After-effects reported by women following colposcopy, cervical biopsies and LLETZ: results from the TOMBOLA trial

Authors


Dr L Sharp, National Cancer Registry Ireland, Building 6800, Cork Airport Business Park, Kinsale Road, Cork, Ireland. Email linda.sharp@ncri.ie

Abstract

Objective  Few studies have investigated physical after-effects of colposcopy. We compared post-colposcopy self-reported pain, bleeding, discharge and menstrual changes in women who underwent: colposcopic examination only; cervical punch biopsies; and large loop excision of the transformation zone (LLETZ).

Design  Observational study nested within a randomised controlled trial.

Setting  Grampian, Tayside and Nottingham.

Population  Nine hundred-and-twenty-nine women, aged 20–59, with low-grade cytology, who had completed their initial colposcopic management.

Methods  Women completed questionnaires on after-effects at approximately 6-weeks, and on menstruation at 4-months, post-colposcopy.

Main outcome measures  Frequency of pain, bleeding, discharge; changes to first menstrual period post-colposcopy.

Results  Seven hundred-and-fifty-one women (80%) completed the 6-week questionnaire. Of women who had only a colposcopic examination, 14–18% reported pain, bleeding or discharge. Around half of women who had biopsies only and two-thirds treated by LLETZ reported pain or discharge (biopsies: 53% pain, 46% discharge; LLETZ: 67% pain, 63% discharge). The frequency of bleeding was similar in the biopsy (79%) and LLETZ groups (87%). Women treated by LLETZ reported bleeding and discharge of significantly longer duration than other women. The duration of pain was similar across management groups. Forty-three percent of women managed by biopsies and 71% managed by LLETZ reported some change to their first period post-colposcopy, as did 29% who only had a colposcopic examination.

Conclusions  Cervical punch biopsies and, especially, LLETZ carry a substantial risk of after-effects. After-effects are also reported by women managed solely by colposcopic examination. Ensuring that women are fully informed about after-effects may help to alleviate anxiety and provide reassurance, thereby minimising the harms of screening.

Introduction

Over 250 000 smears annually are reported as showing a low-grade abnormality in the UK NHS Cervical Screening Programmes (CSPs).1 Over recent years, there has been an increasing tendency to refer women with low-grade cytology to colposcopy for further investigation and treatment, if required, by excision or destruction.1 While several studies have reported psychosocial morbidity among women referred for colposcopy (reviewed in2), data on physical after-effects (such as pain or bleeding) and other consequences (such as impact on menstruation) experienced by women following colposcopy and related interventions are scant. Women referred for colposcopy following low-grade abnormal cytology may be managed by a range of options, including colposcopic examination without further intervention, colposcopy and punch biopsies, or colposcopy and LLETZ. We are not aware of any studies comparing the occurrence of after-effects between these different groups. Such data would make an important contribution to the debate concerning the relative benefits and harms of different management options at colposcopy.3

Of the few available studies on after-effects, some relate to cone biopsy, cold knife conisation or cryosurgery, 4–6 which are now rarely used in women with low-grade cytology. Several other studies concern LLETZ, but, of these, some are from the time when this technique was only starting to be commonly used and others relate to clinical practice in single centres and may not be widely generalisable.7–9 Some studies pertain to complications reported by the clinicians who conducted the procedures,10–12 rather than those reported by women themselves and the perspectives of these two groups might differ considerably.13 In addition, most of the clinicians’ reports relate to the time during and immediately after the procedure and consider major events or complications such as haemorrhage, which are well-recognised risks of LLETZ.14 Although such data are clinically important, they do not capture other important after-effects that occur in the subsequent days and weeks. While such after-effects may not always result in women coming into contact with health services again, they potentially impact on psychosocial wellbeing and quality-of-life15 and, as such, are important consequences of screening.

We compared the frequency of after-effects (pain, bleeding and discharge) and the impact on subsequent menstruation reported by women with low-grade cytology attending colposcopy and managed solely by colposcopic assessment, by punch biopsies or by LLETZ.

