CANCER RISK (STUDY IV)

As with LS, details (inpatient/outpatient diagnosis and primary/secondary diagnosis) were missing in 278 women (2%) with LP, and therefore the cohort for the cancer assessment included 13 100 women and 140 179 person-years. Of the women, 11 011 (84%) were diagnosed in outpatient setting, and LP was a primary diagnosis code in 12 311 (94%) of the women, and the details of the diagnosis did not affect the cancer risks.

A total of 1 520 cancers were diagnosed in the whole cohort - more than the expected number of 1 326.

The overall cancer risk was consequently increased (SIR 1.15, 95% CI 1.09-1.20) (Table 16).

Table 16. Observed (Obs) and expected (Exp) numbers of cancer cases, standardized incidence ratios (SIR) and 95% confidence intervals (CI) among women with diagnosis of lichen planus by site during 1969-2014.

Risk of head and neck cancers (study IV)

The SIRs for different head and neck cancers were elevated: 5.17 for lip cancer (95% CI 3.06-8.16), 12.4 for tongue cancer (95% CI 9.45-16.0), 7.97 for cancer of the oral cavity (95% CI 6.79-9.24), and 3.47 for laryngeal cancer (95% CI 1.13-8.10) (Table 16). The risks of these cancers remained elevated regardless of the time interval between the diagnoses of LP and cancer.

Risk of other cancers (study IV)

The risk was increased for both esophageal cancer (SIR 1.95, 95% CI 1.17-3.04) and vulvar cancer (SIR 1.99, 95% CI 1.18-3.13 (Table 16). The risk for non-melanoma skin cancer was elevated, but not

MORTALITY (STUDY III)

Because the division inpatient/outpatient diagnosis or primary/secondary diagnosis did not affect the results in the study of the cancer risk of LP women, the whole cohort with 13 378 women with 143 040 person-years of follow-up was included in the mortality study. A healthy patient effect was detected within the cohort (Appendix 3); thus, the results from the follow-up of less than five years are ignored.

Most deaths within the cohort were due to circulatory diseases, malignancies, and dementing disorders (Table 17). The all-cause mortality was increased compared to the population: 1 813 observed vs. 1 701 expected deaths yield an SMR of 1.07, 95% CI 1.02-1.11. This is the result of an elevation in mortality from diseases of different organ systems: infections (SMR 1.78, 95% CI 1.14–2.64), bronchitis and emphysema (SMR 1.57, 95% CI 1.09–2.17), digestive diseases (SMR 1.39, 95% CI 1.09–1.75), respiratory diseases (SMR 1.31, 95% CI 1.07–1.57), and cancer (SMR 1.14, 95% CI 1.03–1.25).

The risk of death was increased from the following malignancies: cancers of the mouth and pharynx (SMR 6.50, 95% CI 4.32–9.39), cancer of tongue (SMR 7.25, 95% CI 3.13–14.3), Hodgkin lymphoma (SMR 6.73, 95% CI 1.83–17.2), and non-Hodgkin lymphoma (SMR 1.68, 95% CI 1.11–2.44) (Table 17).

Table 17. Observed (Obs) and expected (Exp) numbers of deaths, standardized mortality ratios (SMR) and 95% confidence intervals (CI) for selected causes of death of women with lichen planus within a follow-up of more than five years.

Cause of death Obs Exp SMR 95% CI

Endocrine, nutritional and metabolic diseases 27 22 1.25 0.82–1.81

Diabetes mellitus 22 18 1.22 0.76–1.83

Dementia, Alzheimer’s disease 223 232 0.96 0.84–1.09

Diseases of circulatory system 766 761 1.01 0.94–1.07

Ischemic heart diseases 429 403 1.07 0.97–1.16

Other heart diseases 76 73 1.04 0.82–1.29

Cerebrovascular diseases 185 198 0.94 0.81–1.07

Diseases of respiratory system 105 80 1.31 1.07–1.57

Bronchitis, emphysema 35 22 1.57 1.09–2.17

Diseases of digestive system 73 52 1.39 1.09–1.75

Diseases of genitourinary system 28 19 1.44 0.96–2.08

Alcohol related diseases and accidental poisoning by alcohol

11 19 0.59 0.30–1.06

Accidents and violence 68 62 1.09 0.85–1.38

Suicide 9 9 1.00 0.46–1.90

DISCUSSION

This thesis presents, for the first time, nationwide data not only on the incidence of lichen sclerosus (LS) and lichen planus (LP), but also on that of cancer risk and mortality of women with these lichens. For this thesis, all women with a diagnosis of LS or LP made in Finnish specialized health care have been followed through registries from 1969 to 2014.

According to our data, 1.6% of Finnish women will develop LS by the age of 80. The association of LS with vulvar squamous cell carcinoma (SCC) was confirmed, and an elevated risk of other vulvar malignancies was also suggested. Another novel association with vaginal cancer and a reduced risk of cervical cancer were observed. The mortality of LS women was decreased compared to that of the Finnish general female population. The reason for the decrease remains, for the present, unknown.

LP was more common – the cumulative incidence was 2.3% by age 80. An association with cancers in the oral region was confirmed, and new associations with cancers of the esophagus, larynx, and vulva were seen. The overall mortality of LP women was increased, with excess mortality from diseases of different organ systems, cancer of the mouth and pharynx, and lymphomas. These results lead to speculations of both local and systemic involvement of inflammation in LP.

The similarities and differences of LS and LP are thus accentuated. Both are fairly common diseases of elderly women, but the peak in incidence is 10 years earlier in LP compared to LS. Both increase the cancer risks of affected tissues, and LP patients show increased mortality from various causes. Therefore, the effects of LP seem more extensive and complex.

