Download Cervical cancer: screening

There is so much we don’t know in medicine that could make a difference,
and often we focus on the big things, and the little things get forgotten. To
highlight some smaller but important issues, we’ve put together a series of
pearls that the Red Whale found at the bottom of the ocean of knowledge!
Cervical cancer: screening
“I think there has been a mistake with my smear result, it says borderline but they don’t want to recheck for
another 3 years, please can I have another test in 6 months like last time?”
In the past, this young woman’s borderline result had been followed by treatment for CIN2 six months later so I
could understand her anxiety! The NHS cervical cancer screening programme has recently changed to
incorporate triage of borderline or mild dyskaryois cytology results with testing for HPV. We will consider the
rationale behind this and implications for women with positive or negative HPV test results.
National Cervical Screening Programme statistics and screening intervals
In 2013, coverage of eligible women (who had had at least one test in the last 5y) was 78.3%.
Aside from a slight increase in 2009, coverage has fallen since 2003 (when it was 81.2%).
If coverage of 80% can be achieved, it is estimated that a reduction in death rates of 95% is possible in the long term.
Cervical screening call and recall is now managed centrally.
Screening intervals:
Age group
Frequency of screening
25
First invitation
25–49
3 yearly
50–64
5 yearly
65+
Only women who have had no test since age 50 or a recent
abnormal test
Cervical screening terminology
Sometimes the terminology used to report cervical screening outcomes can be confusing and terms are used interchangeably.
Cervical smear (cytology) results report the level of dyskaryosis (changes seen within the cervical cells).
They may be reported as normal, borderline, mild, moderate or severe dyskaryosis. These reflect increasing abnormality in
the cervical cells. Moderate or severe dyskaryosis may also be referred to as ‘high grade dysplasia’ and mild dyskaryosis as
‘low grade dysplasia’.
The extent of CIN (cervical intraepithelial neoplasia) can only be determined at biopsy (taken at colposcopy). CIN reflects the
degree of invasion of the abnormal cells.
CIN1 is a one-third depth invasion, CIN3 is full thickness invasion.
The level of dyskaryosis is suggestive, but not indicative, of the level of CIN that will be found at colposcopy, e.g. it is possible
to have moderate dyskaryosis and CIN3, etc.
HPV triage and cervical cancer screening
From April 2011, HPV testing was incorporated into the NHS National Cervical Screening Programme. The interpretation of results
and appropriate follow-up is now more complicated, as is evident in the Public Health England screening protocol algorithm
published in April 2014 which has no less than four flow charts and a whole page of codes and abbreviations!
The key points are:
Cytology is still carried out as the primary screening test.
Women with moderate or severe dyskaryosis are referred straight for colposcopy without HPV testing.
Women with borderline or mild dyskaryosis are tested for HPV (the test is done on the same sample used for cytology).
If the HPV test is positive then the woman is offered colposcopy.
If the HPV test is negative then she can return to the regular screening programme.
Women who receive colposcopy treatment for CIN 1, 2 or 3 now have a HPV ‘test of cure’ at 6m.
Negative HPV result and risk of cervical cancer
A large RCT published in the BMJ compared screening by cytology and HPV testing to cytology alone (BMJ 2014;348:g130). The
women were selected from the Swedish National Screening Programme and followed up for 13y. The protocol was slightly different
to that now in use in the UK because women with normal cytology but positive HPV at baseline were retested for HPV at 12m.
The trial found that the negative predictive value for HPV and cytology-based screening was higher than for cytology alone
throughout follow-up.
The negative predictive value of normal cytology and a negative HPV test at baseline for development of CIN2 or worse over
10y was 99.5%.
This means <1% of women with these two negative results went on to develop severe cervical abnormalities.
What do I tell women about their HPV result?
Some types of HPV, especially types 16 and 18 are found in 99% of cervical cancers. The cervical cancer screening
programme tests for these high risk subtypes and women with a positive result have an increased risk of cervical
abnormalities that may progress to cervical cancer.
