SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Note
  10. References

Objective  To compare the adequacy of cervical cytology sampling by a newly designed cervical cell sampler (the implement) with a commonly used, extended tip cervical scraper, namely the Jordan's spatula, by assessing the quality of cervical smears obtained.

Design  Prospective randomised controlled trial, with the approval of the local Ethics Committee after informing the Medical Devices Agency in the UK.

Setting  Colposcopy Clinic in a District General NHS Trust Hospital in London.

Population  Women attending the Colposcopy Clinic and having a good command of the English language (to avoid any misinterpretation of the information and enabling them to give informed consent) were invited to take part. Two hundred and twenty patients were recruited.

Methods  The women were randomised into two groups, each of them having two smears at the same time, half of them having their first smear with the Jordan's spatula and the other half having the first smear with the new implement.

Main outcome measure  Good quality smears, assessed by evidence of effective sampling of the transformation zone, including immature metaplastic cells and also endocervical cells.

Results  Fifty-four percent of smears taken by the new implement showed good quality smears, compared with 37% taken by Jordan's spatula, a statistically significant result with P value <0.001.

Conclusion  A larger number of good quality smears, judged by evidence of sampling of transformation zone and presence of endocervical cells, were obtained with the new implement. The difference was statistically significant. This would enhance the ability to detect dyskaryosis and increase adequate smear rates, reducing unnecessary recall of patients for repeat smears.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Note
  10. References

Despite the success of the cervical cancer screening programme in reducing cancer of the cervix by at least 20% and mortality from it by >40% since its introduction in the UK in 1988,1 there still remains room for improvement. It is estimated that between 16% and 36% of smears are false negative in CIN.2 This could be due to either cytological misinterpretation or, more likely, sampling error, due to insufficient number of representative cells obtained or failure of the device/operator to scrape/brush the entire surface of the cervix.

Several sampling devices remain in use. One of the major drawbacks common to all of them is the inability to properly sample cells from the transformation zone, endocervix and the whole of the ectocervix, irrespective of size, shape and position of the cervix. To overcome this drawback, a new implement has been developed and its efficacy compared with a commonly used extended tip device, namely the Jordan's spatula (Fig. 1). Improved cervical sampling will lead to better quality smears and hopefully decrease both false negative and inadequate smear rates.

image

Figure 1. Jordan's spatula.

Download figure to PowerPoint

The implement is a flat spatula with a heart-shaped loop at either end (Fig. 2). One end, in addition, has an extended tip/prong (Fig. 3), while the other end is without a prong (Fig. 4). Both loops have a bristle-carrying side that acts as an endocervical, transformation zone and ectocervical brush. The other side of the loop acts as a scraper of the same area. The device thus acts as a scraper and a brush at the same time, providing two different smears in one. The sampling surface has got microscopic transverse grooves acting as cellular reservoirs. Moreover, gentle axial pressure on the spatula widens the diameter of the sampling surface, up to 6 cm if required, which is useful in case of wider lesions and satellite lesions on the ectocervix (Fig. 5). Each side of the loop acts independently of the other to conform to the shape of any cervix (Fig. 6), irrespective of its position (Fig. 7), size or shape. Also, there is a mechanism, including a ball on one side with an opposing socket on the other side to effect locking and reduce the size of the loop (Fig. 8). The prong of the spatula can be manually angled to negotiate awkwardly pointing cervices (Fig. 9). A notch is present proximally on each side of the loop to enable it to be cut off, if required.

image

Figure 2. The new implement.

Download figure to PowerPoint

image

Figure 3. The new implement: magnified view of the pronged end.

Download figure to PowerPoint

image

Figure 4. The new implement: non-pronged end for vault smears.

Download figure to PowerPoint

image

Figure 5. Gentle axial pressure widens the diameter of the sampling surface upto 6 cm.

Download figure to PowerPoint

image

Figure 6. The loop conforms to the shape of the cervix.

