Pediatr Infect Dis J, 7:626-629, 1988 0891·3668/ 88/ $02.00/ 0 Copyright © 1988 by Williams & Wilkins

VoL 7, No. 9 Printed in U.S.A.

Comparison of three topical antimicrobials for acute bacterial conjunctivitis

JACOB A. LOHR, MD, ROBERT D. AUSTIN, BA, MOSES GROSSMAN, MD, GREGORY F. HAYDEN, MD,

GAIL M. KNOWLTON, MS AND SHARON M. DUDLEY, BS

One hundred fifty-eight patients, 21 years of age or less, presenting with culture-positive (Haem oph ilus influenzae or Streptococcus pneumoniae) conjunctivitis were treated with trimethoprim-polymyxin B (TP), gentamicin sulfate (GS) or sodium sulfacetamide (SS) ophthalmic solution for 10 days. Clinical re­sponse at 3 to 6 days after start of therapy was similar for all test agents: 26 of 55 (47%) pa­tients cured, 25 of 55 (45%) improved for TP; 28 of 57 (49%) cured, 26 of 57 (46%) improved for GS; and 19 of 46 (41%) cured, 22 of 46 (48%) improved for SS. Clinical response at 2 to 7 days after completion of therapy was also similar: 46 of 55 (84%) patients cured, 5 of 55 (9%) improved for TP; 50 of 57 (88%) cured, 5 of 57 (9%) improved for GS; and 41 of 46 (89%) cured, 2 of 46 (4%) improved for SS. Bacterio­logic response at 2 to 7 days after completion of therapy was similar for all antimicrobials: 44 of 55 (83%) patients for TP; 39 of 57 (68%) for GS; and 33 of 46 (72%) for SS.

INTRODUCTION

Purulent conjunctivitis is common in childhood and is a source of concern for parents and physicians. The etiology is bacterial (generally Haemophilus in{luenzae or Streptococcus pneumoniae) in approximately 50% of cases.1 Antimicrobial therapy enhances the eradi­cation of these organisms from the conjunctivae and shortens the duration of the clinical disease.2 The role of Staphylococcus aureus in nontraumatic conjuncti­vitis remains controversia1.1,3 This study compared the clinical and bacteriologic responses to a new

Accepted for publication March 2, 1988. From the Department of Pediatrics, University of Virginia Chil­

dren's Medical Center, Charlottesville, VA (JAL, GFH, SMD); Burroughs-Wellcome Company, Research Triangle Park, NC (RDA, GMK); and the University of California-San Francisco, San Francisco, CA (MG).

Reprints not available.

626

ophthalmic antimicrobial preparation, trimethoprim­polymyxin B (TP) , with the responses to gentamicin sulfate (GS) and sodium sulfacetamide (SS) in the treatment of bacterial conjunctivitis.

MATERIALS AND METHODS

Pharmacologic agents used. Trimethoprim hem­isulfate, 1.0 mg/ml, and polymyxin B, 10 000 units/ml (Polytrim®; Burroughs-Wellcome Company), genta­micin sulfate, 3.0 mg/ml (Garamycin®; Schering Cor­poration) and sulfacetamide sodium, 100 mg/ml (So­dium Sulamyd®; Schering Corporation) were the ophthalmic solutions used. The antimicrobial agents chosen for comparison with trimethoprim-polymyxin B were selected because of a similar spectrum of in vitro antibacterial activity,4 solution formulation and common use.

Population. Patients between the ages of 2 months and 22 years of either sex and any race who at the time of an acute illness visit were diagnosed as having conjunctivitis were eligible for admission to the study. Patients were recruited for study at two sites: Univer­sity of Virginia Children's Medical Center, Charlottes­ville, V A; and University of California-San Francisco, San Francisco, CA.

Study entry. The clinical diagnosis of conjuncti­vitis was based on at least three of the following criteria (adapted from Jarudi et a1.5

): ocular symptoms (e.g. itching, burning, photophobia or foreign body sensation); yellowish discharge or mattering; a poly­morphonuclear neutrophilic response on Giemsa- or Gram-stained scraping ofthe conjunctiva; an environ­mental history of close exposure to bacterial conjunc­tivitis; or a history of inadequately treated bacterial conjunctivitis. The clinical diagnosis had to be con­firmed with a laboratory diagnosis of bacterial infec­tion.

