Results of the search

The electronic literature searches conducted in June 2010 identified 836 potentially relevant references for this review. After duplicate review of the titles and abstracts, we classified 748 references as definitely exclude, 23 as definitely include, and 65 as unsure. Seventeen of the 65 references assessed as unsure were letters or editorials that did not report original data and were excluded. We obtained full‐text copies of the 48 remaining references classified as unsure and reviewed them in duplicate. Of those, we excluded 40 and included eight.

A manual search of other resources, including reference lists of included studies and citation index databases, yielded four additional potentially relevant full‐text references for this review. Of those four references, we included two and excluded two from this review.

In the 2011 publication of this review (Gharaibeh 2011), we included 26 studies as reported in 33 publications excluded 41 studies that were in 42 publications.

After revising and updating the electronic searches as of August 2013, we identified 460 additional references for review (Figure 1). After duplicate review of the titles and abstracts, we classified 438 references as definitely exclude and 22 as unsure. We obtained full‐text copies of the references classified as unsure and reviewed them in duplicate. Seventeen of the references were in non‐English languages, and we identified colleagues who read the languages to assist with assessing the articles in duplicate. Of the 22 references reviewed in full, we excluded 20, one was a reference for a study already included in the review, and we included one as a new study in the review.

Figure 1

---

---

Study flow diagram for 2013 update of literature searches.

A manual search of other resources, including reference lists of included studies and citation index databases, yielded four additional potentially relevant full‐text references for this review. Of those four references, we excluded one and the remaining three were from studies already included in this review. In all, we included 27 studies reported by 38 publications in this review and excluded 62 studies reported by 63 publications.

Included studies

The 27 studies included in this review are described in the 'Characteristics of included studies' table. Twenty of the included studies were randomized controlled trials (RCTs), and seven used a quasi‐randomized method to assign participants to treatment groups. The review outcomes reported by the included studies are listed in Table 1.

All but two of the studies restricted entry to patients with primary traumatic hyphema; Welsh 1971 also included patients with perforated globes that had been sutured and were treated as closed globe injuries, and Palmer 1986 also included some patients with secondary hemorrhage. Most studies included all age groups, although some studies excluded very young children (e.g. less than four or seven years) (Farber 1991; Kutner 1987; Marcus 1988; Pieramici 2003; Vangsted 1983; Welsh 1983), and one study included children only (Kraft 1987). Of studies reporting demographic data, the mean age of participants ranged from 10 to 32 years, and the proportion of male participants ranged from 67% to 100%. Studies took place in a number of different countries: three in China; two each in Iran, Sweden, and South Africa; one each in Denmark, Israel, and Malaysia; and the remainder in Canada and the US. The race of participants varied by country, and nine studies reported 50% or more black participants.

Three types of antifibrinolytic agents (epsilon‐aminocaproic acid (aminocaproic acid), tranexamic acid, and aminomethylbenzoic acid) were investigated in the included studies. Other types of pharmaceuticals investigated by the studies included in this review were corticosteroids; including prednisone, prednisolone, hydrocortisone, and cortisone; conjugated estrogen; aspirin; and topical mydriatics and miotics. One study compared traditional Chinese medicine (TCM) (Yunnan Baiyao) versus antihemorrhagic agents. Nonpharmaceutical interventions included the use of monocular or binocular patching, eye shields, bed rest, and elevation of the head. The primary outcome for all but three studies was the risk of a secondary hemorrhage.

Excluded studies

There were 62 excluded studies. The reasons for exclusion are described in the 'Characteristics of excluded studies' table. We excluded 43 studies because the study design was not a randomized or controlled clinical trial; nine studies because they included nontraumatic hyphema cases and did not report outcomes for traumatic hyphema cases separately; seven studies because no original data were presented; and three studies because they investigated interventions outside the scope this review (e.g. surgical interventions and patient education interventions).

Allocation

Twenty of the 27 studies included in the review were RCTs. Seven studies specified using computerized randomization to generate the allocation sequence and one study used a randomization list; we judged these eight studies as having a low risk of sequence generation bias (Figure 2). Twelve of the 20 RCTs did not report methods of allocation; therefore, we assessed these studies as having an unclear risk of sequence generation bias. Of the 20 included randomized trials, eight reported that allocation concealment was implemented: one study used sealed numbered envelopes, two studies used coded bottles, and five studies maintained the randomization code at a pharmacy or other central study center; and 12 studies did not report methods of allocation concealment. The remaining seven studies were controlled clinical trials but did not use randomization to assign participants to treatment. Participants were allocated by alternation for four studies, and by date of admission in one study. The method of allocation was not reported in the remaining two controlled clinical trials.

Figure 2

---

---

Methodologic quality summary: review authors' judgments about each methodologic quality item for each included study. Green: low risk of bias; red: high risk of bias; yellow: unclear risk of bias.

Blinding

Twelve of the 20 included RCTs were double‐masked (participants and investigators), placebo‐controlled trials. One study investigating two doses of oral aminocaproic acid was also double‐masked (Palmer 1986). Participants and treating physicians were partially masked in two studies in which there was only one placebo‐control group for two intervention groups that had different treatment regimens (Karkhaneh 2003; Rahmani 1999). In both of these studies, it was noted that the ophthalmologists and outcome assessors were not involved in participant treatment and were masked to the treatment groups. The interventions of interest in two studies precluded masking; the first study compared aspirin three times daily versus observation only (Marcus 1988), and the second study compared bed confinement versus walking and oral tranexamic acid three times daily (Vangsted 1983). Two studies did not mention whether or not masking occurred (Liu 2002; Wang 1994), and authors of one study reported that no masking was done (Zi 1999).

Masking of participants was not possible because of the type of interventions in four of the seven quasi‐randomized studies included in this review (Edwards 1973; Rakusin 1972; Read 1974; Zetterstrom 1969), and not reported in one (Bedrossian 1974). Masking of participants with the use of placebo pills could have been implemented, but was not achieved in the remaining two quasi‐randomized studies (Sukumaran 1988; Varnek 1980). Masking of outcome assessors was not reported or not clear in all seven quasi‐randomized studies.

Incomplete outcome data

Attrition rates for included studies were minimal due to the nature of the condition and treatment regimens. Typically, treatment duration for traumatic hyphema at the time the studies were completed comprised one week or less, and hospitalization was frequently implemented. Eighteen of the 27 included studies reported no exclusions or losses to follow‐up, and thus used ITT analyses. Of the nine studies that excluded participants from the analysis, four studies excluded only one or two participants due to an adverse effect of treatment (Crouch 1997; Kutner 1987; Palmer 1986), treatment failure (Palmer 1986), or loss of a patient's medical record (McGetrick 1983). The remaining five studies did not conduct ITT analyses, although all reported the number of exclusions and losses to follow‐up.

Selective reporting

The risk of a secondary hemorrhage was reported as a primary outcome in all but five of the included studies; in two studies, time to resolution of the hyphema was reported as the primary outcome (Bedrossian 1974; Zi 1999), in another two studies, secondary hemorrhage was reported as a secondary outcome with no primary outcome identified (Edwards 1973; Read 1974), and in the fifth study, absence of secondary hemorrhage was part of the composite outcome of being "cured" (Wang 1994). All investigators except Zi et al. and Wang et al. reported results for secondary hemorrhage. There were four included studies in which the risk of reporting bias was unclear; due to the lack of study details available in the abstract, and no full version being published (Christianson 1979), because study outcomes were not clearly stated in the publication (Liu 2002; Wang 1994), and because only results for secondary hemorrhage were reported, although VA and IOP were measured throughout the duration of the study (Marcus 1988).

Other potential sources of bias

We detected no other potential sources of bias in 18 of the included studies. We classified four studies as having an unclear risk of other bias because the publications had poor descriptions of study methods and results (Christianson 1979; Liu 2002; Marcus 1988; Wang 1994). In two studies, some participants were selected to receive surgery either at recruitment (Rakusin 1972), or after having been assigned to a treatment group (Read 1974). We classified three studies as having an unclear risk of other bias because they were funded by pharmaceutical companies that either manufactured the drug being investigated in the study or that supplied study drug (Karkhaneh 2003; Pieramici 2003; Welsh 1983).

Antifibrinolytics versus control

Aminocaproic acid versus placebo

Visual acuity (Analysis 1.1;Analysis 1.2;Analysis 1.3;Analysis 2.1)

Two of the studies evaluating aminocaproic acid measured long‐term VA at nine months or from six months to 2.5 years after discharge (Crouch 1976; Kraft 1987). Neither study found a difference in the proportion of participants who achieved useful final VA, defined as VA between 20/20 and 20/40 (Analysis 1.1). Kraft 1987 reported that 17 of 24 (70.8%) participants who had been assigned to aminocaproic acid had VA between 20/20 and 20/40, compared with 20 of 25 (80%) participants assigned to placebo. Similar results were reported by Crouch 1976, with 25 of 32 (79%) participants assigned to drug versus 18 of 27 (67%) participants assigned to placebo achieving useful VA. The summary OR for these two studies indicated no significant difference (OR 1.11, 95% CI 0.47 to 2.61).

No study observed a difference in VA measured at two weeks or less after the hospital admission. At the time of discharge, Kutner 1987 observed VA of 20/40 or better in 14 of 21 (67%) participants in the oral aminocaproic acid group and in 10 of 13 (77%) participants in the placebo group. Similarly, Pieramici 2003 reported that 10 of 24 (42%) participants in the topical aminocaproic acid group and 13 of 27 (48%) participants in the placebo group had VAs of 20/40 or better seven days after study enrolment. Neither study result was significant (Analysis 1.2; Analysis 2.1). Although Karkhaneh 2003 did not report on the proportion of participants with good VA, they did report that there was no significant difference in VA between topical aminocaproic acid‐treated participants and placebo‐treated participants after two weeks of follow‐up.

Two additional studies evaluated final VA with the time of measurement including both short‐ and long‐term time points ranging from five days to 3.4 years (Teboul 1995), or from zero to nine months (McGetrick 1983). Forty‐six of 48 (95.8%) children in the aminocaproic acid group and 44 of 46 (95.6%) children in the placebo group had good final VA in Teboul 1995. McGetrick 1983 reported that the number of participants with final VA of 20/40 or better was 22 of 28 (78.6%) in the aminocaproic acid group and 14 of 21 (66.6%) in the placebo group. The summary OR for final VA of 20/40 or better for these two studies was 1.56 (95% CI 0.53 to 4.56; Analysis 1.3).

