Dental Student, Harvard
Hello, everyone. My name is Yu-Chi Cheng, and I am a second year dental student at Harvard. Today we’re going to be talking about evidence based dentistry and what we can learn by looking at implant data over a long period of time. All this research began with a very simple discussion at a lunch table about bisphosphonates. We had a visiting surgeon who expressed concerns over the use of Bicon implants in patients taking bisphosphonates and other antiresorptive medications. But Dr. Morgan said, well, Bicon has been doing it for the past 25 years with seemingly no issue. So he turned to me and said, Yu-Chi, can you look at the data and see if that is true? So we looked at the data and what we found was that approximately five percent of all of the implants that were being placed at the implant dentistry center in post menopausal female patients are in patients who are taking antiresorptive drugs. That’s three forty four implants in the past twenty years. On top of that, there is an additional eight percent of patients in this demographic with untreated osteoporosis who should be taking these antiresorptive drugs. And we saw a striking result. We saw that both the bisphosphonates cohorts and the general population control group had better implant survival rates than patients who had osteoporosis or osteopenia, but were not taking bisphosphonates. So what does this mean? This means that bisphosphonates actually have a protective effect when it comes to Bicon dental implants. When we see a patient who has osteoporosis, we see that as being associated with a poor implant survival outcome, but that effect can be rescued back to the baseline by taking an oral or IV antiresorptive medication. So why is that? If we think about how bisphosphonates work, they help treat osteoporosis by inhibiting the activity of osteoclasts. We know that bone remodeling and bone turnover is mediated by a balance of osteoblasts and osteoclasts, where osteoblasts deposit bone and osteoclasts degrade bone. We also know that when you screw an implant into alveolar bone, that results in a displacement of the bone in order to accommodate the threads. The pressure caused by implant threads biting into the bone activates osteoclasts, which eventually remodels the bone into appositional bone. That’s why threaded implants have the concept of primary stability, which is the initial stability of the implant that rapidly declines because of this bone remodeling and osteoclastic activity. And since you have osteoclastic activity as a core part of the implant healing process, That process is susceptible to disruption by bisphosphonates. In comparison, a Bicon implant heals by a fundamentally different principle. Instead of osteoclastic activity, you have this fibrin clot inside the osteotomy where the Bicon implant is placed. And this fibrin clot attracts osteoblasts from the surrounding bone, which then migrate into the fibrin clot and form newly deposited Haversian bone. This osteoblastic activity is not affected by bisphosphonates. And that is why Bicon implants we think have been doing so well in patients who are under antiresorptive therapy. So in summary, we see that Bicon implants are not harmed by oral bisphosphonates and other antiresorptive drugs. In fact, they might even benefit from them. And this of course contradicts the current narrative. In twenty twenty two, we published these results in the clinical oral investigations journal, and it has gotten some attention in more recent review articles. In fact, this has become one of the most frequently cited papers published by our group. So we talked about bisphosphonates, which supposedly enhances bone metabolism. And as we found out, it also enhances the survival of implants. Now let’s talk about something that does the opposite. This is a molecule that as every medical student learns cause calcium to leach out of bones. Yet, when you ask a dentist, what do you think about it? Everyone says, oh, it enhances bone metabolism. I’m talking about vitamin D. This picture by the way, is a photo of a vitamin D rapid test that we use at the implant dentistry center in Boston. It requires just a tiny drop of blood and it returns a result within fifteen minutes. I cannot tell you how many times I have stabbed my finger just to calibrate this device, but at least now I can really say that I have given my blood, sweat and tears for dentistry. Here’s a figure from a well known review article in the New England Journal of Medicine. To summarize this very complicated chart, first, the major role of 1,25-dihydroxyvitamin D, also known as calcitriol, which we will subsequently refer to as vitamin D is to maintain blood levels of calcium and phosphorus in the normal range via intestinal absorption so that normal bone mineralization can occur. Second, and more important for our discussion, when calcium levels are low, vitamin D cooperates with parathyroid hormone to increase bone resorption and release stored calcium from the skeleton. With that in mind, let’s look at the data. We compared the vitamin D levels of patients with failing versus healthy implants. And we saw that patients with blood vitamin D levels greater than seventy nanograms per milliliter, which we define as hypervitaminosis D, these patients had a twenty one point one fold increase in the risk of implant failure or severe peri implant bone loss. We also see that patients with low vitamin D levels or hypovitaminosis D is also correlated with a three point nine fold increase in risk, which while significant is not as striking as the very high correlation shown with hypervitaminosis D. Compared to the green line on the left figure, which is where patients whose vitamin D