neoplasia
Tiersma, S.
Citation
Tiersma, S. (2005, June 21). Psychosocial factors and cervical intraepithelial neoplasia. Retrieved from
https://hdl.handle.net/1887/2703
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion ofdoctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/2703
Psychosocial factors
and
cervical intraepithelial neoplasia
PROEFSCHRIFT
ter verkrijging van de graad van Doctor aan de Universiteit Leiden, op gezag van de Rector Magnificus Dr. D.D. Breimer,
hoogleraar in de faculteit der Wiskunde en Natuurwetenschappen en die der Geneeskunde, volgens besluit van het College voor Promoties ter verdediging op dinsdag 21 juni 2005 klokke 14.15 uur
door
Ellen Stella Maria Tiersma
Promotiecommissie
Promotores: Prof. Dr. G.J. Fleuren Prof. Dr. A.A.W. Peters
Co-promotor: Dr. A.P. Visser (Helen Dowling Instituut, Rotterdam) Referent: Prof. Dr. A.J.J.W. Vingerhoets (UvTilburg)
Overige leden: Prof. Dr. A.P.M. Heintz (UMC Utrecht) Prof. Dr. C.J.M. Melief
Prof. Dr. J.B.M.Z. Trimbos
Prof. Dr. H.B.M. van de Wiel (Groningen UMC)
Contents
CHAPTER 1 Introduction
CHAPTER 2 Revision of histology
CHAPTER 3 Visualising scanning patterns of pathologists in the grading of cervical intraepithelial neoplasia
(J Clin Pathol 2003;9:677-80)
CHAPTER 4 Psychosocial factors and the grade of cervical intraepithelial neoplasia: a semi-prospective study (Gynecol Oncol 2004;92:603-10)
CHAPTER 5 Psychosocial factors and the course of cervical intraepithelial neoplasia: a prospective study (Gynecol Oncol, in press)
CHAPTER 6 The relation between repression and precursors of cervical cancer: a semi-prospective study
(Submitted for publication)
CHAPTER 7 Reasons for early interventions by gynaecologists in a clinical follow-up study on cervical intraepithelial neoplasia
(Submitted for publication)
CHAPTER 1
Introduction
‘Mens sana in corpore sano’ is a famous saying and a popular belief. Vice versa, many people believe that the mind may influence the body. The science studying this latter relationship is known as
‘Psychoneuroimmunology’ (PNI) as psychosocial factors may exert an influence on health by way of a psychoneuro(endocrino)immunologic pathway.1;2 The role of the PNI pathway is especially conceivable in
virally-induced diseases, such as cervical cancer.
World-wide, cervical cancer is the second most common type of female cancer, accounting for about 10% of all newly diagnosed cancers in women.3;4 The precancerous stage of cervical cancer, known as dysplasia or
cervical intraepithelial neoplasia (CIN) can be detected by cytological screening.5-7 Unfortunately, unlike in many developed countries, in the third
world virtually no cytological screening programs are available and cervical cancer accounts for the highest cancer mortality in some of these countries.3
During the last decades, the causal relationship between the human papillomavirus (HPV) and cervical cancer has been solidly established.8;9 In
over 99% of the cervical cancers oncogenic HPV types are detected and these “high risk HPV types” are strongly related to CIN as well as to progression of CIN to cervical cancer.9 However, epidemiological studies
indicating that infection with oncogenic HPV types is far more common than CIN, suggest the necessity of cofactors in cervical carcinogenesis.10 Among
the cofactors with a proven relationship to cervical cancer are genetic changes and an insufficient immune system.11-13 Psychosocial factors may
indirectly influence cervical cancer by affecting the immune system.2;14
This thesis describes the results of studies on the influence of psychosocial factors on the grade and course of CIN in patients with abnormal cytology, as well as issues involved in the grading of CIN and the conduct of a multi-centre clinical study.
Pathogenesis of CIN and cervical carcinoma
studied partly during physical examination, by spreading the external os of the cervical canal.15;16 The cervical canal is covered with columnar epithelium,
whereas the ectocervix is mostly covered with non-keratinizing squamous epithelium, which is continuous with squamous epithelium of the vaginal wall. Over the years, the border of the columnar epithelium and squamous
epithelium, called the Squamocolumnar Junction (SCJ), migrates to the endocervical canal. CIN originates from the transformation zone, which is the area between the original SCJ and the new SCJ. In this area the columnar epithelium is replaced by squamous epithelium as part of the normal physiological process, known as squamous metaplasia. Infection with HPV affects the developing immature metaplastic cells of the transformation zone. CIN can be viewed as the interaction of high risk papillomavirus and
immature metaplastic epithelium.17
Two main types of cervical cancer are distinguished. The most frequent type is squamous carcinoma (75-80% of the cervical cancers) originating from the (metaplastic) squamous epithelium. Adenocarcinoma of the cervix (15% of the cervical cancers) originates from the columnar epithelium. In a small percentage of cases adenosquamous carcinoma is
FIGURE 1 Epithelial surfaces of the uterine cervix
columnar epithelium
squamous epithelium
new squamocolumnar junction
original squamocolumnar junction
ectocervix endocervix
diagnosed, which involves both cell types.18 The study described in this thesis
was limited to squamous cervical carcinoma and its precursors.
CIN grading
CIN is histopathologically classified as mild (CIN 1), moderate (CIN 2) or severe dysplasia and carcinoma in situ (CIN 3), depending on the
proportion of the epithelial layer showing dysplastic characteristics.19 These
dysplastic characteristics of CIN involve abnormal cellular proliferation and maturation, and nuclear atypia. According to the Armed Forces Institute of Pathology (AFIP) the lesion is classified as CIN 1 if the proliferation is
confined to the lower third of the epithelium. When the proliferation of atypical parabasal cells involves between one third and two thirds of the thickness of the epithelium, the lesion is classified as CIN 2 and more than two thirds CIN 3. The presence of significant nuclear atypia raises the grade of a lesion from respectively CIN 1 to CIN 2, or CIN 2 to CIN 3.20
Due to the morphologic diversity of CIN, a significant lack of intra-and interobserver reproducibility in the grading of CIN exists.20-23
FIGURE 2 The natural history of cervical dysplasia in the transformation zone. CIN 1, 2, 3, micro-invasive and invasive squamous cell carcinoma.
I II IIIa IIIb IV V
very mild mild moderate severe in situ
micro-invasive invasive
carcinoma dysplasia
CIN 1 CIN 2 CIN 3
Natural history of CIN
The grading system assumes that CIN represents a continuum of stages that may lead to the development of invasive squamous carcinoma.19
The assumption of a continuum of stages is supported by follow-up data that show a tendency of some dysplastic lesions to progress over time to a more advanced stage of CIN.24-27 However, not all cases of CIN progress. Natural
history studies conducted prior to the availability of treatment for CIN lesions suggested that most early CIN lesions (50-85%) will persist or possibly regress.28-30 The probability of progression to invasive cancer increases with
the severity of CIN.28;29;31 In those women who develop cervical cancer,
malignant progression is rarely rapid, more commonly taking many years or decades.17 This process is strongly influenced by the human papillomavirus.
