Picture a 76-year-old woman with a painful knee. Every time she bends her knee, she hears a grinding noise behind the kneecap. Her MRI shows why. The cartilage under her kneecap is broken, and bone rubs against bone. She has tried a full year of anti-inflammatory pills and physical therapy. The next option on the table is a knee replacement, and she is not ready for it.

She was one of eight patients in a study I led in Belo Horizonte, Brazil, between 2012 and 2016. We cleaned each damaged joint with arthroscopy. Then we injected cells taken from the patient’s own bone marrow. One year later, the average knee function score had doubled. Her follow-up MRI showed cartilage covering the kneecap again.

Eight patients is a small number, and I will be honest about what that means. But this study answers questions my patients ask me all the time. What exactly goes into the knee? How does the lab prepare it? And what happens in the two years after the injection?

Why the kneecap joint wears out

The patellofemoral joint is the place where the back of your kneecap slides along a groove at the end of the thigh bone. Doctors often pay more attention to the inner and outer parts of the knee. Yet according to a 2011 randomized trial by Hunter and colleagues, arthritis affects this compartment in about 65% of people with knee osteoarthritis.

You feel it in ordinary moments. Stairs hurt. Getting up from a low sofa hurts. Squatting to pick something up from the floor becomes a planned event. Over time, people walk less and give up sport. Slowly, they lose independence.

The root of the problem is cartilage biology. Cartilage is the smooth, white layer that covers the ends of your bones. It has no blood vessels, so repair cells cannot reach it easily. Damage that stays inside the cartilage layer usually never heals. Damage that goes deeper, into the bone below, does heal, but with fibrocartilage. This scar-like tissue is less smooth than the original hyaline cartilage and handles load less well.

When pills and physical therapy fail, surgeons can choose among several techniques:

1. Microfracture: the surgeon makes tiny holes in the bone, so blood and marrow cells can reach the defect.
2. Multiple drilling: the same idea, with a thin drill instead of a pick.
3. Abrasion: the surgeon scrapes the surface to trigger bleeding and a repair response.
4. Mosaicplasty: small plugs of healthy cartilage and bone move from a low-load area to the damaged one.
5. Realignment procedures: operations that change how the kneecap tracks in its groove, for example by rebuilding the medial patellofemoral ligament.

For the kneecap, all of these give limited results. Three reviews published between 2010 and 2013 tried to find the best surgical option for patellofemoral damage. None of them reached a clear answer.

Arthroscopy is the other common choice. The surgeon inserts a small camera and thin instruments through two tiny cuts. Then the surgeon smooths the damaged cartilage, removes loose fragments and washes the joint. The risk is low, and many patients feel better for a while. However, the long-term benefit of arthroscopy in arthritis is doubtful. Many patients still receive it because they are too young, or not yet ill enough, for a knee replacement.

That gap between “the pills don’t work” and “I’m not ready for a new knee” is where my patients stood. The idea for this study came from that gap.

What is inside a bone marrow cell mix

Bone marrow is the soft tissue inside large bones such as the hip. It makes your blood cells every day. When a lab spins a marrow sample through a special liquid, it can separate a layer called bone marrow mononuclear cells, or BMMCs.

BMMCs are a crowd, not a single type of cell. The mix includes:

• Mesenchymal stem cells (MSCs), which can turn into bone, cartilage or fat cells
• Hematopoietic stem cells, the parent cells of every blood cell you have
• Endothelial progenitor cells, which help build new blood vessels
• Monocytes and macrophages, immune cells that clean up damaged tissue
• Platelets, neutrophils and a few fat cells

Here is the part that surprises most people. MSCs, the cells most of us picture when we hear “stem cells”, are rare. They make up between 0.01% and 0.001% of BMMCs. Imagine a stadium with 100,000 people in it. Only somewhere between 1 and 10 of them would be MSCs.

So why use BMMCs at all? The first reason is practical. Labs can grow pure MSCs in culture and multiply them, and in studies these cultured cells work better. But culture takes weeks and needs a certified laboratory that follows good manufacturing practice rules. That makes it expensive and heavily regulated. BMMCs need no culture. We can collect them, separate them and give them back to the same patient.

