Introduction

Identifying subgroups of patients who benefit less from physiotherapy treatment is a crucial preparatory step to be able to early adjust treatment expectation and treatment programme. The primary treatment modality for patients with Peripheral Arterial Disease (PAD) is physiotherapy and practice results show a high variety in treatment outcome.

Peripheral Arterial Disease in the leg(s) is a manifestation of narrowing or blockage of arteries (atherosclerosis) that causes poor blood flow to the lower extremities. The clinical spectrum of patients ranges from asymptomatic atherosclerosis to those with the common symptoms of Intermittent Claudication (IC), e.g. tight, aching, or squeezing pain in the calf, thigh or buttock. Symptoms occur when the delivery of oxygen is inadequate to meet the metabolic requirements of the active skeletal muscle [1]. Leg pain is relieved by rest. Intermittent Claudication leads to a moderate or severe impairment in walking ability. In order to improve walking ability, Supervised Exercise Therapy (SET) by a physical therapist is recommended as the primary training modality for this patient group [2–4]. Walking limitation is measured by the Initial Claudication Distance (ICD), the distance at which the patient experiences the onset of claudication pain and the Absolute Claudication Distance (ACD), the distance at which the pain forces the patient to stop walking [3].

A few small scale studies indicate that patient characteristics and comorbid conditions might lead to considerably poorer treatment outcomes during and after SET [5–10]. For example, Katzel et al. [6] (N = 141) pointed out that current smokers had a shorter time to onset of claudication pain compared to former or never smokers (188± 28 s vs. 173±25 s and 309± 54 s, respectively; p = 0.023). A reduced improvement after conservative treatment by risk factor modification and physical training for female patients (hazard ratio (HR) 0.51; p = 0.052) was presented by Amighi et al. [7] (N = 181). Dolan et al. [8] (N = 460) observed that after adjusting for age, PAD patients with diabetes type 2 had a shorter mean 6-min walking distance (1,040 vs. 1,168 feet; p<0.001). A recent study of Farah et al. [9] (N = 415) explored that Cardiovascular Disease (CVD) was associated with lower ACD and ICD (OR = 2.45; CI:1.11–5.39 and OR = 1.88; CI:0.86–4.09, respectively) and age, body-mass index (BMI), pulmonary disease and smoking behavior were predictive variables for ACD that were presented by Kruidenier et al. [10]. Taken together, from a scientific point of view these studies seem to provide compelling evidence that specific patient characteristics and comorbidity influence physiotherapy treatment success in patients with IC. However, all previous investigations were based on small numbers of patients (maximum N = 460 [7]) largely recruited in the hospital setting. By the use of our large cohort we want to check whether the hampering effects of patient-related variables and comorbidity remain in place in a large population of patients with IC that follow SET in the usual care setting. Moreover, the large sample size of our cohort makes it possible to use a multilevel model through which it can be acknowledged that patients received treatment by different physiotherapists [11].

If the results of our large physiotherapy cohort show that the inverse association between patient characteristics and comorbidity (as identified in earlier small scale studies) remains, subgroups of patients that benefit less from SET can most certainly be identified. The aim of this study is to investigate the association between age, gender, BMI, smoking behavior, orthopedic comorbidity, neurologic comorbidity, cardiac comorbidity, vascular comorbidity, pulmonic comorbidity, internal comorbidity and initial claudication distance (ICD) during and after supervised treadmill exercise at 1, 3, 6 and 12 months in a large physiotherapy cohort study.

Methods

Results

Descriptive data of participants

The inclusion criteria were met by 2995 patients. The average age of patients was 67 years. Comorbidities that were present in more than half of all patients were vascular comorbidity (62%), internal comorbidity (54%) and cardiac comorbidity (49%). Mean initial claudication distance (ICD) at baseline was 270m and after 12 months of SET 667m. The characteristics of the study population are described in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of the study population.

https://doi.org/10.1371/journal.pone.0146828.t001

The results of the three-level mixed linear regression analysis showed at a 1% significance level that being female (p = 0.000), advanced age (p = 0.000) and a high BMI (p = 0.000) lead to less improvement in ICD after SET. Cardiac comorbidity seemed to be associated with improvement in ICD at a 5% level (p = 0.011). In addition, all three interaction terms, time by age (p = 0.000), time by vascular comorbidity (p = 0.023) and time by BMI (p = 0.023) were significantly associated with ICD progress (Table 2). Interactions gender*time and smoking*time were tested but were not significant.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Final mixed linear model.

https://doi.org/10.1371/journal.pone.0146828.t002

For females the predicted ICD is 41m less than for males after 12 months of SET. For patients with a cardiac comorbidity, the predicted ICD is 30m less after 12 months compared to patients without a cardiac comorbidity. Both interpretations apply when all other variables are kept constant.

