Introduction
Cancer is a leading cause of death and morbidity worldwide. Early detection through screening and eradication of precancerous lesions is one of the few strategies proven to reduce the risk of late-stage cancer and related morbidity and mortality (1,2). The increased cancer survival observed in high-income countries may be attributed partly to better preventive care, universal screening programs, better early detection methods, and improvements in cancer treatment (3).
Cancer is a significant cause of mortality and morbidity in CKD, with approximately twofold higher prevalence than the general population. Cancer prognosis in CKD is poor; <30% of patients on dialysis or with kidney transplants survive 5 years after cancer diagnoses (4). The increased risk of cancer and cancer-related death seems to be specific for urinary tract, viral-related, digestive, and breast cancers (5,6). Despite the higher incidence and poorer outcomes, some have questioned the long-term benefits and costs of routine screening for patients on dialysis or with advanced CKD due to their limited life expectancies (7,8).
Guideline recommendations for regular cancer screening are the same for people with and without CKD (9). For example, women with CKD should start breast cancer screening at age 50 years old and continue every 2 years until age 74 years old (10). Similarly, cervical cancer screening should commence at age 21 years old (or when sexually active) and continue every 3 years until age 70 years old (11). Currently, mass screening programs for breast and cervical cancers are available in most developed countries irrespective of CKD status, but no data describe how often these screenings are used in routine care for patients with CKD. As part of a broad research program assessing the value of cancer screening in CKD, the aim of this study was to examine patterns of breast and cervical cancer screening in women stratified by kidney function and age and identify factors potentially associated with receiving screening.
Materials and Methods
Design and Setting
In Ontario, over 13 million people have universal access to publically funded health care. We conducted a retrospective, population–based, cohort study from 2002 to 2013 using the Ontario, Canada administrative health care databases held at the Institute for Clinical Evaluative Sciences (ICES). Datasets were linked using unique encoded identifiers and analyzed at the ICES. This study was approved by the institutional review board at Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada. The study was conducted and reported according to the Reporting of Studies Conducted Using Observational Routinely Collected Health Data guideline (12) (Supplemental Table 1).
Data Sources
We used seven linked Ontario–wide administrative databases and two outpatient laboratory databases. We used the Registered Persons Database to collect information on demographics, including birth and death dates, for all Ontarians with a valid health card. The Canadian Organ Replacement Register contains information on all kidney transplant recipients and patients receiving chronic dialysis. This information was used to examine CKD status in combination with two laboratory databases: Dynacare, a large outpatient laboratory service provider across Ontario, and outpatient laboratory values from a linked hospital network in southwestern Ontario. We used the Ontario Cancer Registry, which includes information on incident cancer diagnoses and cancer mortality, to exclude those with prior cancer diagnoses.
Previous health care use, diagnoses, and procedures data were ascertained from the Canadian Institute for Health Information: Discharge Abstract Database and the National Ambulatory Care Reporting System, which include information on hospitalizations and emergency department visits, respectively. The Ontario Health Insurance Plan database collects information on all physician reimbursement claims and was used for covariate and outcome data to capture physician billings for screening. The Ontario Breast Screening Program database contains screening data from those enrolled in the Cancer Care Ontario prevention program and was used to capture breast cancer screening.
Patient Selection
Our study involved two separate cohorts to address breast and cervical cancer screening. The breast cancer screening cohort included women between the ages of 50 and 74 years old, and the cervical cancer screening cohort included women between the ages of 21 and 70 years old. These age restrictions follow guidelines for breast and cervical cancer screening in Ontario (13–15).
Patients entered the cohort between April 1, 2002 and March 31, 2010. The cohort entry date was on the basis of the date of their kidney status, which included kidney transplantation, first evidence of dialysis, or first evidence of stable kidney function. To evaluate stable kidney function in women not receiving renal RRT, we used outpatient serum creatinine laboratory values and the CKD Epidemiology Collaboration equation to calculate the eGFR (16). Stable kidney function was defined as two eGFR values within 5 ml/min per 1.73 m2 or ≤5% of each other separated by at least 3 months but <1 year. We further categorized patients with stable kidney function as (1) no CKD/CKD stage 1 or 2, (2) CKD stage 3a, (3) CKD stage 3b, (4) CKD stage 4, and (5) CKD stage 5 with no evidence of dialysis (Supplemental Table 2 shows definitions of CKD stages).
Patients were categorized into only one group on the basis of the following order: kidney transplantation, dialysis, or stable kidney function. A patient with a functioning kidney transplant and history of dialysis was counted in the kidney transplantation group. Similarly, patients with both dialysis and stable kidney function were included in the dialysis group. We excluded individuals from both cohorts with prior kidney transplantation. For the breast cancer screening cohort, we also excluded patients with breast cancer screening before the age of 49 years old before cohort entry (as a proxy for family history of breast cancer). We also excluded women with hysterectomies from the cervical cancer screening cohort.
Covariates of Interest
We collected baseline data on demographics, including age, residential status, neighborhood income quintile (proxy for socioeconomic status), dialysis vintage (among the dialysis group), and long–term care residence (in the last 90 days). We looked back 5 years from cohort entry for evidence of diabetes, hypertension, myocardial infarction, hemorrhagic stroke, ischemic stroke, dementia, and depression. We looked back 1 year for evidence of general practitioner visits. Overall comorbidity was estimated using the Charlson comorbidity index (17–19).
Ascertainment of Outcomes
The primary outcome was cancer screening during the follow-up period determined on the basis of the recommended frequency of cancer screening. Within the breast cancer screening cohort, we followed patients for up to 2 years for evidence of at least one mammography. For the cervical cancer screening cohort, we followed participants for up to 3 years for evidence of at least one Papanicolaou test (Supplemental Table 2 shows outcome definitions).
Statistical Analyses
We calculated the 2- and 3-year cumulative incidences for breast and cervical cancer screening, respectively, using the exponential equation. We also estimated incidence rates of screening defined as the rate per 1000 person-years of follow-up. Cumulative incidences and incidence rates were censored for death. Screening results were also stratified by age to observe any modification in the association between kidney function and screening rates (50–60, 60–70, and 70–74 years old for breast; 21–30, 31–40, 41–50, 51–60, and 61–70 years old for cervical).
