Antivirals started after the neonatal window still improved hearing in congenital CMV

Design

systematic review and single-arm meta-analysis of pooled proportions

Population

8 studies, 192 patients with congenital CMV treated with ganciclovir or valganciclovir after the first month of life

Primary outcome

hearing outcomes and viral load after postneonatal antiviral treatment

Effect

urine viral load -3.67 log at 2 months (95% CI -4.33 to -3.02); hearing normalised in 42.3% (13.5-77.5) and improved in 38.5% (16.8-66.0) of affected ears; deterioration 5.5%; grade 3-4 neutropenia 3.7%

Treatment for congenital cytomegalovirus is licensed and studied in the first month of life, which leaves the large group diagnosed later - after a failed newborn hearing screen chased up slowly, or after hearing loss is noticed in infancy - without an evidence-based option. This review pooled eight studies and 192 patients treated with ganciclovir or valganciclovir started after the neonatal period.

Viral load fell: pooled mean difference at two months was -3.67 log in urine (95% CI -4.33 to -3.02) and -0.80 in plasma (-1.08 to -0.52). Among ears with hearing loss, hearing normalised in 42.3% (95% CI 13.5-77.5) and improved in 38.5% (16.8-66.0), while 5.5% deteriorated (2.2-13.3). Neutropenia was uncommon - 9.1% grade 1-2 and 3.7% grade 3-4.

The caveats are large enough to state before the numbers are used. These are single-arm pooled proportions with no control group, so the natural history of congenital CMV hearing loss - which fluctuates and can improve without treatment - is not subtracted. Heterogeneity was extreme (I² above 84% for both hearing outcomes), and the confidence interval for normalisation runs from 13% to 78%, which is barely an estimate. What this does justify is not writing off a child diagnosed at three or six months. In India, where newborn hearing screening coverage is patchy and confirmatory testing often delayed, that group is not small.

• Do not assume the treatment window has closed because a child with congenital CMV is past one month
• Confirm the diagnosis properly before treating - after three weeks of age, CMV detection no longer distinguishes congenital from postnatal infection without stored newborn samples or dried blood spots
• Monitor the neutrophil count during treatment; grade 3-4 neutropenia occurred in about 1 in 27
• Set expectations honestly - the proportion who improve is genuinely uncertain and some ears deteriorate anyway
• Keep audiological follow-up going regardless of whether antivirals are given

Why it matters

It reopens a treatment question for the children who are diagnosed late, which in much of the world is most of them.

Don't overread it

No control arm and extreme heterogeneity - improvement after treatment cannot be separated from the fluctuating natural history of the condition.

The statistics, in plain English

These are single-arm pooled proportions: everybody was treated, so there is nothing to compare against, and congenital CMV hearing loss is known to fluctuate on its own. I² values above 84% mean the studies disagreed with each other almost completely, and the confidence interval of 13.5% to 77.5% for hearing normalisation reflects that - the true figure could be one ear in seven or three in four. The viral load reduction is the most solid finding here, and viral load is a surrogate, not hearing.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.