Methods

The TOMBOLA trial

The study was nested within the colposcopy arm of the TOMBOLA trial (Trial Of Management of Borderline and Other Low-grade Abnormal smears), a UK multicentre randomised controlled trial, full details of which are described elsewhere.16 Women aged 20–59 years, resident in Grampian, Tayside or Nottingham with a recent low-grade smear taken as part of routine screening were invited to participate in TOMBOLA. If they consented, they were randomised to either cytological surveillance (repeat cytology in primary care) or a colposcopy examination. Within the colposcopy arm, women were further randomised to immediate treatment by LLETZ or punch biopsies with recall for treatment if the biopsy showed CIN2 or more severe disease. Women only received these interventions if the transformation zone was colposcopically abnormal: otherwise, they underwent a colposcopic examination only. None of the women had a cervical cytology sample or an HPV test at colposcopy.

Development of questionnaires on after-effects

We developed two self-completion questionnaires to capture post-colposcopy effects experienced by women. Content was informed by extensive review of the literature and augmented by clinical opinion. The draft questionnaires were pre-tested among TOMBOLA participants for acceptability, ease of completion and face validity17 and modified as required prior to use.

The first questionnaire collected details of any pain/discomfort, bleeding and discharge, together with the duration (in number of days) and severity. Severity of pain was recorded on a 5-point scale ranging from ‘mild’ to ‘very severe’. Severity of bleeding and discharge was also recorded on a 5-point scale ranging from ‘light’ to ‘very heavy’. For women who had a colposcopic examination only, or colposcopy and punch biopsies which showed no CIN or CIN1, or colposcopy with immediate LLETZ, the questionnaire was sent approximately 6 weeks after the colposcopy visit. Women recalled for treatment following punch biopsies reported as CIN2 or worse were sent the questionnaire approximately 6 weeks after the treatment visit. The second questionnaire relating to menstruation was sent to all women approximately 4 months (16 weeks) after the colposcopy visit. Women were asked whether they had had any periods since the colposcopy and if so, whether the first period differed from ‘usual’ in terms of timing, flow, duration and discomfort. For both questionnaires, women who did not respond were sent a maximum of two reminders, 2 weeks apart.

Copies of the questionnaires are available from the authors on request.

Statistical analysis

We analysed women in three groups based on the management procedure received: (1) colposcopy examination only; (2) colposcopy and punch biopsies (but no LLETZ); and (3) colposcopy and a LLETZ (some of whom initially had punch biopsies). For each individual after-effect (i.e. pain, bleeding, discharge), we computed crude frequency rates in each management group. As the characteristics of women differed between the groups, we used logistic regression methods to adjust the rates for factors significantly associated with reporting each after-effect. Potential confounders included age, trial centre, cytology result and a range of socio-demographic and psychosocial factors collected by questionnaire. Variables were included in the models if they were significant (P < 0.05) on likelihood ratio tests. Model goodness-of-fit was checked using the Hosmer and Lemeshow test.18 We repeated these analyses for affect-effects reported to be ‘moderate’ or worse.

For each after-effect, duration was computed as the proportion of women still experiencing the after-effect one, two, three, etc days after the intervention. The distributions of duration were compared between groups, on a pair-wise basis, using the Kolmogorov–Smirnov test.

Women were included in the analysis of effects on menstruation if they reported having had at least one period since their colposcopy visit. We compared reported changes in timing, flow, duration and discomfort between management groups using chi-square tests. We created a variable summarising ‘any change’ to women’s usual menstrual pattern based on these four factors. Crude rates of any changes to menstruation are presented because adjusting for potential confounders had very little impact on the estimates.

All analyses were carried out in STATA version 10.0 (StataCorp LP, College Station, TX, USA).19

Results

The first questionnaire was sent to 929 women; 751 were completed and returned (response rate 80%). Of the respondents, 401 had colposcopy only, 165 had colposcopy and punch biopsies and 185 had colposcopy and LLETZ. Fourteen percent of women in the LLETZ group had punch biopsies taken before their LLETZ. Characteristics of participants, overall and by group, are shown in supporting information Table S1. Overall, 38% of respondents were aged 20–29 years, 27% were aged 30–39, 25% were aged 40–49 and 10% were aged 50–59. Twenty-two percent had a mild smear at recruitment to TOMBOLA and 78% had a smear showing borderline nuclear abnormalities (BNA). Less than 5% had had another BNA smear in the 3 years before recruitment. Fifty-seven percent of women were parous. One-third of women were current users of the oral contraceptive pill. Fifty percent had never smoked, 18% were ex-smokers and 31% current smokers.

