Why I’m So Optimistic About Cartilage Healing, Take 2

I got a comment from a reader below this post. If you remember, I was looking at a two-year study that showed a surprising number of improvements in cartilage defects (well, I found the number surprising anyway). This reader was less enthusiastic:

If I am interpreting the study correctly, most of the defects of the patella actually progressed. For younger people, perhaps for the majority of this site readers, this is bad news. Perhaps, I am wrong, but it seems to me that most of us here suffer from chondromalacia, of one degree or another, and as it is about patella cartilage damage, there isn't much joy in that study. There's another one from 2008, where the level of degradation of patella lesions was high compared to all the investigated knee compartments and the percentage of cases where improvement of a lesion was observed was abysmal.

First, let me address a couple of quick things: (1) Yes, I’ve seen a study too where defects in the patella cartilage didn’t improve as often as defects in cartilage elsewhere in the knee – but still, there were some instances where they did improve. (2) To clarify, the study I reference in the post isn’t looking at just defects in the cartilage behind the patella, but rather, throughout the knee.

Okay, a quick recap:

The table below is from a study, “Factors Affecting Progression of Knee Cartilage Defects in Normal Subjects Over Two Years.” The 86 people who participated had MRIs done of their knees at the start of the study, then two years later.

The condition of each subject’s cartilage was graded for five different knee compartments (at baseline, and after two years). The scoring again goes like this:
Grade 0 = normal
Grade 1 = focal blistering
Grade 2 = irregular surface and loss of thickness of less than 50%
Grade 3 = deep ulceration with loss of thickness of more than 50%
Grade 4 = full-thickness wear of cartilage with bone exposed

Okay, so did most of the defects in the study progress? Well, yeah. Of course. But it doesn’t matter because you have to adjust for the “floor” and “ceiling” effect.

In this case, “floor” means a defect can’t get worse. “Ceiling” means it can’t get better.

Example: If a defect is graded “0” at baseline, two years later, it can only be “0” or worse. It can’t get any better than 0. There is no -1! Conversely, a defect graded “4” at baseline can only stay the same or (if indeed cartilage can heal) can get better.

Now, look at the number of defects that have a “ceiling” effect (grade of 0) or a “near ceiling” effect” (i.e., defects initially graded 1).

There are 117 that start out with a “ceiling” effect (just add the numbers in the first row) and 196 with a “near ceiling” effect (that's the second row). So for this group of 509 defects, there’s a high chance they’re going to get worse. Sure enough, we find a whopping 389 got worse. Awful, right?

Not at all.

Look at the bottom of the table. Here, our attention turns to the “floor” effect (defects with an initial grade of 4) and “near floor” effect (initial grade of 3). There are 5 defects with a “floor” effect and 14 with a “near floor” effect.

How many defects improved? Only 10 – which seems like a small number compared with 389, but consider that we started with only 19 (yes, I know, “small sample size” alert).

Of course, when you look at the number of changes overall, many more defects got worse because most faced the ceiling effect. But look at raw percentages, and the story becomes more interesting:

Defects that started in the “ceiling” or “near ceiling” effect categories
Got worse: 76%
Stayed the same or got better: 24%

Defects that started in the “floor” or “near floor” effect categories
Got worse: 26%
Stayed the same or got better: 74%

Wow! Almost a perfect inversion!

Now, why does this matter (once again) if you really care about cartilage healing (which, again, you shouldn’t obsess about in the first place, because pristine cartilage isn’t a sine qua non for eliminating knee pain).

Because if cartilage really can’t heal, all those defects graded 3 or 4 should be staying the same or getting worse; three-quarters of them shouldn’t improve!

One last fun thing in closing: is there a line in the table (you’ve probably already spotted it) where we can escape the “floor” and “ceiling” effects as defined here? Sure: at baseline, 88 defects landed smack dab in the middle of the table, receiving a score of 2. That means a loss of thickness of less than 50 percent.

Now notice what happened to them two years later. Yes, eight were found to be worse. But more than three times as many, or 27, improved to a grade of 1.

So that’s why I see the glass as half full (even though, for the umpteenth time, don’t obsess over cartilage healing!).