To calculate the improvement in ICD from baseline to IC at 1, 3, 6 and 12 months for the interactions with time the following regression equation was used:

In general terms, per year increase in age (range 33–93) ICD improvement is 8 meters less after 12 months. Per unit increase of BMI (range: 16–46) the improvement of ICD is 10 meters less after 12 months and the presence of vascular comorbidity leads to 85 meters less improvement after 12 months of SET. In all three scenarios all other variables are hold constant. Fig A, Fig B and Fig C in S2 File provide plots of predicted values for the total sample per subgroup of age, BMI and vascular comorbidity to visualize interactions with time.

To check the robustness of our findings two additional analysis were performed. First, we eliminated BMI from the model (30% missing values), resulting in 1254 more patients in the analysis. The results of this analysis were in agreement with the results of the mixed linear model described in Table 2. Second, in view of non-normality of ICD the analysis was repeated by using a mixed logistic regression model after dichotomizing ICD as 1 if ICD = 1600m and as 0 if ICD <1600m. In accordance with the linear analysis, the results showed that being female (p = 0.000), advanced age (p = 0.000) and a high BMI (p = 0.000) were associated with shorter ICD at all-time points. Furthermore, for current smoking (p = 0.031) and vascular comorbidity there seemed to be a negative association with ICD (p = 0.006). None of the interaction terms were significant, probably due to loss of power by dichotomization.

Discussion

This study showed that female gender, advanced age, a high BMI and cardiac comorbidity are associated with less improvement in Initial Claudication Distances (ICD) after 1, 3, 6 and 12 months of Supervised Exercise Therapy (SET). In addition, the interaction time by age, time by BMI and time by vascular comorbidity were significantly associated with ICD. Additional analysis showed that there might be a negative association between vascular comorbidity and ICD and well as between cigarette smoking and ICD.

Some limitations are worth noting. First, due to the cross-sectional design of this study no causality can be inferred. Second, the absence of a control group and a randomisation procedure prevents interpretation of time effects as treatment effects. Time effects may also have been detected due to spontaneous recovery or treatments patients follow or experience outside the physiotherapy setting. Third, comorbidity was registered at the first visit, hence before the start of SET. Therefore, it was not possible to investigate whether the presence of comorbid diseases changed over time. Fourth, the exact training program for each individual patient was not recorded. In general, the recommendations stated in the guideline Intermittent Claudication (IC) of the Royal Dutch Society for Physiotherapy are used to treat patients with PAD. However, we have no information about whether the general recommendation was adjusted for an individual patient. Therefore, the influence of variance in treatment on ICD after SET could not be investigated.

Noteworthy is the fact that this study highlights a significant association with ICD for age, gender, BMI, and cardiovascular comorbidity at one go. When looking at previous research, the hampering association of all these factors already has been established [5–10]. However, it was spread throughout different studies, each presenting the adverse effect of one or two patient characteristics and/or comorbidity on walking distance in patients with IC. A possible explanation might be that the sample sizes of previous studies were relatively small (maximum N = 460 [6]) in comparison with our study (N = 2995). Furthermore, for this study patients were not intentionally recruited, but data from EMRs was used, which allowed the research team to investigate which patient characteristics and comorbidity affect ICD in the standard physiotherapy care setting.

