Incidence and clinical characteristics of infantile conjunctivitis in a Western population

Conjunctivitis is a common ocular condition among infants, generally characterized by injection of the conjunctival vessels with or without discharge. The causes are diverse and include chemical, allergic, bacterial, or viral and may be acute or chronic.¹ Without a timely diagnosis and subsequent treatment, conjunctivitis can result in ocular complications and disseminated infection.^2,3 Although there are numerous studies concerning the bacterial pathogens in neonatal conjunctivitis, ^2–15 there are few reports on the incidence and clinical features of conjunctivitis in the first year of life. The purpose of this study is to describe the incidence and clinical characteristics of infantile conjunctivitis diagnosed over a 10-year period in Olmsted County, Minnesota.

Methods

Study population

The medical records of all patients ≤ 12 months of age diagnosed with any type of conjunctivitis while residing in Olmsted County, Minnesota, from January 1, 2005, through December 31, 2014, were retrospectively reviewed. Olmsted County is located 90 miles (145 km) southeast of Minneapolis. Approximately 70 percent of the county’s residents (124,277) lived within the Rochester city limits in 2000. The racial demographic of Olmsted Country residents was 82.0% Caucasian, 3.7% Asian, 3.2% Black or African-American, 0.3% American Indian or Alaska Native, 2.6% for other, and 7.6% for unknow. Hispanic or Latinos of any race accounted for 2.4% of the population.

Case Ascertainment and Diagnostic Criteria

Mayo Clinic and Olmsted Medical Center institutional review boards granted approval for this study. Patient records were identified using the Rochester Epidemiology Project (REP), a medical records-linkage system constructed to collect every interaction of patient medical care in Olmstead County, Minnesota.¹⁶ Olmsted County is relatively isolated from surrounding urban areas; as a result, Mayo Clinic or Olmsted Medical Group and affiliated hospitals supply the plurality of medical care provided to county residents.¹⁷ A central electronic patient record connects all medical care information received by the individual including demographics, medical diagnoses, and surgeries.

Cases of conjunctivitis were ascertained using the International Classification of Diseases, Ninth Revision (ICD-9) diagnostic codes for conjunctivitis and red eye. Only patients living in Olmsted County at the time of diagnosis, diagnosed within the study period from 2005 through 2014, and ≤ 12 months of age at the time of diagnosis were included. Each patient chart was individually reviewed for demographics, diagnoses, and relevant clinical characteristics. Diagnoses were confirmed based on the information contained within the medical record of each patient. The medical records of children diagnosed with conjunctivitis within 30 days of birth underwent a more detailed review including analysis of ocular surface cultures and therapeutic interventions. Among a total of 2674 individually-reviewed records, 394 (14.7%) were excluded for an incorrect diagnosis, 94 (3.5%) for denying research authorization, and 11 (0.4%) for not residing within Olmsted Country. The data was organized using Research Electronic Data Capture (REDCap), a digital software designed to collect research data.

Incidence rates were estimated using the age- and sex-specific population figures from the 2005 through 2014 Olmsted County, Minnesota census data. Population figures for each year between the census years was calculated using a linear interpolation of the data. The 95% confidence intervals (95% CIs) for overall incidence were calculated assuming Poisson error distribution Statistical analysis was performed using SAS Version 9.2 (SAS Institute; Cary, North Carolina).

Results

A total of 2175 infants were diagnosed with conjunctivitis during the 10-year period, yielding an incidence of 10422 per 100,000 children or 1 in 10 infants in the first year of life. The historical and clinical characteristics of the 2175 children are shown in Table 1. The mean age at diagnosis was 4.9 months (range, 1 day to 12 months) and 1001 (46.0%) were female. The diagnosis of conjunctivitis was made by a primary care provider in 1786 (82.1%) and occurred in both eyes in 1180 (54.3%). The incidence of conjunctivitis within 30 days of life in this cohort was 2759 per 100,000 or 1 in 36 cases (2.7%), with the mean age of diagnosis at 19 days. Less than half (42.2%) were females and bilateral involvement occurred in 222 (38.5%). Concurrent neonatal ocular disorders included dacryostenosis in 160 (27.8%)

TABLE 1.

