Long-term Somatic Disease Risk in Adult Danish Cancer Survivors | Oncology | JAMA Oncology | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Adjusted Hazard Ratios (HRs) and 95% CIs for Hospitalization for Somatic Diseases Among Danish Survivors of Breast, Lung, Prostate, and Colon Cancer and Cancer-Free Matched Comparison People
Adjusted Hazard Ratios (HRs) and 95% CIs for Hospitalization for Somatic Diseases Among Danish Survivors of Breast, Lung, Prostate, and Colon Cancer and Cancer-Free Matched Comparison People

Diseases included the 11 main International Statistical Classification of Diseases and Related Health Problems, Tenth Revision diagnostic groups.

Figure 2.  Adjusted Cumulative Incidence Risk Difference (RD) and 95% CIs for Hospitalization 5 Years After Diagnosis for Somatic Disease Among Danish Survivors of Breast, Lung, Prostate, and Colon Cancer and Cancer-Free Matched Comparison People
Adjusted Cumulative Incidence Risk Difference (RD) and 95% CIs for Hospitalization 5 Years After Diagnosis for Somatic Disease Among Danish Survivors of Breast, Lung, Prostate, and Colon Cancer and Cancer-Free Matched Comparison People

Diseases included the 11 main International Statistical Classification of Diseases and Related Health Problems, Tenth Revision diagnostic groups.

Figure 3.  Adjusted Hazard Ratios (HRs) and 95% CIs for Hospitalization for New Primary Cancers Among Danish Survivors of Breast, Lung, Prostate, and Colon Cancer and Matched Cancer-Free Comparison People
Adjusted Hazard Ratios (HRs) and 95% CIs for Hospitalization for New Primary Cancers Among Danish Survivors of Breast, Lung, Prostate, and Colon Cancer and Matched Cancer-Free Comparison People

Solid lines represent HRs, and the 2 dashed lines represent 95% CIs. P values are for test of hypothesis of constant HR over time.

Figure 4.  Adjusted Hazard Ratios (HRs) and 95% CIs for Hospitalization for Infectious and Parasitic Diseases Among Danish Survivors of Breast, Lung, Prostate, and Colon Cancer and Matched Cancer-Free Comparison People
Adjusted Hazard Ratios (HRs) and 95% CIs for Hospitalization for Infectious and Parasitic Diseases Among Danish Survivors of Breast, Lung, Prostate, and Colon Cancer and Matched Cancer-Free Comparison People

Solid lines represent HRs, and the 2 dashed lines represent 95% CIs. P values are for test of hypothesis of constant HR over time.

Supplement.

eTable 1. Descriptive statistics of the study participants

eTable 2. Results of Cox models for the effect of the 12 most frequent cancer types (exposed) as compared with cancer-free comparisons (unexposed) on the incidence of somatic disease

eFigure 1. Adjusted HRs and 95% CIs for hospitalization for somatic diseases in the 11 main ICD-10 diagnostic groups in Danish survivors of malignant melanoma, bladder, rectum and brain cancer and cancer-free matched comparison people

eFigure 2. Adjusted HRs and 95% CIs for hospitalization for somatic diseases in the 11 main ICD-10 diagnostic groups in Danish survivors of non-Hodgkin lymphoma, pancreas and kidney cancer and leukemia and cancer-free matched comparison people.

eFigure 3. Adjusted cumulative incidence risk difference and 95% CIs for hospitalization five years after diagnosis for somatic disease in the 11 main ICD-10 diagnostic groups among Danish survivors of malignant melanoma, bladder, rectum and brain cancer and cancer-free matched comparison people.

eFigure 4. Adjusted cumulative incidence risk difference and 95% CIs for hospitalization five years after diagnosis for somatic disease in the 11 main ICD-10 diagnostic groups among Danish survivors of non-Hodgkin lymphoma, pancreas and kidney cancer and leukemia and cancer-free matched comparison people.

eFigure 5. Adjusted HRs and 95% CIs for hospitalization for new primary cancers among Danish survivors of malignant melanoma, bladder, rectum, and brain cancer, non-Hodgkin lymphoma, pancreas and kidney cancer and leukemia and matched cancer-free comparison people.

eFigure 6. Adjusted HRs and 95% CIs for hospitalization for infectious and parasitic diseases among Danish survivors of malignant melanoma, bladder, rectum, and brain cancer, non-Hodgkin lymphoma, pancreas and kidney cancer and leukemia and matched cancer-free comparison people.

eFigure 7. Adjusted HRs and 95% CIs for hospitalization of survivors of the 12 most frequent cancer types for diseases of the blood and blood-forming organs as compared with matched cancer-free comparison people.

eFigure 8. Adjusted HRs and 95% CIs for hospitalization of survivors of the 12 most frequent cancer types for endocrine, nutritional and metabolic diseases as compared with matched cancer-free comparison people.

eFigure 9. Adjusted HRs and 95% CIs for hospitalization of survivors of the 12 most frequent cancer types for diseases of the nervous system and senses organs as compared with matched cancer-free comparison people.

eFigure 10. Adjusted HRs and 95% CIs for hospitalization of survivors of the 12 most frequent cancer types for diseases of the circulatory system as compared with matched cancer-free comparison people.

eFigure 11. Adjusted HRs and 95% CIs for hospitalization of survivors of the 12 most frequent cancer types for diseases of the respiratory system as compared with matched cancer-free comparison people.

eFigure 12. Adjusted HRs and 95% CIs for hospitalization of survivors of the 12 most frequent cancer types for diseases of the digestive system as compared with matched cancer-free comparison people.

eFigure 13. Adjusted HRs and 95% CIs for hospitalization of survivors of the 12 most frequent cancer types for diseases of the skin and subcutaneous tissue as compared with matched cancer-free comparison people.

eFigure 14. Adjusted HRs and 95% CIs for hospitalization of survivors of the 12 most frequent cancer types for diseases of the musculoskeletal system and connective tissue as compared with matched cancer-free comparison people.

eFigure 15. Adjusted HRs and 95% CIs for hospitalization of survivors of the 12 most frequent cancer types for diseases of the genitourinary system as compared with matched cancer-free comparison people.