Women with a negative test result have a low risk of developing cervical cancer so can safely return to the normal screening
programme.
Most sexually active women come into contact with HPV at some time in their life. It can be spread by any close sexual
contact, not just penetrative sex and the use of condoms provides only limited protection.
There is nothing you can prescribe to ‘cure’ the infection. In most cases the woman’s immune system will clear the infection
within 1–2y.
HPV as the primary test for cervical cancer screening
If HPV status can be used to determine risk and follow-up interval for women with cervical screening abnormalities detected by
cytology, could it be used effectively as the primary screening tool for all women undergoing cervical screening?
This meta-analysis in the Lancet pooled data from four large European RCTs that compared HPV with cytology-based cervical
screening (Lancet 2014;383:524).
Overall, HPV-based screening prevented 60–70% more invasive cervical cancers than cytology-based screening.
The meta-analysis suggested that 5y intervals with HPV-based screening are safer than 3-y intervals with cytology-based
screening.
However, the results only apply to women from the age of 30y as there were insufficient data to draw conclusions about
younger women.
A cost-effectiveness analysis (BMJ 2012;344:e670) suggested that primary HPV testing followed by cytology only if HPV status is
positive was the most cost-effective approach for women aged >30y in the UK.
Future screening in the UK
In May 2013, the UK cervical screening programme started running pilots for HPV primary screening in six laboratories in England.
The pilot includes all women aged 25–64 being recalled by the cervical screening programme.
Women who are HPV negative do not have any cytology and are recalled after 3 years (if aged 25–49y) or 5 years (if aged
50+).
The results of the pilot will help to determine whether primary HPV screening is adopted throughout the UK – watch this space!
HPV as test of cure after CIN treatment
After colposcopy treatment for CIN, women in the UK screening programme are offered a repeat test of HPV status to assess
whether they have cleared the virus and to stratify their on-going risk of cervical cancer.
This is a beneficial addition to cytology follow-up. A recent large study from the Netherlands showed that after treatment for CIN,
even after three consecutive normal cervical smears, the risk of cervical cancer remained elevated about 4x greater than if a
patient's index smear was normal (BMJ 2012;345:e6855). This was either because abnormal cells were left behind after colposcopy
treatment or persistent HPV infection. Identifying women with persisting HPV infection identifies the highest risk group who will
continue to need more frequent smears.
This has been demonstrated to be cost-effective (BMJ 2012;345:e7086). This UK-based economic modelling study suggests that
HPV test of cure would be more cost-effective than cytology-only follow-up and recommend that this is implemented across the
whole UK screening programme.
Does cervical screening work?
It is well established that women participating in organised cervical screening have a lower incidence of invasive cervical cancer (OR
0.21; CI 0.16–0.28) (J Natl Cancer Inst 2008;100:622).
Five year survival is also improved, but these data are often criticised because if you detect cancer by screening, you potentially
introduce lead time bias (i.e. people appear to survive longer purely because the date of diagnosis was sooner).
This large Swedish cohort study used modelling techniques to determine statistical cure and showed that cervical screening (BMJ
2012; 344:e900):
Improved cure rates of cervical cancer in those detected by screening compared with those presenting symptoms:
Screening detected cure rate: 92% (CI 75–98%)
Symptomatic cure rate: 66% (CI 62–70%)
This was not due to 'lead time bias' and exceeded the benefits of detection at earlier stage alone.
75% of deaths in this study occurred in women who had not attended screening.
The Swedish cervical screening programme is organised in a very similar way to the UK and so these data are probably transferable
to our population.
Why are we not screening under 25 year olds?
The National Screening Committee (www.nsc.nhs.uk) revised its policy on the age range for cervical screening. This was on the
basis of Cancer Research UK research and recommendations.
The absolute risk of cervical cancer in under 25 year olds is very low.
The risk of false positives and unnecessary treatment is greatest in this young age group.
