ARTICLE |
a Department of Pediatrics, Sophia Children's Hospital
c Department of Medical Microbiology and Infectious Diseases, Erasmus MC University Medical Center, Rotterdam, Netherlands
b Department of Pediatrics, Division of Infectious Diseases, State University of New York Downstate Medical Center, Brooklyn, New York
d Department of Ophthalmology, Rotterdam Eye Hospital, Rotterdam, Netherlands
| ABSTRACT |
|---|
|
|
|---|
METHODS. At the Sophia Children's Hospital and Rotterdam Eye Hospital, 2 cohorts of infants <3 months of age presenting with conjunctivitis were studied, 1 retrospectively (July 1996 to July 2001) and 1 prospectively (September 2001 to September 2002). Laboratory diagnosis was based on bacterial culture and polymerase chain reaction for C trachomatis.
RESULTS. C trachomatis was detected in 27 (64%) of 42 retrospectively studied infants and 14 (61%) of 23 prospectively studied infants. Mucopurulent discharge was present in 35 (95%) of 37, swelling of the eyes in 27 (73%) of 37, conjunctival erythema in 24 (65%) of 37, respiratory symptoms in 14 (38%) of 37, and feeding problems in 5 (14%) of 37 infants respectively. Before microbiological diagnosis, general practitioners prescribed antichlamydial antibiotics locally to 5 (12%) of 41 and systemically to 4 (10%) of 41 infants who tested positive for chlamydia, and ophthalmologists prescribed to 21 (51%) of 41 and 7 (17%) of 41, respectively.
CONCLUSIONS. C trachomatis was the major cause of bacterial conjunctivitis in this population. Clinically, differentiation from other pathogens was not possible. Many infants who tested positive for chlamydia did not receive appropriate antibiotic treatment.
Key Words: Chlamydia trachomatis neonatal conjunctivitis
Abbreviations: REH—Rotterdam Eye Hospital SCH—Sophia Children's Hospital GP—general practitioner PCR—polymerase chain reaction
The occurrence of Chlamydia trachomatis infection in infants is directly related to the prevalence of maternal urogenital infections and vertical transmission rates.1–4 The overall risk for infants born to women with untreated chlamydial infection is
50% to 75%, with infection occurring at
1 anatomic site. Conjunctivitis may occur in 20% to 50% of infected infants, nasopharyngitis in
70%, and pneumonia in 5% to 20%.5 C trachomatis has become the most frequent identifiable cause of neonatal conjunctivitis in many countries.6–9 The majority of chlamydial conjunctivitis cases heal spontaneously during the first few months of life. However, untreated persistent infections can lead to acute discomfort and distress for both infant and mother, as well as to chronic eye disease. Simultaneous silent infection of the respiratory tract may cause acute or chronic respiratory disease.3,10,11
Screening of pregnant women for C trachomatis infection was recommended by the Centers for Disease Control and Prevention more than a decade ago.12,13 In 2004 the Dutch National Health Council advised against routine chlamydial screening of Dutch pregnant women, because local data with respect to chlamydial infection in pregnant women and the contribution of C trachomatis to neonatal disease (pregnancy outcome, conjunctivitis, and respiratory tract infection) provided insufficient evidence to support screening.14,15 Recently we reported a prevalence of C trachomatis infection in pregnant women of 6.4%.16 Most of these infections had been missed in routine care, suggesting that C trachomatis transmission to infants may often remain unnoticed.
The main objective of this study was to establish whether C trachomatis was a cause of neonatal conjunctivitis in infants referred to hospital-based care in Rotterdam, the second largest city in the Netherlands. In addition, we evaluated the clinical presentation of and prescribed treatment for chlamydial conjunctivitis compared with other infections.
| MATERIALS AND METHODS |
|---|
|
|
|---|
Microbiological Diagnosis
Eye swabs were taken for routine bacterial culture, including gonococcal culture, and chlamydial and gonococcal polymerase chain reaction (PCR) (Cobas Amplicor, Roche Molecular Diagnostics, Pleasanton, CA). Specimens were obtained by swabbing the conjunctiva of the everted lower eyelid using a sterile Dacron swab. Bacteriologic cultures were processed according to standard procedures for aerobic bacteria in a clinical microbiology laboratory. Chlamydial and gonococcal PCRs were performed according to the manufacturer's instructions.