Human papillomavirus infection
In the 19th century, the relatively low incidence of cervical cancer observed in virgins and nuns versus the high frequency of cervical cancer in prostitutes lead to the hypothesis of cervical cancer being a sexually
transmitted disease.32;33 Subsequently, the idea of the role of a sexually
transmitted disease in the development of cervical cancer was supported by several studies on the epidemiological profile of women with cervical cancer showing strong associations with promiscuity and early age of first sexual intercourse.34-42 Several infectious agents were proposed over the years
including syphilis, gonorrhoea, and type 2 herpes simplex virus (HSV-2).43
Zur Hausen first proposed a role for HPV in the development of cervical cancer.44 In the last decade, the role of HPV as the major infectious etiologic
agent in cervical carcinogenesis has been firmly established, human papillomaviruses being detected in virtually all cervical cancers.9;45
To date, over 100 different human papillomavirus types have been described, depending on the nucleotide sequence of the HPV genome. HPV types can be classified in a cutaneous and a mucosal group on the basis of the site of infection.46 In cervical carcinogenesis only HPV types of the
mucosal group are involved. These mucosal viruses fall into two broad categories: low risk or non-oncogenic types, associated with cervical condylomas and CIN 1; and high risk or oncogenic types found in CIN 2 and CIN 3 lesions, and cervical cancers.47 The most prevalent type in cervical
subtypes are HPV type 18 (15%), HPV type 45 (9%), HPV type 31 (6%) and HPV type 33 (3%).48
Exposure to HPV is a very common event, especially in young sexually active women as most infections occur early in one’s sexual life.49-51
Risk factors for HPV infection in sexually active women are age, years since first sexual intercourse and number of partners in the last years.50-55 In the
Netherlands, overall-, high risk and low risk HPV prevalence in cytological normal cervical smears were reported to be 4.6%, 3.3% and 1.0%
respectively.56 With regard to age, a peak prevalence of 19.6% for all HPVs
was found in women 25-29 years of age, which declined to a mean of 4.3% in women over 30 years.56 For Finnish females the lifetime risk of contracting
at least one HPV infection between ages 20 and 79 was reported to be 79%,57 whereas the lifetime risk for British females to contract an HPV
infection was estimated to be 80-85%.58 The vast majority of HPV infections
are cleared by the host immune system after 6 to 8 months and never present as warts or neoplasia.52;59-64
Only 20% of the infected women will develop CIN 1 or 2 and even a smaller percentage will develop progressing CIN or cancer.52;65 Persistent
presence of high risk HPV and a sustained viral load, are associated with a significantly increased risk of developing abnormal cytology, as well as the development, maintenance and progression of CIN disease.9;52;61;66-74 In fact,
only persistent infections with oncogenic HPV types lead to high grade CIN lesions and result in cervical cancer.75 Clearance of high risk HPV on the
other hand, is reported to precede regression of cervical lesions by an average of 3 months.76 Why some HPV infections are cleared and others are
not, remains subject of study.
HPV and cervical carcinogenesis
The HPV infection cycle is closely adjusted to the cell cycle of its epithelial target, the squamous epithelial cell (keratinocyte).77 Initiation of
productive HPV infection does not influence cell division nor does it disturb the differentiation process of the keratinocyte, while it moves up the
multilayered epithelium. As HPV infection does not destruct the keratinocytes, the chance of evading immediate detection and elimination by the immune system exists.77
In benign and low-grade cervical lesions, the HPV genome is
maintained in the episomal state, i.e. free in the nucleus. Only after persistent HPV infection functional integration of viral DNA into the host-cell
chromosomes may occur and the associated lesions appear to have an extremely low rate of spontaneous remission.78
Proteins encoded by HPV may contribute to malignant transformation by interference with regulators of cell growth. The viral genes of the HPV are divided in two functional groups, the early (E) genes and the late (L) genes, and a non-coding long control region (LCR). The early and late regions have several open reading frames (ORFs) resulting in translation of functional proteins. The early genes encode proteins that are expressed before the onset of viral replication, whereas the late genes (L1 and L2) encode the structural components of the viral particle. New HPV viral particles are released by wear and tear of the top layer of the epithelium during the final stage of the infection cycle. When HPV integrates into the host
chromosomes, the E2 ORF, that regulates transcription of E6 and E7, is disrupted, allowing the continued production of the E6 and E7 oncoproteins.79
High risk HPV E7 proteins bind and inactivate the Retinoblastoma (Rb) protein, whereas high risk HPV E6 proteins bind p53 and direct its rapid degradation.47;80 Both p53 and pRb have tumour suppressor functions and
are key regulatory switches in cell cycle regulation and apoptosis.80;81
Therefore, this process leads to impaired tumour-suppressor-gene function, resulting in abnormal DNA repair, decreased apoptosis, and eventual cell immortalisation.
Several molecular studies have supported the multi-step somatic mutation theory of carcinogenesis,82-84 implying that multiple molecular events
are needed before transformation to malignancy occurs.85 These events may
include proto-oncogen activation, inactivation of tumour suppressor genes and DNA repair genes. Infection with HPV is believed to be an early event in the multi-step process of cervical cancer, followed by the sequential
occurrence and selection of several molecular events, which may include mutations causing chromosomal alterations, loss of heterozygosity, proto-oncogene activation and telomerase activation.80;86 The consistent loss of
indication of the presence of a tumour suppressor gene, whereas telomerase activation may inhibit programmed cell death by apoptosis in case of DNA damage.80;87 The multi-step nature of cervical carcinogenesis is illustrated by
the longer period of time elapsing before CIN lesions may eventually progress to cancer.
Immunity to HPV related lesions of the cervix
Immunosurveillance is believed to play an important role in the control of genital HPV infections. In the majority of infected women HPV infection is believed to be transient as result of an anti-viral immune response that is capable of clearing viral infection.
