The impact of a community intervention to improve cervical cancer screening uptake in the Amazon region of Brazil

ABSTRACT CONTEXT AND OBJECTIVE: In the northern region of Brazil, cervical cancer is the most important cause of cancer-related deaths among women. There is considerable likelihood, however, that official incidence and mortality figures are greatly underestimated. The aim of this study was to estimate the repercussions from improvement in cervical cancer screening programs on the incidence of pre-invasive and invasive cervical lesions in a municipality in this region. DESIGN AND SETTING: This was a quasi-experimental study that assessed process dimensions relevant to the program objectives. The study comprised a sample of 2,226 women seen at primary healthcare units in Cruzeiro do Sul, a small city in the Brazilian Amazon region, from April 2003 to July 2004. METHODS: Women were recruited through local radio advertisements and by oral communication from the investigators. The women answered a structured questionnaire and underwent pelvic examination, which included Papanicolaou (Pap) smears and naked-eye inspection of the cervix after applying diluted acetic acid. Women with positive Pap smears or abnormal gynecological examination were referred for colposcopy and possible biopsy, diathermic large loop excision of the transformation zone or conization. RESULTS: The results obtained were compared with historical official data retrieved from the Brazilian Ministry of Health's database. Intervention resulted in a 40% increase in positive Pap smears and detection of cancer was nine times higher than had been observed in routine screening. CONCLUSIONS: Detection of pre-invasive and invasive cervical lesions in the intervention group was remarkably higher than among women seen during routine screening.


OBJECTIVE
The purpose of the present study was to estimate the changes in the incidence of pre-invasive and invasive cervical neoplasia that a possible improvement in access to health services would cause, in a small and isolated province of the Amazon region, in northern Brazil.

Study site Study site
This study was carried out in the city of Cruzeiro do Sul, which is located in the extreme west of the Brazilian Amazon basin, in the State of Acre.It has been estimated that, in 2001, the municipality had 67,441 inhabitants, 3 52% of them living in rural areas.The city is accessible mainly by river or air transportation, because during the rainy season the main roads to the city remain unserviceable.Illiteracy is very common: approximately 40% of the population cannot read or even sign their own names. 2Economic development is negatively affected by the region's geographical position, in the middle of the Amazon forest, and by several concurrent social problems.

Community-level intervention Community-level intervention
Women were recruited through local advertisements, by radio broadcasting and oral communication; the latter was performed by the investigators themselves.To be eligible to enter the study, women had to: a) provide informed consent by signing (or placing their fi ngerprints on) the informed consent form (which was read out by one of the investigators when the women were unable to read it themselves); b) have a cervix uteri; c) be outside of the bleeding period of the menstrual cycle; d) have not had sexual intercourse or applied vaginal creams over the three days prior to consultations.Women in this last condition were advised to schedule a new appointment in a few days thereafter.Women were excluded if they: a) had undertaken a Pap test during the last year prior to this consultation; b) had undertaken cervical cauterization, biopsy or conization during the last year prior to this consultation; c) were pregnant; d) had had cervical neoplasia (any type) in the past.There is considerable likelihood, however, that offi cial incidence and mortality fi gures are greatly underestimated.The aim of this study was to estimate the repercussions from improvement in cervical cancer screening programs on the incidence of pre-invasive and invasive cervical lesions in a municipality in this region.

DESIGN AND SETTING:
This was a quasi-experimental study that assessed process dimensions relevant to the program objectives.The study comprised a sample of 2,226 women seen at primary healthcare units in Cruzeiro do Sul, a small city in the Brazilian Amazon region, from April 2003 to July 2004.

METHODS:
Women were recruited through local radio advertisements and by oral communication from the investigators.The women answered a structured questionnaire and underwent pelvic examination, which included Papanicolaou (Pap) smears and naked-eye inspection of the cervix after applying diluted acetic acid.Women with positive Pap smears or abnormal gynecological examination were referred for colposcopy and possible biopsy, diathermic large loop excision of the transformation zone or conization.