Written informed consent to participate in the study was obtained. A careful history and eye exami­nation were performed. The severity of 21 signs and 8 symptoms were graded and recorded (see "Clinical evaluation").

Patients were excluded from enrollment in the study

Vol. 7, No.9 LOHR ET AL. 627

for any of the following reasons: (1) use of any ophthalmic or systemic antimicrobial agent in the preceding 72 hours; (2) evidence of concomitant fun­gal, viral or tuberculous pathology of either eye; (3) concomitant use of systemic or topical corticosteroids, tranquilizers, antihistamines or antimicrobial agents. (4) history of hypersensitivity to any of the study medications (TP, GS or any sulfonamide); (5) more than six episodes of conjunctivitis during the preced­ing year.

Medication. Patients were assigned to receive either TP, GS or SS ophthalmic solutions according to a double blinded randomization schedule. Neither the patients nor the investigators were aware of which of the similarly bottled and similar appearing drugs the patients were receiving. One drop of medication was administered to the affected eye(s) every 3 hours, while the patient was awake, for 10 days (5 to 6 applications/day) .

Follow-up. Patients returned for a first follow-up 3 to 6 days after starting treatment and a final follow­up 2 to 7 days after completion of therapy. Signs and symptoms were evaluated at each follow-up visit; a repeat eye culture was obtained at the final visit. Compliance was assessed on the basis of a medication administration log completed by the patient or guard­ian.

Clinical evaluation. At each visit the investigator rated the presence and severity of 8 symptoms (itch­ing, burning, foreign body sensation, photophobia, grittiness, watery discharge, purulent discharge, eye­lids "stuck together" in the morning) and 21 signs (eyelid: edema, erythema, tenderness; eyelid margin: scales, erythema, ulcerations, meibomitis; palpebral conjunctiva: erythema, chemosis, papules, membrane, conjunctival hemorrhage; bulbar conjunctiva: diffuse or angular hyperemia; cornea: infiltrate, ulcer, kera­titis (punctate), edema, filaments; tear film cells; and iritis).

Each factor was rated on a 4-point scale (0 = absent, 1 = mild, 2 = moderate, 3 = severe). An average severity index (ASI) was calculated for each patient by dividing the sum of these individual sign and symp­tom scores by the number of items scored. Many patients were rated on the full battery of 29 signs and symptoms; however, some patients, especially young children with whom an investigator was unable to communicate verbally, were rated on fewer symptoms.

Microbiologic evaluation. At the initial and final visits, an eye swab specimen for culture was obtained with a calcium alginate applicator by swabbing over the lower palpebral conjunctiva of each affected eye. The swab was immediately rolled on the surface of a 5% sheep's blood agar plate and a chocolate agar plate and processed according to standard bacteriologic methods. The plates were examined 24 and 48 hours

later for growth of bacterial pathogens. Isolated orga­nisms were identified according to standard methods.

Statistical analysis. The ASIs among patients in the three treatment groups at the follow-up visits were compared using a general linear models procedure. Three pairs of comparisons were made: TP us. GS; TP us. SS; and GS us. SS. These multiple comparisons required that we adjust the "significant" value of alpha downward using the Bonferroni procedure6 such that p = 0.0167.

At the follow-up visits patients with an ASI of zero (i.e. no signs or symptoms) were considered "cured." Patients whose ASI had decreased to a value <50% of the pretreatment value were considered "improved." Patients whose ASI was ~50% to 100% of the pre­treatment value were considered "unchanged." Pa­tients whose ASI was >100% of the pretreatment value were considered "worse." Patients whose con­junctivitis was "worse" at the first follow-up visit were dropped from their treatment group and were consid­ered "failures" in the final follow-up clinical analysis, although final bacteriologic cultures were not ob­tained. Patients who were "unchanged" or "worse" at the final follow-up visit were considered "failures" in the final follow-up analysis.