Download figure to PowerPoint

image

Figure 7. The loop conforms to the position of the cervix.

Download figure to PowerPoint

image

Figure 8. The new implement showing the lock closed.

Download figure to PowerPoint

image

Figure 9. Prong can be manually angled.

Download figure to PowerPoint

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Note
  10. References

Two hundred and twenty patients attending the Colposcopy Clinic during the period from January to June 2002 were recruited and were randomised into two groups, each of them having two smears at the same time. Picking their own numbers blindly from a jar randomised patients into two groups. Those who had even numbers were allocated to have the first smear taken by Jordan's spatula and the second by the new implement, while those who had odd numbers were allocated to have their first smear taken by the new implement and the second by the Jordan's spatula. Both the slides from each patient, numbered only as 1 and 2, were then sent to the cytopathologist. The cytopathologist was thus unaware as to which device was used for the smear. To avoid sampling bias, the smears were taken by five different people and the following had to be reported by the smear takers for each patient.

 New implementJordan's spatula
Ease of use of spatula/implement by the person taking the smear
Very easy11
Easy22
Difficult33
 
Discomfort to patient
Comfortable11
Uncomfortable22
Painful33
 
Bleeding from taking the smear
None11
Slight22
Moderate/heavy33

The cytopathologist then had to report the following on each slide:

 Slide 1Slide 2
Assessment of overall cellularity (i.e. cells occupying percentage of slide)
Good11
Fair22
Poor33
 
Blood staining
Nil11
Slight22
Moderate/heavy33
 
Evidence of transformation zone sampling
Endocervical cellsEE
Metaplastic cellsMM
Cervical mucusCC
 
Bethesda Criteria of smear quality as evidence of transformation zone sampling
0–1 group of 5 cells00
2–5 groups++
5–10++++
>10++++++

The data were collated and analysed over a period of six months. The endpoints quantified were good quality smears assessed by evidence of transformation zone sampling, including metaplastic and endocervical cells and the presence of cervical mucus (E, M and C, respectively) according to the BSCC (British Society for Clinical Cytology) criteria. The slides were also judged by the Bethesda System,3 using the presence of more than five groups per clusters of cells (++ and +++) as evidence of adequate/good quality smears.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Note
  10. References

The following results (%) were achieved:

 New implementJordan's spatula
Ease of use of spatula/implement by person taking smear
Very easy6787.3
Easy3012.2
Difficult30.5
 
Discomfort to patient
Comfortable4780
Uncomfortable5320
Painful00
 
Bleeding from taking the smear
None6182
Slight3818
Moderate/heavy10

The cytological results (%) of the smears were as follows:

 New implementJordan's spatula
Overall cellularity
Good6054
Fair3742
Poor34
 
Blood staining
Nil7269
Slight2121
Moderate/heavy710
 
Evidence of transformation zone sampling
Endocervical cells (E)6760
Metaplastic cells (M)7063
Cervical mucus (C)8979
 
Evidence of transformation zone sampling (Bethesda Criteria)
0–1 group of cells (O)2532
2–5 (+)1729
5–10 (++)2623
>10 (+++)3216

Two hundred and twenty patients (one patient had a double cervix) entered the study and had two smears, one using the standard Jordan's spatula and the other using the new implement. One hundred and ten patients (of which one patient had a double cervix) were randomised to have the Jordan's spatula first to obtain their smears and 110 were randomised to have the new implement first to obtain their smears.

For the 221 smears that were done using the new implement, 119 (54%) showed evidence of transformation zone using the BSCC (British Society for Clinical Cytology) criteria, based on the presence of E, M and C. This was in contrast to just 81 (37%) smears with E, M and C when using the Jordan's spatula. Table 1 shows the distribution of the 221 pairs of smears showing evidence of transformation zone (E, M and C) by type of device. It is clear from this table that 62 pairs of smears showed E, M and C when using the new implement, but did not show such a positive result with the Jordan's spatula. This is in contrast to just 24 pairs that showed a positive result using the Jordan's spatula but a negative result using the new implement. The odds of showing E, M and C when using the new implement is thus 2.58 that of the Jordan's spatula with 95% confidence interval of 1.59–4.33, a statistically significant result with P value <0.001 (McNemar test). Adjusting for the order in which a smear was performed with each device does not alter the results as expected since half of the patients were randomised to have one of the two devices used first.

