The Epidemiology of Cervical Cancer in Khon Kaen, Northeast Thailand

Abstract

Kohdunkaulasyövän esiintyvyys ja ennuste Thaimaassa

Tämä väitöskirja koostuu kohdunkaulasyövän epidemiologiaa käsittelevistä tutkimuksista Khon Kaenin alueella Koillis-Thaimaassa. Työ koostuu kolmesta toisiinsa liittyvästä osasta: kohdunkaulasyövän kansanterveydellisen merkityksen arvioinnista (kohdunkaulasyövän ilmaantuvuus vuosina 1985-1999), kolmesta eri tutkimusasetelmaan perustuvasta epidemiologisesta tutkimuksesta (sairaalapohjainen sekä kaksi kohortinsisäistä tapaus-verrokkitutkimusta) sekä hoidon tulosten arvioinnista potilaiden eloonjäämisen perusteella.

Tutkimuksen kolmen eri osan tulokset voidaan tiivistää seuraavasti:

Kohdunkaulasyövän ilmaantuvuus Khon Kaenissa on melko korkea ja tilanne on pysynyt lähes muuttumattomana viimeisen 15 vuoden ajan. Syöpä todetaan useammin levinneessä vaiheessa kuin Yhdysvalloissa ja Euroopassa. Khon Kaenin suurimman ilmaantuvuuden alueella syövän yleisyys on kaksinkertainen pienimmän esiintyvyyden alueeseen verrattuna. Tapauksista 80% on levyepiteelisyöpiä. Kunnes papilloomavirusrokote on saatavissa, parhaat keinot kohdunkaulasyövän torjumiseksi ovat varhainen toteaminen ja hoito.

Kolmen epidemiologisen tutkimuksen tulokset vastasivat pääosin muualla tehtyjä tutkimuksia: raskauksien lukumäärä ja ensisynnytysikä liittyivät kohdunkaulasyövän vaaraan. Toisaalta monen muun perinteisen vaaratekijän, kuten ikä ensimmäisen yhdynnän aikana, seksikumppaneiden määrä ja mahdolliset aiemmat sukupuolitaudit, ei voitu osoittaa vaikuttavan kohdunkaulasyövän vaaraan Thaimaassa. Poikkeavat tulokset voivat johtua tutkittavien aineistojen pienestä koosta tai siitä, että Khon Kaenin väestössä kohdunkaulasyövän syytekijät ovat erilaisia. Sairaalapohjaisen aineiston etuna oli potilaiden ja verrokkien suuri määrä sekä keräämisen vaivattomuus ja alhaiset kustannukset. Asetelma on kuitenkin altis harhalle joka saattaa vääristää tuloksia. Altistuslähtöinen seurantatutkimus on luotettavampi ja sen avulla voidaan tutkia vaikutusta moniin eri sairauksiin. Tämä lähestymistapa vaatii kuitenkin paljon aikaa ja resursseja maissa joissa ei ole tietolähteenä yhdisteltäviä rekistereitä.

Kohdunkaulasyöpäpotilaiden viiden vuoden elossaololuku Khon Kaenissa on alempi kuin vauraammissa maissa. Tärkein ennustetekijä on sairauden levinneisyysaste, mikä viittaa siihen, että varhaistoteaminen on hoitomuotoa tärkeämpää eloonjäämisen kannalta. Väitöskirjassa verrataan myös eri menetelmiä eloonjäämislukujen arvioinnissa.

This dissertation describes a series of investigations into the epidemiology of cancer of the cervix uteri (cervical cancer) in Khon Kaen Province, Northeast Thailand. There were three linked components: an evaluation of the magnitude of problem presented by cervical cancer (incidence of cervical cancer during 1985 to 1999), investigation of the risk factors in this population, using different study designs (hospital-based case-control study and two nested case-control studies within a cohort study), and a study of the clinical outcome (in terms of survival) of cervical cancer patients. The background, methods, and results were presented, and the implications of the results discussed.

There were 1848 cases of invasive cervical cancer in Khon Kaen during 1985 to 1999. The incidence was moderately high, with little change over the last 15 years. The overall age-adjusted incidence rates was 16.8 per 100 000 person-years. There was about a two-fold variation in incidence between districts within the province. About 15% of cases were in localised stage, which means that cases are diagnosed much later in the course of disease than in developed countries such as the United States and Europe. 80% of cases were squamous cell carcinomas. Until a vaccine is available, which may prevent infection with HPV, control through early detection and treatment remains the best approach reducing incidence and death.

The hospital-based case-control study consisted of 179 cases and 179 matched controls. The Khon Kaen cohort study began at 1990. Up to 2001 there were 24 723 subjects had been enrolled, including 16 652 women. Blood specimens were obtained from 10 954 women and Pap smears; cervical cells from 10 073 women and have been stored. Follow-up of this cohort was carried out by linkage of the study database with the Khon Kaen population-based cancer registry. Two case-control studies nested within this cohort were carried out: the first with prevalence sampling; 32 cases and 128 controls (matched 1:4) and the second with incidence density sampling; 27 cases and 108 controls (matched 1:4). The results from the three studies of risk factors gave results broadly similar to those of studies elsewhere. Consistently, number of pregnancies and age at having first child were associated with the risk of cervical cancer. However, the usual risk factors of age at first sexual intercourse, number of sexual partners and a history of sexually transmitted disease among the subjects were not observed to be associated with risk of cervical cancer in Khon Kaen. The negative findings may be due to the small size of the studies, or to differences in the epidemiology - particularly with respect to infection with human papillomavirus (HPV) - in the Khon Kaen population. The hospital-based case-control study has the advantage of large numbers of cases and controls, and is simple and cheap to conduct. But, the design is open to various biases that may have influenced the results. The cohort study is a classical design in epidemiology, the results are much less influenced by potential bias in measurement of exposure or outcome, and multiple outcomes can be studied. However, this design needs a lot of financial support and is time consuming in countries where these studies cannot rely on existing linkable registries.

The study of survival was based on a total of 630 invasive cervical cancer cases registered during the period 1985­1990. Those registered as death certificate only cases were excluded, leaving 601 cases for the analysis. Patients were followed up until death or date of loss to follow-up or 31 December 1995 (closing date). The registry used both passive and active measures to establish the vital status (alive/dead) of cancer patients. To investigate the potential bias introduced by failure to trace all registered patients, loss-adjusted survival, using prognostic variables for age, stage of disease, treatment and place of residence was estimated by logistic regression methods. The cumulative survival; actuarial survival, loss-adjusted survival and relative loss-adjusted survival rates were calculated. The overall five-year cumulative actual survival of cervical cancer in Khon Kaen was 56.8%, loss-adjusted survival was 54.7% and the relative loss-adjusted survival was 57.4%. The survival of cervical cancer patients in Khon Kaen, Thailand is rather low compared to that recorded in developed countries. Extent of disease at diagnosis is the most important determinant of survival, implying that early detection is more important in improving outcome than improvement in the modality of treatment. The importance of different methods for estimating survival was addressed. Adjusting the estimate for cases lost to follow-up provided results that were similar to those obtained with the actuarial method, suggesting that the conventional use of the actuarial estimate in the calculation of the relative survival is satisfactory for this population.