We defined the follow-up time as time from cohort entry to the first screening event. Those who did not receive screening were censored at the end of follow-up or time of death. Cox proportional hazards models were built to obtain hazard ratios (HRs) for screening across kidney function in both cohorts and included interaction terms for age. We used univariable and multivariable hazard regression models to examine potential factors associated with screening for breast and cervical cancers. Statistical analyses were conducted using SAS software (version 9.4; SAS Institute Inc.). Two–sided P values <0.05 were considered statistically significant.
Results
Baseline Characteristics
A total of 141,326 and 324,548 women were included in the breast and cervical cancer screening cohorts, respectively (Supplemental Figure 1). Among patients in the breast cancer screening cohort, 15,383 had CKD stages 3a and 3b, 984 had CKD stages 4 and 5, 3499 were receiving dialysis, and 325 were transplant recipients. For patients in the cervical cancer screening cohort, 10,999 had CKD stages 3a and 3b, 723 had CKD stages 4 and 5, 3811 were receiving dialysis, and 950 had a kidney transplant. Table 1 (breast) and Table 2 (cervical) show the baseline characteristics of each cohort. The mean (SD) ages of women within the breast and cervical cancer screening cohorts were 64.4 (±6.7) and 49.4 (±12.9) years old, respectively. Charlson comorbidity index scores and comorbidities varied across the different groups. Over 60% of women on dialysis had a Charlson score above 3 compared with <40% of women with kidney transplants and women with CKD stages 3a and 3b. Women on dialysis were more likely to have hypertension, diabetes, and cardiovascular disease than those with CKD not on dialysis.
Table 1. -
Baseline characteristics of the breast cancer screening cohort (
n=141,326)
| Patient Characteristics |
CKD |
|
|
| CKD, n=121,135 |
Stages 3a and 3b, n=15,383 |
Stages 4 and 5, n=984 |
Dialysis,
a
n=3499 |
Transplant, n=325 |
|
Age groups, yr
|
|
|
|
|
|
| 50–59 |
32,157 (26.5) |
84 (8.5) |
1059 (6.9) |
661 (18.9) |
122 (37.5) |
| 60–69 |
59,698 (49.3) |
404 (41.1) |
6691 (43.5) |
1618 (46.2) |
156 (48.0) |
| 70–74 |
29,280 (24.2) |
496 (50.4) |
7633 (49.6) |
1220 (34.9) |
47 (14.5) |
| Age, yr
b
|
64.4 (6.7) |
68.3 (5.0) |
68.0 (5.6) |
65.7 (6.3) |
62.1 (6.4) |
|
Income quintile,
c
|
|
|
|
|
|
| 1 (Lowest) |
24,655 (20.4) |
270 (27.4) |
3310 (21.5) |
1038 (29.7) |
73 (22.5) |
| 2 |
26,890 (22.2) |
225 (22.9) |
3583 (23.3) |
834 (23.8) |
64 (19.7) |
| 3 |
25,083 (20.7) |
199 (20.2) |
3214 (20.9) |
653 (18.7) |
71 (21.8) |
| 4 |
22,928 (18.9) |
169 (17.2) |
2810 (18.3) |
536 (15.3) |
63 (19.4) |
| 5 (Highest) |
21,579 (17.8) |
121 (12.3) |
2466 (16.0) |
438 (12.5) |
54 (16.6) |
|
Rural residential status
d
|
10,022 (8.3) |
119 (12.1) |
1703 (11.1) |
483 (13.8) |
30 (9.2) |
|
Comorbidities
|
|
|
|
|
|
| Charlson comorbidity index
e
|
|
|
|
|
|
|
0–1
|
113,794 (93.9) |
564 (57.3) |
12,857 (83.6) |
206 (5.9) |
33 (10.2) |
|
2
|
4610 (3.8) |
142 (14.4) |
1287 (8.4) |
659 (18.8) |
151 (46.5) |
|
≥3
|
2731 (2.3) |
278 (28.3) |
1239 (8.1) |
2364 (67.6) |
141 (43.3) |
| Diabetes |
37,503 (31.0) |
557 (56.6) |
6135 (39.9) |
2303 (65.8) |
117 (36.0) |
| Hypertension |
79,053 (65.3) |
866 (88.0) |
12,897 (83.8) |
3178 (90.8) |
283 (87.1) |
| Myocardial infarction |
1729 (1.4) |
83 (8.4) |
558 (3.6) |
505 (14.4) |
9 (2.8) |
| Hemorrhagic stroke |
196 (0.2) |
<6 (0.6) |
40 (0.3) |
21 (0.6) |
<6 (1.8) |
| Ischemic stroke |
787 (0.6) |
34 (3.5) |
271 (1.8) |
211 (6.0 |
8 (2.5) |
| Dementia |
3079 (2.5) |
687 (4.5) |
60 (6.1) |
264 (7.5) |
8 (2.5) |
| Depression |
11,670 (9.6) |
1608 (10.5) |
129 (13.1) |
496 (14.2) |
37 (11.4) |
|
Health care utilization
|
|
|
|
|
|
| General practitioner visits in the previous 1 yr |
|
|
|
|
|
|
Mean (SD)
|
9.07 (7.66) |
10.49 (9.06) |
12.85 (12.80) |
14.97 (18.06) |
5.92 (7.20) |
|
Median (IQR)
|
7 (5–11) |
8 (5–13) |
10 (6–16) |
9 (4–19) |
4 (1–8) |
| Long–term care status in previous 90 d |
1094 (0.9) |
201 (1.3) |
21 (2.1) |
174 (5.0) |
1 (0.3) |
Values are n (%) unless otherwise noted. IQR, interquartile range.
aThe mean (SD) dialysis vintage was 1.20 (2.01) years. The median (IQR) dialysis vintage was 0.3 (0.3–1.1) years.
bValues are mean (SD).
cIncome quintile category 3 includes the missing values.
dRural defined as population <10,000.
ePatients with no hospitalizations (i.e., missing Charlson score) were coded as zero.