The immune system consists of a non-specific (innate) arm and an Antigen (Ag) specific (adaptive) arm. Both arms are involved in the defence against viruses. The supposed importance of the innate immunity in the defence against HPV infections is underlined by the increased Natural Killer (NK) cell activity associated with the regression of HPV induced cervical lesions.88 The adaptive immunity, however, has the advantage of affording
specifically targeted responses and immunological memory. The adaptive immune system is divided into humoral (B-cells) and cellular immunity (T-cells).89 Humoral immune responses (mostly IgG) against HPV-16 E2, E4 ,
E6 and E7, as well as L1 and L2 have been reported.90-94 Recently, in women
prophylactic vaccination with virus-like particle L1 vaccines appeared to induce high levels of neutralising antibodies and effectively prevent incident and persistent infections with either HPV 16 or 18.95 These results are highly
promising, as a reduced risk of persistent infection will most probably decrease the incidence of cervical cancer. Yet, antibodies are not important effectors of regression of established HPV infections.96 The role of cellular
immunity in cervical carcinogenesis is supported by a higher frequency of HPV related lesions observed in immunocompromised individuals, such as transplant recipients, patients with acquired immunodeficiency syndrome, or congenital disorders of the immune system.97-100 Moreover, regressing genital
warts are infiltrated with increased numbers of activated T helper and cytotoxic T cells, suggesting that the spontaneous regression of most premalignant lesions is likely mediated by cellular immune responses.101
Indeed, healthy subjects show strong T helper cell type 1 (Th1) memory responses against HPV 16102;103, while women with CIN 3 or cancer display a
observations further indicate a potential role for Th cells in protection against HPV16-induced progressive disease. In rabbits papillomavirus-specific T-cell immunity was shown to protect against malignant transformation of virus-induced papillomas105;106 and in mice HPV16-specific T-cell immunity can
prevent the outgrowth of HPV16+ tumours.107;108 In addition, the influence of
T-cell immunity on HPV-related diseases is underscored by the results of vaccination studies. Effective HPV16 specific T-cell responses were detected in mice applying candidate vaccines targeting the E6 and E7 oncogenes. The possibility of therapeutic vaccination against HPV 16 positive tumours was demonstrated by the capacity of some of these vaccines to eradicate the tumour.109-112
Cofactors
Potential cofactors in cervical carcinogenesis include HLA type, immunosuppression or immuno-suppressive drugs, cigarette smoking, other sexual transmittable disease (STD) infections, especially HIV, hormonal effects of oral contraceptives and pregnancy, dietary deficiencies and chronic inflammation. Most of these cofactors appear to influence progression to CIN 3 by causing persistent HPV infection.47;62
It has become clear that most of the sexual behaviour parameters that were linked to cervical cancer in the past, such as the number of sexual partners, are merely reflecting the probability of HPV exposure.9 Age at first
sexual intercourse is currently being assessed.9
It has been proposed that the developing cervix (at the age of peri-menarche) or the healing cervix (as a consequence of deliveries, cervical trauma, or any other STD infection) are high risk situations for an HPV infection to reach the basal layer and establish a persistent infection.9;38;113
Multiparity has been associated with an increasing risk for cervical cancer. 114-119 However, less visible and no significant effects were reported as well.120-122
The mechanisms by which cigarette smoking may affect cervical cancer in women with HPV infection are not conclusively established, but may involve a direct effect of the tobacco metabolites, or an indirect effect related to immunosupression or reduced dietary antioxidants.118;123 A largely
consistent association between smoking and cervical cancer in the order of one to threefold is found in studies that adjusted for HPV DNA or restricted analyses to HPV positive women.9 The odds ratios tended to be higher in
more advanced grades of neoplasia, and in several studies a dose-response relationship with the amount of tobacco consumed was seen.114;115;121;123;124
Although a causal relationship between cigarette smoking and cervical cancer is considered plausible, the extent to which cigarette smoking can be considered as independent of HPV is not definitively assessed.125
The evidence for an association of cervical cancer with the use of oral or other hormonal contraceptives is not entirely consistent.9;114;116;121;126;127
Biases related to sexual behaviour, screening, and other factors could not be ruled out as possible alternative explanations for the associations found in some of the studies.127
Some STDs, including HIV, have repeatedly been reported to be associated with cervical cancer.99;128;129 Also, antibodies to HSV-2130;131 and
non-specific inflammatory changes132 have been related to increased risk for
cervical cancer in HPV positive women. Some have described HSV-2 as a potential co-factor in the pathogenesis of cervical cancer, either by
establishing or expanding HPV infection within the transformation zone.133
Other environmental risk factors under evaluation are socioeconomic status134 and nutritional factors.135 Socioeconomic status may be
Cervical screening and treatment of CIN in the Netherlands The present cervical cancer screening program in the Netherlands consists of 5-yearly cervical smears by the general practitioner in women aged 30-60 years.136 If cytology shows moderate dyskariosis (PAP 3A2) or
more, or twice very mild dyskariosis (PAP 2) or mild dyskariosis (PAP 3A1) (with a 6-month interval), women are referred to the gynaecologist for colposcopic examination. Biopsies are taken from all colposcopically suspect areas. If histology shows a CIN 2 or 3 lesion, the patient can be treated by a large loop excision of the transformation zone (LLETZ), laser evaporation, cryocoagulation, cone biopsy or hysterectomy, depending on the preference of both the patient and the physician, as well as on the experience of the latter. In general a LLETZ is preferred, as it can be performed in an outpatient procedure and enables histologic examination of the tissue removed. A CIN 1 lesion is usually not treated, but subjected to follow-up by half-yearly smears. Although a LLETZ is a relatively simple and safe procedure, complications such as excessive bleeding or infections may occur and therefore one should guard for over treatment. A cure rate of more than 90% can be achieved by a LLETZ, and cytological follow-up is advised after 6, 12, and 24 months to detect any residual or recurrent disease.137-140
The protocol for the clinical studies described in this thesis was based on the former screening program, which was still current at the start of the study. Women had cervical smears taken every three years and were referred in case of twice a PAP 3A1 (with a 3-month interval) or one PAP 3A2 or higher. CIN 2 lesions were often not treated, but subjected to follow-up by colposcopy and cervical smears.
Colposcopy
During physical examination the cervix can be examined in closer detail with the use of a colposcope, which is a microscope providing
application. During colposcopy the solution is applied repeatedly, as the effect vanishes after approximately one minute. A iodine solution (Schiller’s test) may additionally be applied to make the new SCJ visible. Iodine stains the glycogen in normal squamous epithelium brown, but has no effect on the glandular mucosa without glycogen.15;16
Psychoneuroimmunology
In the 19th century, Russian physicians investigated the hypothalamic function and influence on the immune system in rodents.141 This was the first
sustained program of research on brain-immune system interactions and involved Pavlovian conditioning of “immune reactions”. (Reviewed by Hull, Kopeloff, Korneva and Ader.)1;142-144 In the 1930’s, Selye, who is considered
as the founder of stress research, showed that physical stimuli which can damage an organism (later called ‘stressors’) inhibit the activity of some components of the immune system.145 In the 1970’s, more and more the
central nervous system, and the endocrine and immune system were considered as interactive units with regard to psychological and physical wellbeing.146 The term psychoimmunology was introduced by Solomon in
1964 in his paper “Emotions, immunity and disease: A speculative theoretical integration”.147 In 1975, a resurgence in interest in psychoneuroimmunology
was caused by Ader and Cohen, demonstrating classical conditioning of immune function in rats, thus following in the Russian foot steps.1;148 After
feeding the rats with a distinctively-flavoured drinking solution,
chemotherapeutics were injected and subsequently immune measures were taken. When the experiment was repeated with the distinctively-tasting substance but without injection of chemotherapeutics, the same
immunophenotypic changes were observed, indicating that a certain type of ‘learning’ experience had taken place with regard to the immune system.
