Symptoms with or because of Kidney Failure?

Individuals with CKD experience substantial physical and mental symptoms (1). These symptoms can affect quality of life, activities of daily living, social functioning, and mortality (1,2). Despite the high burden of symptoms in CKD, many nephrology providers demonstrate limited appreciation of their patients’ symptoms (3). Several factors likely contribute to this, including the limited scientific understanding of symptom etiology, unclear ownership of management, and limited treatment strategies with proven effectiveness in rigorous studies (2). Fortunately, the shift toward patient-centered care has reinvigorated the field of symptom research in patients with CKD (4,5).

The pathogenesis of symptoms in patients with CKD remains a critical but poorly understood process. A better understanding of the pathways underlying symptom development in our patients may provide opportunities to develop new treatments or refashion old therapies for new purposes. In this issue of CJASN, Wulczyn et al. (6) provide epidemiologic evidence to further our understanding of symptoms in patients with CKD. The investigators used longitudinal data from the Chronic Renal Insufficiency Cohort (CRIC) to examine the association between one measure of kidney function, eGFR, and symptom severity in over 3600 patients (including approximately 3300 with longitudinal symptom scores). Symptoms were assessed annually using the symptom items from the Kidney Disease Quality of Life (KDQOL) instrument, which captures how bothersome symptoms are on a five-point Likert scale (from “not at all” to “extremely”). Symptom scores were then scaled from zero to 100, with a one-point Likert scale change representing a 25-point symptom score change. The investigators focused on six symptoms attributable to the accumulation of uremic toxins (fatigue, anorexia, pruritis, nausea, paresthesia, and pain) and examined how longitudinal changes in an individual symptom associated with annual measures of eGFR and other time-varying covariates.

The study represents an excellent use of CRIC with its diverse recruitment (Black participants and Hispanic participants comprised 40% and 13% of the cohort, respectively), high prevalence of CKD stages 3a to 4 (approximately 85% of patients) at baseline, and long follow-up (median of 7 years) (6). As expected, the symptom burden in the cohort was high, with a prevalence of over 40% for pain, fatigue, paresthesia, and pruritis. CKD progression was common, with an average decrease in eGFR of 1.34 ml/min per 1.73 m² per year, and over 500 participants progressed to an eGFR <15 ml/min per 1.73 m².

Annual changes in symptoms were modest—generally less than one point (on a 100-point scale) per year (6). When stratifying by baseline symptoms, the annual change in symptom scores remained stable, with the largest changes seemingly driven by regression to the mean (Supplemental Table 4 in ref. 6). Heterogeneity was noted, with one of three participants demonstrating worsening in at least one symptom by an average of five points per year, which likely represents a one-point or greater change in the Likert scale response given the median follow-up. Conversely, over one in four participants demonstrated an improvement in at least one symptom by this same amount. Overall, 75% of patients had stable symptoms or at least as many symptoms improve as worsen during follow-up (Supplemental Table 5 in ref. 6). Examining longitudinal trends in the KDQOL symptom subscale scores (capturing 11 symptoms) resulted in comparable findings.

In the multivariable linear mixed effects model, eGFR had a significant but modest association with longitudinal symptom severity (6). For example, a 5-ml/min per 1.73 m² decrement in eGFR was associated with worsening in fatigue severity of <1.5 points (using the zero- to 100-point scale). Although baseline eGFR appeared to modify the effect of a decrement in eGFR, the association was again modest (e.g., a 5-ml/min/1.73 m² decrement in eGFR worsened fatigue by 1.15 versus 0.4 points when baseline eGFR was 20 versus 50 ml/min per 1.73 m², respectively). In a separate automated backward regression model, the investigators found that multiple factors associated with increasing symptom scores over time, including younger age, unemployed status, lower income, comorbidities including diabetes and vascular diseases, polypharmacy, anemia, and hypoalbuminemia (Supplemental Table 9 in ref. 6). Interestingly, aside from the baseline symptom score, the factor with the strongest association with longitudinal symptom scores was depressive symptoms (Beck’s Depression Index score of >14).

