A Description of Advanced Chronic Kidney Disease Patients in a Major Urban Center Receiving Conservative Care

Abstract

Background:

Conservative/palliative (nondialysis) management is an option for some individuals for treatment of stage 5 chronic kidney disease (CKD). Little is known about these individuals treated with conservative care in the Canadian setting.

Objective:

To describe the characteristics of patients treated with conservative care for category G5 non-dialysis CKD in a Canadian context.

Design:

Retrospective chart review.

Setting:

Urban nephrology center.

Patients:

Patients with G5 non-dialysis CKD (estimated glomerular filtration rate <15 mL/min/1.73 m²).

Measurements:

Baseline patient demographic and clinical characteristics of conservative care follow-up, advanced care planning, and death.

Methods:

We undertook a descriptive analysis of individuals enrolled in a conservative care program between January 1, 2009, and June 30, 2015.

Results:

One hundred fifty-four patients were enrolled in the conservative care program. The mean age and standard deviation was 81.4 ± 9.0 years. The mean modified Charlson Comorbidity Index score was 3.4 ± 2.8. The median duration of conservative care participation was 11.5 months (interquartile range: 4-25). Six (3.9%) patients changed their modality to dialysis. One hundred three (66.9%) patients died during the study period. Within the deceased cohort, most (88.2%) patients completed at least some advanced care planning before death, and most (81.7%) of them died at their preferred place. Twenty-seven (26.7%) individuals died in hospital.

Limitations:

Single-center study with biases inherent to a retrospective study. Generalizability to non-Canadian settings may be limited.

Conclusions:

We found that individuals who chose conservative care were very old and did not have high levels of comorbidity. Few individuals who chose conservative care changed modality and accepted dialysis. The proportions of engagement in advanced care planning and of death in place of choice were high in this population. Death in hospital was uncommon in this population.

What was known before

Conservative care is an option for the treatment of advanced kidney disease in patients who do not wish to pursue renal replacement therapy.

What this adds

Comprehensive conservative care for patients with advanced kidney disease in a Canadian setting emphasizes advanced care planning and achieves a low risk of death in hospital. However, information technology infrastructure to prospectively study these patients is lacking.

Background

Chronic kidney disease (CKD) is common, affecting about 12.5% of Canadian adults,¹ and is associated with an increased risk of comorbidity, prolonged hospitalization, and mortality.² While older CKD patients are more likely to die than to progress to category G5 non-dialysis CKD,^3-5 those patients who do develop G5 non-dialysis CKD must decide on a treatment modality, which may or may not include renal replacement therapy. As G5 non-dialysis CKD patients tend to feature a poor quality of life,⁶ and as dialysis may not be life-prolonging in older and comorbid G5 non-dialysis CKD patients,^7-9 conservative care may be the preferred modality of treatment.

Conservative care, also known as palliative or supportive care, has been defined by the World Health Organization as “an approach that improves the quality of life of patients and their families facing the problem associated with life-threatening illness, through the prevention and relief of suffering by means of early identification and impeccable assessment and treatment of pain and other problems, physical, psychosocial and spiritual,”¹⁰ which is also the definition recognized by Kidney Disease: Improving Global Outcomes (KDIGO).¹¹ Renal conservative care programs (CCPs) have emerged across the world with the goal of better addressing the end-of-life needs of advanced CKD patients.^11,12 However, information about patients choosing conservative care for G5 non-dialysis CKD in a Canadian context is lacking. A characterization of this population would help guide the development of CCPs in Canada.

The objective of this study was to describe the characteristics of patients choosing conservative care for the treatment of G5 non-dialysis CKD in a Canadian context. We conducted a retrospective cohort study involving patients enrolled in a CCP in an urban Canadian nephrology center. We carried out a descriptive analysis based on patient characteristics and information regarding conservative care follow-up and death.

Methods

We designed a single-center retrospective cohort study to characterize renal patients enrolled in our CCP. It involved a chart review of patient records via an electronic data source. Adult patients with G5 non-dialysis CKD enrolled in our CCP between January 1, 2009, and June 30, 2015 in Calgary, Alberta, Canada, were included in this study. The onset date of G5 non-dialysis CKD was defined as the first of the two earliest consecutive estimated glomerular filtration rate (eGFR) measurements at least three months apart of less than 15 mL/min/1.73 m², whereby all subsequent eGFR measurements remained less than 15 mL/min/1.73 m². We excluded patients whose eGFR was unknown or unmeasured.

