# Late deaths after childhood cancer, what causes them, and the proof that gentler treatment worked

Source: https://onco.cc/technologies/rejuv-paed-late-mortality/  
OnCo record `rejuv-paed-late-mortality` (Technology). Data CC BY-NC 4.0, attribute "Data from OnCo (onco.cc)"; commercial use needs a licence.

## TL;DR

Five-year survivors still die earlier than their peers, but much less than they did. Fifteen-year mortality among American five-year survivors fell from 12.4 per cent for children treated in the early 1970s to 6.0 per cent for those treated in the 1990s, and the fall tracks the radiotherapy and anthracycline that were taken out of the protocols.

## Summary

This is the single clearest demonstration in oncology that reducing the intensity of a curative treatment can be measured decades later in lives.

The Childhood Cancer Survivor Study followed 34,033 people who survived at least five years after a childhood cancer treated between 1970 and 1999, median follow-up 21 years. Of 3,958 deaths, 1,618 (41 per cent) were attributable to health-related causes: 746 from subsequent neoplasms, 241 cardiac, 137 pulmonary and 494 other. Fifteen-year all-cause mortality fell from 12.4 per cent for diagnoses in the early 1970s to 6.0 per cent in the 1990s, and health-related mortality from 3.5 to 2.1 per cent, with falls in death from subsequent neoplasm, cardiac and pulmonary causes. The exposures fell alongside: cranial radiotherapy for acute lymphoblastic leukaemia was given to 85 per cent of children in the 1970s, 51 per cent in the 1980s and 19 per cent in the 1990s; abdominal radiotherapy for Wilms tumour to 78, 53 and 43 per cent; chest radiotherapy for Hodgkin lymphoma to 87, 79 and 61 per cent; and anthracycline exposure fell too. Reduction in exposure was associated with reduced late mortality among survivors of acute lymphoblastic leukaemia and Wilms tumour.

The British population-based cohort shows the same thing from a different direction and adds the long tail. Among 17,981 five-year survivors diagnosed 1940 to 1991 and followed to 2006, there were 3,049 deaths, a standardised mortality ratio of 10.7 (95 per cent confidence interval 10.3 to 11.1), still threefold higher than expected 45 years from diagnosis. What changes is the cause. The absolute excess from recurrence fell from 97 extra deaths per 10,000 person-years at 5 to 14 years to 8 beyond 45 years; the excess from second primary cancers rose from 8 to 58 and from circulatory causes from 2 to 29 over the same intervals. Beyond 45 years, recurrence accounted for 7 per cent of the excess deaths and second cancers plus circulatory disease for 77 per cent.

The 2016 extension to 34,489 survivors diagnosed 1940 to 2006 found 4,475 deaths, 9.1 times expected (8.9 to 9.4), 64.2 excess deaths per 10,000 person-years (62.1 to 66.3), and that those treated in 1990 to 2006 experienced 30 per cent of the excess deaths of those treated before 1970. Among survivors aged 60 or more, excess deaths from circulatory causes exceeded those from second cancers, which is why cardiovascular risk management belongs in survivorship care and not only cancer screening.

What comes back, and when: nothing here reverses, but the trend is the argument for every de-escalation trial running now. A child treated today carries a materially smaller late risk than the cohorts measure, and the honest way to say that to a parent is to name the exposure that was removed.

## Fields

- Kind: Technology
- Status: established
- Last checked: 2026-10-02
- Tags: rejuvenation; survivorship; paediatric; late-effects; evidence:strong
- Principle: Late mortality in survivors is dominated by two mechanisms with long latencies: radiation carcinogenesis, which produces second cancers at a rate rising with dose and with time since exposure, and radiation and anthracycline injury to the heart and vessels, which becomes clinically apparent when ordinary cardiovascular ageing removes the reserve. Both scale with the dose given, so reducing the dose reduces the hazard decades later.
- Strengths: Two independent cohorts, one hospital-based and one population-based, agree on the trend; The cause of excess death is known by era and by time since diagnosis, so follow-up can be aimed; A direct, measured return on treatment de-escalation
- Limitations: Cohorts beginning in 1940 and 1970 describe treatments that are no longer given; Registration data capture death well and illness poorly; Mortality after newer treatments, including immunotherapy and cell therapy in children, has no long-term cohort yet