Using modelling data (Int J Cancer 2009;124:461) the same research group looked at what would happen to a woman aged 20–24y
with CIN3 by the time she was 25 and attended first screening:
1.5% would progress to invasive cancer.
50% of CIN3 would have regressed over the same time period (hence unnecessary treatment would be given if it had been
detected by screening).
This was supported by a modestly sized case–control study (BMJ 2009;339:b2968) that looked at the age-specific effectiveness of
cervical screening). They found:
There was no beneficial effect of screening in those aged 20–24y. In other words, in this age group, those who had been
screened were just as likely to develop cervical cancer as those who had not!
Screening was highly effective in older women, showing a 60% reduction in cancers at age 40y and an 80% reduction in
cancers at age 60y.
Note: this does NOT mean that young women do not develop cervical cancer, just that cervical screening is not an effective way of
detecting them. Young women presenting with symptoms should be taken seriously and investigated. HPV triage may have a role to
play in the future.
What about women who have never been sexually active?
Their risk is very low.
99% of cervical cancers are caused by HPV infection.
They are still eligible for cervical screening but may choose to decline.
Women who have been sexually active at any time carry some risk.
Lesbian and bisexual women
A literature review commissioned by the NHS Cervical Cancer Screening Programme in 2009 found that:
Uptake of cervical screening is up to 10 times lower in this group.
Prevalence of HPV amongst lesbian and bisexual women is 3–30%.
Prevalence amongst lesbian women who report never having heterosexual intercourse is 19%.
HPV can be transmitted through lesbian sexual contact.
30% of lesbian women had been told by their GP they did not need a smear test.
We should encourage lesbian and bisexual women to participate in screening and vaccination in the same way as
heterosexual women.
Is self-testing possible?
Self-taken cervicovaginal swab
This Dutch study looked at whether sending serial non-responders a self-testing kit for collection of cervicovaginal material and
screening this for HPV status may help to improve coverage (BMJ 2010;340:c1040). If the test came back positive for HPV, women
still need a smear test for cytology examination. They found:
Only 25% of the non-responders returned their kit. 75% remained non-responders.
99% of the returned kits were suitable for HPV testing.
Of the women identified as being high risk HPV-positive, 80% agreed to cytology.
This may not be the best use of resources as the conversion rate is still pretty poor.
Urine testing for HPV
Could urine testing for HPV offer a better self-testing alternative? This systematic review and meta-analysis (BMJ 2014;349:g5264)
with an accompanying editorial (BMJ 2014;349:g5542) looked at the accuracy of urine HPV detection compared to detection of HPV
from cervical sampling. The meta-analysis found that:
Urine detection of any HPV had a sensitivity of 87% (CI 78–92%) and specificity of 94% (CI 82–98%).
Urine detection of high risk HPV had a sensitivity of 77% (CI 68–84%) and specificity of 88% (CI 58–97%).
The only factor that explained heterogeneity between studies was the type of urine sample. First void urine samples were
significantly more sensitive than random or mid-stream samples.
The meta-analysis suggests that using first void urine may be an alternative to cervical sampling for HPV detection in women who
will not attend for cervical sampling. However, it is not as sensitive as cervical sampling and, crucially, we don't know if nonresponders to cervical screening invitations would accept this as an alternative test.
In addition, further research with long-term follow-up is needed to determine the predictive value of urine-detected HPV for CIN or
cervical cancer before this could be considered as part of the cervical cancer screening programme.
Adverse obstetric outcomes after CIN treatment
Preterm delivery
Two large observational studies and a meta-analysis in the Lancet and BMJ 7y ago established a link between treatment of CIN and
adverse obstetric outcomes, particularly preterm delivery (Lancet 2006;367:489 and BMJ 2008;337:a1284). As a result of these
studies, we were encouraged as GPs and midwives to refer women who had had cervical treatment for shared obstetric care so that
cervical length could be checked and consideration given to placing a cervical suture. However, with both these studies, concern
remained that the increased risk seen may be as a result of confounding factors, e.g. smoking, persisting HPV infection. In addition,
most previous studies were not UK-based.
