It has been demonstrated in recent studies that obstructive sleep apnea (OSA) is the most prevalent sleep disorder in patients with osteoarthritis (OA), and thus the current study aimed to investigate the influence of OSA on knee extensor torque, pain, stiffness, and physical function in men with low-grade knee OA.
The study included 60 male volunteers, aged 40 to 70 years, allocated into four groups: Group 1 (G1) Control (n = 15): without OA and without OSA; Group 2 (G2) (n = 15): with OA and without OSA; Group 3 (G3) (n = 15): without OA and with OSA; and Group 4 (G4) (n = 15) with OA and with OSA.
All volunteers were examined using knee radiographs and polysomnography, responded to the Western Ontario McMaster Osteoarthritis Index (WOMAC) questionnaire, and completed a test on an isokinetic dynamometer to evaluate peak isometric knee extensor torque, both concentric and eccentric (90°/s and 180°/s).
Regarding the data from the WOMAC questionnaire (for pain, stiffness, and physical function), it was observed that G4 showed higher values compared to G1 or G3. For the concentric isometric and isokinetic peak knee extensor torque, lower values were observed in G4 compared to G1 or G3.
Patients who have knee OA in the early grades, when associated with OSA, have higher changes of the peak extensor torque, pain, stiffness, and physical function, compared with patients who did not have OSA.
Silva A, Mello MT, Serrão PR, Luz RP, Ruiz F, Bittencourt LR, Tufik S, Mattiello SM. Influence of obstructive sleep apnea in the functional aspects of patients with osteoarthritis. J Clin Sleep Med. 2018;14(2):265–270.
Current Knowledge/Study Rationale: Both obstructive sleep apnea and osteoarthritis are more prevalent in older populations. Thus, knowledge of the association of these two conditions may help impaired patients recover in relation to their functional capacity and the musculoskeletal system.
Study Impact: The current study confirms a relationship between sleep apnea and osteoarthritis that negatively affects functional capacity and the musculoskeletal system. Therefore, these findings are extremely important for clinical practice, because they must be taken into account for the rehabilitation process of patients suffering from musculoskeletal problems.
Osteoarthritis (OA) is one of the most common global causes of disability and affects approximately 15% of the world population.1 OA is the most common joint disease and a major cause of chronic pain and disability,2 with the knee as one of the most affected regions. Patients with OA report pain, stiffness, and functional limitations3 and frequently complain regarding their sleep patterns.4–6 Recent studies have shown that the most common sleep disorder in patients with OA is obstructive sleep apnea (OSA).7,8
Sleep changes or restrictions can increase musculoskeletal pain and lead to a poor quality of life,9 thus causing losses in the musculoskeletal system, such as muscular atrophy.10 By contrast, patients with OA, particularly in the more advanced stages, may have sleep complaints because of pain.11 However, the effects on the musculoskeletal system, particularly in patients with knee OA, have not yet been investigated.
Quadriceps muscle weakness may be a predictor of pain and physical dysfunction in this population.12 A decrease in the isometric, concentric, and eccentric peak torque of the knee extensor in patients with OA has been demonstrated. This may be associated with a negative correlation between concentric and eccentric peak torque knee extensor and pain, stiffness, and physical function, as assessed using the Western Ontario McMaster Osteoarthritis Index (WOMAC).13
Considering that poor sleep quality negatively affects patients with OA, the purpose of this study was to investigate the influence of OSA on knee extensor torque, pain, stiffness, and physical function in men with low-grade knee OA. We hypothesized that the association between OA of the knee and OSA will result in deficits in functions and pain of the knee.
This study is a part of a cross-sectional study about the relationship of OSA and patients with knee OA, conducted at the Center for Psychobiology and Exercise (CEPE), and Sleep Institute in the city of São Paulo (Brazil). All participants signed an informed consent, and the study was approved by the Ethics and Research Committee of the Federal University of São Carlos (CEP No 109/2011) and registered on ClinicalTrials.gov (identifier: NCT01422967).
The volunteers were recruited through study dissemination in local print and electronic media. The inclusion criteria were male sex, age between 40 and 70 years, diagnosis of knee OA according to the clinical criteria recommended by the American College of Rheumatology,14 and a grade I or II severity according to the Kellgren Lawrence classification15 determined by radiographic examination. Initially, 111 volunteers were enrolled: 37 did not meet the inclusion criteria and 74 volunteers fulfilled the inclusion criteria. Of the 74 who fulfilled inclusion criteria, 14 dropped out (conflicts with the work schedule [n = 6], health problems [n = 4], travel [n = 2] and personal problems [n = 2]) during the study. Therefore, the final sample consisted of 60 volunteers separated into four groups based on the results of the data collected from the radiographs and sleep tests (see below): Group 1 (G1) Control (n = 15): without OA and without OSA; Group 2 (G2) (n = 15): with OA and without OSA; Group 3 (G3) (n = 15): without OA and with OSA; and Group 4 (G4) (n = 15) with OA and with OSA.
