Good afternoon. I'd like to thank Dr. Nadim, Dr. Dieb, and Dr. Spheer for inviting me to be with you today. Unfortunately, wars divide and regress the advancement of science, but meetings such as this one that has become a notable event to be reconned on can only advance education, science, and care for our patient. What I'll cover today is FRAX calculators in the MENA region. I have nothing to disclose. I'll start with FRAX in general, the principle and assumption behind the model development and validation, requirements to develop country-specific FRAX model, move on to describing the MENA country with FRAX model, overview them, how robust are they, and what are their limitations, and then conclude with what do you do if you do not have a country-specific calculator. There are two recommendations, one from the IACD-IOF 2011 FRAX International Initiative, and then the FRAX website recommendation, and then conclude with some remarks. FRAX is a web-based tool that was launched in 2008 that provides calculators for the assessment of fracture probability in men and women. It is widely accessible on the website. It's similar to other calculators to calculate for breast cancer for the Gale model and cardiovascular disease development for the Framingham, the Raynald, and other scores. The approach uses easily obtained risk factors to estimate 10-year fracture probability for the hip or for a major osteoporotic fracture, and the estimate can be used alone or with femoral neck BMD to enhance fracture prediction. So, what does FRAX do? FRAX can use BMD with risk factors, knowing that fracture probabilities can be developed with risk factors alone, and these are age, gender, BMI, current smoking, alcohol more than three units, prior fracture, parental hip fracture, glucocorticoid rheumatoid arthritis. These can be used alone or in combination with femoral neck BMD. If one is using femoral neck BMD with FRAX risk factors and adds a checkbox for secondary causes of osteoporosis, that secondary cause does not count anymore because its effect on fracture risk is through the femoral neck BMD. Now, this data is actually showing from meta-analyses how these risk factors increase risk independent of BMD. In blue is without BMD risk assessment, and in orange is with BMD, and you can see that the risk estimates, which are anywhere from 1.8 to maybe 2.5-fold higher with the risk factors, do not change when you add BMD. And what that tells you, therefore, is these affect risk fracture probability over and above BMD. There are many references describing how these risk factors were validated in meta-analyses, how the model was developed, present on the FRAX main website under references. So, the development of FRAX is a two-stage process that includes first determining the impact of risk factors on fracture probability, which I just showed you through large multinational cohorts, identification of common risk factors and meta-analyses, and then superimposing the resulting risk algorithm on the epidemiology of hip fracture, hip or major, and mortality for each country. The primary cohorts included baseline and follow-up data. They were a total of nine. They were based mostly in the US, in European countries, including UK, Germany, Canada, and two cohorts from Germany. 46,000 women and men, 70% women, approximately 4,000 osteoporotic fracture, of which 850 were at the hip. Once this model was developed, it was validated in an independent population-based cohort of 11 total cohorts, again, based in the US, in Japan, in Australia, in Europe, with four-fold higher number of women, 230,000, 1.2 million person-years, 18,000 osteoporotic fracture, of which almost 3,400 were at the hip. The requirement to build country-specific fracs include country-specific mortality, high quality national age and gender-specific crude hip fracture incidence rate per 100,000. They should be recent, representative, including ethnic group. Ideally, incidence rate should be also available for fractures at the humerus, forearm, and vertebral to calculate MOF fracture probability. And in the absence of the above, fracs assumes that age and sex-specific ratio of non-hip to hip fracture are equivalent to those in TAMO, Sweden. There are limitations to fracs. I can't go over them in details. It does not account for all risk factors, such as falls, biochemical markers, weight loss, bone loss rate, et cetera. Fracs plus try to address some of those. It lacks details on some risk factors. For example, smoking is yes, no. There is no PAC years incorporated. Steroids is yes, no. There is no high-dose steroid for many, many years. Prior fractures, if it's one or multiple, it's the same. Parental history of hip fracture, if it's both parents or just one, it's the same. Importantly, it