Time to resolution of primary hemorrhage (Analysis 1.4;Analysis 2.2)

In general, the hyphemas in participants assigned to aminocaproic acid took longer to clear than those in participants assigned to placebo or control groups. Christianson 1979 noted that drug‐treated hyphemas tended to take longer to clear compared with controls but reported that it was significant only among hyphemas filling more than half of the anterior chamber. Of the five remaining studies using oral aminocaproic acid, the mean time to resolution of the primary hemorrhage ranged from 4.1 to 6.7 days in the aminocaproic acid group and 2.4 to 6.3 days in the placebo group among all participants. Two studies evaluated the time to clear the initial hyphema after excluding participants who rebled (Crouch 1976; Kraft 1987). In both studies, the group receiving aminocaproic acid took longer to clear the initial hyphema than the group receiving placebo (4.0 days versus 2.8 days in Crouch 1976, and 5.3 days versus 2.6 days in Kraft 1987). In Kraft 1987, the time to resolution appeared to be associated with initial hyphema severity, with larger initial hyphemas taking longer to resolve. The longer resolution times for drug‐treated groups were statistically significant as reported in the Kraft and Teboul studies individually; however, there were insufficient data available to perform a meta‐analysis. In contrast, in McGetrick 1983 the mean time to resolution was longer in the placebo than the aminocaproic acid group.

The mean time to resolution of the primary hemorrhage in participants receiving topical aminocaproic acid in Karkhaneh 2003 was 11.1 days (standard deviation (SD) 4.7) versus 9.3 days (SD 4.2) in the participants in the placebo group (P value = 0.07). Pieramici 2003 reported no significant difference in time to clearance of the primary hyphema between topical aminocaproic acid‐treated participants and placebo‐treated participants. However, these studies included all participants, including those who had a secondary hemorrhage.

Risk of secondary hemorrhage (Analysis 1.5;Analysis 2.3;Table 2)

