|
Permission to reprint granted by The Lancet
University of Zimbabwe/JHPIEGO Cervical Cancer Project (Members listed at end of paper)
Direct Correspondence to:
Zimbabwe Office, Department of Obstetrics and Gynaecology, University of Zimbabwe, PO Box
A178, Avondale, Harare, Zimbabwe; or,
CECAP Office, JHPIEGO Corporation, 1615 Thames Street, Suite 200,
Baltimore, MD 21231, USA
Background Naked-eye visual inspection of the cervix with acetic-acid wash
(VIA), or cervicoscopy, is an alternative to cytology in screening for cervical cancer in
poorly resourced locations. We tested the sensitivity, specificity, and predictive value
of VIA done by nurse-midwives in a less-developed country.
Methods Women were screened by six trained nurse-midwives in a two-phase,
cross-sectional study at 15 primary-care clinics in Zimbabwe. VIA and Pap smears were done
concurrently, and their sensitivity and specificity compared. Colposcopy, with biopsy as
indicated, was used as the reference test to allow a direct comparison of the test
unaffected by verification bias.
Findings 10934 women were screened. In phase II, 2148 (97·5%) of the 2203
participants for whom there was a screening result also had a reference test result. Also
in phase II, VIA was more sensitive but less specific than cytology. Sensitivity (95% CI)
was 76·7% (70·3-82·3) for VIA and 44·3% (37·3-51·4) for cytology. Specificity was
64·1% (61·9-66·2) for VIA and 90·6% (89·2-91·9) for cytology.
Interpretation The high sensitivity of VIA shows that the test could be valuable
in detection of precancerous lesions of the cervix. However, there are costs to the
patient and system costs associated with high numbers of false-positive results, so
attention should be given to improving the specificity of VIA.
Lancet 1999; 353: 869-73
In most less-developed countries, cervical-cancer screening programmes are small-scale
or non-existent.1-3 Consequently, there are few opportunities to diagnose
precancerous disease, and most patients present with invasive disease at an advanced
stage.3-5 In some less-developed countries Pap-smear-based screening is
available, but usually only in urban areas or in the private health sector that serves a
small proportion of the female population. Screening programmes based on Pap smears
require technical capabilities and systems for transportation, communication, follow-up,
and training that are beyond the capacity of healthcare infrastructure in most
less-developed countries.2-6 Thus, other methods of cervical-cancer screening
provision have been investigated.1,6-8
One such method is visual inspection with acetic acid (VIA). The cervix is washed with
acetic acid and then inspected by eye for evidence of disease (also known as cervicoscopy,
or direct visual inspection). This has potential advantages over traditional screening
techniques in poorly-resourced locations--there is immediate feedback of test results to
the patient and, importantly, treatment can be provided immediately after the test.9-11
Given the potential significance of VIA, we undertook a field-based study in Zimbabwe
to screen more than 10000 women with both the Pap smear and VIA. The primary endpoint was
to assess the specificity and sensitivity of VIA done by non-physicians in a primary-care
setting. A secondary objective was to compare these test qualities with those of the Pap
smear, the current screening method in Zimbabwe.
Participants
Our study was a cross-sectional, screening test study that took place between October,
1995, and August, 1997. Women attending 15 primary-care clinics in Chitungwiza and the
greater Harare area, Zimbabwe, were invited to attend a health-education talk on cervical
cancer for the purpose of recruitment into the study. The talk was given every morning
while women were waiting to be seen for other health matters. After the talk, those
interested in being screened were invited to take part in our study. Women aged between 25
and 55 years, who were not pregnant and had no previous history of cervical cancer or
hysterectomy, were eligible for enrolment.
Study design
The study had two phases: in both, participants were interviewed by a trained female
nurse-midwife who used a standardised questionnaire. The study objectives were explained,
and verbal consent was obtained after an informed-consent statement was read out.
Immediately after the participant's history was taken, both screening tests were done.