Table 2. -
Baseline characteristics of the cervical cancer screening cohort (
n=324,548)
| Patient Characteristics |
CKD |
|
|
| No CKD, n=308,065 |
Stages 3a and 3b, n=10,999 |
Stages 4 and 5, n=723 |
Dialysis,
a
n=3811 |
Transplant, n=950 |
|
Age groups, yr
|
|
|
|
|
|
| 21–30 |
32,091 (10.4) |
55 (0.5) |
<6 (0.8) |
137 (3.6) |
72 (7.6) |
| 31–40 |
53,055 (17.2) |
172 (1.6) |
28 (3.9) |
260 (6.8) |
183 (19.3) |
| 41–50 |
75,939 (24.7) |
519 (4.7) |
64 (8.9) |
539 (14.1) |
253 (26.6) |
| 51–60 |
79,749 (25.9) |
2320 (21.1) |
153 (21.2) |
1056 (27.7) |
266 (28.0) |
| 61–70 |
67,231 (21.8) |
7933 (72.1) |
471 (65.1) |
1819 (47.7) |
176 (18.5) |
| Age, yr
b
|
48.8 (12.9) |
62.8 (7.4) |
61.0 (9.4) |
56.8 (11.3) |
48.7 (12.0) |
|
Income quintile
c
|
|
|
|
|
|
| 1 (Lowest) |
60,119 (19.5) |
2343 (21.3) |
200 (27.7) |
1189 (31.2) |
228 (24.0) |
| 2 |
64,607 (21.0) |
2472 (22.5) |
161 (22.3) |
869 (22.8) |
191 (20.1) |
| 3 |
64,176 (20.8) |
2250 (20.5) |
148 (20.5) |
710 (18.6) |
187 (19.7) |
| 4 |
62,455 (20.3) |
2062 (18.7) |
122 (16.9) |
581 (15.2) |
175 (18.4) |
| 5 (Highest) |
56,708 (18.4) |
1872 (17.0) |
92 (12.7) |
462 (12.1) |
169 (17.8) |
|
Rural residential status
d
|
19,196 (6.2) |
1198 (10.9) |
63 (8.7) |
527 (13.8) |
85 (8.9) |
|
Comorbidities
|
|
|
|
|
|
| Charlson comorbidity index
e
|
|
|
|
|
|
|
0–1
|
296,197 (96.1) |
9096 (82.7) |
443 (61.3) |
266 (7.0) |
122 (12.8) |
|
2
|
7837 (2.5) |
971 (8.8) |
93 (12.9) |
879 (23.1) |
451 (47.5) |
|
≥3
|
4031 (1.3) |
932 (8.5) |
187 (25.9) |
2666 (70.0) |
377 (39.7) |
| Diabetes |
62,883 (20.4) |
4255 (38.7) |
360 (49.8) |
2339 (61.4) |
296 (31.2) |
| Hypertension |
118,726 (38.5) |
8659 (78.7) |
630 (87.1) |
3357 (88.1) |
789 (83.1) |
| Myocardial infarction |
1902 (0.6) |
343 (3.1) |
39 (5.4) |
456 (12.0) |
24 (2.5) |
| Hemorrhagic stroke |
282 (0.1) |
28 (0.3) |
<6 (0.8) |
20 (0.5) |
<6 (1.8) |
| Ischemic stroke |
816 (0.3) |
176 (1.6) |
11 (1.5) |
199 (5.2) |
10 (1.1) |
| Dementia |
3227 (1.0) |
354 (3.2) |
29 (3.2) |
198 (5.2) |
11 (1.2) |
| Depression |
36,252 (11.8) |
1368 (12.4) |
100 (13.8) |
597 (15.7) |
110 (11.6) |
|
Health care utilization
|
|
|
|
|
|
| General practitioner visits in the previous 1 yr |
|
|
|
|
|
|
Mean (SD)
|
8.61 (7.58) |
10.32 (9.20) |
12.35 (12.39) |
13.46 (16.89) |
6.14 (8.00) |
|
Median (IQR)
|
7 (4–11) |
8 (5–13) |
9 (5–15) |
8 (3–17) |
4 (2–8) |
| Long–term care status in previous 90 d |
905 (0.3) |
101 (0.9) |
13 (1.8) |
127 (3.3) |
3 (0.3) |
Values are n (%) unless otherwise noted. IQR, interquartile range.
aThe mean (SD) dialysis vintage was 1.20 (2.10) years. The median (IQR) dialysis vintage was 0.3 (0.3–1.0) years.
bValues are mean (SD).
cIncome quintile category 3 includes the missing values.
dRural defined as population <10,000.
ePatients with no hospitalizations (i.e., missing Charlson score) were coded as zero.
Breast Cancer Screening
Figure 1 and Table 3 show the 2-year cumulative incidence, incidence rates, and HRs of breast cancer screening by kidney function and age. Overall, women with more advanced CKD were less likely to undergo breast cancer screening (HR of dialysis versus no CKD was 0.32; 95% confidence interval [95% CI], 0.30 to 0.35). The 2-year cumulative incidences were 60.8% for women without CKD, 53.6% for CKD stages 3a and 3b, 36.5% for CKD stages 4 and 5, 26.0% for dialysis, and 53.0% for transplant. Over the 2-year follow-up, 1.5% of patients died: no CKD (0.7%), CKD stages 3a and 3b (1.9%), CKD stages 4 and 5 (7.3%), and dialysis (26.7%). Among kidney transplant recipients, 6.2% died, and 6.2% experienced graft failure during follow-up.
;)
Figure 1.: Two-year cumulative incidence of breast cancer screening in women with and without CKD.