Open in table viewer

Table 2. Outcomes by initial hyphema severity

Study	Severity scale	Reported severity	Secondary hemorrhage	Other outcomes
Oral aminocaproic acid vs. control
Christianson 1979	NR	NR	NR	Time to resolution of the primary hyphema was significantly longer (P value < 0.05) for patients receiving drug in which the hyphema filled more than ½ of the anterior chamber
Crouch 1976	Blood filling < ⅓ of anterior chamber	Reported no statistically significant differences across groups	NR	NR
	Blood filling ⅓ to ½ of anterior chamber
	Blood filling > ½ to ¾ of anterior chamber
	Blood filling > ¾ to total of anterior chamber, but excluded total hyphema
Kraft 1987	Blood filling < ⅓ of anterior chamber	30/49 (61%) participants; 13/24 (54%) in drug group; 17/25 (68%) in placebo group	1/3 (33%) secondary hemorrhage (in placebo group)	Excluding secondary hemorrhages, mean time to resolution of 3.4 days in drug group (range 1‐11 days); mean time to resolution of 2.2 days in placebo group (range 1‐4 days)
	Blood filling ⅓ to ½ of anterior chamber	14/49 (29%) participants; 9/24 (37.5%) in drug group; 5/25 (20%) in placebo group	1/3 (33%) secondary hemorrhage (in drug group)	Excluding secondary hemorrhages, mean time to resolution of 7.1 days in drug group (range 6‐9 days); mean time to resolution of 4.0 days in placebo group (range 3‐4 days)
	Blood filling ½ or more of anterior chamber	5/49 (10%) participants; 2/24 (8.3%) in drug group; 3/25 (12%) in placebo group	1/3 (33%) secondary hemorrhage (in drug group)	Excluding secondary hemorrhages, time to resolution of 10 days in drug group: mean of placebo 4.3 days (range 3‐5 days)
Kutner 1987	Mean hyphema height	2.2 mm (SD 1.7, n = 21) in drug group; 1.7 mm (SD 1.0, n = 13) in placebo group	"All who rebled had initial hyphemas of 15% or less"	NR
McGetrick 1983;	Mean hyphema height	100% (28/28) hyphemas in drug group were < 25% of anterior chamber; 86% (18/21) hyphemas in placebo group were < 25% of anterior chamber	1 secondary hemorrhage in drug group; 6 secondary hemorrhages in placebo group	NR
Teboul 1995	Blood filling < ⅓ of anterior chamber	88/94 (94%) participants; 44/48 (92%) in drug group; 44/46 (96%) in placebo group	1 secondary hemorrhages in drug group and 2 in placebo group	NR
	Blood filling ⅓ to ½ of anterior chamber	6/94 (6%) participants; 4/48 (8%) in aminocaproic acid group; 2/46 (4%) in placebo group	No rebleeds	NR
Topical aminocaproic acid vs. control
Karkhaneh 2003	Blood filling < ¼ of anterior chamber; excluded microscopic hyphemas	65/80 (81%) participants; 34/41 (83%) in drug group; 31/39 (79.5%) in placebo group	Reported no effect of hyphema size on secondary hyphema (RR 0.7, 95% CI 0.2 to 2.5)	NR
	Blood filled ¼ to ½ of anterior chamber	14/80 (18%) participants; 7/41 (17%) in drug group; 7/39 (18%) in placebo group
	blood filling > ½ of anterior chamber; excluded total or blackball hyphemas	1/80 (1%) participants; 0/41 in drug group; 1/39 (2.5%) in placebo group
Pieramici 2003	Mean hyphema height in mm	1 mm (SE 0) in drug group (range 0‐4 mm); 2 mm (SE 0) in placebo group (range 0‐8 mm)	Size of primary hyphema in 2 participants with secondary hemorrhages in drug group: 0.3 and 1 mm; in 8 participants in the placebo group: 0.8, 0.9, 1, 1.4, 1.8, 2, 2, and 4.5 mm	NR
Low‐dose vs. standard‐dose aminocaproic acid
Palmer 1986	Mean hyphema height in mm	1.7 mm (SD 2.0, range 0.1‐9.9) in low‐dose group (n = 25); 1.5 mm (SD 2.2, range 0.1‐9.9) in standard‐dose group; 1.5 mm in standard‐dose group (n = 33)	1 secondary hemorrhage in low‐dose group; 5 secondary hemorrhages in standard‐dose group	NR
Oral vs. topical aminocaproic acid
Crouch 1997	Blood filling < ⅓ of anterior chamber	44/64 (69%) participants	NR	NR
	Blood filling ⅓ to ½ of anterior chamber	6/64 (9%) participants
	Blood filling > ½ to ¾ of anterior chamber	8/64 (13%) participants
	Blood filling > ¾ to total of anterior chamber	6/64 (9%) participants
Tranexamic acid vs. control
Rahmani 1999	Microscopic, but excluding patients with unlayered microscopic hyphemas	17/238 (7%) participants; 6/80 (7%) in aminocaproic acid group; 4/78 (5%) in prednisolone group; 7/80 (9%) in placebo group	2/43 (5%) secondary hemorrhages	NR
	Blood filling < ¼ of anterior chamber	173/238 (72%) participants; 56/80 (70%) in aminocaproic acid group; 61/78 (78%) in prednisolone group; 56/80 (70%) in placebo group	30/43 (70%) secondary hemorrhages
	Blood filling ¼ to ½ of anterior chamber	36/238 (15%) participants; 13/80 (16%) in aminocaproic acid group; 10/78 (13%) in prednisolone group; 13/80 (16%) in placebo group	7/43 (16%) secondary hemorrhages
	Blood filling >½ of anterior chamber; excluded total hyphemas	12/238 (5%) participants; 5/80 (6%) in aminocaproic acid group; 3/78 (4%) in prednisolone group; 4/80 (5%) in placebo group	4/43 (9%) secondary hemorrhages
Sukumaran 1988	Hyphema height of 0‐1 mm	8/35 (23%) participants; 4/17 (24%) in drug group; 4/18 (22%) in control group	NR	NR
	Hyphema height of 2‐3 mm	12/35 (34%) participants; 6/17 (35%) in drug group; 6/18 (33%) in control group
	Hyphema height of 4‐5 mm	10/35 (29%) participants; 5/17 (29%) in drug group; 5/18 (28%) in control group
	Hyphema height of 6‐7 mm	5/35 (14%) participants; 2/17 (12%) in drug group; 3/18 (17%) in control group
Vangsted 1983	Hyphema height of 1 mm	10/112 (9%) participants; 8/59 (14%) in drug group; 2/53 (4%) in control group	NR	NR
	Hyphema height of 2 mm	33/112 (29%) participants; 15/59 (25%) in drug group; 18/53 (34%) in control group
	Hyphema height of 3 mm	37/112 (33%) participants; 18/59 (31%) in drug group; 19/53 (36%) in control group
	Hyphema height of 4 mm	18/112 (16%) participants; 9/59 (15%) in drug group; 9/53 (17%) in control group
	Hyphema height of 5 mm	9/112 (8%) participants; 6/59 (10%) in drug group; 3/53 (6%) in control group
	Hyphema height of 6 mm	4/112 (4%) participants; 3/59 (5%) in drug group; 1/53 (2%) in control group
	Hyphema height of 7 mm	None in either group
	Hyphema height of 8 mm	1/112 (1%) participants; 0/59 (0%) in drug group; 1/53 (2%) in control group
Varnek 1980	Mean hyphema height in mm	2.0 mm in drug group (n = 102); 2.1 mm in control group (n = 130)	1.0 mm in 2 participants in drug group with a secondary hemorrhage; 2.2 mm in 12 participants in control group with a secondary hemorrhage	NR
Welsh 1983	Mean of proportion of anterior chamber area filled with blood	68% in drug group (n = 19); 63% in placebo group (n = 20)	NR	NR
Aminomethylbenzoic acid vs. control
Liu 2002	Blood filling < ⅓ of anterior chamber and level is lower than the inferior boarder of pupil	47/92 (51%) participants; 31/60 (52%) in drug group; 16/32 (50%) in control group	NR	NR
	Blood filling ½ of anterior chamber and level is higher than the inferior border of the pupil, but not exceeding the median line	30/92 (33%) participants; 19/60 (32%) in drug group; 11/32 (34%) in control group
	Blood filling > ½ of anterior chamber or filling the entire anterior chamber	15/92 (16%) participants; 10/60 (17%) in drug group; 5/32 (16%) in control group
Oral corticosteroids vs. control
Spoor 1980	0‐33% of anterior chamber area filled with blood	38/43 (88%) participants; 21/23 (91%) in prednisone group; 17/20 (85%) in placebo group	2/4 (50%) secondary hemorrhages	1. 30 hyphemas resolved in 5 days or less; 8 hyphemas resolved in more than 5 days 2. 34 patients with final visual acuity between 20/20 and 20/50
	> 33% to 75% of anterior chamber filled with blood	5/43 (12%) participants; 2/23 (9%) in prednisone group; 3/20 (15%) in placebo group	2/4 (50%) secondary hemorrhages	1. 1 hyphema resolved in 5 days or less; 4 hyphemas resolved in more than 5 days 2. 5 patients with final visual acuity between 20/20 and 20/50
Rahmani 1999	See above under "Tranexamic acid vs. control"
Topical corticosteroids
Zetterstrom 1969	Mean hyphema height in mm	2.5 mm in topical corticosteroid group (n = 58); 3.5 mm in control group (n = 59)	No patient with secondary hemorrhage in topical corticosteroid group; 4 patients with secondary hemorrhage in control group	NR
Antifibrinolytics vs. oral corticosteroids
Farber 1991	Microscopic	24/112 (21%) participants; 11/56 (20%) in aminocaproic acid group; 13/56 (23%) in prednisone group,	3/8 (38%) secondary hemorrhages; 2 in aminocaproic acid group; 1 in prednisone group	NR
	Hyphema height 0.1‐3.9 mm	80/112 (71%) participants; 41/56 (73%) in aminocaproic acid group; 39/56 (70%) in prednisone group	4/8 (50%) secondary hemorrhages; 1 in aminocaproic acid group; 3 in prednisone group
	Hyphema height 4.0‐5.9 mm	4/112 (4%) participants; 3/56 (6%) in aminocaproic acid group; 1/56 (2%) in prednisone group	No secondary hemorrhages in either group
	Hyphema height 6.0‐11 mm	2/112 (2%) participants; 0/56 (0%) in aminocaproic acid group; 2/56 (4%) in prednisone group	No secondary hemorrhages in either group
	Total hyphema	2/112 (2%) participants; 1/56 (2%) in aminocaproic acid group; 1/56 (2%) in prednisone group	1/8 (12%) secondary hemorrhage; 1 in aminocaproic acid group; none in prednisone group
Rahmani 1999	See above under "Tranexamic acid vs. control"
Conjugated estrogens vs. control
Spaeth 1966	Blood filling < 20% of anterior chamber	55/85 (65%) participants; 28/39 (72%) in estrogen group; 27/46 (59%) in control group	13/20 (65%) secondary hemorrhages; 8 in estrogen group; 5 in control group	NR
	Blood filling 20‐40% of anterior chamber	17/85 (20%) participants; 5/39 (13%) in estrogen group; 12/46 (26%) in control group	4/20 (20%) secondary hemorrhages; 1 in estrogen group; 3 in control group
	Blood filling 40‐60% of anterior chamber	5/85 (6%) participants; 2/39 (5%) in estrogen group; 3/46 (7%) in control group	1/20 (5%) secondary hemorrhage; 0 in estrogen group; 1 in control group
	Blood filling 60‐80% of anterior chamber	2/85 (2%) participants; 1/39 (3%) in estrogen group; 1/46 (2%) in control group	No secondary hemorrhages in either group
	Blood filling > 80% of anterior chamber	6/85 (7%) participants; 3/39 (8%) in estrogen group; 3/46 (7%) in control group	2/20 (10%) secondary hemorrhages; 1 in estrogen group; 1 in control group
Cycloplegics vs. miotics
Bedrossian 1974	Hyphema height of 1 mm	20/58 (34%) participants; 10/28 (36%) in the cycloplegic group; 10/30 (33%) in the miotic group	1/1 (100%) secondary hemorrhage (in cycloplegic group)	Mean time to resolution in cycloplegic group of 1.9 days (SD 1.4); mean time to resolution in miotic group of 2.5 days (SD 1)
	Hyphema height of 2 mm	22/58 (38%) participants; 10/28 (36%) in the cycloplegic group; 12/30 (40%) in the miotic group	No secondary hemorrhages in either group	Mean time to resolution in cycloplegic group of 3.3 days (SD 1.8); mean time to resolution in miotic group of 4.2 days (SD 1.3)
	Hyphema height of 3 mm	12/58 (21%) participants; 6/28 (21%) in the cycloplegic group; 6/30 (20%) in the miotic group	No secondary hemorrhages in either group	Mean time to resolution in cycloplegic group of 3.2 days (SD 1.9); mean time to resolution in miotic group of 4.0 days (SD 1.1)
	Hyphema height of 4 mm	4/58 (7%) participants; 2/28 (7%) in the cycloplegic group; 2/30 (7%) in the miotic group	No secondary hemorrhages in either group	Mean time to resolution in cycloplegic group of 2.5 days (1 resolved on day 2 and 1 on day 3); mean time to resolution in miotic group of 4.0 days (1 resolved on day 3 and 1 on day 5)
Aspirin vs. no aspirin
Marcus 1988	Reported that "the two groups were comparable with respect to age, cause, and extent of hyphema" and that 2 of 3 eyes with a secondary hemorrhage in the aspirin group (n = 23) had an initial total hyphema, while of the 2 eyes with a secondary hemorrhage in the control group (n = 28), 1 had 30% and 1 had almost total hyphema			NR
Traditional Chinese medicine vs. control treatment
Wang 1994	Any level	No significant differences between groups	NR	Proportion of patients who were "cured" (defined as the resolution of the primary hemorrhage after 5 days of treatment, visual acuity of 0.7 or better after resolution of the primary hemorrhage, and no recurrence of bleeding for 1 week following resolution of the primary hemorrhage) was 29/45 (64%) in the TCM group and 10/38 (26%) in the control group
Monocular vs. binocular patching
Edwards 1973	Blood filling < ⅓ of anterior chamber	42/64 (66%) participants; 21/35 (60%) in the monocular patching group; 21/29 (72%) in the binocular patching group	7/14 (50%) secondary hemorrhages; 4 in the monocular group; 3 in the binocular group	62% (13/21) of patients with final visual acuity of 20/50 or better in the monocular group; 71% (15/21) of patients with final visual acuity of 20/50 or better in the binocular group
	Blood filling ⅓ to ½ of anterior chamber	14/64 (22%) participants; 9/35 (26%) in the monocular patching group; 5/29 (17%) in the binocular patching group	7/14 (50%) secondary hemorrhages; 4 in the monocular group; 3 in the binocular group	57% (8/14) of patients with final visual acuity of 20/50 or better in the monocular group; 62% (5/8) of patients with final visual acuity of 20/50 or better in the binocular group
	Blood filling ½ or more of anterior chamber	8/64 (12%) participants; 5/35 (14%) in the monocular patching group; 3/29 (10%) in the binocular patching group
Ambulatory vs. conservative treatment
Read 1974	Blood filling < ⅓ of anterior chamber	79/137 (58%) participants; 47/71 (66%) in the ambulatory group; 32/66 (48%) in the conservative group	16/30 (53%) secondary hemorrhages; 9 in the ambulatory group; 7 in the conservative group	NR
	Blood filling ⅓ to ½ of anterior chamber	28/137 (20%) participants; 11/71 (16%) patients in the ambulatory group; 17/66 (26%) patients in the conservative group	5/30 (17%) secondary hemorrhages; 4 in the ambulatory group; 1 in the conservative group
	Blood filling ½ but not total anterior chamber	19/137 (14%) participants; 8/71 (11%) patients in the ambulatory group; 11/66 (17%) patients in the conservative group	6/30 (20%) secondary hemorrhages; 3 in the ambulatory group; 3 in the conservative group
	Total hyphema	11/137 (8%) participants; 5/71 (7%) patients in the ambulatory group; 6/66 (9%) patients in the conservative group	3/30 (10%) secondary hemorrhages; 2 in the ambulatory group; 1 in the conservative group
Elevation of head vs. laying flat
Zi 1999	Blood filling < ½ of anterior chamber and level was lower than the inferior boarder of pupil	36/74 (49%) participants; 18/35 (51%) patients with elevation of the head; 18/39 (46%) patients laying flat	NR	NR
	Blood filling ½ of anterior chamber and level was higher than the inferior border of the pupil	19/74 (26%) participants; 6/35 (17%) patients with elevation of the head; 13/39 (33%) patients laying flat	NR	NR
	Blood filling > ½ of anterior chamber or filling the entire anterior chamber	19/74 (26%) participants; 11/35 (31%) patients with elevation of the head; 8/39 (21%) patients laying flat	NR	NR
Other
Rakusin 1972 *	Blood filling < ½ of anterior chamber	213 participants	NR	1. 4% (8/213) of patients with elevated intraocular pressure across all patients 2. 22% (47/213) of patients with complications 3. 78% (166/213) of patients with final visual acuity better than 20/60
	Blood filling > ½ of anterior chamber	157 participants	NR	1. 85% (133/157) of patients with elevated intraocular pressure across all patients 2. 78% (123/157) of patients with complications 3. 28% (44/157) of patients with final visual acuity better than 20/60

* Rakusin 1972 reported severity for entire study population rather than by trials of topical corticosteroids, cycloplegics vs. miotics, monocular vs. binocular patching, and ambulatory vs. conservative treatment. See under "Other".

CI: confidence interval; n: number of participants; NR: not reported; RR: risk ratio; SD: standard deviation; SE: standard error.

Data from eight studies, all RCTs comparing aminocaproic acid versus placebo, reported results on the risk of secondary hemorrhage (Christianson 1979; Crouch 1976; Karkhaneh 2003; Kraft 1987; Kutner 1987; McGetrick 1983; Pieramici 2003; Teboul 1995). Participants who were assigned to receive aminocaproic acid, either orally or topically, experienced a secondary hemorrhage less often compared with participants receiving placebo. This association was stronger when oral aminocaproic acid was used (OR 0.25, 95% CI 0.11 to 0.57; Analysis 1.5) than when topical aminocaproic acid was used (OR 0.42, 95% CI 0.16 to 1.10; Figure 3; Analysis 2.3). Because an ITT analysis was not performed in two studies of oral aminocaproic acid, each of which excluded one participant from analysis (Kutner 1987; McGetrick 1983), we performed a sensitivity analysis to assess the effect of excluding these studies. Excluding these two studies resulted in a nonsignificant effect of aminocaproic acid (OR 0.41, 95% CI 0.16 to 1.09).