## Sources

- Wikipedia: https://en.wikipedia.org/wiki/Cancer_survival_rates
- Reduction in late mortality among 5-year survivors of childhood cancer (CCSS) (NEJM 2016): https://doi.org/10.1056/NEJMoa1510795
- Long-term cause-specific mortality among survivors of childhood cancer (BCCSS) (JAMA 2010): https://doi.org/10.1001/jama.2010.923
- Long term cause specific mortality among 34,489 five year survivors of childhood cancer in Great Britain (BCCSS) (BMJ 2016): https://doi.org/10.1136/bmj.i4351
- The Childhood Cancer Survivor Study: an NCI-supported resource for outcome and intervention research (JCO 2009): https://doi.org/10.1200/JCO.2009.22.3339

## Connected records

- trials: [Childhood Cancer Survivor Study (CCSS)](https://onco.cc/trials/ccss/)
- collections: [British Childhood Cancer Survivor Study (BCCSS)](https://onco.cc/collections/bccss/)
- roadmaps: [Paediatric oncology roadmap: cooperative-group cures → engineered immunity → drugs developed for children first](https://onco.cc/roadmaps/paediatric-oncology-roadmap/), [Recovery and rejuvenation roadmap: cure is not enough → survivorship gets a name → the cohorts that measured the cost → exercise proven as treatment → biological ageing measured and sold → repair, if anyone funds it](https://onco.cc/roadmaps/rejuvenation-roadmap/), [Supportive care and survivorship roadmap: making treatment bearable → proving it extends life → caring for tens of millions afterwards](https://onco.cc/roadmaps/survivorship-roadmap/)
- cancers: [Acute lymphoblastic leukaemia](https://onco.cc/cancers/all-leukemia/), [Childhood cancers (all types)](https://onco.cc/cancers/childhood-cancers/), [Hodgkin lymphoma](https://onco.cc/cancers/hodgkin-lymphoma/), [Medulloblastoma](https://onco.cc/cancers/medulloblastoma/), [Wilms tumour (nephroblastoma)](https://onco.cc/cancers/wilms-tumor/)
- fronts: [Recovery & Rejuvenation](https://onco.cc/fronts/rejuvenation/), [Supportive Care & Survivorship](https://onco.cc/fronts/supportive-care/)
- technologies: [How much illness childhood cancer survivors carry, and at what age](https://onco.cc/technologies/rejuv-paed-chronic-disease-burden/), [Proton therapy](https://onco.cc/technologies/proton-therapy/), [Second cancers after childhood cancer: the risk by treatment, and why it is falling](https://onco.cc/technologies/rejuv-paed-second-cancers/), [Survivorship care and late-effects surveillance](https://onco.cc/technologies/survivorship-care-plan/), [The heart after anthracyclines and chest radiotherapy in childhood](https://onco.cc/technologies/rejuv-paed-heart/)
- terms: [Cure is not enough: when late effects stopped being an afterthought](https://onco.cc/terms/rejuv-history-cure-is-not-enough/), [Late effects and survivorship toxicity](https://onco.cc/terms/late-effects/), [Radiotherapy](https://onco.cc/terms/radiotherapy/)
- drugs: [Cyclophosphamide](https://onco.cc/drugs/cyclophosphamide/), [Doxorubicin](https://onco.cc/drugs/doxorubicin/)
- bottlenecks: [Registries count diagnoses and deaths, and not what treatment left behind](https://onco.cc/bottlenecks/rejuv-agenda-late-effects-are-not-counted/), [Survivorship and late effects are neglected](https://onco.cc/bottlenecks/b-survivorship/)

---
JSON: https://onco.cc/api/v1/entities/rejuv-paed-late-mortality.json