Polished / Machined Collars: When Bone Doesn’t Read Textbooks

Polished / Machined Collars: When Bone Doesn’t Read Textbooks

There was a time when we were certain.

Certain that bone does not attach to smooth surfaces.

Certain that polished collars were the enemy of stability.

Certain enough to transform assumptions into protocols.

Looking back, I don’t think we were wrong.

I think we were honest with what we knew at that time.

I remember very clearly a discussion many years ago — a cold Tuesday, a very cold Tuesday — in downtown Manhattan, at the Marion marquee, during an Astra Tech meeting. Michael Norton was giving a lecture, and all the international students of Tarnow — and even Tarnow himself — were there. One of those moments that stays with you, not because of the conclusion, but because of the tension.

One side defended that bone simply cannot attach to smooth surfaces. The other argued that it will, under the right circumstances. It wasn’t an argument of egos; it was an argument of biology interpreted through different lenses.

According to Hermann and Buser’s studies, the tendency was indeed that bone could not adhere to the polished neck from Straumann. Impossible — the biological width would get you, and bone loss would be “your destiny.” On the other side, Michael would argue that a hollow cage into bone, made of these polished collars, would definitely integrate. Sigmund Stall, Abrahamsson, Albrektsson — they were all quoted and thrown into the heated discussion.

We, the students, looked like tiny rabbits 🐇 watching all this.

Take a look at some histology provided by Straumann:

1 – collar of the titanium grade 4 tissue-level classical implant (the one used in the Hermann and Buser articles)

2 – collar of the tissue-level TLX tulip-shaped (titanium–zirconium)

3 – collar of the TLX NT (titanium–zirconium)

4 – collar of the machined collar of the Brånemark original screw

5 – section of the machined part of the body of a Brånemark original screw

Pretty different, right?

And now look at the values given by the SRA components.

Later, language changed. What we once called polished became polished / machined. A subtle but important semantic shift. Almost an admission that the surface was not what we thought it was. Not fully polished. Not fully rough. Somewhere in between. Bone, as usual, didn’t care about the name.

What the evidence showed — and continues to show — is that subcrestal placement of these collars leads to bone loss. Predictable. Reproducible. Uncomfortable. And yet, clinically, something else started to happen. Something that didn’t fully fit the narrative.

I saw it in Catarina’s case.

I saw it again in the father of Princess Jasmin.

And suddenly, the question was no longer what we believe, but what we observe.

The bone seemed to tolerate — even accept — that smooth or machined surface. Not in a romantic way. Not by “attaching” as we like to define attachment. But by behaving. By stabilizing. By not collapsing in the way we had been taught to expect.

Years later, materials changed. Pure titanium gave way to titanium–ceramic alloys. Macrogeometries evolved. The tulip shape was replaced by straighter profiles. The possibility of placing these implants subcrestally became real — not theoretically, but clinically.

And then we hit the contradiction.

A subcrestal implant needs an emergence profile.

But the traditional tissue-level components were never designed for that reality.

From the beginning, many of us said it out loud: this doesn’t make sense. You cannot ask a component designed for supracrestal behavior to behave subcrestally without consequences. At first, the idea was resisted. Slowly, as often happens, biology and clinical reality won. Not by force, but by repetition. By outcomes. By time.

Today, the portfolio is complete. And it’s not because industry suddenly became visionary — it’s because bone kept answering honestly.

What almost no one talks about is that, along the way, titanium grade 4 of the old tissue-level implants was replaced by titanium–zirconium. This may have an impact on how cells — and the immune response — behave around the implant. Here may lie one of the features that could be a game changer in crestal bone performance, together with subcrestal positioning.

So yes — details matter.

And with respect to that, the connection unit is also something that may play a role in what we are seeing clinically. Tissue-level and TLX implants were generally placed supracrestally, so the discussion of the polished collar was more academic than truly clinical for Straumann. But for Nobel Biocare, it was different. The classical external hexagon, due to the cylindrical body, could actually be buried into bone, allowing bone to interact with a polished surface.

With the NT platform, both in TLX and TLC implants, we now have the same model of implant placement but with a totally different surface and biomaterial. Connection therefore also becomes an issue. Who can argue that the classic 2 mm of bone loss was not due to a platform-matched, primitive hexagon connection, rather than to a lack of bone attachment to the polished collar?

Take a look at the different connections:

1 – tissue-level internal connection

2 – TLX internal connection

3 – TLX NT internal connection

4 – Brånemark original connection

And now a summary of the whole complex and bone placement:

1 – tissue-level internal connection with supracrestal placement

2 – TLX supracrestal placement with TorcFit connection

3 – TLX NT subcrestal internal conical connection with platform-switch components

Having said this, I believe there is no scientific evidence in the current literature that fully explains bone performance around a tissue-level NT titanium–zirconium implant with a tight conical connection placed subcrestally.

What we have are assumptions.

The first assumption: tissue-level implants with machined / polished collars should be used because, if bone loss occurs, the part exposed to the oral cavity has fewer chances of being colonized by bacteria, and therefore — biological plausibility… very dangerous territory — fewer chances of being lost due to peri-implantitis.

A second assumption could be formulated: subcrestal implants made of titanium–zirconium alloy will lose around 2.5 mm of bone due to biological width formation, but because of a more biocompatible surface, there will be a reduced immune response — less interleukin 1 and 6 — and therefore less marginal bone remodeling. Also, less plaque accumulation, and therefore slower progression of peri-implant disease compared to pure titanium implants (a feature we already know from ceramic implants).

A third premise could be that, due to the tight connection, subcrestal placement of NT implants, together with platform switching, leads to more favorable bone behavior compared to implants with unstable microgaps.

So which of these details matters the most?

Or, in other words, are we witnessing a combination of all these factors?

Can we actually build the statement like this: bone adheres to a polished collar — not because it is polished, but because immune response is lowered due to the zirconium element of the Roxolid alloy, and because subcrestal placement of NT implants, together with a tight conical connection and platform switching, allows minimal bacterial leakage?

By lowering immune response and bacterial load, the whole complex becomes better protected from the oral environment, potentially leading to a decrease in peri-implantitis incidence.

This is not a post about being right.

It’s a post about being attentive.

Bone doesn’t read textbooks.

Bone reacts to surfaces, forces, blood supply, and time.

And when our concepts don’t align with that, bone doesn’t negotiate.

Detail matters here. Surface finish matters. Placement depth matters. Indication matters. And most of all, humility matters.

Because progress in implant dentistry is rarely a revolution.

More often, it’s the quiet moment when we admit that biology was ahead of us all along.