The Causation of “Bisphosphonate-Associated” Atypical Femur Fractures

Antares Placido topographer CSO had been used to measure 50 eyes corneas at
Ophthalmological Clinic of the Greek Red Cross hospital «Korgialeneio - Benakeio in
collaboration with University of West Attica and their refractive maps were taken. The
refractive power map was taken before cataract surgery with the method of phacoemulsification.
The keratometric measurements for 3 & 5mm were recorded before and
one month after surgery. The correlation of the K1, K2 readings before and one month
after the phacoemulsification procedure, do not cause changes in the refractive power of
the two main meridians of the cornea of patients, in order to alter the final astigmatism
refractive post-operative effect, to a statistically significant degree.


Introduction
Osteoporotic bone fragility is the result of a natural process resulting from the loss of estrogen associated with aging.
"Osteoporosis" and "osteopenia," as these terms are universally used today, do not conform to the typical definition of disease. Over the last 25 years, the definition of osteoporosis has slowly evolved into a conflicting group of several independent meanings. The most common definition of osteoporosis used today conflates mutually exclusive terms. For many years, the diagnosis of osteoporosis depended upon the prevalence of a fracture. This event, examined in the context of the individual's physical examination, could be interpreted in appropriate circumstances as due to an intrinsic lack of resistance to fracture. i.e., osteoporosis. However, in 1992, a group convened by the World Health Organization (WHO) and meeting in Milan, agreed to define osteoporosis, not only by a prevalent fracture, but also as a bone density 2.5 standard deviations below that of an average 30-year-old white woman. This decision transformed the treatment of a specific clinical illness into a "disease" based on the supposed discovery of a risk, introducing bone mineral density as a surrogate for clinical disease. Subsequently, "osteopenia" became defined, by default, as a bone density lower than one standard deviation below but greater than 2.5 standard deviations below that of an average 30-year-old woman. The adventitious nature of this introduction of so-called osteopenia into clinical medicine has been publicly acknowledged [1] by one of the most prominent scientists involved in many of the Phase 3 bisphosphonate clinical trials as inappropriate. The decision to use these definition limits "Some…heretofore moved by the reasons of Galen and Colombo…will yeeld." William Harvey 1628 was arbitrary and was perhaps initially only intended to establish degree of risk, and not to afford those terms diagnostic significance [2,3]. Yet these criteria have been widely popularized to imply that millions of apparently healthy persons now have a disease called osteoporosis requiring treatment. These millions were immeasurably increased by the additional attribution of disease status to the mineral densities which were higher than, but close enough to, those assigned to osteoporosis.

The Search for a Cure
The definition of individuals with an appropriate low-impact prevalent fracture or a newly discovered risk was concurrent with the avid search in the pharmaceutical industry for a treatment. The number of individuals meeting the new definition represented a vastly increased market. Ironically, since that time there has been little, quite sporadic cooperation between specialists concerned with the former definition of osteoporosis (a prevalent fracture resulting from negligible trauma), to wit, the orthopedic surgeons, and the several industries proffering medical treatment. Most of the embrace of these risk-aversive preventatives has been with the specialists who have become ever more familiar with the current emphasis on prevention.

Establishing a Meaningful Definition of Causation
Sir Austin Bradford Hill was a Professor of Biostatistics at the University of London. In 1965, he gave a plenary lecture [4] at the first meeting of the newly created section of Occupational Medicine of the Royal Society of Medicine. Leaving aside many of the niceties which have complicated this issue over the centuries, he proposed a practical method to enable moving from a recognition of association between factors and disease to the conclusion of causation. This is exactly the step which the Task Forces, convened by the American Society of Bone and Mineral Research (ASBMR) to evaluate the association between the bisphosphonate drugs and the complications of osteonecrosis of the jaw [5] and atypical femur fractures [6,7] declined to take, despite the fact that they finally acknowledged that the evidence for the association was "robust" [7]. Bradford Hill proposed a set of well-defined criteria upon which such a conclusion could be reliably reached. He argued his criteria as sufficient despite the semantics indulged in by a long list of philosophers addressing causation. His point was that once an association has been recognized, and possibly before the frequency of that association has been established and almost certainly before the mechanism of action of the association is known, it was The Association requirement is met because bisphosphonate use is associated with a detectable increase in the incidence of nonhip femoral fractures. As noted above, the second ASBMR Task Force has recognized this association and called it "robust" [7]. These sorts of datasets have continued to accumulate. As an illustration: One meta-analysis [11] of all the association (incidence) studies  [12] disclosed an incidence of such fractures in ever-users vs never-users of bisphosphonates of 4.3-fold. In users of 3 or more years, the incidence increased to 9.6-fold.
