Human papillomavirus testing and the management of women with mildly abnormal cervical smears: an observational study

BMJ 2001; 322 doi: (Published 14 April 2001) Cite this as: BMJ 2001;322:893
  1. Gemma Rebello, consultant cytopathologista,
  2. Nick Hallam, consultant virologistb,
  3. George Smart, consultant gynaecologista,
  4. David Farquharson, patient services director, reproductive medicinea,
  5. Jane McCafferty, colposcopy sistera
  1. aColposcopy Clinic, Royal Infirmary of Edinburgh, Lothian University Hospitals NHS Trust, Edinburgh EH3 9YW
  2. bRegional Clinical Virology Laboratory, City Hospital, Lothian University Hospitals NHS Trust, Edinburgh EH10 5SB
  1. Correspondence to: N Hallam
  • Accepted 11 October 2000

Editorial by Manos

Following publication of a report in Health Technology Assessment the NHS is running a pilot scheme screening women for human papillomavirus if they have a mildly dyskaryotic or borderline smear.1 How reliable is testing for human papillomavirus as a marker for high grade disease in those with mildly abnormal smears?

Participants, methods, and results

Three hundred and thirty three consecutive new patients (aged 17 to 61 years, median 30 years) referred for colposcopy with persistent borderline or mildly dyskaryotic smears and who consented to the study were tested for human papillomavirus (high risk types only) with the Digene Hybrid Capture assay HC II (Abbott Laboratories, Maidenhead), by using cervical brush specimens placed in Digene transport medium, and were treated by large loop excision of the transformation zone.

The table summarises test performance, and table A on the BMJ's website presents results by age, smear history, and test cut off.2 Subjects aged under 30 years (166) were more likely than older subjects (167) to test positive for human papillomavirus (79% (131/166) v 45% (75/167); χ2=39.4, df=1, P<0.001) and (dependent on this) to have cervical intraepithelial neoplasia grade 2 or 3 (high grade disease) (43% (71/166) v 27% (45/167); χ2=8.5, df=1, P<0.01).

Performance of human papillomavirus testing in detecting subjects with high grade disease according to age, smear history, and cut off value of test

View this table:


A proportion of patients with mildly abnormal smears will harbour high grade disease.3 Identifying this subgroup to target appropriate management is an important clinical issue. Colposcopy itself does not identify high grade disease reliably and we asked whether testing for human papillomavirus might improve diagnostic accuracy. We found that testing for human papillomavirus had a higher overall sensitivity (93% [1 pg/ml cut off] to 85% [4 pg/ml]) than specificity (55% [1 pg/ml] to 62% [4 pg/ml]) in detecting those with high grade disease, thus limiting its usefulness as a surrogate marker (table). This differential was greater for younger women, for those with mild dyskaryosis, and at 1 pg/ml cut off (table). Testing performed best with negative predictive values (prevalence adjusted) of 96% in older women and 94% overall (1 pg/ml).

The report in Health Technology Assessment concludes that “the clearest role for HPV testing at the moment is in the management of women with borderline or mildly dyskaryotic smears. In particular, those aged above 30 years who test positive for high risk types could be referred immediately for colposcopy, while those younger than 30 years who test negative could receive less-intensive surveillance.” The report supports limited introduction of testing for human papillomavirus, which should be carefully monitored, and encourages further research, including assessing the safety of returning to routine screening women with borderline or mild smears who test negative for human papillomavirus. Extrapolating our results to these guidelines would mean that 55% of those aged 30 years or over who test positive and are referred would have high grade disease but such disease might be missed in a small proportion of those testing negative and not referred (4% for older subjects, 11% for younger ones), this being more significant for younger women if they then face less intensive surveillance. We recognise, however, that our study population, with repeated abnormal smears, differs from that at triage for a single abnormal smear when the prevalence of high grade disease would probably be lower.

Manos et al report that human papillomavirus testing is useful in triaging those with atypical squamous cells of undetermined significance (approximates to borderline change)4; others are evaluating this but believe that testing has limited potential in triaging low grade squamous intraepithelial lesions (approximates to mild dyskaryosis).5 We found an overall test sensitivity and specificity of 86% and 76% for 75 subjects with borderline smears, 94% and 39% for 117 with mild dyskaryosis, and 95% and 54% for 141 with both smear types (1 pg/ml). We welcome the NHS pilot scheme while advising caution in the clinical use of testing for human papillomavirus, especially at a single point in time.


We thank nursing and medical colleagues at the colposcopy clinic for their assistance.

Contributors: GR had the idea for the study and reviewed the histology. NH performed human papillomavirus testing, wrote the paper, and will act as guarantor. All the authors were clinically involved in carrying out the study and contributed to the final version of the paper.


  • Funding The Royal Infirmary of Edinburgh Endowment Fund. Abbott Laboratories kindly provided human papillomavirus testing kits at a discount.

  • Competing interests GR and NH have been sponsored by Digene Diagnostics to attend several conferences.

  • Embedded Image Table A appears on the BMJ's website


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract