Daily Step Counts Are Associated with Hospitalization Risk in Pulmonary Arterial Hypertension

To the Editor:

Patients with pulmonary arterial hypertension (PAH) often have decreased exercise capacity, which is consistently associated with adverse outcomes and reduced right ventricular (RV) function. The 6-minute-walk distance (6MWD) is associated with clinical outcomes in PAH but has important limitations as a clinical tool and trial endpoint (1, 2). In contrast, little is known about the relationship between uncoached daily activity (e.g., step counts) and clinical outcomes in PAH (3). Step counts and 6MWD correlate only moderately, suggesting the two measures capture different, nonoverlapping information about activity. We hypothesized that decreased step counts would be associated with adverse PAH outcomes, including decline in RV function and increased risk of hospitalization. Some results were presented at the American Thoracic Society 2021 Scientific Sessions (4).

Methods

We recruited adult patients with idiopathic, heritable, or associated PAH between October 2017 and January 2020 and followed them until March 15, 2021. Functional class I–III patients on stable PAH therapy for 3 months were included. Participants wore a Fitbit Charge 3 device for a 2-week run-in period of a behavioral intervention trial. Average step counts for this study were determined by taking the mean daily step count for Week 2 of the run-in. Week 1 steps were discarded to account for a possible observer effect. 6MWD and echocardiography were performed at the conclusion of the activity monitoring period. Subsequent 6MWD and RV function measurements were obtained as part of routine clinical care. To examine how baseline step counts and 6MWD associated with outcomes during follow-up, we used ordinary least regressions for continuous outcomes and logistic regression for binary outcomes. We performed time-to-event analyses for the first hospitalization and medication escalation during follow-up. All regressions included baseline step counts (or 6MWD) as the main predictor and were adjusted for age and sex. The ordinary least and logistic regressions were also adjusted for baseline outcome level and the time elapsed between baseline and follow-up measures. To improve interpretability of the results, hazard ratios, odds ratios, and estimated differences reflect a comparison between patients at the 75th and 25th percentiles of baseline step counts (or 6WMD). Medication escalation was defined as either an unplanned increase in the dosage or addition of a new PAH-targeted therapy.

Results

Forty-one participants completed the initial monitoring period. Thirty-six were female (85%) with a median age of 46.6 years (interquartile range [IQR], 40.0–56.7). Most were World Health Organization (WHO) functional class II (63%). The median baseline daily step count was 4,656 steps (IQR, 3,649–6,256). Baseline 6MWD, fractional area change, and tricuspid annular plane systolic excursion were 427 m (IQR, 360–480), 29% (IQR, 20–33), and 2.1 cm (IQR, 1.9–2.3), respectively. Median follow-up from baseline to last medical contact was 2.2 years (IQR, 1.6–3.0) for a total of 92 person-years (PY) of follow-up. Follow-up echocardiograms were performed a median of 1.4 years after baseline testing (IQR, 1.0–2.1). In total, 11 (27%) of participants were hospitalized over the course of the study and 1 (2%) died.

We observed a significant negative association between greater baseline daily step counts and the risk of hospitalization (hazard ratio [HR], 0.26; 95% confidence interval [CI], 0.08–0.89; P = 0.03; Table 1). The median step count for those hospitalized was 3,899 (IQR, 2,425–4,463) compared with 5,367 (IQR, 3,700–6,548) in those not hospitalized (Figure 1; P = 0.03). Only 1 patient out of 15 (3.6/100 PY) who averaged greater than 5,500 steps was hospitalized, whereas 10 out of 26 (27.7/100 PY) who averaged ⩽5,500 steps was hospitalized (P = 0.04). In contrast, baseline 6MWD was not associated with risk of future hospitalization in our cohort (HR, 1.93; 95% CI, 0.61–6.08; P = 0.26). Higher daily step counts (odds ratio, 0.38; 95% CI, 0.16–0.92; P = 0.03) and 6MWD (odds ratio, 0.39; 95% CI, 0.16–0.96; P = 0.04) were both associated with lower odds of having a worsening functional class over the follow-up interval. We did not observe any association between baseline daily step counts or 6MWD with fractional area change (P = 0.51 and 0.45, respectively). There was a numerical but nonsignificant trend toward worsening tricuspid annular plane systolic excursion with lower baseline step counts (P = 0.12). The HR of requiring an escalation of PAH medications with greater baseline daily step counts was 0.68 but did not meet statistical significance (P = 0.38). Baseline 6MWD was not associated with any of the other measured parameters over the follow-up interval. Results were similar (although not statistically significant) in the subgroup with idiopathic and heritable PAH (data not shown).