Frequency of after-effects

Table 1 shows crude and adjusted rates of pain, bleeding and discharge for each management group. After adjusting for confounders, the frequency of each after-effect was lowest among women who had colposcopy only and highest among those who had a LLETZ. In the colposcopy only group, between 14 and 18% reported experiencing each after-effect. Around half of women managed by punch biopsies reported pain (unadjusted rate: 55%; adjusted rate: 53%) compared with around two-thirds of those who had LLETZ (adjusted rate: 67%). The results for discharge were similar (adjusted rates: 46 and 63% respectively). In contrast, almost as many women reported bleeding in the biopsy only group as in the LLETZ group (adjusted rates: 79 and 87% respectively).

Table 1.   Crude and adjusted prevalence of pain, bleeding and discharge, by management group—% with 95% CI
 Colposcopy examination onlyColposcopy and punch biopsiesColposcopy and LLETZ
  1. *Adjusted for trial centre, age, ever had children, current pill use, Hospital Anxiety and Depression Scale (HADS) anxiety score at colposcopy.

  2. **Adjusted for trial centre, age, ever had children.

  3. ***Adjusted for trial centre, age.

  4. ****Adjusted for age, current pill use, HADS anxiety score at colposcopy.

  5. *****Adjusted for ever had children, HADS anxiety score at colposcopy.

  6. ******Adjusted for age, Multi-dimensional Health Locus of Control Scale (MHLCS) powerful others score, Eysenck neuroticism score.

Pain
Any pain
 Crude19.7 (15.9–23.9)55.2 (47.2–62.9)67.0 (59.7–73.7)
 Adjusted*18.3 (14.6–22.7)52.7 (44.3–61.0)66.8 (58.9–73.8)
Moderate or more severe pain
 Crude6.0 (3.9–8.8)28.7 (21.9–36.2)39.6 (32.4–47.1)
 Adjusted**5.3 (3.5–8.1)28.0 (21.1–36.0)32.7 (25.6–40.6)
Bleeding
Any bleeding
 Crude15.2 (11.8–19.1)79.4 (72.4–85.3)85.4 (79.5–90.2)
 Adjusted***14.1 (11.0–18.0)79.1 (72.0–84.8)87.3 (81.6–91.4)
Moderate or more severe bleeding
 Crude3.0 (1.6–5.2)22.0 (15.9–29.1)52.5 (44.9–60.0)
 Adjusted****3.0 (1.7–5.3)21.4 (15.5–28.6)52.9 (45.2–60.4)
Discharge
Any discharge
 Crude15.7 (12.3–19.6)45.5 (37.7–53.4)64.9 (57.5–71.7)
 Adjusted*****15.4 (12.1–19.4)46.3 (38.6–54.1)63.0 (55.5–70.0)
Moderate or more severe discharge
 Crude5.0 (3.1–7.6)17.6 (12.1–24.3)44.3 (37.0–51.2)
 Adjusted******5.3 (3.4–8.1)14.3 (9.5–20.9)41.8 (34.0–50.1)

Four hundred-and-seventy-five women (63%) reported one or more after-effects. Of these, 39% had experienced only one (pain: 10%; bleeding: 17%; discharge: 12%), 33% had experienced two (pain plus bleeding: 19%; pain plus discharge: 4%; bleeding plus discharge: 9%) and 29% had experienced all three.

Severity of after-effects

Figure 1 parts A–C show, for each after-effect, frequency according to severity, with severity expressed on a five-point scale ranging from very mild/light to very severe/heavy. For each after-effect, the distributions of severity varied significantly between the three management groups. Women who had LLETZ had the most severe after-effects and women who had colposcopy only had the least severe after-effects. After adjusting for confounders, 3% of women in the colposcopy only group reported moderate or more severe bleeding, with 5% reporting moderate or more severe pain or discharge (Table 1). In the LLETZ group, over 40% of women experienced moderate or more severe discharge and over 50% experienced moderate or more severe bleeding. Less than half as many women in the biopsy only group reported these after-effects (moderate or worse discharge: 14%; moderate or worse bleeding: 21%). The frequency of moderate or more severe pain was similar in the biopsy only and LLETZ groups (28 and 33% respectively).