Table 3. -
Two-year cumulative incidences, incidence rates, and HRs of breast cancer screening by CKD stage and age
| Age Groups and CKD Stage |
2-yr Cumulative Incidence, % |
Incidence Rate per 1000 person-yr |
HR (95% CI) |
|
Total cohort
|
|
|
|
| CKD |
|
No CKD
|
60.8 |
468.4 |
1.0 (reference) |
|
CKD stages 3a and 3b
|
53.6 |
383.7 |
0.82 (0.80 to 0.84) |
|
CKD stages 4 and 5
|
36.5 |
227.4 |
0.49 (0.44 to 0.54) |
| Dialysis |
26.0 |
150.5 |
0.32 (0.30 to 0.35) |
| Transplant |
53.0 |
377.5 |
0.81 (0.70 to 0.94) |
|
50–59 yr old
|
|
|
|
| CKD |
|
No CKD
|
55.7 |
406.7 |
1.0 (reference) |
|
CKD stages 3a and 3b
|
47.0 |
317.2 |
0.78 (0.71 to 0.85) |
|
CKD stages 4 and 5
|
34.4 |
210.8 |
0.52 (0.36 to 0.75) |
| Dialysis |
34.8 |
214.1 |
0.53 (0.46 to 0.61) |
| Transplant |
47.5 |
322.0 |
0.80 (0.61 to 1.03) |
|
60–69 yr old
|
|
|
|
| CKD |
|
No CKD
|
66.6 |
547.8 |
1.0 (reference) |
|
CKD stages 3a and 3b
|
60.7 |
466.3 |
0.85 (0.83 to 0.88) |
|
CKD stages 4 and 5
|
43.4 |
284.9 |
0.52 (0.45 to 0.61) |
| Dialysis |
28.9 |
170.4 |
0.31 (0.28 to 0.34) |
| Transplant |
56.3 |
414.1 |
0.76 (0.62 to 0.93) |
|
70–74 yr old
|
|
|
|
| CKD |
|
No CKD
|
54.3 |
391.8 |
1.0 (reference) |
|
CKD stages 3a and 3b
|
47.9 |
325.7 |
0.84 (0.81 to 0.87) |
|
CKD stages 4 and 5
|
31.1 |
186.3 |
0.48 (0.41 to 0.57) |
| Dialysis |
17.0 |
93.0 |
0.24 (0.21 to 0.28) |
| Transplant |
56.2 |
412.5 |
1.05 (0.71 to 1.55) |
A lower incidence of screening in women with advanced CKD compared with no CKD was most pronounced in older age groups (interaction P value <0.001). For women ages 50–59 years old, the 2-year cumulative incidence for breast cancer screening varied from 34.8% (women on dialysis) to 55.7% (women without CKD) (HR, 0.53; 95% CI, 0.46 to 0.61). Corresponding HRs for screening among women ages 60–69 and 70–74 years old were 28.9%–66.6% (HR, 0.31; 95% CI, 0.28 to 0.34) and 13.9%–53.0% (HR, 0.24; 95% CI, 0.21 to 0.28), respectively.
Cervical Cancer Screening
Figure 2 and Table 4 show the 3-year cumulative incidences, incidence rates, and HRs of cervical cancer screening by CKD stage and age. Compared with women without CKD, cervical cancer screening was significantly lower in women with more advanced CKD (HR of dialysis compared with no CKD was 0.31; 95% CI, 0.29 to 0.33). Overall, the 3-year cumulative incidences of cervical cancer screening were 76.1% for women without CKD, 46.5% for CKD stages 3a and 3b, 34.0% for CKD stages 4 and 5, 34.6% for dialysis, and 59.5% for transplant. Over the 3-year follow-up, 1.0% of patients died: no CKD (0.5%), CKD stages 3a and 3b (2.9%), CKD stages 4 and 5 (8.0%), and dialysis (29.7%). Among kidney transplant recipients, 5.1% died, and 5.5% experienced graft failure during follow-up.
Figure 2.: Three-year cumulative incidence of cervical cancer screening in women with and without CKD.
Table 4. -
Three-year cumulative incidences, incidence rates, and HRs of cervical cancer screening by CKD stage and age
| Age Group and CKD Stage |
3-yr Cumulative Incidence, % |
Incidence Rate per 1000 person-yr |
HR (95% CI) |
|
Total cohort
|
|
|
|
| CKD |
|
No CKD
|
76.1 |
477.1 |
1.0 (reference) |
|
CKD stages 3a and 3b
|
46.5 |
208.7 |
0.47 (0.46 to 0.48) |
|
CKD stages 4 and 5
|
34.0 |
138.3 |
0.32 (0.28 to 0.36) |
| Dialysis |
34.6 |
141.4 |
0.31 (0.29 to 0.33) |
| Transplant |
59.5 |
301.1 |
0.65 (0.60 to 0.71) |
|
21–30 yr old
|
|
|
|
| CKD |
|
No CKD
|
85.9 |
653.1 |
1.0 (reference) |
|
CKD stages 3a and 3b
|
66.7 |
366.5 |
0.60 (0.43 to 0.83) |
|
CKD stages 4 and 5
|
63.2 |
333.3 |
0.35 (0.11 to 1.07) |
| Dialysis |
54.9 |
265.1 |
0.44 (0.35 to 0.56) |
| Transplant |
51.5 |
241.2 |
0.41 (0.29 to 0.57) |
|
31–40 yr old
|
|
|
|
| CKD |
|
No CKD
|
87.3 |
687.0 |
1.0 (reference) |
|
CKD stages 3a and 3b
|
68.5 |
384.9 |
0.59 (0.49 to 0.72) |
|
CKD stages 4 and 5
|
81.3 |
559.3 |
0.77 (0.51 to 1.16) |
| Dialysis |
53.2 |
253.0 |
0.40 (0.33 to 0.48) |
| Transplant |
71.6 |
331.2 |
0.64 (0.54 to 0.76) |
|
41–50 yr old
|
|
|
|
| CKD |
|
No CKD
|
83.5 |
601.5 |
1.0 (reference) |
|
CKD stages 3a and 3b
|
71.4 |
417.1 |
0.72 (0.64 to 0.79) |
|
CKD stages 4 and 5
|
58.8 |
295.3 |
0.52 (0.37 to 0.72) |
| Dialysis |
52.8 |
250.3 |
0.44 (0.39 to 0.50) |
| Transplant |
63.0 |
331.2 |
0.58 (0.49 to 0.68) |
|
51–60 yr old
|
|
|
|
| CKD |
|
No CKD
|
73.9 |
447.5 |
1.0 (reference) |
|
CKD stages 3a and 3b
|
60.7 |
311.1 |
0.72 (0.68 to 0.76) |
|
CKD stages 4 and 5
|
39.9 |
169.7 |
0.40 (0.31 to 0.53) |
| Dialysis |
41.7 |
179.8 |
0.41 (0.37 to 0.46) |
| Transplant |
61.1 |
314.5 |
0.72 (0.61 to 0.84) |
|
61–70 yr old
|
|
|
|
| CKD |
|
No CKD
|
53.6 |
255.8 |
1.0 (reference) |
|
CKD stages 3a and 3b
|
39.9 |
170.0 |
0.68 (0.66 to 0.71) |
|
CKD stages 4 and 5
|
25.2 |
96.7 |
0.39 (0.33 to 0.48) |
| Dialysis |
19.7 |
73.2 |
0.29 (0.26 to 0.32) |
| Transplant |
40.5 |
172.9 |
0.68 (0.53 to 0.87) |
Older age significantly modified the association between kidney function and cervical cancer screening (interaction P value <0.001). The 3-year cumulative incidences of screening among women ages 21–30 years old were 54.9% for people on dialysis and 85.9% for people without CKD (HR, 0.44; 95% CI, 0.29 to 0.57). Corresponding numbers for people ages 31–40, 41–50, 51–60, and 61–70 years old were 53.2%–87.3% (HR, 0.40; 95% CI, 0.33 to 0.48), 52.8%–83.5% (HR, 0.44; 95% CI, 0.39 to 0.50), 41.7%–73.9% (HR, 0.41; 95% CI, 0.37 to 0.46), and 19.7%–53.6% (HR, 0.29; 95% CI, 0.26 to 0.32), respectively.