The notion that the central nervous system is related to the human psyche is not hard to accept, considering the fact that the psyche is seated in the brain. Two pathways bridge the brain and the immune system: autonomic nervous system activity and neuroendocrine outflow from the pituitary. By now, there is evidence from a number of fields pointing toward bi-directional link between neural activity and altered immune responses.14;149;150 Studies in
rodents showed a causal relationship between a variety of stressors and subsequent suppression or enhancement of immune responses.150;151
and examination stress in medical students are associated with altered measures of immune reactivity, and in some instances with altered health status, but these two processes have not yet been linked causally.2 Changes
in immune function would constitute a critical link between psychosocial factors and an altered susceptibility to or progression of disease. However, this chain of psychophysiological events has not yet been firmly established. It remains to be conclusively demonstrated that an altered resistance to disease is a direct result of biologically relevant changes in immune function induced by psychosocial or stressful life experiences.14
Psychosocial factors and disease
Psychosocial factors have been subject of study in relation to several diseases and infections, such as diseases concerning the cardiovascular system, ulcerative colitis, infection with HIV or herpes, plane colds, auto-immune arthritis and allergies. With respect to cancer initiation and progression most studies were done on breast cancer patients, but
melanoma, lung cancer, rectal cancer and other types of cancer were studied as well. Types of cancer that are sensitive to hormonal and immunological factors, such as breast cancer and melanoma and virally-induced types of cancer such as cervical cancer, may be influenced via a neuro-hormonal-immunological pathway and may have preference in this field of study. Some evidence was found that helplessness and repression of negative emotions are factors promoting cancer initiation and progression, while denial/
minimising seemed to be associated with a favourable diagnosis. A low level of social support, chronic depression and having experienced loss events were sometimes linked to unfavourable diagnosis.152 However, there is a
number of negative studies on each of these subjects as well. Possible explanations for discrepancies between the studies may be that the number of patients studied was too low, biomedical and sociodemografic factors were not always controlled for, nor was sick role bias always taken into account.153
Additionally, psychosocial factors may show their effect only in the presence of other factors, such as older age, or in people with a biomedical risk factor such as smokers. Therefore, psychosocial factors should also be studied in relation with each other and in relation with relevant biomedical and
Psychosocial factors in relation to cervical cancer and its precursors
Whereas for many diseases it is difficult to study the role of psychosocial factors on the pathogenesis, the easily accessibility of the cervix offers a good possibility for study and follow-up.
Earlier studies reported a relationship between hopelessness, anxiety, depression (psychological distress), an increased use of repression and denial (more ‘passive coping’) and cervical cancer.154-156 Two studies on
psychosocial factors in relation to CIN reported a positive relationship between negatively-rated life events and CIN grade.157;158 A third study
showed an opposite relationship.158-160 It was suggested that differences in
socioeconomic status between the populations studied might have contributed to the discrepancy in these findings. Another explanation was given by differences in coping style; an active coping style was considered to protect patients against stress caused by negative life events.
Stressor-support-coping model for CIN and distress
Based on the earlier studies on psychosocial factors and CIN a stressor-support-coping model was developed by Goodkin to study the influence of life events, social support, and coping style on CIN.161;162 In this
model the influence of these psychosocial factors can be studied both separately as well as in interaction with each other or in interaction with biological factors such as HPV. In figure 4 the possible relationships between life events, social support, coping style and CIN grade are depicted. In the stressor-support-coping model, distress is both an outcome and a predictor variable, as the influence of life events, lack of social support, and coping style is considered to be, at least partly, mediated through effects on
psychological distress.150 Distress is the response that an individual develops
facing a stressor and is affected by the perception of the level of threat that the stressor poses.163
Sick role bias
associated with negative emotions, which may cause a bias when negative emotions, psychosocial problems or associated concepts such as
hopelessness are studied as predicting variables. This bias is known as sick role bias and may be reduced by gathering psychosocial data before the diagnosis is known. However, the suspicion of a patient that she has cancer may still play a role and should be controlled for.164 In addition, distress
caused by the abnormal cervical smear may be a measure to control for sick role bias. In studies on psychosocial factors in CIN the risks of the
occurrence of sick role bias are likely lower than in studies on psychosocial factors in cancer patients.
Outline of the thesis
This thesis presents clinical studies on the influence of psychosocial factors on the grade and course of CIN. The stressor-support-coping model described above was applied.
From February 1996 to February 1998, 393 women with an abnormal cervical smear of PAP 3A or higher were enrolled in the cross-sectional study at the colposcopy clinics of three participating hospitals: Leiden University Medical Centre, Medical Centre Haaglanden, location Westeinde Hospital and Leyenburg Hospital (two non-academic teaching hospitals in The Hague). All participants had a colposcopically-directed biopsy as part of the routine procedure.
As part of the ‘internal validity control’ of the clinical studies, all histological material was re-examined. In Chapter 2, we describe the results of this histopathological revision of the CIN diagnoses in the cross-sectional study. During the ‘revision sessions’, the grading criteria were discussed and consensus on the diagnoses was reached. From these histopathological revision sessions arose the idea of visualising pathologists’ scanning patterns in the grading of CIN. To this end, we used EyeCatcher, an eye gaze tracking device, described in Chapter 3, to visualise and compare the scanning patterns of five pathologists while they graded two projections of CIN. Although a number of studies based on physicians’ eye gaze exist for radiology and dermatology, studies on eye gaze in pathology, one of the main areas within medicine centering on the study of images, are rare. In fact, this was the first study to actually visualise the scanning patterns of pathologists. We created density cloud images as well as graphical representations of the scanning patterns, and a questionnaire and interview provided information on the following steps in the diagnostic process of CIN.
the grade of CIN, controlling for HPV status and sick role bias.
Chapter 5 focuses on the longitudinal clinical study and describes the results of the follow-up of 94 women with a CIN 1 or 2 outcome in the cross-sectional study. These women were followed by half-yearly colposcopy and cervical smears during a maximum period of 2.25 years. A shorter set of questionnaires on the psychosocial factors was completed during the follow-up visits, followed by an interview on the impact of the life events, as well as on changes in lifestyle and gynaecological factors. Negatively-rated life events, social support, and coping style were studied separately and in interaction, in relation to time till progression and time till regression of the CIN lesion respectively. HPV status and sick role bias were controlled for.
In Chapter 6, we describe the results of a separate study on non-expression of emotions (repression) in relation to the CIN grade after the abnormal cervical smear. For this purpose, a sub sample of the participants in the cross-sectional study was asked to fill out additional questionnaires on repression. Repression was studied separately and in interaction with life events, social support, age and HPV.
During the multi-centre clinical studies, violations of the protocol occurred. In Chapter 7, we determined early interventions that occurred in discordance with the protocol during the follow-up study. The reasons for these interventions were registered, and acceptable and unacceptable reasons were distinguished. The histopathogical outcome after the early interventions was recorded and the patients’ influence on the non-adherence to the protocol was studied.
CHAPTER 2
Revision of histology
E. Stella M. Tiersma, MD
Abstract
Objective: The grading of cervical intraepithelial neoplasia (CIN) is prone to interobserver variation. We subjected all cervical biopsies taken as part of a multi-centre clinical study on CIN to histopathological revision. The revision was performed to control for a possible effect of interobserver as well as interlaboratory variation in the grading of CIN on the study outcome. Methods: All histological material derived from colposcopically-directed biopsies in 393 women with an abnormal cervical smear in three participating hospitals was revised by a panel consisting of two pathologists with a broad experience in CIN grading and the author of this thesis. Consensus was reached by discussion. The criteria for CIN grading as described by the Armed Forces Institute of Pathology (AFIP) were applied.
