The relative risk of second primary cancers in Austria’s western states: a retrospective cohort study

Baseline characteristics and distribution of first primary cancers by cancer type

Distribution of gender, site and time of occurrence of second primary cancers

Age at diagnosis of the first primary cancer

SIR of second primary cancer by type of first primary cancer and sex

SIR of all cancers combined except prostate cancer was 1.00 (95% CI 0.96–1.05) for females and significantly decreased 0.90 (95% CI 0.86–0.93) in men. After exclusion of prostate cancer SIR was significantly increased at 1.10 (95% CI 1.05–1.15). The SIR for second primary cancers varied substantially according to the type of first primary cancer. There was a significantly increased SIR for second primary cancers in men after head/neck/larynx cancer (SIR 1.88; 95% CI 1.67–2.11), kidney cancer (SIR 1.22; 95% CI 1.04–1.42) and bladder cancer (SIR 1.20; 95% CI 1.07–1.34), see Table 3. Amongst women there was a significantly increased SIR for second primary cancers after head/neck/larynx cancer (SIR 1.74; 95% CI 1.30–2.28), lung cancer (SIR 1.58; 95% CI 1.19–2.05), cervical cancer (SIR 1.40; 95% CI 1.14–1.70), thyroid cancer (SIR 1.40; 95% CI 1.11–1.75) and kidney cancer (SIR 1.29; 95% CI 1.03–1.59), see Table 3, while women after breast cancer (SIR 0.82; 95% CI 0.75–0.89) and men after prostate cancer (SIR 0.68; 95% CI 0.64–0.72) had a significantly decreased SIR of developing a second primary cancer. Figures showing the SIR of further entities can be found in the supporting material section (see Additional file 1).

Role of prostate cancer

The incidence of prostate cancer more than doubled in Tyrol in 1993 and some 5 years later in Vorarlberg due to the introduction of PSA screening in men aged 45 to 79. Prostate cancer accounts for about 1/3 of all cancer cases in males in our cohort and therefore had a major impact on the estimates for all cancer sites combined.

Therefore we were interested in an estimate for all cancer sites combined, except prostate cancer. Analysing this, the SIR in males for all cancer sites except prostate was slightly increased (10%) and this increase was statistically significant.

When comparing our results with other study results this observation should be kept in mind because the mix of cancer sites varies in some extent between countries [9, 15]. Our data with a reduced SIR for second primary cancers in patients with prostate cancer is in accordance to data published by Coyte et al. [8].

An inclusion of prostate cancer may alter the results due to radiation therapy. In Austria prostate-specific antigen (PSA) screening allows prostate cancer to be detected in a very early stage, achieving a very good prognosis. The underlying aetiology of developing a second primary cancer after prostate cancer may be related to various factors, including treatment modality. More than 50% of the small intestine tumours were carcinoid malignancies, suggesting possible hormonal influences. An excess of pancreatic cancer may be due to pathogenic variants, which predisposes to both [17].

Role of definition

The criteria for defining second primary cancers have evolved over time and sometimes differ among studies. Using rules, registries are able to discriminate between new cases and metastases of an existing malignancy. Definitions are critical when analysing the SIR for second primary cancer. Internationally, both the definition of the International Association of Cancer Registries (IACR) and the International Agency for Research on Cancer (IARC) [1] as well as the rules of the Surveillance Epidemiology and End Results (SEER) Program [2] definitions are widely used and some registries use their own definitions [8, 9].

The IARC rules are more exclusive. Irrespective of time only one tumour is registered for an organ, unless there are histological differences. In contrast, North American cancer registries use the SEER rules that take account of histology, site, laterality and time since initial diagnosis to identify multiple primary cancers [2].

Coyte et al. demonstrated the implication of these differing definitions: for an aggregation of 10 cancer sites in the Scottish study, applying the IARC definition led to a SIR of 0.86 and SEER definition to a SIR of 1.0 [8]. At least for the Scottish data, the absolute difference in SIR is at about 0.15. Our estimates for all cancer sites combined at SIR 1.00 (0.96–1.05) for females and SIR 0.90 (0.86–0.93) for males are in the range reported in different studies, namely from 1.08 to 1.3 [9, 18–20]. Therefore our estimates for all cancer sites combined are in line with published data by previous studies [8–10]. This observation was applicable for site specific results, e.g. increased SIR for primary cancer in the head & neck cancer, kidney, bladder and thyroid and reduced SIR in prostate. For female breast published results are inconsistent [7, 9] but the Scottish data from Coyte et al. are in line with our lowered SIR in women with breast cancer as a first primary [8].

Role of age

Taking into account the variation of age specific incidence during the follow-up period in our method we found a consistent pattern of higher SIR for second primary cancer in younger patients (SIR 1.24 for patients aged below 50) and lower SIR in patients aged 65 and higher. Some of the results for SIR were significantly higher in the lowest age group (15–49 years). According to increased SIRs in head/neck/larynx, cervix and prostate cancer as well as in the all cancers combined group a more dense surveillance may be warranted. This observation is in line with results reported by previously published data [9] and has clinical implications such as more dense surveillance in younger cancer patients [9], but also the fact of their longer life expectancy. The risk compared to the age-matched general population was higher in survivors at younger ages, but within the survivor population, increasing age is still associated with increased cancer risk. In an other study it has been shown that adolescents and young adult cancer survivors, who survive more than 5 years have a higher relative risk of secondary malignant neoplasms compared with younger or older cancer survivors [21]. In addition we would like to notice that we calculated age adjusted SIRs.