Table 1.  Distribution of smear pairs showing evidence of transformation zone by type or device (BSCC criteria). Values are given as n and (%).
New implement smears showing evidence of transformation zone (E, M and C)Jordan's spatula smears showing evidence of transformation zone (E, M and C)
YesNoTotal
Yes5762119 (54%)
No2478102
Total81 (37%)140221

Similarly, using the American/Bethesda classification to assess the 221 smears that were done using the new implement, 128 smears (58%) were satisfactory, showing evidence of an endocervical/transformation zone component with more than five groups of cells (++ or +++). This was in contrast to just 86 (39%) smears with evidence of transformation zone (++ or +++) when using the Jordan's spatula. Table 2 shows how many of the 221 pairs of smears showed evidence of transformation zone (++ or +++) by which type of device. It is clear from this table that 57 pairs of smears showed evidence of transformation zone (++ or +++) when using the new implement but did not show such a positive result with the Jordan's spatula. This is in contrast to just 15 pairs that showed a positive result using the Jordan's spatula but with a negative result using the new implement. The odds of showing evidence of transformation zone (++ or +++) when using the new implement is thus 3.80 that of the Jordan's spatula with a 95% confidence interval of 2.12–7.23, which is a statistically significant result with a P value of <0.001 (McNemar test). Adjusting for the order of which smear was performed first does not alter the result as expected.

Table 2.  Distribution of smear pairs showing evidence of transformation zone by type or device (Bethesda criteria). Values are given as n and (%).
New implement smears showing evidence of transformation zone (++ or +++)Jordan's spatula smears showing evidence of transformation zone (++ or +++)
YesNoTotal
Yes7157128 (58%)
No157893
Total86 (39%)135221

Table 3 summarises the effect of the order of the smear on both outcomes, the percentage of smears showing E, M and C and the percentage showing (++ or +++). For the first outcome (E, M and C), it seems that the first smear whether performed by the Jordan's spatula or the new implement had the worse outcome than the second smear. Overall, 41% of all first smears showed E, M and C as opposed to 49% using the second smear. However, this difference did not reach statistical significance, P= 0.09. For the second outcome (++ or +++), the first smear whether performed by the Jordan or the new implement had significantly worse outcome than the second smear. Overall, 43% of all first smears showed (++ or +++) as opposed to 54% using the second smear, a statistically significant result, P= 0.022.

Table 3.  Effect of order of smear on the outcome variables. Values are given as n/total (%) or RR [95% CI].
 First smearSecond smearRR [95% CI]P
Smears showing evidence of transformation zone (E, M and C) using
Jordan's spatula35/111 (32)46/110 (42)  
New implement56/110 (51)63/111 (57)  
Either spatula91/221 (41)109/221 (49)0.83 [0.68–1.03]0.09
 
Smears showing evidence of transformation zone (++ or +++) using
Jordan's spatula34/111 (31)52/110 (47)  
New spatula61/110 (55)67/111 (60)  
Either spatula95/221 (43)119/221 (54)0.80 [0.66–0.97]0.022

Only 10 cases were exactly similar between the new implement and Jordan's spatula in their results and another 21 cases were almost similar except for very minor differences. There is a discrepancy in the grading between the two slides taken from each patient (25.34% of the cases). The discrepancy ranged from negative to borderline, borderline to mild dyskaryosis and mild dyskaryosis to moderate dyskaryosis (i.e. one grade up or down in severity between one smear report and another). Using the new implement, only three smears (1.3%) were inadequate, with no evidence of transformation zone or endocervical cells, compared with four (1.8%) with the Jordan's spatula.