Figure 1.

 (A) Severity of pain reported, by management group. (B) Severity of bleeding reported, by management group. (C) Severity of discharge reported, by management group.

Within the biopsy only group, women who had between three and five punch biopsies were more likely than women who had one or two punch biopsies to report moderate or more severe pain (37% versus 23%; z = −1.97, P = 0.049) or moderate or more severe bleeding (25% versus 19%), although this latter difference was not statistically significant (z = −0.92, P = 0.358).

Duration of after-effects

The duration of after-effects (of any severity) by management group are shown in Figure 2A–C. For each after-effect, duration was shortest for women managed by colposcopy only. For pain, the duration was similar among those managed by biopsy only and those managed by LLETZ (D = 0.150, P = 0.230). Bleeding and discharge were of significantly longer duration among women who had a LLETZ than among women who had biopsies only (bleeding: D = 0.547, P < 0.001; discharge: D = 0.538, P < 0.001).

Figure 2.

 (A) Duration of pain: proportion of women still experiencing pain by days since procedure. (B) Duration of bleeding: proportion of women still experiencing pain by days since procedure. (C) Duration of discharge: proportion of women still experiencing discharge by days since procedure.

Menstruation after colposcopy

The second questionnaire was completed by 641 women (response rate 69%). Of these, 533 had had a period since their colposcopy visit. A further six women reported that they would normally have had periods, but had not had one since their last appointment; two of these women were pregnant and, of the remaining four, two were in the colposcopy only group, one in the biopsy only group and one in LLETZ group.

Table 2 shows women’s responses to questions relating to the first period after colposcopy. Twenty-nine percent (95% CI 23.4–34.2%) of women in the colposcopy only group reported some change to their first period post-colposcopy compared to 43% (95% CI 33.3–52.5%) in the biopsy only group and 71% (95% CI 62.4–78.1%) in the LLETZ group (chi-square (2df) = 68.41, P < 0.001). A similar trend across the management groups was seen when timing, flow, duration and discomfort were considered separately.

Table 2.   Prevalence of changes in menstruation* after colposcopy, by management group—% with 95% CI
 Colposcopy examination onlyColposcopy and punch biopsiesColposcopy and LLETZ
  1. *Relates to first period after colposcopy.

  2. **Any change to timing, flow, duration or discomfort.

Timing
Later than usual5.0 (2.7–8.2)13.0 (7.3–20.8)18.2 (12.2–25.7)
Same as usual85.8 (81.2–89.7)78.7 (70.0–86.0)70.8 (62.4–78.3)
Earlier than usual9.2 (6.1–13.2)8.3 (3.9–15.2)10.9 (6.3–17.4)
 Chi-square (4df) = 20.43, P < 0.001
Flow
Lighter than usual6.0 (3.6–9.5)5.7 (2.1–11.9)10.1 (5.6–16.4)
Same as usual82.9 (78.0–87.1)69.8 (60.1–78.3)42.0 (33.7–50.7)
Heavier than usual11.0 (7.6–15.3)24.5 (16.7–33.8)47.8 (39.3–56.5)
 Chi-square (4df) = 78.37, P < 0.001
Duration
Shorter than usual8.2 (5.2–12.0)10.4 (5.3–17.8)8.7 (4.6–14.7)
Same as usual83.0 (78.1–87.2)76.4 (67.2–84.1)51.4 (42.8–60.0)
Longer than usual8.9 (5.8–12.8)13.2 (7.4–21.2)39.9 (31.6–48.5)
 Chi-square (4df) = 64.56, P < 0.001
Discomfort
Less uncomfortable than usual4.2 (2.2–7.3)2.8 (0.6–8.0)4.3 (1.6–9.2)
Same as usual87.3 (82.8–90.9)79.4 (70.1–86.7)54.3 (45.7–62.8)
More uncomfortable than usual8.5 (5.5–12.4)17.8 (11.0–26.3)41.3 (33.0–50.0)
 Chi-square (4df) = 66.48, P < 0.001
Any change**
Crude prevalence28.6 (23.4–34.2) 42.7 (33.3–52.5)70.7 (62.4–78.1)
 Chi-square (2df) = 68.41, P < 0.001

When women were asked about subsequent periods, 31% (95% CI 22.9–39.1%) of those treated by LLETZ said that these were different from usual compared to 14% (95% CI 8.0–21.9%) managed by biopsies and 16% (95% CI 11.6–20.5%) who had a colposcopy examination only (chi-square (2df) = 15.46, P < 0.001).