INCIDENCE OF LICHEN

Some research in the literature reports LS and LP as rare diseases, while others describe them as frequent.

This is probably a result of incidence and prevalence studies made in highly selected populations, most often in patients of single, specialized clinics, which causes difficulties in generalizing the results on a larger scale.

Another problem in assessing the frequency of LS or LP is underdiagnosis. Patients probably underreport and physicians underrecognize them when either lichen affects the vulvar area or LP affects the more rare sites (106,208-210). Additionally, the problem with studying LP – whether its epidemiology or any other aspects - is its diversity of clinical presentation resulting in care fragmenting to many different specialties with little cooperation.

We report the first nationwide incidence rates of LS and LP in this thesis – 19/100 000 woman-years for LS and 28/100 000 for LP. A recent Finnish study with similar methodology assessed the incidence of psoriasis and hidradenitis suppurativa (HS) in the population: 28/100 000 and 3/100 000 (211). Thus, LP seems to be as common and LS rarer than psoriasis, but both are more frequent than HS. LP is diagnosed in somewhat younger women than LS: the incidence peaks in 65- to 69-year-olds in LP and 75- to 79-year-olds in LS.

The only comparable incidence study on LS was conducted in two provinces in the Netherlands: All histological diagnoses were gathered from a national Pathology Registry, and an incidence of 16/100 000 woman-years was calculated 2006-2011 (10). This is an underestimate to some extent, at least, due to inclusion of histological diagnoses only, which does not limit our study. That said, our data miss the diagnoses made in the primary health care and private sector. According to a German health insurance company, general practitioners (GP) made 12% of all LS diagnoses (208), but such information and the number of patients diagnosed in the private sector in Finland is unknown.

The incidence of LS rose from 14 to 22/100 000 in 2003-2012. The aforementioned Dutch study also found an increasing incidence from 1991-1995 to 2006-2011: from 9 to 16/100 000. This may reflect a true rise in incidence due to some change in the environment. Another explanation could be improving diagnostics - patients seeking care more eagerly and physicians being more aware of vulvar diseases.

The fact that LS is most often diagnosed in postmenopausal women is well known and also obvious in our data. The mean age at diagnosis in this study was 60 years, somewhat higher than the reported age at presentation in a British study – 48 years (8). The median age at diagnosis was 60, similar to our median, 63 years, in the incidence study from the Netherlands (10). A highly variable diagnostic delay is characteristically presented in studies: from 0 to 51 years with a mean of 5 years (8,48). The length of the diagnostic delay in Finnish women with LS is not known.

“A second peak in incidence” is often cited to occur in prepubertal girls (87,88), and it is evident from a clinical perspective but is not well documented in research. The aforementioned Dutch study reported that the incidence of LS was low in girls (10). All guidelines discourage biopsy taking in girls (87,88), and the low incidence in the Netherlands reflects the restriction to include histological diagnoses only. Our study, however, shows a slight elevation in incidence in girls between the ages of 5 and 9 (7/100 000).

Low estrogen levels characterize both prepuberty and postmenopause. The incidence peaks of LS during these periods have led researchers to look into possible hormonal contributions to the etiology of LS.

Research has found differences in the sex hormone levels and their receptors in the serum and on skin in LS patients (79,212). Still, therapies with topical estrogen, testosterone, or progesterone are found inefficient (75,213); thus, the findings’ significance is uncertain. It is also possible, that the LS incidence peaks in the hypoestrogenic states of life because LS may be more symptomatic in a hypoestrogenic vulva’s skin.

The LS of fertile-aged women has received little attention in the literature. The incidence in this age group is indeed low in our data but rises with age. The incidence equals the much-cited “peak in incidence in girls” in 30- to 35-year-old women and exceeds that in older age groups. Therefore, physicians should keep a high index of suspicion of LS even in women throughout their reproductive years.

Prevalence studies, mainly of the oral subtype, dominate the literature regarding LP. This thesis calculated the incidence rate of any LP in women diagnosed in secondary or tertiary health care and found it to be 28/100 000 woman-years. A British study evaluated the incidence of any LP in patients treated by sentinel general practitioners: 38/100 000 woman-years (30). This figure, however, is not comparable to ours due to the differing study populations. No studies exist on the incidence of cutaneous LP or on the rare subtypes of mucosal LP. The incidence of oral LP has been evaluated in three studies and shows marked

(34). An incidence of 188/100 000 was found in a screening program in a Japanese city (33). A total of 4 300 women were screened, and 19 cases of oral LP were found, most of them of the reticular subtype.

The largest incidence, 250/100 000, was found among 5 000 Indian women in the 1960s and 1970s (32).

Special care must be undertaken when interpreting this result, since the smoking habits – especially the large number of people using chewing tobacco - are rather unique to that study population. The incidence of oral LP in this study was 90/100 000 among women not smoking or chewing tobacco.

Studies report either the age of onset of symptoms of LP or the age at presentation. The age of onset of cutaneous LP is during the 5th decade, and that of oral and vulvar LP during the 6th decade of life (26,27,35,52,95,104) , whereas the age at presentation for mucosal LP is during the 5th or the 6th decade (25,26,35,36,52,101,104) . The mean age at diagnosis in our data ― 57 years ― is comparable to these.

The diagnostic delay in oral or cutaneous LP has not further been elaborated in the aforementioned studies, but it is 3-5 years in the vulvar, esophageal, and otic subtypes (21,35,36,52,106). The length of the diagnostic delay in Finnish women with LP of different subtypes is unknown.

The incidence of LP in children was found to be low. Literature concerning LP in children is scant;

historically, it has usually been considered a rare disease before adulthood. According to newer evidence, however, it may be more common in children, especially in certain ethnic groups, i.e., in children from India and Africa (29,214).