Genetic and behavioural risk factors

Possible reasons for an increased SIR for second primary cancers in cancer survivors are genetic and behavioural risk factors [11, 22, 23], treatment of the first primary cancer radiotherapy and chemotherapy, and more intense surveillance of prevalent cancer cases [24]. Lifestyle factors such as smoking (risk factor for head and neck/lung/bladder/kidney) and alcohol consumption are risk factors for a number of cancers. A lack of risk factor data in our cohort limits us to speculation regarding correlations. However, in our analysis the majority of primary cancer sites with increased SIR is nicotine-associated. Of course changing modifiable lifestyle factors like e.g. to quit smoking, will reduce the risk of second primary cancer but also risk of other diseases [25]. However, there is little knowledge on whether cancer survivors in fact are successful to change their habits and we have no data on this. There is some evidence that a cancer diagnosis in adults may have a positive influence on smoking and diet but a negative influence on exercise [26].

Therapy as a risk

Radiation is a risk factor to neighbouring organs of the first primary cancer site. About half of all cancer patients receive radiotherapy at some stage of their disease in developed countries, and at least for some cancer sites like Hodgkin lymphoma, breast cancer, and some gynaecologic malignancies such as vulvar and endometrial cancer it has been shown that radiotherapy causes second primary cancers. These are lung, breast, stomach and thyroid cancer after Hodgkin lymphoma, contralateral breast, lung and oesophagus after breast cancer and leukaemia and any other secondary malignancy after vulvar or endometrial cancer, respectively [27–33].

SIRs and rates of secondary malignancies in high-risk populations have been influenced also by changes in chemotherapy protocols. Chemotherapy-sensitive tissues such as bone marrow, epithelial cells of the gastrointestinal tract and hair follicles are most likely to begin carcinogenesis, therefore the development of leukaemia and lymphoma as secondary hematologic cancers seem to be the greatest long-term risk to cancer survivors after chemotherapy [29–34].

Future effort of research might focus on the complex area of molecular mechanisms of second cancer development. In times of targeted therapies it becomes increasingly important to incorporate factors in our decision making process that might be able to predict the susceptibility of patients to both acute and chronic toxicity, including second primary cancers. This might offer opportunities to individualize therapy, to maximize therapeutic benefit and to minimize serious late toxicity [35].

Surveillance matters

Increased surveillance after a first primary cancer leads to earlier detection of second primary cancers. For example routine use of ultrasound has been shown to dramatically increase thyroid cancer incidence on a population level and hence we would also expect a higher detection rate of thyroid cancer as second primary cancer [36].

Why the risk of second primary cancers is so important

Age-specific mortality rates for chronic diseases are driven by changes in exposure to risk factors and by availability of screening systems and treatment. The risk of cancer after cancer in the overall population might be expected to rise because of persisting effects of genetic and behavioural risk factors, long-term side effects of chemotherapy and radiotherapy, and improved diagnostics [8]. Even if cancer incidence and survival rates remain stable, the number of cancer survivors in the United States will increase by 31%, to about 18.1 million, by 2020 [37]. Because of the aging of the U.S. population, the largest increase in cancer survivors over the next 10 years will be in the age group 65 and older. If new tools for cancer diagnosis, treatment, and follow-up continue to be more expensive, medical expenditures for cancer could reach as high as $207 billion [37]. Policies and programs modifying behavioural and environmental factors to reduce the burden of cancers are key [38].

These data also fit to different other high-income countries and therefore also for Austria. We are also confronted with an aging population. Due to a growing number of cancer survivors this becomes an increasing health concern also in Austria [39], as these patients may impact the overall quality of long-term care in this growing population, like elsewhere [40].

As more and more patients are surviving a cancer, preventing both recurrence and development of second primary cancers is a major goal of national health plans, as they are cost intensive in treatment and care [41]. Taking care of cancer survivors is becoming a challenge for health programs. Cancer survivors could benefit from a coordinated public health effort to support them, as they face numerous physical, psychological, social, spiritual, and financial issues throughout their diagnosis and treatment and the years thereafter. Support depends on the national health care system of the respective country. By preventing secondary diseases or recurrence of cancer and with it improving quality of life for each survivor, many of these issues could be successfully and more focally addressed. Patterns of secondary cancers as shown by our analysis would be helpful for deciding where to focus efforts. One focus of course must be primary prevention as it has already been shown to be the most effective way to fight cancer [42] and the reduction of exposure to key behavioural and environmental risk factors is key to prevent a substantial proportion of deaths form cancer [38].

Strengths and limitations

The strength of our study is the high degree of data completeness of both registries over the full study period and the strict definition of second primary cancer. Furthermore, this is the first study with Austrian data on an array of entities. The limitations are the low population number which causes broader confidence intervals and limits the conclusions to draw especially for site specific results, the lack of registering key information on risk factors and more detailed information on treatment which in consequence does not allow us to analyse the impact of these factors on the risk of second primary cancer.