1.
Rowland  JH, Bellizzi  KM.  Cancer survivorship issues: life after treatment and implications for an aging population.  J Clin Oncol. 2014;32(24):2662-2668. doi:10.1200/JCO.2014.55.8361PubMedGoogle ScholarCrossref
2.
Aziz  NM.  Cancer survivorship research: state of knowledge, challenges and opportunities.  Acta Oncol. 2007;46(4):417-432. doi:10.1080/02841860701367878PubMedGoogle ScholarCrossref
3.
Gjerstorff  ML.  The Danish Cancer Registry.  Scand J Public Health. 2011;39(7)(suppl):42-45. doi:10.1177/1403494810393562PubMedGoogle ScholarCrossref
4.
Pedersen  CB.  The Danish Civil Registration System.  Scand J Public Health. 2011;39(7)(suppl):22-25. doi:10.1177/1403494810387965PubMedGoogle ScholarCrossref
5.
Lynge  E, Sandegaard  JL, Rebolj  M.  The Danish National Patient Register.  Scand J Public Health. 2011;39(7)(suppl):30-33. doi:10.1177/1403494811401482PubMedGoogle ScholarCrossref
6.
Charlson  ME, Pompei  P, Ales  KL, MacKenzie  CR.  A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.  J Chronic Dis. 1987;40(5):373-383. doi:10.1016/0021-9681(87)90171-8PubMedGoogle ScholarCrossref
7.
Baadsgaard  M, Quitzau  J.  Danish registers on personal income and transfer payments.  Scand J Public Health. 2011;39(7)(suppl):103-105. doi:10.1177/1403494811405098PubMedGoogle ScholarCrossref
8.
Jensen  VM, Rasmussen  AW.  Danish Education Registers.  Scand J Public Health. 2011;39(7)(suppl):91-94. doi:10.1177/1403494810394715PubMedGoogle ScholarCrossref
9.
Andersen  PK.  Decomposition of number of life years lost according to causes of death.  Stat Med. 2013;32(30):5278-5285. doi:10.1002/sim.5903PubMedGoogle ScholarCrossref
10.
Overgaard  M, Andersen  PK, Partner  ET.  Regression analysis of censored data using pseudo-observations: an update.  Stata J. 2015;15:809-821. doi:10.1177/1536867X1501500313Google ScholarCrossref
11.
Carver  JR, Shapiro  CL, Ng  A,  et al; ASCO Cancer Survivorship Expert Panel.  American Society of Clinical Oncology clinical evidence review on the ongoing care of adult cancer survivors: cardiac and pulmonary late effects.  J Clin Oncol. 2007;25(25):3991-4008. doi:10.1200/JCO.2007.10.9777PubMedGoogle ScholarCrossref
12.
Crossen  JR, Garwood  D, Glatstein  E, Neuwelt  EA.  Neurobehavioral sequelae of cranial irradiation in adults: a review of radiation-induced encephalopathy.  J Clin Oncol. 1994;12(3):627-642. doi:10.1200/JCO.1994.12.3.627PubMedGoogle ScholarCrossref
13.
Plummer  C, Henderson  RD, O’Sullivan  JD, Read  SJ.  Ischemic stroke and transient ischemic attack after head and neck radiotherapy: a review.  Stroke. 2011;42(9):2410-2418. doi:10.1161/STROKEAHA.111.615203PubMedGoogle ScholarCrossref
14.
Abayomi  J, Kirwan  J, Hackett  A.  The prevalence of chronic radiation enteritis following radiotherapy for cervical or endometrial cancer and its impact on quality of life.  Eur J Oncol Nurs. 2009;13(4):262-267. doi:10.1016/j.ejon.2009.02.007PubMedGoogle ScholarCrossref
15.
Lange  MM, van de Velde  CJ.  Urinary and sexual dysfunction after rectal cancer treatment.  Nat Rev Urol. 2011;8(1):51-57. doi:10.1038/nrurol.2010.206PubMedGoogle ScholarCrossref
16.
Samper-Ternent  R, Zhang  D, Kuo  YF, Hatch  S, Freeman  J.  Late GI and bladder toxicities after radiation for uterine cancer.  Gynecol Oncol. 2011;120(2):198-204. doi:10.1016/j.ygyno.2010.10.017PubMedGoogle ScholarCrossref
17.
Schneider  EC, Malin  JL, Kahn  KL, Ko  CY, Adams  J, Epstein  AM.  Surviving colorectal cancer: patient-reported symptoms 4 years after diagnosis.  Cancer. 2007;110(9):2075-2082. doi:10.1002/cncr.23021PubMedGoogle ScholarCrossref
18.
Buttros  DdeA, Nahas  EA, Vespoli  HdeL, Uemura  G, de Almeida  BdaR, Nahas-Neto  J.  Risk of metabolic syndrome in postmenopausal breast cancer survivors.  Menopause. 2013;20(4):448-454.PubMedGoogle Scholar
19.
Bosco  C, Bosnyak  Z, Malmberg  A, Adolfsson  J, Keating  NL, Van Hemelrijck  M.  Quantifying observational evidence for risk of fatal and nonfatal cardiovascular disease following androgen deprivation therapy for prostate cancer: a meta-analysis.  Eur Urol. 2015;68(3):386-396. doi:10.1016/j.eururo.2014.11.039PubMedGoogle ScholarCrossref
20.
Beijers  AJ, Mols  F, Vreugdenhil  G.  A systematic review on chronic oxaliplatin-induced peripheral neuropathy and the relation with oxaliplatin administration.  Support Care Cancer. 2014;22(7):1999-2007. doi:10.1007/s00520-014-2242-zPubMedGoogle ScholarCrossref
21.
Hershman  DL, Lacchetti  C, Dworkin  RH,  et al; American Society of Clinical Oncology.  Prevention and management of chemotherapy-induced peripheral neuropathy in survivors of adult cancers: American Society of Clinical Oncology clinical practice guideline.  J Clin Oncol. 2014;32(18):1941-1967. doi:10.1200/JCO.2013.54.0914PubMedGoogle ScholarCrossref
22.
Jim  HS, Phillips  KM, Chait  S,  et al.  Meta-analysis of cognitive functioning in breast cancer survivors previously treated with standard-dose chemotherapy.  J Clin Oncol. 2012;30(29):3578-3587. doi:10.1200/JCO.2011.39.5640PubMedGoogle ScholarCrossref
23.
Koppelmans  V, Breteler  MM, Boogerd  W, Seynaeve  C, Schagen  SB.  Late effects of adjuvant chemotherapy for adult onset non-CNS cancer; cognitive impairment, brain structure and risk of dementia.  Crit Rev Oncol Hematol. 2013;88(1):87-101. doi:10.1016/j.critrevonc.2013.04.002PubMedGoogle ScholarCrossref
24.
Holmes  HM, Nguyen  HT, Nayak  P, Oh  JH, Escalante  CP, Elting  LS.  Chronic conditions and health status in older cancer survivors.  Eur J Intern Med. 2014;25(4):374-378. doi:10.1016/j.ejim.2013.12.003PubMedGoogle ScholarCrossref
25.
Leach  CR, Weaver  KE, Aziz  NM,  et al.  The complex health profile of long-term cancer survivors: prevalence and predictors of comorbid conditions.  J Cancer Surviv. 2015;9(2):239-251. doi:10.1007/s11764-014-0403-1PubMedGoogle ScholarCrossref
26.
Keating  NL, Nørredam  M, Landrum  MB, Huskamp  HA, Meara  E.  Physical and mental health status of older long-term cancer survivors.  J Am Geriatr Soc. 2005;53(12):2145-2152. doi:10.1111/j.1532-5415.2005.00507.xPubMedGoogle ScholarCrossref
27.
Gallicchio  L, Kalesan  B, Hoffman  SC, Helzlsouer  KJ.  Non-cancer adverse health conditions and perceived health and function among cancer survivors participating in a community-based cohort study in Washington County, Maryland.  J Cancer Surviv. 2008;2(1):12-19. doi:10.1007/s11764-008-0046-1PubMedGoogle ScholarCrossref
28.
Ng  HS, Roder  D, Koczwara  B, Vitry  A.  Comorbidity, physical and mental health among cancer patients and survivors: an Australian population-based study.  Asia Pac J Clin Oncol. 2018;14(2):e181-e192. doi:10.1111/ajco.12677PubMedGoogle ScholarCrossref
29.
Berry  NM, Miller  MD, Woodman  RJ,  et al.  Differences in chronic conditions and lifestyle behaviour between people with a history of cancer and matched controls.  Med J Aust. 2014;201(2):96-100. doi:10.5694/mja13.10701PubMedGoogle ScholarCrossref
30.
Sarfati  D, Koczwara  B, Jackson  C.  The impact of comorbidity on cancer and its treatment.  CA Cancer J Clin. 2016;66(4):337-350. doi:10.3322/caac.21342PubMedGoogle ScholarCrossref
Original Investigation
March 7, 2019