Treatment
Antibiotics that were prescribed and considered to be effective against C trachomatis were erythromycin, azithromycin, clarithromycin, tetracycline and doxycycline, and chloramphenicol.
Data Collection
We collected retrospective data through a systematic review of medical charts. We used a standardized questionnaire and laboratory investigations to collect prospective data. Recorded variables included the following: age at presentation, gender, complaints (conjunctival erythema, swelling of the eyelids, mucopurulent discharge, unilateral/bilateral involvement, and respiratory or feeding problems), physical examination, diagnostic tests, diagnosis, and therapy by GPs and ophthalmologists before microbiological diagnosis.
Statistical Analysis
We used SPSS 10.0.0 (SPSS Inc, Chicago, IL) for our analyses. We used uncorrected
2 tests to compare categorical variables. We calculated risk ratios (with 95% confidence intervals) to examine factors associated with a diagnosis of chlamydia conjunctivitis.
| RESULTS |
|---|
|
|
|---|
Diagnostic Tests
Five of 64 retrospectively studied infants were clinically diagnosed with bacterial conjunctivitis without performing a laboratory test. The remaining 59 infants had a bacterial culture done, including gonococcal culture; 42 infants also had a chlamydial PCR, and 17 infants had a gonococcal PCR. With inclusion of all of the tests, 50 (85%) of 59 tested infants had a positive result. All 23 of the prospectively studied infants had bacterial and gonococcal cultures done, as well as a chlamydial and gonococcal PCR. Nineteen infants (83%) had a pathogen detected. Table 1 shows the frequency of isolated species. Detection of C trachomatis was significantly higher than of other pathogens (P < .001), with similar rates in retrospectively and prospectively studied infants: 27 (64%) of 42 and 14 (61%) of 23, respectively. Three infants in the retrospective cohort with Haemophilus influenzae had a second pathogen diagnosed.
|
|
|
Follow-up
At follow-up, no more conjunctivitis was observed, and no symptoms or signs of respiratory tract infection were present.
| DISCUSSION |
|---|
|
|
|---|
The main limitation of this study was that we could only include infants with conjunctivitis who were referred to specialist hospitals. Other limitations were the incomplete diagnostic workup in the retrospective cohort and missing data on clinical symptoms in a few infants. Also, the number of children in analyses was relatively small, too small to draw firm conclusions. Although we cannot show what proportion of all neonatal conjunctivitis is caused by C trachomatis, we have shown that it is an important cause of (persistent) conjunctivitis. Determining the incidence of chlamydial neonatal conjunctivitis and transmission rates would require a prospective study testing all pregnant women and their newborns.
The identification rate of pathogenic bacteria in our study was 84%, which is higher than in most reports.17–20 In addition, C trachomatis was the organism isolated most often. This is likely to be related to our study design and to the Dutch referral system, in which mainly infants with persistent conjunctivitis are referred to ophthalmologists. However, we may even have underestimated the contribution of chlamydia to persistent neonatal conjunctivitis, because we excluded the infants with a clinical diagnosis of viral conjunctivitis. Underdiagnosis of C trachomatis persistent conjunctivitis may also be suggested by the relatively low observed number of 5 cases per year in the retrospective study versus 14 in the prospective year. However, the latter increase may also reflect an actual increase of chlamydial infection in Dutch adults.21
To put our results into context, we related the findings from our prospective study to obstetric data from the SCH in that year. There were 1648 deliveries in total, of which 731 were in women <30 years of age. With an antenatal prevalence of 6.4%,16
47 women would be expected to have C trachomatis infection. With an estimated transmission rate of 50% to 75% and 20% to 50% of these developing conjunctivitis,5 we would expect 9 to 24 infants with chlamydial conjunctivitis, which corresponds with our findings. This may suggest that all of the infants with conjunctivitis were detected. However, not all of the studied infants were born to women delivering at the SCH, but also in other hospitals in Rotterdam or at home.
The clinical presentation of chlamydial conjunctivitis has been extensively described by others, and our study shows similar results to previous reports.3,4,9,22–24 Most infants who tested positive for chlamydia presented between the ages of 1 and 6 weeks (because of the slow reproductivity of the organism) and with all 3 of the symptoms (erythema, swelling, and discharge).