A case was considered evaluable for evidence of drug efficacy ifthe patient complied with the correct dosage regimen (fewer than 20 doses missed), if the eye cul­tures were obtained and processed correctly and if the clinical outcome could be strictly classified as "cured," "improved" or "failed." In the case of bilateral disease one eye was randomly selected for analysis via a computer-generated random schedule.

RESULTS

Three hundred thirty-seven patients qualified for enrollment in the study (Table 1). One hundred thirty­two (39%) of these patients were dropped from the study before the first follow-up because of no bacterial growth (79 patients) or growth of organisms other than H. in{luenzae or S. pneumoniae on initial culture (53 patients).

The remaining 205 patients were culture-positive for H. in{luenzae, S. pneumoniae or both. Forty-seven of the 205 patients (23%) could not be considered in the final analysis because they were lost to follow-up (28 patients), had inadequate medication administra­tion (medication stopped, medication lost or 2::20 doses missed) (10 patients) or received a concomitant anti­bacterial medication (9 patients). The one patient with an adverse experience and the one patient who was a therapeutic failure at the first follow-up were dropped from their treatment groups but were included in the 158 patients considered in the final analysis. Mild lid edema, unassociated with other evidence of drug tox­iCity, was considered an adverse experience in the one

628 TOPICAL ANTIMICROBIALS FOR CONJUNCTIVITIS September, 1988

TABLE 1. Number of enrolled patients by treatment group who were eliminated from or included in the final analysis

Treatment Group

Trimethoprim - Gentamicin Sodium Total polymyxin B sulfacetamide

No. of patients enrolled in study 117 110 110 337 No. of patients eliminated from final analysis

Culture No bacterial growth 26 28 25 79 Nonpathogen(s)" only 24 12 17 53

Lost to follow-up 10 7 11 28 Inadequate medication administrationb 1 1 8 10 Concomitant antibacterial medication 1 5 3 9 Total no. 62 53 64 179

No. of patients dropped from treatment group during study but considered in final analysis

Adverse experience 1 0 0 1 Therapeutic failure 1 0 0 1 Total no. 2 0 0 2

No. of patients who completed study 53 57 46 156 No. of patients in final analysis 55 57 46 158

(Z Gram-negative, facultatively anaerobic rods; Gram-negative rods, unclassified: Escherichia sp.; Citrobacter freundii; Serratia sp.; Haenwphilus sp.; Haemophilus parain{luenzae; Pasteurella sp.; Neisseria sp.; Branhamella catarrhalis; Moraxella sp.; Moraxella lacunata; Moraxella osloensis; Moraxella atlantae; Moraxella nonliquefaciEns; Moraxella-like cocci; Acinetobacter calcoaceticu.s; Gram-positive cocci; Staphylococcus sp.; Staphylococcus aureus; Staphylococcus albus (epidermidis)j Streptococcus viridans; Streptococcus bovis; Streptococcus anhemolyticus; Sarcina sp.; Corynebacterium sp.; Corynebacterium ulcerans; Candida albicansj Trichosporon sp.; rare Gram-negative rod (probably Pseudomonas vesicularis).

b Medication stopped, medication lost or ~20 doses missed.

patient even though it was present before medication use. The patient considered to be a therapeutic failure was clinically worse but had no evidence of drug toxicity.

The mean severity index for each ofthe three groups at each of the clinical evaluations is reported in Table 2. No significant differences existed among the groups for anyone visit.

The patients in the three study groups were com­parable as to age, gender, race and pathogen (Table 3).

Clinical and bacteriologic responses by treatment groups are shown in Table 4. At the first follow-up­clinical cure or improvement was seen in 92, 95, and 89% of the patients treated with TP, GS and SS, respectively. At the final follow-up, the number of patients clinically cured, improved or failed was not statistically different for the three treatment groups (P > 0.1 by Fisher's exact test). Of the patient's considered clinical failures, all were unchanged; none was worse and none had an adverse experience. In addition no patient experienced spread from unilateral to bilateral disease.

The overall bacteriologic response was not statisti­cally different for the three treatment groups (83, 68, and 72% for TP, GS and SS, respectively) (P > 0.1 by Fisher's exact test). However, the bacteriologic cure rate for H. influenzae infections appeared different for TP (81%) us. GS (57%) (P = 0.03) and for TP (81%) us. SS (57%) (P = 0.06).