Although this remarkable study has multiple strengths as highlighted above, there are several limitations. First, longitudinal assessments were performed annually. This timing is unlikely to capture acute worsening in symptoms in the weeks immediately preceding dialysis initiation for patients on the precipice of requiring KRT. Second, as the authors acknowledge, eGFR is only one measure of kidney function. Third, patients’ subjective realities are frequently influenced by the contrast of their expectations and lived experience. Because expectations often change over time (i.e., adaptation), patients’ may perceive their symptoms as less severe when worsening occurs gradually. Finally, the well-respected investigators used data available to them from CRIC and focused on six symptoms (of the 11 in the KDQOL symptom subscale) they felt were related to uremic toxins. Advances in CKD symptom research will be accelerated when investigators use standardized instruments that have been rigorously developed; have been found to be reliable, valid, and responsive to change; and have a defined minimal clinically important difference. In addition, using formal approaches to prioritize symptoms on the basis of consensus (either patients’ and/or expert clinicians and researchers) should be encouraged to enhance standardization and reproducibility (2,4,5). Nonetheless, using data collected from CRIC, the investigators have expanded our understanding of longitudinal symptom severity in patients with nondialysis-dependent CKD.

What conclusions should we draw from the relatively modest association of eGFR decline and worsening symptoms? Should symptom research in CKD pivot away from a search for uremic toxins? A recent longitudinal study examining symptoms in patients with kidney failure (80% undergoing maintenance dialysis) paints a more complex picture (7). Three months after a kidney transplant, symptom severity markedly diminished for cramps, pruritis, xeroderma, and fatigue by approximately 12–18 points (7) using the same zero to 100 scale as in Wulczyn et al. (6). Whether this is explained by improvements in uremic toxin levels, reductions in polypharmacy, cessation of dialysis, or improvements in other kidney failure–associated conditions (e.g., psychosocial factors) is unknown. Clearly, a working allograft provides substantial symptom relief to patients with kidney failure receiving dialysis. However, allograft recipient scores on the KDQOL symptom scale only improved from the 50th percentile for patients on dialysis at the time of transplant to the 75th percentile for patients on dialysis at 3 months post-transplant, when symptoms were at their nadir (Table 4 and Supplemental Table 1 in ref. 7). Although transplant-related factors may contribute to symptoms, we should accept that for patients with CKD, uremia is only one of the “thousand natural shocks that flesh is heir to” (8).

Moving forward, we must acknowledge that our understanding of symptom pathogenesis in patients with CKD is rudimentary (5). We should avoid labels that prejudge complex processes and are likely to lead to misnomers and confusing terminology, like uremic symptoms. Such terminology may also imply we are focused on the biochemical nature of symptoms rather than patients’ experiences (9). One hope in searching for uremic toxins that affect symptoms should be developing new insights on activating pathways shared with other contributing factors. Our patients experience food insecurity, housing instability, comorbidities, mood and anxiety disorders, polypharmacy, socioeconomic disadvantage, un- or underemployment, and social isolation. Each of these factors may promote the inflammatory, immune, and/or hormonal disturbances underlying symptom perception (10,11). In addition, we must acknowledge that our patients are dynamic. They may implement coping strategies and behavioral adaptations that allow them to live meaningfully and adjust to their symptoms. We should celebrate their resilience.

Disclosures

K. Abdel-Kader reports serving in an advisory or leadership role for BMC Nephrology and CJASN and reports other interests or relationships with the National Kidney Foundation Education committee.

Funding

K. Abdel-Kader is supported by National Institute of Diabetes and Digestive and Kidney Diseases grant R01DK116957.

Author Contributions

K. Abdel-Kader wrote the original draft and reviewed and edited the manuscript.