The insignificance of bond-strength

by Nelson Gendusa, DDS
Director - Research

What I’m going to say next is so blasphemous, I may have to wear a disguise when I attend next year’s IADR Research Meeting.

The more I compare published research with clinical experiences, the more convinced I become that bond strength data are pretty much humbug. (That is “humbug” as in “without much clinical significance.”)

I’m not referring to outright fraud, where advertisers or lecturers with some axe to grind jigger the data. Heaven knows there’s enough of that, but I’m talking about ALL bond tests - even those performed by the most conscientious, pure-in-heart researcher.

You see, bonding studies are based on the assumption that there’s a correlation between tensile bond or shear bond strength and clinical success. In fact, there isn’t. Or at least I’ve never found any. The few researchers who have tried to predict performance based on bond results have failed miserably. Offhand, I can think of at least 4 studies that concluded that there is no relationship between bond tests and clinical results.^{1, 2, 3, 4} And I don’t know of one solitary study that suggested there is a relationship.

Funny thing, the people who pore over bond-strength data when selecting a bonding agent generally consider themselves “evidence-based” dentists. Yet, nobody ever asks for any evidence that bond numbers have anything to do with clinical success. I guess that’s because it’s just so darn “obvious” that they do.

Let’s look at how a typical shear test is performed. (You tell ME how close it is to your technique when you place a real restoration.)

1) The extracted teeth used to determine bond-strength are primarily human 3rd molars or bovine anteriors.

There’s considerable research showing that bond strength varies significantly according to the specific tooth you’re bonding to.* Third molars may not be representative.

In real practice relatively few restorations are placed in wisdom teeth ... fewer still in cow teeth.

2) The teeth used in bonding research are often carefully selected to be caries-free. Even in this day of quantum cosmetics, I suspect most bonded restorations are still placed precisely because the teeth were carious. Research is now showing that the sound dentin below caries is significantly harder to bond to than pristine dentin in teeth that are “caries free”.

3) The surface of the tooth is removed to expose a flat dentin surface. In most laboratories the occlusal portion is cut off, which means the tubules tend to run perpendicular to the bonded surface. And the C-factor is lowest it can be.

In real life no one bonds to a flat surface. The direction of the tubules cannot be controlled. And the C-factor is, well a factor.

4) The surface is ground using a 600 grit polisher. Let’s see a show of hands. How many of you polish your preps prior to bonding?

5) The actual bonding procedure is performed on a horizontal dentin surface with the sample sitting on a lab bench. One study found that bond strength to dentin dropped by 50% when the surface was vertical during bonding rather than horizontal!⁵ In bond tests there’s generally no humidity like the oral cavity. No saliva or blood to contaminate the surface.

6) The “dentist” who bonds the material may be a student or perhaps a technician who has never used the bonding agent before. Many researchers never have an opportunity to climb the technique “learning curve” for any of the products they handle. (Here’s a neat trick we manufacturers use when we submit products for “independent” comparative research. We try to find a dentist who uses the product in private practice. That way there’s a good chance our product will be used properly - and the other products won’t.)

7) The samples are thermocycled from 5°C to 55°C. Some researchers t-cycle as few as 100 times, others t-cycle as many as 50,000 times. (I just read an article by some hyperactive researchers who cycled their sample 1.2 million times!) Thermocycling isn’t remotely related to anything that has ever happened in anybody’s mouth since the beginning of time.

8) A purely directional force (either tensile or shear) is applied to the composite until it is dislodged. In real life, dislodging forces are complex and in many cases, unpredictable.

9) The researcher can use a number of different test techniques. These different techniques yield wildly different bond strengths. Even worse, a bonding agent may seem to perform much better using one test than another. This means that a rank-order of adhesives according to their bond strength might change if the researcher simply used a different test technique.⁶

I’ve occasionally queried researchers about the significance of bond strength. I’ve pointed out that glass ionomer restorations hang in there year after year even though in the lab they provide shear bond strengths of just 3 or 4 MPa. With the exception of Class V’s, today’s restorations don’t fall out no matter what they are bonded with.

When I pose “Gendusa’s Glass Ionomer Conundrum” to researchers, they almost invariably respond that high bond strength isn’t necessary to retain the restoration. It’s necessary to preserve tight margins and prevent gapping. One researcher had even calculated the minimum shear bond strength needed to overcome the stress caused by composite shrinkage. (It was between 17 and 18MPa ... and he had a page and a half of calculations to prove it.)

Well, here are some studies I’d like all you “evidence-based bonditos” to think about before you start throwing around your next bond-strength number.

A joint team from Japan and Australia studied how the quality of composite margins and the amount of gapping is influenced by different factors ...⁷
• the speed of cure (fast vs slow-start)
• the design of the cavity prep (low c-factor vs high c-factor)
• the bonding agent used (Amalgambond^® vs Clearfil LinerBond^® vs ClearFil PhotoBond^®)

I wouldn’t be doing my job if I didn’t mention that Amalgambond beat the other bonding agents every way they could be beaten. When restorations were bonded with Amalgambond they showed much less marginal leakage and much better adaptation to the cavity walls. And this was true whether the composite was fast-cured ... or cured with a pulse-delay light. It was true whether the composite was placed in shallow preps ... or deeper preps that put a lot of stress on the adhesive.

Nothing surprising there ... except for this.

In laboratory bond tests Amalgambond consistently shows lower bond strength than the other two agents.

Let me say that again. Amalgambond (lower bond strength) prevented margins from pulling away from the tooth and prevented the curing composite from pulling away from the cavity walls much better than PhotoBond (higher lab bond strength) or LinerBond 2 (much higher lab bond strength.)

Other research conducted in Europe found precisely the same thing with Touch&Bond. Despite the fact that its bond-strength numbers were about middle-of-the-pack, Touch&Bond^® showed the highest percentage of complete margins and the lowest incident of gapping of all the commercial adhesives tested.^{8, 9}

THE POINT: If you’re choosing your bonding agent based on bond strength data, you’re operating under a delusion. You might as well be choosing it based on the color of the package. It’s an understandable delusion, I’ll grant you that. And it’s a delusion that’s shared by many others. But it’s a delusion nevertheless.

The fact is, successful bonding is a multidimensional endeavor that can’t be reduced to a single number. It’s related to the modulus of elasticity of the agent ... toughness of the film ... the molecular density of hybrid layer ... the depth of monomer penetration ... the degree of cure ... and a host of other characteristics that are not measured in a “bond test.”