A case-control study based on the National Health and Nutrition Examination Survey to evaluate the effects of human papilloma virus on bone health in women
Curation statements for this article:-
Curated by eLife
eLife assessment
The study presents a useful finding on a potential link between two common public health issues, namely HPV infection and osteoporosis. The evidence supporting the claims of the authors is incomplete as a more robust statistical methodology would have strengthened the study. The work will be of interest to epidemiologists working on associations of infectious diseases with other health issues.
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
Abstract
Background
Globally, both human papillomavirus (HPV) infection and osteoporosis (OP) are more prevalent in women than in men. It remains unclear whether HPV has an impact on bone health.
Methods
This case-control study was based on data from the National Health and Nutrition Examination Survey (NHANES). Comparable datasets were created via the nearest neighbor propensity score matching (PSM) method (1:2). The Welch two-sample t test was used to analyze the association between HPV infection and bone mineral density (BMD). Restricted cubic spline (RCS) and Kendall’s tau-b tests were used to explore the effect of HPV infection type on BMD.
Results
BMDs in the legs and lumbar spine were lower in subjects infected with HPV than in uninfected subjects. RCS analysis showed that the larger the number of cooccurring HPV types in a woman, the lower the BMD was. In addition, four HPV types were negatively associated with leg BMD, and 14 HPV types were negatively associated with lumbar spine BMD. HPV types 53, 59, and 89 had effects on both leg and lumbar spine BMDs.
Conclusions
HPV infection was associated with a decrease in BMD. Appropriately designed trials can help determine whether interventions to prevent HPV infection will have a protective effect on BMD.
Funding
This research was not supported by any specific grant from any funding agency in the public, commercial or not-for-profit sector.
Article activity feed
-
eLife assessment
The study presents a useful finding on a potential link between two common public health issues, namely HPV infection and osteoporosis. The evidence supporting the claims of the authors is incomplete as a more robust statistical methodology would have strengthened the study. The work will be of interest to epidemiologists working on associations of infectious diseases with other health issues.
-
Reviewer #1 (Public Review):
The authors conducted a case-control study in the NHANES database and found that women who tested positive for HPV infection had lower bone mineral density (BMD) measures at the spine and at the hip. A major strength is the novelty of the association that they are reporting. Major weaknesses include not controlling for covariates that might account for the association between HPV and osteoporosis; unclear definition of the hip (described as "leg") BMD; and unclear methodology used for the propensity score matching and correlations. These weaknesses mean that it is unclear whether the authors' results support their conclusions. The impact of the work on the field and the utility of the methods and data to the community is therefore limited.
-
Reviewer #2 (Public Review):
To explore their dataset, the authors first identify all eligible women (n = 4673) in the database queried and use propensity score matching (PSM) to match group A (not infected by HPV) with group B (infected by HPV) for several covariates thought to affect bone mineral density (e.g.: age, smoking, alcohol). After PSM, no significant difference for selected covariates can be detected between the two groups.
Because they add matched their groups for relevant covariates possibly affecting bone mineral density, the authors then use Welch two-sample t-test to compare bone mineral densities of leg and lumbar spine between group A and group B, and detect significantly lower bone mineral densities for participants infected by HPV, group B. Here, the statistical approach chose by the author seems limited, and …
Reviewer #2 (Public Review):
To explore their dataset, the authors first identify all eligible women (n = 4673) in the database queried and use propensity score matching (PSM) to match group A (not infected by HPV) with group B (infected by HPV) for several covariates thought to affect bone mineral density (e.g.: age, smoking, alcohol). After PSM, no significant difference for selected covariates can be detected between the two groups.
Because they add matched their groups for relevant covariates possibly affecting bone mineral density, the authors then use Welch two-sample t-test to compare bone mineral densities of leg and lumbar spine between group A and group B, and detect significantly lower bone mineral densities for participants infected by HPV, group B. Here, the statistical approach chose by the author seems limited, and although PSM had been applied to match group earlier in the analysis pipeline, the reader could expect the statistical approach to be more robust, i.e. accounting for other covariates, like a linear mixed model.
Then, the authors analyse each HPV subtype independently and use Kendall's tau-b correlation test to estimate a correlation between a given HPV subtype and bone mineral density. To apply this test, the authors had to transform the bone mineral density to a binary variable, i.e. greater or equal to 1. Here again, the statistical approach does not control for any of the bone mineral density potentially affecting covariates. Also, the authors' study performed 32 Kendall's tau-b correlation tests and did not seem to correct for multiple testing.
Finally, the authors use the Restricted cubic spline model to establish a non-linear relationship between the number of infected HPV subtypes and bone mineral density.
The authors had set the aim to explore the association between HPV and bone mineral density. Unfortunately, due to possibly not high enough robustness of statistical approaches used in this manuscript, it does not seem sufficient to establish a clear association between HPV infection status and a lower bone mineral density. However, given the database the authors have created, it is believed that they have all the tools needed to pursue their aim.
-