Results: In 23% of the cases the histopathological revision resulted in an altered diagnosis compared to the original diagnosis of the pathology
department of the participating hospital. For ”hospital 1” the altered diagnoses mainly showed a fall in CIN grade, for “hospital 2” an equally high percentage of altered diagnoses showed a higher CIN grade as a lower CIN grade and for “hospital 3” in most cases a rise in CIN grade was noted. The lowest percentage of altered diagnoses was found for diagnosing CIN 3, followed by ‘no CIN’.
Introduction
Various studies have shown that the grading of cervical dysplasia is prone to both inter- as intraobserver variation.21;22;165-173 The Cervical
Intraepithelial Neoplasia (CIN) terminology proposed by Richart categorises mild dysplasia as CIN 1, moderate dysplasia as CIN 2, and severe dysplasia and carcinoma in situ combined as CIN 3.174 Most studies on inter- and
intraobserver reproducibility have concluded that agreement diagnosing CIN 3 and higher is fair to good, whereas the reproducibility of CIN 1 and
especially CIN 2 is poor.21;22;165;171;175
The Bethesda system, promulgated in the United States in 1989, combines CIN 1 and cellular evidence of HPV effects as Low Grade Squamous Intraepithelial Lesions (LSIL) and CIN 2 and 3 as High Grade Intraepithelial Lesions (HSIL).176 The interobserver agreement applying the
Bethesda system in the grading of cervical dysplasia, is reported to be better, but still poor.169;170 Only agreement within the HSIL group was reported to be
excellent.177 It has been argued that the Bethesda system causes over
treatment ignoring the natural history of cervical dysplasia by combining CIN 2 with CIN 3 and raises the costs of management of lesser degrees of cervical epithelial abnormalities.178 In the Netherlands the CIN terminology is
generally accepted and applied.
Recently, we performed a multicentre clinical study on the influence of psychosocial factors on the grade of CIN in women with an abnormal cervical smear. Because of the known interobserver variation in CIN grading, we performed a histopathological revision of all biopsy material, to reduce a possible effect of interlaboratory or interobserver variation on the study outcome in secondary analyses. This chapter describes the results of the histopathological revision in the clinical study.
Methods
to participate in the study.
All biopsy samples were initially graded at the pathology departments of the participating hospitals. These original CIN diagnoses were later revised by a panel consisting of two pathologists with a broad experience in CIN grading and the author of this thesis. All biopsies were first revised by one pathologist and the author. The second pathologist was consulted in case of discrepancies. In these cases consensus was reached by further discussion. The criteria for CIN grading as described by the Armed Forces Institute of Pathology (AFIP) were applied.20
Results
In table 1 the results of the histopathological revision of the biopsy samples are summarised for each hospital site. In some cases the diagnosis after histopathological revision was reported as “insufficient quality for diagnosis” or as “not representative”, i.e. when the transformation zone was not included while no signs of CIN were found. The histopathological revision resulted in an altered diagnosis in 23% of the cases. In the academic hospital (“hospital 1”) the diagnosis was altered in 19% of all cases, with a lower CIN grade as outcome in 70% of these cases. In “hospital 2” (one of the non-academical teaching hospitals) revision led to a different diagnosis in 23% of all revised cases, resulting in a higher CIN grade in 38% of these cases and a lower CIN grade in 38% of these cases. In “hospital 3” (the other non-academic teaching hospital) the revision led to an altered diagnosis in 31% of the cases, resulting in a higher CIN grade in 57% of these cases. In table 2 the percentages of altered diagnoses for the different CIN grades in the original diagnoses are shown.
Discussion
The results of the histopathological revision of the biopsy samples of the clinical study show remarkable differences between the hospitals sites. Whereas for “hospital 1” the altered diagnoses mainly showed a fall in CIN grade, for “hospital 3” in most cases a rise in CIN grade was noted.
seemed to exist for diagnosing CIN 3, followed by ‘no CIN’. Pieters described that pathologists working at the same pathology department did not
necessarily show more interobserver agreement than pathologists from separate pathology departments. On the other hand, he reported a tendency of some pathology departments to upgrade CIN lesions compared to other pathology departments.179 This tendency was also noted in our sample. In
practice, this tendency might mean that women with the same type of CIN lesion are treated in some hospitals whereas this is not the case in other hospitals.
Interobserver variance will predominantly lead to over treatment, as virtually all women with a CIN lesion are subjected to cytological follow-up and will finally be treated in case of initial under treatment. In general, treatment takes place by a large loop excision of the transformation zone (LLETZ) and less frequently by a cold knife conisation. Although these procedures are relatively simple and safe, complications such as excessive bleedings and infection may occur and over treatment should therefore be prevented. However, interobserver variance may also lead to under treatment, as women who are incorrectly diagnosed as having no CIN, are lost to follow-up for a maximum period of five years, in the current Dutch cervical screening program. In those patients a CIN 1 lesion is most likely, which has a high tendency of regression. Still, in 10% of the patients with a CIN 1 lesion progression to CIN 3 may occur, and in 1% progression to invasive cancer.30
The interobserver variation in CIN grading may partly be explained by the fact that an arbitrary division is made in what is actually a continuous process.166 Moreover, CIN grading is dependent on both cellular atypia and
proliferation rate and thus on the importance attributed to these criteria by each pathologist.166 Interobserver agreement in the grading of CIN was
reported to improve after reaching theoretical consensus on which grading criteria to apply by discussion.167 Agreement might further benefit from
practical consensus, reached by training sessions behind the microscope, although this last step did not lead to a significant improvement in
interobserver agreement in a study by de Vet.168 Nevertheless, such
CHAPTER 3
Visualising scanning patterns of pathologists in
the grading of cervical intraepithelial neoplasia
E. Stella M. Tiersma, MD
Alexander A.W. Peters, MD, PhD Heralt A. Mooij, PhD
Gert Jan Fleuren, MD, PhD
Abstract
Objective: To investigate how effectively eye tracking devices can visualise scanning patterns of pathologists, for application in studies on diagnostic decision making.
Methods: EyeCatcher, an eye tracking device, was applied to visualise and compare the scanning patterns of five pathologists while they graded two projections of cervical intraepithelial neoplasia. Density cloud images were created from the scanning patterns. A questionnaire and interview provided information on the following steps in the diagnostic process.
Introduction
The study of images plays a crucial role in pathology. However, diagnoses based on images are often prone to interobserver variation. Interobserver variation may arise at three levels, which are not only influenced by prior knowledge and experience, but also depend on the applied diagnostic criteria. The first level concerns the visual search of the image, the second level involves the interpretation of the perceived visual information, and finally, the third level concerns the way the collected information is combined to reach a diagnosis. Investigation of each of these levels is necessary to study the origin of the interobserver variation.
In radiology, earlier studies analysed the scanning patterns applied by physicians to detect and interpret breast and lung tumours.180-183 Recently, a
study in this field was also performed in pathology, concentrating on the development of visual diagnostic expertise in breast pathology.184 In this
study, the subjects’ search of a slide was captured on video through the microscope. However, in contrast to the radiology studies, no eye tracking systems were applied. Eye tracking systems enable the visualisation of the eye points of gaze of a subject and provide possibilities to create
computerised graphic representations of scanning patterns.
