Although true preinvasive lesions are CIN 3s, and there is a relatively low agreement between pathologists on CIN 2 diagnosis, we preferred for comparison CIN 2 + lesions instead of CIN 3+.  Besides, there is more than 25% risk of underdiagnosing CIN3’s as CIN 2’s and if we ignore CIN 2 lesions, about one quarter of preinvasive lesions would have been missed in the comparison. 
In our analysis, cytological examination contributes 5.4% to HPV screening to detect CIN2 + lesions. Moreover, 6 invasive cancers were detected (0.7% of the study population) with cytological evaluation which otherwise would have been missed if we had opted for only HPV based screening. Excluding colposcopy indications for HPV 16, 18 and double positivity of other hrHPVs one year apart, cytological examination did not significantly increase CIN 2 + detection rate of single HPV positivity other than 16 and 18. HPV-negative cervical cancer accounts for approximately 5% of all cervical cancer cases. A study from USA enrolling more than one million women, has found that 1,8% and 0.1% of HPV negative patients with cytological abnormality had CIN 2+, and invasive cancers respectively. The difference in our study may be attributed that Turkey has recently started HPV based screening program for women above 30 years old, and many older population has not been screened before. The average age of our study population is 45, whereas the screening program in abovementioned study country starts at 25 (no average age of the study is mentioned), Older patient age at diagnosis of cervical cancer is associated with decreased positivity of HPV testing.[19–21] On the other hand, some of the HPV negative CIN 2 + lesions may be “false-negative” that may be caused by non-high risk HPV viruses like 6, 11, 42 and 70 which may be higher in screened population, especially when the vaccination rates are very low as in our study population.
In patients with only a cytological abnormality, cytological abnormality with single other hrHPV, or negative cytology and double other hrHPV positivity (without HPV 16 and 18); we detected 159 (19%) CIN 2 + and 64 (7.6%) CIN 3 + lesions. In other words, if we had been opted HPV-based screening for only HPV 16 and 18 without cytology, 19% of all CIN 2 + lesions and 7.6% of all CIN 3 + lesions would have been missed in our study group. In a nation-wide analysis in Turkey, among 521 CIN3 + patients, authors have concluded that HPV-based screening for only HPV 16 and 18 without cytology could have been missed 12,7% of CIN3 + patients. Although the incidence of HPV 33 is low in our study population (%3.4 in all HPV infections), 23% of HPV 33 infections is associated with CIN 3 and 20% with CIN 2; whereas these associations for HPV 16 are 16% for CIN 2 and 16% for CIN 3 (Table 4). The low numbers of HPV 33 infections and relatively high percentage of CIN 2 + lesions for HPV 33 may be due to the reference algorithm that more persistent (two samples one year apart) HPV 33 infections are referred for colposcopy, in comparison with HPV 16 and 18, where single positivity is enough for colposcopic evaluation.
Viral loads of HPV 33 seem to be at least 7-fold higher than HPV 16 and 18 in a study of population-based cervical screening (POBOSCAM ) trial’s sub-population.  In a study from Sweden using a different real-time PCR procedure than the previous one, also demonstrated a markedly higher viral load of HPV 33 in comparison to HPV 16 and 18.  Increased viral loads may be associated with an increases risk of pre-invasive disease, and decreased viral loads seem to be a consequence of an effective clearance by immune system.  A study from Southwestern China demonstrated that HPV 16, 58 and 33 were most prevalent subtypes in CIN 2 and CIN 3, and the viral loads of these three strains were corelated with the severity of the lesions.  Maybe, the lower incidence of HPV 33 masks its potential for pre-cancer lesions.
For Turkish screening program, extensive HPV genotyping is done since 2014 for mainly epidemiologic mapping.  Therefore, our findings about HPV 33 predicting CIN 2 + lesions better than HPV 18 in our study group may contribute to further studies for planning vaccination and screening policies such as considering HPV 33 as an immediate colposcopy indication or preferring nonavalent vaccine in specific populations. The prevalence of the most common HPV types nationwide was as; 16, 51, 31, 52 and 18 respectively ; and 16, 18, 52, 51 and 31 in our subpopulation respectively, which is in harmony with nationwide results. The actual prevalence of other hrHPV types is expected to be much higher, because the only referred individuals were the ones that at least one year the infection persisted.
Local differences for prevalence and maybe genetic differences among individuals effecting virulence of specific subtypes of HPV should be considered. Possible mechanisms related to the immune response facilitates faster elimination of infection by certain HPV types in specific populations.  This necessitates creating population-based policies about HPV triage tests or combination tests.
HPV 16 was the most common HPV type referred and was the most common type in co-infection with more than one type of hrHPV in our study. We found that HPV 16 tends to be more frequent in younger age, which is concordant with the literature. [27, 28] This high-lightens the prophylactic effect of HPV 16 vaccines in younger population.
We demonstrated that increasing total number of HPV subtypes did not correlate with the occurrence of CIN 2 + lesions. There are conflicting studies investigating whether multiple HPV infections were associated with more CIN 2 + lesions compared to single HPV type infections,[29–31] According to our results, there seems a lack of a synergistic effect of simultaneous multiple HPV infections for high grade lesions.
A strength of this study is all colposcopies were performed by the same certified gynecologic oncology specialist, completed 3 years of sub-specialization program of The Health Ministry of Turkey and all the data are recorded in text. There was a low diagnostic excision rate (3.1%) and acceptable mean number of biopsies taken (2.14). A limitation of the study may be that there is no external reviewer for pathology results.
Unfortunately, our data for other hrHPV is confined to double positivity at least one year apart or concomitant with a cytologic abnormality, and we could not calculate CIN 2 + lesion ratio in HPV 33 positive patients in a single examination without cytological abnormality due to nationwide referring algorithm. HPV 33 seems more tumorigenic than other hrHPV types and in this regard, future prospective studies are needed questioning whether a single HPV 33 positivity is feasible for a direct colposcopy indication.