Discussion

Strengths and weaknesses

To our knowledge, this study is the first to compare post-colposcopy physical effects in women managed by colposcopic examination only, cervical punch biopsies and LLETZ. It was population-based and nested in a pragmatic RCT within the UK CSPs, so generalisability of the results to the CSPs is likely to be high. Overall, 52% of eligible women participated in TOMBOLA,16 although younger women were less likely to take part than older women. Generally, we found that age was inversely associated with reporting after-effects. Thus, although the questionnaire response rates were high, we may have under-estimated the overall frequency of after-effects among women undergoing colposcopy, punch biopsies and LLETZ. On the other hand, women who had not experienced any post-colposcopy effects may have been less likely to complete the questionnaires, meaning that our estimates of the frequency of after-effects could be slightly too high.

A major strength of this study—and one of the reasons that it provides an important contribution to the evidence-base—is that it is one of the few studies to have considered after-effects from the perspective of women themselves, rather than from the perspective of the treating clinician. A study by Johnson and Crompton,13 of women with cervical atypia treated by laser surgery, illustrated the discordance between women’s own assessment of pain and the assessment of the supporting nurse and the surgeon. Within our study, women may have reported similar after-effects differently (for example, what one woman reported as light bleeding may have been considered by another to be heavy bleeding). This is not a limitation as the impact of the management interventions and any associated after-effects on a woman’s life is likely to depend upon her perception of the experience.

The frequency of after-effects reported by women varied by trial centre, even after adjusting for other factors such as women’s age. We know there were variations in practice between centres (for example, one centre routinely used sultrin cream while others did not), and it is plausible that these might account, at least in part, for the between-centre variations in after-effects. The effects of such practice variations on women’s experiences post-colposcopy are not well understood and need further investigation. In the current study, colposcopy, punch biopsies and LLETZ procedures were performed by a limited number of colposcopists, all of whom were accredited by the British Society for Colposcopy and Cervical Pathology. We are not aware of any evidence that after-effects experienced by women relate to characteristics of the colposcopist (such as experience).

Management received and after-effects

The most important factor predicting reporting of after-effects, of any type, was the type of procedure(s) the women had undergone. Frequency increased with increasing invasiveness of the intervention. The higher rates in women who had LLETZ compared with punch biopsies is unsurprising and corresponds with the findings of Doyle et al.,8 who reported that the proportion of women who had bleeding 2-weeks after excisional treatment increased with increasing size of excision.

There is compelling evidence that the most important predictor of diagnostic accuracy of punch biopsies is the number of biopsies taken.20 Our data suggest that women who have more biopsies are more likely to report after-effects, particularly pain. We, therefore, agree with suggestions that the increased performance offered by multiple biopsies needs to be weighed against increased discomfort and risk of complications for women.21

Management of low-grade smears by punch biopsies at colposcopy requires that a proportion of women will be recalled for treatment, by LLETZ or another procedure. Therefore, a further concern about this management policy might be that it would place women at greater risk of after-effects either because each procedure carries an independent risk or because having had previous punch biopsies increases the likelihood of after-effects following a LLETZ. Further examination of our data tentatively suggests that the frequency of pain might be higher in women who have punch biopsies followed by LLETZ compared with those who have had LLETZ only (81% versus 65%), but this was based on only 26 women who had punch biopsies prior to LLETZ. In addition, a similar pattern was not seen for bleeding or discharge. The issue of after-effects is a new dimension in the debate about the relative benefits and harms of management by ‘see and treat’ versus punch biopsies and selective recall. Our findings suggest that the experience of after-effects in women who have multiple punch biopsies taken, or undergo punch biopsies then LLETZ, should be a consideration in this debate.