The second reason is biological. The other cells in the mix seem to matter. A 2014 study in sheep by Song and colleagues found that BMMCs regenerated cartilage in arthritic knees despite their tiny share of MSCs. The effect was weaker than with cultured MSCs, but it was real. The authors suggested that the non-MSC cells play an active part. Blood-forming cells release proteins called cytokines and growth factors. These proteins are chemical messages that help the few MSCs survive and multiply.

Think of a building site. MSCs are the few skilled builders. The other cells are the suppliers who keep materials coming. With a small team of builders and good support, the work still gets done, only more slowly.

Age also changes what your marrow contains, a topic I covered in how aging affects bone marrow, muscle, and fat stem cells. For a broader overview of marrow-based knee treatments, you can read bone marrow stem cells for knee pain relief.

How we treated the knee, step by step

We ran the study at the Nucleus of Orthopedics and Traumatology in Belo Horizonte. The local ethics committee approved it (approval number 01/2012), and every patient signed an informed consent form. We operated on all eight patients between June 2012 and January 2014.

Who could take part. Adults aged 30 to 80 with patellofemoral arthritis confirmed by MRI and X-rays, and negative tests for autoimmune rheumatic disease. We excluded people with cancer in the previous five years and women who were pregnant or breastfeeding. We also excluded active neurological disease, uncontrolled diabetes or thyroid disease, and heart or lung disease that needed medication. Blood tests had to be negative for syphilis, Chagas disease, hepatitis B and C, HIV and HTLV.

Here is the procedure, in the order the patient lived it:

1. Marrow collection. The patient lay face down. We numbed the back of the hip bone, the posterior superior iliac crest, with 20 ml of 1% lidocaine. With a special needle, we drew marrow into four 20 ml syringes, each holding 1 ml of heparin to stop clotting. We collected 52 ml on average (between 30 and 87 ml). Two patients asked us to stop earlier because of pain and anxiety.
2. Laboratory processing. The samples stayed refrigerated, and the lab processed them 12 to 18 hours later. A density liquid called Ficoll and an automated separator (Sepax) isolated the mononuclear cells. The lab washed the cells twice in saline with 20% of the patient’s own albumin, a blood protein, and spun them at 400 g for 10 minutes. A blood cell counter then measured how many cells each sample held.
3. Anesthesia. Each patient received an epidural block. A cuff on the thigh, inflated to 300 mmHg, kept the knee free of blood for about 15 minutes.
4. Arthroscopy. Through two small cuts beside the kneecap, we inspected the whole knee to rule out arthritis in the other compartments and meniscus tears. A motorized shaver removed the damaged cartilage under the kneecap until the healthy bone below was exposed. We then washed the joint with saline.
5. Cell injection. Under direct view of the camera, we placed a needle at the upper outer edge of the kneecap. We removed the camera, stitched the small cuts and applied a bandage. Finally, we injected 20 ml of cells through that needle.

After surgery, one rule applied to everyone: no anti-inflammatory drugs and no cortisone for three months. We did not want any medicine to interfere with the response of the cells in those first weeks.

What happened after one and two years

The group had four men and four women. Their average age was 52.5 years, with the youngest at 37 and the oldest at 76. Seven had the problem in the right knee, one in the left. All eight had grade 4 damage on the Outerbridge scale, which means the cartilage had worn away completely and bone was exposed. One patient had type 2 diabetes controlled with medication. Because of their knees, none of them did any physical activity.

The cell doses were not equal. Everyone received the same 20 ml, but the amount of marrow collected varied, and so did the concentration. Patient 3 gave 87 ml of marrow and received the richest sample. Patient 8 gave 30 ml and received the poorest one, almost eight times less concentrated.

To measure knee function, we used the Tegner-Lysholm knee score. It rates the knee from 0 to 100 using eight everyday items, such as pain, limping, swelling and stair climbing. We also used the SF-36, a questionnaire that measures quality of life in eight areas. Patients filled both in before treatment, after one year and after two years. Nobody dropped out.

