Trabecular Bone Score as an Index of Bone Fragility in Patients with Acromegaly

Introduction

Growth hormone (GH) and insulin-like growth factor 1 (IGF-1) physiologically exert a beneficial role in skeletal health, enhancing bone formation and remodelling [1]. Based on these actions, acromegaly has long been neglected as a condition potentially associated with skeletal fragility [2^, 3]. This paradigm, however, has been revisited in recent years, as an increasing body of evidence demonstrated that GH and IGF-1 excess may cause abnormalities in trabecular and cortical bone microstructure, with an overall increase in bone turnover and decrease in bone strength, and ultimately a higher risk of fractures [4^-8].

Despite these advances in knowledge, however, the management of skeletal fragility in acromegaly remains challenging. Recent evidence suggests that standard bone-active therapies may be effective in preventing vertebral fractures in patients with active acromegaly [9]. However, guiding treatment choices based on the assessment of fracture risk according to standard algorithms might be inappropriate in acromegalic patients, as one of the most relevant “classical” determinants of skeletal fragility, i.e., the bone mineral density (BMD) does not clearly predict the occurrence of vertebral fractures, which in this setting may arise even in the presence of normal BMD values [5^-8]. In fact, due to its peculiar pathophysiology, acromegalic osteopathy is not characterized by a significant impairment of BMD [5^, 10]; in acromegalic patients, BMD values at the lumbar spine, femoral neck, and total hip are at least comparable or even higher than those of control subjects [5]. Therefore, alternative parameters of skeletal fragility should be sought, in place of BMD, for a better fracture risk assessment in acromegalic patients [10].

In recent years, the trabecular bone score (TBS) has emerged as a novel index of bone micro-architectural health and skeletal fragility [11^, 12]. TBS is a textural metric that captures the mean rate of gray-level variations in lumbar spine dual-energy X-ray absorptiometry (DXA) images [11^, 12]; a high TBS value reflects a better trabecular bone microstructure, while a low TBS value indicates a trabecular microstructure impairment [11^, 12].

Within the setting of acromegaly, some explorative studies noted that, contrary to BMD, TBS values might be lower in acromegalic patients with respect to control subjects [13^-18]; in other papers, however, no significant differences in TBS between acromegalic patients and controls could be found [19^, 20]. Other aspects of this topic are also matter of active debate, such as the role – within this context – of gonadal status [14^, 16^, 19] and acromegaly disease activity [17^, 18]. Finally, and most importantly, the actual performance of TBS as a predictor of fracture risk in acromegalic patients still needs to be fully established and prospectively validated on large patient cohorts.

A previous meta-analysis [21], which included articles published until September 2020, did not find any significant differences in TBS between acromegalic patients and controls; however, this result was limited by the low number of available studies; moreover, no data assessing the relationship between TBS and fragility fractures have been presented by the authors, as only one study evaluating this outcome was available at that time [18]. Within the last 2 years, additional research has been published in this regard, both about the difference in TBS between patients with and without acromegaly [13^, 15^-17] and about the comparison of TBS values between acromegalic patients with and without prevalent fragility fractures [13].

The aim of this systematic review and meta-analysis was, thus, to evaluate all published data about TBS as an index of skeletal fragility in patients with acromegaly, in order to provide an up-to-date synthesis of the current knowledge in this regard and, more specifically, to quantitatively summarize the available evidence within the areas of uncertainty that we have previously outlined.

Methods

Results

Search Results

A total of 218 records were identified in the initial literature search. Removal of duplicates led to an overall pool of 134 studies. An accurate title or abstract revision was sufficient to exclude 122 articles as not pertinent or not fulfilling our pre-specified inclusion or exclusion criteria. The remaining 12 studies were assessed in full-text for eligibility [13^-20^, 25^-28]; among these, 1 did not report any quantitative data on TBS [28], while 3 others evaluated different outcomes from those of our interest [25^-27]; the remaining 8 met all criteria for being included in the final analysis [13^-20] (Fig. 1), encompassing 336 acromegalic patients and 490 non-acromegalic controls. All included studies [13^-20] reported relevant data about the first outcome of interest (i.e., the comparison of TBS values between acromegalic patients and controls). Two studies [13^, 18] also reported relevant data about the second outcome of interest (i.e., the relationship, within acromegalic patients, between TBS values and fragility fractures); both of them, in particular, evaluated the cross-sectional relationship between TBS and prevalent fragility fractures; on the other hand, the longitudinal relationship between baseline TBS values and incident fragility fractures at follow-up was not assessed in any study.

