Patient-Reported Symptoms and Subsequent Risk of Myocardial Infarction in Chronic Kidney Disease

Abstract

Background and objectives 

Although patient-reported symptoms often precede acute presentations of cardiovascular disease, patients with nondialysis-requiring CKD are less likely to have typical symptoms of atherosclerotic disease when presenting with acute myocardial infarction. However, the associations between typical atherosclerotic symptoms and subsequent risk of myocardial infarction are unknown in ambulatory patients with CKD.

Design, setting, participants, & measurements 

To determine whether typical atherosclerotic symptoms are associated with risk for subsequent myocardial infarction in people with CKD, we examined participants from the Chronic Renal Insufficiency Cohort Study. Chest pain, shortness of breath, and inability to climb stairs were evaluated annually using the Kidney Disease Quality of Life Instrument. Associations between categorical time-updated symptoms and physician-adjudicated incident myocardial infarction were assessed using Cox regression models.

Results 

Among 3910 participants (mean age of 58±11 years; mean eGFR =44±15 ml/min per 1.73 m²), there were 476 incident myocardial infarctions over a median follow-up period of 10.4 years (interquartile range, 5.36–12.6 years). Median time from symptom assessment to incident myocardial infarction was 213 days (interquartile range, 111–333 days). Compared with no symptoms, mild, and moderate or worse, symptoms of chest pain (hazard ratio, 1.30; 95% confidence interval, 1.01 to 1.67; and hazard ratio, 1.70; 95% confidence interval, 1.27 to 2.27, respectively) and shortness of breath (hazard ratio, 1.37; 95% confidence interval, 1.10 to 1.70; and hazard ratio, 1.33; 95% confidence interval, 1.05 to 1.69, respectively) were significantly associated with greater risks for subsequent myocardial infarction. Participants reporting mild and severe limitations in climbing stairs (versus no limitation) had significantly higher adjusted risk of myocardial infarction (hazard ratio, 1.44; 95% confidence interval, 1.10 to 1.89; and hazard ratio, 1.89; 95% confidence interval, 1.44 to 2.49, respectively).

Conclusions 

In a large ambulatory cohort of adults with CKD, symptoms of atherosclerotic cardiovascular disease were strongly associated with a higher risk for subsequent myocardial infarction.

Podcast 

This article contains a podcast at https://www.asn-online.org/media/podcast/CJASN/2022_03_17_CJN12080921.mp3

Visual Abstract

Introduction

Patient-reported symptoms are key to diagnosing atherosclerotic cardiovascular disease and often precede the acute presentation of myocardial infarction. In the general population, such symptoms include chest pain, shortness of breath, and inability to climb stairs (^123–4). These symptoms may raise clinical suspicion for atherosclerotic cardiovascular disease, prompting additional workup (^56–7). Patients with CKD have a higher risk of atherosclerotic cardiovascular disease and myocardial infarction versus those without CKD and are more likely to experience adverse outcomes from these events; indeed, patients with CKD are more likely to die of cardiovascular disease than they are to progress to kidney failure (^{89101112–13}). However, previous work has shown that similar to patients with diabetes, patients with CKD are less likely to report chest pain and other typical symptoms of myocardial infarction during acute presentations of myocardial infarction (^141516–17).

Prior studies in patients with CKD have focused on symptoms experienced by patients during presentation for acute myocardial infarction rather than associations between these symptoms and future cardiac events. These data may guide clinicians on the utility of routine symptom assessment in ambulatory patients. Specifically, it is unknown whether atherosclerotic symptoms of chest pain, shortness of breath, and difficulty climbing stairs are associated with higher risk for subsequent myocardial infarction in patients with nondialysis-requiring CKD, as they may lack specificity in patients with CKD or diabetes (¹,^{1819202122–23}). To investigate this knowledge gap, we examined whether typical atherosclerotic symptoms were associated with higher risk for subsequent myocardial infarction in people with nondialysis-requiring CKD. We hypothesized that these symptoms would not be associated with future myocardial infarction in patients with CKD.