The high frequency of after-effects reported by women who only had a colposcopy examination is noteworthy and important. This finding suggests that among women attending for colposcopy following a low-grade smear and who have no colposcopically visible abnormality, one in three will experience pain, bleeding or discharge and one in eight will report that this was ‘moderate’ or more severe. Given the number of colposcopies conducted in the CSPs each year (122 000 referrals to colposcopy in 2007–08 in England alone 22), this represents significant morbidity and is an important consequence of screening for those who participate. In the US, the perception that a smear is painful was found to be a strong predictor of non-adherence to screening recommendations.23 It is therefore plausible that the experience of pain (or other after-effects) might impact negatively on adherence with follow-up post-colposcopy or with future participation in screening. Further investigation of this is warranted.

In a study of women managed by loop excision, Paraskevaidis et al.15 found a link between bleeding and consultations with medical professionals. In our study, 14% of women who reported after-effects (most of whom had had LLETZ) had sought professional advice in relation to these, from various sources including their GP, practice nurse, pharmacist and NHS Direct. This suggests that as well as constituting a cost to women, after-effects also have a cost for the NHS.

Frequency of specific after-effects

As regards individual after-effects, previous studies have focussed on women managed by LLETZ following colposcopy.7–10,12,15,19 Findings of these studies vary considerably. There are several explanations for these variations, including differences between studies in who recorded/reported the after-effects, the method and timing of data collection, the definitions of after-effects and characteristics of participants. In the current study, 67% of those treated by LLETZ reported pain. This compares with 41%7 and 69%9 reported in two UK studies that were undertaken in colposcopy clinics and used similar methodology to our study. The first and largest of these two studies was conducted in a single department7 and variations in practice that might influence reporting of after-effects are likely to have been less than in the current study; this may in part account for the lower reported occurrence of pain in that study than in our study. The proportions of women managed by LLETZ in the current study who reported having experienced bleeding (86%) and discharge (65%) were comparable to those from these two previous studies (bleeding: 72%7 and 79%9; discharge: 69%7 and 72%9).

In common with our finding that seven out of ten women who had undergone LLETZ experienced changes in their menstrual pattern, Lopes et al.7 and Williams et al.9 also observed that substantial proportions of women reported menstrual changes. For example, 32% noted a change with regard to timing,7 while 47% in one study7 and 19% in the other9 reported changes in relation to flow. We are not aware of any previous studies of menstrual pattern changes after punch biopsies (four in ten women in TOMBOLA) or colposcopic examination only (three in ten).

Implications and conclusion

This study demonstrates that colposcopic examination, punch biopsies and LLETZ are not trivial interventions for women and carry a substantial risk of after-effects. The consistency between our findings for LLETZ and those of previous studies in UK colposcopy clinics7,9 implies that our findings for women managed by punch biopsies and those who undergo colposcopic examination are likely to be generalisable to the NHS CSPs. The findings are therefore relevant, not only for women undergoing colposcopy but also for colposcopy clinic staff and primary care staff who may have to deal with the consequences of colposcopy and related interventions.

The NHS CSP colposcopy leaflet for women provides little information about the after-effects that may be experienced following punch biopsies or LLETZ.24 To better prepare women for colposcopy and related interventions, we would suggest that information provided to women could be improved to include, for example: the risks of experiencing different after-effects (pain, bleeding, discharge and menstrual changes) with each procedure, including colposcopic examination only; advice on what to do if after-effects occur (even if that is to say that these are ‘completely normal’); and some indication of when it would be appropriate to consult a pharmacist or medical professional. Such information might also note that many women, even those who have a LLETZ, will not experience any after-effects and that this too is ‘normal’. Providing this type of information would help ensure that women are fully informed about the procedures and their likely consequences. It would also let women prepare for what might happen afterwards and the potential impact of this on their lives. Ultimately, this may help alleviate anxiety and provide reassurance in women attending colposcopy, thereby minimising the harms of screening.

Disclosure of interests

None of the members of the writing and analysis group for this paper have any interests to declare.

Contribution to authorship

Linda Sharp contributed to study design and questionnaire development; directed statistical analysis; drafted initial version of paper with SC; and drafted final version of paper. Seonaidh Cotton contributed to study design and questionnaire development; directed statistical analysis; drafted initial version of paper with LS; and commented on final version of paper. Claire Cochran undertook literature review and initial statistical analysis and contributed to interpretation and revisions of paper. Nicola Gray contributed to study design and questionnaire development, interpretation of data and revisions of papers. Julian Little contributed to study design, interpretation of data and revisions of paper. Keith Neal contributed to overall design of TOMBOLA, and revisions of paper. Maggie Cruickshank contributed to questionnaire development, interpretation of data and revisions of paper. All authors reviewed and approved the final version of the paper.