Open multimedia modal

Characteristics of the Included Studies

Table 1 summarizes the basic study characteristics. With respect to the outcomes of interest for this meta-analysis, all included studies had a cross-sectional design [13^-20]. The acromegalic patient group was represented by subjects with newly diagnosed acromegaly in two studies [14^, 15] and by subjects with controlled/cured acromegaly in two studies [13^, 20], while the remaining four studies [16^-19] did not apply any restriction in terms of disease activity. The matching criteria adopted for the selection of the control group included age in seven studies [13^-16^, 18^-20], sex in six studies [13^-16^, 19^, 20], and BMI in six studies [14^-16^, 18^-20]; gonadal status was considered as an adjunctive matching criterion in one study [19], while menopausal state and smoking in another one [15].

Open multimedia modal

Comparison of TBS Values between Acromegalic Patients and Controls

As previously stated, eight studies [13^-20] reported relevant data to compare TBS values between acromegalic patients and controls. Overall, TBS values were significantly lower in acromegalic patients compared to controls (−0.089, 95% CI: [−0.111, −0.067], p < 0.01), with no significant between-study heterogeneity (I² = 35.78%, p = 0.17) (Fig. 2). These results were consistently confirmed also after stratifying the data according to the gonadal status of patients and controls, whenever possible (Fig. 3). At subgroup analysis, no differences emerged when subdividing the studies according to the selection criteria adopted by each in terms of acromegaly disease activity (p = 0.51) (Fig. 4). Among the 4 studies in which patients with any disease activity were enrolled [16^-19], only two [17^, 18] provided a head-to-head comparison of TBS values between patients with active and controlled/cured acromegaly, without highlighting any significant difference (p = 0.98) (online suppl. Fig. 1).

Open multimedia modal

Open multimedia modal

Open multimedia modal

As an ancillary analysis, in order to exclude possible biases arising from the specific criteria adopted for enrolment of the control group, data regarding the comparison between acromegalic patients and controls of BMD values at the lumbar spine, femoral neck, and total hip were also extracted from the included studies whenever available. The statistical pooling of the results did not highlight any significant difference in BMD values between acromegalic patients and controls, either at the lumbar spine (−0.019, 95% CI: [−0.046, 0.008], p = 0.16), femoral neck (0.012, 95% CI: [−0.013, 0.037], p = 0.35), or total hip (0.003, 95% CI: [−0.026, 0.031], p = 0.84) (online suppl. Fig. 2).

Relationship, within Acromegalic Patients, between TBS Values and Fracture Risk

Two studies [13^, 18] reported relevant data to evaluate the relationship, within acromegalic patients, between TBS values and prevalent fragility fractures in a cross-sectional design. The presence of fractures was systematically assessed, in both studies, by performing lateral spine imaging (either with a densitometer [18] or by conventional radiography [13]) in all enrolled patients. Overall, TBS values were significantly lower in acromegalic patients with vertebral fractures than in unfractured ones, with a pooled mean difference of −0.099 (95% CI: [−0.166, −0.032], p < 0.01) (Fig. 5).

Open multimedia modal

No studies, on the other hand, evaluated, in a longitudinal design, the relationship between TBS values at baseline and the occurrence of incident vertebral fractures at follow-up. One study [27] reported prospective data about the relationship between incident fractures and TBS percentage change during follow-up, displaying a greater reduction over time of TBS values in patients with incident vertebral fractures compared to unfractured ones; however, data about the relationship between incident fractures and TBS values at baseline were not available, and therefore, the study did not fit our eligibility criteria for inclusion in the meta-analysis.

