticals, Genentech, Jerini, Merck, NotalVision, Novartis, OSI Eyetech, Othera, Pfizer, QCT, Schering-Plough, Regeneron, and TargeGen; been a consultant for NotalVision, Genentech, Oxigene, Potentia, and AstraZeneca; received travel expenses from Genentech; and received honoraria from/been on the speakers' bureau for Pfizer. Dr Gross has received honoraria from/been on the speakers' bureau for ISTA. Dr Wells has been a consultant for Allergan, OSI Eyetech, and Pfizer; received honoraria from/been on the speakers' bureau for and received travel expenses from OSI Eyetech and Pfizer; and received a research grant from Pfizer.

Funding/Support: This work was supported through cooperative agreements EY14231, EY14269, and EY14229 from the National Eye Institute and the National Institute of Diabetes and Digestive and Kidney Diseases.