Figure 3

---

---

Forest plot of comparison: 1 Oral aminocaproic acid versus placebo, outcome: 1.5 Secondary hemorrhage.

Of the six studies comparing oral aminocaproic acid versus placebo, four excluded participants with sickle cell trait (Kraft 1987; Kutner 1987; McGetrick 1983; Teboul 1995). Crouch 1976 reported that eight participants had sickle cell trait, although the trial investigators do not say to which group these participants were assigned. The one participant who had a secondary hemorrhage in the aminocaproic acid group and two of the nine participants who had a secondary hemorrhage in the placebo group also had sickle cell trait. Of the eight participants with sickle cell trait, five rebled. Similarly, in the topical aminocaproic acid versus placebo studies, only Pieramici 2003 reported that two participants in the aminocaproic acid group and one in the placebo group had sickle cell trait but again they did not report on the rebleed rate for participants with sickle cell trait/disease.

Initial hyphema severity was reported in almost all studies. Most investigators reported initial hyphema severity by the proportion of anterior chamber filled with blood or by the height of the hyphema in millimeters. There did not appear to be any overall pattern in the proportion of participants who had a secondary hemorrhage within groups defined by initial hyphema severity. Some studies reported no effect of initial hyphema size on secondary hemorrhages (Karkhaneh 2003), or that all secondary hemorrhages occurred in initially less severe hyphemas (Kutner 1987; Teboul 1995), while other studies found evidence of a higher proportion of secondary hemorrhages when the initial hyphema was more severe (Kraft 1987).

Time to rebleed (Analysis 1.6;Analysis 2.4)

Five of the six studies that studied oral aminocaproic acid reported data on the time between the initial injury and a secondary hemorrhage. Of the 10 participants who had a secondary hemorrhage in Crouch 1976, the one participant in the aminocaproic acid rebled on day one, and the nine placebo‐treated participants rebled between days two and seven. Of the three participants in Kraft 1987 who experienced a secondary hemorrhage, the two who had received aminocaproic acid had a rebleed on days three and four, and the placebo‐treated participant rebled on day four. All three participants who rebled in Kutner 1987 were in the placebo group and rebled on day two. In the one aminocaproic acid‐treated participant who rebled in McGetrick 1983, the secondary hemorrhage occurred on day four, and three of the five participants in the placebo group rebled on day three, one on day five and one on day six. Of the three participants who rebled in Teboul 1995, one rebled on day two (placebo), one rebled on day six (aminocaproic acid), and one rebled on day seven (placebo).

The mean time to rebleed in the five participants receiving topical aminocaproic acid who rebled in Karkhaneh 2003 was 3.2 days (SD 0.5) versus 3.0 days (SD 0.8) in the seven participants who rebled in the placebo group (P value = 0.18). Pieramici 2003 reported that, of the participants in their study who rebled, those receiving topical aminocaproic acid took longer to rebleed (one participant on day six) compared with those receiving placebo (eight participants; range in days two to six). However, this result was observed after excluding one participant in the aminocaproic acid group who had taken aspirin and rebled on day three.

Overall, there appeared to be little difference in the time for a secondary hemorrhage to occur although the small numbers of events makes statistical testing unreliable.

Risk of corneal bloodstain (Analysis 1.7;Table 3)

Open in table viewer

Table 3. Risk of corneal bloodstaining

Study	Test intervention	No. with outcome/No. in group	Control intervention	No. with outcome	Total No./No. with outcome
Aminocaproic acid
Crouch 1976	Oral aminocaproic acid	0/32	Placebo	2/27	2/59
Crouch 1997	Oral aminocaproic acid	0/29	Topical aminocaproic acid	0/35	0/64
Tranexamic acid
Vangsted 1983	Tranexamic acid	0/59	Bed rest only	0/53	0/112
Varnek 1980	Tranexamic acid	1/102	Conservative treatment	0/130	1/232
Prednisone/cortisone
Spoor 1980	Oral prednisone	NR	Placebo	NR	1/43
Zetterstrom 1969	Atropine plus cortisone eyedrops	0/58	Conservative treatment	1/59	1/117
Estrogen
Spaeth 1966	Estrogen	2/39	Placebo	2/46	4/85
Nondrug medical interventions
Edwards 1973	Monocular patching	1/35	Binocular patching	1/29	2/64
Read 1974	Moderate ambulatory activity, patching and shielding of injured eye	5/71	Bed rest with elevation of the head, bilateral patches and eye shield	4/66	9/137

NR: not reported.

One study examining oral aminocaproic acid reported outcomes for corneal bloodstain (Crouch 1976). Two participants in the placebo group who also had secondary hemorrhages required surgery "due to increased intraocular pressure and early corneal bloodstaining."

Risk of peripheral anterior synechiae formation (Table 4)

Open in table viewer

Table 4. Risk of peripheral anterior synechiae

Study	Test intervention	No. with outcome/No. in group	Control intervention	No. with outcome	Total No./No. with outcome
Aminocaproic acid
Crouch 1997	Oral aminocaproic acid	NR	Topical aminocaproic acid	NR	4/64
Prednisone
Spoor 1980	Oral prednisone	0/23	Placebo	0/20	0/43
Conjugated estrogen
Spaeth 1966	Conjugated estrogens	NR	Placebo	NR	15/85
Nondrug medical interventions
Read 1974	Moderate ambulatory activity, patching and shielding of injured eye	NR	Bed rest with elevation of the head, bilateral patches and eye shield	NR	9/137

NR: not reported

Crouch 1976 reported that 14 participants experienced PAS formation in the study cohort. Although the difference between groups was reported to be nonsignificant, the number of participants for each group were not reported.

Risk of glaucoma or elevated intraocular pressure (Analysis 1.8;Analysis 1.9;Analysis 2.5;Table 5)

Open in table viewer

Table 5. Risk of elevated intraocular pressure

Study	Test intervention	No. with outcome/No. in group	Control intervention	No. with outcome	Total No./No. with outcome
Aminocaproic acid
Kraft 1987	Oral aminocaproic acid	1/24	Placebo	1/25	2/49
Kutner 1987	Oral aminocaproic acid	1/21	Placebo	3/13	4/34
Teboul 1995	Oral aminocaproic acid	3/48	Placebo	3/46	6/94
Pieramici 2003	Topical aminocaproic acid	2/24	Placebo	1/27	3/51
Palmer 1986	Standard‐dose oral aminocaproic acid	2/33	Low‐dose oral aminocaproic acid	0/26	2/59
Tranexamic acid
Vangsted 1983	Tranexamic acid	8/59	Bed rest only	6/53	14/112
Varnek 1980	Tranexamic acid	8/102	Conservative treatment	7/130	15/232
Rahmani 1999	Tranexamic acid	12/80	Placebo	12/80	24/160
Welsh 1983	Tranexamic acid	1/19	Placebo	2/20	3/39
Prednisone/cortisone
Spoor 1980	Oral prednisone	0/23	Placebo	0/20	0/43
Rahmani 1999	Oral prednisone	9/78	Placebo	12/80	21/158
Zetterstrom 1969	Atropine plus cortisone eyedrops	3/58	Conservative treatment	2/59	5/117
Nondrug medical interventions
Edwards 1973	Monocular patching	3/35	Binocular patching	0/29	3/64
Read 1974	Ambulation	17/71	Bed rest	19/66	36/137
Zi 1999	Laying on right and left lateral position	7/39	Laying in semi‐reclining position	8/35	15/74

NR: not reported.

Three studies reported the number of participants with elevated IOP in oral aminocaproic acid and placebo groups (Kraft 1987; Kutner 1987; Teboul 1995). None of the studies included participants with sickle cell disease/trait. Teboul 1995 reported that six participants (three in each group) developed transient increases in IOP that did not persist following discharge (OR 0.96, 95% CI 0.18 to 5.00). Kraft 1987 reported that two participants (one in each group) had IOP greater than 25 mmHg at follow‐up and Kutner 1987 reported that four participants (one in the aminocaproic group and three in the control group) had elevated IOP at time of discharge (summary OR 0.35, 95% CI 0.06 to 1.98) (Analysis 1.8).

One study involving topical aminocaproic acid reported a nonsignificant increase in the number of participants using aminocaproic acid who had elevated IOP during the seven‐day trial compared with participants using placebo (OR 2.36, 95% CI: 0.20 to 27.85) (Pieramici 2003). This study enrolled three participants (6%) with sickle cell disease/trait, but it was not clear if any of these participants developed elevated IOP. The other study involving topical aminocaproic acid reported no significant differences in initial or final IOP between treatment groups (Karkhaneh 2003).

Risk of optic atrophy (Analysis 1.10;Table 6)

Open in table viewer

Table 6. Risk of optic atrophy

Study	Test intervention	No. with outcome/No. in group	Control intervention	No. with outcome	Total No./No. with outcome
Aminocaproic acid
Crouch 1976	Oral aminocaproic acid	0/32	Placebo	2/27	2/59
Crouch 1997	Oral aminocaproic acid	0/29	Topical aminocaproic acid	0/35	0/64
Tranexamic acid
Varnek 1980	Tranexamic acid	1/102	Conservative treatment	0/130	1/232
Cortisone
Zetterstrom 1969	Atropine plus cortisone eyedrops	0/58	Conservative treatment	1/59	1/117
Nondrug medical interventions
Read 1974	Moderate ambulatory activity, patching and shielding of injured eye	NR	Bed rest with elevation of the head, bilateral patches and eye shield	NR	8/137

NR: not reported.

Crouch 1976 reported that two participants (7.4%) in the placebo group, and none in the aminocaproic acid group developed optic atrophy. This difference was not statistically significant.