Materials and Methods

Study Population

We used National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) public repository data for the Chronic Renal Insufficiency Cohort (CRIC) Study, which enrolled 3939 participants between June 2003 and August 2008 at seven clinical centers across the United States (Ann Arbor/Detroit, MI; Baltimore, MD; Chicago, IL; Cleveland, OH; New Orleans, LA; Philadelphia, PA; and Oakland, CA) (²⁴,²⁵). All data and materials have been made publicly available at the NIDDK Central Repository and can be accessed through request at https://repository.niddk.nih.gov/studies/cric/. Details on study design, baseline characteristics, and inclusion and exclusion criteria have been published previously (²⁴). All participants had annual in-person study visits, including interviews, physical examination, laboratory assessments, and cardiovascular testing. All participants provided written informed consent; the study protocol was approved by institutional review boards at each participating site. For these analyses, we excluded 29 participants without follow-up, leaving 3910 eligible participants. No participants were excluded for missing symptom data.

Symptom Assessment by the Kidney Disease Quality of Life Instrument

The Kidney Disease Quality of Life (KDQOL) Instrument is a validated 36-question form assessing symptoms, functional capacity, and overall quality of life in the prior 4 weeks in patients with CKD (²⁶). The KDQOL questionnaire was administered to CRIC participants at each annual visit. Participants were asked the degree to which chest pain and shortness of breath had bothered them over the past 4 weeks. Possible responses were “not at all,” “somewhat,” “moderately,” “very much,” and “extremely” bothered. As few participants answered “very much” or “extremely” bothered, we combined these responses with “moderately bothered.” We renamed “somewhat bothered” as “mild bother.” Our final categories for chest pain and shortness of breath were “no bother,” “mild bother,” and “moderate or worse bother.” For limitations in climbing stairs, possible responses were “yes, limited a lot,” “yes, limited a little,” and “no, not limited at all.” For consistency, we categorized these as “severe limitation,” “mild limitation,” and “no limitation,” respectively. In case of missing symptoms (numbers of visits were 31 for chest pain, 45 for shortness of breath, and 76 for inability to climb stairs), the visit preceding the missing visit was analyzed instead.

Determination of Myocardial Infarction

Our primary outcome was hospitalization for the first fatal or nonfatal myocardial infarction during the follow-up period. Every 6 months, participants were asked if they had been hospitalized for a myocardial infarction; medical records from qualifying encounters were queried. Myocardial infarction was defined as a typical rise and either slow or rapid fall in cardiac enzymes along with typical symptoms of myocardial ischemia or electrocardiogram changes compatible with this diagnosis. In the absence of elevated cardiac enzymes, myocardial infarction was defined by typical ischemic symptoms with ST-segment elevations on electrocardiogram or new fixed perfusion abnormalities on nuclear perfusion imaging versus prior examination with a new corresponding wall motion abnormality on functional evaluation, such as echocardiography compared with prior examination. Patient-reported hospitalizations triggered retrieval of medical records as above, which were reviewed and adjudicated by two physicians (²⁴).

Covariates

Participants provided information on their sociodemographic characteristics, medical history, medication usage, and lifestyle behaviors at baseline. Race and ethnicity were by self-report and categorized as non-Hispanic White, non-Hispanic Black, Hispanic, and other. Baseline coronary artery disease, heart failure, and stroke were determined by self-report. BP and body mass index (BMI) were assessed using standard protocols (²⁷). Diabetes mellitus was defined as a fasting glucose >126 mg/dl, a nonfasting glucose >200 mg/dl, or use of insulin or other antidiabetic agents. Alcohol use was dichotomized as none versus any in the past 12 months. Tobacco use was dichotomized as active use versus not at time of cohort entry.

Serum creatinine was measured by an Ortho Vitros 950 (Raritan, NJ) at the CRIC Central Laboratory using a standardized enzymatic method (²⁸). eGFR was calculated via the Chronic Kidney Disease Epidemiology Collaboration equation (²⁹). Additional measurements included hemoglobin and 24-hour proteinuria.