Historical and Clinical Characteristics of 2175 Infants Diagnosed With Conjunctivitis in Olmsted County. Minnesota, from 2005 to 2014

Characteristic	Diagnosed ≤30 d(n = 576)	Diagnosed ≤ 12 mo(n=2175)
Female	243 (42)	1001 (46)
Premature birtha (<37 wk of gestational age)	47 (8)	52(2)
Age at diagnosis, mean (range)	19.5	4.9
	(0–3 d)	(0–12 mo)
Diagnosis made by
Primary careb	479(83)	1786 (82)
Emergency physician	22(5)	160(7)
Ophthalmologist	10(2)	30(1)
Otherc	65(11)	199(9)
Laterality
Right eye	168 (29)	506 (23)
Left eye	186 (32)	489(22)
Bilateral	222 (39)	1180(54)
Other ocular disorder
Nasolacrimal duct obstruction	160 (278)
Preseptal cellulitis	1 (0.2)
Subconjunctival hemorrhage	2 (0.3)

Unless otherwise noted. values are n(%).

^a<37 weeks⁺ gestational age.

^bPrimary care includes family practice and pediatrician.

^cPhysical medicine and rehabilitation, neonatal medicine, internal medicine, allergy and asthma, acute care, sleep medicine, physician assistant, dermatologist, hospice and palliative medicine, or nurse practitioner.

Ocular surface cultures were obtained in 111 (19.3%) of the 576 diagnosed in the first 30 days. The microbiology cultures were positive in 35 (31.5%) with 4 (3.6%) isolating more than one organism as shown in Table 2. A total of 14 (12.6%) were Gram positive, 11 (9.9%) were Gram negative, 3 (2.7%) showed chlamydia, 0 (0.0%) showed gonococcus, and 9 (8.1%) were suspected to be bacterial but did not show a specific organism. The management of conjunctivitis for children ≤ 30 days of age, where recorded, included topical antibiotics in 523 (90.8%) and simple observation in 47 (8.2%). Of the topical antibiotics, Polymyxin B/Trimethoprim or Erythromycin antibiotics were prescribed in 421 (73.1%). When recorded, 98.6% percent of resolving symptoms were observed within 14 days. None of the infants were known to have permanently decreased vision as a result of their conjunctivitis.

TABLE 2.

Etiology of Cultures and Treatment From 576 Patients Diagnosed With Conjunctivitis ≤30 Days of Life

Characteristic	Number (%)
No culture	465 (80.7)
Negative culture	45 (40.5)
Positive ocular culturesa	36 (32.4)
Bacterial, unspecified	9(8.1)
Haemophilus	5 (4.5)
Streptococcus pyogenes	4 (3.6)
Coagulase-negativo staphylococci	4 (3.6)
Alpha-hemolytic streptococci	4 (3.6)
Staphylococcus aureus	3 (2.7)
Pseudomonas aeruginosa	3(2.7)
Chlamydia	3 (2.7)
Streptococcus pneumoniae	2 (1.8)
Rothia mucilaginosa	1 (0.9)
Serratia marcescens	1 (0.9)
Enterobacter	1 (0.9)
Haemophilus parainfluenzae	1 (0.9)
Gonococcal	0(0)
Unknown or undocumented	30 (27.0)
Chemical	1 (0.9)
Treatment
Observation	47 (8.16)
Topical antibiotics	523 (90.80)

^aCultures may have grown with more than 1 causative agent.