Ophthalmologists more often prescribed antibiotics that were effective against chlamydial infection than GPs. Still, only half of the infants with chlamydial conjunctivitis received effective antibiotics before microbiological diagnosis, and <20% received systemic treatment. In the prospective study, even fewer infants received systemic treatment. This probably reflects the agreement made with ophthalmologists to refer infants to a pediatrician after diagnosis. We cannot draw any conclusions about chlamydial conjunctivitis in primary health care from our study. Acute conjunctivitis may be correctly treated by GPs with topical antibiotics according to Health Care Insurance Board guidelines. However, chlamydia is probably not being considered in the differential diagnosis by some GPs, because only 12% of infants with chlamydia had received antibiotics that were active against C trachomatis.
Our results may suggest the need to institute eye prophylaxis for ophthalmia neonatorum, which is not practiced in the Netherlands. However, we do not want to advocate the application of neonatal eye prophylaxis, because prophylaxis does not prevent all chlamydial neonatal conjunctivitis, and the absence of conjunctivitis as an indicator of chlamydial infection may delay the proper diagnosis of (silent) chlamydial infection at other sites. Another motivation against the decision to start routine eye prophylaxis is that this may lead to overtreatment of newborns in a country with easy access to medical care. We would rather like to use our findings to indicate the need for proper (systemic) antibiotic treatment of newborns who test positive for chlamydia and screening of pregnant women.
Chlamydial screening for pregnant women is not standard practice in the study area or in the rest of the Netherlands. Testing is only done when clinically warranted. In the prospective year of the study, only 1% of 1648 deliveries were tested for C trachomatis, of which 4 tested positive. These figures may also reflect the underestimation of chlamydial infection in pregnant women. The value of screening may warrant additional discussion.16 To prevent 1 case of chlamydial conjunctivitis, 31 to 78 pregnant women need to be screened. Assuming a specificity of
99.8%,16 this could result in 0.06 to 0.16 false-positive mothers and 2 to 5 truly positive mothers (and their partners); furthermore, this could result in the prevention of 1.4 to 2.5 nasopharyngitis cases and 0.2 to 1.0 pneumonia cases. Previously we described the costs of chlamydia screening of pregnant women by different DNA-isolation methods in individual and pooled urine samples.16 The cost per chlamydia case detected when using the Cobas Amplicor on individual urine samples was
275 and when using a combined method of isolation with the MagNA Pure bacterial DNA-isolation kit (Roche Molecular Diagnostics, Pleasanton, CA), and subsequent amplification and detection by Cobas Amplicor on pooled urine samples, as we did, was
108.
In the absence of a chlamydial screening program, we urge clinicians to have a higher index of suspicion for neonatal and maternal chlamydial infection. Furthermore, we recommend that, in areas where prenatal chlamydial screening and treatment of pregnant women is not routine practice, infants with signs of conjunctivitis that persist for >72 hours while applying frequent normal saline eye irrigation should have a full microbiological evaluation including Neisseria gonorrhoeae and C trachomatis. While awaiting laboratory results, empiric treatment in this population should include erythromycin eye drops or ointment to both eyes. We recommend that confirmed chlamydial conjunctivitis should be treated systemically with erythromycin suspension or alternatively azithromycin.13,25
Current national treatment guidelines for neonatal chlamydial conjunctivitis are from the Health Care Insurance Board and recommend tetracycline eye drops (with or without oral erythromycin or tetracycline).26 The current Centers for Disease Control and Prevention recommendation, however, is to treat all neonatal chlamydial conjunctivitis systemically with a 14-day course of erythromycin suspension.13 Systemic treatment has been demonstrated to be more effective than topical treatment, and infants with conjunctivitis are often infected at other sites as well.27
| CONCLUSIONS |
|---|
|
|
|---|
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
Address correspondence to Ingrid G.I.J.G. Rours, MD, MMed, Department of Pediatrics, Erasmus Medical Centre, Sophia Children's Hospital, Dr Molewaterplein 60, 3015 GJ Rotterdam, Netherlands. E-mail: rours123{at}planet.nl
Dr Rours is the principle investigator and first author and was responsible for statistics; Dr Hammerschlag is a coauthor and adviser and was responsible for revision; Dr de Groot is a co-author and was responsible for revision and promotion; Dr de Faber is a co-author and was responsible for revision and logistics at the Rotterdam Eye Hospital; Dr Verbrugh is a co-author and was responsible for revision and promotion; Dr Verkooyen is a co-author and was responsible for revision and co-promotion; and Dr Ott is a co-author and was responsible for revision and statistics.
The authors have indicated they have no financial relationships relevant to this article to disclose.
| REFERENCES |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||