RESULTS:
The results obtained were compared with historical offi cial data retrieved from the Brazilian Ministry of Health's database.Intervention resulted in a 40% increase in positive Pap smears and detection of cancer was nine times higher than had been observed in routine screening.
CONCLUSIONS: Detection of pre-invasive and invasive cervical lesions in the intervention group was remarkably higher than among women seen during routine screening.A total of 2,226 women were enrolled, and consultations were carried out in the primary healthcare units of Cruzeiro do Sul, from April 2003 to July 2004.In addition to the consultations performed for the purposes of this study, another 3,780 routine Pap tests were carried out in Cruzeiro do Sul while the study was underway.

Consultation routine Consultation routine
After signing the informed consent form, the women responded to a structured questionnaire and were subjected to pelvic examination, with inspection of the cervix and collection of samples for the Pap test.As a fi nal step, 3.0% diluted acetic acid was applied to the cervix, and the cervix was re-examined using the naked eye.Abnormal results from the cervical inspection or Pap test prompted referral for colposcopy, with possible biopsy or diathermic excision of the transformation zone, depending on the characteristics of the lesion.

Pap smear Pap smear
Pap smears were taken using an Ayre spatula and endocervical brush.They were fi xed in 95% ethanol and stained using the modifi ed Papanicolaou method.The fi nal cytological diagnosis was obtained using the Bethesda System, 4 and the results were classifi ed as normal/infl ammatory, atypical squamous cells (ASC), atypical glandular cells (AGC), low-grade squamous intraepithelial lesion (LSIL), high-grade squamous intraepithelial neoplasia (HSIL) or suggestive of cancer.The cytological slides from smears collected from patients enrolled to this study were read by an experienced cytopathologist, based at the cytology laboratory of Universidade Estadual de Campinas.On the other hand, patients screened normally, within the offi cial program, had their Pap tests performed by local laboratories in the State of Acre.For statistical purposes, normal/infl ammatory results were categorized as negative, and ASC, LSIL or HSIL as positive.

Histology Histology
Biopsy and conization specimens were fi xed in 10% phosphate-buffered formalin, embedded in paraffi n, and stained with hematoxylin and eosin (HE).Biopsies were analyzed according to the World Health Organization (WHO) criteria 5 and classifi ed as negative, CIN 1, CIN 2, CIN 3 or invasive squamous carcinoma.

Statistical methods Statistical methods
Box plot representations of the mean, median and standard deviations of age at diagnosis for each pre-neoplastic or neoplastic lesion were produced.The performance estimators, or the fi rmed disease was even higher among women with HSIL, since 81.1% of them had CIN 2 or worse.All the women with Pap smears suggestive of cancer did indeed have high-grade disease (Table 1).The majority (69.4%) of the CIN cases occurred in women below 40 years of age, and disease severity increased in parallel with age (Figure 1).
The proportion of cytology abnormalities in this series was 40% higher (ratio: 1.40) than what has been offi cially reported for the city of Cruzeiro do Sul. 3However, during the intervention period of the study, the number of HSIL or worse Pap tests was four times higher (ratio: 4.1) than the offi cial historical fi gures, and the proportion of cancer cases detected in the population was nine times higher (ratio: 9.3) than what was observed from routine screening (Table 2).