We investigated whether eye gaze tracking devices could help to study diagnostic decision making in pathology and focused on the grading of cervical intraepithelial neoplasia (CIN), one of the fields within pathology often prone to interobserver variation.21;185-188 We combined eye gaze registration
with a questionnaire and an interview to gain information on the influence of each criterion which contributes to the final diagnosis.
Methods
The eye gaze tracking system, EyeCatcher, used in this study uses infrared light reflections from the eye. Five pathologists experienced in histopathological grading of CIN were asked to grade two CIN lesions while wearing a light weight helmet. The pathologists sat in front of a screen displaying successively two images of the CIN lesions (figures 1 and 2). The helmet was mounted with an infrared camera and infrared light source, a mirror for reflection of the infrared light, and a miniature video camera (figure 3). The infrared light was reflected by the mirror on to the eye and
that infrared light is not visible to the human eye, the pathologists were unaware of the light beam.
This video image of the eye showed the position of the pupil and the location of the ‘hot spot’. The hot spot is the cornea’s reflection of the infrared light beam. From these two features, the computational part of the system determined the orientation of the eye in relation to the head. The miniature video camera was focused on the projection of the CIN lesion and the video image coming from the camera moved in synchronisation with the
pathologist’s head.
The data representing the eye orientation were merged with the images from the miniature video camera. Thus, a video image of the CIN lesion was obtained with a pair of cross hairs indicating the points of gaze in time.
The video images and eye orientation data in digital form were used in a computer analysis to calculate the X and Y co-ordinates of the points of gaze of the pathologists. Using these co-ordinates, a graphic representation of the scanning pattern was created in which dots represented the points of gaze every 20 milliseconds. These graphics were blurred using a mathematical calculation to obtain density clouds representing the total amount of attention for each area. By use of colours, areas of high attention were highlighted.
In a briefing beforehand, the pathologists were presented with the Armed Forced Institute of Pathology (AFIP) criteria for the grading of CIN and were requested to grade only according to these descriptions.20
Each pathologist was assigned 45 seconds to grade the lesion and was
requested to say “yes” as soon as he had reached a diagnosis. At 45 seconds, the CIN projection was covered and the pathologist was asked to answer several questions concerning the interpretation of the projected lesion. The same procedure was repeated for the second CIN lesion.
Afterwards, the pathologists were asked to comment on the test, criteria, and diagnoses during a structured interview.
With regard to the three levels in the diagnostic process, the EyeCatcher was used to register the first step, the scanning pattern, whereas the
FIGURE 2 Cervical intraepithelial neoplasia lesion 2; haemotoxylin and eosin staining.
FIGURE 3
questionnaire and interview provided information on the following two steps, the interpretation and combination of the information, particularly with regard to the respective grading criteria.
R
esultsRegistration of the points of gaze
Figure 4 shows the density cloud images of the scanning pattern of each pathologist for both CIN lesions until time of diagnosis.
Lesion 1
In the first lesion, a prominent feature was present in the form of a mitotic figure in the middle of the projection. For all pathologists this was an
important point of focus. Only the graphics of pathologist 3 and 4 also revealed other areas of high attention.
Pathologists 2 and 5 examined the image in a scanning matter shortly focusing on many different points. This is represented by a large amount of light blue cloud scattered over the image. Pathologists 1, 3 and 4 examined the image in a more direct way and focused on specific points only. In contrast to the dark blue and green/red clouds, there are not many light blue areas present in the graphics of these pathologists. This direct way of looking is most apparent for pathologist 4, who took only seven seconds to gather the information for diagnosis in a highly selective manner.
Lesion 2
All pathologists needed more time to reach a diagnosis for lesion 2 than for lesion 1. For all pathologists except pathologist 3, more green/red areas are present in the graphics for lesion 2 than for lesion 1, which means that for lesion 2 the attention was more equally spread over the entire image. In lesion 2 as a whole, more blue areas are present in the graphics in accordance with the longer diagnostic time.
Pathologists 1, 3 and 4 showed more or less the same direct way of looking at lesion 2 as at lesion 1, whereas pathologist 3 showed special attention to the higher regions of the image.
Registration of the order of the scanning procedure
The order in which the images were analysed can be studied by
of the five pathologists often exhibited noticeable differences. Figure 5 shows a clear example of two different structures of scanning patterns of
pathologists (1 and 4) looking at lesion 2.
The eye gaze of pathologist 1 was initially directed at the bottom of the lesion, then moved up to the surface and subsequently slowly moved down again with mostly horizontal movements. Pathologist 2 showed a tidy structure of both horizontal and vertical movements. The movements of the eye gaze below may indicate a quick scanning of the basement membrane.
Interpretation and diagnostic decision making
As shown in table 1, the highest variation in the interpretation of the criteria was found for the level of nuclear atypia and the level of abnormal maturation. The answers concerning the mitotic activity did not vary, whereas the answers concerning the atypical character of the mitoses varied for both lesions (data not shown). For lesion 1 all pathologists but one reported the presence of atypical mitoses. Pathologist 2 refrained from answering. For lesion 2, only pathologist 3 and 4 reported the presence of atypical mitoses.
On three occasions, the final diagnosis was not reached in accordance with the AFIP criteria. One of the pathologists graded lesion 1 as CIN 3, although he reported the presence of abnormal maturation, nuclear atypia, and mitoses limited to no more than two thirds of the epithelium. Another pathologist graded lesion 1 as CIN 2, although reporting the presence of abnormal maturation and nuclear atypia extending to three thirds of the epithelium. The same pathologist graded lesion 2 as CIN 1 in the presence of nuclear atypia in three thirds of the epithelium.
Discussion
We found that the pathologists studied the CIN lesions in different ways, with regard to the areas they examined, the amount of study time for each area and the order in which the areas were studied. The amount of time needed to reach a diagnosis also varied considerably.
between the pathologists were remarkable, especially considering the fact that all pathologists were requested to grade according to the same criteria.
The judgement of the levels to which abnormal proliferation and nuclear atypia were present varied considerably. Most of the questions raised by the pathologists in the questionnaire and interview concerned the interpretation of the level to which nuclear atypia was present, in addition to the extent to which it should influence the final diagnosis. The AFIP states that if
pronounced nuclear atypia is noted in the presence of normal maturation in the same layer of the epithelium, and if this nuclear atypia is not the kind of atypia associated with productive human papillomavirus (HPV) infection, the CIN level should be upgraded.20 However, most pathologists found it difficult
to follow this advice, especially during their judgement of lesion 2. They doubted whether the nuclear atypia was HPV related and whether the nuclear atypia was substantial enough to upgrade the CIN level. This was the reason for the spread in diagnoses for lesion 2.