DISCUSSION

Purulent conjunctivitis in children customarily has been treated with topical antibiotics. However, the effectiveness of such therapy was not studied in a controlled manner until Gigliotti et a1. 2 examined 102

TABLE 2. Patients' mean average severity index· by treatment groupb at each visit

ASI

Time of Visit Trimethoprim- Gentamicin Sodium polymyxin

(n= 57) sulfacetamide (n= 53) (n = 46)

Pretreatment 0.65 0.72 0.65 (0.2-1.2 ± (0.3-1.5 ± (0.3-1.3 ± 0.24)' 0.23) 0.22)

First follow-up 0.10 0.10 0.12 (0.0-1.0 ± (0.0-0.7 ± (0.0-0.5 ± 0.16) 0.13) 0.15)

Final follow-up 0.03 0.04 0.03 (0.0-0.6 ± (0.0-0.8± (0.0-0.6 ± 0.09) 0.15) 0.12)

o Index of 8 symptoms and 21 signs associated with Haemophilus influenzae and/or Streptococcus pneumoniae-associated conjunctivitis.

b Patients dropped from their treatment group because they were "worse" at first follow-up (one patient) or because of an adverse reaction (one patient) were not included in the tsble because an AS! could not be determined for them at the final follow-up.

C Numbers in parentheses, range ± SD.

TABLE 3. Patients' age, gender, race and pathogen by treatment group

No. of Patients

Trimethoprim- Gentamicin Sodium polymyxin (n = 57) sulfacetamide

(n = 55) (n = 46)

Age 0-2 years 28 (53)" 31 (54) 27 (59) 3-12 years 20 (36) 18 (32) 15 (32)

13-21 years 7 (11) 8 (14) 4 (9) Gender

Male 28 (51) 28 (49) 21 (46) Female 27 (49) 29 (51) 25 (54)

Race Caucasian 23 (43) 24 (42) 21 (46) Black 14 (23) 11 (19) 11 (24) Other 18 (34) 22 (39) 14 (30)

Pathogen Haemophilus influenzae 33 (60) 30 (53) 21 (46) Streptococcus pneumoniae 22 (40) 25 (44) 25 (54) Both 0(0) 2 (3) 0(0)

o Numbers in parentheses, percent.

Vol. 7, No.9 LOHR ET AL. 629

TABLE 4. Patients' clinical and bacteriologic responses by treatment group

No. of Patients

Trimethop~im. Gentamicin Sodium, polymyxm (n = 57) sulfacetamIde (n = 55) (n = 46)

I. First follow-up Clinical response"

Cured 26 (47)" 28 (49) 19 (41) Improved 25 (45) 26 (46) 22 (48) Unchanged 2 (4) 3 (5) 5 (11) Worse 1 (2) 0(0) 0(0) Adverse experience 1 (2) 0(0) 0(0)

II, Final Follow-up Clinical response

Cured 46 (84) 50 (88) 41 (89) Improved 5 (9) 5 (9) 2 (4) Failed (unchanged) 4 (7) 2 (3) 3 (7)

Bacteriologic response ' Haemophilus influenzae

Organism eradicated 26 (81) 17 (57) 12 (57) Organism persisted 6 (19) 13 (43) 9 (43) Total infections with 32 (100) 30 (100) 21 (100)

Haemophilus influ-enzae

Streptococcus pneumoniae Organism eradicated 18 (86) 21 (84) 21 (84) Organism persisted 3 (14) 4 (16) 4 (16) Total infections with 21 (100) 25 (100) 25 (100)

Streptococcus pneu-moniae

Both organisms Organisms eradicated 0(0) 1 (50) 0(0) Organism(s) persisted 0(0) 1 (50) 0(0) Total infections with 0(0) 2 (100) 0(0)

both organisms Total

Organism(s) eradicated 44 (83) 39 (68) 33 (72) Organism(s) persisted 9 (17) 18 (32) 13 (28) Total infections 53 (100) 57 (100) 46 (100)

o Cured = ASI (see text) of zero; improved - ASI of <50% of the pretreatment value; unchanged = ASI of 2:50% to 100% of the pretreatment value; Worse s ASI of >100% of the pretreatment value; adverse experience = suspected drug toxicity.