Adverse effects (Analysis 1.11;Table 7;Table 8)

Open in table viewer

Table 7. Risk of other ocular events

Study	Outcome	Test intervention	No. with outcome/No. in group	Control intervention	No. with outcome	Total No./No. with outcome
Aminocaproic acid
Crouch 1997	Conjunctival/corneal foreign body sensation	Topical aminocaproic acid	4/35	Oral aminocaproic acid	0/29	4/64
	Transient punctate corneal staining	Topical aminocaproic acid	3/35	Oral aminocaproic acid	0/29	3/64
Tranexamic acid
Varnek 1980	Vitreous and retinal hemorrhage	Tranexamic acid	5/102	Conservative treatment	5/130	10/232
	Traumatic cataract	Tranexamic acid	2/102	Conservative treatment	0/130	2/232
Nondrug medical intervention
Read 1974	Traumatic cataract	Moderate ambulatory activity, patching and shielding of injured eye	NR	Bed rest with elevation of the head, bilateral patches and eye shield	NR	8/137
	Vitreous hemorrhage	Moderate ambulatory activity, patching and shielding of injured eye	NR	Bed rest with elevation of the head, bilateral patches and eye shield	NR	11/137
	Commotio retinae	Moderate ambulatory activity, patching and shielding of injured eye	NR	Bed rest with elevation of the head, bilateral patches and eye shield	NR	4/137
	Occluded pupil	Moderate ambulatory activity, patching and shielding of injured eye	NR	Bed rest with elevation of the head, bilateral patches and eye shield	NR	2/137
	Optic atrophy with nasalization of optic cup	Moderate ambulatory activity, patching and shielding of injured eye	NR	Bed rest with elevation of the head, bilateral patches and eye shield	NR	4/137
	Optic atrophy without nasalization of optic cup	Moderate ambulatory activity, patching and shielding of injured eye	NR	Bed rest with elevation of the head, bilateral patches and eye shield	NR	8/137

NR: not reported.

Open in table viewer

Table 8. Risk of nonocular adverse effects

Study ID	Comparison	Type of complication	Results
Aminocaproic acid
Kraft 1987	Oral aminocaproic acid vs. placebo	Nausea	Drug group: 8 of 24; placebo group 1 of 25
Kutner 1987	Oral aminocaproic acid vs. placebo	Nausea or vomiting	Drug group: 6 of 21; placebo group: 0 of 13
		Light headedness	Drug group: 7 of 21; placebo group: 1 of 13
		Systemic hypotension	Drug group: 4 of 21; placebo group: 1 of 13
		Total complications	Drug group: 10 of 21; placebo group: 1 of 13
McGetrick 1983	Oral aminocaproic acid vs. placebo	Nausea or vomiting	Drug group: 6 of 28; placebo group: 0 of 20
		Diarrhea	Drug group: 2 of 28; placebo group: 0 of 20
		Muscle cramps	Drug group: 1 of 28; placebo group: 0 of 20
Pieramici 2003	Topical aminocaproic acid vs. placebo	Systemic hypotension	Drug group: 3 of 24; placebo group: 3 of 27
Crouch 1997	Oral vs. topical aminocaproic acid	Dizziness, nausea, vomiting	Oral group: 5 of 29; topical group: 1 of 35
Palmer 1986	Low‐dose vs. standard‐dose oral aminocaproic acid	Nausea or vomiting	Low‐dose group: 5 of 25; standard‐dose group: 9 of 33
		Dizziness and hypotension	Low‐dose group: 0 of 25; standard‐dose group: 5 of 33
		Syncope	Low‐dose group: 0 of 25; standard‐dose group: 2 of 33
		Diarrhea	Low‐dose group: 1 of 25; standard‐dose group: 0 of 33
		Rash or pruritis	Low‐dose group: 1 of 25; standard‐dose group: 2 of 33
		Hot flashes	Low‐dose group: 1 of 25; standard‐dose group: 0 of 33
		Dry mouth or nose	Low‐dose group: 1 of 25; standard‐dose group: 0 of 33
Farber 1991	Oral aminocaproic acid vs. oral prednisone	Any adverse event	Aminocaproic acid group: 0 of 56; prednisone group; 0 of 56
Tranexamic acid
Welsh 1983	Tranexamic acid vs. placebo	Nausea	Drug group: 1 of 19; placebo group: 0 of 20
Rahmani 1999	Tranexamic acid vs. placebo	Nausea	Drug group: 0 of 80; placebo group: 0 of 80
Aminomethylbenzoic acid
Liu 2002	Oral aminomethylbenzoic acid vs. placebo	Nausea and vomiting	Drug group: 7 of 60; placebo group: NR

NR: not reported.

Nausea and vomiting occurred significantly more often in participants treated with oral aminocaproic acid than in participants treated with placebo. In three studies that reported the occurrence of nausea and vomiting in the aminocaproic acid group compared with the placebo group, the summary OR was 11.76 (95% CI 2.59 to 53.46; Analysis 1.11) (Kraft 1987; Kutner 1987; McGetrick 1983).

In addition, McGetrick 1983 reported that two participants experienced diarrhea and one participant had muscle cramps; all were in the group treated with oral aminocaproic acid. No participants in Kutner 1987 had diarrhea or muscle cramps, but 10 (45%) of the participants in the aminocaproic acid group had at least one complication compared with only one participant (8%) in the placebo group (P value < 0.02). Other than nausea and vomiting, complications reported in Kutner 1987 included light‐headedness and systemic hypotension. Systemic hypotension was also observed in 13% of participants in the topical aminocaproic acid group versus 11% of participants in the placebo group in Pieramici 2003.

Duration of hospitalization (Analysis 1.12)

The duration of hospitalization was reported by two studies, although not enough details were provided to perform a meta‐analysis. McGetrick 1983 reported that the mean duration of hospitalization was 5.7 days for the aminocaproic acid group and 7.3 days for the placebo group. The difference was not statistically significant. This trend was the reverse in Teboul 1995, in which the aminocaproic acid group had a longer hospital stay (7.3 days) compared with the placebo group (5.4 days) (P value < 0.001).

Low‐ versus standard‐dose aminocaproic acid

Visual acuity (Analysis 3.1)

Only one study (Palmer 1986) compared low‐dose (50 mg/kg) versus standard dose (100 mg/kg) of oral aminocaproic acid, so we did not perform meta‐analyses for any outcome. Although "final" VA was measured, the time from injury to final VA was not reported. Final VAs of 20/40 or better were attained by 16 of 25 (64.0%) participants receiving low‐dose aminocaproic acid and by 25 of 32 (78.1%) participants receiving standard‐dose aminocaproic acid. These results were not statistically different (P value = 0.24).

Time to resolution of primary hemorrhage (Analysis 3.2)

No significant difference was reported between groups regarding time to resolution of the primary hemorrhage. The mean time for resolution of the primary hemorrhage was 3.1 days (SD 2.3) in the low‐dose group and 3.3 days (SD 1.8) in the standard‐dose group (Analysis 3.2).

Risk of secondary hemorrhage (Analysis 3.3;Table 2)

The investigators reported that one of 25 (4.0%) eyes receiving low‐dose aminocaproic acid rebled, and five of 33 (15.2%) eyes receiving the standard dose of aminocaproic acid rebled. These results were not statistically different (P value = 0.20). Participants with sickle cell trait were excluded from this study, and there did not appear to be an effect of initial hyphema severity on the rate of secondary hemorrhage.

Time to rebleed (Analysis 3.4)

The one participant who rebled in the low‐dose group rebled on day four. Of the five participants who rebled in the standard‐dose group, one did so on day two, two on day three, and two on day six.

Risk of corneal bloodstain

Palmer 1986 did not report this outcome.

Risk of peripheral anterior synechiae formation

Palmer 1986 did not report this outcome.

Risk of glaucoma or elevated intraocular pressure (Analysis 3.5;Table 5)
Two participants in the standard‐dose group experienced elevated IOP requiring surgical intervention. No elevated IOP was observed in the low‐dose group; however, the groups were not statistically different (P value = 0.36).

Risk of optic atrophy
Palmer 1986 did not report this outcome.

Adverse effects (Analysis 3.6;Table 8)

There were no significant differences in adverse events reported between groups (Table 8). Nausea or vomiting was reported in five participants in the low‐dose group and in nine participants in the standard‐dose group (P value = 0.52). Dizziness and hypotension were reported in five participants in the standard‐dose group, and syncope was reported in two participants in the standard‐dose group. Other adverse events in the low‐dose group included diarrhea and dry mouth or nose, each had one participant. Rash or pruritis was reported in one participant in the low‐dose group and in two participants in the standard‐dose group.

Duration of hospitalization (Analysis 3.7)

The duration of hospitalization was not statistically different between groups. The mean hospital stay was 5.4 days (SD 1.1) in the low‐dose group and 5.5 days (SD 1.4) in the standard‐dose group (P value = 0.76).

Oral versus topical aminocaproic acid

Visual acuity (Analysis 4.1)

Results for final (short‐term) VA were reported by Crouch 1997. Final VAs of 20/40 or better were attained by 20 of 29 (85.7%) participants receiving oral aminocaproic acid and by 30 of 35 (69.0%) participants receiving topical aminocaproic acid. These results were not statistically different (P value = 0.11).

Time to resolution of primary hemorrhage

Crouch 1997 did not report this outcome.

Risk of secondary hemorrhage (Analysis 4.2)

We did not perform meta‐analysis because only one study compared oral with topical aminocaproic acid (Crouch 1997). The number of secondary hemorrhages was not statistically different between groups: one of 29 (3%) eyes in the oral group versus one of 35 (3%) eyes in the topical group (P value = 0.89). Two participants in each of the treatment groups had sickle cell trait, but there was no report on the rate of secondary hemorrhage by this condition or by initial hyphema severity.

Time to rebleed (Analysis 4.3)

Crouch 1997 reported that the secondary hemorrhage in the participant in the oral aminocaproic acid group occurred on day three and the secondary hemorrhage in the participant in the topical aminocaproic acid group occurred on day five.

Risk of corneal bloodstain (Analysis 4.4;Table 3)

No incident corneal bloodstaining was reported in either the oral or topical aminocaproic acid groups (Crouch 1997).

Risk of peripheral anterior synechiae formation

Crouch 1997 reported that four participants experienced PAS formation, but the number of participants for each group were not reported.

Risk of glaucoma or elevated intraocular pressure

Crouch 1997 did not report this outcome.

Risk of optic atrophy (Analysis 4.5;Table 6)

No incident optic atrophy was reported in either the oral or topical aminocaproic acid groups (Crouch 1997).

Adverse effects (Analysis 4.6;Table 7;Table 8)

There were no significant differences in adverse events reported between groups. Of the 35 participants in the topical aminocaproic acid group, four reported feeling a conjunctival or corneal foreign body sensation, three experienced transient punctate corneal staining, and one had dizziness, nausea, and vomiting on two occasions. Five of the 29 participants in the oral aminocaproic acid group had dizziness, nausea, and vomiting (Analysis 4.6).