depends on the adequacy of the epidemiologic information. As you know, there are limited country models available, but it's been increasing exponentially over time. And the model was developed for untreated patients. There is some validation in small studies. This is the website as of last week. The model is available now in 86 countries, Middle East and Africa 16, 31 languages. Let's move on to see what's happening in our region. So this is now Fracs, Middle East and Africa. And we've broken them down by nine countries in the Middle East and three countries in North Africa, a total of 12, Abu Dhabi, Iran, Jordan, Kuwait, Lebanon, Palestine, Qatar, Saudi Arabia, Syria, and for North Africa, Egypt, Morocco, and Tunisia. And what I've plotted here on the y-axis is as we progress over years, and on the x-axis, the specific countries. And as you can see, Lebanon was first launched a year after the launch of Fracs, Jordan came shortly after, Tunisia in 2011, Palestine 2012, Morocco 2013, Abu Dhabi 2016, Syria 2019, Qatar, Saudi 2021, and Egypt 2024. Now, let's look at some of the crude fracture incidence rate. In this graph on the y-axis is hip fracture per 100,000. H category are on the x-axis and the color codes are for Lebanon in the greens, Saudi in yellow, Morocco in red, Kuwait in maroon, Qatar in gray, and Egypt in blue. And as you can see, at the earlier years, the curves are relatively closer to each other, but they certainly do diverge when you get to above 70 years, and especially in the extreme ages. The same trend is seen in males, although the fracture rates are lower. The female to male ratio is anywhere between 1.5, 1.8 to 2.2, depending on the country. Now, how does this variation in fracture incidence rate affect Fracs-derived probability? This is the AUB Fracs MENA project we started a few years back. We collected data from our densitometry unit on all patients who walked in who were between age 40 and 90. The period was January 2019 to January 2020. We collected 2,500 cases. These were, of course, approved by the IRB. The population was on the average 65 years of age plus or minus 10, BMI 27, mostly female, 92 percent. 18 percent reported a parental history of hip fracture, 25 percent smoking, 14 percent taking glucocorticoids, 11 percent had secondary osteoporosis, which does not count with a BMD, if it's present. And, of course, all these cases had BMD. Rheumatoid arthritis, 1 percent. Alcoholism was negligible. 50 percent had no risk factors, so it's a relatively healthy population. One-third had one risk factor from Fracs and only 6 percent had two or more risk factors. This busy slide shows on the left the MOF Fracs by age category, less than 50, 50 to 70, more than 70. The color codes are for all the countries. And then on the right is the hip fracture incidence rate, again, by age group with all these countries. As you can see, there are some variations that I'd like to explore further in specific countries. So what I'm showing you here is the Levant, meaning Lebanon, light green, Syria, darker green, Jordan, maroon, Palestine, in dark blue. On the left is the MOF. On the right is the hip. This is the overall population, 90 percent women. And as you can see, the variation in MOF is very low, less than 2 percent until age 70 and gets to 5 percent above age 70. For hip fracture, the variation is even lower in the younger age group between these countries, and it's at the most less than 3 percent for the older population. What about GCC country? The same pattern. The differences are relatively minor between Kuwait in pink, Qatar in darker pink, and KSA in brown. The differences are less than 1.5 percent of the younger age group and at the maximum 4 percent in the older age groups. And for the hip, of course, the same pattern and a maximum difference of 2 percent for the hip. If we go to a higher risk population, subjects with previous fracture and more than two risk factors, the numbers are getting smaller, of course, and we want to scrutinize the discrepancy. Again, as the age goes up, the discrepancy in fracs in the Levant country goes up. It can be up to 7 percent if you look at MOF, and it could be 6 percent if you look at hip. And for Qatar, same pattern. The differences are relatively small for the younger age group. They get to be between 5 and 7 percent for hip and MOF for the older age group. So frac's regional differences are not huge. They're certainly larger when you get to a higher risk population. And remember, the numbers for the high risk in our study group is relatively small. What are the limitations of MENA frac's calculator? Well, first of all, the low fracture numbers, especially at younger ages in these countries with very large confidence intervals, short period to capture hip