Two more recent observational studies (one cohort BMJ 2012;345:e5174 & e5847 (editorial) and one case–control (BMJ
2014;349:g6223)) have looked at the risk of preterm delivery after treatment for CIN in 12 NHS centres in the UK where colposcopy
is highly regulated and guideline driven and LLETZ (large loop excision of transformation zone) is the treatment of choice.
The cohort study found that for women who had had treatment (either punch biopsy or LLETZ):
The risk of preterm delivery was increased.
9% of all births in the treatment group were preterm compared with 6.7% of all births in England over the same period.
This suggests that treatment in UK (compared to other developed countries) presents a lower risk, probably because of the
techniques used. However, the risk is still greater than in the background population.
The authors comment that this excess risk could be due to the treatment effect or to confounding factors which they were
unable to correct for, such as ethnicity, smoking and immune status, which may affect the risk of preterm delivery.
The case–control study looked at the risk of preterm delivery in women after punch biopsy or cervical excision between 1988 and
2011. Most (92%) of the women with a single excision were treated with LLETZ. Women with a preterm (<37w gestation) birth were
matched with women with term births (≥38w). They excluded women with cervical cancer, certain medical conditions and high risk
pregnancies. They found:
Overall, the risk of preterm birth was greater in women who had an excision compared to a punch biopsy (RR 1.38, CI 1.1–
1.72).
However, when the women were grouped by depth of excision there was no significant difference in risk between women who
had a punch biopsy and those having a small (< 10mm) excision.
The risk of preterm delivery increased with the depth of excision.
Women with excisions >15mm depth had a significantly increased risk of subsequent preterm delivery compared to those with
excision <10mm (see table).
Depth of excision
Absolute risk of preterm
birth (95% CI)
RR preterm birth compared
to small excision (<10mm)
Risk in general UK women, for
comparison
6.7%
Punch biopsy
7.2% (5.9–8.5%)
0.96 (0.73–1.27)
Small (<10mm)
7.5% (6–8.9%)
1 (reference)
Medium (10–14mm)
9.6% (7.7–11.5%)
1.28 (0.98–1.68)
Large (15–19mm)
15.3% (10.5–20.1%)
2.04 (1.41–2.96)
Very large (≥20mm)
18% (10.7–25.1%)
2.4 (1.53–3.75)
The risk in women with multiple excisions appeared to be explained by the total depth of tissue removed.
There was no difference in risk of preterm delivery by type of excision once the results were adjusted for the depth of excision.
Restricting analysis to women with spontaneous preterm births showed a similar association.
The results suggest that small excisions (<15mm in depth) do not increase the risk of preterm delivery compared to a punch
biopsy.
Future fertility and miscarriage
A recent systematic review and meta-analysis has looked at whether treatment for CIN has any effect on future fertility or pregnancy
outcomes (BMJ 2014;349:g6192). The meta-analysis found that:
Treatment for CIN did not adversely affect future chances of conception.
The rates of miscarriage overall and in the first trimester were similar between treated and untreated women.
There was a significantly increased risk of second trimester miscarriage in treated women (1.6%) compared to untreated
women (0.4%), RR 2.6 (CI 1.45–4.67).
The eight studies that reported on second trimester miscarriage included a total of 16 558 treated women.
Most of these were in one large study in which patients had treatment with knife, laser or LLETZ. It was not possible to
stratify results based on the treatment used.
There was no significant inter-study heterogeneity in results supporting the validity of the overall finding.
Due to the observational nature of the studies, it is possible that the observed increased risk is due to other confounding
factors. For example, the authors suggest it is possible that immunological variations may make some women more
vulnerable to ascending vaginal infection and persistent HPV which would influence the risk of both CIN and preterm birth.
What does this mean in practice?
We can reassure women that, based on current evidence, their future fertility should not be affected by treatment for CIN.
Treatment for CIN is associated with an increased risk of 2nd trimester miscarriage.