The following covariates were obtained at each annual visit: age, eGFR, systolic and diastolic BP, BMI, smoking status, alcohol use status, hemoglobin, comorbidity status (including diabetes, heart failure, atrial fibrillation, and chronic obstructive pulmonary disease [COPD]), and medication use, including angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, aspirin, statins, β-blockers, and diuretics.

Statistical Analyses

Study variables were described by prevalence of comorbid conditions or medication use; continuous variables were described by mean (SD) values unless there was a significantly skewed distribution, in which case they were described using median (interquartile range [IQR]) values.

Our primary outcome was time to first myocardial infarction during follow-up, censoring at time of death, loss to follow-up, withdrawal of consent, or end of administrative follow-up on December 31, 2017, whichever came first. We calculated incidence rates for myocardial infarction by dividing the total number of observed events by the sum of follow-up time. We described incidence rates by symptom severity (along with differences between each incidence rate and the incidence rate for myocardial infarction among those with no reported symptoms) and across categories of eGFR (>60, 45–59, 30–44, and <30 ml/min per 1.73 m²). We calculated 95% confidence intervals (95% CIs) for each incidence rate using a nonparametric bootstrap approach with 2000 replicates (³⁰).

We modeled each symptom as a categorical predictor with no symptoms or limitation as the referent group. Our first category was “mild bother” from chest pain or shortness of breath or “mild limitation” in climbing stairs; the second was “moderate or worse bother” for chest pain or shortness of breath or “severe limitation” in climbing stairs. We utilized a time-updated approach; for each myocardial infarction, we considered the symptoms reported at the visit preceding the event. Cox regression models using the counting process formulation of Andersen and Gill (³¹) were used to estimate the association of each categorical symptom with myocardial infarction. We calculated a P value for trend across categories of symptom severity for each model. To understand the separate effects of various categories of covariates, we fit a series of nested Cox models with sequential adjustment for time-updated potential confounders. Our base model was unadjusted. Model 1 was adjusted for age, sex, race, and ethnicity. Model 2 was adjusted for the factors in model 1 plus prior myocardial infarction or revascularization, stroke, peripheral vascular disease, diabetes, systolic and diastolic BP, atrial fibrillation, heart failure, BMI, COPD, eGFR, proteinuria, and smoking status. Model 3 was adjusted for the components of model 2 plus time-updated medication use (coronary vasodilators, calcium channel blockers, and β-blockers). Covariates were chosen by biologic plausibility for potential confounding effects.

In secondary analyses, we tested for multiplicative interaction with each of the following terms: prior myocardial infarction or revascularization, eGFR (modeled continuously and categorically), sex, diabetes, and proteinuria.

Sensitivity Analyses

We performed three sensitivity analyses. First, we tested for additive interactions between our symptoms of interest by calculating the relative excess risk due to interaction term (³²). We generated 95% CIs for each term using a nonparametric bootstrap approach with 2000 replicates (³⁰).

Second, we stratified our findings by diabetes status because patients with diabetes are less likely to report typical symptoms of myocardial infarction (¹⁶,¹⁷). We described the incidence rates of myocardial infarction by symptom severity stratified by diabetes status and tested for associations between each symptom and myocardial infarction stratified by diabetes status similar to the primary analysis.

In the third sensitivity analysis, to understand how patient-reported symptoms changed over time, we graphically described changes in each symptom from the preceding visit. As changes in symptom severity may be clinically applicable, we assessed associations between each symptom and myocardial infarction with additional adjustment for changes in symptoms from the preceding annual study visit. This duration was chosen given differing length of follow-up time by each patient and because a 1-year change in symptoms seemed most clinically applicable.

Missing data for medical history and medication usage, where observed, were addressed by using the preceding response. Missing continuous covariates were multiply imputed using chained equations via the mice package in R (³³). Missingness was low overall (at baseline, 2% for COPD, 5% for proteinuria, and <1% for all other potential covariates). The most common missing covariate in time-updated data was proteinuria (49% missing), followed by eGFR (14% missing); patients with diabetes appeared more likely to have missing covariates.