Long-term Somatic Disease Risk in Adult Danish Cancer Survivors

Author Affiliations
  • 1Survivorship Unit, Danish Cancer Society Research Center, Copenhagen, Denmark
  • 2Unit of Statistics and Pharmacoepidemiology, Danish Cancer Society Research Center, Copenhagen, Denmark
  • 3Childhood Cancer Research Group, Danish Cancer Society Research Center, Copenhagen, Denmark
  • 4Department of Clinical Medicine, Faculty of Health, Aarhus University, Aarhus, Denmark
  • 5Department of Urology, Aarhus University Hospital, Aarhus, Denmark
  • 6Cancer Epidemiology & Population Health, King’s College London, London, England
  • 7The Danish Clinical Registries, Aarhus, Denmark
  • 8Copenhagen Prostate Cancer Center, Department of Urology, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark
  • 9Department of Oncology, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark
  • 10Department of Clinical Oncology & Palliative Care, Zealand University Hospital, Naestved, Denmark
JAMA Oncol. 2019;5(4):537-545. doi:10.1001/jamaoncol.2018.7192
Key Points

Question  What are the incidence and temporal pattern of somatic diseases that require hospitalization in adult cancer survivors?

Findings  In this population-based study of 458 646 cancer survivors and 2 121 567 cancer-free comparison people with up to 17 years of follow-up after cancer diagnosis, cancer survivors had a higher risk of hospitalization for a broad range of incident diseases after the cancer diagnosis than did cancer-free comparison people, mostly pronounced in the first years after the cancer diagnosis.

Meaning  The diversity of health problems in adult cancer survivors suggests the importance of close posttreatment monitoring in the primary and secondary health sectors.

Abstract

Importance  Long-term health effects of cancer in adult survivors are a major concern; however, it is difficult to differentiate between the consequences of cancer and cancer treatment and those of normal aging or comorbidity.

Objectives  To provide an overview and investigate the temporal pattern of hospitalizations for medically verified incident somatic disease in adult survivors compared with cancer-free comparison people, taking into consideration pretreatment comorbidity and the socioeconomic position of the participants.

Design, Setting, and Participants  In this nationwide, population-based cohort-cohort study, 458 646 survivors of the 12 most frequent first primary cancers listed in the Danish Cancer Registry between January 1, 1997, and December 31, 2014, and 2 121 567 matched cancer-free comparison people were identified from the Danish Central Population Registry. Hospitalizations for somatic diseases after cancer diagnosis or study entry were identified from the National Patient Register and stratified according to the 11 main diagnostic groups in the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision. A cohort of cancer survivors was formed for each combination of cancer type and diagnostic group, with a corresponding group of cancer-free people, resulting in 132 unique cohorts. Data analysis was performed from September 1, 2017, to January 15, 2018.

Main Outcomes and Measures  Risk of hospitalization and the temporal pattern of incidence were analyzed in Cox proportional hazards regression models. Cumulative incidence proportions were calculated by the pseudo-observation method.

