Preoperative Clinical Characteristics
There were 101 eyes of 101 patients with stage III (n=65,64.4%) or IV (n=36, 35.6%) IMH that underwent PPV combined with ILM peeling and air tamponade. No serious postoperative complications happened. 26 patients (25.7%) were men, and 75 patients (74.3%) were women. One eye (1.0%) was pseudophakic and 100 eyes (99.0%) were phakic.
Comparison of the Preoperative Clinical Characters of the Closed and Unclosed Groups
81 eyes (80.2%) got a closed IMH (Type I) after one operation (defined as Group A). Postoperative unclosed IMHs were found in 20 eyes (19.8%) (defined as Group B). There were 50 (61.7%) stage III IMHs and 31(38.3%) stage IV IMHs in group A and 15 (75.0%) stage III IMHs and 5 (25.0%) stage IV IMHs in group B. Patients in group A and B did not differ significantly in age (64.1±4.4 years versus 67.0±4.5 years; P=0.32), duration of symptoms (8.3±11.1 months versus 8.3±3.9 months, P=0.38), preoperative BCVA (1.1±0.4 logMAR versus 1.0±0.4 logMAR, P=0.68), preoperative IOP (15.5±2.7 mmHg versus 16.0±3.5 mmHg, P=0.46), and axial length (23.4±0.7mm versus 22.9±0.7mm; P=0.24)(Table 1).
Minimum and maximum diameter of IMH was also analyzed. The mean minimum diameter was significantly smaller in group A than in group B (597.6±120.1mm vs. 703.6±116.1mm, P<0.01). The mean maximum diameter was not significantly different between the two groups (1144.6±182.0mm vs. 1232.2±186.3mm, P=0.58) (Table 1).
In patients with a disease duration of more than 6 months, the mean minimum diameter was significantly smaller in group A than in group B (572±140 mm vs. 688±89 mm, P<0.01). The mean maximum diameter was not significantly different in the two groups (1032±202.0 mm vs. 1195±84 mm, P=0.08).
Analysis of HDR in the Two Groups
The mean HDR was significantly smaller in group A than in group B (0.5±0.1 vs. 0.6±0.1, P=0.03).
The area under the receiver operating characteristic curve of preoperative HDR for differentiating postoperative closed and unclosed MH was 0.680 (95% confidence interval 0.559–0.801, P=0.01). ( Figure 1). When the HDR cutoff value was set at 0.6, the sensitivity was 75.0%, with the specificity 67.9%.
In patients with a disease duration of more than 6 months, the mean HDR was also 0.5±0.1 in group A, significantly (P=0.02) smaller than 0.6±0.1 in group B. In these patients with long duration, when the cutoff value of HDR was set at 0.6, the sensitivity for predicting postoperative closure was 78.6%, and the specificity was 70.5%.
With the HDR cutoff value 0.6, we classified IMHs into A-type (HDR < 0.6, n=61) and H-type (HDR³0.6, n=40). There was significantly more A-type IMHs in group A than in group B (67.9% vs. 30.0%, P=0.002). Overall, the mean minimum diameter of IMHs was significantly smaller in A-type than that in H-type (567.0±108.2 mm vs. 697.3±110.9 mm, P<0.01). Subgroup analysis demonstrated In group A, the minimum diameter of A-type IMHs was also significantly smaller than that of H-type (559.0±102.9 mm vs. 679.1±114.6 mm , P<0.01), but not in group B (639.7±137.5 mm vs. 731.0±98.7 mm , P=0.11). For A-type IMHs, both the mean minimum diameter (559.0±102.9 mm vs. 639.7±137.5 mm, P=0.1) and HDR (0.5±0.1 vs. 0.5±0.02, P =0.4) were not significantly different between group A and B. Similarly, for H-type IMHs, the mean diameter (679.1±114.6 mm vs. 731.0±98.7 mm, P=0.2) and HDR (0.6±0.1 vs. 0.6±0.04, P =0.3) were not significantly different between group A and B. (Table 2)
In a multivariate logistic regression model, the post-operative closure of IMH was significantly associated with preoperative minimum IMH diameter (P=0.02) and HDR (P=0.04), but not significantly associated with gender(P=0.97), age(P=0.15), duration of disease(P=0.28), preoperative visual acuity(P=0.08), or axial length (P=0.29).
Comparison of closure rate of two types of IMHs
The postoperative closure rate was significantly higher in A-type IMHs than that in H-type IMHs (90.2% vs. 65.0%, P<0.01).