All analyses were performed using R 4.0.2 (R Foundation for Computing, Vienna, Austria).

Results

Characteristics of the Study Population

Among 3910 participants, the mean (±SD) age was 58 (±11) years. In total, 1761 (45%) were women; 1633 (42%) were White, and 1639 (42%) were Black. The mean eGFR was 44 (±15) ml/min per 1.73 m²; the median proteinuria was 0.18 (IQR, 0.07–0.91) g/d. Overall, 855 participants (22%) reported prior myocardial infarction or revascularization, 1895 (48%) had diabetes, and 513 (13%) used tobacco at enrollment. In total, 1665 (43%) participants used aspirin and 2139 (55%) used statins at enrollment. At baseline, 3112 participants (80%) noted no bother from chest pain; 2140 (55%) participants noted no bother from shortness of breath. Limitations in climbing stairs were common; 1348 (35%) participants reported no limitations, although 1328 (34%) and 1171 (30%) reported mild and severe limitations in climbing stairs, respectively (Table 1).

Table 1. - Characteristics of study participants with CKD overall at their initial study visit (n=3910)

Variable	Value
Age, yr, mean (SD)	58 (11)
Women, N (%)	1761 (45)
Race and ethnicity, N (%)
Non-Hispanic White	1633 (42)
Non-Hispanic Black	1630 (42)
Hispanic	494 (13)
Other	153 (4)
eGFR, ml/min per 1.73 m2, mean (SD)	44 (15)
eGFR≥60 ml/min per 1.73 m2, N (%)	586 (15)
eGFR=45–59 ml/min per 1.73 m2, N (%)	1182 (30)
eGFR=30–44 ml/min per 1.73 m2, N (%)	1418 (36)
eGFR=15–30 ml/min per 1.73 m2, N (%)	720 (18)
eGFR<15 ml/min per 1.73 m2, N (%)	4 (0.1)
24-h urine protein, g/d, median (IQR)	0.18 (0.07–0.91)
24-h urine sodium, mg/d, median (IQR)	3447 (2466–4620)
Diabetes mellitus, N (%)	1895 (48)
History of heart failure, N (%)	378 (10)
History of atrial fibrillation, N (%)	662 (17)
Prior myocardial infarction or revascularization, N (%)	855 (22)
Prior stroke, N (%)	390 (10)
Peripheral vascular disease, N (%)	260 (7)
COPD, N (%)	124 (3)
Systolic BP, mm Hg, mean (SD)	128 (22)
Diastolic BP, mm Hg, mean (SD)	71 (13)
Body mass index, kg/m2, mean (SD)	32.1 (7.80)
Current smoker, N (%)	513 (13)
Alcohol use, N (%)	2457 (63)
Hemoglobin, g/dl, mean (SD)	12.6 (1.78)
LDL cholesterol, mg/dl, mean (SD)	103 (36)
HDL cholesterol, mg/dl, mean (SD)	48 (15)
Angiotensin-converting enzyme inhibitor/angiotensin receptor blocker, N (%)	2670 (68)
Aspirin, N (%)	1665 (43)
Statins, N (%)	2139 (55)
β-Blockers, N (%)	1915 (49)
Diuretics, N (%)	2314 (59)
Baseline symptom burden, N (%)
Chest pain, N (%)
No bother	3112 (80)
Mild bother	515 (13)
Moderate or worse bother	265 (7)
Shortness of breath, N (%)
No bother	2140 (55)
Mild bother	979 (25)
Moderate or worse bother	772 (20)
Limitations in climbing stairs, N (%)
No limitations	1348 (35)
Mild limitations	1328 (34)
Severe limitations	1171 (30)

IQR, interquartile range; COPD, chronic obstructive pulmonary disease.

