Crown or Onlay: Preserving More of the Tooth

In direct comparisons, survival did not differ significantly between crowns and onlays, but long-term trials are lacking.

A 2022 meta-analysis (Wang and colleagues) pooled observational studies of full crowns against tooth-coloured onlays or partial crowns on back teeth, finding no statistically significant survival difference at 1 or 3 years. Its one randomised trial, not pooled, ran 1 year with no failures in 38 onlays or 40 crowns.

Among pressed lithium disilicate restorations on back teeth in one private practice, without random allocation, bulk fractures or large chips needing replacement affected 16 of 1,782 full crowns and 6 of 610 partial-coverage restorations over up to 16.9 years. Malament and colleagues (2021) concluded there was no statistically significant survival difference.

Vagropoulou and colleagues (2018), reviewing 9 studies, said no meaningful comparison between restoration types was possible, put 5-year survival above 90%, and reported a relatively high failure rate due to decay and ceramic fractures.

Other reviews cover one treatment each, so their figures cannot be compared. Morimoto and colleagues (2016) estimated 92% to 95% survival at 5 years and 91% at 10 for ceramic inlays, onlays and overlays, with failures from fractures or chipping (4%), nerve complications (3%), decay (1%) and debonding (1%). A 2026 review of 64 studies on front and back teeth put 5-year crown survival between 90.4% (feldspathic or silica-based ceramic) and 98.5% (monolithic lithium disilicate).

Over 15 years, 55 of 228 extensive bonded pressed ceramic coverages, partial and complete, failed, 39% on non-vital teeth and 20.9% on vital (van Dijken and Hasselrot, 2010). A one-dentist series of 74 monolithic lithium disilicate crowns in 12 patients, with median follow-up of 12.8 years, had 13 failures (van den Breemer and colleagues, 2017).

Composite onlays have thinner evidence. Bustamante-Hernández and colleagues (2020), pooling ceramic, hybrid and composite onlays, estimated 94.2% survival over 2 to 15 years, lower over time and with composite, urging caution over study quality. A single composite onlay trial, direct or indirect, reported 87% overall at 5 years (Angeletaki and colleagues, 2016).

A crown is commonly recommended when a lot of tooth is missing, but the guides cited here set no single cut-off. Bristol Dental School's guide gives these reasons: a lot of tooth tissue missing from decay or fracture, a large filling needing further protection, root canal treatment calling for a strong restoration, or replacing a failing crown. Onlays, it adds, can be used where a larger amount of natural tooth is left.

What about a root-filled tooth?

The British Endodontic Society's guide says to consider each case individually. On a root-filled back tooth, it says, cusp coverage, such as an onlay or crown, will be required in most cases once one side wall is missing, and should be used if two or more are.

Which restoration suits is less clear: a 2022 review (Mannocci and colleagues) says defined guidance is lacking. A 9-year trial of lithium disilicate partial crowns, on root-filled teeth with at least 50% of the tooth above the root left and patients reporting no bruxism (tooth grinding), found tooth survival of 85.7% to 88.8%. Another 2022 review reports partial restoration failure on root-filled teeth rising from 4.32% at 2 to 4 years to about 20.94% at 12 to 30 years. Neither study compared them with full crowns. Timing is covered under crowns after root canal treatment.

Does grinding change the choice?

Grinding is a concern for both, and the evidence is mixed. A 2025 scoping review, largely retrospective and over half on implants, calls bruxism a risk factor for failed direct and indirect restorations, indirect monolithic zirconia excepted. A 2026 trial of 109 patients, 93 in its 3-year groups, detected no effect of sleep bruxism on monolithic lithium disilicate or zirconia molar crowns over 3 years, with complication-free success of 81.5% and 85.7% with bruxism, 95.2% and 95.8% without.