Improving polymerase chain reaction diagnostic rates for herpes simplex keratitis: results of a pilot study

Abstract

Background

Laboratory confirmation is crucial for diagnosis and management of herpes simplex virus (HSV) keratitis. However, the sensitivity of polymerase chain reaction (PCR) in keratitis is low (25%) compared with that of mucocutaneous disease (75%). We developed an educational intervention aimed at improving the diagnostic yield of PCR.

Methods

The medical records of keratitis cases seen at the emergency department of a London tertiary ophthalmic referral hospital over two distinct periods, before and after an educational program on swab technique, were reviewed retrospectively.

Results

A total of 252 HSV cases were included. Increases in the laboratory-confirmed diagnosis of HSV-1 were observed, in both first presentations (11.1%–57.7%) and recurrent cases (20%–57.6%). The rate of positive HSV-1 PCR in eyes with an epithelial defect increased from 19% pre-intervention to 62% post intervention. Notably, 3% were positive for varicella zoster virus DNA, and there was a single case of Acanthamoeba keratitis.

Conclusion

Our results suggest that, with proper swabbing technique, PCR may be more sensitive than previously reported.

Introduction

Herpes simplex virus (HSV) keratitis is the primary cause of corneal blindness in high-income countries and a leading cause of blindness worldwide.1 The estimated annual global incidence of HSV keratitis is 1.7 million, and 10% of affected eyes have a vision worse than 6/30 (20/100) after 10 years.2,3 There are limited data on UK incidence, however.

Establishing early identification and laboratory confirmation of HSV keratitis is crucial to optimizing management and outcome. Reliance on clinical features alone may lead to diagnostic uncertainty or error. Clinical misdiagnosis of HSV keratitis occurs frequently, particularly when terminal bulbs and epithelial infiltrates are absent; similar lesions occur with both varicella zoster virus (VZV) and Acanthamoeba keratitis.4 It has been reported that 4%–16% of clinically diagnosed HSV keratitis is VZV when polymerase chain reaction (PCR) testing is performed (Lies Remeijer, personal communication).5

Although viral culture remains the gold standard investigation, nucleic acid amplification testing (NAAT), typically PCR, is more commonly used for laboratory diagnosis. PCR is widely available, and positive results provide reassuring evidence that the clinical diagnosis is correct, with implications for treatment and prognosis. However, the sensitivity of PCR for corneal samples in HSV keratitis is often low, reported at around 25%, compared with mucocutaneous disease at 75%.6,7 The purpose of the current pilot study was to establish the proportion of PCR-positive samples among clinically diagnosed HSV keratitis cases before and after an educational intervention, including proper swabbing technique.

Subjects and Methods

This study was approved by the Clinical Audit Committee of Moorfields Eye Hospital and was compliant with the tenets of the Declaration of Helsinki. The medical records of HSV keratitis cases attending the Emergency Department at Moorfields Eye Hospital, a tertiary ophthalmic referral hospital, over two distinct three-month periods—from September 1, 2019 to November 30, 2019, and October 1, 2022 to December 31, 2022—were reviewed, retrospectively. An educational program was delivered between the two study periods. There was no change in the laboratory testing techniques between the study time points.

Consecutive patients with a coded discharge diagnosis of herpes simplex keratitis were identified via electronic patient records. All patients presenting to the department with a clinical diagnosis of HSV keratitis were included. A clinical diagnosis was made on the basis of presence of symptoms and signs consistent with HSV infection, namely, an injected, painful, photophobic eye, often with blurred vision, and a dendritic epithelial defect or coarse punctate staining and/or stromal edema, particularly if disciform in shape. Rapid NAAT using PCR was performed on initial presentation for HSV-1, HSV-2, and VZV DNA. Samples were processed by an external laboratory (Micropathology Ltd, Coventry, UK). On follow-up review by a corneal specialist, if there was high clinical suspicion of HSV, the diagnosis would remain even in the presence of a negative swab. Where an alternative diagnosis was made, cases were excluded.

The educational interventions were aimed at all clinicians in the department (ophthalmologists, optometrists, and nurses) and included lectures on HSV keratitis that highlighted the importance of taking a swab sample for all presentations and, reinforcing the need to debride the epithelium when taking a swab. A video demonstrating correct swab technique was distributed and made available online (see Video 1). These educational interventions were implemented over a 1-month period in August 2022 and reinforced by senior clinicians after this time.

Video 1.

Download video file ^{(42.8MB, MOV)}

Video demonstrating correct swabbing technique. LINK TO VIDEO]

As a retrospective, nonrandomized pilot study, formal statistical analysis was neither planned nor performed.