Predictors of Breast and Cervical Cancer Screening
Table 5 summarizes the multivariable models for factors associated with breast and cervical cancer screening stratified by kidney function. Among the CKD stages 3a and 3b group, older age, lower income, long–term care residence, and greater comorbidity (Charlson comorbidity score, diabetes, dementia, and myocardial infarction) were associated with less breast cancer screening. The significant associations of older age and long-term care with lower breast cancer screening were also observed for women on dialysis along with depression and greater dialysis vintage. Older women on dialysis were less likely to undergo breast cancer screening compared with their younger counterparts. For every 10 years of age, the adjusted HR for breast cancer screening for women on dialysis was 0.65 (95% CI, 0.59 to 0.73). Increasing morbidity was independently associated with less breast cancer screening. For every 1-U higher Charlson score, the adjusted HRs for screening mammography were 0.87 (95% CI, 0.83 to 0.92) in women on dialysis and 0.93 (95% CI, 0.92 to 0.94) in those without CKD. Other than long–term care residence, there were no other significant factors for breast cancer screening in transplant recipients.
Table 5. -
Multivariable model for predictors of breast and cervical cancer screening
| Variables of Interest |
CKD Categories at Baseline |
|
|
| No CKD |
CKD Stages 3a and 3b |
CKD Stages 4 and 5 |
Dialysis
a
|
Transplant |
|
Breast cancer
|
|
|
|
|
|
| Age per 10 yr |
1.04 (1.03 to 1.05) |
0.86 (0.82 to 0.89)
b
|
0.84 (0.70 to 1.00)
b
|
0.65 (0.59 to 0.73)
b
|
1.38 (1.08 to 1.76)
b
|
| Rural
c
|
1.03 (1.00 to 1.06) |
1.02 (0.95 to 1.09) |
1.16 (0.85 to 1.58) |
0.91 (0.73 to 1.13) |
1.57 (0.95 to 2.60) |
| Income quintile 1 (lowest; versus quintile 3)
d
|
0.85 (0.83 to 0.87)
b
|
0.91 (0.85 to 0.98)
b
|
0.82 (0.60 to 1.12) |
1.00 (0.81 to 1.25) |
0.93 (0.58 to 1.48) |
| Income quintile 2 |
0.96 (0.94 to 0.98)
b
|
1.03 (0.96 to 1.10) |
0.79 (0.57 to 1.09) |
1.03 (0.82 to 1.28) |
1.04 (0.65 to 1.67) |
| Income quintile 4 |
1.07 (1.04 to 1.09)
b
|
1.12 (1.04 to 1.20)
b
|
0.87 (0.62 to 1.23) |
1.11 (0.87 to 1.42) |
0.72 (0.43 to 1.18) |
| Income quintile 5 (highest) |
1.16 (1.13 to 1.19)
b
|
1.26 (1.17 to 1.35)
b
|
0.90 (0.62 to 1.30) |
1.28 (1.00 to 1.65) |
1.23 (0.78 to 1.94) |
| Charlson per 1 U |
0.93 (0.92 to 0.94)
b
|
0.90 (0.88 to 0.93)
b
|
0.95 (0.86 to 1.04) |
0.87 (0.83 to 0.92)
b
|
0.97 (0.81 to 1.14) |
| Diabetes |
0.79 (0.78 to 0.81)
b
|
0.74 (0.71 to 0.78)
b
|
0.85 (0.68 to 1.08) |
0.97 (0.81 to 1.16) |
0.73 (0.49 to 1.10) |
| Hypertension |
0.99 (0.98 to 1.01) |
0.97 (0.92 to 1.03) |
1.08 (0.77 to 1.52) |
1.02 (0.80 to 1.29) |
1.15 (0.70 to 1.88) |
| Myocardial infarction |
0.78 (0.72 to 0.83)
b
|
0.80 (0.69 to 0.93)
b
|
0.76 (0.44 to 1.29) |
0.96 (0.74 to 1.25) |
1.05 (0.35 to 3.11) |
| Ischemic stroke |
0.87 (0.78 to 0.97)
b
|
0.88 (0.71 to 1.08) |
0.83 (0.40 to 1.70) |
1.28 (0.91 to 1.81) |
0.74 (0.23 to 2.42) |
| General practitioner visits |
1.00 (1.00 to 1.00) |
1.01 (1.00 to 1.01) |
1.00 (0.99 to 1.01) |
1.00 (0.99 to 1.00) |
1.00 (0.98 to 1.03) |
| Long-term care |
0.12 (0.10 to 0.15)
b
|
0.18 (0.11 to 0.29)
b
|
0.34 (0.08 to 1.43) |
0.47 (0.25 to 0.87)
b
|
21.3 (2.05 to 220.8)
b
|
| Dementia |
0.78 (0.74 to 0.83)
b
|
0.76 (0.66 to 0.87)
b
|
0.73 (0.42 to 1.26) |
0.75 (0.52 to 1.10) |
1.14 (0.41 to 3.22) |
| Depression |
0.96 (0.93 to 0.98)
b
|
0.99 (0.92 to 1.07) |
1.44 (1.06 to 1.95) |
0.74 (0.58 to 0.95)
b
|
1.00 (0.60 to 1.68) |
|
Cervical cancer
|
|
|
|
|
|
| Age per 10 yr |
0.84 (0.83 to 0.84)
b
|
0.71 (0.69 to 0.73)
b
|
0.65 (0.58 to 0.73)
b
|
0.73 (0.70 to 0.77)
b
|
0.90 (0.84 to 0.97)
b
|
| Rural
c
|
0.88 (0.86 to 0.89)
b
|
0.87 (0.79 to 0.96)
b
|
1.12 (0.68 to 1.83) |
0.90 (0.75 to 1.09) |
0.96 (0.71 to 1.30) |
| Income quintile 1 (lowest; versus quintile 3)
d
|
0.84 (0.82 to 0.85)
b
|
0.88 (0.80 to 0.97)
b
|
0.73 (0.49 to 1.09) |
0.84 (0.70 to 1.00) |
0.84 (0.65 to 1.09) |
| Income quintile 2 |
0.94 (0.93 to 0.95)
b
|
1.06 (0.97 to 1.16) |
1.18 (0.80 to 1.75) |
0.92 (0.77 to 1.12) |
0.85 (0.65 to 1.12) |
| Income quintile 4 |
1.09 (1.07 to 1.10)
b
|
1.26 (1.15 to 1.38)
b
|
1.07 (0.71 to 1.64) |
0.95 (0.77 to 1.17) |
0.89 (0.67 to 1.17) |
| Income quintile 5 (highest) |