We had two interesting cases during our study. One patient was referred because of previous inadequate smears. Colposcopy showed CIN II. Her first smear was with Jordan's spatula, which was reported yet again as inadequate. Her second smear was with the new implement, which was reported as mild dyskaryosis. The other interesting case mentioned earlier was the patient who had a double cervix.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Note
  10. References

For a smear to be considered as adequate, samples of cells from the endocervix, transformation zone and ectocervix must be present in the smear.4 An extended tip spatula was chosen for comparison as it has been shown in various studies to be better than other smear collection devices for collection of endocervical cells, detection of dyskaryosis and adequate smear rates. Extended tip spatulae of various designs yield better results than short tip spatulae.5 Whether a cervical brush or a cervical scraper gives a better result is debatable. Apparently, the Cervex brush offers no advantage over the Aylesbury spatula (which is an extended tip spatula) in reducing the incidence of inadequate smears.6 Besides, it is advocated that primary screening should not be carried out with an endocervical brush alone, as such smears may be composed of endocervical cells only and may not sample mature squamous cells or transformation zone epithelium.7 The most effective combination appears to be with the simultaneous use of the cytobrush with an extended tip spatula.5 The new implement has the advantages of both these devices in one. The new implement would also prove particularly effective in women who have been treated for CIN, where sampling of the transformation zone may be difficult and requiring the use of an endocervical brush in addition to the spatula.

Even with extended tip spatulae or when a combination of devices is used, the diameter of the cervix that is sampled is defined by the radius of the device, which is why it is advisable to perform several strokes over the surface of the cervix especially if the transformation zone is extensive.8 However, with the new implement, the diameter can be extended from a 3 cm locked position, up to twice that diameter, that is, 6 cm in open position, (this would have the advantage of sampling a larger area of the cervix with a higher chance of sampling small isolated cervical lesions).

Many studies have shown that using two smears at the same time yield better quality smears.9–14 Taking two specimens increases the chance of getting positive smears by 50%.9 The prevalence of cytological dysplasia and more severe abnormality seems to be 86% higher with duplicate smears than that based on the first smear alone.12 This very good result is not shared by Luthy et al.13 who reported a higher incidence of only 26.3%.

As a result of the design of the new implement, it effectively takes two smears simultaneously with better results, as shown in our study. Taking two smears always yields better results and the second smear seems to be of better quality than the first, the reason being that the first smear contains a larger number of already desquamated cells and the second represents more cells dislodged from deeper layers. This explains why 16–36% of single smears are false negative.2“Serious doubt is therefore raised about the sensitivity of the single cervical Pap smear in detecting abnormal cervical cytology”.12 However, taking two smears at the same time for screening is prohibitively expensive due to duplication. The evaluative sampling errors between the two smears are no greater than 10%11 and are therefore not cost effective. The new implement has the advantage of having an extended tip and being a scraper on one side of the loop and a brush on the other side, thus equivalent to taking two smears in one.

Studies have also suggested that during smear taking, use of a second spatula is more successful at sampling endocervical cells.10 The design of our study helps to overcome this bias, with half the patients having been randomised to have their first smear with the new implement and the other half first with the Jordan's spatula. Subject variation bias (as some cervices may be difficult to sample) was overcome due to the fact that each patient had a smear with both devices. We used the ability of the device to sample the transformation zone and collect endocervical cells as endpoints, as the presence of endocervical cells is a valid and convenient surrogate for the ability to detect dyskaryosis.15 We intentionally did not include any histology results because our primary aim was to assess the quality of smears obtained and not their correlation with histology. Cytological examination has detection accuracy rather than diagnostic accuracy, and there is imperfect correlation between the severity of epithelial abnormality predicted by cytological findings and that found histologically in cervical tissue.6 Moreover, biopsies were not universally taken and therefore any comparison would have been defective.