Discussion

In this paper, we provided an updated systematic review and meta-analysis that specifically assessed the role of TBS as an index of bone structure quality and fracture risk in patients with acromegaly. Compared to a previous work [21], which was based on articles published until September 2020 and could not retrieve any significant results, additional research has been published and could be included in our analysis. This allowed for greater statistical power in assessing the difference in TBS values between acromegalic patients and controls and for finer and more detailed sub-analyses; in addition, for the first time, the relationship between TBS values and prevalent fragility fractures could also be assessed.

In recent years, the correlation between TBS values and fracture risk has already been demonstrated in various settings of primary and secondary osteoporosis [11^, 12]. In primary osteoporosis, TBS predicts both prevalent and incident fragility fractures [11]. In fact, in retrospective or cross-sectional studies, TBS has been shown to be consistently lower in subjects with an osteoporotic fracture than in those without [29^-33]. A similar result holds true also when prospectively assessing the risk of incident fractures, as TBS was proven to be at least as good as BMD in predicting hip fractures, vertebral fractures, and all fractures [34^-36]. In secondary osteoporosis, the role of TBS in assessing skeletal fragility and fracture risk seems to be even more prominent [12]. In fact, in these patients, the measurement of BMD often lacks sensitivity in predicting fracture risk [12]. For example, the increase in fracture risk associated with glucocorticoid exposure occurs before major bone loss can be measured by DXA [37]; this is also true for other causes of secondary osteoporosis, such as diabetes mellitus and asymptomatic primary hyperparathyroidism, in which bone mineralization is relatively well preserved [38^-41], and the increase in fracture risk is not clearly governed by the expected relationship with BMD [38^-41]. In these conditions, skeletal fragility is more related to subtler changes in bone micro-architecture [42], which can be undetectable when simply evaluating BMD; nevertheless, they may be more clearly unveiled by TBS, which has been shown to outperform DXA for fracture risk prediction in many of these clinical contexts [43^-45].

Within the setting of acromegaly, our meta-analysis highlighted the presence of lower TBS values in affected patients compared to unaffected controls. On the contrary, in the same studies, no overall differences in BMD values could be found between the two groups at any skeletal site; this latter outcome is in line with previous literature [5^, 21], in which BMD was already shown to be either unaffected (at lumbar spine) or even slightly increased (at femoral neck and total hip) in acromegalic patients with respect to non-acromegalic controls. From a pathophysiological point of view, the finding of lower TBS values in acromegalic patients is coherent with the available data about the impairment in trabecular bone micro-architecture that is typical of acromegalic osteopathy: histomorphometric analyses, in fact, demonstrated a reduced trabecular bone volume, increased trabecular separation, and impaired trabecular biomechanical competence in acromegalic patients compared to controls [46^, 47]; moreover, a reduced trabecular bone density and trabecular thickness were also displayed at high-resolution peripheral quantitative computed tomography (HR-pQCT), both at the distal radius and at the distal tibia [48^, 49].

Gonadal status is known to be an important factor affecting bone strength by directly influencing osteoblast and osteoclast function; in women and men, in fact, oestradiol and testosterone, respectively, help in maintaining a correct balance between bone formation and bone resorption [50^, 51]. Patients with acromegaly due to GH-secreting adenoma frequently present with pituitary hormone deficiencies, either due to compression of normal pituitary tissue by the adenoma or due to pituitary damage following surgery/radiotherapy, when performed [52]. Thus, when evaluating bone-related features in acromegalic patients, a possible interplay between GH dysregulation and gonadal status should always be considered. This point has been assessed by some of the studies included in this meta-analysis, though on relatively limited patient cohorts [14^, 16^, 19]. When pooling these data, notably, the impairment in TBS values in acromegalic patients was consistently confirmed both in eugonadal and hypogonadal patients. Thus, the detrimental role of GH hypersecretion on trabecular bone micro-architecture appeared to be evident irrespective of gonadal status.

