With 493,000 new cases and 273,000 deaths annually,1 cervical cancer represents the second most common cancer in women worldwide and is a major public health issue. In allowing the detection and treatment of its precursors, screening programs in developed countries based on cervical cytology have facilitated a dramatic reduction in the burden of cervical cancer. However, the effectiveness of the cervical cytology is limited, and colposcopy-directed cervical biopsies remain the “gold standard” for the diagnosis of cervical premalignancy. With specificity estimated between 48% and 69% and a sensitivity from 85% to 96%,2 the diagnostic accuracy of colposcopy is also limited and depends crucially on the ability to fully visualize the transformation zone,3 specifically its upper limit, the squamocolumnar junction. To improve the visibility of the transformation zone and, therefore, the accuracy of colposcopy, authors such as Cartier4 have recommended performing colposcopy during the second week of the cycle (ie, follicular phase), and patients whose transformation zone is not fully visible at the initial colposcopic examination are advised by some colposcopists to reschedule for the follicular phase of the cycle. Such practice has not been supported by objective evidence.
This study attempted to estimate if the follicular phase of cycle was associated with a better rate of visibility of the upper limit of the transformation zone. We also examined the influence of factors such as age, parity, smoking, and combined oral contraception on the transformation zone visibility.
MATERIALS AND METHODS
In this retrospective study, data from all patients who attended our colposcopy clinic between November 2003 to March 2007 and stored in our computerized system (MediScan; Polartechnics Ltd, New Malden, Surrey, UK) were examined. Only data from the first attendance were considered. Patients whose date of last menstrual period was not reported were excluded from the study. Because they are amenorrheic, we excluded women who were pregnant or postmenopausal. Patients with amenorrhea of more than 60 days were also excluded.
All colposcopies were performed using standard techniques with one of our two colposcopes (150 FC; Carl Zeiss, Oberkochen, Germany), irrespective of the menstrual cycle time. Colposcopies were performed by one of our three colposcopists who are certified by the British Society for Colposcopy and Cervical Pathology or by trainees under supervision. Specific attention was paid to reporting the transformation zone type according to the classification system adopted by the International Federation of Cervical Pathology and Colposcopy.5–7 If necessary, an endocervical speculum was used to allow the visibility of the upper limit of the transformation zone. The transformation zone was classified as type 1 when completely ectocervical and fully visible. Patients whose transformation zone upper limit presented some endocervical component but could be fully seen were considered as type 2. If necessary, an endocervical speculum was used to allow the visibility of the upper limit of the transformation zone. When not fully visible due to endocervical extends, the transformation zone was classified as type 3 and colposcopy was considered unsatisfactory. As a routine procedure, the colposcopist in charge of each patient completed a questionnaire to provide data on epidemiologic and clinical features. Age, parity, use of combined oral contraceptive method, smoking status, and the menstrual cycle date were recorded.
We examined the rate of type 1 transformation zone according to the cycle time at the first colposcopic visit. We further assessed the influence of age, parity, smoking status, and use of combined oral contraceptive method. Data were analyzed using the χ2 and logistic regression. A multivariable analysis was conducted to identify variables having an independent effect on the type of the transformation zone. P<.05 was considered significant.
From November 2003 to March 2007, 1,695 patients attended our colposcopy clinic. A total of 347 patients were excluded from the study because the last menstrual period was not recorded, 52 because they were postmenopausal, 12 because they were pregnant at the time of their first attendance, and 36 because of amenorrhea of more than 60 days. Finally, 1,248 patients were included. Patients' characteristics are summarized in Table 1. None of our patients were referred after an initial colposcopic evaluation. There were no differences between patients' characteristics in the follicular phase and patients in the remainder of the menstrual cycle.
Type 1 and 2 transformation zone was observed in 71.2% and 26.1% of patients, respectively (Table 2). Colposcopy was unsatisfactory due to type 3 transformation zone in 2.7% of patients. Only 8.4% of patients had a history of previous treatment elsewhere (Table 1). Among these patients, unsatisfactory colposcopy rate was 11.3%, with a type 1 and 2 transformation zone rate of 51.5% and 37.1%, respectively. No significant difference was found between the rate of type 1, 2, or 3 transformation zone observed in patients who were examined during the second week of their menstrual cycle and other times (P=.581) (Table 2). Table 3 shows the percentage of women with a type 1 transformation zone at each week of the menstrual cycle. This rate was the lowest in women examined during the first 7 days of the cycle, and the highest in women on day 22 or later. Compared with women in the first week of their menstrual cycle, those on the 22nd day or later were significantly more likely to have a type 1 transformation zone at colposcopic examination (OR 1.6, 95% CI 1.1–2.2, P=.029, logistic regression using day 1 to 7 as referent). Because 8% of participants were seen at day 29 or more into their menstrual cycle, we repeated the above analysis omitting those whose cycle was longer than 28 days. The results were similar, with a significant increase in likelihood of type 1 transformation zone in the last 7 days of the cycle (OR 1.8, 95% CI 1.2–2.6, P=.006).
