Response to Letter to the Editor: “Association of Vitamin D Deficiency With Peripheral Arterial Disease: A Meta-Analysis of Literature Studies” Free

We read with interest the letter by Sanguankeo and Upala (1), and we carefully evaluated comments to our meta-analysis evaluating the association of vitamin D levels with peripheral artery disease (PAD) (2).

We agree with authors that it is important to clarify that type of study design when performing the Newcastle-Ottawa score (NOS) (3). This is the reason why in the online appendix (2), readers can find two separate NOS tables: one for prospective studies and one for retrospective ones. Although, by a general point of view, we also agree that combining studies with a different design might be a source of heterogeneity, it is noteworthy that in our meta-analysis heterogeneity was significant for studies evaluating levels of vitamin D in subjects with PAD as compared with controls without PAD (I²: 86.5%) but no significant heterogeneity was found when analyzing prevalence of PAD in subjects with vitamin D insufficiency (I²: 0%) and in those with vitamin D deficiency (I²: 7.65%) as compared with subjects with normal levels of vitamin D. Nevertheless, in all cases we performed sensitivity analyses only including prospective studies and we entirely confirmed results. Furthermore, with the aim to address any potential source of heterogeneity, we performed specific meta-regression analyses and sensitivity analyses only including high quality studies and studies specifically defining PAD according to ankle brachial index.

It is also relevant to highlight that although absolute risk and attributable risk should be calculated using incidence rates (4), some methodological studies suggested that these values can be evaluated also in case-control studies (5). However, it is important to remember that, in the latter case, they represent a prevalence and not risk. Accordingly, both in results section and in the discussion, we always used terms “prevalence” and “association” and avoided the term “risk.”

As to the issue about differences between OR and relative risk (RR), the Cochrane handbook (6) clearly reports that “risk” describes the probability with which a health outcome will occur, whereas “odds” is the ratio of the probability that a particular event will occur to the probability that it will not occur. In our meta-analysis, we presented all results as ORs, also considering that the difference between odds and risk is limited when an event is rare (7). The reason is that small numbers for the outcome will not affect the calculations very much because they exert little influence on the denominators in the RR calculation (7). However, to fully address the issue raised by Sanguankeo and Upala (1), we have now used the formula provided by the Cochrane handbook (8) to convert ORs in RRs and we found RRs similar to ORs reported in our meta-analysis (2). In detail, for patients with vitamin D insufficiency, the reported OR of peripheral artery disease was 1.09, and the calculated RR was 1.10. Similarly, for patients with vitamin D deficiency, the reported OR of peripheral artery disease was 1.48, and the calculated RR was 1.50.

Overall, differences among studies included in a meta-analysis are frequent. By a methodological point of view, it is important to implement all statistical strategies to assess potential sources of heterogeneity. Thus, it is important to always read with great attention both primary analyses and subgroup analyses to have more complete information from a meta-analysis.

Abbreviations:

 
• NOS

  Newcastle-Ottawa score

 
• PAD

  peripheral artery disease

 
• RR

  relative risk

Acknowledgments

Disclosure Summary: The authors have nothing to disclose.

References