Dear Editor,
I read with interest the study by Martinez-Zammora et al. (1), which reported an increased incidence of thrombotic events in patients with antiphospholipid syndrome with recurrent spontaneous abortion (APS-RSA
group) in comparison with 3 different control groups. Although their conclusions seem to be valid, some comments may be appropriate.
This study is labeled as 'case-control', but it is actually a retrospective cohort study. Case-control studies have a different design, where researchers select cases and controls based on the presence or absence of the outcome, respectively, and look back in time for
disproportionate exposures (2). Cohort studies follow groups defined by different exposures and observe the incidence of the outcomes. In the retrospective cohort design, the researcher starts the study at the time
follow-up has already been completed (3). This kind of mislabeling of study design has been frequently observed in medical journals, and may cause confusion among readers (2).
Given that this is a cohort study, the relative risk (RR) should have been calculated instead of the odds ratio. Odds ratios frequently generate larger differences between exposed and non-exposed groups (especially when
dealing with non-rare events), therefore overestimating the observed effect (4). However, considering the longer follow-up time of the APS-RSA group in comparison to the tRSA and uRSA groups, perhaps the incidence rate ratio (5) would be more adequate than the RR to compare de risk of
thrombosis in this study.
The authors should have made the exclusion criteria (like history of thrombotic events previously to the beginning of the follow-up period) clearer in methods. Also, the plotting of the Kaplan-Meier curves seems to
be inaccurate, since the curves that represent cumulative incidences in the tRSA and uRSA groups should have ended when the patients with longer permanence in the cohort were censored or had presented an event (6). In the abstract, the reporting of "12 year cumulative thrombotic incidence rate" gives the wrong idea that the mean follow-up time was 12 years.
Despite the problems of analysis and reporting, this study has adequate internal validity and problably external validity (albeit only Spanish Caucasian individuals were included), representing an advance in
the knowledge about the antiphospholipid syndrome.
References
1. Martinez-Zamora MA, Peralta S, Creus M, et al. Risk of thromboembolic events after recurrent spontaneous abortion in antiphospholipid syndrome: a case-control study. Ann Rheum Dis 2011 [Epub ahead of print].
2. Grimes DA. "Case-control" confusion: mislabeled reports in obstetrics and gynecology journals. Obstet Gynecol 2009;114:1284-6.
3. Euser AM, Zoccali C, Jager KJ, et al. Cohort studies: prospective versus retrospective. Nephron Clin Pract 2009;113:214-7.
4. Case LD, Kimmick G, Paskett ED, et al. Interpreting measures of treatment effect in cancer clinical trials. Oncologist 2002;7:181-7.
5. Pearce N. Effect measures in prevalence studies. Environ Health Perspect 2004;112:1047-50.
6. Clark TG, Bradburn MJ, Love SB, et al. Survival analysis part I: basic concepts and first analyses. Br J Cancer 2003;89:232-8.
Conflict of Interest:
None declared
Dear Editor,
I read with interest the study by Martinez-Zammora et al. (1), which reported an increased incidence of thrombotic events in patients with antiphospholipid syndrome with recurrent spontaneous abortion (APS-RSA group) in comparison with 3 different control groups. Although their conclusions seem to be valid, some comments may be appropriate.
This study is labeled as 'case-control', but it is actually a retrospective cohort study. Case-control studies have a different design, where researchers select cases and controls based on the presence or absence of the outcome, respectively, and look back in time for disproportionate exposures (2). Cohort studies follow groups defined by different exposures and observe the incidence of the outcomes. In the retrospective cohort design, the researcher starts the study at the time follow-up has already been completed (3). This kind of mislabeling of study design has been frequently observed in medical journals, and may cause confusion among readers (2).
Given that this is a cohort study, the relative risk (RR) should have been calculated instead of the odds ratio. Odds ratios frequently generate larger differences between exposed and non-exposed groups (especially when dealing with non-rare events), therefore overestimating the observed effect (4). However, considering the longer follow-up time of the APS-RSA group in comparison to the tRSA and uRSA groups, perhaps the incidence rate ratio (5) would be more adequate than the RR to compare de risk of thrombosis in this study.
The authors should have made the exclusion criteria (like history of thrombotic events previously to the beginning of the follow-up period) clearer in methods. Also, the plotting of the Kaplan-Meier curves seems to be inaccurate, since the curves that represent cumulative incidences in the tRSA and uRSA groups should have ended when the patients with longer permanence in the cohort were censored or had presented an event (6). In the abstract, the reporting of "12 year cumulative thrombotic incidence rate" gives the wrong idea that the mean follow-up time was 12 years.
Despite the problems of analysis and reporting, this study has adequate internal validity and problably external validity (albeit only Spanish Caucasian individuals were included), representing an advance in the knowledge about the antiphospholipid syndrome.
References
1. Martinez-Zamora MA, Peralta S, Creus M, et al. Risk of thromboembolic events after recurrent spontaneous abortion in antiphospholipid syndrome: a case-control study. Ann Rheum Dis 2011 [Epub ahead of print].
2. Grimes DA. "Case-control" confusion: mislabeled reports in obstetrics and gynecology journals. Obstet Gynecol 2009;114:1284-6.
3. Euser AM, Zoccali C, Jager KJ, et al. Cohort studies: prospective versus retrospective. Nephron Clin Pract 2009;113:214-7.
4. Case LD, Kimmick G, Paskett ED, et al. Interpreting measures of treatment effect in cancer clinical trials. Oncologist 2002;7:181-7.
5. Pearce N. Effect measures in prevalence studies. Environ Health Perspect 2004;112:1047-50.
6. Clark TG, Bradburn MJ, Love SB, et al. Survival analysis part I: basic concepts and first analyses. Br J Cancer 2003;89:232-8.
Conflict of Interest:
None declared