In addition to the timing of colposcopy, we examined the effect of age, combined oral contraception, parity, and smoking status on the probability of finding a type 1 transformation zone. This probability declined with the age, with an odds ratio of 0.59 associated with a 10-year increment in age (P<.001) and with smoking status (OR 0.55 for current smokers versus nonsmokers, P<.001). The use of combined oral contraception significantly increased the likelihood of finding a type 1 transformation zone (OR 2.7, P<.001). Parous women had a lower probability of having a type 1 transformation zone than nulliparous women (OR 0.47, P<.001), which further declined with increasing parity (P<.001).
In multivariable logistic regression analysis, age, current smoking, and parity were still associated with a lower probability of observing a type 1 transformation zone, whereas use of a combined oral contraception was associated with an increased probability (Table 4). Adjusted for these factors, the increased probability of type 1 transformation zone in week 4 of the cycle was no longer statistically significant.
Table 5 illustrates the combined effects of these factors by showing the probability for a 30-year-old woman having a type 1 transformation zone, calculated by logistic regression. The highest probability (86%) found was for a nonsmoking nulliparous women taking combined oral contraception (95% CI 0.82–0.89), whereas only 54% (95% CI 0.48–0.61) of parous smokers not using combined oral contraception are likely to achieve a type 1 transformation zone.
The main impetus for proposing the International Federation of Cervical Pathology and Colposcopy transformation zone classification system5–7 was to introduce a standard nomenclature whereby transformation zone description and treatment could be reliably compared. The literature describing treatment of cervical intraepithelial neoplasia did not previously have a standard transformation zone nomenclature and had reference to excision, cone biopsy, or conization, inter alia. These terms mean different things to different researchers. Furthermore, the excision of a small type 1 transformation zone is considerably less than that of a large type 3. The transformation zone classification system of the International Federation of Cervical Pathology and Colposcopy aspires to clarify and standardize nomenclature. It is hoped that the use of a standard transformation zone type classification might reduce the subjective nature that is inherent in colposcopic examination and that using the system may result in more valid comparison of therapeutic interventions.
We reviewed our data to determine what effect the time of cycle and other variables had upon the type of transformation zone reported in women attending our colposcopy clinic for the first time. A satisfactory colposcopic examination depends crucially on complete visualization of the transformation zone,3 in particular its upper limit, that is, the squamocolumnar junction, where most dysplastic epithelial lesions are located.8,9 In this work, we report a large series of colposcopic examinations detailing the factors that influence the determination of transformation zone type.
Colposcopic examination in the follicular phase was not associated with better chance of fully visualizing the upper transformation zone limit. Surprisingly, the only evident effect was that the last week of the cycle was associated with a higher chance to observe a fully ectocervical squamocolumnar junction (ie, type 1 transformation zone). Indeed, using multivariable logistic regression analysis, we determined that this finding was no longer statistically significant (Table 5). Our data reveals no advantage to timing a colposcopy examination and contradicts traditional teaching and belief that the squamocolumnar junction shifts outwards during the menstrual cycle. It is accepted that unopposed estrogen affects cervical morphology, and this may be clinically useful in women having various gynecological interventions, such as intrauterine coil insertion, colposcopy, or hysteroscopy. Unfortunately, the effects do not appear to move the squamocolumnar junction ectocervically at colposcopy. Scheduling appointments for a particular time of the cycle may, therefore, increase the risk of nonattendance and is inconvenient for patients and clinic staff. It appears to have no role in routine colposcopy practice.
An unexpected finding of our study is the association of smoking with involution or retraction of the squamocolumnar junction into the endocervical canal. Smoking is only already known as a risk factor for cervical intraepithelial neoplasia and squamous cell carcinoma.10 Our data reveals a statistically negative effect of smoking on the likelihood of finding a type 1 transformation zone. Smoking patients are, therefore, exposed to a potential risk of colposcopic complication because of the difficulties in evaluating and treating endocervical transformation zones.
Our results show a significant influence of parity on the transformation zone visibility. Accepted teaching is that previous pregnancy increases the rate of squamocolumnar junction visibility. Reporting the colposcopic examination of 174 pregnant women, Ostergaard reported that colposcopic examination revealed an 80% likelihood of being able to see the squamocolumnar junction in parous women who underwent colposcopy during pregnancy.11 He reasoned that the predominant effect of previous pregnancy is to stimulate the endocervical migration of this junction. Our data would support this view.
The effects of age12–15 and oral contraception15,16 on squamocolumnar junction visibility has also been described and are confirmed by our data. Using a 10-year interval, we found a significant negative association between age and squamocolumnar junction visibility. Independent of the menopause per se, age reduces the chance for a patient to achieve a type 1 transformation zone.
