Search the internet and you may find estimates of vertebral compression fractures (VCF) in the US. There are statistics out there and they are often quoted by doctors and research articles. However, what most people don’t realize is that these numbers are based on estimates and that those estimates are over 20 years old.
We have learned a great deal about osteoporosis in the last two decades, but we’ve only begun to investigate the tip of the iceberg. At the same time, there is a steady increase in not just the overall US population, but the number of patients at risk for osteoporosis due to age.
When outdated numbers are used, it can make the problem seem smaller than it actually is. My goal is to help people understand how common vertebral compression fractures are, because they are a significant cause of pain and suffering as well as death. As it turns out, vertebral compression fractures each year in the United States actually numbers around 900,000. Of those, around 300,000 probably present to the doctor. That’s about 20% higher than widely quoted statistics, which roughly correlates to the increase in population over the last decade.
If you’d like to know more about how I came to that number, or just more about fractures, then read on. There’s also a great tip to help doctors identify symptomatic fractures at the end of this article.
Osteoporosis By The Numbers
Osteoporosis is a huge problem, particularly with an aging population. It is estimated that there are at least 10 Million people in the United States that suffer from osteoporosis. About 1 in 5 are men and 80% are women. Osteoporosis affects people by creating a level of low bone mass – that puts people particularly at risk for getting fractures from just everyday normal activities.
There are another estimated 34 million people in the US who have “low bone mass.” While a screening test may show “low bone mass” or “osteopenia” these people are also at increased risk of getting fractures. It’s a myth that “osteopenia” or “low bone mass” is something different from osteoporosis. In reality, bone density is a spectrum from clearly normal to clearly abnormal. The term “osteopenia” is an artifact from an arbitrarily used cut off point when measuring bone mass with DEXA.
The problem with bone mineral density testing and DEXA is that these are tools designed to help predict who will have a fracture. While a radiologist may ‘diagnose’ you with osteoporosis or osteopenia, predicting the fracture risk of an individual is difficult. There are many people with severe osteoporosis that don’t have fractures. Likewise, there are people with normal bone mass who get fractures because they have problems with bone quality from medications, medical conditions and lifestyle choices. This is generally referred to as secondary osteoporosis.
How Common Are Vertebral Compression Fractures (VCF)?
That brings me to the topic of vertebral compression fracture (VCF). How do we predict those? There is some good data supporting increased risk once someone experiences a VCF. Once you have one VCF, your risk for a second VCF increases by a factor of 5. Once you have a second, it increases to about 10X increased risk. By the third VCF, the risk is about 72X or 7200% higher than someone with no fractures. At that point, it’s really not a question of “if” but rather “when” another fracture is going to occur.
So how many VCFs are there in the US each year? The most commonly quoted statistic is about 750,000 VCFs per year. This comes from data from the National Osteoporosis Foundation back in 1997 (Melton et al, 1997). However, it’s important to remember that in the last two decades, the US population has increased from around 273 Million in 1997 to an estimated 327 million in 2018. That’s a 20% increase. So at first thought, we might expect about 20% more fractures, or a total of 900,000 VCFs in the US per year.
How could we test that idea? The best metric we have is information from the CMS dataset. While this method isn’t perfect, it does give a rough idea of how many Medicare patients presented to the doctor with a painful VCF. However, before we get to that, there’s another important consideration. As a rule of thumb, we think that about 2/3 of VCFs are silent—that is the patient either never notices pain or the VCFs heal on their own and never get mentioned to a doctor. So, that leaves about one out of three VCFs where the patient actually ends up seeking medical care.
When I was in training we were using the 750,000 VCFs per year in the US figure. It was less than 10 years old at the time. In fact, that’s the figure that’s still widely quoted, although it is unlikely to be accurate given the change in total population and the shift in age of the US population.
Using the 750,000 figure, based on 273 million US population 1997, we would expect about a VCF to occur in every 364 people in the US per year. That’s 0.275% of people (75×10^4/273×10^6=0.00275). While that figure is an estimate and not perfect, it does give a reasonable way to estimate fractures. However, if only 1/3 come to clinical attention, that means that only 0.0916% or one out of every 1,092 people will experience a symptomatic VCF each year.
Using this figures, and the 2018 US population of 327 million, we could estimate that about 900,000 (898k) VCF occur each year and about 300,000 (299k) come to clinical attention.
How Many Vertebral Compression Fractures Occur In Oklahoma And The Tulsa Area Each Year?
If we use the above information which are derived from estimates from the NOF and generally available population data, that means that in the state of Oklahoma (population of 3.9 million), we would expect about 10,714 VCFs per year. Of that, we would expect about 3,571 of those to present to a doctor, clinic or hospital each year.