The volunteers were first interviewed by the study's principal investigator, and then the resting and exercise electrocardiograms were scheduled with the cardiologist of the CEPE. If results of the electrocardiograms were normal, knee radiographic examinations were scheduled, and the results were analyzed by a rheumatologist of the CEPE. Then, each volunteer was referred to a sleep physician who scheduled the polysomnography testing. After the results were received, a new appointment was scheduled to diagnose OSA. The volunteers then responded to the WOMAC questionnaire and performed a strength test using an isokinetic dynamometer.16 Isokinetic testing was performed on volunteers only after all of these procedures were completed and the subjects were assigned to one of the four groups. In total, the volunteers participated in all stages of the study over five visits to the laboratory. The evaluations were always conducted simultaneously between 2:00 pm and 6:00 pm and performed by a blind evaluator. At the end of each evaluation, the volunteers received a report and explanation regarding the results.
Radiographs were performed for the anteroposterior and mediolateral positions of both knees to confirm the presence of OA: the presence of minimal osteophytes of uncertain significance and the presence of defined osteophytes without reduction of the intra-articular space.15
To record the polysomnography through the night, the Embla S7000 (Embla Systems, Inc., Reykjavik, Iceland) was used in the Laboratory of Sleep (Sleep Institute, São Paulo, Brazil). All recording sensors were attached to the patient in a noninvasive manner with tape or elastic bands. The following physiological variables were monitored simultaneously and continuously: a four-channel electroencephalogram (C3-A2, C4-A1, O1-A2, O2-A1), two-channel electro-oculogram (EOG) (EOG-left-A2, EOG-right-A1), four-channel surface electromyography (submental region muscle, tibialis anterior muscle, masseter region, and seventh intercostal space) and single-channel electrocardiogram (modified lead V1). Airflow detection was conducted via two channels through a pair of thermal sensors (single channel) and nasal pressure (single channel) and respiratory effort of the thorax (single channel) and abdomen (single channel). Using inductance plethysmography, snoring (single channel), the position (single channel), oxygen saturation and a single pulse oximeter were monitored via EMBLA. The sleep stages, arousals, respiratory events, and leg movements were analyzed according to the criteria established by the scoring manual published by the American Academy of Sleep Medicine.17 The OSA diagnosis was obtained by a sleep specialist medical doctor, and volunteers with OSA presented an apneahypopnea index (AHI) of 5 to 15 events/h and at least one clinical complaint of snoring, sleepiness or apnea or an AHI greater than 15 events/h, regardless of symptoms.
The Western Ontario McMaster Universities Osteoarthritis Index (WOMAC) is a self-reported questionnaire designed to assess problems experienced by individuals with lower limb OA. Consisting of three domains: pain, stiffness, and physical function. The score was expressed by a Likert scale, in which none = 0, low = 25, moderate = 50, severe = 75 and very severe = 100.13 The maximum score on each section was expressed as a percentage, with higher scores indicating greater pain, stiffness, and physical dysfunction.
Muscle torque was assessed using an isokinetic dynamometer (Biodex Multi-Joint System 3, Biodex Medical Inc., Shirley, New York, United States). The isokinetic dynamometer was calibrated according to the manufacturer's instructions. Prior to the evaluation, the volunteers warmed up on a stationary cycle ergometer for 5 minutes with a 75 W load and constant speed of 20 km/h followed by general stretching of the lower limbs. Isometric, concentric, and eccentric knee extensor torque evaluations were performed. An evaluation of the isometric torque of the knee extensor was performed at 60° flexion with 5 maximal isometric contractions and each contraction lasting 10 seconds with 5 minutes of rest between each contraction. The concentric and eccentric evaluations were conducted at speeds of 90 and 180°/s (5 and 10 maximum contractions, respectively) with an interval of 5 minutes between each speed.18 Both knees were evaluated. The order of the tests and choice of participants initially evaluated were randomized.
The evaluations were performed with the volunteer sitting and stabilized in the equipment's chair by belts around the chest and pelvic regions and the knee flexed at 90°. The mechanical axis of rotation of the dynamometer was aligned with the lateral epicondyle of the femur, and the resistance was applied distally at the ankle joint 5 cm above the medial malleolus.13 The volunteers were instructed to keep their arms crossed in front of their chest during the test to avoid compensation. Before each assessment, the volunteers performed three sub-maximal contractions to become familiar with the procedures. During the contractions, a verbal command was used to encourage patients to produce maximum torque.13
The torque data, measured in Nm, were normalized by body mass (in kg) using the following formula: (Nm/kg) × 100. For the statistical analysis, the mean peak torque was used for the isometric, concentric, and eccentric isokinetic knee extensors of the OA limb (or the most affected limb in bilateral cases).19 For volunteers without OA, a randomized lottery was conducted to choose the limb examined in the statistical analysis.