The low rate of unsatisfactory colposcopic examination due to type 3 transformation zone observed in our study was a surprise to us and is worthy of comment. Previous cervical treatment, such as cervical conization or large loop excision of the transformation zone, is accepted as a major risk factor for unsatisfactory colposcopy due to cervical stenosis and endocervical squamocolumnar junction.17 Because the inclusion criterion was that the patient should be a first attendee, only 8.4% of patients had a history of previous treatment elsewhere (Table 1). This characteristic of our study population could, therefore, explain this situation. Moreover, because our primary ambition was to relate the time of cycle to the transformation zone type, we did not include in our study menopausal women, who are known to have higher probability of unsatisfactory colposcopy.15,18
Our work might be limited by the inherent biases of all retrospective analyses. However, we believe the specific design of our study dramatically reduced those biases. We considered only computerized stored data from the first attendance at our colposcopy clinic without any follow-up data. Moreover, the retrospective analysis allowed a reduction in the risk of observation bias. Colposcopists, indeed, even if not blinded to the timing of colposcopy, because they were not told about our interest in the influence of timing, were probably not influenced in the way they reported the transformation zone type. Finally, the main limitation of our study is that, of the 1,647 patients who could have been included, 347 (21.1%) were excluded because the date of the last menstrual period was missing.
Our study confirms the role of age and combined oral contraception on the position and visibility of the upper limit of the transformation zone. It also highlights the influence of smoking and parity. A completely ectocervical and fully visible transformation zone (type 1) is, therefore, more likely to be observed in a young nulliparous, nonsmoking woman using an combined oral contraception. A specific timing for colposcopy does not improve the visibility of the transformation zone and should not, therefore, be recommended.
1. Sankaranarayanan R, Ferlay J. Worldwide burden of gynaecological cancer: the size of the problem. Best Pract Res Clin Obstet Gynaecol 2006;20:207–25.
2. Mitchell MF, Schottenfeld D, Tortolero-Luna G, Cantor SB, Richards-Kortum R. Colposcopy for the diagnosis of squamous intraepithelial lesions: a meta-analysis. Obstet Gynecol 1998;91:626–31.
3. Hopman EH, Kenemans P, Helmerhorst TJ. Positive predictive rate of colposcopic examination of the cervix uteri: an overview of literature. Obstet Gynecol Surv 1998;53:97–106.
4. Cartier R, Cartier I. Practical colposcopy. 3rd ed. Paris, France: Laboratoire Cartier; 1993.
5. Bm RH, Woo W, Walker P. Internal Federation for Cervical Pathology and Colposcopy nomenclature and treatment for cervical preinvasive disease. J Low Genit Tract Dis 2006;10:51–4.
6. Prendiville W. Recent innovations in colposcopy practice. Best Pract Res Clin Obstet Gynaecol 2005;19:779–92.
7. Zlatkov V, Kostova P. New revision of the international colposcopic terminology: Barcelona–2002 [in Bulgarian]. Akush Ginekol (Sofia) 2003;42:44–6.
8. Thornton WN Jr, Fox CH, Smith DE. The relationship of the squamocolumnar junction and the endocervical glands to the site of origin of carcinoma of the cervix. Am J Obstet Gynecol 1959;78:1060–73.
9. Marsh M. Original site of cervical carcinoma; topographical relationship of carcinoma of the cervix to the external os and to the squamocolumnar junction. Obstet Gynecol 1956;7:444–52.
10. Comparison of risk factors for invasive squamous cell carcinoma and adenocarcinoma of the cervix: collaborative reanalysis of individual data on 8,097 women with squamous cell carcinoma and 1,374 women with adenocarcinoma from 12 epidemiological studies [published erratum appears in Int J Cancer 2007;120:2525]. Int J Cancer 2007;120:885–91.
11. Ostergard DR. The effect of pregnancy on the cervical squamocolumnar junction in patients with abnormal cervical cytology. Am J Obstet Gynecol 1979;134:759–60.
12. Singer A. The uterine cervix from adolescence to the menopause. Br J Obstet Gynaecol 1975;82:81–99.
13. Wetrich DW. An analysis of the factors involved in the colposcopic evaluation of 2194 patients with abnormal Papanicolaou smears. Am J Obstet Gynecol 1986;154:1339–49.
14. Costa S, Nuzzo MD, Rubino A, Rambelli V, Marinelli M, Santini D, et al. Independent determinants of inaccuracy of colposcopically directed punch biopsy of the cervix. Gynecol Oncol 2003;90:57–63.
15. Saunders N, Anderson D, Sheridan E, Gilbert L, Sharp F. Endoscopic localization of the squamocolumnar junction before cervical cone biopsy in 284 patients. Cancer 1990;65:1312–7.
16. Milam MR, Pollock JW, Nick AM, Scott T, Jones HW 3rd. The effect of hormonal contraception on the adequacy of colposcopic examination of the cervix. Am J Obstet Gynecol 2005;192:1368–9.
17. Baldauf JJ, Dreyfus M, Wertz JP, Cuenin C, Ritter J, Philippe E. Consequences and treatment of cervical stenoses after laser conization or loop electrosurgical excision [in French]. J Gynecol Obstet Biol Reprod 1997;26:64–70.
18. Toplis PJ, Casemore V, Hallam N, Charnock M. Evaluation of colposcopy in the postmenopausal woman. Br J Obstet Gynaecol 1986;93:843–51.