fracture, partially representative population. I'll talk about that in a minute. It does not take into account secular trends in hip fracture and changes in mortality over time. And none of these frac's calculators have been validated in our region with prospective cohorts. And of course, the assumption that the MOF hip is the same across country, and I don't think this is necessarily valid if we scrutinize the data. Let's look at some more additional information on some of these countries. The only two countries that have their registries based on a Ministry of Public Health population-based approach is from Lebanon and Kuwait. Our study in Lebanon so far is spanning 12 years. The frac's was developed based on data in the earlier year. We may update it. And now we have data up to 17 years. We're actually updating it to 2023. Kuwait, four years, 209 to 212, so nothing recent. Qatar, three years. Saudi Arabia, Egypt, two years each, and Morocco, four years. Importantly for Qatar, it was based on the Hamad medical system, which is supposed to be the principal public health provider in Qatar, but we don't know how representative that is. Is it 60, 70, 80% of the population? Saudi Arabia took many hospitals. And what it actually mentions is that the catchment hospitals captured in this study constitute 12% of the total Saudi population. For Egypt, it was exclusively based on two FLS centers, one in the north, one in the south. And for Morocco, it was several public and private hospital, but we don't know whether they capture a specific and large proportion of the population. Now, what about change in fracture rates over time? So this is a study we started extending our original period of 2006 to 2008, on which our frac's calculator was developed, into several years to 2017. And what we've plotted are the earlier years in blue, and the latest year is in purple. And you can see there's quite a drop in the rate across age group, at least starting 70 and above. This is a representation for both males and females averaging three years together. Women are in red, and there is a gradual decrease in fracture rate, maybe a slight bump up in 2015, 2017, and a gradual steady increase in men. We're now extending the data to 2023. So there are secular trends in hip fracture. And basically, this is not unique to Lebanon, there are secular trends worldwide. Now, if we go beyond Lebanon and stick to the region, I've shown fracture rates per 100,000 in women on the left, and in men on the right. And what I'm doing here is showing the same country in old versus new studies. And you can see systematically for Lebanon, Kuwait, and Morocco, fracture rates have been going down in men. And on the right hand in blue, again, similarly, standardized hip fracture rate in men have been going down, at least when scrutinized in these countries. Now, what about the assumption of MOF to hip ratio being the same as Malmo, Sweden? This is actually when we co-chaired the International Fracture Initiative, we wanted to look at that assumption. So we've tabulated MOF to hip ratio in Sweden, which is the population on which everything else is flowing, because many countries don't have MOF fractures. And we're looking at rates in USA and in Switzerland. This is for men, these fracture rates ratio are not very different. If you look across the lines for the various age group in men, but they certainly are for women. Looking at USA and Switzerland for 50 to 55 wide variation, 60 to 64 wide variation, 65 to 59, sorry, 69 wide variation, it could be anywhere from two to three-fourths rate ratios, and even up to four in certain age group. Now, we went back to the chart in 2021. And Dr. Shakhtoura in our group actually did a systematic review, looking at major osteoporotic fracture to hip fracture incidence rate ratio. This is a northern European female, this is the ratio on the y-axis, these are the age group on the x-axis, and she plotted them. And as you can see, they are not identical. These are color codes for various countries. And then what we did is we actually applied statistical tests to assess whether these rate ratio were different from Sweden at the various age group in women. And the countries represented that were scrutinized are on the longitudinal axis in the column, and then the age group are across the x-axis. Now, if you have a wide area, that means that the rate ratios do not differ in Japan compared to Sweden, except for the older women. Similarly, in Australia, for that study, they didn't differ. Conversely, they are different in the light gray boxes, meaning being lower than Sweden, in Denmark, in Finland, in Iceland, in Italy, in the Netherlands, and in the UK, in most age group, not all. And it's higher as indicated by the dark area in Japan in the