Studying the density cloud images one can speculate to a certain extent about the criteria studied in the specific areas for both lesions. In lesion 1, the mitotic figure in the middle is obviously a very recognisable criterion. In lesion 2, more red/green areas are present in the graphics. This indicates that the attention of the pathologists was more equally spread over the entire image, probably because of the lack of a main point of attraction. With regard to the other areas that received more than average attention, the criteria studied can also be speculated about. Particularly in the upper layers of the lesions, it TABLE 1 Interpretation and diagnostic decision making
for lesion 1 and 2
is very likely that the nuclear atypia was the criterion studied, because the cells here have irregular nuclei, although they are differentiating rather well. Pathologist 3 in particular often focused on these upper layers of the epithelium, most evidently for lesion 2. He was one of the three
pathologists who questioned the nuclear atypia afterwards during the interview.
We found an impressive spread of diagnoses. The broad range of diagnoses found is especially remarkable because only one selected image was shown for each case. In a normal clinical situation,
pathologists often base their judgement on many more images and select the most relevant area(s) within the biopsy, which probably contributes towards the interobserver variation. In our study, by selecting one single image, the number of potential influencing variables was minimised, thus allowing us to study the origin of the interobserver variation as closely as possible.
Hypothetically, the wide spread in diagnoses found in our study may arise from the differences in scanning patterns. However, the different scanning patterns are probably a reflection of differences in the way the images were interpreted and the way the grading criteria were applied by the pathologists.
By showing the areas of high attention during the visual search by the pathologist, eye tracking devices provide an excellent basis for further discussion on the interpretation and grading criteria. Eye tracking devices as such may be a valuable asset in pathology, for detailed studies on diagnostic decision making, in addition to the development of training and quality control programs for pathologists.
Acknowledgements:
CHAPTER 4
Psychosocial factors and the grade of cervical
intraepithelial neoplasia: a semi-prospective
study
E. Stella M. Tiersma, MD Marije L. van der Lee, MA Alexander A.W. Peters, MD, PhD Adriaan P. Visser PhD
Gert Jan Fleuren, MD, PhD Bert Garssen PhD
Karin M. van Leeuwen, MD, PhD Saskia le Cessie, PhD
Karl Goodkin MD, PhD
Abstract
Objective: To study the influence of psychosocial factors on the grade of cervical intraepithelial neoplasia.
Methods: The influence of psychosocial factors on the grade of cervical intraepithelial neoplasia (CIN) was studied in a group of 342 patients with an abnormal cervical smear. Participants completed a set of questionnaires after colposcopically-directed biopsy before knowing the biopsy result. Negatively-rated life events, social support, and coping style were studied in relation to distress and grade of CIN. Infection with human papillomavirus (HPV) types was controlled for, as well as sick role bias caused by suspicion of having cervical cancer and distress due to the abnormal cervical smear.
Results: Negatively-rated life events, lack of social support and emotional coping were significant predictors for level of distress. No significant
relationship was found, however, between the psychosocial factors and grade of CIN.
Introduction
Cervical carcinogenesis is a multifactorial process strongly related to the human papillomavirus (HPV).8;9;128 Other important factors involved in the
carcinogenic process are genetic changes and an inefficient host immune system.189;190
Although the role of psychosocial factors in the initiation and progression of cancer has been studied in many well-designed prospective studies, their role is still uncertain.153;191 Psychosocial factors may exert an influence by way
of a psychoneuroimmunologic (PNI) pathway,161;162 which is a particularly
tempting hypothesis for cervical carcinogenesis because of the involvement of HPV in the development of this disease. Psychosocial factors can affect the immune function,150 whereas an optimal functioning immune system may
reduce the influence of HPV. Insight into the role of psychosocial factors may contribute to the development of preventive strategies for cervical cancer.
Findings from earlier studies suggested that hopelessness, anxiety, depression (psychological distress) and an increased use of repression and denial (more ‘passive coping’) were related to cervical cancer.154-156
One of the potential problems of studies on psychosocial factors in cancer patients may be a bias caused by the patient’s knowledge or suspicion of having cancer. The assumption of a patient that he or she has cancermay be associated with negative emotions.164 This could cause ‘sick role bias’ if
negative emotions, psychological problems or associated concepts such as hopelessness, are chosen as predicting variables. The possibility of sick role bias is strongly reduced in cervical intraepithelial neoplasia (CIN), the precancerous stage of cervical cancer, in particular when the psychosocial data is gathered before knowledge of diagnosis (‘semi-prospective design’), as patients do not yet know whether they are affected. Nevertheless, it is always recommendable to control for sick role bias caused by the suspicion of having cancer and by distress caused by the abnormal cervical smear.
In two studies on psychosocial factors in CIN, a positive relationship between negative life events and CIN grade was found.157;158 However, a third
study showed the opposite.158;159 The discrepant findings were explained by
the differences in the socioeconomic status between the samples and by a difference in coping style; an active coping style was considered to protect the patients against stress caused by negative life events. Based on the findings in the earlier studies on CIN and cervical cancer154-159 a
stressor-support-coping model was developed.161;162 This model presumes that
unsatisfactory social support and the use of a more passive coping style, are associated with a higher level of distress and a greater likelihood of
progression of CIN (figure 1). Distress is both a predictor and an outcome variable in this model, as the effect of psychosocial factors on health may, at least in part, be mediated through effects on psychological distress.150
In this present study, we applied the stressor-support-coping model in a semi-prospective setting to test whether psychosocial factors can predict the level of distress and the grade of CIN in patients with an abnormal cervical smear.
Methods
During the period from February 1996 to February 1998, 393 patients were enrolled in the study at three participating hospitals: Leiden University Medical Centre and two non-academic teaching hospitals (Leyenburg Hospital and Westeinde Hospital, The Hague). The inclusion criteria were a colposcopically-directed biopsy as a result of an abnormal cervical smear of mild to moderate dyskaryosis (Pap 3A) or higher, no concomitant cancer, and sufficient Dutch language skills. Patients were excluded if they had had treatment in the cervix within the last two years, a recent pregnancy up to six weeks prior to study entrance, current symptoms referable to cervical cancer or related complications, an abnormal cervical smear based on cylindrical abnormalities only, a history of or current major psychiatric diagnosis, alcohol or substance dependence or any previous history of intravenous drug use. A standard colposcopic assessment with acetic acid and iodine solutions was carried out by the colposcopist, after which written informed consent from the
patients was acquired. The study was approved by the Medical Ethical Committees of the participating hospitals.
Participants were asked to complete a set of questionnaires immediately after the colposcopy, or at least within two weeks thereafter and before knowing the biopsy result. The set of questionnaires was composed of separate questionnaires on life events, social support, coping style, distress, and biographical factors, including gynaecological history, lifestyle and sick role bias. During an interview, directly following completion of questionnaires, the impact of the life events was assessed. The average time to complete the questionnaire and interview was approximately 45 minutes.