Results  A total of 2 580 213 people were investigated, of whom 458 646 were cancer survivors (mean [SD] age, 69 [11.6] years; 230 793 [50.3%] male and 227 853 [49.7%] female) and 2 121 567 were comparison people (mean [SD] age, 69 [11.5] years; 1 054 465 [49.7%] male and 1 067 102 [50.3%] female). More cancer survivors vs comparison people had comorbid conditions at the time of cancer diagnosis or study entry (Charlson Comorbidity Index ≥1: 19% vs 13%). Overall, the risk of hospitalizations for somatic diseases was significantly higher for cancer survivors in almost all diagnostic groups (eg, diseases in the nervous system among breast cancer survivors: hazard ratio, 1.20; 95% CI, 1.17-1.22; diseases in the respiratory system in lung cancer survivors: hazard ratio, 5.85; 95% CI, 5.63-6.07; and diseases in blood and blood-forming organs in prostate cancer survivors: hazard ratio, 2.60; 95% CI, 2.50-2.71).

Conclusions and Relevance  The findings suggest that adult survivors of the 12 most common cancers are at significantly higher risk for a broad range of somatic diseases that require hospitalization compared with matched cancer-free comparison people. The results of this study suggest the importance of close, targeted monitoring for new somatic disease during follow-up care of cancer survivors.

Introduction

With the increasing number of cancer survivors worldwide, the long-term health consequences after the end of primary treatment are a major concern.1 Most information about the pattern of health problems after cancer treatment in adults originates from clinical cohorts of adult survivors defined by tumor site and stage or treatment, often without comparison with cancer-free people.2 Among the many elderly survivors of adult cancer, it is difficult to differentiate between the consequences of the cancer and its treatment and the pathological aspects of normal aging, including comorbidity. Knowledge of long-term health consequences is frequently requested by the survivors, their families, and practitioners and is of great value to society as well.

To date, no large population-based study has been conducted to characterize the overall somatic disease risk among survivors of adult cancer. Using the unique, complete, nationwide Danish health registries, we studied the incidence and temporal pattern of somatic diseases that required hospitalization among Danish cancer survivors and compared them with those among cancer-free people in a longitudinal study, taking into account the role of comorbidity present at cancer diagnosis.

Methods

We used a cohort-cohort design with individually matched study participants. The Danish Data Protection Board approved the study protocol. All data linkages were performed in accordance with the Danish Act on Processing of Personal Data with subsequent amendments and European regulations. There was no contact with any of the persons involved. Informed consent was not required because data were stored and analyzed on servers at Statistic Denmark, ensuring the highest degree of confidentiality and with encrypted personal identification numbers. The servers at Statistic Denmark require a multistage login procedure with personal passwords in several steps.

A database of 458 646 adult survivors of the 12 most frequent first primary cancers reported to the Danish Cancer Registry between January 1, 1997, and December 31, 2014, was established. The registry contains information on all individuals in Denmark with a diagnosis of cancer since 1943, including the type of cancer and date of diagnosis. To ensure accurate linkage to the Danish population–based registries, the unique Danish personal identification number of each person, assigned to all Danish residents at the time of birth, was retrieved from the registry.3 To be included in the database, survivors had to be alive 1 year after cancer diagnosis, had to be at least 40 years or older at the time of diagnosis, and had to have 1 of the following cancer types (International Statistical Classification of Diseases and Related Health Problems, Tenth Revision [ICD-10] code): breast (C50), lung (C33, C34), prostate (C61), colon (C18), malignant melanoma (C43), bladder and urinary tract (C65–C68, D09, and D41.4), rectum (C19, C20), brain and central nervous system (C70–C72, D32, D33, D42, and D43), non-Hodgkin lymphoma (C82–C85, C96), pancreas (C25), kidney (C64), or leukemia (C91-C95). People who died or emigrated within 1 year of their cancer diagnosis or since the entry date for comparison people were excluded. Twelve subcohorts were then formed to investigate each cancer type. Data analysis was performed from September 1, 2017, to January 15, 2018.

Comparison People

To account for rates of morbidity in the general Danish population, we matched 5 cancer-free comparison people from the nationwide Danish Central Population Registry4 (n = 2 121 567) on age, sex, and income for each cancer survivor at time of cancer diagnosis.

Information on Somatic Disease Risk After the Cancer Diagnosis

The personal identification numbers of each cancer survivor and cancer-free comparison person were linked to the Danish National Patient Register5 to obtain a full history of inpatient hospitalizations and outpatient hospital contacts (hereafter hospitalizations) for somatic disease after the cancer diagnosis or study entry through December 31, 2014. The register was established in 1977 and contains medically verified diagnoses coded according to ICD-8 and ICD-10 for all somatic diseases that led to hospitalization since 1978 and, since 1995, outpatient visits.

The outcomes were 115 disease categories or diagnoses grouped according to the 11 main diagnostic groups of the ICD-10 system: infectious and parasitic diseases; malignant neoplasms (new primary cancer); diseases of blood and blood-forming organs; endocrine, nutritional, and metabolic diseases; diseases of the nervous system and sense organs; diseases of the circulatory system; diseases of the respiratory system; diseases of the digestive system; diseases of the skin and subcutaneous tissue; diseases of the musculoskeletal system and connective tissue; and diseases of the genitourinary system.

Information on Comorbidity and Covariates

To determine comorbid conditions before study entry in all survivors and cancer-free comparison persons, a 3-year history of information on conditions present before the cancer diagnosis or study entry was obtained. Charlson Comorbidity Index (CCI) scores were computed to measure comorbidity by cumulative addition of up to 19 conditions,6 excluding the cancer in question. Information on covariates, such as highest attained educational level, income, and cohabitation status, was obtained from the Danish social registers,7,8 administered by Statistics Denmark and updated annually since 1981.

Statistical Analysis

Each combination of cancer type and outcome (main diagnostic group) yielded a subcohort of cancer survivors and their corresponding comparison person free of the outcome disease for 132 unique subcohorts for analyses. For each subcohort, only people without a hospital contact for any diagnosis in the diagnostic group in question before study entry were included. All people were followed up from 1 year after the date of cancer diagnosis until a first hospitalization for a diagnosis in the diagnostic group in question, death, emigration, December 31, 2014, or a cancer diagnosis in the comparison person, whichever came first. There was no loss to follow-up because of the complete information in the nationwide, Danish population–based registries.

