Statistics from Altmetric.com
Newton et al (Gut 2001;49:282–7) described a retrospective study on bone mineral density (BMD) in a large cohort of patients with primary biliary cirrhosis (PBC). The authors concluded that osteoporosis is not a specific complication of PBC, but certain weaknesses in the study design do not support this conclusion.
The authors did not include age and sex matched controls from the general population, or control groups with different types of liver disease.
A proper methodological design comparing osteoporosis in PBC and in a normal population should calculate the standardised incidence ratio of osteoporosis for the two cohorts by comparing the observed incidence versus the expected incidence. The calculation should include 95% confidence intervals according to exact Poisson limits.1
A logistic regression analysis including risk factors for osteoporosis (that is, age, menopausal status, smoking habits, alcohol consumption, etc) should have been performed.
The major drawback however is the lack of control populations with different types of liver disease. New data are emerging in the literature concerning this field. In particular, several recent studies (including one of our own)2–4 have demonstrated that PBC in itself does not exert a negative influence on BMD. Thus we agree with Newton et al that osteoporosis in PBC should be revisited. In fact, analysis of the literature enables the following conclusions to be drawn. There are several osteoporosis risk factors common to liver disease, aging processes, or genetic variability, as well as cholestasis related risk factors, that are obviously not specific for PBC (table 1). The pathogenesis of osteoporosis is multifactorial, increasing with advancing age, and influenced by genetic factors, and it may be that liver disease accelerates bone resorption through various mechanisms.
We read with interest the letter of Professor Floreani in which she agrees that it is timely to revisit the perceived dogma that patients with primary biliary cirrhosis (PBC) are predisposed to osteoporosis. It was pleasing to note that our findings compare favourably with work from her group. We would also like to take this opportunity to draw attention to two further studies which have been presented in abstract form since the submission of our manuscript, which confirm that there is no increase in prevalence of osteoporosis1 and no increase in fracture risk in PBC populations taken as a whole compared with appropriately matched normal controls.2 A further study has also been published recently3 describing bone mineral density in a selected series of patients with PBC in whom the proportion with osteoporosis (defined by T score) was comparable with that seen in our series. We were pleased to note that in this prospective series other risk factors for osteoporosis were examined. They concur with our finding that increased age is an independent risk factor, although they do not present mean Z score data (bone mineral density data controlled for both sex and age). Their group, we would argue, was younger and had more severe disease than patients in our series, whom we would regard as more representative of the whole PBC population.
In our study we demonstrated that although 85/272 (31%) patients had osteoporosis, as defined by the WHO (T <−2.5) at the time of their first DEXA scan, mean Z score at the neck of femur was −0.1 and at the lumbar spine 0.1. As Z scores represent bone mineral density compared with an age and sex matched population, this suggests that the prevalence of osteoporosis seen in PBC is a reflection of the fact that this is predominantly a disease of postmenopausal women who show a generalised increased prevalence of osteoporosis. The use of Z scores implicitly controls our data for age and sex norms.
We agree with Professor Floreani that readdressing the question of osteoporosis prevalence in other chronic liver diseases (both cholestatic and non-cholestatic) would be of interest but we feel that this is not relevant to the current study.
Given the very real problems experienced by some PBC patients as a result of osteoporotic fracture (particularly in the early post-transplant period), further study of the aetiology is appropriate (although the retrospective nature of the current study makes the suggested logistic regression analysis inappropriate). The message that we (and more recently others) have been communicating is that the search for risk factors for osteoporosis should not be focused on liver disease specific factors but could more usefully be directed at more generalised population risk factors.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.