The CCP was unique to this nephrology center. Patients were enrolled in the CCP after opting to forgo dialysis or after withdrawal of dialysis if they were not actively dying. The CCP involved patients who were referred by a nephrologist who had been previously following the patients at a multidisciplinary CKD clinic (generally involving patients with eGFRs below 30 mL/min/1.73 m²). A team with an interest in palliative care followed these patients, including a nephrologist with fellowship training in palliative medicine, an advanced care planning nurse clinician, and dedicated conservative care nurses (a full-time equivalent registered nurse divided between 2 individuals). A social worker and dietician were also available as needed. The CCP provided patient education about G5 non-dialysis CKD and about a trajectory without dialysis. Because patients chose enrollment into the program, the CCP provided active medical care that was specific to the goals and wishes of patients and their families. Symptom control was a priority. Patient follow-up was individualized and variable. Patients were followed in clinic and between clinic visits with phone calls and home visits. This program offered facilitated advanced care planning, which included a Goal of Care designation, though patients could have chosen to not engage in advanced care planning. It also offered bereavement follow-up with family members. Patients were encouraged to maintain relationships with their primary care providers.

Data were obtained from the prospective database of the Southern Alberta Renal Program.¹³ This database kept a record of patient demographics, blood work results, physician consultation notes, and multidisciplinary progress notes.

In this chart review, we sought to describe patient characteristics (eg, demographics, comorbidities, and laboratory results) at the onset of G5 non-dialysis CKD. We used the Charlson Comorbidity Index score for G5 non-dialysis CKD patients¹⁴ to describe the level of comorbidity burden. We considered a late referral to the multidisciplinary CKD clinic to be one that occurred after onset of or at ≤90 days before the onset of G5 non-dialysis CKD. We then described the details of CCP participation until the end of the study (eg, duration of follow-up and proportion of deceased patients by the end of the study). Finally, we characterized those patients who died while in the program, including details about advanced care planning and details of death. The details of death were already collected by the conservative care physician and nurses prior to this chart review. We determined the cause of death to be due to uremia if there was no other apparent cause of death and the course of illness included the typical constellation of symptoms associated with uremia (eg, encephalopathy, myoclonus, nausea, anorexia, and pruritus). The Dialysis Quality of Dying Apgar^15,16 was used since the beginning of the CCP to rate the quality of death, out of a maximum score of 10. A higher score represented a higher quality of death. Despite the limited use of this tool in the literature,¹⁷ we employed the Dialysis Quality of Dying Apgar as it was the only such tool specific to G5 non-dialysis CKD patients available at the time of the program’s inception, and because of its ease of use.

The results of this chart review were analyzed descriptively in terms of mean (standard deviation), median (interquartile range [IQR]), and proportions for numerical data. The University of Calgary’s Conjoint Health Research Ethics Board granted ethics approval for this study.

Results

Among 241 patients enrolled in the CCP since the study start date, 63 deceased and 24 alive patients (by the end of the study date) did not meet our inclusion criteria as they had not reached G5 non-dialysis CKD. Therefore, we included 154 patients in this chart review for a descriptive analysis (Figure 1). We summarized the patients’ baseline characteristics at the time of study enrollment (onset of G5 non-dialysis CKD; Table 1). The mean age was 81.4 ± 9.0 years. The most common CKD etiologies were hypertension/ischemia and diabetes mellitus. About one-third of patients had a late referral to the multidisciplinary CKD clinic. The mean Charlson Comorbidity Index score was not high. Few patients previously underwent dialysis or kidney transplantation. The CKD lab parameters that were measured at the onset of G5 non-dialysis CKD are also listed in Table 1.

Figure 1.

Study cohort.

Note. CCP = conservative care program; CKD = chronic kidney disease.

Table 1.

Baseline Patient Characteristics in the Renal Conservative Care Program in Calgary, Alberta.

Characteristic at G5 non-dialysis CKDa onset	All patients (n = 154)
Age, mean (SD), years	81.4 (9.0)
Male, n (%)	71 (46.1)
Smoking history, n (%)	65 (42.2)
BMI, mean (SD), kg/m2	27.8 (6.3)
CKD etiology, n (%)b
Hypertension/ischemia	81 (52.6)
Diabetes mellitus	67 (43.5)
Reflux/obstruction	13 (8.4)
Glomerulonephritis	4 (2.6)
Polycystic kidney disease	1 (0.6)
Otherc	17 (11.0)
Unknown	15 (9.7)
Charlson Comorbidity Index score, mean (SD)	3.4 (2.8)
Comorbidities, n (%)b
Diabetes with end-organ damage	69 (44.8)
Congestive heart failure	40 (26.0)
Cerebrovascular disease	39 (25.3)
Myocardial infarction	38 (24.7)
Chronic pulmonary disease	32 (20.8)
Peripheral vascular disease	27 (17.5)
Dementia	22 (14.3)
Diabetes without end-organ damage	17 (11.0)
Peptic ulcer disease	12 (7.8)
Metastatic cancer	7 (4.5)
Rheumatological disease	7 (4.5)
Leukemia	1 (0.6)
Lymphoma	1 (0.6)
Moderate or severe liver disease	1 (0.6)
Late referral to multidisciplinary CKD clinic, n (%)	54 (35.1)
Prior kidney transplant, n (%)	1 (0.6)
Prior dialysis, n (%)	8 (5.2)
Hemodialysis	6 (75)
Peritoneal dialysis	2 (25)
Duration on dialysis, median (IQR), months	10 (2.8-45.2)
Laboratory measurements, mean (SD)
Hemoglobin, g/L	108 (16)
Phosphate, mmol/L	1.5 (0.3)
Calcium, mmol/L	2.4 (0.2)
Calcium phosphate product, mmol2/L2	3.7 (0.8)
Parathyroid hormone, ng/L	187 (172)