For each combination of cancer type and outcome, Cox proportional hazards regression models were applied to estimate the hazard ratios (HRs) and 95% CIs for hospitalization among the cancer survivors compared with the comparison people with time since diagnosis as the underlying timescale and stratified by age group at entry. Because matching was performed on the overall cohort level, analyses of subcohorts were adjusted for sex, age, and year of cancer diagnosis or study entry, income, marital status, and CCI (grouped as 0, 1, and ≥2) to avoid residual confounding.

To investigate the absolute difference in the incidence of somatic diseases that required hospitalization and thereby examine the additional risk of hospitalization for each outcome after most patients had completed the 5-year Danish follow-up program, cumulative incidence proportions of disease 5 years after cancer diagnosis or study entry were estimated (cases per 100 people), with death as a competing event. Pseudo-observations were calculated, and linear regression models were fitted to compare the cumulative incidence proportion in the cancer survivors and the cancer-free comparison people, with adjustment for sex, age, period at entry, income, marital status, and CCI.9,10 A generalized linear model was applied with the identity link function, and SEs were estimated with a robust sandwich estimator.

The temporal pattern of the incidence of hospitalization for somatic disease after study entry was investigated by splitting the time since inclusion (ie, 1 year after diagnosis) into 3 time bands (1-2 [early survivorship period], 3-11 [long-term survivorship], and ≥11 years [very long-term survivorship]), and the association with cancer was estimated in each band. A likelihood ratio test with 2 df was used to test whether the cancer association was constant or varied over time. A continuous function of time was modeled in the Cox proportional hazards regression model with restricted cubic splines instead of the 3 piecewise constant estimates in time bands. All P values were from 2-sided tests (likelihood ratio test), and level of significance was P > .05

Data analysis was conducted with Stata, version 14 (StataCorp) and the user-written command –stpci- for pseudo-observations. Plots were constructed using the ggplot2 package in R (R Foundation for Statistical Computing).

Results

A total of 2 580 213 people were investigated, of whom 458 646 were cancer survivors (mean [SD] age, 69 [11.6] years; 230 793 [50.3%] male and 227 853 [49.7%] female) and 2 121 567 were comparison people (mean [SD] age, 69 [11.5] years; 1 054 465 [49.7%] male and 1 067 102 [50.3%] female). More cancer survivors than comparison people had comorbid conditions at the time of cancer diagnosis or study entry (CCI ≥1: 19% vs 13%) (eTable 1 in the Supplement).

Except for some of the estimates for survivors of malignant melanoma and for people with new primary cancers after some cancer types, the risk of hospitalization among cancer survivors was significantly higher than that among cancer-free comparison people (eg, diseases in the nervous system among breast cancer survivors: hazard ratio, 1.20; 95% CI, 1.17-1.22; diseases in the respiratory system in lung cancer survivors: hazard ratio, 5.85; 95% CI, 5.63-6.07; and diseases in blood and blood-forming organs in prostate cancer survivors: hazard ratio, 2.60; 95% CI, 2.50-2.71); this was seen for all diagnostic groups regardless of cancer type (Figure 1 and eFigures 1 and 2 in the Supplement). Additional information on the number of events, person-years at risk, and levels of statistical significance is given in eTable 2 in the Supplement.

Having had cancer was associated with increased absolute excess risk (adjusted cumulative incidence proportion) of hospitalization for a new somatic disease 5 years after diagnosis overall, with small-to-moderate differences between survivors of cancers in many of the diagnostic groups and cancer-free comparison people after adjustment (Figure 2 and eFigures 3 and 4 in the Supplement). Exceptions include a high absolute excess risk of hospitalization with respiratory system disease in survivors of lung cancer (risk difference [RD], 16.75; 95% CI, 15.89-17.61), leukemia (RD, 11.92; 95% CI, 10.79-13.04), and non-Hodgkin lymphoma (RD, 8.94; 95% CI, 7.72-1.15); infections and parasitic disease in survivors of leukemia (RD, 10.12; 95% CI, 9.19-11.06) and non-Hodgkin lymphoma (RD, 7.49; 95% CI, 6.49-8.49); nervous system disease in survivors of brain cancer (RD, 12.43; 95% CI, 11.03-13.84); and digestive system disease in survivors of colon (RD, 9.75; 95% CI, 9.00-10.50), rectum (RD, 12.95; 95% CI, 12.01-13.88), and pancreatic cancer (RD, 9.97; 95% CI, 7.22-12.71) (Figure 2 and eFigures 3 and 4 in the Supplement).

Investigation of the pattern of hospitalization for diseases in the 11 main diagnostic groups for which the risk differed significantly with time showed higher HRs in the first years after diagnosis, followed by a decrease or leveling; however, the rates were still significantly higher than those in the cancer-free comparison group (Figure 3, Figure 4, and eFigures 5–15 in the Supplement). An overall association between the first cancer and new primary cancers could be concluded only for survivors of cancers of the breast (test for 1 overall cancer effect; HR at 10 years since inclusion, 1.15; 95% CI, 1.06-1.25; P < .001) and lung (HR, 1.45; 95% CI, 1.12-1.87; P < .001), indicating that the risk of a new primary cancer of these types was significantly higher during the period studied than in the cancer-free comparison group (Figure 3 and eFigure 5 in the Supplement). For the other cancer types investigated, the risk of new primary cancer remained stable over time. Survivors of prostate cancer had a lower HR of new primary cancer in each time band compared with comparison people (HR at 1-2 years after inclusion, 0.58; 95% CI, 0.54-0.62; HR at 3-10 years after inclusion, 0.70; 95% CI, 0.66-0.73; and HR at ≥10 years after inclusion, 0.71; 95% CI, 0.58-0.87; P < .001) (Figure 3 and eFigure 5 in the Supplement).

