You Can’t Graft Your Way Out of a Prosthetic Problem

A patient points at a grey shadow above an anterior implant crown and asks what happened. The reflex is to book a connective tissue graft. That reflex is where most of these cases are lost — not because grafting is wrong, but because a coronally advanced flap has nothing to advance over when the crown itself is the obstacle. The restoration that helped create the deficiency is still sitting in the way, and no amount of surgical technique gets around it.

The literature is blunt about this. Coverage of peri-implant soft tissue dehiscences with a flap and graft alone has historically landed between 40% and 66% — and in Burkhardt's prospective cohort, where coverage decayed from 75% at one month to 66% at six, complete coverage was achieved in none of the sites (Burkhardt 2008; Anderson 2014). The same defects treated with a staged prosthetic-surgical protocol reach 96.3% mean coverage at one year and 99.2% at five, with complete coverage in 75% and 79% of sites respectively (Zucchelli 2013; Zucchelli 2018). The technique is not what changed. The order of operations did. This issue walks the staged sequence — diagnose, reshape, graft, re-restore — through a single case, and closes with the part nobody enjoys hearing: most of these are preventable at placement.

Disease first, and this is a hard gate. Every classification and decision tree in this space applies only to a clinically healthy implant. Following the 2017 World Workshop, an implant with bleeding on gentle probing, probing depth ≥ 6 mm, and bone levels ≥ 3 mm apical to the coronal portion of the intraosseous surface is diseased, not deficient (Berglundh 2018). Treating a peri-implantitis lesion with a bilaminar flap buries the problem under a graft. Probe, radiograph, and be honest about what you find before anything else happens.

Then classify by two things: where the crown sits, and how tall the papillae are. Zucchelli and co-workers built the reference system around the buccolingual position of the implant-supported crown, and then subdivided by papilla height (J Periodontol 2019). Class I is a margin in the right place with the grey showing through — a thickness problem, not a position problem. Class II is an apically positioned margin where the crown profile still sits inside the imaginary curve connecting the adjacent teeth. Classes III and IV are apically positioned margins where the crown profile sits outside that curve; the two are separated only by where the implant head itself lands once the crown is off, palatal (III) or facial (IV) to the line joining the adjacent gingival margins.

The subclass is what actually dictates the sequence. Subclass a means both papilla tips are ≥ 3 mm coronal to the ideal margin position; b means at least one is between 1 and 3 mm; c means at least one is under 1 mm. That single measurement decides whether you can advance a flap at all. Subclass a can go straight to CAF + CTG. Subclass b is the indication for the full prosthetic-surgical approach — you have to grow the papillae before you can borrow from them. Subclass c has no papillae to borrow from, so the site gets treated as edentulous: crown and abutment out, cover screw in, augment under submerged healing.

Alrmali's tree is the version to keep on the wall. Working from Michigan, Alrmali and colleagues distilled the same territory into four destinations — purely prosthetic, purely surgical, combined prosthetic-surgical (with or without abutment removal), and explantation — routed by three evaluations: three-dimensional implant position relative to the alveolar contour, hard tissue status, and soft tissue status, with keratinized mucosa width dichotomized at 2 mm (J Esthet Restor Dent 2023). It is the more practical instrument at the chair, because it forces the position question first. And it names the outcome we avoid saying out loud: when the implant head is frankly facial and the papillae are gone — Zucchelli's Class IVc — the honest answer is removal, not coverage.

Take the crown off and shorten the abutment. In the case shown here — a maxillary left central with an apically positioned margin, visible metal at the collar, and a papilla that would not support advancement — the definitive crown came off, the abutment was reduced in height and hollowed, and a shortened screw-retained provisional went on with a deliberately flat-to-concave subgingival profile. The provisional is intentionally shorter than the final crown. You are not restoring esthetics in this phase; you are vacating space so tissue can move into it.

Then wait, because the papillae are growing. Zucchelli's protocol allows roughly two months between the provisional and the surgery, and the interproximal tissue responds by increasing in both height and thickness. This is the entire point. The coronally advanced flap in Stage 2 is fed and anchored by those papillae — they get de-epithelialized toward the palate and become the bed the flap sutures to. A flap advanced over shallow papillae has no blood supply and no anchorage, which is precisely why the crown-in-situ approach underperforms in these classes.

