Reply
The DOI matches its Crossref record; 7 of the 7 references it cites were found in a registry or an index.
Abstract
Sir—We thank Lambert et al. [1] for comparing and contrasting their recent analysis of tuberculosis (TB) recurrences and concurrent HIV infection [2] with our own [3]. It is true that our meta-analysis did not include every study of TB recurrences in which HIV was not a consideration. We included a total of 41 studies of HIV-uninfected patients and found that the relative risks of recurrence associated with HIV-positivity and with the duration of rifampin treatment did not change with the addition of the last few studies of HIV-uninfected patients that we identified. Rather, the results were determined by the more-limited set of data on HIV-positive patients, consisting of 21 studies. Of the additional studies suggested by Lambert et al. [1], 3 studies [4–6] meet our eligibility criteria and could have been included. These 3 studies each reported on HIV-negative patients. If we add these studies to the meta-analysis, the results remain essentially unchanged, with estimated recurrence rates among HIV-uninfected patients of 1.5 per 100 person-years for recipients of ⩾7 months of rifampin therapy, 2.2 cases per 100 person-years for recipients of 5–6 months of rifampin therapy, and 3.2 cases per 100 person-years for recipients of 2–3 months of rifampin therapy (P = .001). Relative risks of recurrence associated with HIV infection at each of these 3 treatment durations would become 2.1, 1.9, and 4.2, respectively (interaction between HIV-infection status and treatment duration, P = .014). The study by Gonzalez-Montaner et al. [7] did not report the exact duration of follow-up (apart from the statement that follow-up continued for “up to 2 years”), which is an important parameter in standardizing recurrence events across studies in the meta-analysis (as a rate per 100 person-years), and, therefore, that study could not be included. The study by Castelo et al. [8] was not eligible because it included patients regardless of their compliance to therapy and because follow-up duration was not precisely reported. The study by Glynn et al. [9] was potentially very informative because it described recurrences according to the patient's HIV status. However, rates of recurrence could not be calculated according to HIV status because the follow-up duration was reported only as an aggregate for both HIV-infected and HIV-uninfected patients, whereas the duration of follow-up was probably shorter among HIV-infected patients because of their higher mortality rate. Moreover, this study included some children among the patients. Unexplained residual variation between studies (the “heterogeneity” referred to by Lambert et al. [1]) is indeed an important issue in meta-analysis. It is inevitable that some of the variation between studies will not be accounted for by a few potentially explanatory variables. However, the statistically significant effect of the covariate “follow-up duration” in our study would have accounted, at least in part, for the different effects of HIV infection and of the duration of rifampin therapy across studies of relapse (which is concentrated in the early post-therapy period) and reinfection (which is distributed more equally over the post-therapy period). Our decision to focus on overall recurrence rates, rather than on relapses and reinfections, was dictated by the availability and quality of the data. Far more data are available on overall recurrences than on relapses and reinfections. Several biases are apparent in the definition of relapse versus that of reinfection, and those biases preclude drawing solid quantitative conclusions on the basis of these outcomes, as is clear from both recent reviews [2, 3]. We agree with Lambert et al. [1] that separate analyses of the determinants of rates of relapse and reinfection would be clinically and epidemiologically meaningful— and possibly simpler to interpret than our results on overall recurrences— once such data become available in sufficient quantity and quality from future restriction fragment–length polymorphism studies. Nevertheless, the clear influence we observed of the duration of both follow-up and rifampin treatment, which mainly affect relapses, on overall recurrence rates indicates that the majority of recurrences in HIV-positive and HIV-negative patients must be relapses. Combining this evidence with the clear increase in recurrence rate caused by HIV infection and with the stronger influence of the duration of treatment in HIV-positive patients—which can only be explained if HIV infection increases, not only the risk of reinfection, but also of relapse—led us to conclude that HIV infection must increase the risk both of relapse and of reinfection. Although the comparative strengths of these 2 effects cannot be quantified at present, we consider this outcome to be highly relevant clinically, epidemiologically, and for health policy. The increased risks of reinfection and relapse in HIV-positive patients with TB imply that, in countries where H
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Citations
Cites
Bulletin of the World Health Organization
C., N. H. (1955) · The British Journal of Sociology · doi:10.2307/587237 · Cited by 3,601
Recurrence in tuberculosis: relapse or reinfection?
Lambert, M.-L., Hasker, E., Deun, A. V., Roberfroid, D., Boelaert, M., & Van der Stuyft, P. (2003) · The Lancet Infectious Diseases · doi:10.1016/s1473-3099(03)00607-8 · Cited by 225
Korenromp, E. L., Scano, F., Williams, B. G., Dye, C., & Nunn, P. P. (2003) · Clinical Infectious Diseases · doi:10.1086/375220 · Cited by 147
Self-citationLj, G. M., Natal, S., Yongchaiyud, P., Olliaro, P., Abbate, E. H., Mosca, C., Casado, G., Lonardo, M. D., Gerhart-Filho, G., Betjel, I., et al. (1994) · Tubercle and Lung Disease · doi:10.1016/0962-8479(94)90079-5 · Cited by 120
Rifampin and Recurrence of Tuberculosis among Patients Infected with HIV
Lambert, M.-L., Hasker, E. C., Deun, A. V., et al. (2003) · Clinical Infectious Diseases · doi:10.1086/379616 · Cited by 42
COMPARISON OF DAILY AND TWICE-WEEKLY REGIMENS TO TREAT PULMONARY TUBERCULOSIS
Castelo, A., Goihman, S., Dalboni, M. A., et al. (1989) · The Lancet · doi:10.1016/s0140-6736(89)91788-1 · Cited by 38
Martinez, F. D. G., & Hernandez, H. M. (1996) · Tubercle and Lung Disease · doi:10.1016/s0962-8479(96)90115-x · Cited by 3
Research works printed but not linked: 3
- 1997, The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease, 1, 509
- Cao, 1998, The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease, 2, 360
- Harries, 2003, The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease, 7, 616
Cited by
Still being checked.
Metadata
- Authors
- Korenromp, E. L., Scano, F., Williams, B. G., Dye, C., & Nunn, P.
- Affiliations
- World Health Organization (CH)
- Corresponding author
- Scano, F.
- Venue
- Clinical Infectious Diseases
- Publisher
- Oxford University Press (OUP)
- Published
- 2003-12-15
- Volume
- 37
- Issue
- 12
- Pages
- 1720-1722
- ISSN
- 1058-4838 (ISSN-L)
- 1537-6591
- OpenAlex
- W4233315041
- Cited by
- 0
- References
- 7
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Clinical Infectious Diseases (publishedVersion)
- Topics
- Health Sciences › Medicine › Infectious Diseases › Tuberculosis Research and Epidemiology
- Health Sciences › Medicine › Infectious Diseases › Infectious Diseases and Tuberculosis
- Health Sciences › Medicine › Public Health, Environmental and Occupational Health › HIV, TB, and STIs Epidemiology
- Keywords
- tuberculosis relapse, HIV coinfection, rifampicin, tuberculosis prevention, treatment duration, HIV-negative patients, tuberculosis mortality, tuberculosis incidence
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.