There was 46 times greater bisphosphonate use in ST/FS fracture patients in two major Swedish hospitals over 1 year than in comparable patients who were not using the drugs [13]. This ratio was later verified for the entire nation [14]. The Consistency requirement looks to independent observers to be found to report similar experiences with the side-effect in question. The finding of the association has been consistent in many independent studies and separate venues. The ASBMR Task Force [6,7] summarized independent reports from 17 different countries. The wide difference of incidence reported from various sources has been attributed to differences in data collecting techniques with actual radiographic examination [15][16][17] generally finding a greater incidence than retrospective review of diagnostic codes. Some of these differences may also be attributable to observer bias, as the challenge of these fractures has occasioned intensely defensive reactions in some quarters.
The Specificity requirement is met, not because bisphosphonates are the only cause of these femoral fractures, but because most of the causes of these fractures share the same end point to their common mechanism of action. Bradford Hill [4], the FDA-acknowledged authority in this matter, specifically warns that the mechanisms producing any given side effect may be similar even if quite disparate agents or conditions are causative, thus multiple causes may deceptively appear to challenge specificity.
In fact, in my opinion, this particular specificity of mechanism Temporality is met because bisphosphonate-associated fractures occur after the ingestion or administration of bisphosphonates rather than in persons never using the drugs. One analysis [19] showed a steadily increasing incidence of very specific femur fractures since 1999 despite the lack of apparent increase in the hip fracture incidence. A U.S. National Institutes of Health analysis showed that non-hip fractures are increasing while hip fractures are decreasing, even in the face of a 50% decrease in the use of bisphosphonates over the study interval [20].
Satisfaction of the Biological Gradient depends on an adequate understanding of the metabolism of the bisphosphonates.
The bisphosphonate medication used for osteoporosis are not metabolized and a long residence in the skeleton is maintained.
The terminal half-life has been shown to be around 10 years. This expression refers to the persistence of the drug after re-dosing is stopped [21]. This almost unique situation means that gradient issues must be understood through duration of therapy, not dose. The pharmacokinetics of the bisphosphonates differ qualitatively from the typical kinetics of drug metabolism.
Any dose saturates available binding sites, but only additional doses expose new sites to saturation, the prior dose remaining where it attached. The burden of the drug thus accumulates over time [22]. However, the reported decreasing risk of a second attributable complete fracture if the drug is stopped when the fracture occurs is not necessarily evidence of a withdrawal gradient [23,24]. This risk decrease is clearly not connected with a meaningful reduction of the drug burden, but rather with the circumstances surrounding the development of the second fracture. Most complete ST/FS fractures initiate laterally as insufficiency incomplete cracks [25]; the potent stimulation of resorption and formation activity which normally (in the absence of bisphosphonate) combats progress of such a fracture line is inhibited by circulating bisphosphonate which is only present to a large degree when re-dosing is occurring.
The occurrence of the initial fracture is likely associated most times with cessation of re-dosing. If it is not, the incidence of a second fracture is greater [26].
A very well-done recent study [27] establishes beyond a reasonable double the relationship between duration of therapy and risk. The Plausibility requirement is met by the many consistent mechanisms whereby the use of the drug could result in bone fragility. The ASBMR Task Force has acknowledged seven plausible mechanisms of action [6,7]. All of these are quite reasonable possibilities. One of the plausible mechanisms alluded to but not expanded in the ASBMR Task Force report relates to an important anatomic parameter long recognized to influence the toughness or energy-to-fracture of bone [28,29]. The bone mineral density distribution is a measure of the heterogeneity of the bone mineral matrix [30]. The wider the distribution, the more varied the hemi osteon and osteon density will be. Bone remodeling is a discrete and localized process which is initiated at new locales continuously, each completed remodeling being followed by a continuous but not linear mineralization influx into the tiny newly minted bone packets over a period of time often measured in years.
As the initiation process unrolls, these micro loci will necessarily have differing mineral densities depending on how much time has passed since mineralization began.
The result is a three-dimensional patchwork of tiny bone parcels which differ in density from their immediate neighbors.
These interfaces between these new parcels, quite appropriately termed "cement lines", act as barriers to crack propagation [31].
Furthermore, the lateral femoral shaft in untreated individuals in the steady state harbors more intrinsic microdamage than the medial shaft [32][33][34]. This sets up the lateral shaft for a fracture and probably results from and adds to the anatomic stress concentration long known to maximize there [35]. Bisphosphonatetreated patients have a more homogeneous bone due to the marked decrease in new remodeling initiations and the resultant permitted movement of all parcel densities toward a uniform maximum.
In addition to this, the femoral shaft has been shown [35] to be more homogeneous than the hip to begin with, and exaggeration of this tendency by bisphosphonates has been shown by biopsy of fractures in that region during orthopedic repair. All of these factors combine to make it very plausible indeed that the ST/FS should be the expected major location of these breaks.