DISCUSSION
The primary purpose of this study was to evaluate the consequences of a communitybased intervention for the screening of women who had not had a Pap test over the preceding 12 months, in a remote municipality in the Amazon region.In this intervention study, in comparison with the offi cial historical data and routine screening in Cruzeiro do Sul, a remarkably higher number of abnormal Pap tests were found and many more cases of cancer were detected.It should be noted, however, that invasive cancer detection rates generally exceed the cancer incidence rates in the population, percentages of positive screening tests across different categorical strata, were compared with historical data retrieved from the Brazilian Ministry of Health's database. 1The cancer detection rate was calculated per 1,000 women screened.
Cervical inspection revealed abnormalities in 21.2% of the women, and 541 women were examined by means of colposcopy.Abnormal fi ndings from colposcopy were found in 26% of the women that had undertaken the examination, and 55.8% of these women also had a positive Pap test (data not shown).
Visually detectable lesions were found in a few women with negative cytology: six cases of CIN 1, ten of CIN 2 or 3 and one case of cancer.The majority (78.6%) of the women with AGC were found to be free of histologically-confi rmed disease, and the same occurred in 63.6% of the women with ASC.Three (6.8%) of the women with ASC had cancer, and 13 (29.6%) of these were diagnosed with CIN 2 or 3.Among women with LSIL, six (37.5%) were diagnosed with CIN 2 or 3.The proportion of women with histologically-con-because some cancers would remain asymptomatic in the absence of screening.In the present series, the cancer detection rate (4.49cases per 1,000 women screened) was 30 times higher than the offi cially estimated incidence of the disease in Cruzeiro do Sul (14.53 per 100,000 women). 3n countries where the female population is screened appropriately, it has been estimated that less than 1% of Pap smears should harbor a cytological abnormality. 4,6In the present study, 5.4% of the Pap smears were considered to be abnormal.This can be attributed to inadequate previous screening (if any), and to the criterion of only including women who had not had a Pap test during the 12 months prior to enrollment.In addition, the personnel involved in the study were well-trained and performed rigorous quality control at all stages of the screening, from recruiting the women to their treatment.
In alignment with the high rate of Pap abnormalities found in this study, 21.2% of the women also had abnormalities in their cervix that were visible to the naked eye.In India and some African nations, investigators have analyzed the possible role of visual inspection of the cervix with acetic acid (VIA), alone or in combination with Pap smears, as a screening tool.][9] High sensitivity is probably the most desirable feature in a screening test.The sensitivity of Pap smears, when the test is performed under ideal conditions (i.e.adequate collection, slides immediately fi xed and experienced cytologists) barely exceeds 70%. 10,11On the other hand, it is also desirable that the number of false-positive results be reduced to a minimum, thereby avoiding unnecessary referrals for colposcopy.It has been well established that colposcopy places a heavy burden on healthcare systems, because it can only be performed by trained physicians, thus restricting its usage in underdeveloped regions.In the present study, abnormal colposcopy fi ndings were present in 55.8% of the women who had altered Pap smears.However, it has to be mentioned that, in contrast with the expected fi gures, the number of women with high-grade disease was exceptionally high.The correlation between Pap smear abnormalities and the real presence of cervical disease is not very favorable for low-grade cytological fi ndings.In the ASCUS-LSIL Triage Study (ALTS), only 15% of women with ASC were diagnosed with signifi cant histological lesions, and 50% of those with LSIL had abnormal cervical patterns visible via colposcopy. 9t is unreasonable to suppose that the cervical cancer screening defi ciencies in the remote and impoverished regions of Brazil could be resolved in the short term.Cytology-based screening programs demand expensive and complex infrastructure.In addition, Pap smears do not yield results immediately, which is especially problematic in sparsely populated regions with severe transportation shortcomings, like most of the Brazilian Amazon region.
As reported and well discussed in previous Indian, African and even Brazilian reports, VIA might be an acceptable alternative to cytology in low-resource settings with high incidence of cervical neoplasia, considering the logistic and economic circumstances usually encountered in such regions.This is certainly the case in the geographic region where the present study was conducted.For instance, as advocated by Sankaranarayanan et al., VIA allows diagnosis and treatment of cervical abnormalities on a single occasion.This would clearly help overcome one of the main problems in the Amazon region: the huge distances between settlements

Figure 1 .
Figure 1.Age-related distribution of fi nal diagnoses of cervical intraepithelial neoplasia (CIN) in 2,226 women examined in a municipality in the Amazon region Diagnosis Neg = negative; AGC = atypical glandular cells; ASC = atypical squamous cells; LSIL = low-grade squamous intraepithelial lesion; HSIL = high-grade squamous intraepithelial neoplasia; CIN = cervical intraepithelial neoplasia.

Table 1 .
Final diagnoses from the cytology results of 2,226 women screened for cervical cancer