Tranexamic acid versus control

Visual acuity (Analysis 5.1)

We analyzed data from five studies reporting results comparing tranexamic acid versus control (Rahmani 1999; Sukumaran 1988; Vangsted 1983; Varnek 1980; Welsh 1983). Three studies were RCTs, and two were quasi‐randomized controlled clinical trials. Short‐term VA was reported by four of these studies. VA was measured by Rahmani 1999 at the time of discharge (range five to 15 days); 41 of 77 (57%) participants in the tranexamic acid group had VA of 20/40 or better compared with 35 of 79 (44%) participants in the placebo group. These results were not statistically different (P value = 0.23). However, we did not perform an ITT analysis because VA measurements were missing for three excluded participants in the tranexamic acid group, and for one excluded participant in the control group. Sukumaran 1988 reported that all participants had a final VA of 20/30 or better with the exception of one participant in the control group. The time of measurement for final VA was not reported but participants were followed up for only one week. Vangsted 1983 reported that all 59 participants in the tranexamic acid group had VA between 20/20 and 20/40 two weeks after the initial trauma. In the control group, all 53 participants had VA between 20/20 and 20/50 two weeks after the initial trauma. A meta‐analysis of these three studies showed no statistically significant effect of tranexamic acid (OR 1.65, 95% CI 0.91 to 2.99; Figure 4). In addition, Varnek 1980 reported mean VAs of 0.9 in both the tranexamic acid and control groups at day five after the trauma. VA was not reported by Welsh 1983.

Figure 4

---

---

Forest plot of comparison: 5 Tranexamic acid versus control, outcome: 5.1 Short‐term visual acuity from 20/20 to 20/40.

Time to resolution (Analysis 5.2)

Rahmani 1999 found no significant difference for time to primary resolution between groups who received tranexamic acid (mean 4.0 days, SD 2.2) versus placebo (mean 3.7 days, SD 1.6) after excluding participants who had secondary hemorrhages. Sukumaran 1988 also found no difference in time to resolution between groups, but included participants with and without secondary hemorrhages in the analysis (tranexamic group; mean 4.0, SD 2.4 versus control group; mean 3.9, SD 2.4). Although Welsh 1983 did not report time to resolution of the primary hyphema directly, the group estimated the daily rate of improvement in the hyphema by calculating the geometric mean of the per cent area of the hyphema remaining at each day following injury. These calculations indicated that tranexamic acid‐treated hyphemas cleared faster than those treated with placebo.

Risk of secondary hemorrhage (Analysis 5.3;Table 2)

All five studies reported the risk of a secondary hemorrhage. Using a fixed‐effect model, the summary OR comparing oral tranexamic acid to placebo or control was 0.25 (95% CI 0.13 to 0.49). This result was significant with P value < 0.05 and no statistical heterogeneity detected (I² = 0%) (Figure 5).

Figure 5

---

---

Forest plot of comparison: 5 Tranexamic acid versus control, outcome: 5.3 Secondary hemorrhage.

No study that evaluated tranexamic acid reported on the presence of sickle cell trait. Two of the studies had all white populations, thus it would be unlikely for any participant to have this condition (Rahmani 1999; Varnek 1980). Although initial hyphema severity was reported by all investigators, only Rahmani 1999 reported the proportion of secondary hemorrhages in groups defined by the severity of the initial hyphema, finding no effect of severity on rebleed rate. Varnek 1980 reported that the initial size of the hyphemas that underwent secondary hemorrhage was 1.0 mm (one secondary hemorrhage) in the study group and 2.2 mm (12 secondary hemorrhages) in the control group.

Time to rebleed (Analysis 5.4)

Three studies reported the time interval between the initial injury and the time of the secondary hemorrhage (Rahmani 1999; Sukumaran 1988; Varnek 1980). In Rahmani 1999, the mean time to rebleed in eight participants who experienced a secondary hemorrhage in the tranexamic acid group was 3.4 days (SD 0.7) compared with 3.8 days (SD 1.0) in the 21 participants who rebled in the placebo group. This difference was reported as not significant. In Sukumaran 1988, rebleeding occurred between days two and three in the participants who rebled in either group, and Varnek 1980 reported that the secondary hemorrhage took place at day three in the two participants in the tranexamic group who experienced this event. The time to rebleed ranged from day two to day seven in the 12 participants who rebled in the control group.

Risk of corneal bloodstain (Analysis 5.5;Table 3)

Two studies reported corneal bloodstaining as an outcome. Vangsted 1983 observed corneal bloodstaining in one participant in the control group of 53, and Varnek 1980 reported observing no corneal bleeding in either the tranexamic acid group or the placebo group.

Risk of peripheral anterior synechiae formation

This outcome was not reported by any study comparing tranexamic acid with control.

Risk of glaucoma or elevated intraocular pressure (Analysis 5.6;Table 5)
Four of the five studies reported the number of participants with transient increases in IOP in each group following the treatment period (Rahmani 1999; Vangsted 1983; Varnek 1980; Welsh 1983). None of the studies reported including participants with sickle cell disease/trait. Rahmani 1999 defined elevated IOP as greater than 21 mmHg during the hospital stay and requiring medical or surgical treatment or both. Vangsted 1983 and Varnek 1980 defined transient elevated IOP as 25 mmHg or greater. Welsh 1983 did not define IOP by a pressure level but stated that three participants required surgery for elevated IOP. The summary OR was 1.23 (95% CI 0.70 to 2.16) when comparing tranexamic acid versus control (Figure 6). In addition, Vangsted 1983 reported no instances of secondary glaucoma.

Figure 6

---

---

Forest plot of comparison: 5 Tranexamic acid versus control, outcome: 5.6 Incidence of glaucoma or increased intraocular pressure (IOP).

Risk of optic atrophy (Analysis 5.7;Table 6)
Varnek 1980 reported one incident of optic atrophy in the tranexamic acid group and none in the placebo group.

Adverse effects (Analysis 5.8;Table 8)
Welsh 1983 reported that one of 19 participants receiving tranexamic acid complained of nausea. Rahmani 1999 reported that medical staff observed no adverse events in either the drug‐treated or control group.

Duration of hospitalization (Analysis 5.9)
Three studies reported on the length of hospitalization (Rahmani 1999; Vangsted 1983; Varnek 1980). The mean hospital stay for participants receiving tranexamic acid in Rahmani 1999 was six days (SD 1.6), and that of participants in the control group was 6.3 days (SD 1.8). This difference was not significant. Vangsted 1983 reported that the mean length of hospitalization for the tranexamic acid group was six days compared with seven days for the control group. The length of hospitalization for the tranexamic acid group in Varnek 1980 was 6.8 days compared with 6.5 days for the control group.

One study reported the mean number of days off work (Vangsted 1983). The mean period off work for the tranexamic acid group was 17 days compared with 20 days for the control group.

Aminomethylbenzoic acid versus placebo

We did not perform meta‐analysis because only one study (Liu 2002) compared aminomethylbenzoic acid with placebo.

Visual acuity
Liu 2002 did not report this outcome.

Time to resolution of primary hemorrhage
Liu 2002 did not report this outcome.

Risk of secondary hemorrhage (Analysis 6.1)
Liu 2002 reported that participants treated with oral aminomethylbenzoic acid were less likely to rebleed compared with participants treated with placebo (OR 0.07, 95% CI 0.01 to 0.32).

Time to rebleed
Liu 2002 did not report this outcome.

Risk of corneal bloodstain
Liu 2002 did not report this outcome.

Risk of peripheral anterior synechiae formation
Liu 2002 did not report this outcome.

Risk of glaucoma or elevated intraocular pressure
Liu 2002 did not report this outcome.

Risk of optic atrophy
Liu 2002 did not report this outcome.

Adverse events (Table 8)
Of the 60 participants who received oral aminomethylbenzoic acid, seven reported nausea and vomiting. Adverse events for the placebo group were not reported.

Duration of hospitalization
Liu 2002 did not report this outcome.

Corticosteroids versus control

Visual acuity (Analysis 7.1;Analysis 7.2;Analysis 8.1)
Two studies compared oral corticosteroids versus placebo. VA outcomes between studies could not be combined because they were assessed at different follow‐up times and participants were divided by cut points into different levels of VA. Spoor 1980 reported that 21 of 23 (91%) participants in the prednisone group achieved final VA between 20/20 and 20/50 compared with 18 of 20 (90%) participants in the placebo group (P value = 0.88). In Rahmani 1999, short‐term VA was compared for participants in each treatment group. At time of discharge (range five to 12 days), 40 of 75 (53%) participants in the corticosteroid group had VA of 20/40 or better compared with 35 of 80 (44%) participants in the placebo group. These results were not statistically different (P value = 0.23).

Two studies administering topical corticosteroids reported short‐term VA. Again, the VA outcomes could not be combined because different cut points were used across studies (Rakusin 1972; Zetterstrom 1969). Rakusin 1972 reported that six of 13 (46%) participants assigned to corticosteroid eyedrops and 13 of 21 (62%) participants assigned to the control eyedrops achieved short‐term VA better than 20/60. Zetterstrom 1969 reported that 56 of 58 (97%) participants in the corticosteroid group had final VA of 0.9 (between 20/20 and 20/25), and 53 of 59 (90%) in the control group achieved VA better than 0.7 (about 20/30). At discharge, mean VA in the group assigned to corticosteroids was 0.96, compared with 0.91 in the control group.

Time to resolution of primary hemorrhage (Analysis 7.3;Analysis 8.2;Table 2)
In one of the two studies that evaluated oral corticosteroids, Spoor 1980 reported means of 4.4 days and 4.5 days for the resolution of primary hemorrhage in groups receiving prednisone and placebo, respectively. This result remained nonsignificant when we excluded participants who rebled from the analysis. Spoor 1980 reported that the time to resolution was shorter in hyphemas that were less severe initially. Rahmani 1999 also found no significant difference for time to primary resolution in participants who had not experienced a secondary hemorrhage and were assigned to prednisolone (mean 3.5 days, SD 1.8) or placebo (mean 3.7 days, SD 1.6). In the one study evaluating topical corticosteroids that measured time to resolution of primary hemorrhage, Rakusin 1972 reported that the primary hyphema was resolved within one week in 10 of 13 (77%) participants assigned to corticosteroid eyedrops and in 16 of 21 (76%) participants assigned to the control group.