The average Lysholm score was 48.3 before treatment. It reached 97.3 at one year and 96.7 at two years. The jump from baseline was statistically significant (p<0.001). In plain words, the chance that a difference this large appeared by luck is less than 1 in 1,000. Between year one and year two, the change was not significant. The benefit held.

Knee function and quality of life before treatment and after two years

The averages hide some striking details. This is how the answers to single questions changed:

• Pain: at the start, 7 patients had constant pain and 1 had pain after walking less than 2 km. At one year, 7 had no pain at all. At two years, 6 had no pain and 2 felt it only with heavy exercise.
• Limping: 4 limped slightly and 4 heavily. At one and two years, none limped.
• Locking: 4 had occasional and 4 had frequent locking of the knee. Afterwards, nobody did.
• Swelling: all 8 had swelling. Afterwards, nobody did.
• Stairs: 6 climbed one step at a time and 2 could not climb at all. At two years, 6 had no difficulty and 2 had slight difficulty.
• Squatting: nobody could bend past 90 degrees. At two years, 5 squatted with no difficulty and 3 with slight difficulty.

The SF-36 showed the same pattern. The physical function score rose from 10 to 90. The pain score went from 20 to 80 at one year and 100 at two years, and on this scale a higher number means less pain. Social life went from 33 to 100, and the emotional role score from 25 to 87.5. Two areas kept improving in the second year: general health rose from 67 to 82, and mental health from 68 to 76.

During the physical exam before treatment, every patient had severe pain and loud crackling when we pressed on the kneecap. At one year, six had no crackling and two had a much lighter one. Four patients reported recurring pain at the lower tip of the kneecap during the two years. It went away with stretching, cold packs and muscle strengthening. Nobody had an infection or any complaint related to the injection.

Then came the MRI scans. Six patients had shown an increased T2 signal in the kneecap, a sign of stress in the cartilage and the bone below. In all six, the cartilage layer looked clearly better, with no increase in that signal. The 76-year-old woman from the opening is one of them. Her first scan showed broken cartilage over the kneecap. Her scan one year later showed a continuous layer again. The swollen spots in the bone under the cartilage also became smaller.

These numbers look impressive. Before anyone books an appointment, it helps to know what a study of eight people can and cannot prove.

What eight patients can and cannot prove

I wrote this study, and I can list its weak points better than anyone. Here they are, in plain words.

First, the sample is small and there was no control group. Every patient received both arthroscopy and cells. So I cannot tell you how much of the benefit came from the cleaning and how much from the cells. My reading is this: arthroscopy alone tends to give relief that fades, while our scores at two years were almost the same as at one year. That durability makes me think the cells contributed. Only a randomized trial, with one group receiving arthroscopy alone, can prove it.

Second, questionnaires are subjective. A patient who hopes to improve may rate the knee more kindly. The MRI findings and the physical exam partly balance this, but not completely.

Third, the doses differed. The richest sample was nearly eight times more concentrated than the poorest. Interestingly, even the patients who received fewer cells improved. That is encouraging, but it also means we still don’t know the right dose.

Fourth, an MRI is not a biopsy. The images showed cartilage covering the kneecap again. They cannot tell us whether that tissue is true hyaline cartilage or the weaker fibrocartilage. Only a small tissue sample under the microscope could answer that.

Other groups have seen similar signals. In a 2006 case report by Centeno and colleagues, a man with a severely worn hip received marrow cells twice, one month apart. His hip extension improved by 15 degrees, and his MRI suggested partial regrowth of the joint surface. In 2006, Slynarski and colleagues treated 14 patients with cartilage defects using fresh bone marrow under a patch of bone lining. Of these, 57% improved clearly and returned to normal activity within three months, and 13 had smooth joint surfaces on MRI. Still, 17% could not return to their previous level of sport.

How might the cells work? Animal studies point to two effects. Cell therapy lowers inflammatory molecules such as TNF-α and prostaglandin E2. It also switches on the genes for type II collagen and aggrecan, two building blocks of healthy cartilage, and switches off MMP-13, an enzyme that breaks cartilage down. If these terms are new to you, my articles on cartilage proteoglycans and on how inflammation damages joints in osteoarthritis explain them step by step.