Each woman was placed in a modified lithotomy position on an examination table. An
unlubricated bivalve speculum was inserted into the vagina by the nurse-midwife and a
cervical cytology specimen was obtained with a wooden Ayres spatula. The nurse-midwives
were trained to scrape the cervix around the entire transformation zone to obtain an
adequate specimen. Immediately after a cytology specimen was obtained, the nurse-midwife
cleansed away any excess mucus thoroughly with a saline-soaked swab, and applied a
solution of 4% acetic acid to the cervix with a cotton-tipped applicator. With the aid of
a handheld flashlight, the nurse-midwife then visually inspected the whole cervix.
Categories of VIA findings (panel) were recorded on the study questionnaire.
| Panel: VIA classification categories |
| Normal |
Smooth, pink, uniform, featureless |
| Atypical |
Cervicitis (inflammation, red spots)
Discharge
Ectropion
Polyp
|
| Abnormal |
White plaques
Ulcer
Acetowhite epithelium |
| Cervical cancer |
Cauliflower-like growths
Fungating mass
|
In phase I of the study, if the VIA assessment showed an abnormal
result, the woman was scheduled for colposcopy by a study coordinator. By use of a
systematic random sampling scheme, every tenth woman with a normal or atypical VIA
assessment was also scheduled for colposcopy. If the Pap smear was abnormal, attempts were
made to tell the woman of the result by home visit, letter, and communication through
non-study clinical staff at the 15 clinics. If contacted, the woman was encouraged to make
an appointment for a colposcopy. Colposcopy was done in Harare by three University of
Zimbabwe faculty members (ZMC, JK, NH). If no lesion was found during colposcopy, the
patient was reassured and asked to return to her local clinic every year for routine
follow-up. Women assessed as abnormal on colposcopy (confirmed, if indicated, by biopsy)
were referred for appropriate treatment and followed up according to standard local
clinical protocol.
Phase II of the study began about 1 year after the start of phase I, and continued
until the end of the study. Phase II differed from phase I primarily in that recruitment
took place on a given day, and on the next day all screening and diagnostic tests were
done. Thus, in phase II all test-positive and all test-negative women were sent for the
reference test. Transport was provided to a Harare clinic, where colposcopy was done by
two investigators (EN, ZMC). In phase II, a specimen was obtained from all women to be
tested for human-papillomavirus. Human-papillomavirus data were collected to assess the
specificity and sensitivity of that test as a single screening test in a less-developed
country, and to assess the usefulness of human-papillomavirus tests as an adjunct to VIA
screening. These findings will be presented elsewhere.
Outcome measures
The reference standard used in our study was colposcopy with biopsy (whenever the
latter was clinically indicated)--a commonly used and accepted reference standard for
cervical-cancer screening studies.9,10,12-16 We defined two thresholds of
disease: a low threshold of low-grade squamous intraepithelial lesion (LGSIL) or worse on
colposcopy (or their equivalents on biopsy); and a high threshold of high-grade squamous
intraepithelial lesion (HGSIL) or worse. LGSIL or more was taken as a positive test for
the Pap smear, and, consistent with local cytology norms, atypical squamous cells of
undetermined significance and atypical glandular cells of undetermined significance were
taken to be test-negative. We defined a positive test for VIA as either abnormal or
cancer, as shown by acetowhite lesions and other visual markers of (pre)cancerous lesions
of the cervix.
Quality control
In both phases of the study, Pap smears were analysed by cytotechnicians in Harare, who
were unaware of the VIA results. All positive smears and a 10% random sample of smears
assessed as negative by each cytotechnician were forwarded for a second assessment by the
study cytopathologist (RM). In addition, in phase I of the study a certified
cytopathologist at the Johns Hopkins Bayview Medical Center, Baltimore, MD, USA, reviewed
a sample of 10% of all negative and all positive smears. In phase II, all slides reviewed
by the local cytopathologist were sent for review by the US cytopathologist. Study
protocol called for colposcopy results to be recorded without knowledge of the VIA or Pap
smear findings, and for biopsies to be read without knowledge of any test result. The
study was approved by the ethics committee of the Zimbabwe Research Council and the
Institutional Review Board of the Johns Hopkins Bayview Medical Center, USA.
