S I G N

Guideline topic: Management of post menopausal bleeding

Evidence table: Diagnosis of endometrial cancer and other abnormalities in postmenopausal women taking tamoxifen

Reference

Study type

SIGN evidence level

Number of subjects

Diagnostic test

Reference standard

Inclusion criteria

Outcomes

Sensitivity
(95% CIs)

Specificity

(95% CIs)

Additional comments

Franchi et al (1999)

Cross sectional

2-

163 (46 vaginal bleeding)

TVUS

endometrial thickness >9mm

Hysteroscopy +/- biopsy

Breast cancer; post-menopausal;

endometrial thickness >4mm

Endometrial disorders

87 (69-100)

For vaginal bleeding:

88 (76-100)

78 (72-84)

For vaginal bleeding:

70 (50-90)

Logistic regression indicated ORs of abnormal findings: vaginal bleeding 7.5; endometrial thickness 4.6; and duration of tamoxifen use 2.0.

Tesoro et al (1999)

Retros-pective analysis

2-

80 (12 vaginal bleeding)

TVUS + sonohystero-graphy if >2.5mm

D&C or endometrial biopsy

Breast cancer; post-menopausal

Endometrial disorders

88 (75-100)

19 (2-36)

Not all patients had reference standard

Timmerman et al (1998)

RCT

1-

52 (none vaginal bleeding)

TVUS + sonohystero-graphy if > 4mm

Hysteroscopy and biopsy

Breast cancer; post-menopausal; asymptomatic

Endometrial polyps and cancer

93 (83-100)

88 (75-100)

1 case endometrial cancer missed by hysteroscopy but detected by TVUS and hysterectomy (included in recalculated outcomes); 35 women actually underwent sonohysterography

Cecchini et al (1998)

Cohort study

2-

1010 (no initial bleeding)

Annual TVUS endometrium >5mm

Biopsy if abnormal

Breast cancer; post-menopausal

Endometrial cancer

N/A

N/A

1 cancer detected at initial screen; 4 interval cancers – at least 2 presented with bleeding

Tepper et al (1997)

Cohort study

2-

68 (none vaginal bleeding)

Sonohystero-graphy

hysteroscopy

Endometrial thickness > 8 mm on TVUS

‘Any abnomality’; complex hyperplasia

Abnormality:

91 (79-100)

Complex:

33 (0-86)

Abnormality:

98 (94-100)

Complex:

68 (57-79)

No blinding; results may have been influenced by sequencing of tests; unrepresentative sample; small numbers

Love et al (1999)

Cross sectional

2-

357 (none vaginal bleeding); 130 controls

TVUS > 5mm

hysteroscopy

Breast cancer; asymptomatic

Detection of endometrial abnormalities

N/A

’46% false positive rate’

Hysteroscopy only if TVUS abnormal; no blinding; TVUS less painful than hysteroscopy

Gerber et al (2000)

Cohort

2-

247 (none VB); 98 controls

TVUS every 6mths for 5 yrs >10mm or >5mm plus irregular morphology

Hysteroscopy / D&C

Breast cancer; asymptomatic

Detection of endometrial abnormalities

N/A

N/A

1 cancer detected in 52 asymptomatics versus 2/20 with vaginal bleeding; no blinding; hysteroscopy only if TVUS abnormal or bleeding

Barakat et al (2000)

Cohort

2-

159 (none VB)

EB every 6mths for 2 yrs; F/B every 12mths for 3 yrs

Hysteroscopy / D&C

Breast cancer; asymptomatic

Detection of endometrial abnormalities

N/A

N/A

48 ‘nonassessable’; no blinding; investigation only if abnormalities or symptoms; no cancers detected (but under-powered)