1.18 (1.17 to 1.20)
b
|
1.38 (1.26 to 1.51)
b
|
1.39 (0.89 to 2.16) |
1.12 (0.90 to 1.38) |
1.13 (0.86 to 1.47) |
| Charlson per 1 U |
0.94 (0.93 to 0.94)
b
|
0.91 (0.88 to 0.94)
b
|
0.94 (0.83 to 1.06) |
0.88 (0.83 to 0.92)
b
|
0.92 (0.83 to 1.01) |
| Diabetes |
0.77 (0.76 to 0.78)
b
|
0.70 (0.66 to 0.75)
b
|
0.76 (0.56 to 1.04) |
0.96 (0.82 to 1.11) |
0.91 (0.71 to 1.16) |
| Hypertension |
0.89 (0.88 to 0.90)
b
|
0.88 (0.82 to 0.94)
b
|
0.94 (0.65 to 1.38) |
1.30 (1.09 to 1.57) |
0.94 (0.75 to 1.18) |
| Myocardial infarction |
0.74 (0.69 to 0.79)
b
|
0.80 (0.64 to 0.99)
b
|
0.98 (0.47 to 2.03) |
0.73 (0.55 to 0.97)
b
|
0.91 (0.47 to 1.76) |
| Ischemic stroke |
0.78 (0.70 to 0.87)
b
|
0.73 (0.54 to 1.00)
b
|
0.51 (0.07 to 3.69) |
1.11 (0.77 to 1.60) |
0.55 (0.17 to 1.72) |
| General practitioner visits |
1.00 (1.00 to 1.00) |
1.00 (0.99 to 1.00) |
1.00 (0.98 to 1.01) |
1.00 (0.99 to 1.00) |
1.01 (1.00 to 1.02) |
| Long-term care |
0.13 (0.11 to 0.17)
b
|
0.15 (0.06 to 0.40)
b
|
0.48 (0.06 to 3.64) |
0.47 (0.24 to 0.93)
b
|
0.89 (0.11 to 7.47) |
| Dementia |
0.76 (0.72 to 0.80)
b
|
0.68 (0.54 to 0.84)
b
|
0.78 (0.34 to 1.79) |
0.71 (0.48 to 1.06) |
0.65 (0.22 to 1.87) |
| Depression |
0.97 (0.96 to 0.98)
b
|
1.02 (0.93 to 1.11) |
1.03 (0.69 to 1.54) |
0.99 (0.83 to 1.19) |
0.97 (0.74 to 1.28) |
aDialysis vintage per 1 yr for breast cancer and for cervical cancer were 0.89 (0.85 to 0.93) and 0.86 (0.83 to 0.90), respectively. Both dialysis vintages had significant associations.
bSignificant associations.
cRural defined as population <10,000.
dIncome quintile categories are quintiles 1 (lowest) to 5 (highest).
Across all kidney function groups, increasing age was also independently associated with less cervical cancer screening. For every 10 years of age, the adjusted HRs for cervical cancer screening in women on dialysis, women with CKD stages 4 and 5, and women with kidney transplants were 0.73 (95% CI, 0.70 to 0.77), 0.65 (95% CI, 0.58 to 0.73), and 0.90 (95% CI, 0.84 to 0.97), respectively. Similar to breast cancer screening, there was an independent association between greater comorbidity (Charlson comorbidity score, diabetes, hypertension, dementia, and myocardial infarction) and less cervical cancer screening among the CKD stages 3a and 3b group. Additional factors significantly associated with less cervical cancer screening for the CKD stages 3a and 3b group were lower income and long–term care residence. Among women on dialysis, there were significant associations between higher Charlson scores (HR, 0.88; 95% CI, 0.83 to 0.92 for every 1-U higher Charlson score), long–term care residence, and greater dialysis vintage. Apart from younger age, there were no other predictors for cervical cancer screening in transplant recipients.
Discussion
Using data from a large population–based cohort in Ontario, Canada, we showed that women with CKD were substantially less likely to undergo breast and cervical cancer screening compared with women without CKD. Despite the heightened risk of cervical cancer in young kidney transplant recipients, the 3-year cumulative incidences of cervical cancer screening were 52% in those ages 21–30 years old and 60% among transplant recipients. This association was modified by CKD stages and age. Older age, higher comorbidity, and lower income were associated with lower incidences of cervical and breast cancer screening in women with moderate to advanced CKD. We found that women with the greatest risk of death (older women on dialysis) had the lowest rates of breast and cervical cancer screening.
We also found that women with CKD and within the highest-income quintile were more likely to receive breast and cervical cancer screening. This is not surprising, because differences in socioeconomic status in cancer screening are well documented and may not necessarily reflect barriers, such as direct costs, but may include a lack of knowledge about the importance of early detection, cultural differences, and new immigration (20). The relatively low uptake of cervical cancer screening in the young and at–risk dialysis and transplanted population was also expected. A study in the United Kingdom reported <10% of the transplant population of women followed recommended guidelines for cervical cancer screening (21).