level are within normal limits, and the implant survived at a ninety five percent rate over twenty years, both the hypo and hyper vitamin D, vitamin D had worse implant survival outcomes over the same period of time at seventy three point six percent and seventy three point seven percent respectively. Interestingly, this effect of reduced implant survival is only present when implant survival is correlated with the absolute level of vitamin D. It is not correlated with the absence or presence of vitamin D supplementation. So if a patient is taking vitamin D supplements, but just maintaining a normal level of vitamin D, we should not be too worried about them. I also want to mention the bone loss pattern in patients with hypervitaminosis D. We see very significant early implant bone loss of around two millimeters within the first year after implant placement. And we see a continued maintenance of that compromised marginal bone level over the lifetime of the implant. You can read more about this research in the article that we published in the Journal of Dentistry last year. So having talked so much about bone loss, let’s talk about how we can make bone gain happen. And this is a study of one hundred and seven implants placed in the molar regions of ninety five healthy patients. Here we can see one illustrative example of maintaining bone level around the top of the implant shoulder, where bone gain is involved two years after loading. In this graph, we see three different modalities of different factors can affect marginal bone levels. On the left, we see that initial marginal bone level really has no effect on long term bone levels. Regardless of where the implant bone level starts at, the implant has a way of reaching equilibrium where on average the bone level will gravitate towards the top of the implant over time. So if you start out with a low bone level, as long as the implant is placed properly, there is nothing to worry about. The bone gain will happen and the bone level will tend towards the top of the implant. However, this does not mean that the implant can be placed anywhere willy nilly. In the middle, we see that if the implant is placed too shallow relative to the crestal bone, then it becomes very difficult for bone levels to reach the top of the implant. Last but not least on the right, we see a really interesting phenomenon where implants placed in patients who are taking NSAIDs, specifically the eighty one milligram baby aspirin that many patients take for cardiovascular reasons, see this mitigation of the early implant bone loss that is so often seen around implants. This is so far the only factor we have identified that can successfully prevent the early implant bone loss from happening within the first year of implant placement. And we can see that the extra bone that is preserved within that first year in patients taking NSAIDs is then maintained for many, many years. So maybe there is a benefit for patients to be on NSAIDs, specifically aspirin around the time of implant placement, which we should look into in the future. And these results can also be found in our publication shown here. Today, we’re gonna talk about the results of the statistical analyses in our multicenter fibula graft study. We’ll talk a little bit about the primary outcomes of this study, implant survival, implant success and prosthesis outcomes. Then we’ll talk about the use of this exciting new analysis called manifold learning, which is a type of machine learning. And we’re applying this technique to analyze implant outcomes. And lastly, we’ll talk about the effects of prosthesis construction on bone gain. This study included forty five patients who received one hundred and seventy implants in total followed up over the course of fourteen years. We included patients who received implants in free fibula flaps due to a variety of reasons, including tumor, trauma, or extreme atrophy of the jaw. When we look at the primary outcomes of this study, we see high implant survival and success rates across fourteen years of follow-up. The implant survival rate was 89.7 % and the implant success rate was 82.7 %. A number that is on par with what has been previously reported for other types of implants in the literature, which is roughly an annual failure rate of 0.02. We see in our latest analysis that implant survival in the maxilla and mandible are similar, but the really striking part of this data is when it comes to prosthesis survival and success. When we look at other studies that have reported prosthesis outcome, we see much lower numbers like forty two point nine percent in the study of fifty six patients and even twenty percent in a case series, supposedly due to the difficulty of restoring implants placed in these fibula grafts. But here we see very high numbers at fourteen years the rate of prosthesis survival is ninety one point seven percent and the rate of prosthesis success is eighty point three percent. And these high numbers really highlight the excellent restorative capability of the implant system. Cox regression is not my favorite analysis, but reviewers love it. So it’s worth going through the results. I’m going to single out osteoporosis and past use of anti resorptive medications here, which somehow is correlated with improved implant outcomes. And I’ll discuss why that is the case in my separate lecture. We also see that tooth loss from trauma is correlated with a better outcome compared to tooth loss due to tumors. And we see some other covariates that are also unfavorable for implant survival, including hypertension, advanced age, lateral bone augmentation, nothing surprising. What I’m really excited about is this manifold learning analysis. It’s a