Psychosocial predictors
Five psychosocial predictor variables were used in the analyses: social support, total number of negative life events and three coping styles:
‘problem solving coping’, ‘palliative coping’ and ‘emotion-focused coping’. The total number of life events over the past year was assessed by the Dutch version of the Life Experience Survey (LES).192 One question was added
concerning negative life events that had occurred more than a year ago but which still had a strong negative impact on the patient’s life. Following the completion of the LES, patients were asked about the impact of the indicated life events on their lives. This impact could either be negative, neutral, or positive. The total number of negatively-rated life events was used as a measure in the analyses. Social support was assessed by a modified version of the Social Support Questionnaire (SSQ, Dutch adaptation by Bleiker).193 A
second questionnaire on social support (Van Sonderen List) was added to test the validity of the applied modified SSQ. It appeared that the SSQ total score was significantly related to the Availability of Support scale (r= 0.54) and the Satisfaction with Support scale (r= 0.62) of the Van Sonderen List.194
To measure the patient’s coping capacity, the 19-item version of the Dutch Utrecht Coping List (UCL)195 was used. The psychometric characteristics of
the LES, SSQ, Van Sonderen List and the UCL appeared to be adequate. 192-195
Control variables
The Health Behaviour Questionnaire196 was used to assess demographic
these potential confounding variables have been chosen based on earlier research.161 Some variables were added because of their association with
immune function (physical exercise hours, sleep quality, alcohol use, substance use, caffeine intake).161;162 Further, we controlled for possible
effects of language or culture on the data collection and for possible effects of the hospital site on the CIN diagnosis by the different pathology
departments. A list of all control variables concerning the distress outcome and the CIN outcome is shown in table 1.
Sick role variables
The sick role bias variables were measured by three questions. The first two questions were ‘How would you estimate your chances of having cervical cancer?’ and ‘To what extent did you feel distressed during the first week after you learned about the abnormal cervical smear?’ Patients were asked to answer on a visual analogue scale, ranging from 0% to 100%. The third question concerned the patient’s perception of the gynaecologist’s opinion during colposcopy. In response to the question ‘How did the gynaecologist judge your cervix during colposcopy?’, the possible answers were (1) don’t know, (2) normal, (3) abnormal, (4) very abnormal.
Outcome measures
The first outcome measure, distress, was assessed by using the total mood disturbance score on the Dutch version of the Profile Of Mood States (POMS).205The second outcome measure was the ‘CIN grade’. The histology
outcome ranged from ‘no CIN’, CIN 1 (mild dysplasia), CIN 2 (moderate dysplasia), CIN 3 (severe dysplasia) to micro-invasive carcinoma (less than 3 mm depth) and invasive carcinoma. The original diagnoses from the hospital pathology department where the patients entered the study were used in the primary analysis. The biopsies were revised by a panel consisting of the first author and two pathologists with a broad experience in CIN grading. All biopsies were revised by the first author and one pathologist. The other pathologist was consulted in case of discrepancies. A secondary analysis was performed using the revised diagnoses.
HPV typing
microscopy. The CPI and CP-IIG consensus primers are directed against the E1 open reading frame, which is highly conserved among the HPV genome and enables the detection of a broad spectrum of different HPV types.197 An
initial
β
-globin PCR was performed on all material to confirm suitable quality of the DNA for PCR.198 After purification using the ‘EasyPrep’ kit (Pharmacia),the consensus-PCR products were subjected to sequence analysis to determine the HPV genotype.
For nucleotide sequence analysis and comparisons, the programs Seqed and Fasta of the Genetics Computer Group sequence analysis software package (GenBank DNA database, version 8.1, Wisconsin) were used. For the regression analysis, patients with a high risk HPV type were placed in the ‘high risk HPV group’ and patients with a low risk HPV type and patients with a negative HPV result were placed in the ‘low risk HPV/HPV negative group’.46
Statistical analysis
Reduction in number of control variables
Analyses to reduce the number of control variables were performed separately for the distress and the CIN outcome. Spearman’s correlations between the two outcome measures and all potential control variables were used to select only those variables that were significantly associated with outcome at a p< 0.20 level. To keep only independent control variables, a further selection was made based on the results of a linear regression analysis (p< 0.20) using the remaining control variables to predict the outcome variable (table 1).161 The sick role bias variables and the number of
prior cervical smears were not considered as ordinary control variables and were entered separately in the final regression analyses and without prior selection.
Testing the hypotheses
main analyses were repeated, while each of the three sick role bias variables was added in separate analyses, with the number of prior cervical smears. In a final series of analyses, we tested whether psychosocial factors exert their influence only in interaction. To that end, the interaction terms ‘life events with active coping’ and ‘life events with social support’ were added to the predictor variables in additional analyses.
Results
A total of 393 patients were included in the study. A group of 31 women, who were originally included and assessed, were later excluded from the analyses because they appeared to fulfil the exclusion criteria. A further 20 women were excluded from the analyses because a valid histology report was lacking. This reduced our study sample to 342 patients. There were no significant differences in any of our predictor or control variables between the excluded group and our final study sample group, except for a small
difference in mean age (M= 39.3 years for the excluded group and M= 36.2 years for the analysed sample, p< 0.05. Data not shown).
Histology was obtained by a punch biopsy on 304 patients, by an endocervical curettage on six patients, by a Large Loop Excision of the Transformation Zone (LLETZ) on 31 patients and by a conisation on one patient.
The original CIN diagnoses of these 342 women were: ‘no CIN’ (n= 71), CIN 1 (n= 72), CIN 2 (n= 71), CIN 3 (n= 124), micro-invasive carcinoma (n= 1) and invasive squamocellular carcinoma (n= 3). Four samples that were repeatedly negative in the ß-globin PCR were considered of insufficient quality for HPV analysis and were withdrawn from the regression analyses on the CIN outcome. Another ten patients were withdrawn from the regression analyses on the CIN outcome as the biopsy tissue no longer contained the CIN lesion. Of the 328 samples amplified with the CPI/IIG primers, 242 (74%) were positive. The results of the sequence analysis of the 242 positive samples showed 229 samples with high risk HPV types (HPV 16, 18, 31, 33, 35, 45, 51, 52, 53, 56, 58, 59, 66, 68, 70) and five samples with low risk HPV types (6, 11, 26 and 55). In eight cases the sequence could not be identified. These eight cases were also withdrawn from the regression analyses on the CIN outcome.
following control variables for distress as the outcome measure: socioeconomic status, whether or not born in the Netherlands, vitamin supplement intake, smoking, current drug use, and physical exercise. For CIN grade as the outcome measure, the following control variables were selected: whether or not born in the Netherlands, HPV group, age, and hospital site. Although socioeconomic status was not selected in our procedure, it was decided to maintain it as a control variable, because socioeconomic status affected CIN grades in other studies.
Regression analysis on level of distress
For the 342 women, a multiple linear regression analysis was performed on the level of distress. As predicted by the stressor-support-coping model, the analytical model with the psychosocial predictor variables and selected control variables significantly predicted the level of overall psychological distress (F for the total model= 24.5; p< 0.001), accounting for a total of 31% of the variation (table 2, analysis II). A higher number of negative life events and a higher score on emotional coping were associated with higher levels of distress (b= 7.98; p< 0.01 and b= 1.00; p< 0.01 respectively), and more social support (b= -1.86; p< 0.01) was associated with lower levels of distress. ‘Palliative coping’ and ‘problem solving coping’, however, did not significantly predict distress. When the psychosocial predictor variables alone were entered in the model, 26% of the variation in distress outcome was predicted (table 2, analysis I). In additional analyses, interaction terms, number of cervical smears, and sick role bias variables were added to the main analysis (table 2, analyses III).
The following control variables and sick role bias variables