Cancer screening in people with chronic illness, including CKD, is complex. CKD not only increases the risk of cancer but is also a predictor of poor outcomes, driven largely by breast and urinary tract cancers (5). Despite this, literature suggests that routine screening should be avoided in patients on dialysis and those not on the transplant donor waiting list because of reduced life expectancy and therefore, lower likelihood of screening benefit, even if cancer was diagnosed early (7,8). Routine screening in those with limited life expectancy not only has health care resource use implications but may also induce harms through subsequent diagnostic investigations, overdiagnoses, and overtreatment.
Our findings regarding the current cancer screening patterns within this large population are consistent with the recommendations by the Choosing Wisely campaign endorsed by the American Society of Nephrology and other specialty colleges worldwide (22). This initiative aims to improve patient safety and quality care and reduce wasteful resources and unnecessary tests, which may lead to more harm than benefits in patients with chronic diseases. The lower screening rates, particularly for those with comorbidities and those on dialysis, may also reflect patients’ preferences for preventive medicine in the context of chronic illness (23). Apart from cancer, women with CKD have other related comorbidities with shared risk factors, including diabetes, hypertension, cardiovascular disease, and peripheral vascular disease, and over 40% of those with ESRD have more than three comorbidities (24). This is particularly relevant for older women, many of who may have underlying cognitive impairment (25,26) and may not have the ability and resources to cope with complex disease management. A previous study reported cancer prevention and screening as low priority to those on dialysis, and an over-riding attention to CKD care was observed (27).
On the contrary, younger women with kidney transplants should be encouraged to undertake routine screening and a proactive approach to preventive health care. Younger women on dialysis or who have received a kidney transplant are at increased risk of cancers. Population-based studies have reported excess risks of cervical cancers by at least 2.5 and seven times in younger women on dialysis and with kidney transplants, respectively (28). Furthermore, the risk of breast cancer–related death in women with CKD was twofold higher than in women without CKD (5). Early detection through cancer screening may modify the longer–term cancer prognoses in women with CKD. The lower rates of screening among younger women may, therefore, represent a missed opportunity to improve patient care.
Our study has several strengths. We studied a large population–based cohort of over 300,000 women in Ontario, Canada with and without CKD. This is also the first comparative study that examined screening rates for breast and cervical cancers in women across the full spectrum of CKD with different comorbidity profiles. We have also provided 2- and 3-year cumulative incidences of breast and cervical cancer screening in accordance with Canadian guidelines. There are also potential limitations. Proteinuria and albuminuria laboratory data were not available for all participants. As such, we may have underestimated the true prevalence of those with early and moderate CKD. Also, only baseline CKD stages were available. We were unable to evaluate the changing status of kidney function over time and its effect on screening. Detailed information regarding reasons for screening or nonscreening in this cohort was not available. Lastly, we did not have information on potential residual confounders, including medication use, lifestyle, ethnicity, and cognition.
We have shown that women on dialysis and women with advanced CKD are substantially less likely to undergo breast and cervical cancer screening. Understanding the patterns of screening may help target potential interventions, particularly for the underscreened and hard to reach but at-risk population. Trial-based evidence has proven efficacy and efficiency of routine screening for breast and cervical cancers in the general population. For those with chronic illness, such as CKD, because of the complexity in disease management, the decision to screen should be shared between clinicians, patients, and caregivers, considering patients’ preferences and priorities. Efforts should also be directed at understanding potential barriers and enablers for cancer screening in the younger at–risk population, such as those with kidney transplants and those with longer life expectancies.
Disclosures
None.
Acknowledgments
We thank Gamma-Dynacare Laboratories for providing access to their data, and we thank the team at London Health Sciences Centre, St. Joseph’s Health Care, and the Thames Valley Hospitals for providing access to the Cerner laboratory data. Infrastructure support was provided by the Lilibeth Calberto Kidney Clinical Research Unit.
This study was conducted by the Institute for Clinical Evaluative Sciences (ICES) Western facility through the ICES Kidney, Dialysis and Transplantation (KDT) Research Program. The ICES is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). Core funding for the ICES Western is provided by the Academic Medical Organization of Southwestern Ontario (AMOSO), the Schulich School of Medicine and Dentistry (SSMD), Western University, and the Lawson Health Research Institute (LHRI). The ICES KDT team is supported by a grant from the Canadian Institutes of Health Research (CIHR). A.X.G. was supported by the Dr. Adam Linton Chair in Kidney Health Analytics.
The opinions, results, and conclusions reported in this paper are those of the authors and independent from the funding sources. No endorsement by the ICES, the AMOSO, the SSMD, the LHRI, the CIHR, or the MOHLTC is intended or should be inferred. The opinions expressed in this publication are those of the authors/researchers and do not necessarily reflect the official views of the Public Health Agency of Canada. Parts of this material are on the basis of data and information compiled and provided by Canadian Institute for Health Information (CIHI). However, the analyses, conclusions, opinions, and statements expressed herein are those of the authors and are not necessarily those of CIHI. Parts of this material are on the basis of data and information provided by Cancer Care Ontario (CCO). The opinions, results, view, and conclusions reported in this paper are those of the authors and do not necessarily reflect those of CCO. No endorsement by CCO is intended or should be inferred.