Personnel were trained and assessed, and clinical standardisation was ensured at the
start of the study. Refresher training was provided for the nurse-midwives in speculum
insertion and Pap collection, followed by practical training in VIA over 3 days. The
nurse-midwives were familiarised with the naked-eye appearance of the cervix in various
states of health and disease. A pictorial atlas was used for reference during training and
during the study.17
Over 5 days, the study cytotechnicians took part in a review course designed to
standardise their skills run by the study cytopathologist (RM) who continued to work with
the technicians for a week. In addition, the US-based cytopathologist visited Zimbabwe for
1 week at the beginning of the study to work with the cytotechnicians and to standardise
assessment categories (modified Bethesda system) with the local cytopathologist.
Statistical analysis
The initial sample size for the study was set at 24 000 patients, to allow statistical
precision of 0·05. We assumed a 60% sensitivity, the presence of high-grade lesions in
5·6% of patients, and that 10% of all participants who tested negative on all screening
would receive the reference test. The overall sample-size estimate was lowered markedly to
12 000 women at the start of phase II of the study, when high-grade lesions were found in
around 10% of patients--protocol changes meant that 100% of all test-negative women on
screening would be referred for the reference test.
Univariate analyses were done for all questionnaire study variables. Analyses of
sensitivity, specificity, and predictive value were done by use of standard formulae for
these test qualities.18 The rate of disease (pre-cancer and cancer) detection
was calculated as the number of true-positive results detected by the test divided by the
total number of women with a screening result from that test.19 Exact binomial
95% CIs were calculated for predictive evaluation.18
Our study enrolled 10 934 women (8731 in phase I, 2203 in phase II). 22 other women
were recruited but were not included in the analysis because important identifying
information was missing. In phase I, 1584 (18·1%) women had colposcopy (and biopsy as
indicated) as a result of a positive VIA or Pap smear screen, or because they were
selected as one of the test-negative women to receive the reference-standard test. In
phase II, the reference standard test was done on 2147 (97·5%) of the 2203 women for whom
there was a Pap smear or VIA result (Table 1).
Table 1: Test results by study
phase
| |
Phase I
(n=8731) |
Phase II
(n=2203) |
VIA
negative |
VIA
positive |
T
o
t
a
l |
VIA result missing |
VIA
negative |
VIA
positive |
T
o
t
a
l |
| Pap X |
Pap -
|
Pap + |
Pap X |
Pap -
|
Pap + |
Pap X
|
Pap + |
Pap X |
Pap -
|
Pap + |
Pap X |
Pap -
|
Pap + |
Pap X |
| Missing |
261 |
5712 |
461 |
50 |
507 |
156 |
7147 |
4 |
0 |
2 |
31 |
0 |
2 |
15 |
2 |
56 |
| Negative |
20 |
217 |
155 |
34 |
561 |
182 |
1169 |
15 |
1 |
35 |
1128 |
70 |
14 |
570 |
107 |
1940 |
| Positive |
6 |
38 |
99 |
12 |
95 |
165 |
415 |
1 |
0 |
3 |
31 |
14 |
3 |
80 |
75 |
207 |
| Total |
287 |
5967 |
715 |
96 |
1163 |
503 |
8731 |
20 |
1 |
40 |
1190 |
84 |
19 |
665 |
184 |
2203 |
Pap X = Pap result
missing
Pap - = Pap
result negative
Pap + = Pap result positive |
Most of the women were married, all were sexually active, most had used a method of
family planning at some time, and few (<15%) had ever been screened for cervical cancer
in the past (Table 2). Most participants were in their late 20s or early 30s, and had
completed either primary or secondary school. These statistics are characteristic of women
who attend primary-care clinics in Zimbabwe. There was little difference in baseline
characteristics between study phases, except for variables related to sexually transmitted
infections. The difference in experience of such infections reflects increased
self-selection by women who came to or returned to the clinic specifically for screening
in the last year of the study.
