C-reactive protein, but not neutrophil-lymphocyte ratio, is inversely associated with muscle strength only in older men: NHANES 1999 – 2002

To evaluate the association of inflammation (C-reactive protein (CRP) and neutrophil-lymphocyte ratio (NLR) levels) with muscle strength in older adults. We also aimed to evaluate whether these associations are sex-specific. A cross-sectional study was performed with data from the National Health and Nutrition Examination Survey (NHANES) 1999 – 2000 and 2001 – 2002. A total of 2387 individuals over 50 years of both sexes were evaluated, according to the eligibility criteria for the strength test. Muscle strength was measured by Kinetic Communicator isokinetic dynamometer; while the NLR was obtained by the ratio of the total neutrophil for lymphocyte count and CRP was quantified by latex nephelometry. Linear regression analyses, crude and adjusted for confounders, were used to estimate the coefficients and 95 % confidence intervals for peak strength (muscle strength) by tertiles of NLR and CRP. There was no association between NLR and peak strength for both sexes. CRP levels were inversely associated with peak force in men [2nd tertile β = (cid:0) 3.33 ( (cid:0) 15.92; 9.25); 3rd tertile β = (cid:0) 24.69 ( (cid:0) 41.18; (cid:0) 8.20), p for trend = 0.005], but not in women [2nd tertile β = (cid:0) 3.22 ( (cid:0) 15.00; 8.56); 3rd tertile β = (cid:0) 9.23 ( (cid:0) 28.40; (cid:0) 9.94), p for trend = 0.332]. In conclusion, NLR levels were not associated with muscle strength in both sexes. CRP levels were inversely associated with muscle strength in older men, but not in women, suggesting that the association between inflammation and muscle strength in older adults can be sex-specific.


Introduction
Muscle strength (measured by isokinetic strength of knee extension) decreases with aging by 14 % per decade (Frontera et al., 2000;Hughes et al., 2001), which can increase the risk of overall mortality (Li et al., 2016;Peterson Zhang et al., 2016). The reduction in strength can occur due to several reasons, such as anabolic resistance (a condition that decreases the quality or quantity of muscle), low protein intake, sedentary lifestyle (Burd et al., 2019;Paulussen et al., 2021), diabetes mellitus, smoking (Stenholm et al., 2010), increased oxidative stress (Baumann et al., 2016;Fulle et al., 2004) and inflammation (Tuttle et al., 2020). Regarding inflammation, it is a biological response of the immune system to different stimuli; however, when it occurs for a prolonged period a reduction in muscle strength is generally observed (Ferrucci et al., 1999;Singh and Newman, 2011) mainly due to an increase in catabolic state (Tuttle et al., 2020) and anabolic resistance (Breen et al., 2013). However, although increased inflammation has been considered one of the causes of muscle strength loss (Ferrucci et al., 2002), the evidence is still controversy (Goldeck et al., 2016;Koshikawa et al., 2020;Legrand et al., 2013) and more studies are needed.
C-reactive protein (CRP), a marker of acute and chronic inflammation, has been used to predict low physical performance among the elderly population (Yoshida et al., 2010), since it seems to be inversely associated with muscle strength (Tuttle et al., 2020). Besides the association between CRP and muscle strength, several other inflammatory markers have been inversely associated with muscle strength, such as Interleukin-6 (IL-6) and Tumor Necrosis Factor-α (TNF-α) (Tuttle et al., 2020). Tuttle et al. (2020) demonstrated that these inflammatory markers are inversely associated with strength independent of the sex, but a stronger association was observed in men, suggesting that there could be a sex-specific association. Nevertheless, Tuttle et al. (2020) also observed that the male population included in the meta-analysis had higher prevalence of chronic diseases when compared with women, which can be a confounder for the analyses. Therefore, more studies are needed evaluating the association between inflammation and strength separating by sex and also adjusting for important confounders.
Neutrophil and lymphocyte ratio (NLR) has been associated with pro-inflammatory cytokines (Guthrie et al., 2013;Motomura et al., 2013) and can also be used as an inflammatory marker. In patients with chronic diseases, NLR is associated with the risk of death from cardiac events (Shah et al., 2014), but it is also related with the risk for sarcopenia and low muscle strength in cancer patients (Borges et al., 2021;Mainardi et al., 2020;Proctor et al., 2011). Then, despite the NLR is a biochemical exam with low-cost and accessible in clinical practice, its applicability to detect the low muscle strength has not yet been evaluated in epidemiological studies evaluating a representative sample.
Thus, the present study aimed to evaluate the association of CRP and NLR levels with muscle strength in older adults derived from National Health and Nutrition Examination Survey (NHANES) 1999-2002. We also aimed to evaluate whether these associations are sex-specific. We hypothesized that high CRP and NLR levels would be associated with low muscle strength; and that the associations would be stronger in men.