Incidence Rates of Myocardial Infarction by Symptom Severity and Estimated Glomerular Filtration Rate

Among 3910 participants, 476 myocardial infarctions were observed over a median follow-up time of 10.4 (IQR, 5.36–12.6) years. The median time from symptom assessment to myocardial infarction was 213 (IQR, 111–333) days.

Chest Pain.

Among participants reporting “no bother” from chest pain in the 4 weeks preceding the study visit, the incidence rate for myocardial infarction was 1.19 (95% CI, 1.06 to 1.31) per 100 patient-years (Table 2). The incidence of myocardial infarction was higher with greater chest pain regardless of the level of eGFR (Figure 1).

Table 2. - Number and incidence rate of myocardial infarction by symptom severity

Symptom and Severity	Number of Myocardial Infarctions	Incidence Rate (95% Confidence Interval) per 100 Patient-Years	Difference in Incidence Rate (95% Confidence Interval) from Referent Group
Chest pain	476
No bother	338	1.19 (1.06 to 1.31)	Referent
Mild bother	80	1.86 (1.46 to 2.25)	0.67 (0.25 to 1.09) a
Moderate or worse bother	57	2.49 (1.86 to 3.12)	1.30 (0.66 to 1.94) a
Shortness of breath
No bother	206	1.03 (0.89 to 1.16)	Referent
Mild bother	145	1.72 (1.44 to 1.99)	0.69 (0.38 to 1.00) a
Moderate or worse bother	124	1.90 (1.56 to 2.24)	0.87 (0.50 to 1.24) a
Inability to climb stairs
No limitation	84	0.68 (0.53 to 0.83)	Referent
Mild limitation	167	1.37 (1.16 to 1.58)	0.68 (0.42 to 0.95) a
Severe limitation	224	2.13 (1.86 to 2.41)	1.45 (1.13 to 1.77) a

^aIndicates a statistically significant difference.

Figure 1.: Incidence rate of myocardial infarction by chest pain severity and eGFR category (error bars represent the upper bounds of the 95% confidence interval).

Shortness of Breath.

Among participants reporting “no bother” from shortness of breath, the incidence rate for myocardial infarction was 1.03 (95% CI, 0.89 to 1.16) per 100 patient-years (Table 2). The incidence of myocardial infarction was higher with increasing shortness of breath regardless of the level of eGFR (Figure 2).

Figure 2.: Incidence rate of myocardial infarction by shortness of breath severity and eGFR category (error bars represent the upper bounds of the 95% confidence interval).

Inability to Climb Stairs.

Participants with no limitations in climbing stairs had an incidence rate of myocardial infarction of 0.68 (95% CI, 0.53 to 0.83) per 100 patient-years (Table 2). Regardless of eGFR level, the incidence of myocardial infarction was higher with increasing limitations in climbing stairs (Figure 3).

Figure 3.: Incidence rate of myocardial infarction by the limitation in climbing stairs and eGFR category (error bars represent the upper bounds of the 95% confidence interval).

Association of Time-Updated Symptoms with Myocardial Infarction

Chest Pain.

In unadjusted models, “mild bother” and “moderate or worse bother” from chest pain were associated with 1.56-fold (95% CI, 1.22 to 1.99) and 2.09-fold (95% CI, 1.58 to 2.76) higher risk for myocardial infarction, respectively. After adjusting for potential confounders, “mild bother” and “moderate or worse bother” remained significantly associated with myocardial infarction (hazard ratio [HR], 1.30; 95% CI, 1.01 to 1.67; and HR, 1.70; 95% CI, 1.27 to 2.27, respectively) (Table 3).

