Chlamydia antibody titres

Ben W.J. Mol1, Fulco van der Veen and Patrick M.M. Bossuyt

Department of Clinical Epidemiology & Biostatistics and Centre for Reproductive Medicine, Academic Medical Centre, PO Box 22700, 1105 AZ Amsterdam, The Netherlands

Land et al. (1998) recently reported on the use Chlamydia antibody testing in subfertile patients. The authors state that they compare Chlamydia antibody titres (CAT) and findings at laparoscopy in consecutive patients that attended their fertility clinic, and recommend a cut-off level for CAT of 1:32 or 1:64, depending on the clinical context.

Since in their study the decision for laparospcopy was partially based on the result of the CAT, the series are not as consecutive as the authors want us to believe. The fact that patients with a CAT <8 were much less likely to undergo laparoscopy introduces verification bias, which not only is likely to interfere with the shape of the constructed receiver operating characteristic curves, but also has impact on the interpretation of cut-off levels (Begg and Greenes, 1983Go).

A second point of concern is that the authors repeatedly dichotomize the CAT in a positive and a negative result, thereby ignoring the option of calculating likelihood ratios for each level of the test. The latter is considered to be one of the major advantages of the use of such test parameters (Sackett et al., 1991Go). We recalculated likelihood ratios for several levels of the CAT from the data provided by Land et al. (1998) using the most severe definition of tubal disease, i.e. extensive periadnexal adhesions and/or occlusion of both tubes (Table IGo).


View this table:
[in this window]
[in a new window]
 
Table I. Likelihood ratios (LR) as derived for different levels of Chlamydia antibody titres (CAT)
 
The clinical application of this recalculation is demonstrated in Table IIGo, in which likelihood ratios of different CAT levels are combined with different pre-test probabilities for tubal pathology to calculate post-test probabilities, assuming mutual independence between patient characteristics and the CAT (Sackett et al., 1991Go). Pre-test probabilities for tubal pathology are obtained from a cohort of 359 patients who underwent hysterosalpingography (HSG) in which tubal pathology is defined as double-sided tubal pathology at HSG (Mol et al., 1997Go). Whereas Land et al. (1998) propose a single cut-off level, Table IIGo shows that the use of various levels of test-results provides a useful differentiation as compared to dichotomising the test result. As stated by Land et al. (1998) differences in clinical context such as female age or duration of subfertility, as well as patient's preferences can subsequently be taken into account to decide at which post-test probability hysterosalpingography or laparoscopy should be performed.


View this table:
[in this window]
[in a new window]
 
Table II. Calculation of post-test probabilities (%) for tubal pathology by combining pre-test probabilities and the likelihood ratios calculated in Table IGo
 

Notes

1 To whom correspondence should be addressed Back

References

Begg, C.B. and Greenes, R.A. (1983) Assessment of diagnostic tests when disease verification is subject to selection bias. Biometrics, 39, 207–215.[ISI][Medline]

Land, J.A., Evers, J.L.H. and Goossens, V.J. (1998) How to use Chlamydia antibody testing in subfertility patients. Hum. Reprod., 13, 1094–1098.[Abstract]

Mol, B.W.J., Swart, P., Bossuyt, P.M.M., Van der Veen, F. (1997) Is hysterosalpingography an important tool in predicting fertility outcome? Fertil. Steril., 67, 663–669.[ISI][Medline]

Sackett, D.L., Haynes, R.B., Guyatt, G.H. and Tugwell, P. (1991) Clinical Epidemiology: A Basic Science for Clinical Medicine. Little Brown and Company, Boston, Toronto, London, UK.


 
Jolande A. Land1 and Johannes L.H. Evers

Department of Obstetrics and Gynaecology, Academisch Ziekenhuis Maastricht

Valère J. Goossens

Department of Medical Microbiology, Academisch Ziekenhuis Maastricht PO Box 5800, 6202 AZ Maastricht, The Netherlands

Dear Sir,

We note the work of Mol et al. (Mol and Bossuyt, 1995Go; Mol et al., 1996aGo,bGo,cGo; 1997aGo,bGo; Mol and Van der Veen, 1997Go) and read with interest their letter to the editor regarding our recent Chlamydia manuscript (Land et al., 1998Go). We appreciate their comments, and agree that verification and selection bias are hard to prevent in clinical studies, unless one is prepared to perform the complete fertility investigation of a patient on a single day. Another example for this selection bias can be found in the second part of their present letter (Table II), where the authors refer to their own previous study (Mol et al., 1997cGo) in which they failed to do an hysterosalpingography (HSG) in all patients at the very moment they entered their clinics but still made inferences regarding the predictive value of HSG.

Regarding their second concern: research is data reduction, and we therefore would like to leave it up to the reader whether he prefers our dichotomized single cut-off levels or the individual likelihood ratios per test result, as proposed in their present letter to the editor.

1 To whom correspondence should be addressed Back

References

Land, J.A., Evers, J.L.H. and Goossens, V.J. (1998) How to use Chlamydia antibody testing in subfertility patients. Hum. Reprod., 13, 1094–1098.[Abstract]

Mol, B.W.J. and Bossuyt, P.M.M. (1995) Surgical treatment of ectopic pregnancy. [Letter to the editor.] Lancet, 346, 638–639.[ISI][Medline]

Mol, B.W.J. and Van der Veen, F. (1997) A study of ruptured tubal ectopic pregnancy. [Letter to the editor.] Obstet. Gynecol., 90, 866–867.[Free Full Text]

Mol, B.W.J., Pajkrt, E., Van Lith, J.J. and Bilardo, C.M. (1996a) Screening for fetal trisomies by maternal age and fetal nuchal translucency thickness at 10 to 14 weeks of gestation. [Letter to the editor.] Br. J. Obstet. Gynaecol., 103, 1051–1052.[ISI][Medline]

Mol, B.W.J., Van der Veen, F., Redekop, K. and Bossuyt, P.M.M. (1996b) The value of diagnostic tests in infertility. [Letter to the editor.] Ned. Tijdschr. Geneeskd., 140, 279–280.

Mol, B.W.J., Van der Veen, F. and Bossuyt, P.M.M. (1996c) Time to pregnancy – after ectopic. [Letter to the editor.] Fertil. Steril., 66, 172–173.[Medline]

Mol, B.W.J., Hajenius, P.J., Ankum, W.M. et al. (1997a) Comparative costs of methotrexate and laparoscopic surgery. [Letter to the editor.] Hum. Reprod., 12, 1603–1604.[ISI][Medline]

Mol, B.W.J., Van der Veen, F., Lijmer, J. and Bossuyt, P.M.M. (1997b) Observer variation and clinical decision making. [Letter to the editor.] Fertil. Steril., 68, 381–384.[ISI][Medline]

Mol, B.W.J., Swart, P., Bossuyt, P.M.M. and Van der Veen, F. (1997c) Is hysterosalpingography an important tool in predicting fertility outcome? Fertil. Steril., 67, 663–669.[ISI][Medline]