Design: systematic review with single-arm meta-analysis of proportions
Population: 192 children across eight studies treated with ganciclovir or valganciclovir for congenital cytomegalovirus after the first month of life
Primary outcome: hearing outcomes, with viral load and neutropenia
Effect: hearing normalised in 42.29% of affected ears (95% CI 13.48 to 77.51) and improved in 38.48%; deterioration 5.47%; urine viral load mean difference −3.67 log (95% CI −4.33 to −3.02)

Treatment of congenital cytomegalovirus is licensed and studied as a neonatal intervention, which leaves the child identified at four months or two years — the common situation when newborn screening is not universal — without an evidence base. Eight studies covering 192 patients treated with ganciclovir or valganciclovir after the first month of life were pooled in a single-arm meta-analysis.

Viral load fell over the first two months: pooled mean difference −3.67 log for urine (95% CI −4.33 to −3.02, I²=75.4%) and −0.80 for plasma (95% CI −1.08 to −0.52, I²=31.6%). Among ears with hearing loss, hearing normalised in 42.29% (95% CI 13.48 to 77.51, I²=84.3%) and improved in 38.48% (95% CI 16.77 to 66.00, I²=88.2%); deterioration occurred in 5.47% (95% CI 2.15 to 13.25). Neutropenia was uncommon — grade 1 to 2 in 9.09%, grade 3 to 4 in 3.67%.

This matters in India specifically, because universal newborn hearing screening is uneven and congenital cytomegalovirus is often diagnosed late, after a failed screen or a delayed-speech referral. The current answer to those families is that the treatment window has closed. This suggests it may not have, and that is worth a conversation with paediatric infectious diseases rather than a default refusal.

But the design will not carry a treatment decision on its own. There is no control arm, and congenital cytomegalovirus hearing loss fluctuates spontaneously — it improves and deteriorates without any intervention, which is precisely why a single-arm proportion of 42% normalising cannot be attributed to the drug. The very wide intervals say the same thing.

Raise antiviral treatment with paediatric infectious diseases for the late-diagnosed child rather than assuming the window has closed
Document serial audiometry before and during any treatment — fluctuation is the norm and is the main confounder
Monitor the full blood count; neutropenia was uncommon but grade 3 to 4 occurred in about 1 in 27
Confirm the diagnosis is congenital rather than postnatally acquired before treating, which late presentation makes harder
Counsel families that this is observational evidence without a control group

Don't overread it

No control group and very high heterogeneity — hearing in congenital cytomegalovirus fluctuates on its own, so these proportions cannot be attributed to the antiviral.

The statistics, in plain English

Heterogeneity of 84.3% and a confidence interval from 13.48% to 77.51% around a 42.29% normalisation rate means the studies disagree profoundly — the pooled figure is close to uninformative as a prediction for any individual child. A single-arm meta-analysis has no comparator, so none of these proportions is a treatment effect. The most solid number is the fall in urine viral load, whose interval is narrow, but viral load is a surrogate, and the hearing outcome is what matters.