Table 3. - Associations between categorical time-updated symptoms of atherosclerosis and myocardial infarction (n=3909)

Symptom and Severity	Hazard Ratio (95% Confidence Interval)
	Unadjusted	Model 1	Model 2	Model 3
Chest Pain
No bother	1.0 (referent)	1.0 (referent)	1.0 (referent)	1.0 (referent)
Mild bother	1.56 (1.22 to 1.99) a	1.63 (1.28 to 2.08) a	1.36 (1.06 to 1.74) a	1.30 (1.01 to 1.67) a
Moderate or worse bother	2.09 (1.58 to 2.76) a	2.22 (1.67 to 2.94) a	1.80 (1.35 to 2.40) a	1.70 (1.27 to 2.27) a
P value for trend across symptom severity	<0.001	<0.001	<0.001	<0.001
Shortness of breath
No bother	1.0 (referent)	1.0 (referent)	1.0 (referent)	1.0 (referent)
Mild bother	1.67 (1.35 to 2.06) a	1.62 (1.31 to 2.00) a	1.38 (1.11 to 1.71) a	1.37 (1.10 to 1.70) a
Moderate or worse bother	1.85 (1.48 to 2.31) a	1.80 (1.44 to 2.26) a	1.37 (1.08 to 1.74) a	1.33 (1.05 to 1.69) a
P value for trend across symptom severity	<0.001	<0.001	0.004	0.008
Inability to climb stairs
No limitation	1.0 (referent)	1.0 (referent)	1.0 (referent)	1.0 (referent)
Mild limitation	1.98 (1.53 to 2.57) a	1.86 (1.43 to 2.43) a	1.47 (1.12 to 1.92) a	1.44 (1.10 to 1.89) a
Severe limitation	3.10 (2.41 to 3.98) a	3.00 (2.32 to 3.87) a	1.95 (1.48 to 2.56) a	1.89 (1.44 to 2.49) a
P value for trend across symptom severity	<0.001	<0.001	<0.001	<0.001

Model 1 is adjusted for sex, self-reported race and ethnicity, and time-updated age. Model 2 is adjusted for the components of model 1 plus time-updated prior myocardial infarction or revascularization, prior stroke, diabetes, systolic and diastolic BP, heart failure, atrial fibrillation, body mass index, chronic obstructive pulmonary disease, eGFR, current smoking status, and log-transformed proteinuria. Model 3 is adjusted for the components of model 2 plus time-updated use of coronary vasodilators, calcium channel blockers, and β-blockers.

^aIndicates a statistically significant confidence interval.

Shortness of Breath.

In unadjusted models, “mild bother” and “moderate or worse bother” from shortness of breath were associated with myocardial infarction versus “no bother” (HR, 1.67; 95% CI, 1.35 to 2.06; and HR, 1.86; 95% CI, 1.48 to 2.31, respectively). After adjustment, “mild bother” and “moderate or worse bother” remained significantly associated with myocardial infarction (HR, 1.37; 95% CI, 1.10 to 1.70; and HR, 1.33; 95% CI, 1.05 to 1.69, respectively) (Table 3).

Inability to Climb Stairs.

In unadjusted models, “mild limitation” and “severe limitation” in climbing stairs were significantly associated with myocardial infarction versus “no limitation” (HR, 1.98; 95% CI, 1.53 to 2.57; and HR, 3.10; 95% CI, 2.41 to 3.98, respectively). After adjustment, “mild limitations” and “severe limitations” in climbing stairs remained associated with myocardial infarction (HR, 1.44; 95% CI, 1.10 to 1.89; and HR, 1.89; 95% CI, 1.44 to 2.49, respectively) (Table 3).

Secondary Analyses

Interactions by the presence of prior myocardial infarction or revascularization, eGFR (continuous and categorical), proteinuria, and sex were not statistically significant for the associations of time-updated chest pain, shortness of breath, and inability to climb stairs and risk of subsequent myocardial infarction (all P>0.05) (Supplemental Table 1).

Sensitivity Analyses: Additive Interactions

The relative excess risk due to interaction for chest pain and shortness of breath was −0.00 (95% CI, −0.84 to 0.83); for chest pain and inability to climb stairs, it was −0.18 (95% CI, −1.64 to 1.29), and for shortness of breath and inability to climb stairs, it was −0.80 (95% CI, −2.05 to 0.44) (Supplemental Table 2), suggesting no additive interaction.