I’ve used this methodology for years for talks because the 1997 estimates are now two decades old. More specifically, I practice in Tulsa, Oklahoma and people often ask me how many VCFs there likely are. Current population estimates for the Tulsa metropolitan statistical area (MSA) and combined statistical area (CSA) are 953k and 1.12M respectively. Based on the current population, this would suggest that each year we have about 2,600 VCFs in the metro area, of which about 900 would be expected to come to clinical attention. In the larger CSA region those estimates would be 3,100 and 1,000 respectively. The Tulsa CSA includes other areas including Muskogee and Bartlesville.
Is There Support For This Updated Estimate In The Literature?
These numbers always make me wonder how good we are at identifying these fractures – and are these patients getting treated? Also, has anyone updated these numbers from the 1997 data? It seems that’s the number most people use in their statistics and talks, but it’s out of date. Is there any support for the numbers that I have been using in the literature?
As it turns out, in 2018, a Ong et al published an interesting article in Osteoporosis International titled “Were VCF patients at higher risk of mortality following the 2009 publication of the vertebroplasty “sham” trials?” One thing they found is that after the NEJM articles critical of vertebroplasty were published, that fewer patients with VCF were treated. As a result, mortality in this group of patients increased. That’s because both balloon kyphoplasty and even vertebroplasty have been shown to dramatically decrease mortality in symptomatic VCF patients. In fact, this study showed that 85% of these VCF patients died in the following 10 years. That’s a 10 year survival of 15% which is lower than many cancers.
Another interesting factoid from Ong et al (2018) is that by searching the CMS dataset, they found that there were 239,325 Medicare patients diagnosed with VCF in 2005. That would be the 1/3 of patients who present to the doctor. So I looked at the US population in 2005, which was estimated at 296 million.
If we go back to my methodology, that based on the NOF 1997 estimates that 1/364 people (man/woman/child) get a VCF, then we would have expected about 813,000 total VCFs in 2005. Of those, 1/3 or about 271,000 VCF would be expected to be symptomatic and actually present to a doctor.
Comparing my population-based methodology estimate of 271,000 to the findings of 239,000 Medicare patients of the Ong et al study, these are within about 13% of each other. So, it seems that the 1997 NOF estimates may be pretty close to what we see based on claims data.
Why Is This Information Useful?
If you’ve made it this far, you probably have an interest in osteoporosis and public health. There are several important facts that we can deduce from this information.
First, we can assume that about 1/364 of the population will sustain a VCF in any given year. Considering the 1997 NOF number was an estimate, this gives reasonable validation to that estimate and serves as a way to estimate VCF incidence in a population, at least in the US.
Second, it seems that the old rule of thumb of roughly 2/3 VCF being clinically silent is true. Certainly, using the claims data from CMS in the Ong et al study, this seems to be a reasonable assumption.
Third, this simple estimates of 1/364 and 1/1,092 for the incidence of all VCF and symptomatic VCF respectively appear valid and should be used for predicting these events.
This information should find applications in public health, particularly at a system level. This may help payers, government and health care providers estimate prevalence of these conditions and respond accordingly. If CMS used these numbers as benchmarks to get patients treated for osteoporosis, that could significantly decrease the severity and burden of disease and result in significant savings to the US healthcare system. Optimally, it would be nice to see these used as targets for number of expected osteoporosis patients to treat at a healthcare system and even an individual primary care practice level.
Fracture Estimates For Primary Care Doctors
The following is speculative, but for my primary care colleagues, it’s worth considering.
Let’s say that you’re a family practice provider. This could work for other PCPs, but I like family practice because these estimates are population based (every man, woman and child) and probably fit best for a family practice. If you’re an internal medicine doctor who sees mostly geriatric patients, these numbers are too low.
However, let’s say that you’re a family practice provider and have 3,000 patients in your practice. Using these estimates, that means that you probably have 8 patients that developed a vertebral compression fracture last year. How many did you actually diagnose? Likewise, you probably had 3 patients with a symptomatic vertebral compression fracture. Those are the patients that are really suffering and typically have 7-10/10 pain.
If you didn’t identify that many last year, it’s likely that they were mistaken for a flare of back pain. It’s not your fault. They are difficult to diagnose and the tests we are recommend are pretty low quality for accuracy. For example, it’s well known that imaging misses over half of fractures. And while STIR on MRI is generally considered pretty specific when called, there are some studies that suggest that using STIR on MRI misses up to 2/3 of acute symptomatic VCFs. So using STIR on MRI may be actually worse than plain x-rays. In fact, using STIR to ‘age’ or ‘date’ a vertebral compression fracture has NEVER been clinically validated in controlled trials. So this practice is based on empiric data rather than evidence based medicine.