The inverse association between anthropometric indices like gender, age and BMI and walking distance has been previously studied. To illustrate, Troosters et al. [21] highlighted that 66% of the variability in walking distance assessed by the six-minute walk test in healthy participants aged 50–85 years is attributable to age, gender and BMI. Applying these finding to patients with IC, a systematic review of Robeer et al. [22] showed that older patients respond less favourably to SET compared to their younger counterparts. Changes in the cardiovascular system that occur with aging might explain why advanced age reduces the improvement in ICD after SET. Alterations in cardiovascular physiology like a reduction in a heart’s peak capacity to pump blood by 5% to 10% per decade might cause a diminished aerobic capacity with increasing age. Furthermore, with advanced age, the elasticity of arteries decreases and vascular walls thicken. This results in a slower exchange between oxygen and waste products [23]. As intermittent claudication is already associated with a decreased blood flow to the lower extremities particular during times of increased metabolic demand such as during SET, above mentioned age-related cardiovascular adaptations might worsen muscle oxygen supply. Thus, the combination of PAD and age-related cardiovascular adaptations might cause oxygen ischemia faster; the muscle may not receive as much oxygen as needed to support physical activity and this might explain why ICD is lower in older patients with IC compared to their younger counterparts. The capacity to perform aerobic exercise also depends on the ability of the heart to augment its output to the exercising muscles. This ability is often impaired in patients with cardiac comorbidity. Results of the present study show that patients with cardiac comorbidity show fewer improvements in ICD after SET, which might be caused by inadequate blood flow to the active skeletal muscle secondary to impaired cardiac output. Research has shown that patients with cardiac diseases may achieve less than 50% of the maximum cardiac output attained by individuals without cardiac disease [24].

Regarding the inverse association between a high BMI and less improvement in ICD after SET, previous research reported that BMI is a predictor of walking ability and functional capacity. Overall, it seems that increased body mass, either around the hip and thighs (gynoid obesity) or around the belly (android obesity) reduces walking ability and has a negative impact on functional capacity. The results of studies whether high fat mass or low lean mass are more important in the association between obesity and walking capacity remain contradictory. A study by Jankowski et al. [25] and a study by Davison et al. [26] found that a higher fat-mass is more important than low lean mass in the association with mobility-related disability. In contrast, a study of Newman et al. [24] showed an independent effect of low lean mass on walking ability.

The results of the present study also report that females show less improvement in ICD compared to males. Microvascular gender differences in patients with IC have been previously reported and might explain the differential improvement in ICD between men and women [27,28]. A lower haemoglobin oxygen saturation of the calf muscles combined with a more impaired microvascular function seems to cause greater muscle ischemia during SET in women compared to men [27,28]. Other general gender-based biochemical differences like higher levels of oxidative stress and inflammation that are common in postmenopausal women might also contribute to lower improvements in ICD [27]. Nevertheless, other differences such as motivation, willingness to walk through claudication pain and pain tolerance threshold might be additional explanatory factors.

To check the robustness of our findings among others mixed logistic regression analysis was performed. This analysis was the only supplementary analysis reporting that addionally current smoking and vascular comorbidity seemed to be associated with less improvement in ICD after SET. Nevertheless, previous research reported that current smoking is associated with poorer measures of exercise capacity during treadmill testing, including lower peak oxygen update and earlier onset of claudication pain during walking. A study of Afaq et al. [29] found that current smokers have significantly lower rates of tissue oxygenation at peak claudication distance, which might explain the lower ICD improvements in current smokers compared to non- or past-smokers found in this additional analysis. Peripheral factors that impact walking ability and therefore might explain lower ICD due to vascular comorbidity may include abnormalities in the endothelial function like vasodilatory capacity during physical activity and distribution of cardiac output. Since only one of the additional analysis showed current smoking and vascular comorbidity being predictors of ICD, these two findings should be interpreted with caution. Due to dichotomization of the outcome variable we lost power in the analysis, which might also explain why none of the interaction terms was significant.

In conclusion, the present study showed that being female, advanced age, high BMI and cardiac comorbidity are associated with less improvement in ICD after SET. Because this large prospective usual care cohort study showed that the association of all these factors remained in place, subgroups of patients who benefit less from SET have been most certainly identified. Taking care of these patients who benefit less from SET is a crucial preparatory step to be able to early adjust treatment expectations and treatment programme.

Supporting Information

Acknowledgments

We want to thank Bart Bongers for his support and insightful comments related to exercise physiology. He supported us to explain why certain patient characteristics and comorbidities may impede improvements in initial claudication distance.