A U-shaped curve, indicating significantly increased HRs at the beginning and end of the period (+10 years), was found for some combinations of cancer and diagnostic group (eg, infectious and parasitic disease in survivors of bladder cancer, pancreatic cancer, or non-Hodgkin lymphoma [Figure 4 and eFigure 6 in the Supplement]; diseases of the blood and blood-forming organs in survivors of rectum cancer [eFigure 7 in the Supplement]; endocrine disease in survivors of lung or bladder cancer [eFigure 8 in the Supplement]; diseases of the nervous system in survivors of lung cancer or non-Hodgkin lymphoma [eFigure 9 in the Supplement]; respiratory disease in survivors of non-Hodgkin lymphoma [eFigure 11 in the Supplement]; digestive system disease in survivors of prostate or bladder cancer or non-Hodgkin lymphoma [eFigure 12 in the Supplement]; diseases of the musculoskeletal system and connective tissue in survivors of prostate or kidney cancer [eFigure 14 in the Supplement]; and genitourinary system disease in survivors of prostate cancer [eFigure 15 in the Supplement]).

Discussion

In this large, nationwide, population-based cohort study of disease risk in more than 458 000 adult Danish cancer survivors of the 12 most common cancers, the overall pattern was that, regardless of cancer type, survivors were at significantly higher risk of a first hospitalization for diseases in all but 1 of 11 major diagnostic groups compared with a matched cancer-free comparison group up to 17 years after the cancer diagnosis. Overall, the risks of hospitalization in most diagnostic groups and cancer types were higher in the first few years after diagnosis than those in cancer-free comparison people, after which the risk decreased or leveled with time, although the rates were still significantly higher in the following years than those in the comparison group.

To our knowledge, no previous study has been conducted with such an extensive overview and comprehensive information for inpatient and outpatient hospitalizations for new diseases acquired after cancer. The study thus adds important knowledge to the epidemiology of the long-term somatic disease risk experienced by adult cancer survivors overall.

Few studies are directly comparable to ours, although numerous previous studies11-23 have reported increased risks of somatic diseases among certain cancer survivors or after specific cancer treatments. For many of the diseases under investigation in the current study, the results were consistent with the well-known consequences of cancer and its treatment. For example, radiation insult to noncancer tissue may increase the risks of late cardiovascular and pulmonary disease in survivors of cancers in which the thorax is exposed to radiation.11 Chronic vasculopathy attributable to radiation may affect the medium and large intracranial and extracranial arteries and induce, for example, ischemic stroke or neurobehavioral sequelae in survivors of cancers that involves treatment to the head.12,13 Persistent bowel and bladder toxic effects may arise in survivors of cancers who receive treatment to the pelvis.14-17 Endocrine therapy may be associated with metabolic disturbances or diabetes,18,19 and certain forms of chemotherapy may increase the risks of chronic neuralgia,20 neuropathy,21 and cognitive impairment.22,23

Few studies24,25 have investigated the overall somatic disease risk in adult cancer survivors, and most26-29 were based on self-reported data, had a cross-sectional design, had few study participants, collapsed the findings for all cancer types, and lacked discrimination between comorbidity acquired before and after the cancer diagnosis. The results of these studies nevertheless point to a substantial disease burden among cancer survivors. In one of the larger studies,24 a US cross-sectional telephone survey that involved more than 18 000 cancer survivors and 94 000 control individuals 65 years or older, more cancer survivors reported having 2 or more chronic conditions and poorer health outcomes than similar people without cancer, as identified with a number of burden measures for all cancer types and times since diagnosis. Participants were asked about only a few conditions known to be common in the general population and not conditions specific to cancer survivors, which may have reduced the observed differences between survivors and controls. A study25 in which account was taken of comorbidity before and after the cancer diagnosis gave similar results for 1527 disease-free long-term survivors of breast, prostate, colorectal, and gynecologic cancers. They reported a mean of 1.9 medical conditions (95% CI, 1.8-2.0) after cancer, with survivors of breast cancer reporting the most conditions and survivors of prostate cancer the least. In other cross-sectional studies of cancer survivors and cancer-free control people, more survivors reported circulatory and musculoskeletal conditions and endocrine disturbances than did age- and sex-matched control people26-28 despite having similar lifestyle habits and body mass index.29

Our study design differed from that of most earlier studies and has several strengths, including the nationwide, population-based, longitudinal cohort design and high statistical precision of a comprehensive range of risk estimates owing to the large numbers of survivors and comparison people and the long follow-up. Further strengths include precise linkage with several unique Danish population–based registries with no loss to follow-up and with medically verified hospital diagnoses, which eliminated the possibility of selection bias and differential reporting. Because we included both inpatient and outpatient contacts, we included the full spectrum of patients irrespective of organizational change of hospital care throughout the study period. We had detailed information on comorbidity for all cohort members 3 years before the study entry, and we adjusted for important confounders, such as socioeconomic status.

Overall, survivors of colon, bladder, brain, and kidney cancer and leukemia had modestly higher estimated risks of hospitalization for new primary cancers than did cancer-free comparison people. For the other cancers investigated, the HRs were equal to or lower than those in the comparison group. This study provided support that new primary cancers with a long induction period are rare outcomes in adult cancer survivors, many of whom have shorter lifespans; thus, the proportion of the disease risk attributable to new primary cancers was smaller than that of other somatic diseases.

We counted only the first hospitalization for each group of diagnoses; thus, our estimates of incident morbidity were conservative because people are often at risk for other conditions in the same organ system owing to shared risk factors, unhealthy lifestyle, or genetic predisposition. Although we indirectly removed some of the association attributable to shared lifestyle risk factors by adjusting for preexisting comorbidity and socioeconomic position, the higher morbidity with some diseases, such as pulmonary diseases in lung cancer survivors, may be attributable to cancer treatment but may also represent conditions that are not associated with cancer but are associated with shared risk factors, such as smoking, about which we had no information. Future studies with detailed information on cancer treatment, lifestyle, and genetic disposition should be conducted to determine how risk factors shared among diseases affect the disease burden in cancer survivors.25,30

Limitations

The limitations of this study include lack of information on treatment, which obviated investigation of associations among treatment regimens, treatment doses or combinations, and risks of somatic disease after treatment. The aim of the study, however, was to review the long-term risk of incident somatic disease in adult cancer survivors, including the frequency and distribution of severe somatic diseases that required hospitalization rather than to investigate the complex associations between cancer treatments and late health consequences. Other limitations include lack of information on risk factors, such as lifestyle, which may confound the risk estimates.

