This study has shown that the age-standardized incidence of CRC is falling, while the incidence of EOCRC in Aotearoa New Zealand is continuing to rise. This increase in EOCRC incidence is predominantly associated with distal colonic and rectal cancers, with smaller increases seen in proximal colonic cancers. In patients aged 50–79, however, there are ongoing decreases in CRC incidence across all sites. Concerningly, in contrast to the total population, crude incidence of CRC in Māori is rising, driven by increases in incidence of EOCRC in Māori patients.

Comparing these results to data published by Gandhi et al [2] that described incidence rates in Aotearoa New Zealand from 1995 to 2012, rates of increase of EOCRC appear to be rising. The rates of increase of rectal cancer incidence in women is accelerating from 13% per decade to 25% per decade, while rates of increase in distal colonic cancer in men has risen from 14% per decade to 35% per decade, and likewise rates of increase of rectal cancer incidence in men has risen from 18% per decade to 32%. This study has also shown statistically significant increases in other sites that previously failed to reach significance, with rates of distal colonic cancers in women rising by 29% per decade (p = < 0.0005), and rates of proximal colonic cancers in women rising by 14% (p = < 0.05) and 20% in men (p = < 0.005). Our modelling predicts that if these trends continue to 2030, there will be 361 cases of EOCRC diagnosed in Aotearoa New Zealand, an increase of 92 cases per year when compared with numbers seen in 2020. These numbers increase to a potential 524 total cases per year by 2040, an additional 255 cases of EOCRC per year.

Our data shows that early onset colorectal cancers are predominately sigmoid and rectal cancers as has been described in other studies [2, 16]. This is, however, the first data to also show significant increases in proximal colonic cancers here in Aotearoa New Zealand. This is consistent with trends seen in Europe over the past 25 years [17]. The evolving pattern of disease will have significant implications with regards to potential screening methods, with proximal cancers being missed by flexible sigmoidoscopy.

While a recent metanalysis found that there was no material difference in cancer specific survival in EOCRC compared to older patients [7], strategies are needed to help improve outcomes in EOCRC. One issue affecting EOCRC internationally is delay to diagnosis [5, 18], with delays contributing to more advanced stage at diagnosis, and resulting in poorer outcomes [19]. These delays may be driven by patients not seeking medical advice for symptoms, or doctors not appreciating the epidemiological shift in the pattern of this disease and, therefore, not investigating young patients appropriately. One study found that the largest contributor to delays in diagnosis in rectal cancer in patients under 50 years was patients not seeking medical advice, with a median time from symptom onset to seeing a doctor of 121 days, compared to only 21 days in patients over 50 years [18]. Improving patient education about when to seek medical attention may help improve these delays. Bowel Cancer New Zealand’s recent campaign, ‘Never too young’, is an example of an organisation helping to improve patient education about what signs to look out for [20]. Likewise, improving physician education to ensure timely workup of symptomatic patients once they have presented is also crucial. There is no specific data on whether there is different survival in EOCRC in Aotearoa New Zealand or on delays to diagnosis in those aged under 50 years in Aotearoa New Zealand, and these are both areas where further research would be beneficial.

Lowering the age of screening will also help, through both prevention of some cancers and earlier diagnosis of others. Aotearoa New Zealand has only recently started screening patients over 60 years of age, and the exact impact of this in the screened population has yet to be established. In contrast to many other countries around the world that have been screening people over the age of 50 years, including Canada, UK, and Germany, with the USA and Australia recently screening from age 45 years [21,22,23,24]. Several American guidelines, including the American Cancer Society (ACS) [25], National Comprehensive Cancer Network (NCCN) [26], American College of Gastroenterology (ACG) [27] and the US preventative task force now recommend starting screening of average risk individuals at 45 years of age [28, 29]. While there is no direct evidence that screening below the age of 50 years will reduce EOCRC, countries such as Italy, Austria and Japan have been screening patients in their 40’s since the 1980s [24], and they are some of the few countries internationally where EOCRC rates have been declining [3, 17, 30, 31].

We acknowledge that there will be issues with resourcing, and health systems need to balance providing access to colonoscopy for screening programmes, without increasing wait times for at-risk or symptomatic populations. This, however, should not limit goals for standards of care from being set, and increasing endoscopy capacity should be a priority for health systems to reach these goals. In the long term, it will save the health system money due to reduced treatment costs as a result of earlier diagnosis of CRC. Modelling studies from both USA and Canada have shown screening from 45 years, or as low as 40 years to be cost effective [32, 33]. Because of this, we believe that screening for average-risk individuals in Aotearoa New Zealand should start at age 45. It is important to note that while this will only reduce the impact in those eligible for screening, this makes up a large proportion of all EOCRC, with 44% of all EOCRC diagnosed during our study period occurring in those aged 45–49 years (see Supplementary Table 4). Looking forward, as methods of diagnosis improve, such as through the development of biomarkers, screening could potentially selectively target younger higher-risk patients [34]. Timely diagnosis of younger symptomatic individuals is also critical. Given that the largest increases in EOCRC incidence were seen in distal colonic and rectal cancers, clinicians may consider flexible sigmoidoscopy as a tool to investigate per rectal bleeding in young patients.