That said, however, it is important to note that there is no consensus that the bisphosphonate effects eventuating in insufficient bone are confined to the femur. Indeed, the original poster [36] presentation revealing this problem to the world in 2003 described similar events in other cortical locations. Around 30% of the cases I reviewed with Jennifer Schneider in 2012 [26] also sustained fractures of the metatarsal bones and we also had occasion later to review many other non-femur sites where the same general fracture types have occurred under the influence of the bisphosphonates [37]. In addition to these and the ASBMR-cited mechanisms, several others may be cited. A plausible theory has been advanced [38] that micro-crack propagation through the homogeneous medium created by chronic bisphosphonate therapy can result in one master crack that may propagate unrestrained to eventuate into a macro-fracture. The lateral initiation particular site of the injury is the result of the maximum standing stress produced in this area due to the shape of the femur. This distribution of force has been well-understood by anatomists for the last hundred years [35].
Many workers in this field, especially in the Scandinavian countries, suggest that the origin of all these fractures is in a form of stress (un-displaced) fracture that differs from the athletic stress fracture in that the latter occurs in adequate bone subjected to chronic excessive force whereas the former occurs in bone rendered inadequate as mentioned above and subjected only to normal force. [39] For this reason, the term "insufficiency fracture" has become favored as it better describes the operative mechanism.
The coherence requirement is met because there is no conflict between these mechanisms and the laws of physics and biology.
The reported mechanisms have been the product of application of the scientific method in reputable laboratories around the world.
Experiments performed at the material level indicate a decreased toughness of the bone after bisphosphonate exposure [40]. Femur fractures have been demonstrated in bisphosphonatetreated dogs [41]. Bisphosphonates have been experimentally shown to increase the proportion of non-enzymatic glycation products (crosslinks) in bone collagen, an effect associated with decreasing toughness [42]. Bisphosphonates have been shown to decrease cortical modeling, an effect which likely compromises the integrity of the femoral shaft [43]. Animal experiments have also shown [44] that the decrease in bone toughness occasioned by a bisphosphonate is in experimental animals can be reversed by prior or concomitant treatment with corticosteroids, thus quite possibly exonerating corticosteroids from contributing to the fracture problem even though the impact of steroids on the formation phase of bone is well known. Animal experiments have also shown that the presumed inverse correlation between bone mineral density and fracture risk is misleading and false [45]. Finally, physical chemical experiments have been [46] and can be conducted that may confirm that the exposure of bone to phosphorus smoke results in the binding of pyrophosphoric acid, a bisphosphonate analog, to human bone, thus reproducing the femur fragility occurring in one of the analogous contexts mentioned below.
The Analogy requirement was left somewhat optional by Bradford Hill as he recognized that analogies might not always be available. The example he gave was rather weak. But there have been some striking analogies to the drug-induced osteonecrosis and atypical femur fractures pointed out once an understanding developed that effects of bisphosphonates paralleled those of the analogous pyrophosphoric acid [47]. The discovery of the antiresorptive effects of the drugs was preceded by discovery of the antiresorptive effects of pyrophosphoric acid in physiologic conditions in three laboratories in the interval 1968-1980 [48][49][50]. Pyrophosphoric acid is an abundant product of human metabolism. As described above, it was research in the chemistry and physiology of pyrophosphoric acid that led to the interest in bisphosphonates, which are structurally analogous to it. Pyrophosphoric acid and the bisphosphonates share the pharmacologic properties of antiresorption and antimineralization. Like the osteopetrosis described above, there is a group of human genetic anomalies which result in increased levels of pyrophosphoric acid due to an inherited enzyme dysfunction which normally metabolizes the moiety. This syndrome, in the least extreme form -adult hypophosphatasia, has been recognized [57] as producing fractures that resemble the atypical femoral fractures caused by the bisphosphonates. Again, this presents an excellent satisfaction of Bradford Hill's requirement of rational analogy. Again ironically, the Fleisch group studied [58] this syndrome secondary to their original interest is pyrophosphoric acid. They quantified the increased levels of pyrophosphoric acid in affected individuals, but they never speculated, at least in print, as to the possibility of induction of the fractures suffered in hypophosphatasia by treatment with their acknowledged analog.
One more analogy is currently available. Workers, exposed to the fumes of burning white phosphorus, in phosphorus match factories in the latter half of the 19th century occasionally developed osteonecrosis of the jaw and/or low trauma femoral fractures. Analogous to the hypophosphatasia victims, who were exposed internally to excess pyrophosphoric acid, these workers were exposed to an external source of the same molecule, since the fresh smoke from burning white phosphorus has been shown [59] to consist largely of the two substances shown by Fleisch [48] to have antiresorptive properties: pyrophosphoric acid and mixed polyphosphates. The details of this interesting analogy have been presented [60] elsewhere.