Sir, we read with considerable interest the paper by Malden et al.1 The authors report that a dental extraction can increase this risk of ONJ by a factor of up to seven. However, they do not comment on two studies published last year which confirmed the association of dental extractions and ONJ development.3,4 The first study from our institution confirmed an at-least 16-fold increased risk for ONJ following dental extraction.4 The other, coming from the MD Anderson Cancer Center, demonstrated a 10 to 53-fold risk for ONJ following dental extraction.3

The authors propose that whenever possible, patients should be encouraged and counselled to stop smoking. Despite the well-known association of smoking with periodontal disease, based on existing evidence it may not be appropriate to recommend quitting smoking to reduce risk for ONJ. In the study from our institution, we were not able to detect an increased risk for ONJ among smokers.4 The latter study encompasses American Society of Clinical Oncology (ASCO) Level III evidence.

The authors propose that extractions in cases where discontinuation of bisphosphonates has been instigated for 12 months or more would be expected to carry a reduced risk of ONJ. In patients who are going to experience surgical treatment for ONJ, discontinuation of BP is still a matter of debate. BP are reported to have an up to ten years long clearance time from calcified tissue.5,6 Some studies reported that besides having discontinued BP for at least six months prior to surgical treatment for ONJ, no obvious outcome improvement was recorded6 whilst others concluded that withdrawal is not recommended.5 A recent study included discontinuation in the treatment protocol, however, the authors could not answer whether discontinuation had a positive effect on the outcome.7 In a recent manuscript we describe a series of molecular mechanisms implicated in BP interaction with hard and soft tissues, which are thought to require minimal concentration of BP in the extra-cellular fluid,8 thereby suggesting that withdrawal of BP could hardly have some positive effect on outcome. Based on these data we have revised treatment protocols in our institution to exclude BP withdrawal and possibly prevent systemic disease deterioration following withdrawal, at least until further evidence is presented. Therefore, the authors' expectation that discontinuing BP could reduce risk for ONJ development, may not be supported by the existing evidence.

Malden et al. suggest that when possible, periodontal issues could be addressed prior to any extractions.1 Recent evidence by Estilo et al.9 and also unpublished data from our institution support this argument. Oral hygiene and periodontal disease were not found to be associated with increased risk for ONJ development and we therefore suggest that non-surgical periodontal treatment may be a safe modality.

N. Malden, C. Beltes and V. Lopes respond: We thank Drs Kyrgidis and Vahtsevanos for their constructive criticism.

One of our main aims with the paper was to support general practitioners in the provision of dental treatment (particularly tooth extractions) in the low risk bisphosphonate group. We accept the high incidence of reported extraction associated BONJ cases from Greece but observe that these have almost exclusively been reported from the high risk (malignancy) group. Compared to a dental extraction a smoking habit may be a relatively weak risk factor for development of BONJ. The study referenced by Kyrgidis et al. (Ref 4 ) included 20 subjects with ONJ with 40 controls and may not include an adequate subject number to demonstrate a possible link between BONJ and smoking (under-powered study). In consideration of the association of smoking and dry socket a recent study10 looked at >800 extractions in 469 subjects and was convincing in establishing a link with dry socket development.

More powerful studies looking at smoking and BONJ will no doubt follow in time. In the UK clinicians are under an obligation to counsel patients regarding smoking cessation and we would consider it a part of the informed consent process prior to any surgical procedure.

Our comments on drug discontinuation were primarily aimed at the low risk (alendronic acid) group and we apologise for any ambiguity. Ten years clearance time? Once bisphosphonates have been administered it is unlikely that they will ever be completely cleared from the skeleton, but this is of no relevance. The following paper which we referenced in our original article is a multi-centre double blind randomised controlled trial; it looked at the effects of continuing or stopping alendronic acid after five years of administration. 11 It demonstrated a gradual but immediate sustained increase in the level of bone turn-over markers from the time of discontinuation. At five years discontinuation of alendronate, bone turn-over had returned to close to that of pre-treatment levels (ten years previously). Now we accept that these marker levels are giving a general picture of skeletal turnover and may not be representative of what is happening to the jaws. The mandible and maxilla may have been preferentially and uniquely affected and may not be sharing in a recovery of bone turnover. It was interesting, however, that the FLEX trial did not show that the recovery of bone turnover was accompanied by a significant decline in bone mineral density, BMD or an increase in low impact fracture risk. In such cases the benefit to the jaws of drug discontinuation could be judged to outweigh the risks associated with systemic disease deterioration. However, in the high risk (malignancy group) the potential benefit of drug discontinuation to the jaws will generally be out-weighed by the risks of systemic disease deterioration. Case by case multidisciplinary assessment may be prudent.

We consider that these arguments highlight the need for us to clearly distinguish between the low and high risk groups when discussing management protocols. In fact one could argue that they should be considered as separate disease entities. The more challenging group for us at the moment, as mentioned in our paper, is the intermediate cases who are often receiving steroids and possibly cytotoxic drugs for non-malignant conditions.

We await further publications from Dr Kyrgidis's unit with interest.