Results

Following the educational intervention, the number of swabs performed for suspected HSV keratitis increased from 22.5% (29/129) to 48% (59/123). The rate of HSV-1 DNA–positive results increased for both new diagnoses, 11.1% (1/9) to 57.7% (15/26) and recurrent diagnoses, 20% (3/20) to 57.6% (19/33). See Table 1. In eyes with an epithelial defect present when the sample was taken, the rates increased from 19% (4/21) pre-intervention to 62% (45/75) post-intervention. Notably, 2% (1/51) and 3% (2/75) of swabs were positive for VZV DNA in the pre-intervention and post-intervention groups, respectively. In addition, one patient in the post-intervention group, clinically diagnosed as HSV keratitis was diagnosed with Acanthamoeba keratitis by in vitro confocal microscopy (IVCM) during follow-up. In the pre- and post-intervention groups, 8 and 10 cases, respectively, were excluded, because of miscoding, missing notes, or alternative diagnoses being made on follow-up review in a specialist corneal clinic (eg, recurrent corneal erosion syndrome).

Table 1.

PCR swab results before and after educational intervention

Discussion

In our study cohort, education on correct sampling techniques improved HSV keratitis diagnosis on PCR from 19% to 62% of clinically suspected cases. To our knowledge, this is the first report on the effects of an educational program on the diagnosis of viral keratitis by swab PCR. Previous attempts to improve diagnosis have focused on the PCR of tears, corneal scrapes, or corneal impression membranes.8–10 Our study shows a notable improvement using a simple and efficient intervention, easily replicated at minimal cost, using readily available materials. Based on the results of this pilot study, we propose that ophthalmology centers provide clinician education on proper swab technique and take swab samples in all cases of suspected viral keratitis to allow for the assessment of diagnostic impact in their patient population.

It is paramount to attain PCR confirmation of diagnosis at first presentation, or recurrent presentation where not previously confirmed, so that in the event of complicated or recurrent disease, including stromal or endothelial disease, where PCR diagnosis is more challenging, that the full gamut of medical and surgical management for HSV keratitis can be offered with confidence.

Our results also confirm that VZV keratitis (primary or recurrent) can be clinically misdiagnosed as HSV. The significance for patients with VZV keratitis is that misdiagnosis may lead to poorer outcomes because of inadequate antiviral therapy, because VZV treatment requires double the dose of oral antiviral compared with HSV. In cases of epithelial or stromal HSV keratitis, topical treatment is sufficient, but in epithelial or stromal VZV keratitis, systemic treatment is required; thus, misdiagnosis will always lead to significant antiviral underdosing. The identification of an Acanthamoeba keratitis case that had been misdiagnosed as HSV keratitis reinforces the conclusion of a prior report highlighting the risk of this misattribution.4 It is therefore best practice to confirm all cases of suspected viral keratitis by PCR rather than to rely on a clinical diagnosis alone. Additionally, any contact lens wearers should receive PCR for HSV, VZV, and Acanthamoeba, as well as IVCM where diagnosis is uncertain.

Although the observed improvements in PCR diagnosis were dramatic, there are important limitations to this study. First, it was a small, retrospective pilot study at a single center, with no randomization applied, rendering statistical analysis invalid. Reproduction of these findings in a larger prospective study across multiple sites is advised. Second, behavioral change in the post-intervention arm is likely to have impacted the pre-test likelihood of HSV keratitis in swabbed individuals; in the post-intervention arm, more individuals with clinical features highly consistent with HSV may have been swabbed in addition to borderline or uncertain cases, which may have predominated in the pre-intervention arm. This may have affected the observed sensitivity of PCR diagnosis in the post-intervention arm. It is not possible to assert the relative contribution of this behavior change to the observed increase in PCR-sensitivity compared with the contribution of improved swabbing technique following the educational program. This may benefit from further study. Finally, it is not clear how long the effects of this educational program will persist and whether or how frequently it will need to be repeated; ongoing assessments are planned in this regard. Despite these limitations, the observed sensitivity of PCR diagnosis in the post-intervention arm is far higher than has been previously reported in ophthalmic swab samples and approaches the sensitivity seen in mucocutaneous disease, suggesting that wider use of PCR confirmation in suspected viral keratitis may be justified.

In conclusion, the observed improvements in HSV PCR diagnosis in our pilot study may be attributed to the introduction of an educational program. The results suggest that a simple, low-cost intervention may make PCR diagnosis of HSV keratitis more sensitive than previously reported. Education on correct corneal swab techniques, with an emphasis on debridement being beneficial therapeutically and diagnostically, may result in improved diagnosis for both first and recurrent cases of HSV and VZV keratitis.

Literature Search

PubMed was searched performed on December 29, 2023, without date or language restrictions. Search terms and combinations included the following: herpes simplex virus, HSV, diagnosis, PCR, polymerase chain reaction, and swab technique.