HPV primary screening is an attractive option to health service managers because the results are not subject to inter-observer variation.  Also, in the long run, HPV screening may be cheaper than cytology because the costs of personnel, training, update and quality control may be lower.

Nevertheless, HPV testing also requires equipment, reagents, training, quality control and accreditation - and sensitivity and specificity of different HPV tests is known to vary (Cubie & Cuschieri 2013).

Meta-analyses of HPV primary screening of women over 30 years of age in several countries have indicated higher sensitivity for CIN2+ and CIN3+ compared with cytological screening and it makes sense to have the most sensitive test as the primary test (Arbyn et al. 2012).

Data from ARTISTIC in the UK, which was carried out in laboratories with rigorous quality control of cytology, there was no advantage of HPV testing over cytology in terms of detection of CIN3+ after two rounds of 3-yearly screening (Kitchener et al. 2009). 

An advantage in favour of HPV screening was demonstrated in terms of detection of CIN2+ (1.4% vs. 0.9%) after the third round of screening, which might allow the screening interval to be increased to 5-6 years.  The cumulative rate of CIN2+ in women over 50 years of age was 8% with HPV screening compared with 2% with cytology.

The costs of HPV testing remain too high for low-resource settings but low-cost tests are under development (Qiao et al. 2008).  Self-testing for HPV has been put into effect in Mexico (Lazcano-Ponce et al. 2011). 

A new screening test could improve screening uptake in England, where the NHSCSP is piloting primary HPV testing in women aged 25-64 years (Kitchener 2015).  Including women aged 25-29 years of age in this new test could improve screening coverage, which has been gradually falling and was 62% in 2013-14 (HSCIC 2014).

When prevalence of high-grade CIN and cancer falls with vaccination primary HPV testing is likely to be necessary with or without co-testing or cytology triage (see section on vaccination below).

Challenges of HPV primary screening

The main problem with HPV primary screening is its low specificity and PPV.  False positives relate to reversible HPV infection because existing tests measure HPV infection rather than HPV integration into the host genome.  Even APTIMA, which is an RNA test that aims to detect integrated HPV, has specificity little more than 50% (Arbyn et al. 2013). 

Cytology triage of HPV+ women may be used to decide which women need colposcopy, which places a challenge on cytologists to provide an accurate result with as few false negatives and false positives as possible.

Re-testing of HPV-positive / cytology-negative women after one year gains the advantage of the higher sensitivity of HPV testing and tests persistence of HPV which is essential for progression to CIN2+. Otherwise the sensitivity returns to that of cytology. 

Management of HPV+ women without CIN2+ at colposcopy is the main problem with primary HPV screening unless a test is developed that only measures integrated HPV. 

One solution to this problem is to integrate p16 with HPV testing as has been done in Italy by Carozzi et al. 2013.  In this longitudinal study p16 positivity significantly increased the detection of CIN2+ within 3 years in HPV+ women in whom CIN2+ was not initially found at colposcopy.  Double-negative (HPV and p16) women could safely be returned to routine screening.

Sensitivity of HPV testing is around 85-95%.  Jastania et al. (2006) reported 4.5% false negative CIN2-3 cases, which they quote as the lower end of previous reports ranging from 3.7% to 18.2%.  Although sensitivity is at least as high as cytology, cancers arising after false negative HPV tests may be unexpected and difficult to explain.

False negative HPV tests could be avoided by co-testing with cytology, which would be expensive unless confined to early rounds of screening when prevalence of precancerous changes is high.  A study by Katki et al. (2011) pointed out that 3- or 5-yearly primary HPV screening would be safe for women who were double-negative for cytology and HPV.

A recent study of multiple practices in the US provides further evidence of the greater sensitivity of co-testing compared with cytology or HPV testing alone as a primary screening test in women above 30 years of age (Blatt et al 2015).

The effect of HPV vaccination on cervical cancer screening

National vaccination programmes

In many countries vaccination programmes have already been introduced for 12-13-year-old girls.  Some have introduced vaccination for boys as well as girls.  Girls up to 18 years of age or older have been offered ‘catch-up’ vaccination if they were above 12-13 when vaccination was introduced.

Until recently two vaccines were available and recommended by the European Centre for Disease Control (ECDC 2014); a third was approved by the US Food and Drug Administration in November 2014 and has been recommended by the European Committee for Medicinal Products for Human Use, which is the final step before a license is granted by the European Commission (Sanofi Pasteur MSD (2015).

• Bivalent vaccine against HPV 16 and 18 (Cervarix)
• Quadrivalent vaccine against HPV 10, 11, 16 and 18 (Gardasil)
• Multivalent vaccine against HPV 10, 11, 16, 18, 31, 33, 35, 52 and 58 (Gardasil-9)

All three are active against the commonest oncogenic HPV types in Europe while the third targets a substantially wider group of viruses; Gardasil and Gardasil-9 also prevent genital warts, which are benign and not at risk for cancer.

Effect of vaccination on cervical abnormalities

Vaccination of 12-13-year-old girls was introduced in England in 2008 with an uptake of 80%.  As screening does not start until age 25 years of age in England, the effect on prevalence of cytological abnormalities will not be seen until about 2020.  However, the effect on prevalence of infection with HPV 16 and 18  in women aged 16-24 years has already been reported (Mesher et al. 2013).

Vaccination was introduced for all women aged 12-26 years in Australia between 2007 and 2009 and a decline in cytological abnormalities was reported as early as 2011 where screening starts earlier (Brotherton et al. 2011).  The latest results show prevalence of high-grade cytological abnormalities has fallen by approximately 50% and 33% in women aged below 20 and 20-24 years respectively while prevalence has increased slightly at age 25-29 (Brotherton et al. 2015).

Effect of declining prevalence of abnormalities on accuracy

Declining prevalence of abnormalities could have a deleterious effect on sensitivity as well as PPV.  For different reasons, false negatives as well as false positives could be relatively more frequent. [Link to Chapter 5 - Principles of screening and measurement of accuracy] 

• Lower prevalence of high-grade abnormalities would make their detection by cytological screening more difficult and less interesting for cytotechnologists.  Prevalence of abnormalities has been shown to affect sensitivity of cytological screening: Evans et al. (2010) reported false negative rates of low-prevalence samples to be nearly double those of high-prevalence samples.
• PPV is known to be related to prevalence of a high-grade outcome as the ‘background noise’ of false positives tends to outnumber true positives.  This would to some extent be mitigated by vaccination preventing low-grade and borderline abnormalities caused by high-risk HPV, but lesions caused by low-risk HPV would remain (depending on the type of vaccine).

Cervical screening of unvaccinated women and non-16/18 HPV

Cervical cancer screening will still be needed in vaccinated populations for HPV types other than 16 and 18, until the multivalent vaccines are in widespread use, and for unvaccinated women.

As mentioned above, primary HPV testing is likely to be introduced in vaccinated populations and will be subject to the advantages and challenges already dealt with. 

Cervical cytology will also have an important role in diagnostic triage after primary HPV screening.