The good news is that the most specific test that I’ve ever found for a symptomatic vertebral compression fracture is the closed fist percussion exam (CFPE). This is a simple clinical exam that picks up 87% of symptomatic fractures and is 90% specific (Langdon et al, 2010). If you’re not using this test, you should. I use it in every back pain patient I see to exclude a VCF.
CFPE picks up almost all symptomatic fractures and is definitely superior to MRI. In fact, with only rare exceptions, I never do kyphoplasty unless a patient is CFPE positive.
Why Does This Matter To Me?
As a musculoskeletal radiologist in practice, I alway wonder how many patients are out there suffering unnecessarily from symptomatic VCFs. My grandfather died while suffering from vertebral fractures that were missed in a VA hospital in the 1980s. I hate to think about people like my parents or grandparents suffering and dying unnecessarily because of unrecognized and untreated VCFs. I’ve never had a fracture myself, but over the years I have seen a few more of my family members and friends get them. I’ve also seen thousands of patients suffering from painful VCFs.
Luckily, for patients with painful VCFs, kyphoplasty is a simple, elegant and extremely effective procedure. Success rates for pain relief in the literature are about 94%. There is no other procedure in modern medicine with that high of a success rate for relieving severe back pain. Read that last sentence again.
Granted kyphoplasty ONLY works for a painful vertebral compression fracture and it does NOT treat the underlying osteoporosis causing the fracture. Therefore, in my opinion, any patient treated with kyphoplasty should receive medical osteoporosis treatment.
In my practice, I see patients in Tulsa. We see patients from all over the Tulsa CSA, many statewide, even from outside the state in the region. While there are over a dozen physicians trained in Tulsa to perform kyphoplasty, I seem to consistently perform a larger volume that most places. Part of the reason is that since my training I’ve always focused on identifying, accurately diagnosing and when present, reating vertebral compression fractures and osteoporosis. I treated about 120 VCF patients with kyphoplasty during my fellowship year at the University of Oklahoma. I did more during my year of training than most practicing doctors had done in their career. That’s not to knock anyone, it’s just to state that I have been blessed with a lot of experience in treating fractures due to my training.
When I went out into practice, I quickly found that there was a shortage of doctors treating osteoporosis in Tulsa. I would treat patients with kyphoplasty in the hospital, assuming that their primary care doctor would treat their osteoporosis. Most of the time, I saw the patients back with another fracture and they hadn’t been treated. That’s when I knew that I had to do what was right.
My practice has grown over the years. A lot of that is due to experience as I treat hundreds of acute or subacute symptomatic vertebral compression fractures per year. A larger reason is that I have always focused on doing the right thing and medically treating the underlying osteoporosis that causes these fractures. I continue to see patients who have been treated for osteoporotic fractures but never received any kind of medical treatment for osteoporosis. We try to stop the patients from falling through the cracks before they get more literal cracks in their bones.
Now my approach has been validated in the largest ever trial of kyphoplasty in Medicare aged patients, the EVOLVE trial.
What About Vertebroplasty?
While I mention kyphoplasty, you might ask about vertebroplasty. Vertebroplasty is also effective and does save lives. However, there are multiple studies now that have excellent data showing that kyphoplasty is unequivocally better at saving lives compared to vertebroplasty. Some studies show that vertebroplasty is only about 50% effective at saving lives as kyphoplasty. There are also fewer complications with kyphoplasty than vertebroplasty.
My experience includes treating thousands with kyphoplasty. The main reason is that I also treated hundreds of VCFs with vertebroplasty and clearly saw a pattern of improved outcomes and fewer complications with kyphoplasty. Today, looking at the literature, I cannot ethically recommend vertebroplasty over kyphoplasty for most patients with active, symptomatic VCF. It is also my opinion that if a fracture fails to heal with conservative therapy, then nonsurgical management is absolutely unethical. It’s pretty clear from the literature that kyphoplasty saves lives compared to non-surgical management.
- Ong, Kevin et al. “Were VCF patients at higher risk of mortality following the 2009 publication of the vertebroplasty “sham” trials?” Ostep Int. 2018 Feb;29(2):375-383.
- Langdon et al. “Vertebral compression fractures – new clinical signs to aid diagnosis.” Ann R Coll Surg Eng. 2010 Mar; 92(2): 163–166.
- Melton LJ, 3rd, et al. “Fractures attributable to osteoporosis: Report from the National Osteoporosis Foundation.” J Bone Miner Res. 1997;12:16–23.
– James Webb, MD