Note. CKD = chronic kidney disease; BMI = body mass index.

^aEstimated glomerular filtration rate <15 mL/min/1.73 m².

^bThe total percentages for CKD etiology and comorbidities exceeded 100% because some patients had more than one etiology and comorbidity.

^cOther etiologies include drug-induced, infection, and tubulointerstitial disease.

The details of participation in the CCP are summarized in Table 2. The median duration of participation in the CCP was 11.5 (IQR: 4-25) months. Advanced care planning was initiated before the end of the study in most patients. The majority of patients died in the CCP before the study end date, while few patients switched treatment modality from conservative care to dialysis.

Table 2.

Conservative Care Program Patient Outcomes.

Outcome	All patients (n = 154)
Duration of participation, median (IQR), months	11.5 (4-25)
Advanced care planning initiated before program exit, n (%)	109 (70.8)
Goal of Care designation, n (%)
Medical	70 (45.4)
Comfort	35 (22.7)
Resuscitative	22 (14.3)
Unassigned	27 (17.5)
Reason for CCP exit, n (%)
Death	103 (66.9)
Switch to dialysis	6 (3.9)
End of study	45 (29.2)

Note. IQR = interquartile range; CCP = conservative care program.

Among patients who died during the study period, their outcomes are detailed in Table 3. Advanced care planning was more common in these patients. Most patients died in their preferred place of death, whereas about one-quarter of patients died in hospital (Table 4). The mean Dialysis Quality of Dying Apgar was high, and uremia was the commonest cause of death. Death outcomes based on early versus late referral to the multidisciplinary CKD clinic appeared to be similar.

Table 3.

Death Outcomes Following CCP Participation by Early and Late Referral to the Multidisciplinary CKD Clinic.^a

Death outcome	Deceased patients (n = 103, 67% of all patients)	Late referral (n = 36/103, 35% of deceased patients)	Early referral (n = 67/103, 65% of deceased patients)
Advanced care planning initiated before program exit, n (%)b	82 (88.2)	30 (93.8)	52 (85.2)
Known preferred place of death, n (%)	82 (79.6)	31 (86.1)	51 (76.1)
Match between preferred and actual place of death, %c	81.7	83.9	80.4
Death in hospital, n (%)d	27 (26.7)	9 (25.0)	18 (27.7)
Dialysis Quality of Dying Apgar, mean (SD)	8.9 (1.0)	9.0 (1.0)	9.0 (1.1)
Death due to uremia, n (%)e	52 (63.4)	19 (61.3)	33 (64.7)
Bereavement follow-up for family members, n (%)f	70 (71.4)	27 (79.4)	43 (67.2)

Note. CCP = conservative care program; CKD = chronic kidney disease.

^aEarly referral: at >90 days before onset of G5 non-dialysis CKD. Late referral: after onset of or at ≤90 days before onset of G5 non-dialysis CKD.

^bDue to missing data, the proportions for advanced care planning initiation are with respect to the following sample sizes: deceased patients (n = 93), late referral (n = 32), and early referral (n = 61).

^cWhen both actual and preferred places of death were known.

^dDue to missing data, the proportions for death in hospital are with respect to the following sample sizes: deceased patients (n = 101), late referral (n = 36), and early referral (n = 65).

^eDue to missing data, the proportions for death due to uremia are with respect to the following sample sizes: deceased patients (n = 82), late referral (n = 31), and early referral (n = 51).

^fDue to missing data, the proportions for bereavement follow-up are with respect to the following sample sizes: deceased patients (n = 98), late referral (n = 34), and early referral (n = 64).

Table 4.

Place of Death Among Deceased Patients According to Their Preferred Place of Death (n = 103).