Discussion

Conjunctivitis, or a red, often purulent eye significant enough to present to a medical care provider, occurred in 1 in 10 children less than 1 year of age in this population-based cohort diagnosed over a 10-year period. There was a modest male predominance and at least half of the patients had bilateral involvement. Concurrent nasolacrimal duct obstruction (CNLDO) occurred in one fourth of those occurring in the first 30 days of life and sight-threatening infections were rare.

Neonatal conjunctivitis, defined in this study as any conjunctivitis diagnosed by 30 days of life, occurred in 1 in 36 infants. Prior studies, most of which are not population-based and vary by age at diagnosis, report incidence rates ranging from 0.3% - 17.3%.^{4,7,9,10,12,13} Two of four population-based studies on infantile conjunctivitis only report on chlamydia and gonococcal infections.^8,11 Dannevig et al observed a higher incidence of 18.9% for 1928 infants in Troms County, Norway.¹⁴ Some of the affected children, however, were diagnosed by the caregiver’s history and not by a health care provider. In a prospective, population-based study on infantile conjunctivitis within ten Finnish child welfare clinics, Honkila et al reported an incidence of 1.8%,¹⁵ similar to the findings in the current study. The diagnosing clinician in 4 of 5 infants in this study was a primary care provider which is consistent with prior reports in which approximately 70% of conjunctivitis patients were managed by primary care offices and urgent care centers. ^1,10,16

Although various causes for neonatal conjunctivitis have different ages at onset and pattern of symptoms, the current study did not find sufficient association to identify causative organisms. Honkila et al did not detect meaningful differences in symptoms across bacterial and viral causes of neonatal conjunctivitis.¹⁵ Because the majority of cases were identified and treated by primary care providers, eye exam documentation was limited, and topical cultures were uncommonly performed in this series. Concurrent ocular disorders diagnosed at the time of conjunctivitis included nasolacrimal duct obstruction in one fourth and rare other conditions.

Ocular surface cultures were performed in only 1 in 5 cases of those diagnosed by 30 days of life, from which one third of the cultures identified a causative agent. Consistent with other studies of infantile conjunctivitis, various Gram-positive cocci are the most common etiologies for conjunctivitis in the first 30 days of life. ^5–9,14,15 Prior studies report limitations on culturing neonatal conjunctivitis as cultures often fail to identify an isolate, show non-pathologic growth or normal mixed microbial flora (range 5 – 53%).^4,9–11,14 Pierce et al identified a bacterial cause in only one third of cases when presenting to general practitioners. ¹¹ Chlamydia and gonococcal conjunctivitis were rarely detected in this cohort, similar to the findings from a large population-based study by Honkila et al. and a retrospective 15-year study by Pak et al.^5,15 The prevalence of chlamydia and gonorrhoeae depends on the population studied, standards of hygiene, and use of prophylactic perinatal topical antibiotics.^2,3,9,10,15 The standard practice of prophylactic erythromycin in the birth centers at the study institution likely contributed to the low rate of gonorrhoeae conjunctivitis in this series.

There are several limitations to the findings of this study. Its retrospective design is limited by non-standardized data and irregular follow-up. The incidence rate for infantile conjunctivitis in this cohort may have underestimated the actual value since children with conjunctivitis may have sought care outside Olmsted County or, for those with milder symptoms, simply avoided evaluation altogether. Most infants in this cohort were diagnosed by primary care providers instead of ophthalmologists, which likely resulted in less detailed documentation and a potentially suspect diagnosis. Symptoms compatible with both conjunctivitis and CNLDO were sometimes described without subsequent delineation between the two diagnoses. However, distinguishing between the two was facilitated in this study by retrospectively reviewing each medical record. Finally, due to the relatively uniform population of Olmsted County, the findings from this cohort may not translate to populations outside of Olmsted County.

In this population-based study, a diagnosis of conjunctivitis occurred in 1 in 10 infants in the first year of life with one fourth diagnosed by 30 days of life. Half of the patients were bilaterally affected, nasolacrimal duct obstruction occurred concurrently in one fourth, and infections threatening visual acuity were rare.