Details of ethics approval

Ethical approval was obtained from the joint Research Ethics Committee of NHS Grampian and the University of Aberdeen, the Tayside Committee on Medical Research Ethics and the Nottingham Research Ethics Committee.

Funding

The work was supported by the Medical Research Council (G9700808) and the NHS in England and Scotland.

Acknowledgements

We thank Ian Duncan and Norman Waugh for helpful comments on an earlier version of the manuscript. We are grateful for the cooperation and assistance that we received from NHS staff in the coordinating centres and clinical sites. We thank the women who participated in TOMBOLA.

Appendix

The TOMBOLA Group comprises:

Grant-holders:

University of Aberdeen and NHS Grampian, Aberdeen, Scotland

Maggie Cruickshank, Graeme Murray, David Parkin, Louise Smart, Eric Walker, Norman Waugh (Principal Investigator 2004–08)

University of Nottingham and Nottingham NHS, Nottingham, England

Mark Avis, Claire Chilvers, Katherine Fielding, Rob Hammond, David Jenkins, Jane Johnson, Keith Neal, Ian Russell, Rashmi Seth, Dave Whynes

University of Dundee and NHS Tayside, Dundee, Tayside

Ian Duncan, Alistair Robertson

University of Ottawa, Ottawa, Canada

Julian Little (Principal Investigator 1999–2004)

National Cancer Registry, Cork, Ireland

Linda Sharp

Bangor University, Bangor, Wales

Ian Russell

University of Hull, Hull, England

Leslie Walker

Staff in clinical sites and coordinating centres

Grampian: Breda Anthony, Sarah Bell, Adrienne Bowie, Katrina Brown, Joe Brown, Kheng Chew, Claire Cochran, Seonaidh Cotton, Jeannie Dean, Kate Dunn, Jane Edwards, David Evans, Julie Fenty, Al Finlayson, Marie Gallagher, Nicola Gray, Maureen Heddle, Alison Innes, Debbie Jobson, Mandy Keillor, Jayne MacGregor, Sheona Mackenzie, Amanda Mackie, Gladys McPherson, Ike Okorocha, Morag Reilly, Joan Rodgers, Alison Thornton, Rachel Yeats

Tayside: Lindyanne Alexander, Lindsey Buchanan, Susan Henderson, Tine Iterbeke, Susanneke Lucas, Gillian Manderson, Sheila Nicol, Gael Reid, Carol Robinson, Trish Sandilands

Nottingham: Marg Adrian, Ahmed Al-Sahab, Elaine Bentley, Hazel Brook, Claire Bushby, Rita Cannon, Brenda Cooper, Ruth Dowell, Mark Dunderdale, Dr Gabrawi, Li Guo, Lisa Heideman, Steve Jones, Salli Lawson, Zoë Philips, Christopher Platt, Shakuntala Prabhakaran, John Rippin, Rose Thompson, Elizabeth Williams, Claire Woolley

Statistical analysis: Massoud Boroujerdi, Seonaidh Cotton, Kirsten Harrild, John Norrie

External Trial Steering Committee: Nicholas Day (chair, 1999–2004), Theresa Marteau (chair 2004-), Mahesh Parmar, Julietta Patnick and Ciaran Woodman

External Data Monitoring and Ethics Committee: Doug Altman (chair), Sue Moss, Michael Wells

Writing and analysis group for this paper: Linda Sharp1, Seonaidh Cotton2, Claire Cochran2, Nicola Gray3, Julian Little4, Keith Neal5, Maggie Cruickshank6

1National Cancer Registry Ireland, Cork, Ireland; 2Health Service Research Unit, University of Aberdeen, Aberdeen, Scotland; 3Department of General Practice and Primary Care, University of Aberdeen, Aberdeen, Scotland; 4Canada Research Chair in Human Genome Epidemiology, Department of Epidemiology and Community Medicine, University of Ottawa, Ontario, Canada; 5Department of Epidemiology and Public Health, University of Nottingham, Nottingham, England; 6Department of Obstetrics and Gynaecology, University of Aberdeen, Aberdeen, Scotland

Ancillary