Loading...
Please wait, while we are loading the content...
Similar Documents
Author's response to reviews Title: HIV Associated High-Risk HPV Infection among Nigerian Women Authors:
| Content Provider | Semantic Scholar |
|---|---|
| Author | Akarolo-Anthony, Sally N. Dareng, Eileen O. Famooto, Ayotunde O. Almujtaba, Maryam Olaniyan, Olayinka B. |
| Copyright Year | 2013 |
| Abstract | 1) The first sentence of the abstract “the incidence of cervical cancer has rem ined stable in HIV+ women is misleading and an over-simplification”. We only have good data from western countries with high-levels of screening. We know little about what is happening in area s such as Africa where cART is being rolled out far quicker than screening. The background of the abstract has been changed. It now reads “In developed countries, the incidence of cervical cancer has remained stable in HIV+ women but the preval ence nd multiplicity of high risk HPV (hrHPV) infection, a necessary cause of cerv i al cancer, appears different comparing HIV+ to HIVwomen. Little is known about HIV and HPV co-inf ection in Africa” 2) All HPV infections are more prevalent in HIV+ women, and increase in multiplici y of infection is a consequence, so it does make sense to focus only on multiplicity as a spec ial outcome per se. Our paper focuses on the prevalence of hrHPV types in the presence of HIV infecti on. We did not focus on only multiplicity of infection as a special outcome. 3) The second and third sentences of the conclusion are misinterpretations, as there is no information in this paper on the types that cause pre-cancer and cancer. Infact , it is known that the high prevalence of types in women without lesions (even in HIV+ women – see Clifford et al, AIDS, 2006), are not representative of those that cause cancer. This is because cer t in typ s, namely 16 and 18 are far more carcinogenic than others. See paper comparing HIV+ and HIVcancers in Kenya and South Africa, Int J Cancer, first author de Vuyst. Persistent hrHPV infections have been established as a necessary cause of ervical pre-cancer and cancer. Our conclusions are based on the prevalent hrHPV types in the study populat ion. BACKGROUND 4) I am not aware of any literature criticising the IARC classifi cation of high-risk types, and certainly not the cited paper. Li, N., et al., Human papillomavirus type distribution in 30,848 invasive cervical cancers worldwide: Variation by geographical region, histological type and year of publication. International Journal of Cancer, 2011. 128(4): p. 927-935. Provide evidence criticizing the IARC classification. 5) There quite a lot of data published from Africa on HPV types in HIV+ women. The meta analysis Clifford et al, AIDS, 2006 is a good start, and there have been many more from the region since. These papers were reviewed and up to 7 were referenced. 6) The claim that cervical cancer incidence is on average higher in Western t han Eastern Africa is tenuous. The GLOBOCAN data cited are almost entirely modelled based on very f ew data from a couple of cancer registries in select countries across the continent. In any case, even if such a difference is true , the hypothesis that this is due to HPV type differences is not answered by this study, that is more a c mparison of HIVversus HIV+. The West versus East Africa issue should be entirely dropped. Reports from several studies suggest that regional differences in cervical can er incidence may be due to differences in the prevalence and types of hrHPV between the regions Li. et . al. 2011. Compared to results from East African studies, our findings support the hypothesis that the prevalence and types of hrHPV in West Africa differ and this may explain the di fferent incidence rates of cervical cancer in these populations. RESULTS 7) The prevalence of HR-type (single or multiple) positive women that were HI V+ and HIVis purely driven by the study recruitment source. This proportion should be reported the other w ay round, i.e. the proportion of HIV+ versus HIVwomen that were HR HPV pos. These proportions should be reported in Table 2 (see below) The proportions have been changed. They are now reported according to HIV status as suggested. The proportions are shown in Table 2. 8) In Table 2, overall HPV prevalence and multiple HPV prevalence can be shown. Ra ther th n p values, which do not show the direction of the difference, prevalence ratios with 95% CI s can be shown. The p-values show the reader that HPV prevalence is different between HIV+ and H IVwomen. Prevalence ratios and 95% confidence intervals are shown in Table 3. To make this table different to Figure 1, prevalences and prevalence ratios for i ndividual types could be shown among HPV-positive women only, as in Clifford et al, AIDS, 2006. We have taken Clifford et al, AIDS, 2006 into account. 9) As it stands, Figure 1 is somewhat redundant with Table 2 and could be dropped. Table 2 shows that the statistical relationship of hrHPV by HIV status. Figur e 1 helps the reader easily ascertain the most prevalent hrHPV types by HIV status. Many r e ders find figures more intuitive. 10) Table 3: data are too sparse to show two different models for single and multiple infection, and anyway would not be expected to be different. Hence present one column only for any HRHPV infection. The meaning of the PRs for the risk factors (age, sexual partners , marital status, education and age at first sexual intercourse) are meaningless unless the r ader knows what category is being compared to what. The models are comparing the risk factors for HPV by HIV categories. H IV negative women are the reference category. A legend has been added to the table to indicate this. The confi dence intervals (CI) are presented to help the reader understand the direction of the risk a nd the effect of the sample size on the effect estimates. Since the effect estimates p-values were significant and the CI did were not too wide across models, the sample size was sufficient to show a difference between the groups compared and the data are not stretched too thin ac r ss strata. 11) I am surprised that no HPV35 infections were found in HIV-neg women, as this is a ty pe that is known to be commonly found in HIV-neg women in Africa and Nigeria. The data should be checked for a technical problem. The data has been checked, none of the HIVwomen had HPV35. DISCUSSION 12) Be clearer about when discussing different HPV types in Africa versus the res of the world, and when talking about HIV+ versus HIV-ve in Africa. In general, when citi g other papers, it is important not to mix HPV type distribution in the general population with that in severe lesions or cancer, which have different meanings, for the reasons explained above. The discussion has been modified to make this clear. 13) Drop discussion of West versus East Africa, for reasons stated above. Our results support this hypothesis. However, this is not emphasized in our conclusions. 14) Longitudinal studies of HPV infection are largely impossible due to the require ment to offer treatment and the enormity of the duration and sample size required. Rather, crossse tional comparisons of HPV type distribution across different lesion grades up to cancer c offer a similar, but more efficient, reply to the question. Longitudinal studies of HPV infection are difficult to conduct but not impossible. The y provide the best evidence to investigate HPV genotype-specific risks for cervica l prec ncer and cancer outcomes. 15) The last sentence about stable cervical cancer rates is not based on evidenc e – w badly need studies to show what is happening to cancer in HIV+ women in Africa in the cART era. T h re are some data (e.g. proportions of cervical cancers that are HIV-positive , ee papers from Kenya and South Africa), that suggest an increasing epidemic of cervical cancer as survival is improved with cART. The sentence on stable cancer rates has been changed. The suggestion on studies exam ining cancer among HIV+ women in Africa has been included. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | https://static-content.springer.com/openpeerreview/art:10.1186%2F1471-2334-13-521/12879_2013_2720_AuthorComment_V2.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