Variables with nonparametric distribution were normalized using Z-scores. Levene test was used for the evaluation of intragroup homogeneity. An analysis of the different parameters was performed using a General Linear Model, considering the group as the main factor, and Tukey test was used for multiple comparisons between groups using SPSS version 18 software (IBM Corp, Armonk, New York, United States). The results are presented as the mean ± standard deviation, and the significance level used was P ≤ .05.
Table 1 shows the demographic characteristics of volunteers and observed a homogeneous sample, once the variables age (F56,3 = 0.559; P = .644), height (F56,3 = 2.401; P = .07), weight (F56,3 = 0.606; P = .614) and body mass index (F56,3 = 1.987; P = .126) did not differ between groups. However, for the AHI (F56,3 = 8.191; P = .001; effect size = 0.305; observed power = 0.988), we found a higher AHI in G4, which differs statistically from G1 (P = .013) and G2 (P = .017). We also found an increased number of AHI in G3, which differs significantly from G1 (P = .002) and G2 differ (P = .002).
Demographic characteristics of volunteers.
Demographic characteristics of volunteers.
Table 2 shows the data for the sleep pattern of the volunteers. No differences were observed between groups for the polysomnography parameters except stage N1 sleep and arousals. For stage N1 sleep, a significant difference was observed between groups (F56,3 = 5.365; P = .003; effect size = 0.223; observed power = 0.917) with G4 presenting more stage N1 sleep compared to G1 (P = .053). The G3 also showed more stage N1 sleep compared to G1 (P = .002). Additionally, a significant difference for the arousals (F56,3 = 4.149; P = .010; effect size = 0.182; observed power = 0.827) was observed because G4 showed more arousals compared to G1 (P = .044). The identical result was true for G3, which was different than that for G1 (P = .042).
Sleep parameters obtained by polysomnography.
Sleep parameters obtained by polysomnography.
Figure 1 shows the data for the WOMAC questionnaire (pain, stiffness, and physical function). A significant difference between groups was observed for the pain domain (F56,3 = 5.853; P = .002; effect size = 0.239; observed power = 0.939) with G4 showing higher values compared to G1 (P = .002) or G3 (P = .008). An identical difference was observed for the stiffness domain (F56,3 = 5.604; P = .002; effect size = 0.231; observed power = 0.929), in which G4 showed higher values compared to G1 (P = .004) or G3 (P = .005). A significant difference between groups was also observed for the physical function domain (F56,3 = 6.860; P = .001; effect size = 0.269; observed power = 0.969), in which G4 showed a higher physical function score compared to G1 (P = .001), G2 (P = .034) or G3 (P = .002).
Data are presented as mean ± standard deviation. General Linear Model, followed by Tukey test. Significance was assumed at P ≤ .05: * = compared to G1, § = compared to G2, ¥ = compared to G3. WOMAC = Western Ontario McMaster Osteoarthritis Index
In the analysis of peak knee extensor torque parameters (Table 3), the assessment of isokinetic concentric at 90°/s showed a significant between-group difference (F56,3 = 4.595; P = .006; effect size = 0.198; observed power = 0.867). Lower extensor torque values were observed in G4 compared to G1 (P = .029) or G3 (P = .018). A significant difference between the groups was also observed in the assessment of isokinetic concentric at 180°/s (F56,3 = 10.149; P = .001; effect size = 0.352; observed power = 0.997) with the lowest value of the peak knee extensor torque in G4 compared to G1 (P = .001) or G3 (P = .001) and lower values in G2 compared to in G1 (P = .004) or in G3 (P = .027).
Data on knee extensor torque of volunteers.
Data on knee extensor torque of volunteers.
For the eccentric isokinetic test, no significant difference was observed for the peak knee extensor torque between groups at either speed: 90°/s (F56,3 = 2.108; P = .109; effect size = 0.101; observed power = 0.511) and 180°/s (F56,3 = 2.843; P = .051; effect size = 0.132; observed power = 0.651). Similar to the isometric test, a significant difference was observed between groups (F56,3 = 9.406; P = .001; effect size = 0.335; observed power = 0.995), in which lower torque values were observed for G4 compared to in G1 (P = .012) or in G3 (P = .002) and lower values in G2 compared to in G1 (P = .004) or in G3 (P = .001).
In the current study, the group of individuals with OA and OSA presented a lower peak knee extensor torque in the isometric and concentric assessments and higher values in all domains (pain, stiffness, and physical function) of the WOMAC questionnaire compared to the other groups.