Furthermore, neither the Danish National Patient Register nor the Danish Cancer Registry reliably distinguishes relapses from first primary cancers. Although we accounted for death as a competing risk, it is possible that some of the increased risk of hospitalization for survivors who experienced recurrences can be attributed to continued cancer treatment.

Our data on disease burden were derived from registries that include only the most severe conditions that require hospitalization. We lacked information on less severe conditions treated in the primary health sector, and our results probably underestimate the overall somatic disease risk experienced by adult cancer survivors.

It is possible that our results were affected by medical surveillance of the cancer survivors, which could explain the excess diagnosis in the early follow-up period. However, in the long-term survivorship periods, most estimates were still significantly increased compared with the cancer-free comparison people, suggesting that the increased risk of somatic disease persists many years after diagnosis.

Conclusions

The findings suggest that adult survivors of the 12 most frequent cancers are at significantly higher risk for a broad range of incident somatic diseases that require hospitalization than are matched cancer-free comparison people up to 17 years after diagnosis. Survivors had a generally higher risk of hospitalization for subsequent somatic disease in the first few years after diagnosis. The wide diversity of health problems in adult cancer survivors suggests the importance of close posttreatment monitoring in the primary and secondary health sectors. Our results may inform health care initiatives, such as risk-based, targeted follow-up care with early detection and management of additional illnesses associated with cancer survivorship, depending on the cancer type, treatment received, and presence of comorbidity.

Back to top
Article Information

Accepted for Publication: December 13, 2018.

Published Online: March 7, 2019. doi:10.1001/jamaoncol.2018.7192

Correction: This article was corrected on April 11, 2019, to add an affiliation.

Corresponding Author: Trille Kristina Kjaer, PhD, Survivorship Unit, Danish Cancer Society Research Center, Strandboulevarden 49, Copenhagen DK-2100, Denmark (trille@cancer.dk).

Author Contributions: Dr Kjaer was the principal investigator. Drs Kjaer and Andersen had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Kjaer, Andersen, Winther, Bidstrup, Larsen, Johansen, Dalton.

Acquisition, analysis, or interpretation of data: Kjaer, Andersen, Winther, Borre, Moller, Larsen, Johansen, Dalton.

Drafting of the manuscript: Kjaer, Larsen, Dalton.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Kjaer, Andersen, Moller.

Obtained funding: Kjaer, Johansen.

Administrative, technical, or material support: Johansen.

Supervision: Winther, Bidstrup, Johansen, Dalton.

Conflict of Interest Disclosures: None reported.

Funding/Support: This work was supported by grant 17-B-0285 from The Health Foundation (Helsefonden) (Dr Kjaer).

Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

References
1.
Rowland  JH, Bellizzi  KM.  Cancer survivorship issues: life after treatment and implications for an aging population.  J Clin Oncol. 2014;32(24):2662-2668. doi:10.1200/JCO.2014.55.8361PubMedGoogle ScholarCrossref
2.
Aziz  NM.  Cancer survivorship research: state of knowledge, challenges and opportunities.  Acta Oncol. 2007;46(4):417-432. doi:10.1080/02841860701367878PubMedGoogle ScholarCrossref
3.
Gjerstorff  ML.  The Danish Cancer Registry.  Scand J Public Health. 2011;39(7)(suppl):42-45. doi:10.1177/1403494810393562PubMedGoogle ScholarCrossref
4.
Pedersen  CB.  The Danish Civil Registration System.  Scand J Public Health. 2011;39(7)(suppl):22-25. doi:10.1177/1403494810387965PubMedGoogle ScholarCrossref
5.
Lynge  E, Sandegaard  JL, Rebolj  M.  The Danish National Patient Register.  Scand J Public Health. 2011;39(7)(suppl):30-33. doi:10.1177/1403494811401482PubMedGoogle ScholarCrossref
6.
Charlson  ME, Pompei  P, Ales  KL, MacKenzie  CR.  A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.  J Chronic Dis. 1987;40(5):373-383. doi:10.1016/0021-9681(87)90171-8PubMedGoogle ScholarCrossref
7.
Baadsgaard  M, Quitzau  J.  Danish registers on personal income and transfer payments.  Scand J Public Health. 2011;39(7)(suppl):103-105. doi:10.1177/1403494811405098PubMedGoogle ScholarCrossref
8.
Jensen  VM, Rasmussen  AW.  Danish Education Registers.  Scand J Public Health. 2011;39(7)(suppl):91-94. doi:10.1177/1403494810394715PubMedGoogle ScholarCrossref
9.
Andersen  PK.  Decomposition of number of life years lost according to causes of death.  Stat Med. 2013;32(30):5278-5285. doi:10.1002/sim.5903PubMedGoogle ScholarCrossref
10.
Overgaard  M, Andersen  PK, Partner  ET.  Regression analysis of censored data using pseudo-observations: an update.  Stata J. 2015;15:809-821. doi:10.1177/1536867X1501500313Google ScholarCrossref
11.
Carver  JR, Shapiro  CL, Ng  A,  et al; ASCO Cancer Survivorship Expert Panel.  American Society of Clinical Oncology clinical evidence review on the ongoing care of adult cancer survivors: cardiac and pulmonary late effects.  J Clin Oncol. 2007;25(25):3991-4008. doi:10.1200/JCO.2007.10.9777PubMedGoogle ScholarCrossref
12.
Crossen  JR, Garwood  D, Glatstein  E, Neuwelt  EA.  Neurobehavioral sequelae of cranial irradiation in adults: a review of radiation-induced encephalopathy.  J Clin Oncol. 1994;12(3):627-642. doi:10.1200/JCO.1994.12.3.627PubMedGoogle ScholarCrossref
13.
Plummer  C, Henderson  RD, O’Sullivan  JD, Read  SJ.  Ischemic stroke and transient ischemic attack after head and neck radiotherapy: a review.  Stroke. 2011;42(9):2410-2418. doi:10.1161/STROKEAHA.111.615203PubMedGoogle ScholarCrossref
14.
Abayomi  J, Kirwan  J, Hackett  A.  The prevalence of chronic radiation enteritis following radiotherapy for cervical or endometrial cancer and its impact on quality of life.  Eur J Oncol Nurs. 2009;13(4):262-267. doi:10.1016/j.ejon.2009.02.007PubMedGoogle ScholarCrossref
15.
Lange  MM, van de Velde  CJ.  Urinary and sexual dysfunction after rectal cancer treatment.  Nat Rev Urol. 2011;8(1):51-57. doi:10.1038/nrurol.2010.206PubMedGoogle ScholarCrossref
16.
Samper-Ternent  R, Zhang  D, Kuo  YF, Hatch  S, Freeman  J.  Late GI and bladder toxicities after radiation for uterine cancer.  Gynecol Oncol. 2011;120(2):198-204. doi:10.1016/j.ygyno.2010.10.017PubMedGoogle ScholarCrossref
17.
Schneider  EC, Malin  JL, Kahn  KL, Ko  CY, Adams  J, Epstein  AM.  Surviving colorectal cancer: patient-reported symptoms 4 years after diagnosis.  Cancer. 2007;110(9):2075-2082. doi:10.1002/cncr.23021PubMedGoogle ScholarCrossref
18.
Buttros  DdeA, Nahas  EA, Vespoli  HdeL, Uemura  G, de Almeida  BdaR, Nahas-Neto  J.  Risk of metabolic syndrome in postmenopausal breast cancer survivors.  Menopause. 2013;20(4):448-454.PubMedGoogle Scholar
19.
Bosco  C, Bosnyak  Z, Malmberg  A, Adolfsson  J, Keating  NL, Van Hemelrijck  M.  Quantifying observational evidence for risk of fatal and nonfatal cardiovascular disease following androgen deprivation therapy for prostate cancer: a meta-analysis.  Eur Urol. 2015;68(3):386-396. doi:10.1016/j.eururo.2014.11.039PubMedGoogle ScholarCrossref
20.
Beijers  AJ, Mols  F, Vreugdenhil  G.  A systematic review on chronic oxaliplatin-induced peripheral neuropathy and the relation with oxaliplatin administration.  Support Care Cancer. 2014;22(7):1999-2007. doi:10.1007/s00520-014-2242-zPubMedGoogle ScholarCrossref
21.
Hershman  DL, Lacchetti  C, Dworkin  RH,  et al; American Society of Clinical Oncology.  Prevention and management of chemotherapy-induced peripheral neuropathy in survivors of adult cancers: American Society of Clinical Oncology clinical practice guideline.  J Clin Oncol. 2014;32(18):1941-1967. doi:10.1200/JCO.2013.54.0914PubMedGoogle ScholarCrossref
22.
Jim  HS, Phillips  KM, Chait  S,  et al.  Meta-analysis of cognitive functioning in breast cancer survivors previously treated with standard-dose chemotherapy.  J Clin Oncol. 2012;30(29):3578-3587. doi:10.1200/JCO.2011.39.5640PubMedGoogle ScholarCrossref
23.
Koppelmans  V, Breteler  MM, Boogerd  W, Seynaeve  C, Schagen  SB.  Late effects of adjuvant chemotherapy for adult onset non-CNS cancer; cognitive impairment, brain structure and risk of dementia.  Crit Rev Oncol Hematol. 2013;88(1):87-101. doi:10.1016/j.critrevonc.2013.04.002PubMedGoogle ScholarCrossref
24.
Holmes  HM, Nguyen  HT, Nayak  P, Oh  JH, Escalante  CP, Elting  LS.  Chronic conditions and health status in older cancer survivors.  Eur J Intern Med. 2014;25(4):374-378. doi:10.1016/j.ejim.2013.12.003PubMedGoogle ScholarCrossref
25.
Leach  CR, Weaver  KE, Aziz  NM,  et al.  The complex health profile of long-term cancer survivors: prevalence and predictors of comorbid conditions.  J Cancer Surviv. 2015;9(2):239-251. doi:10.1007/s11764-014-0403-1PubMedGoogle ScholarCrossref
26.
Keating  NL, Nørredam  M, Landrum  MB, Huskamp  HA, Meara  E.  Physical and mental health status of older long-term cancer survivors.  J Am Geriatr Soc. 2005;53(12):2145-2152. doi:10.1111/j.1532-5415.2005.00507.xPubMedGoogle ScholarCrossref
27.
Gallicchio  L, Kalesan  B, Hoffman  SC, Helzlsouer  KJ.  Non-cancer adverse health conditions and perceived health and function among cancer survivors participating in a community-based cohort study in Washington County, Maryland.  J Cancer Surviv. 2008;2(1):12-19. doi:10.1007/s11764-008-0046-1PubMedGoogle ScholarCrossref
28.
Ng  HS, Roder  D, Koczwara  B, Vitry  A.  Comorbidity, physical and mental health among cancer patients and survivors: an Australian population-based study.  Asia Pac J Clin Oncol. 2018;14(2):e181-e192. doi:10.1111/ajco.12677PubMedGoogle ScholarCrossref
29.
Berry  NM, Miller  MD, Woodman  RJ,  et al.  Differences in chronic conditions and lifestyle behaviour between people with a history of cancer and matched controls.  Med J Aust. 2014;201(2):96-100. doi:10.5694/mja13.10701PubMedGoogle ScholarCrossref
30.
Sarfati  D, Koczwara  B, Jackson  C.  The impact of comorbidity on cancer and its treatment.  CA Cancer J Clin. 2016;66(4):337-350. doi:10.3322/caac.21342PubMedGoogle ScholarCrossref
×