In our study, five people including doctors and nurses, who were instructed on a single occasion prior to the use of the new implement for the study, took the smears. All of them found the new implement easy to use and although noted to be difficult in 3% of women as compared with 0.5% with the Jordan's spatula, it was possible to obtain a smear with both devices for each patient. A larger percentage of the patients found the new implement uncomfortable compared with Jordan's. However, discomfort has been recorded in the results, even if it was comparatively marginally so and none of the patients found it painful. It was noted that mild stretching of the vaginal fornices due to a wider diameter of the loop was a contributing factor in women with small cervices. In these cases, discomfort can be overcome by using a lock that narrows the loop. Interestingly, there were a larger percentage of patients with bleeding noted from the cervix after taking a smear with the new implement, although this was not reflected in blood staining of the slides.

Using the new implement, only three smears (1.3%) were inadequate, with no evidence of transformation zone sampling or endocervical cells, compared with four (1.8%) with the Jordan's spatula. These are comparable to the inadequate smear rates reported with the use of liquid based cytology techniques in various studies. Evidence from 17 studies suggested that the proportion of inadequate tests with liquid based cytology ranged from 0% to 8.5% with a median of 0.7%.16 Our next study aims to compare the Cervex brush with the new implement in obtaining specimens for thin prep cytology.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Note
  10. References

Statistically significant larger numbers of good quality smears, judged by evidence of sampling of transformation zone and endocervical cells, are obtained with the new implement. The use of cervical cell sampling devices showing the best cytological results improve the interpretation and validity of cervical smears.17 Thus, use of the new implement would aim to increase adequate smear rates, enhance the ability to detect dyskaryosis 15 and reduce unnecessary recall of patients for repeat smears.5,18

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Note
  10. References

The authors would like to thank Mrs Kathy McLoughlin, colposcopy clinic co-ordinator, whose contact with GPs and patients ensured that the protocol was adhered to, Mrs Sandra Malham, for the essential secretarial work and her help in writing the manuscript, Mrs Helen Edwards and Margaret Wilczynska, for their help in locating results, addresses of patients and their files, and Sister Slobodank Gilfillan and staff nurses Raquel Dumas, Sarah Roscoe and Brigid Burke, for ensuring the smooth running of the research and collection of data.

Note

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Note
  10. References

Ashwood Timber and Plastics (London), who have undertaken to apply for international patent, manufactured the new implement. http://www.ashwood.biz/.

Conflict of Interest Statement

Mr Samuel George designed the new implement, which was manufactured for research purposes by Ashwood Timber and Plastics, who applied for international patents. No monies have exchanged hands yet, however, royalty fees are expected to be paid to Mr. Samuel George and Ealing Hospital NHS Trust on marketing the product.

Funding

The Obstetrics and Gynaecology Department agreed to pay whatever expenses were incurred. We were fortunate enough that everyone who participated or contributed to the research did so wholeheartedly, without payment.

Author Contributions

Mr S. George participated in the design of the new implement, the protocol, the main body of the research and collation of data and has seen and approved the final version. Mr Y. Abrahams participated in the preparation of the protocol and main body of the research work and has seen and approved the final version. Dr S. Karim participated in the cytological examination of all the slides provided and has seen and approved the final version. Ms Anjali Kothari participated in the main body of the research, collation, analysis and computerisation of the data and has seen and approved the final version. Ms Kothari is the corresponding author and has full access to all the data in the study and had final responsibility for the decision to submit for publication.