b Numbers in parentheses, percent. C Patients "worse" at first follow-up (1) or having an adverse experience (1) were

dropped from their treatment groups and could not be evaluated bacteriologically at the final follow-up,

children with purulent conjunctivitis in a randomized, double-blind trial of polymyxin-bacitracin us. placebo treatment. Gigliotti et al. l had previously demon­strated that H. in{luenzae and S. pneumoniae were the only bacterial agents statistically associated with con­junctivitis in children older than 2 to 3 months of age. Therefore in their therapy trial, clinical and bacteri­ologic cure rates were determined for the 66 patients who had purulent conjunctivitis and isolation of H. in{luenzae, S. pneumoniae or both and who received only topical therapy_ Those treated with polymyxin­bacitracin achieved significantly higher clinical cure rates by Days 3 to 5 than did those patients treated with placebo. The clinical cure rates at Days 8 to 10 were not significantly different. The bacterial patho­gen was eradicated by medication in 71% of the pa­tients at Days 3 to 5 and in 79% at Days 8 to 10; the pathogen was recovered in 81 % at Days 3 to 5 and remained present at Days 8 to 10 in almost 70% of the placebo-treated patients. The authors concluded

that acute bacterial conjunctivitis is a self-limited disease, but topical antibiotic therapy with polymyxin­bacitracin shortens the duration of the clinical disease and enhances eradication of the causative organism from the conjunctiva.2

In our study clinical and bacteriologic responses to a new ophthalmic solution formulation of trimetho­prim and polymyxin were compared with the re­sponses to solutions containing either gentamicin sul­fate or sodium sulfacetamide. All three preparations are known to have adequate in uitro activity against H. in{luenzae and S. pneumoniae. 4

Each of the three ophthalmic formulations was as­sociated with high rates of clinical improvement or cure after 3 to 6 days oftherapy (92% for TP, 95% for G8 and 89% for 88). These clinical responses to topical antimicrobial therapy are consistent with those in the controlled study by Gigliotti et al. 2 The overall bacteriologic responses with the three medi­cations were not statistically different and were also consistent with those obtained in the study by Gig­liotti et al.2 However, in our study eradication of H. in{luenzae at the final follow-up appeared better for TP (81 %) than for G8 (57%) (P = 0.03) or 88 (57%) (P = 0.06). The clinical significance of these findings was not apparent.

All test agents showed similar safety profiles with only one adverse experience reported in a patient who received TP therapy. It is questionable that the mod­erate lid edema present in this patient was related to therapy, since the condition was present before TP was started.

Ophthalmic formulations for treating acute con­junctivitis are selected by considering clinical efficacy, bacteriologic cure rates, adverse experiences, cost and ease of administration. The ophthalmic solution con­taining trimethoprim and polymyxin B hastens clini­cal cure, provides early eradication of the responsible bacterial organism and is associated with limited ad­verse experiences. It is a reasonable addition to the list of effective topical ophthalmic antimicrobials.

REFERENCES

1. Gigliotti F, Williams WT, Hayden FG, et al: Etiology of acute conjunctivitis in children. J Pediatr 98:531-536, 1981

2. Gigliotti F, Hendley JO, Morgan J, et al: Efficacy of topical antibiotic therapy in acute conjunctivitis in children. J Pediatr 104:623-626, 1984

3. Brook I: Anaerobic and aerobic bacterial flora of acute conjunc­tivitis in children. Arch Ophthalmol 98:833-835, 1980

4. Ashley KC: The anti-bacterial activity of topical anti infective eye preparations. M ed Lab Sci 43:157-162, 1986

5. Jarudi N, Golden B, Hoyme J, et al: Comparison of antibiotic therapy in presumptive bacterial conjunctivitis. Am J Ophthal­mol 79:790-794, 1975

6. Godfrey K: Statistics in practice: Comparing the means of several groups. N Engl J Med 333:1450-1456,1985