Risk of secondary hemorrhage (Analysis 7.4;Analysis 8.3;Table 2)
We analyzed data from two studies evaluating systemic corticosteroids and reporting results for the risk of secondary hemorrhage (Rahmani 1999; Spoor 1980). Using a fixed‐effect model, the summary OR comparing oral corticosteroids to placebo was 0.61 (95% CI 0.31 to 1.22; Analysis 7.4); however, we did not perform an ITT analysis due to missing data from the exclusion of four participants by Rahmani 1999. A meta‐analysis of secondary hemorrhage including data from Rakusin 1972 (topical corticosteroids versus placebo eyedrops) and Zetterstrom 1969 (topical corticosteroids versus complete bed rest with no simultaneous local therapy) did not show a statistically significant difference (OR 0.27, 95% CI 0.05 to 1.61; Analysis 8.3).

None of the four studies reported on the presence of sickle cell trait.

Rahmani 1999 observed no effect of initial hyphema severity on the proportion of participants with a secondary hemorrhage, but Spoor 1980 found that there was a lower proportion of secondary hemorrhages in participants with less severe initial hyphemas (2/38 (13%) versus 2/5 (40%) where severity was defined as blood filling one‐third versus more than one‐third of the anterior chamber).

Time to rebleed (Analysis 7.5)
In Rahmani 1999, rebleeding occurred a mean of 3.2 days (SD 0.8) from the time of trauma in the 14 participants who rebled in the prednisolone group and 3.8 days (SD 1.0) in the 21 participants who rebled in the placebo group. This difference was reported as not significant. In Spoor 1980, the mean time to rebleed in three participants who experienced a secondary hemorrhage in the prednisone group was 2.3 days compared with 2.6 days in the four participants who rebled in the placebo group. Like the Rahmani study, this difference was not significant.

Risk of corneal bloodstain (Analysis 7.6;Analysis 8.4;Table 3)
One of 43 participants included in Spoor 1980 experienced corneal bloodstaining. The study group in which the bloodstain occurred was not reported. In Zetterstrom 1969, one participant in the control group experienced corneal bloodstaining compared with none in the group receiving corticosteroid eyedrops.

Complications of hyphema, including corneal bloodstaining; pigment on endothelium, anterior lens capsule, or vitreous; posterior synechiae; PAS; anterior chamber blood clots; and fibrous membrane formation, were documented among participants in Rakusin 1972. It was reported that 54% of the corticosteroid group had complications compared with 70% of the control group, although this difference was not significant and the risk of corneal bloodstaining was not reported separately.

Risk of peripheral anterior synechiae formation (Analysis 7.7;Table 4)
Spoor 1980 reported that there was no instance of PAS formation in either group.

Risk of glaucoma or elevated intraocular pressure (Analysis 7.8;Analysis 8.5;Table 5)
Rahmani 1999 reported that nine (11.5%) of 78 participants in the prednisolone group and 12 (15%) of 80 participants in the placebo group had an IOP greater than 21 mmHg during hospitalization that required medical treatment, surgical treatment, or both. This difference was not significant. Two participants studied by Spoor 1980 had elevated IOP that was controlled by acetazolamide therapy alone; one participant was in the prednisolone group, and one was in the control group. No participant in this cohort had IOP greater than 35 mmHg. Five participants in Zetterstrom 1969 developed "elevated" IOP (undefined); three of 58 in the group assigned to topical corticosteroids and two of 59 in the control group (Analysis 7.8).

Risk of optic atrophy (Analysis 8.6)
One incident of optic atrophy was reported by Zetterstrom 1969 in the group of 58 participants assigned to topical corticosteroid eyedrops.

Adverse effects
Rahmani 1999 reported that medical staff observed no adverse events in either the drug‐treated or control groups.

Duration of hospitalization (Analysis 7.9;Analysis 8.7)
In Rahmani 1999, participants treated with prednisolone were hospitalized a mean of 5.9 days (SD 1.4) and participants treated with placebo were hospitalized a mean of 6.3 days (SD 1.8). The mean difference between groups was ‐0.40 days (95% CI ‐0.90 to 0.10).

Zetterstrom 1969 reported duration of hospitalization, finding that the mean length of stay for participants assigned to corticosteroid drops was 5.9 days compared with 8.9 days for participants assigned to the control group.

Oral aminocaproic acid versus oral prednisone

Visual acuity (Analysis 9.1)
We performed no meta‐analysis because only one study compared oral aminocaproic acid versus oral prednisone (Farber 1991). After five days of hospitalization, 10 of 56 (18%) participants in the aminocaproic acid group had short‐term VA of 20/200 or worse compared with seven of 56 (12.5%) participants in the prednisone group. These results were not statistically different (P value = 0.43). Likewise, there was no difference in final VA of 20/40 or better between groups (26 of 56 (46%) participants in the aminocaproic acid group and 31 of 56 (55%) participants in the prednisone group).

Time to resolution of primary hyphema
Farber 1991 did not follow the participants past discharge and so did not report on time to resolution of the primary hyphema. They did report however that "at discharge" (mean time to discharge = five days) 43% of the aminocaproic acid group compared with 75% of the prednisone groups had complete resolution of their hyphema. This difference was statistically significant (P value = 0.001).

Risk of secondary hemorrhage (Analysis 9.2;Table 2)
The risk of secondary hemorrhage was equal for both groups; four eyes out of 56 eyes per group (P value = 1.00). Participants with sickle cell trait/disease were excluded from this study. There did not appear to be an influence of initial hyphema severity on rate of secondary hemorrhage.

Time to rebleed
Farber 1991 did not report this outcome.

Risk of corneal bloodstain
Farber 1991 did not report this outcome.

Risk of peripheral anterior synechiae formation
Farber 1991 did not report this outcome.

Risk of glaucoma or elevated intraocular pressure
No significant differences were reported for mean IOPs at time of discharge between groups.

Risk of optic atrophy
Farber 1991 did not report this outcome.

Adverse events
Farber 1991 did not report this outcome.

Duration of hospitalization
Farber 1991 did not report this outcome.

Conjugated estrogen versus placebo

Visual acuity
VA at time of discharge was partially reported by the one study that compared conjugated estrogen versus placebo (Spaeth 1966). Among all participants, 61% had VA better than 6/12, 30% had VA better than 6/60, and 9% had VA of 6/60 or worse at time of discharge. These results were not reported by treatment groups.

Time to resolution of primary hyphema
Spaeth 1966 did not report this outcome.

Risk of secondary hemorrhage (Analysis 10.1;Table 2)
It was reported that of 39 estrogen‐treated participants, 10 rebled (25.6%) and of 46 placebo‐treated participants, 10 rebled (21.7%). These results were not statistically different (P value = 0.67).

Spaeth 1966 did not report on the presence of sickle cell trait/disease. The risk of secondary hemorrhage by initial hyphema severity did not appear to differ across severity ratings.

Time to rebleed
The time to rebleed, reported not by treatment group but overall, was a mean of 3.5 days after injury with a range of one to eight days.

Risk of corneal bloodstain (Analysis 10.2)
In the estrogen group, two of 39 (5%) participants had corneal bloodstaining compared with two of 46 (4%) participants in the placebo group (OR 1.19, 95% CI 0.16 to 8.86).

Risk of peripheral anterior synechiae formation
Fifteen cases of PAS were reported among all participants; however, the number of cases by treatment group were not reported.

Risk of glaucoma or elevated intraocular pressure
Thirteen cases of secondary glaucoma were reported among all participants; however, the number of cases by treatment group were not reported. Four of these 13 cases occurred prior to secondary hemorrhage.

Risk of optic atrophy
Spaeth 1966 did not report this outcome.

Adverse events
Spaeth 1966 did not report this outcome.

Duration of hospitalization
Spaeth 1966 did not report this outcome.

Cycloplegics versus miotics

Short‐term visual acuity (Analysis 11.1)
Two studies looked at the effect of cycloplegics compared with miotics (Bedrossian 1974; Rakusin 1972). Rakusin 1972 reported that nine of 17 (53%) participants in the homatropine group and 11 of 17 (65%) participants in the pilocarpine group had short‐term VA better than 20/60. Bedrossian 1974 did not report on VA.

Time to resolution (Analysis 11.2;Table 2)
Bedrossian 1974 reported a longer time to resolution with the pilocarpine group (mean 3.6 days, SD 1.3) compared with the atropine group (mean 2.7 days, SD 1.7). The time to resolution showed a slight increase with larger size of initial hyphema. In Rakusin 1972, there was no significant difference between the proportion of participants with absorption within one week between cycloplegic (12/17) and miotic (13/17) groups.

Risk of secondary hemorrhage (Analysis 11.3;Table 2)
In Bedrossian 1974, only one participant experienced a secondary hemorrhage; that participant was in the cycloplegic group and had an initial hyphema height of 1 mm. The one participant with a secondary hemorrhage in Rakusin 1972 was in the group receiving homatropine (Analysis 11.3).

Time to rebleed (Analysis 11.4)
Bedrossian 1974 reported that the time to rebleed in the one individual with a secondary hyphema was two days.

Risk of corneal bloodstain
It was reported that the number of complications of hyphema, including corneal bloodstaining; pigment on endothelium, anterior lens capsule, or vitreous; posterior synechiae; PAS; anterior chamber blood clots; and fibrous membrane formation, were similar in all groups in Rakusin 1972.

Risk of peripheral anterior synechiae formation
It was reported that the number of complications of hyphema, including corneal bloodstaining; pigment on endothelium, anterior lens capsule, or vitreous; posterior synechiae; PAS; anterior chamber blood clots; and fibrous membrane formation, were similar in all groups in Rakusin 1972.

Risk of glaucoma or elevated intraocular pressure
Bedrossian 1974 and Rakusin 1972 did not report this outcome.

Risk of optic atrophy
Bedrossian 1974 and Rakusin 1972 did not report this outcome.

Adverse events
Bedrossian 1974 and Rakusin 1972 did not report this outcome.

Duration of hospitalization
Bedrossian 1974 and Rakusin 1972 did not report this outcome.

Aspirin versus observation

Because only one study compared aspirin versus observation (Marcus 1988), we did not perform a meta‐analysis.

Visual acuity
Marcus 1988 did not report this outcome.