second study from Australia, and in Canada, only in that age group. So it's not as homogeneous. And I think we need to understand more what's happening with these race ratios and that assumption. And Eugene McCloskey actually notified me last month that they actually have a study now, specifically looking at this issue. Now, FRAX is there and is there to stay. MENA countries have increasingly adopted FRAX, as I've shown you. What do we do for countries that do not have a country-specific calculator? Well, I think it's important. We cannot choose any country calculator because there are wide variation in fracture risk worldwide. These are age-standardized hip fracture rate that we derived a while ago. I've highlighted Lebanon and Kuwait here. This is older data, of course, but we're in the lower half. And if we look at more recent data, we're probably going to be in the lower third. So what is the official position of the IOF and IACD on FRAX regarding international differences? I won't show you all the work behind it, but let me just share the recommendations with you. Recommendation 27, in the absence of high-quality national hip fracture data, you may want to choose a country-specific FRAX model that uses hip fracture incidence rate from the surrogate country incorporating country-specific mortality rate. Recommendation 28, in the absence of any hip fracture data, you could potentially develop a FRAX model based on broad categories, high risk, medium risk, low risk across the world, and adjust them to country-specific mortality rate. Has this been done? Well, it looks like it's been done when I inquired with John Canis in Syria. When they introduced their model in 2019, they used Lebanon for hip fracture data and Syria, WHO mortality data. Same thing in Palestine. When they launched their model, they used Jordan for hip fracture data and Palestine for WHO data. This is when the country models were launched, as I've shown you before. We have updated mortality in Lebanon in 2012, and there was an update in 2012 for Jordan, but we're not sure what the update was all about. The FRAX notes usually give you some information. All the blank is because we didn't have any other information. Now, what does FRAX say about not having a country represented? They modified a little bit the ISCD-IUF FRAX recommendation 27. They say, use the country for which the epidemiology of osteoporosis most closely approximates your country. And high-risk country would be Denmark and Sweden. Low-risk country would be Lebanon and China. And of course, they said new models are expected to be made available and lobby your national society to have a country-specific model, or at least to help you build a surrogate model. What about changes in rates and life expectancy? An essential thing is having robust, recent, valid hip fracture rate, but also mortality data. And they state significant changes will affect the accuracy of the model so that FRAX model needs adjustment from time to time. This is an adjustment we did for Lebanon. This is published in Bone in 2014. We actually changed from the mortality WHO in 1999 and lobbied for changing to a more recent 2007. This was because we used 2007 hip fracture rates, and the 10-year FRAX-derived probability of MOF and hip increased substantially. The changes were most pronounced in high-risk subgroups, going with a relative risk estimate increasing by 80% to 120%. So this is not a minor point, and I think it's very important to be cognizant. In conclusion, FRAX is a widely used fracture risk assessment tool. It requires robust country-specific hip fracture incidence and mortality data. The ratio-similarity assumption needs to be scrutinized. It misses on certain risk factor, and FRAX plots may partially remedy to that. Secular trend and fracture and longevity have to be kept in mind. What about our region? Well, our rates are one-third those of worldwide. Our MENA region has 12-country calibrated FRAX calculator. The original data was not always available in peer-reviewed journal. Abu Dhabi is one example. The data is limited in terms of number of fracture, representativeness, and recency of the fracture and mortality data. FRAX for MENA country has not been validated in any countries in prospective cohorts. There are differences in FRAX-derived probability, MOF, and hip between MENA, but they're relatively minor, anywhere from 2% to 3% in low-risk and up to 6% in high-risk for these probabilities. The surrogate country is a reasonable alternative. Now, I think having looked at the regional data, I think a regional FRAX MENA is a possibility that could be explored. I'd like to thank you for attention and invite questions.

FRAX model

fracture risk

MENA region

risk factors

epidemiologic data

regional adaptation