Preferred place of death	Place of death (%)
	Home/LTC	Hospice	Hospital	Othera	Unknown
Known (n = 82/103, 79.6% of deceased patients)
Home/LTC (n = 33), n (%)	21 (63.6)	6 (18.2)	6 (18.2)	0	0
Hospice (n = 26), n (%)	0	23 (88.5)	2 (7.7)	1 (3.8)	0
Hospital (n = 6), n (%)	0	0	6 (100)	0	0
Othera (n = 17), n (%)	0	0	0	17 (100)	0
Unknown (n = 21/103, 20.3% of deceased patients), n (%)	0	3 (14.3)	13 (61.9)	3 (14.3)	2 (9.5)

Note. LTC = long-term care facility.

^aOther places of death included, for example, a friend or relative’s home.

Discussion

A review of patients who chose conservative care in this Canadian nephrology center revealed an older cohort with a low comorbidity index. Most participants were followed in the multidisciplinary CKD clinic for at least 90 days prior to G5 non-dialysis CKD onset. Few of them previously received renal replacement therapy or switched treatment modality from conservative care to dialysis. Their CKD laboratory parameters were reasonably well managed. The majority of patients engaged in advanced care planning. Overall, most deaths occurred in the place of choice, and death in hospital was relatively uncommon. The quality of death was high.

Although Canadian dialysis patients are generally also older,¹⁸ patients receiving dialysis have reportedly demonstrated different outcomes compared with patients enrolled in our CCP. Unlike our conservative care cohort, other G5 non-dialysis CKD patients receive delayed or inadequate end-of-life planning and palliative care service availability.^19-22 The patients in our CCP had a lower proportion of in-hospital death compared with dialysis patients (eg, 44.8% in an American study,²³ 66% in an Australian study,²⁴ and 74% in a Canadian study²⁵) and with the general population in Alberta and Canada (57.5% and 63.4%, respectively²⁶). Although we did not measure the frequency of procedures and hospitalization, other studies found that being on dialysis results in more hospitalization.²⁷ Moreover, although we did not look for an association between comorbidity burden and survival, studies have reported that comorbidity increases the risk of death in dialysis²⁸ and nondialysis²⁹ advanced CKD patients. To this effect, it has been suggested that dialysis is not life-prolonging compared with conservative care in comorbid older patients.³⁰

The outcomes of our CCP suggest that it has been successful in terms of death in place of choice and quality of death. Previous studies have reported that provision of conservative care should emphasize patient preference,^31,32 include a multidisciplinary team,^33,34 involve shared decision making,^22,34,35 and align with local practices and community needs.³⁶ These characteristics of health care delivery hence warrant organized CCPs, such as the case presented in this study, which featured success in allowing most patients to die in their preferred place of death. With a high proportion of bereavement follow-up with family members, we were able to address the experience of grief.³⁷ We were also successful in providing a high quality of death to the deceased cohort. Furthermore, as seen elsewhere, our CCP patients commonly featured survival beyond a year after enrollment and seldom switched to dialysis.²⁴

This study featured several limitations. The starting eGFR for patients at the time of CCP enrollment was heterogeneous. For patients who were followed in the CCP prior to G5 non-dialysis CKD onset, we counted CCP enrollment in this study once the eGFR fell below 15 mL/min/1.73 m². However, other patients were referred to the CCP with lower starting eGFRs. As the data were retrieved from a single data source (PARIS software), it was dependent on the completeness of the information entered in the database (inconsistent degree of patient details among online records). This study is also limited by its retrospective design, with inherent biases in patient selection, as well as its generalizability (single-center Canadian study). Still, it represents an important example of comprehensive conservative care in a Canadian setting where such published experiences are scarce.

The results of this study will help guide the development of other CCPs in Canada. These programs are needed not only in centers where advanced kidney disease is managed by nephrologists but also for the large number of such patients managed by their primary care physicians.³⁸ However, obtaining this information by way of chart review was cumbersome, signifying that renal information systems for conservative care are lacking. Information technology infrastructure needs to be developed for renal CCPs that will allow for patient outcomes to be measured on a systematic and prospective basis. Such infrastructure would facilitate research of this population that may be stronger methodologically, including prospective and larger studies (eg, multicentered when such infrastructure is adopted by other centers). It would also allow us to observe trends in patient outcomes over time. In the future, we would like to also collect and report information on the frequency of procedures and hospitalization, the rate of renal function decline, performance status, frailty, and symptoms. This study underlines, furthermore, the need for a validated tool for measuring the quality of death in G5 non-dialysis CKD patients.³⁹

Conclusion

This report of advanced CKD patients enrolled in a CCP in an urban Canadian center is the first one in Canada to describe their demographic and clinical characteristics as well as death outcomes. These patients were not generally very comorbid and chose palliative care without ever having received dialysis. They engaged in advanced care planning and featured favorable outcomes at the time of death. A better understanding of these patients and how they are managed provides insight into potential improvements in health services delivery for this patient population. However, information technology infrastructure is needed for renal CCPs to facilitate the prospective measurement of patient outcomes.