Time to resolution
Marcus 1988 did not report this outcome.

Risk of secondary hemorrhage (Analysis 12.1)
Marcus 1988 reported that three of 23 (13%) eyes receiving aspirin rebled and two of 28 (7%) eyes receiving observation rebled. These results were not statistically different (P value = 0.49). The study investigators reported that two of the three eyes that rebled in the aspirin group initially had a total hyphema, while of the two eyes that rebled in the control group, one had an initial hyphema of 30% and one an "almost total" hyphema.

Time to rebleed
Marcus 1988 did not report this outcome.

Risk of corneal bloodstain
Marcus 1988 did not report this outcome.

Risk of peripheral anterior synechiae formation
Marcus 1988 did not report this outcome.

Risk of glaucoma or elevated intraocular pressure
Marcus 1988 did not report this outcome.

Risk of optic atrophy
Marcus 1988 did not report this outcome.

Adverse events
Marcus 1988 did not report this outcome.

Duration of hospitalization
Marcus 1988 did not report this outcome.

Traditional Chinese medicine versus control

We did not perform a meta‐analysis for TCM versus control treatment since only one study evaluated these interventions (Wang 1994). The authors of Wang 1994 reported only one outcome: the proportion of patients who were "cured". The outcome of being cured was a composite outcome defined as the resolution of the primary hemorrhage after five days of treatment, VA of 0.7 or better after resolution of the primary hemorrhage, and no recurrence of bleeding for one week following resolution of the primary hemorrhage. One week after completing treatment, 29 of 45 (64%) participants in the TCM group and 10 of 38 (26%) participants in the control group met this criteria for being "cured".

Visual acuity
Wang 1994 did not report this outcome.

Time to resolution
Wang 1994 did not report this outcome.

Secondary hemorrhage (Analysis 12.1)
Wang 1994 did not report this outcome.

Time to rebleed
Wang 1994 did not report this outcome.

Risk of corneal bloodstain
Wang 1994 did not report this outcome.

Risk of peripheral anterior synechiae formation
Wang 1994 did not report this outcome.

Risk of glaucoma or elevated intraocular pressure
Wang 1994 did not report this outcome.

Risk of optic atrophy
Wang 1994 did not report this outcome.

Adverse events
Wang 1994 did not report this outcome.

Duration of hospitalization
Wang 1994 did not report this outcome.

Monocular versus binocular patching

Visual acuity (Analysis 13.2;Table 2)
We identified two studies that compared the use of monocular versus binocular patches (Edwards 1973; Rakusin 1972). Rakusin 1972 reported that 22 of 26 (85%) participants in the monocular group compared with 24 of 27 (89%) participants in the binocular group had short‐term VA better than 20/60. Edwards 1973 reported that 21 of 26 (81%) participants in the monocular group had VA better than 20/50 compared with 20 of 20 (100%) participants in the binocular group, although the time at which VA was measured was not specified. Participants with an initial hyphema filling less than one‐third of the anterior chamber, 67% (28/42) had VA of 20/50 or better compared with 59% (13/22) of those with more severe hyphemas.

Time to resolution
Rakusin 1972 reported that the primary hyphema was resolved within one week in 22 of 26 (85%) participants with monocular patching and in 24 of 27 (89%) participants with binocular patching.

Risk of secondary hemorrhage (Analysis 13.3;Table 2)
In Edwards 1973, there were eight participants each with a secondary hemorrhage from both the group with a patch on both eyes (n = 35; 23%) and the group with a patch only on the injured eye (n = 29; 28%). The proportion of secondary hyphemas was greater in participants with initially more severe hyphemas (32% (seven of 22) versus 17% (seven of 42) for those with an initial hyphema filling less than one‐third of the anterior chamber versus more). The results from Rakusin 1972 also showed no difference between groups on risk of secondary hemorrhage (one of 26 (3.8%) in the group with a monocular patch and two of 27 (7.4%) in the group with binocular patches) (Analysis 13.3).

Time to rebleed (Analysis 13.4)
A mean of three days between injury and secondary hemorrhage was reported for eight individuals in the group with a monocular patch as well as for eight individuals who had a secondary hemorrhage in the group with binocular patches (Edwards 1973).

Risk of corneal bloodstain (Analysis 13.5;Table 3)
One individual in each of the two treatment groups experienced corneal bloodstaining in Edwards 1973.

It was reported that the risk of complications of hyphema, including corneal bloodstaining; pigment on endothelium, anterior lens capsule, or vitreous; posterior synechiae; PAS; anterior chamber blood clots; and fibrous membrane formation, were similar in both groups in Rakusin 1972.

Risk of peripheral anterior synechiae formation
It was reported that the risk of complications of hyphema, including corneal bloodstaining; pigment on endothelium, anterior lens capsule, or vitreous; posterior synechiae; PAS; anterior chamber blood clots; and fibrous membrane formation, were similar in both groups in Rakusin 1972.

Risk of glaucoma or elevated intraocular pressure (Analysis 13.6;Table 5)
In the Edwards 1973 study three participants in the monocular patching group developed secondary glaucoma while none in the binocular patch developed secondary glaucoma (Edwards 1973).

Risk of optic atrophy
Edwards 1973 and Rakusin 1972 did not report this outcome.

Adverse events
Edwards 1973 and Rakusin 1972 did not report this outcome.

Duration of hospitalization
Edwards 1973 and Rakusin 1972 did not report this outcome.

Quality of life
Edwards 1973 noted no difference between groups on the "cooperation index". This index included a number of outcomes including those associated with quality of life (pain, restlessness, activity, and emotional state while in the hospital).

Ambulatory versus conservative treatment

Visual acuity (Analysis 14.1)
Two studies compared ambulatory (i.e. moderate activity allowed) versus conservative treatment, which comprised bed rest alone (Rakusin 1972), or bed rest with elevation of the head, bilateral ocular patches, and a shield over the injured eye (Read 1974). In Read 1974, VA was not reported by treatment group but the authors distinguished between poor VA due to the initial trauma and that due to secondary effects of the hyphema. They stated that poor VA due to hyphema occurred in nine of 71 (13%) participants in the ambulatory group compared with four of 66 (6%) participants in the conservative group. Overall, the proportion of participants with good VA was 104 of 137 (76%) with more participants in the ambulatory group having good VA. In Rakusin 1972, 22 of 26 (85%) participants in the ambulatory group had short‐term VA better than 20/60 compared with 20 of 26 (77%) participants in the conservative group.

Time to resolution of primary hyphema (Analysis 14.2)
Read 1974 reported a mean of 5.8 days between the initial injury and resolution of the hyphema in the ambulatory group compared with 5.6 days in the group receiving bed rest. However, Rakusin 1972 observed a significant difference in the speed of reabsorption. The primary hyphema was resolved within one week in 13 of 26 (50%) participants in the ambulatory group compared with 22 of 26 (85%) participants in the conservative group.

Risk of secondary hemorrhage (Analysis 14.3;Table 2)
Eighteen of 71 (25%) participants in the ambulatory group developed a secondary hemorrhage, and 12 of 66 (18%) participants in the group receiving bed rest did so in Read 1974. This difference was not statistically significant. The proportion of participants with a secondary hemorrhage appeared to be smaller with more severe initial hyphemas (16 of 30 (53%) versus 14 of 90 (16%) for those with an initial hyphema filling less than one‐third compared with one‐third or more of the anterior chamber) (Analysis 14.3).

Time to rebleed
Read 1974 reported that the majority of secondary hemorrhages occurred between day two and day five following injury, although two secondary hemorrhages took place on day seven following the initial injury.

Risk of corneal bloodstain (Analysis 14.4;Table 3)
Nine participants in Read 1974 developed corneal bloodstaining; five of 71 (7%) participants in the ambulatory group and four of 66 (6%) participants in the group receiving bed rest.

It was reported that the risk of complications of hyphema, including corneal bloodstaining; pigment on endothelium, anterior lens capsule, or vitreous; posterior synechiae; PAS; anterior chamber blood clots; and fibrous membrane formation, were similar in both groups in Rakusin 1972.

Risk of peripheral anterior synechiae formation
It was reported that the risk of complications of hyphema, including corneal bloodstaining; pigment on endothelium, anterior lens capsule, or vitreous; posterior synechiae; PAS; anterior chamber blood clots; and fibrous membrane formation, were similar in both groups in Rakusin 1972.

Risk of glaucoma or elevated intraocular pressure (Analysis 14.5;Table 5)
Of the 71 participants in the group that was allowed moderate activity, 17 (23.9%) developed IOP of 25 mmHg or greater while 19 of the 66 (28.8%) participants in the group with bed rest developed IOP during hospitalization in Read 1974.

Risk of optic atrophy
Rakusin 1972 and Read 1974 did not report this outcome.

Adverse events
Rakusin 1972 and Read 1974 did not report this outcome.

Duration of hospitalization
Rakusin 1972 and Read 1974 did not report this outcome.

Elevation of the head versus control

One study compared elevation of the head by assigning participants to a semi‐reclined body position or to laying on their right or left side (Zi 1999).

Visual acuity
Zi 1999 did not report this outcome.

Time to resolution
Time to resolution was compared by level of hyphema. The time to resolution was somewhat shorter for participants with their head elevated compared with those laying flat if the initial hyphema filled up to half of the anterior chamber, but longer if the blood filled more than half (level of blood < one‐half of the anterior chamber: 1.7 days (n = 18) versus 2.8 days (n = 18); level of blood = one‐half of the anterior chamber: 2.2 days (n = 6) versus 3.1 days (n = 13); level of blood > one‐half of anterior chamber: 9.0 days (n = 11) versus 8.0 days (n = 8)).

Risk of secondary hemorrhage
Zi 1999 did not report this outcome.

Time to rebleed
Zi 1999 did not report this outcome.

Risk of corneal bloodstain
Zi 1999 did not report this outcome.

Risk of peripheral anterior synechiae formation
Zi 1999 did not report this outcome.

Risk of glaucoma or elevated intraocular pressure (Table 5)
Fifteen participants developed secondary glaucoma, eight of 35 (23%) in the group in the semi‐reclined position and seven of 39 (18%) in the group laying flat (Zi 1999).

Risk of optic atrophy
Zi 1999 did not report this outcome.

Adverse events
Zi 1999 did not report this outcome.

Duration of hospitalization
Zi 1999 did not report this outcome.