This study has observed increases in the incidence of CRC in Māori, with this change being driven by increases in EOCRC. Additionally, Māori patients aged 50–69 years are not seeing the same decreases in incidence rates that are occurring in the total population. Ongoing sequelae of colonisation includes reduced life expectancy in Māori compared to non-Māori; [35] hence, to keep proportions comparative, the age threshold for older Māori was reduced to 70 years [36]. While incidence rates of CRC in Māori patients remain lower than in non-Māori, due to the different age distribution of the Māori population, with a far higher proportion of younger people, Māori patients are disproportionately affected by early-onset disease, with 30% of diagnoses in Māori women and 25% in Māori men occurring prior to the age of 50 years [37]. Also, Māori, once diagnosed with CRC are more likely to die from their disease than non-Māori [38]. Causes for this are likely multifactorial and there is evidence of inequitable access to early diagnosis, chemotherapy, and treatment in CRC [39]. Māori are also more likely to be diagnosed with stage 4 disease (stage 4 colon cancer 31.6% vs. 22.8%, stage 4 rectal cancer 29.4% vs. 18.1%) [40], which may be due to greater delays to diagnosis. These factors make the rate of increase in Māori patients under 50 extremely concerning and if these increases go unchecked, we will see rates in Māori overtake those of the general population. To help combat this inequity, lowering of the screening age in Aotearoa New Zealand should be extended further for Māori. The fact that CRC incidence in Māori aged 50–69 years is not falling, may also be a result of inequitable access to health interventions, such as bowel cancer screening, or increasing use of diagnostic colonoscopy for symptomatic patients, which has previously been credited with helping drive the decreases in incidence we are seeing in older adults [2].

The factors driving the increase in EOCRC incidence are unknown but are likely to be multifactorial. LOCRC appears to be driven by the interaction of an individual’s microbiome and diet, with conventional risk factors, such as obesity, alcohol, processed meat, sugary drinks, and a ‘Western diet’ (high fat, high meat, and low fibre), and such mechanisms are likely with EOCRC [41,42,43,44,45,46,47], However the mechanism must be different from LOCRC to be driving different changes in incidence. While several bacterial species have already been implicated in adenoma or LOCRC development [48,49,50], data specific to EOCRC suggest that the microbiome in patients with EOCRC is different compared that found in patients with LOCRC and healthy controls [51]. These differences may reflect early-life events and/or ongoing environmental factors, many of which emerged over the past several decades. These include caesarean delivery [52], formula feeding [53], antibiotic use [54], changing diet, synthetic food dyes, MSG high-fructose corn syrup, or microplastics [55].

Managing the impact that cancer diagnosis and treatment has on younger patients also needs to be considered. Recent research has shown that there are large deficits in cancer survivorship care in Aotearoa New Zealand, this having profound impacts on the wellbeing of cancer patients after treatment [56]. This is particularly important in EOCRC patients, with research showing psychosocial impacts being more pronounced in EOCRC compared with LOCRC [57]. This is another area where we believe more focus and funding is critical to improve the quality of life of the increasing numbers of EOCRC patients.

It is however worth emphasising that despite the rising incidence of EOCRC, CRC remains largely a disease of older adults. From 2015 to 2020, an average of 92.5% of CRC was diagnosed in those over the age of 50 years and the absolute incidence rates of EOCRC are a small fraction of those seen in older patients (Table 1). It is also worth considering that for patients over the age of 80 years, the increasing incidence rates may be a result of increased detection, rather than a true increasing prevalence of disease, e.g., increasing incidental findings of CRC picked up on cross sectional imaging for another indication [58]. Whatever the cause, due to the rapidly aging population the volume of CRC diagnoses in patients over 80 years is set to increase dramatically. Our modelling has shown that the number of CRC cases diagnosed over the age of 80 will increase from 801 in 2020 to 1266 as close as 2030, and to a potential 1830 cases per year in 2040. This will mean that CRCs diagnosed over the age of 80 years may account for 34.5% of all CRC cases in 2030, and 43% in 2040. How to manage this will be another huge challenge ahead for our already struggling health system. Discussions around whether this may necessitate rationing of health-care resources for elderly people are already underway, and is a contentious issue that is likely to become more urgent [59,60,61].

A notable strength of this study is that it is the most up-to-date, population-based data describing overall trends in CRC incidence in Aotearoa New Zealand by age, sex, and site of disease. From 1995 it has been compulsory to register all new cancer diagnoses in Aotearoa New Zealand with the New Zealand Cancer Registry, so we can be confident that this study captures the vast majority of new diagnoses of colorectal adenocarcinoma in Aotearoa New Zealand. There are, however, some limitations, namely that cancer registrations can be missed or miscategorised, either with an incorrect type or site of disease; these are problems inherent with using any registry data. The impact of bowel cancer screening has also not been specifically addressed in this paper. It was trialled in one area of Aotearoa New Zealand starting from 2011, and then rolled out nationally from 2019 to 2021. Therefore, it will currently be having an impact on CRC incidence rates in the middle age bracket. There are well-documented problems with undercounting of Māori ethnicity [62] and there was also a change to Māori ethnicity classification in 2006, which may influence calculated incidence in this analysis, and result in underestimation of incidence rates. This research has been approved and guided by input from Māori researchers, and following the steps outlined in the CONSIDER statement [14], however none of the researchers are themselves Māori and we acknowledge this as a limitation. While every care has been taken with the methodology, we urge caution when interpreting these projections as they are sensitive to changes in population structure, due to extrinsic factors, such as, immigration and economic circumstances. For our predictions, we have assumed an ongoing linear trend of all incidence rates, but for EOCRC in recent years, the rate of increase has been increasing, so our predictions in this group may be an underestimation. This also does not take into account any impact of screening, which is likely to reduce the incidence in CRC in older populations. There is currently no data on EOCRC survival in Aotearoa New Zealand and we acknowledge that this as an area where further research is required.