Table 1.

Association between Baseline Daily Step Counts and Baseline 6MWD with PAH Metrics

Outcome Metrics at Follow-Up*	Baseline Daily Step Count			Baseline 6MWD
	Estimated Difference at Follow-Up (95% CI)†	HR or OR (95% CI)‡	P Value§	Estimated Difference at Follow-Up (95% CI)†	HR or OR (95% CI)‡	P Value§
6MWD (n = 39)	15.5 (−6.38 to 37.3)	—	0.159	—	—	—
TAPSE (n = 27)	0.18 (−0.05 to 0.40)	—	0.118	0.14 (−0.09 to 0.37)	—	0.211
FAC (n = 18)	3.05 (−6.67 to 12.77)	—	0.504	2.20 (−3.95 to 8.35)	—	0.447
Functional class (n = 41)	—	0.38 (0.16 to 0.92)	0.031	—	0.39 (0.16 to 0.96)	0.041
Med escalationǁ (n = 40)	—	0.68 (0.29 to 1.60)	0.382	—	0.81 (0.24 to 2.72)	0.736
Hospitalization (n = 41)	—	0.26 (0.08 to 0.89)	0.032	—	1.93 (0.61 to 6.08)	0.262

Definition of abbreviations: 6MWD = 6-minute-walk distance; CI = confidence interval; FAC = fractional area change; HR = hazard ratio; OR = odds ratio; PAH = pulmonary arterial hypertension; TAPSE = tricuspid annular plane systolic excursion.

^*Adjusting for age, sex, baseline values, and time between baseline and follow-up measure.

^† Model-estimated difference comparing 6,526 versus 3,649 daily steps (i.e., the 75th vs. 25th percentiles).

^‡ Comparing 6MWD of 480 versus 360 meters (i.e., the 75th vs. 25th percentiles).

^§ Values for 6MWD, TAPSE, and FAC were calculated using ordinary least regression models. The model for World Health Organization functional class was a logistic regression model; therefore, data represent OR rather than HR. PAH medication escalation and hospitalization values were based on survival analyses; consequently, data represent HR.

^ǁ Defined as unplanned increase in dosage or addition of a PAH-targeted therapy.

Figure 1.

Daily step counts in hospitalized or nonhospitalized patients with PAH. Patients were divided based on whether they were hospitalized (n = 11) or not hospitalized (n = 30) during the median 1.54-year follow-up interval. Box plots were constructed using the median and interquartile ranges for each cohort. △ denotes the mean. Statistical analysis was performed using a Mann-Whitney (Wilcoxon) rank-sum test. PAH = pulmonary arterial hypertension. The asterisk refers to a single data point that is outside of the y-axis range.

Discussion

We found that lower baseline daily step counts measured using an inexpensive, commercially available activity monitor were associated with increased risk of hospitalization and worsening WHO functional class over a median 2.2-year follow-up interval. Although several smaller studies have shown an association between daily activity and quality of life indices, functional class, and 6MWD (5–7), this is the largest study that demonstrates an association between daily activity and hospitalizations in adult patients with PAH.

In contrast to step counts, baseline 6MWD was not predictive of future hospitalizations in this study. Disadvantages of the 6MWD highlighted by others in the PAH field include potential ceiling and floor effects, as well as variable prognostic value in response to therapy (1). Our results are not consistent with prior studies that demonstrated significant associations between 6MWD and hospitalization, such as the Registry to Evaluate Early and Long-term PAH Disease Management (REVEAL) registry. We suspect these varied results may reflect differences in patient demographics. Whereas only 10% of patients in our study were classified as WHO functional class III, the REVEAL registry had 61.5% in class III (8). Furthermore, our cohort was younger and had a longer baseline 6MWD compared with the REVEAL cohort, suggesting that we enrolled a more compensated population (2, 8). Therefore, our findings with respect to the prognostic value are most generalizable to stable patients with PAH.

The major limitations of this study are that it is a single-center, relatively small cohort. Moreover, follow-up echocardiography was at the discretion of the treating physician rather than a protocol, which likely made associations with RV function underpowered. Assessment of hospitalization, however, does not suffer from the same limitation. Although individuals in this study were enrolled in a 12-week behavioral intervention to increase step counts, it is unlikely that the intervention group (n = 20) benefited from a carryover effect given our median follow-up of 2.2 years.

This study suggests that daily step counts may have prognostic value in patients with PAH that warrants further investigation. The widespread use of commercial activity trackers may offer a valuable clinical adjunct to the 6MWD.