There is an increased risk of preterm delivery after treatment for CIN. The depth of excision is important. We can reassure
women that an excision <15mm in depth is not associated with increased risk compared to a punch biopsy.
We should still refer women who have had cervical treatment for shared obstetric care so that cervical length can be
assessed for the individual woman and intervention considered.
Cervical or vaginal cancer risk in those previously treated for CIN
This large cohort study in Sweden followed all women previously diagnosed and treated for CIN3 to see if they developed cervical or
vaginal cancer (BMJ 2014;348:f7361). They found that:
Women with a history of CIN3 had an increased risk of developing cervical or vaginal cancer after the age of 60. The
background risk for cervical or vaginal cancer in women aged 70 was 10.2 per 100,000 women, but in those with previous
CIN3 the incidence was >70 per 100,000 women.
The risk increased with increasing age, being greatest in women treated for CIN3 in later life.
I think the message from this study is to be more alert to any gynaecological symptoms in women with a history
of CIN. Rather than just reaching for the clotrimazole prescription, have a look!
Cervical cancer screening - Summary
All women in the UK aged 25–64y are offered cervical screening by cytology.
More than 99% of cervical cancer is due to high risk HPV infection.
HPV triage for borderline or mild dyskaryosis is now a routine part of the screening
programme.
HPV 'test of cure' is part of follow-up for all women who have colposcopy treatment and
determines future screening interval and whether further colposcopy is needed.
Primary HPV testing to determine the need for cytology is being piloted in the UK.
Under 25 year olds are not offered screening because research has shown that it is not
effective in this age group and the harms outweigh the benefits.
Lesbian and bisexual women should be encouraged to attend screening because they
are at risk of HPV infection and attendance is typically 10 times lower than other women.
Treatment of CIN increases the risk of second trimester miscarriage and preterm
delivery and women who have had cervical treatment should be referred for shared
obstetric care.
Be aware of the considerable increased risk of cervical or vaginal cancer in older women
with a history of CIN3.
The Red Whale | GP Update Course – The one day course for GPs and GP Registrars, by GPs.
(We think it’s the most fun you can have earning CPD credits!)
The GP Update Course brings you up to speed with all the latest issues, guidelines and research. Run entirely by
GPs, we trawl through all the literature and pick out the things that are relevant to everyday general practice and
present it in a lively and entertaining way.
If you book on this course you will get:
•
The GP Update Handbook - comprehensive and fully referenced covering all the most recent research and
guidelines pertinent to primary care but interpreted for real life general practice.
•
12 months FREE access to www.gpcpd.com - our revitalised, fully searchable and most comprehensive online
reference source ever!
•
The GP Update Handbook Online - twice the contents of the paper version and you can earn credits for every
page you read at the click of a button!
•
Focused Learning Activities - activity ideas that amount to much more than 50 hours of CPD credits and more
importantly that will really impact on patient care - not just a series of tick box exercises.
The GP Update Course is completely free from pharmaceutical company sponsorship. This means that you can be
totally reassured that there will be no biasing of the information presented to you and that there will be no reps
there on the day!
Summer 2015 dates:
Bristol - Wednesday May 13
Exeter - Thursday May 14
London - Friday May 15
London - Saturday May 16
Belfast - Wednesday May 20
Newcastle - Wednesday June 3
Sheffield - Thursday June 4
Manchester - Friday June 5
Birmingham - Saturday June 6
Norwich - Wednesday June 10
London - Thursday June 11
Chelmsford - Wednesday June 24
GPs £195, ST1, 2 & 3 and Nurses £150. For more details see www.gp-update.co.uk
We make every effort to ensure the information in these pages is accurate and correct at the date of
publication, but it is of necessity of a brief and general nature, and this should not replace your own good
clinical judgement, or be regarded as a substitute for taking professional advice in appropriate circumstances.
In particular check drug doses, side effects and interactions with the British National Formulary. Save insofar as
any such liability cannot be excluded at law, we do not accept any liability for loss of any type caused by
reliance on the information in these pages.
GP Update Limited