Sensitivity Analyses: Stratified by Diabetes Status

In analyses stratified by diabetes status, chest pain was associated with higher incidence of myocardial infarction irrespective of diabetes status (Supplemental Figure 1, Supplemental Table 3). In those with diabetes, only mild bother from shortness of breath was significantly associated with myocardial infarction versus significant findings in those without diabetes (Supplemental Figure 2, Supplemental Table 3). Mild and severe limitations in climbing stairs were significantly associated with myocardial infarction irrespective of diabetes status (Supplemental Figure 3, Supplemental Table 3).

Sensitivity Analyses: Changes in Symptom Severity

Most participants noted no changes in their symptoms at subsequent annual visits (Supplemental Figures 4–6). Adjusting for changes in symptoms from the visit preceding each symptom assessment did not significantly alter the results of our primary analyses (Supplemental Table 4).

Discussion

In a large ambulatory cohort with CKD, time-updated symptoms of chest pain, shortness of breath, and inability to climb stairs were associated with significantly higher risk of subsequent myocardial infarction after adjusting for numerous possible confounders, including eGFR and proteinuria. These results highlight the possible importance of routine symptom assessment for early manifestations of atherosclerotic disease in ambulatory patients with CKD.

Chest pain and shortness of breath are common in patients with CKD, even without clinically known atherosclerotic disease (^343536–37). In this study, patients reporting mild or worse bother from chest pain or shortness of breath were at higher risk for myocardial infarction versus those reporting no bother from these symptoms. Previous work primarily focused on symptoms experienced by patients with kidney disease when presenting with acute myocardial infarction rather than the presence of symptoms prior to myocardial infarction. For instance, Go et al. (³⁸) demonstrated that patients with CKD are more likely to present with acute myocardial infarction versus angina as their initial presentation of atherosclerotic cardiovascular disease. Among 29,319 patients with stages 4 and 5 CKD in the National Registry of Myocardial Infarction, 40% of patients with CKD reported chest pain when presenting for myocardial infarction versus 62% of patients without CKD (¹⁴). Furthermore, participants with stage 3a CKD or worse (eGFR <60 ml/min per 1.73 m²) were less likely than participants without CKD to report chest pain in the context of myocardial infarction (adjusted odds ratio, 0.48; 95% CI, 0.41 to 0.57) but more likely to report shortness of breath (adjusted odds ratio, 1.73; 95% CI, 1.40 to 2.15) (¹⁵). Although prior studies were retrospective and cross-sectional, we utilized a well-characterized ambulatory CKD cohort to examine the prospective associations of atherosclerotic symptoms with subsequent risk of myocardial infarction.

Inability to climb stairs is a marker of exercise capacity, which has been associated with myocardial infarction in the general population (¹⁸,²⁰). In one study of 9191 participants, reduced exercise capacity was associated with higher risk of myocardial infarction (HR, 2.36; 95% CI, 1.55 to 3.60) (³⁹). Among 1249 participants with CKD in the Cardiovascular Health Study, physical inactivity was associated with cardiovascular mortality (HR, 1.57; 95% CI, 1.22 to 2.02) (⁴⁰). However, the mean eGFR of those participants was 50 ml/min per 1.73 m², potentially limiting generalizability to more advanced CKD (⁴⁰,⁴¹). Low physical activity was also associated with all-cause and cardiovascular mortality in a study of 540 kidney transplant recipients (⁴²). By contrast, our study examined myocardial infarction as an outcome and utilized patient-reported symptoms to assess physical activity (not formal exercise testing), which may be a pragmatic and low-cost approach for incorporation into clinical practice.