What does this mean for you today? If a clinic offers you a cell treatment for your knee, bring these questions to the visit:

• What exactly will you inject? Whole marrow concentrate, separated mononuclear cells, cultured stem cells and platelet-rich plasma are different products with different evidence.
• Does the lab count the cells? In our study, a counter measured every sample.
• Will you treat the damaged surface too, or is it only an injection? Our patients had arthroscopic cleaning in the same session.
• Which published studies support this for my grade of damage? Ask for the journal and the year.
• What is the plan afterwards? Ask about rehabilitation and which medicines to avoid, and for how long.

And if a knee replacement turns out to be the right path for you, good preparation changes the result. I explain how in my guide to rehabilitation before and after knee replacement.

The bottom line for a worn kneecap

Arthritis behind the kneecap is common and hard to treat. Cartilage cannot heal on its own, and the classic surgical techniques give limited results. In our study, eight patients with the most severe damage received arthroscopic cleaning plus an injection of their own bone marrow mononuclear cells.

The results were clear in this small group. The average knee function score went from 48.3 to 96.7 after two years. Constant pain gave way to little or no pain, and nobody limped anymore. Six patients showed better cartilage coverage on MRI, and nobody had a complication from the injection.

The method has practical advantages too. It needs no weeks of cell culture, and the patient receives their own cells. That keeps the risk of contamination and the costs lower than cultured stem cell therapy.

My message is a cautious one. These results justify larger randomized trials, and they do not yet justify promises. If you live with kneecap pain, talk with an orthopedic surgeon about where you stand, and ask the questions listed above. The U.S. National Institutes of Health keeps a list of ongoing research and patient resources on the osteoarthritis page of NIAMS.

One number from the study stays with me. Before treatment, all eight patients struggled on the stairs, and two could not climb them at all. Two years later, six of them climbed with no difficulty at all.

References

1. Guimarães MV, Rios PAGM, Ruiz MA, Noronha JCP, Terra DL, Lana JFSD. Patellofemoral osteoarthritis: treatment with autologous bone marrow mononuclear cells and arthroscopic surgery, a prospective study. J Stem Cell Res. 2018;2(1):1-6.
2. Hunter DJ, Harvey W, Gross KD, et al. A randomized trial of patellofemoral bracing for treatment of patellofemoral osteoarthritis. Osteoarthritis Cartilage. 2011;19(7):792-800.
3. Noyes FR, Barber-Westin SD. Advanced patellofemoral cartilage lesions in patients younger than 50 years of age: is there an ideal operative option? Arthroscopy. 2013;29:1423-34.
4. Mouzopoulos G, Borbon C, Siebold R. Patellar chondral defects: a review of a challenging entity. Knee Surg Sports Traumatol Arthrosc. 2011;19:1990-01.
5. van Jonbergen HP, Poolman RW, van Kampen A. Isolated patellofemoral osteoarthritis: a systematic review of treatment options using the GRADE approach. Acta Orthop. 2010;81:199-05.
6. Steadman JR, Briggs KK, Matheny LM, et al. Ten-year survivorship after knee arthroscopy in patients with Kellgren-Lawrence grade 3 and grade 4 osteoarthritis of the knee. Arthroscopy. 2013;29:220-25.
7. Song F, Tang J, Geng R, et al. Comparison of the efficacy of bone marrow mononuclear cells and bone mesenchymal stem cells in the treatment of osteoarthritis in a sheep model. Int J Clin Exp Pathol. 2014;7(4):1415-26.
8. Centeno CJ, Kisiday J, Freeman M, Schultz JR. Partial regeneration of the human hip via autologous bone marrow nucleated cell transfer: a case study. Pain Physician. 2006;9(3):253-56.
9. Slynarski K, Deszczynski J, Karpinski J. Fresh bone marrow and periosteum transplantation for cartilage defects of the knee. Transplant Proc. 2006;38:318-19.
10. Giannini S, Buda R, Vannini F, et al. One-step bone marrow-derived cell transplantation in talar osteochondral lesions. Clin Orthop Relat Res. 2009;467:3307-20.