Data selection and participants
This is a cross-sectional study developed with individuals from National Health and Nutrition Examination Survey (NHANES) 1999(NHANES) -2000(NHANES) and 2001(NHANES) -2002. NHANES is conducted by Centers for Disease Control and Prevention (CDC) and is a public survey with a representative stratified sample of the non-institutionalized population of the United States of America, which assesses measures of health and nutrition. The participants of NHANES provided a written informed consent in accordance with the National Center for Health Statistics Research Ethics Review Board (NCHS ERB) (protocol #98-12 for NHANES cycle 1999#98-12 for NHANES cycle -2002. In NHANES 1999NHANES -2000NHANES and 2001NHANES -2002 datasets, a total of 21,004 participants were evaluated. In the present study, only individuals aged from 50 to 85 years were included due to strength test eligibility. In addition, participants with missing data referring to biochemical evaluations, anthropometry and dietary intake were excluded. Individuals with implausible peak force velocity (>65 and <55 s) (Kuo et al., 2006) and who did not perform at least four trials in the isokinetic strength test were also excluded. Therefore, 2387 individuals were evaluated in the present study (Fig. 1).

Anthropometric data
Weight and height were evaluated by a trained examiner in the mobile examination center using the recommendations of the Anthropometric Standardization Reference manual (Lohman et al., 1988). Body Mass Index (BMI) was calculated dividing weight per height squared.

Muscle strength
Muscle strength was evaluated through the peak isokinetic knee extensor measured by Kinetic Communicator isokinetic dynamometer (kin ComMP, chattecx Corp., Chattanooga, TN). The individuals were instructed to perform six measurements of muscle strength of the right quadriceps at a speed of 60 • per second. The first three measurements were performed for warm-up/learning, and the last three trials were made for recording the muscle strength. Peak strength was considered the highest value observed from the fourth complete assessment. For this reason, participants who did not complete at least four assessments were excluded from the study (Kuo et al., 2006;Rossato et al., 2020).

Dietary intake
One 24-h dietary recall was applied in a private room by trained interviewers to obtain the dietary intake information. During NHANES 1999-2000and 2001, the 4-step multiple pass was used to evaluated the dietary intake, while in NHANES 2002 the automated 5-steps multiplepass method was used (Prevention CfDCa, 1999-2000Prevention Excluded (missing) (n=18197): Excluded as implausible criteria (n=420): Fig. 1. Flowchart of the sample selection (NHANES 1999(NHANES -2002. CfDCa, 2001CfDCa, -2002. To code and report the NHANES 1999-2000 dietary data, the Food Intake Analysis System, version 3.99; and USDA 1994-98 Survey Nutrient Database were used. USDA Food and Nutrient Database for Dietary Study was used to evaluate the dietary intake of 2001-2002. Intakes of energy (kcal/day), carbohydrate (g/day), protein (g/day and g/kg), lipids (g/day); saturated, polyunsaturated and monounsaturated fats (g/day), cholesterol (mg/day) and alcohol (g/day) were analyzed.

Biochemical parameters
The Beckman Coulter® MAXM instrument in the Mobile Examination Center (MEC) was used to produce a complete blood count on blood specimens and provided a distribution of blood cells for all participants. The methods used to derive complete blood count (CBC) parameters were based on the Beckman Coulter® method of counting and sizing, in combination with an automatic diluting and mixing device for sample processing, and a single-beam photometer for hemoglobinometry. The white blood count (WBC) differential uses VCS technology. The NLR was obtained by the ratio of the total neutrophil to lymphocytes count. CRP was quantified by latex nephelometry.

Covariates of interest
The demographic variables evaluated were age (years), sex (men or women), race (non-Hispanic white or others), marital status (single/ divorced/widowed/never married or married/living as married), annual family income (0 to $19,999, from $20,000 to 54,999 or ever $55,000 per year) and educational level (under high school graduate and some college or over). Health conditions and lifestyle habits included in this study were hypertension (yes: to people diagnosed with high blood pressure by a doctor or other healthcare professional or no), diabetes (yes: if the doctor ever said that the person had or no), arthritis (yes: if the doctor ever said that the person had or no), cancer (yes: if the doctor ever said that the person had or no), smoke (yes: if smoked at least 100 cigarettes in life or no) and menopause for women (yes: if do not had regular periods in past 12 months or no). Physical activity level included moderate, vigorous or muscle strength exercises (yes: if do any moderate, vigorous or muscle strength activities over the past 30 days or no), as previously described (Rossato et al., 2020). Body mass index (BMI, kg/m 2 ); energy intake (kcal/day), protein intake (g and g/kg day) and alcohol consumption (g/day) were also assessed through the application of a 24-h food recall (Tables 1 and 2).