References
1. Gøtzsche PC, Nielsen M: Screening for breast cancer with mammography. Cochrane Database Syst Rev 4: CD001877, 200617054145
2. Eddy DM: Screening for cervical cancer. Ann Intern Med 113: 214–226, 19902115753
3. Dowling EC, Klabunde C, Patnick J, Ballard-Barbash R, Quaine J, Lang A, Silva Correa R, Onysko J, Svoboknik A, Lynge E, Malia N, Anttila A, Sancho-Garnier H, Deitz D, Boncz I, Sigurdsson K, O’Brien T, Rennert G, Paci E, Saito H, Scharpantgen A, Lee WC, Fracheboud J, Cox B, Hofvind S, Bulliard JL, Fidaner C, Moss S, Walker R, Pou G; International Cancer Screening Network (ICSN): Breast and cervical cancer screening programme implementation in 16 countries. J Med Screen 17: 139–146, 2010
4. Miao Y, Everly JJ, Gross TG, Tevar AD, First MR, Alloway RR, Woodle ES: De novo cancers arising in organ transplant recipients are associated with adverse outcomes compared with the general population. Transplantation 87: 1347–1359, 200919424035
5. Iff S, Craig JC, Turner R, Chapman JR, Wang JJ, Mitchell P, Wong G: Reduced estimated GFR and cancer mortality. Am J Kidney Dis 63: 23–30, 201423993153
6. Wong G, Staplin N, Emberson J, Baigent C, Turner R, Chalmers J, Zoungas S, Pollock C, Cooper B, Harris D, Wang JJ, Mitchell P, Prince R, Lim WH, Lewis J, Chapman J, Craig J: Chronic kidney disease and the risk of cancer: An individual patient data meta-analysis of 32,057 participants from six prospective studies. BMC Cancer 16: 488, 201627421889
7. LeBrun CJ, Diehl LF, Abbott KC, Welch PG, Yuan CM: Life expectancy benefits of cancer screening in the end-stage renal disease population. Am J Kidney Dis 35: 237–243, 200010676722
8. Kajbaf S, Nichol G, Zimmerman D: Cancer screening and life expectancy of Canadian patients with kidney failure. Nephrol Dial Transplant 17: 1786–1789, 200212270985
9. Wong G, Chapman JR, Craig JC: Cancer screening in renal transplant recipients: What is the evidence? Clin J Am Soc Nephrol 3[Suppl 2]: S87–S100, 2008
10. Tonelli M, Connor Gorber S, Joffres MCanadian Task Force on Preventive Health C, Dickinson J, Singh H, Lewin G, Birtwhistle R, Fitzpatrick-Lewis D, Hodgson N, Ciliska D, Gauld M, Liu YY: Recommendations on screening for breast cancer in average-risk women aged 40–74 years.[Erratum appears in CMAJ. 2011 Dec 13;183(18):2147]. CMAJ 183(17):1991–2001, 2011
11. Parboosingh EJ, Anderson G, Clarke EA, Inhaber S, Kaegi E, Mills C, Mao Y, Root L, Stuart G, Stachenko S: Cervical cancer screening: Are the 1989 recommendations still valid? National Workshop on screening for cancer of the cervix. CMAJ 154: 1847–1853, 1996
12. Benchimol EI, Smeeth L, Guttmann A, Harron K, Moher D, Petersen I, Sørensen HT, von Elm E, Langan SM; RECORD Working Committee: The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement. PLoS Med 12: e1001885, 201526440803
13. Murphy J, Kennedy EB, Dunn S, McLachlin CM, Fung Kee Fung M, Gzik D, Shier M, Paszat L; Ontario Cervical Screening Program; Program in Evidence-based Care: Cervical screening: A guideline for clinical practice in Ontario. J Obstet Gynaecol Can 34: 453–458, 201222555138
14. Wadden N, Doyle GP: Breast cancer screening in Canada: A review. Can Assoc Radiol J 56: 271–275, 200516579020
15. Warner E, Heisey R, Carroll JC: Applying the 2011 Canadian guidelines for breast cancer screening in practice. CMAJ 184: 1803–1807, 201222966059
16. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration): A new equation to estimate glomerular filtration rate. [Erratum appears in Ann Intern Med. 2011 Sep 20;155(6):408]. Ann Intern Med 150: 604–612, 200919414839
17. Fried L, Bernardini J, Piraino B: Charlson comorbidity index as a predictor of outcomes in incident peritoneal dialysis patients. Am J Kidney Dis 37: 337–342, 200111157375
18. Hemmelgarn BR, Manns BJ, Quan H, Ghali WA: Adapting the Charlson comorbidity index for use in patients with ESRD. Am J Kidney Dis 42: 125–132, 200312830464
19. Liu J, Huang Z, Gilbertson DT, Foley RN, Collins AJ: An improved comorbidity index for outcome analyses among dialysis patients. Kidney Int 77: 141–151, 201019907414
20. Pruitt SL, Shim MJ, Mullen PD, Vernon SW, Amick BC 3rd: Association of area socioeconomic status and breast, cervical, and colorectal cancer screening: A systematic review. Cancer Epidemiol Biomarkers Prev 18: 2579–2599, 200919815634
21. Courtney AE, Leonard N, O’Neill CJ, McNamee PT, Maxwell AP: The uptake of cervical cancer screening by renal transplant recipients. Nephrol Dial Transplant 24: 647–652, 200918952575
22. Williams AW, Dwyer AC, Eddy AA, Fink JC, Jaber BL, Linas SL, Michael B, O’Hare AM, Schaefer HM, Shaffer RN, Trachtman H, Weiner DE, Falk AR; American Society of Nephrology Quality, and Patient Safety Task Force: Critical and honest conversations: The evidence behind the “Choosing Wisely” campaign recommendations by the American Society of Nephrology. Clin J Am Soc Nephrol 7: 1664–1672, 201222977214
23. Dolan JG: Patient priorities in colorectal cancer screening decisions. Health Expect 8: 334–344, 200516266421
24. Tonelli M, Wiebe N, Guthrie B, James MT, Quan H, Fortin M, Klarenbach SW, Sargious P, Straus S, Lewanczuk R, Ronksley PE, Manns BJ, Hemmelgarn BR: Comorbidity as a driver of adverse outcomes in people with chronic kidney disease. Kidney Int 88: 859–866, 201526221754
25. Pereira AA, Weiner DE, Scott T, Chandra P, Bluestein R, Griffith J, Sarnak MJ: Subcortical cognitive impairment in dialysis patients. Hemodial Int 11: 309–314, 200717576295
26. Seliger SL, Weiner DE: Cognitive impairment in dialysis patients: Focus on the blood vessels? Am J Kidney Dis 61: 187–190, 201323318010
27. Williams NC, Tong A, Howard K, Chapman JR, Craig JC, Wong G: Knowledge, beliefs and attitudes of kidney transplant recipients regarding their risk of cancer. Nephrology (Carlton) 17: 300–306, 201222171765
28. Vajdic CM, McDonald SP, McCredie MR, van Leeuwen MT, Stewart JH, Law M, Chapman JR, Webster AC, Kaldor JM, Grulich AE: Cancer incidence before and after kidney transplantation. JAMA 296: 2823–2831, 200617179459