type of machine learning. And what it does is it takes those complex high dimensional data sets, like this three-dimensional representation of a mammoth skeleton here. And we use this technique called UMAP to collapse it into a two dimensional representation here on the right. What’s nice about this technique is that the relative architecture of the dataset is preserved. So the tusks are still attached to the head and the head is still next to the torso, etcetera. And we can do the same for really complex high dimension datasets, like an implant database that has fifty, sixty or so columns of data. And we can collapse the data into a two dimensional representation like the one we see on the right, where each implant is represented by a dot and the dots that represent similar implants are grouped together. And when we color these dots by, let’s say implant success versus failure versus complication, we start to see a pattern. We see that all the implants with failures and complications concentrate in this one region. So now our task is simple. We know that these implants that are susceptible to failure share a set of similar characteristics. That’s why they’re grouped together. Our research question becomes, what are the common factors that underlie these susceptible implants? So what we see is that those implants have characteristically short follow-up durations, which is consistent with survival analysis results. We see that most failed implants failed within the first three years of surgery. We also see that this cohort of susceptible implants fits into a homogenous patient profile that includes the patient’s age being over forty five at the age of surgery, at the time of surgery, the patient presenting with tooth loss due to a tumor, usually receiving a fibula graft and implants in the mandible, and then receiving radiation after implant surgery. And then after all this receiving almost always a partial fixed dental prosthesis. Notably, is no correlation with patient gender and there were no patients of osteoporosis or with any history of anti resorptive treatment who developed complications. So to recap, we see a rather homogenous population of susceptible patients who share a set of similar characteristics. A patient over forty five who presented for surgery due to a tumor, received a fibula flap in the mandible, and then received radiation after implant surgery, and finally received a partial fixed dental prosthesis after all of this. So moving forward, we would want to study this group of patients and see what exactly is going wrong here. And if we figure that out, we could preemptively identify future patients who might fall into this bucket and give them supportive care so that we can prevent complications and failures from happening. Lastly, let’s talk about the effect of prosthesis construction on bone gain. We measured the bone levels around Bicon implants and abutments. And in a previous study about TRINIA on three prostheses that we followed up for over ten years, we saw a bizarre effect. We saw that if we had a longer distal extension, like the very long one you see in the radiograph here, we actually see a better outcome. We see bone gain. And we showed that this effect was significant using statistics. I won’t go into the details of this graph, but basically what it’s showing is that if you have a longer distal extension on your prosthesis, the implant immediately next to it receives a bone gain benefit over time, as long as the distal extension is longer than fifteen millimeters. While the other implants on the other hand were not affected. In the literature, people like to call these distal extensions cantilevers. Dr. Morgan does not like this term and the reason is because a cantilever is defined as a rigid structural object, but TRINIA being a flexible material is by definition not a cantilever. The flexibility of TRINIA is shown by this bottle opener graphic is really nicely demonstrated by all on three prostheses because the distal extension does not result in a force trying to pop out the anterior abutment. Instead, as you can see in our in vitro study, the occlusal forces applied to the distal extensions are evenly distributed throughout the bone surrounding the implant. And the amount of stress and strain that we saw in those in vitro studies corresponded to the level that is traditionally associated with bone gain in the literature. Now the question is, does the same effect occur around implants in vascular grafts? And the short answer is yes, we do see that implants adjacent to distal extensions experienced bone gain effects once the distal extensions are longer than about twelve millimeters. So overall, let’s talk about why Bicon short implants function so well in this context. We see that they work very well in very complex cases, including full arch cases and vascular graft cases. And as we hinted at with the in vitro analysis, Bicon implants have plateaus which provide a greater amount of surface area. And that surface area along with the short implant design allows the forces from occlusal loading to be transmitted to the surrounding bone evenly. This better distribution of compressive and lateral forces is what we believe is ultimately responsible for the robust and excellent outcomes of Bicon implants, not just in everyday single implant scenarios, but also in challenging scenarios like the ones we mentioned today. Lastly, I want to present the published article containing these results that is now available in the Journal of Cranio-Maxillofacial Surgery after a very lengthy review process. And I want to thank everyone in the multi center study group, especially Prof. Rolf Ewers for making this possible. Thank you very much.