Design the flap to be advanced, not just elevated. A trapezoidal envelope with split-full-split thickness elevation gives you the mobility to bring the margin above the abutment convexity and hold it there without tension. The anatomic papillae are de-epithelialized toward the palatal so the advanced flap has vascularized connective tissue to sit on rather than epithelium.

Use a de-epithelialized free gingival graft, not a subepithelial harvest. Zucchelli's group makes this distinction deliberately: a graft de-epithelialized from a free gingival graft is dense collagenous connective tissue with minimal fatty and glandular content, which makes it firmer, more stable, and better at holding volume than a subepithelial harvest (J Clin Periodontol 2010). Around implants, where the peri-implant connective tissue is scar-like — fibers parallel to the surface, fewer fibroblasts, reduced vascularity (Sculean 2017) — that quality difference matters more than it does on a root.

Fix the graft at the level of the contralateral tooth's margin. Not at the current margin, and not where the crown used to end. The reference is the homologous natural tooth, because that is the definition of a successful outcome in this classification. Thickness is the other target: Zucchelli's group puts the working threshold at roughly 2 mm of mucosal thickness before the grey of the abutment stops reading through the tissue (van Brakel 2011).

Close it passively, over the abutment. Primary closure above the abutment convexity is what protects the graft through the first weeks. If the flap needs tension to reach, the prosthetic phase was too short or the abutment was not reduced enough — go back rather than pulling.

Condition first, restore late. After maturation, a new temporary crown conditions the augmented tissue, and the definitive restoration follows roughly 8 to 12 months after surgery. It feels slow. The reason to tolerate it is that the tissue is still moving. Buccal soft tissue thickness increased 1.54 ± 0.21 mm at one year, then gained a further 0.3 mm between years one and five (p < .001), with keratinized tissue height adding another 0.5 mm over the same window. Mean coverage climbed from 96.3% to 99.2% through creeping attachment (Eur J Oral Implantol 2018). A definitive crown delivered too early gets contoured against tissue that has not finished arriving.

Grafted sites hold. Across 15 prospective studies with at least three years of follow-up, CTG-based augmentation at implant sites produced stable or increasing keratinized mucosa width and mucosal thickness over time, stable marginal bone, and peri-implantitis incidence between 0% and 7.1% — while non-augmented implants tended to show continued apical migration of the margin (Stefanini 2023). The graft is not only cosmetic.

The numbers are worse than the field's reputation suggests. The AO/AAP systematic review of 221 studies puts mean PSTD prevalence at 46.2%, with incidence up to 38.3% within five years of loading, in the absence of disease (Tavelli & Barootchi 2025). Restricted to single implants in the esthetic zone, prevalence measures 54.2% at the patient level (Tavelli 2022). This is not a rare complication we occasionally see; it is the default trajectory of an unprotected implant.

The risk factors are almost entirely ours. Limited mucosal thickness, immediate placement, and the absence of soft tissue augmentation at placement drive PSTD; guided surgery, bone grafting at placement, soft tissue augmentation, and adequate keratinized mucosa and thickness protect against it (Tavelli & Barootchi 2025). Add reduced keratinized mucosa width, longer time in function, and increased buccal bone distance from the ultrasonographic data (Tavelli 2022). Every item on the protective list is a decision made on the day of surgery.

Build the restorative half in from the start. The PROS framework — platform switching, restorative abutment design, an optimal internal conical connection, and subcrestal placement — organizes the prosthetic variables that keep peri-implant tissues stable rather than leaving them to be repaired later (Misch, Alrmali, Galindo-Fernandez, Saleh & Wang, Int J Oral Implantol 2025). Abutment height is the underrated member of that set: above 2 mm of transmucosal height, emergence angle largely stops behaving as an independent risk factor, and every additional millimetre of height lowers the odds of any marginal bone loss (Misch et al., Clin Oral Implants Res 2025).