Table 2: Participant characteristics
Characteristic |
Phase I
(n=8731) |
Phase II
(n=2203) |
Total
(n=10 934) |
| Demographic |
| Mean (SD) age (years) |
32·0 (6·5) |
33·2 (7·1) |
32·2 (6·6) |
| Ever married (%) |
96·2 |
94·8 |
95·9 |
| Education (%) |
| None |
3·3 |
6·0 |
3·8 |
| 1-7 years completed |
44·3 |
47·3 |
44·9 |
| 8-12 years completed |
51·0 |
46·5 |
50·1 |
| Higher education |
1·3 |
0·3 |
1·1 |
| Adult literacy |
0·1 |
0·0 |
0·1 |
| Sexual |
| Mean gravidity (SD) |
3·4 (2·1) |
3·3 (2·3) |
3·4 (2·2) |
| Ever used family planning (%) |
91·1 |
86·2 |
90·1 |
| Sexually active in past year (%) |
100 |
100 |
100 |
| STI or suspected STI (%) |
45·5 |
59·6 |
48·3 |
| Current STI symptoms (%) |
67·8 |
83·8 |
71·0 |
| Previous Pap smear (%) |
10·2 |
12·9 |
10·7 |
| STI = sexually transmitted infection. |
VIA assessment was adequately completed for 10 913 women (8731 in phase I, 2182 in
phase II). Adequate Pap smears were obtained from 8348 women in phase I and 2144 in phase
II. In phase I, 20·2% of the women tested positive for VIA and 14·6% were Pap-smear
positive. In the second phase, test-positive rates were 39·8% and 12·6% for VIA and Pap
smear, respectively (Table 3).
Table 3: Results of screening tests
Test |
Numbers Tested (%) |
| Phase I |
Phase II |
Total |
| VIA |
| Normal |
2688 (30·8) |
495 (22·7) |
3183 (29·2) |
| Atypical |
4281 (49·0) |
819 (37·5) |
5100 (46·7) |
| Abnormal |
1747 (20·0) |
857 (39·3) |
2604 (23·9) |
| Cancer |
15 (0·2) |
11 (0·5) |
26 (0·2) |
| Total |
8731 (100) |
2182 (100) |
10 913 (100) |
| Pap smear |
| Normal |
4998 (59·9) |
1300 (60·6) |
6298 (60·0) |
| Inflammation |
1150 (13·8) |
252 (11·8) |
1402 (13·4) |
| ASCUS |
911 (10·9) |
285 (13·3) |
1196 (11·4) |
| AGUS |
71 (0·9) |
38 (1·8) |
109 (1·0) |
| LGSIL |
828 (9·9) |
196 (9·1) |
1024 (9·8) |
| HGSIL |
371 (4·4) |
69 (3·2) |
440 (4·2) |
| Squamous cancer |
19 (0·2) |
3 (0·1) |
22 (0·2) |
| Adenocarcinoma |
0 (0) |
1 (<0·1) |
1(<0·1) |
| Total |
8348 (100) |
2144 (100) |
10 492 (100) |
ASCUS = atypical squamous cells of undetermined significance.
AGUS = atypical
glandular cells of undetermined significance.
LGSIL = low-grade squamous intraepithelial
lesion.
HGSIL = high-grade squamous intraepithelial lesion. |
Among the 8731 women in phase I with a definitive Pap or VIA result there were 305
cases of LGSIL, 398 cases of HGSIL, and 17 cancers detected by the reference standard.
Among the 2182 women in phase II there were 294 cases of LGSIL, 204 cases HGSIL, and three
cancers. Of the 602 cases of HGSIL and 20 cancers, 74·8% were confirmed by biopsy.
Next Page
|
