The current study confirms that patients with a history of PC carry an overall excess risk of over 80% for developing an SPC compared to the general Finnish population. Our findings further indicate that this elevated SPC risk persists even beyond 20 years following the diagnosis of primary PC (SIR 1.92, 95% CI: 1.42–2.54 after 20 years). In addition to an increased cancer risk in the mouth/pharynx and respiratory organs observed among OPC (SIR 4.41 and 3.51, respectively) and HPC patients (SIR 6.91 and 4.80), an elevated cancer risk in digestive organs was noted among OPC patients (SIR 1.83). Apart from the aerodigestive organs and skin, the risk of SPC was not elevated for other anatomical sites among these patients. Male NPC patients exhibited an increased risk for oral/pharyngeal, brain, haematolymphoid, and skin cancers (SIRs 5.14, 6.60, 3.05, and 2.91, respectively).

The risk of SPC in the mouth/pharynx, respiratory, and digestive organs among OPC and HPC patients aligns with results from previous studies and may be explained by the field cancerization effect induced by the chronic use of tobacco and alcohol drinking [11, 12]. Supporting our findings, Morris et al. [13] conducted a study in the US utilizing data from the Surveillance, Epidemiology, and End Results (SEER) database. They analysed 75 087 cases of HNC and reported SIRs of 2.99 (95% CI: 2.88–31.10) and 3.47 (3.27–3.68) for OPSCC and HPSCC, respectively, for any SPC. The study, which had a median follow-up time of 5.8 years, identified elevated risks not only in the mouth/pharynx, respiratory, and digestive organs, but also observed an excess of lymphoma and thyroid cancer among OPSCC and HPSCC patients, respectively. In South Korea, Jung et al. [14]. reviewed data of 5587 OPSCC and 3685 HPSCC patients with mean follow-up periods of 4.2 and 2.6 years, respectively. The authors conveyed SIRs of 1.46 (95% CI: 1.30–1.63) and 1.93 (1.71–2.17) for any SPC among OPSCC and HPSCC patients, respectively, with elevated SPC risks in the mouth/pharynx, respiratory organs, and oesophagus.

In our study the risk of SPC among OPC patients remained elevated in each time period after 1985 with no major differences observed. The prevalence of HPV in OPC has significantly increased over time worldwide. For instance, in Europe, HPV-positive OPC rose from 39.7% (95% CI: 32.8-47.0) before 2000 to 59.0% (95% CI: 30.2-82.7) between 2000 and 2004 [15]. The average latency period from infection to HPV-positive OPC is estimated at around 10 to 30 years. Thus, the SIRs could change with longer follow-up [16].

Women exhibited higher SIRs for most tobacco and alcohol associated SPC sites, especially among OPC patients for SPCs occurring in the respiratory organs aligning with previous studies [17,18,19,20]. Smoking and alcohol use influence the development of SPCs in the upper aerodigestive tract [21, 22], and the observed higher risk for a SPC in women, as described also in previous studies, may be attributable to the intrinsic nature of the SIR analysis, where the observed incidence is compared with the expected number derived from age, sex, and calendar-specific rates in the general population. Given the higher prevalence of smoking and alcohol and the incidence of related diseases in men compared to women [23, 24], the SIRs are consequently higher in women due to the lower rates in the reference population.

In our study 9.7% of NPC patients were diagnosed with an SPC within the first 5 years of follow-up and 12% during the first 10 years. A recent systematic review [10], encompassing 21 articles – among others two from Europe, three from the USA, and 11 from China, Hong Kong, Singapore, and Taiwan – reported a 6.6% occurrence rate (range 1.5–20.2) of SPC with an average follow-up time of 7.7 years. The overall SIR of SPC for all countries and cancer sites combined was 2.0, consistent with our risk estimate. NPC is typically treated with high-dose radiotherapy combined with chemotherapy, commonly cisplatin [25]. The elevated risk of brain cancer observed among NPC patients could be subsequent sequelae from radiotherapy, as the brain would be exposed to the radiation [26]. However, the advent of intensity-modulated radiotherapy (IMRT) as the standard technique for NPC treatment has not only enhanced disease control but also minimized unintended radiation to the brain [27]. IMRT was introduced in the late 1990s and became widely adopted in Finland in the early 2000s. Within our patient cohort, we observed five cases of brain SPCs, with four diagnosed before the widespread use of IMRT, lending support to the hypothesis that radiation played a role in their development. Still, there is also a chance of mistakenly classifying an extensive nasopharyngeal NPC as brain cancer.

Our study reveals an increased incidence of skin cancer after five years among patients with OPC and NPC, mirroring findings in Taiwan where a heightened risk of skin cancer was noted among NPC patients [19]. It could be argued that PCs are not inherently linked to UV radiation exposure, and plausible that shared lifestyle factors, such as outdoor activities, among these patients may contribute to the elevated risk of skin cancer. However, in Nordic countries, where outdoor occupations have not been associated with an increased risk of OPC or NPC cancers, this explanation may not entirely apply [28, 29]. It would make more sense that the elevated skin cancer risk would also be attributable to radiation exposure, which can increase the risk of skin cancer in the irradiated area [30].

Concerning the risk of haematolymphoid cancers among NPC patients, case reports delineate instances of these cancers diagnosed after chemoradiotherapy treatment [31, 32]. Cisplatin has been associated with an elevated risk of secondary leukaemia following treatment of ovarian and testicular cancers [33, 34]. Occupational exposure to specific agents may also contribute to this risk. Indeed, evidence, albeit weak, suggest a link between formaldehyde exposure and NPC, leukaemia, and non-Hodgkin’s lymphoma [35]. Additionally, common genetic susceptibility may play a role in these associations. For instance, polymorphisms in the DNA repair gene XRCC1 have been associated with an elevated risk of both NPC and leukaemia in specific populations [36, 37]. Lastly, the heightened risk of haematolymphoid cancers described may simply reflect enhanced surveillance and detection of these malignancies, exemplified in our study by the elevated risk of haematolymphoid cancers exclusively observed among male patients aged 75 years or older during the initial five years of follow-up.

Studies investigating the risk of SPC following primary PC in Europe have typically relied on small case series, with very few exceptions [2, 3, 10]. To the best of our knowledge, the present cohort represents one of the largest published in Europe thus far with the most extensive follow-up. The underreporting of cancers in the FCR is negligible, as quality assessment studies have demonstrated high coverage (95% for head and neck tumours in 2009–2013) and diagnostic accuracy [38]. One of the major strengths of this study lies in its low risk of misclassifying a recurrence as an SPC, which is attributable to the FCR’s practice of not documenting subsequent cancers that occur at the same site as the index site. Thought, this could also lower the overall SPC risk. Moreover, population-based studies help mitigate selection bias inherent in hospital or clinical series [39].

Within the context of cancer register-based studies, it is crucial to consider the inherent limitations of our results. The absence of data on aetiological factors like tobacco, alcohol consumption, and HPV status, as well as details about primary treatment, especially radio- and chemotherapy regimens, hinders our ability to draw conclusions about their relationship to SPCs. Moreover, assessing the specific impact of HPV vaccination on SPC risk – particularly in patients with OPSCC – will require dedicated evaluation in the future [40]. Additionally, our study is susceptible to misclassification of the exact site, an unfortunate reality in clinical practice. The proximity of some SPC sites to the primary tumour could have led to misdiagnoses of local spread or disease recurrence as SPCs and primary tumour metastases might have been inaccurately classified as SPCs and vice versa. However, this potential diagnostic bias cannot solely explain the association between PC and increased SPC risk, as the risk remained elevated even beyond 20 years post-diagnosis of the primary tumour.