Other Contributors

Dr Muna Abdalla, PhD, lecturer at the London School of Hygiene and Tropical Medicine carried out the statistical analysis. Dr S. Karthikesalingam, MBBS, MRCOG, MFFP, LRCP, MRCS, general practitioner/colposcopist, contributed to recruiting women for the study. Sister Hazel Prendergast, colposcopy specialist nurse, participated in smear taking.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Note
  10. References
  • 1
    Onon T, Shafi M. Developments in cervical screening techniques. Trends Urol Gynaecol Sex Health 1999;4(1):2731 (January/February).
  • 2
    Singer A. Cervical cancer screening: state of the art. In: JonesHW, editor. Bailliere's Clinical Obstetrics and Gynaecology. Cervical Intraepithelial Neoplasia. London: Bailliere Tindall, 1995; 9(1):3964.
  • 3
    Kurman RJ, Solomon D. The Bethesda System for Reporting Cervical/Vaginal Cytologic Diagnoses. New York: Springer-Verlag, 1994.
  • 4
    Consensus in Medicine. Cervical cancer screening: the papsmear, summary of an NIH consensus statement. BMJ 1980;281: 1264 (November 8).
  • 5
    Martin-Hirsch P, Jarvis G, Kitchener H, Lilford R. Collection devices for obtaining cervical cytology samples. Cochrane Database Syst Rev 2000;(3):CD001036.
  • 6
    Dey P, Collins S, Desai M, Woodman C. Adequacy of cervical cytology sampling with the Cervex brush and the Aylesbury spatula: a population based randomised controlled trial. BMJ 1996;313: 721 (September 21).
  • 7
    Cervical Cytopathology—Achievable Standards, Benchmarks for Reporting, Criteria for Evaluating Cervical Cytopathology. Report of a working party set up by the Royal College of Pathologists (RC Path), British Society for Clinical Cytology (BSCC) and NHS Cervical Screening Programme (NHSCSP), 1995.
  • 8
    Onon T, Shafi M. Cervical smears: when and how to take one. Trends Urol Gynaecol Sex Health 1999;4(2):25 (March/April).
  • 9
    Sedlis A, Walters A, Balin H, Hontz A, LoSciuto LA. Evaluation of two simultaneously obtained cervical cytological smears, a comparison study. Acta Cytol 1974;18(4):291 (July/August).
  • 10
    Bourne R, Beilby JOW. Methods and devices, trial of new cervical spatula. Lancet 1976;1330 (June 19).
  • 11
    Davis JR, Hindman WM, Paplanus SH, Trego DC, Wiens JL, Suciu TN. Value of duplicate smears in cervical cytology. Acta Cytol 1981;25(5):533 (September/October).
  • 12
    Shulman JJ, Hontz A, Sedlis A, Walters A, Balis H, LoScuito L. The pap smear: take two. Am J Obstet Gynecol 1975;121(8):1024 (April 15).
  • 13
    Luthy DA, Briggs RM, Buyco A, Eschenbach A.. Cervical cytology, increased sensitivity with a second cervical smear. Obstet Gynaecol 1978;51(6):713 (June).
  • 14
    Buntinx F, Bronwers M. Relation between sampling device and detection of abnormality in cervical smears: a meta-analysis of randomised and quasi-randomised studies. BMJ 1996;313: 1285 (November 23).
  • 15
    Martin-Hirsch P, Lilford R, Jarvis G, Kitchener HC. Efficacy of cervical-smear collection devices: a systematic review and meta analysis. Lancet 1999;17631770 (November 20).DOI: 10.1016/S0140-6736(99)02353-3
  • 16
    Guidance on the Use of Liquid-Based Cytology for Cervical Screening. Technology Appraisal 69; October 2003. Available: http://www.nice.org.uk/article.asp?a1864.
  • 17
    Kohlberger PD, Stani J, Gitsch G, Kieback DG, Breitenecker G. Comparative evaluation of seven cell collection devices for cervical smears. Acta Cytol 1999;43(6):10231026 (November/December).
  • 18
    Du-Plessis JM, Schaetzing AE, Wrantz PA, Louw M. Aylesbury and cervitula spatulas: a comparative study to assess the adequacy of cervical smears. Acta Cytol 2001;45(5):675678 (September/October).

Accepted 12 March 2004