Patients with diabetes are known to demonstrate atypical symptoms of myocardial infarction compared with the typical chest pain (²²,²³). In a sensitivity analysis in this population of patients with CKD, these symptoms were all strongly associated with myocardial infarction in participants without diabetes. In patients with CKD and diabetes, chest pain and inability to climb stairs remained associated with myocardial infarction; however, “moderate or worse bother” from shortness of breath was not significantly associated with myocardial infarction in persons with diabetes. These findings conflict with some seen in the general population, which, for example, have found that patients with diabetes are less likely to report chest pain (^4344–45). Further studies are needed to understand the mechanisms by which CKD and diabetes interact to affect atherosclerotic disease presentation.

Patients with CKD experience an overall high symptom burden, which has been associated with lower quality of life (^343536–37). Previous work suggests that improved recognition of patient-reported symptoms may improve communication with patients and outcomes (³⁷,⁴⁶,⁴⁷). Understanding the burden of symptoms experienced by patients with CKD might lead to interventions focused on improving quality of life (⁴⁸,⁴⁹). It is also possible that patient-reported outcomes themselves may serve as outcomes for therapies in clinical trials involving patients with CKD.

Patient-reported atherosclerotic symptoms and their potential use as warning signs of cardiovascular disease prior to the acute presentation are not well described in people with CKD. Understanding the associations between patient-reported atherosclerotic symptoms and myocardial infarction in patients with CKD may lead to improved risk stratification, rates of diagnosis, and primary prevention for cardiovascular disease in this high-risk population; however, further research on primary prevention strategies are needed. For instance, the International Study of Comparative Health Effectiveness with Medical and Invasive Approaches-Chronic Kidney Disease (ISCHEMIA-CKD) trial recently demonstrated that among patients with advanced CKD and stable coronary disease with moderate to advanced ischemia, an initial invasive strategy was not associated with a lower rate of death or myocardial infarction compared with an initial conservative strategy (⁵⁰). Our findings may additionally underscore the need for optimal medical therapy; for example, only half of our population was on statins. More aggressive treatment may be beneficial in patients with CKD and symptoms of cardiovascular disease.

This study has several strengths. It was conducted in a large cohort with annual symptom assessment, allowing us to examine the associations between time-updated symptoms and future myocardial infarction. We controlled for multiple possible confounders, and the outcome of myocardial infarction was ascertained by rigorous physician adjudication. However, our approach does have several notable limitations. First, the number of participants reporting severe symptoms was low, limiting our ability to study more discrete categories of symptoms. Second, not every participant had annual checks of symptoms (however, the overall missingness of symptoms was low). Third, the KDQOL Instrument is a general quality-of-life index in CKD and was not specifically developed to assess atherosclerotic symptoms. Furthermore, shortness of breath and inability to climb stairs can be nonspecific for atherosclerotic cardiovascular disease in the CKD population (²¹,²³). All comorbidities were self-reported, and it is possible that certain comorbidities (including COPD) were underascertained as a result; we further lack data on the severity of COPD, limiting our ability to adjust for this diagnosis. Similarly, smoking status was dichotomized as “never or previous use” versus “active,” and we were unable to adjust for smoking quantity or prior smoking history. Data regarding symptom assessments for patients at the time of presentation with acute myocardial infarction were not readily available. Finally, the cohort was composed of research volunteers who were closely followed in the clinic, potentially limiting generalizability to other CKD populations.

In conclusion, our study demonstrated significant associations between time-updated typical atherosclerotic symptoms and subsequent myocardial infarction in a large prospective ambulatory cohort of participants with CKD. These findings suggest that routine symptom assessment might identify important risk factors for future myocardial infarction in patients with CKD. Further studies are needed to investigate symptom assessment as a predictive tool for atherosclerotic cardiovascular disease in patients with CKD.

Disclosures

N. Bansal reports serving as an associate editor for Kidney360 and on the editorial board for CJASN. L.R. Zelnick reports consultancy agreements with the Veterans Medical Research Foundation and serving as a statistical editor for CJASN. All remaining authors have nothing to disclose.

Funding

This study was supported by federal support from National Institutes of Health grants R01 DK103612 (to N. Bansal) and T32 DK007467 (to B. Lidgard). Additional support was provided by an unrestricted fund from the Northwest Kidney Centers.