Acromegaly disease activity has also been suggested as a possible determinant of bone fragility in acromegalic patients. Patients with active acromegaly have been shown to be at higher risk of vertebral fractures compared to those with a controlled/cured disease [5]. However, in a recent HR-pQCT study, Duan et al. [53] suggested that the negative impact of acromegaly on bone microstructure may still not be fully reversible. With respect to TBS values, our meta-analysis could not highlight any significant difference in the outcome measure according to acromegaly disease activity. The strength of this result is certainly limited by the low number of available studies; nevertheless, it further supports the hypothesis that – in acromegalic patients – the changes in bone micro-architecture induced by persistent GH excess might not be easily reverted by disease cure or control, except possibly in the very long term.

With respect to fracture risk, our results highlighted an association between TBS values and prevalent vertebral fractures; in fact, when pooling the data from eligible studies, TBS values were significantly lower in acromegalic patients with vertebral fractures than in those without. This finding is in line with the available evidence from various other settings of primary and secondary osteoporosis [11^, 12] and supports the use of TBS as a helpful index of skeletal fragility in acromegalic patients too. Of note, this is an even more relevant result in light of the fact that, as previously demonstrated, BMD is a poorly reliable measure of fracture risk in acromegalic disease [5^-8]; nevertheless, it has to be noted that, to date, no longitudinal data are still available about the relationship between baseline TBS values and the occurrence of fragility fractures at follow-up, which should be the gold standard criterion for the definite establishment of TBS as a reliable predictor of fracture risk.

Our analysis had some limitations. Firstly, the strength of the conclusions was limited by the low number of available studies; this was particularly true for the analysis that examined the relationship, within acromegalic patients, between TBS values and fracture risk, as only two studies provided sufficient data for the assessment of this outcome. Secondly, the quality was limited by that of the included studies; nevertheless, according to the AXIS tool, the risk of bias was generally moderate-to-low, which reassured about the likely small impact of this issue on the final results. Thirdly, patients’ inclusion criteria were partly different between studies, and this could be responsible for a certain degree of heterogeneity in the considered outcomes; however, heterogeneity is a common limitation of all meta-analyses, and appropriate statistical methods – such as the use of a random-effect model – were adopted to account for it. Fourthly, the comparisons of TBS values between groups were all based on crude differences, as derived by univariate analyses; thus, the possible interplay with other predictors could not be evaluated. Fifthly, no studies evaluated, in a longitudinal design, the relationship between TBS values at baseline and the occurrence of incident vertebral fractures at follow-up; thus, prospective data about the utility of TBS as a predictor for future fractures in acromegalic patients are still lacking.

In conclusion, our meta-analysis supports the notion that acromegalic patients have an impaired bone quality, described by lower TBS values with respect to controls; conversely, BMD did not seem to be significantly affected, in line with previous results. Moreover, at cross-sectional evaluations, TBS values were lower in acromegalic patients with prevalent vertebral fractures than in those without. However, prospective data evaluating the capacity of TBS values to predict the future occurrence of incident fractures are still lacking. Overall, thus, our findings suggest that TBS could be valuable in the assessment and management of skeletal fragility in acromegalic patients, especially in light of the poor information provided in this setting by BMD. Nevertheless, larger prospective studies are needed for a more detailed evaluation of TBS as a predictor for future fractures.

Statement of Ethics

An ethics statement is not applicable because this study is based exclusively on published literature.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This research did not receive any specific funding.

Author Contributions

Fabio Bioletto and Marco Barale contributed to work conceptualization, data collection, data analysis, and manuscript writing. Nunzia Prencipe, Alessandro Maria Berton, and Mirko Parasiliti-Caprino contributed to data interpretation and manuscript writing. Valentina Gasco, Ezio Ghigo, Massimo Procopio, and Silvia Grottoli contributed to data interpretation and final draft supervision. All authors approved the manuscript in its final form.

Data Availability Statement

All data generated or analysed during this study are included in this article and its online supplementary materials. Further enquiries can be directed to the corresponding author.