Some studies have reported that OA affects the initiation and maintenance of sleep. Moreover, it is speculated that the decrease in sleep time exacerbates pain in these individuals,11 which might be associated with conditions of joint pain and thus disrupt the sleep-wake cycle.20
Taylor-Gjevre and colleagues8 evaluated the quality of sleep in patients with OA using the Pittsburgh questionnaire and observed that 53% of patients had sleep disturbances. In the current study, the G4 volunteers presented an increased number of arousals and stage N1 sleep compared to the other groups. These results corroborated the literature, as a polysomnography study with 14 patients with OA showed an increase in stage N1 sleep, decreased stage N2 sleep, and increased sleep fragmentation compared to 16 healthy individuals.21 Additionally, the presence of OSA in this group should be observed. OSA caused sleep disruption, which resulted in poor quality and little restorative sleep in this specific population.
Therefore, the pain domain assessed using the WOMAC in the current study was increased in G4. This result was supported by other studies that reported that patients with knee OA presented with increased pain.3,13 However, as previously described, OA pain can have negative effects on sleep,22 and fragmented sleep in patients with OSA exacerbates musculoskeletal pain.23 These results were consistent with those of previous studies that observed that sleep disturbances were common in individuals with OA and associated with increased pain perception and greater fatigue.24
The results of the current study corroborated those of previous studies suggesting that pain-related sleep disturbances lead to a bidirectional relationship between sleep disorders and pain.25 Thus, the current results demonstrate that, in addition to previously present OA pain, OSA appears to exacerbate sleep losses.
The WOMAC questionnaire used in studies involving knees with OA13 has been recently correlated with sleep disturbances.26 Parimi and colleagues26 used the WOMAC questionnaire to show that sleep fragmentation was higher in women with pain in hip joints with OA compared to women without pain.
In addition to the pain domain, the WOMAC also evaluates the stiffness and physical function of patients with OA. In the current study, both domains showed a higher score in G4 compared to the other groups, which suggested an altered perception of stiffness and function because of the illness. These identical results were obtained by Hubley-Kozey and colleagues27 in patients with grade I to III and Serrão and colleagues13 in patients with early stages of OA. Because OA has characteristics of joint degeneration, it has a clinical presentation of pain, stiffness, and impairment of physical function and is expected to present problems with patients' quality of life.28 Similarly, patients with OSA also present with behaviors similar to those of patients with OA because some studies have shown a decrease in the quality of life of these patients.9
The altered perception experienced by these individuals suggests that OSA, identified in G4, may have an effect on OA in these individuals because sleep disorders are associated with impaired daytime function, fatigue, and a reduced quality of life.29 These sleep disturbances have a correlation with knee pain and worsening of self-perceived health, function, and physical performance.4
Regarding physical performance, the OA groups (G2 and G4) showed lower peak knee extensor torque compared to the groups without OA for the concentric isokinetic and isometric assessments in the current study. This result is supported by the literature because the patients with OA have weakness of the quadriceps muscle,30 which causes an increased load on the knee joint and may contribute to increased pain. The literature previously described that the quadriceps muscle is an important structure that absorbs the effects suffered by the knee joint, particularly in the eccentric action of this muscle.31 Serrão and colleagues13 observed an inverse correlation between knee extensor torque and pain, as assessed through the WOMAC questionnaire in patients with OA, which indicated that the lower the extensor torque, the higher the self-reported pain in patients with low-grade OA knees.
Additionally, at speeds of 90°/s in concentric isokinetic test, G4 presented the lowest extensor torque value compared to G1 or G3, thus demonstrating that OSA may be influencing this decrease in muscle torque previously present in low-grade OA. This result supports the hypothesis that sleep deprivation induces muscle atrophy because sleep is important for maintaining a healthy musculoskeletal system,10 thus contributing to a decreased ability to generate force.
In the current study, familiarization with the maximal test on the isokinetic dynamometer was not performed because it was only conducted with the submaximal test on the assessment day. Thus, this aspect can be considered a study limitation.
This study was a cross-sectional investigation and observed that OSA causes sleep fragmentation, therefore enhancing the negative effects presented by patients with OA; one may conclude that patients with OA and OSA had lower peak knee extensor torque and received higher scores in all WOMAC domains. Thus, the sleep negatively affected the musculoskeletal system and, consequently, the functional aspects of patients with OA of the knee previously present at early stages of the disease. Therefore, understanding sleep disorders in patients with OA must be considered in the process of rehabilitation because it can maximize the clinical status presented by the disease.
Work for this study was performed at Sleep Institute of São Paulo, Brazil. All authors have seen and approved the manuscript. The authors report no conflicts of interest.