Statistical analyses
Due to the magnitude of the sample size, NLR and CRP were divided into tertiles to facilitate the interpretation of the results regarding the association between inflammation and muscle strength, like other studies with large sample sizes (Gicevic et al., 2021;Castellanos-Gutiérrez et al., 2021). The demographic characteristics, health conditions and habits, anthropometric data, muscle strength, dietary intake and supplementation, and biochemical parameters were compared according to the tertiles of NLR and CRP using regression analysis. The continuous variables were described as mean and standard deviation, while the categorical variables were described as percentage and confidence interval. Linear regression was used to estimate the coefficients and 95 % confidence intervals (95%CI) for dependent variable peak force (muscle strength) by tertiles of NLR and CRP (independent variables). The analyses were performed without (Model 1) and with adjustments for confounders (Model 2). The variables included as adjustments were age, sex, race, marital status, annual family income, education level, body mass index, energy intake (kcal/day), protein intake (g/day), alcohol consumption (g/day), smoking, menopause status (only for women), physical activity, hypertension, diabetes, arthritis and cancer. Since muscle strength was part of NHANES MEC exam, all analyses were performed using "svy" command to consider the information on the "full sample 4-year MEC exam weight" sample weight, primary sampling units and strata for correct variance estimation. All statistical analyses were performed using Stata 12.0 software (StataCorp LP, College Station, TX, USA). In all analyses p < 0.05 was considered statistically significant.

Individual characteristics
The values of NLR in 1st tertile ranged from 0.27 to 1.65, in 2nd tertile from 1.66 to 2.4, and in 3rd tertile there was a greater variation, ranging from 2.41 to 21.2 (Table 1). A higher prevalence of men, non-Hispanic white, and presence of hypertension and cancer was observed for individuals in the 3rd tertile of NLR when compared with who had lower NLR. In addition, the individuals in the 3rd tertile were older and consumed more energy (kcal), carbohydrate (g), lipid (g), saturated (g/day) and monounsaturated fats (g/day), and cholesterol (mg/day). These individuals also presented higher values of white cell count, neutrophils (% and n), CRP and NLR levels. However, individuals in the 3rd tertile had a lower lymphocyte count (% and n) ( Table 1).
The values of CRP in 1st tertile ranged from 0.01 to 0.16 mg/dL, in 2nd tertile from 0.17 to 0.4 mg/dL, and in tertile 3 there was a greater variation, ranging from 0.41 to 14 mg/dL (Table 2). Comparing the CRP tertiles, the individuals in the 3rd tertile were older, and a lower percentage of men, annual family income, education level and physical activity level was observed. However, a higher number of individuals with hypertension, diabetes, arthritis and smokers was observed in the 3rd tertile. They consumed less energy (kcal), carbohydrate (g), protein (g and g/kg/day), lipid (g), saturated and monounsaturated fats (g/day). These individuals presented lower muscle strength, height (m) and lymphocytes (%); and higher weight, BMI, white cell count, neutrophils (% and g), lymphocytes (g), platelets (%), NLR and CRP levels ( Table 2).

Association between inflammatory markers and peak force
Linear regression analyses showed a negative association between tertiles of NLR and peak strength for men and women, but not for total sample, in the unadjusted model (Supplementary Table 1; Model 1). However, these associations were no longer significant after adjustments for confounders ( Fig. 2a,  A negative association between tertiles of CRP and peak force in the total sample was observed, but no association was found for men and women in the analyses without adjustments (Supplementary Table 2; Model 1). After adjustments for confounders, CRP was negatively associated with strength for total sample and men, but not for women (Fig. 2d, e, f; Supplementary Table 2; Model 2).

Discussion
This study investigated the association of inflammatory markers with muscle strength, and although no association was found between NLR and peak strength, CRP levels were inversely associated with strength specifically in men. Thus, our results suggest that the association between CRP and muscle strength in older adults seems to be sex-specific.
The inverse association between CRP and muscle strength observed in the present study is, at least partially, in agreement with the literature. A recent meta-analysis (Tuttle et al., 2020) assessed the association between inflammatory markers and loss of muscle strength, and found that all the inflammatory markers assessed (CRP, IL-6 and TNFα) were inversely associated with knee extension strength. However, unlike in our study, the associations were statistically significant independent of the sex, even when CRP was evaluated. Nevertheless, a sex-difference cannot be ignored as a stronger negative correlation of IL-6 and TNFα with muscle strength was observed in men (Tuttle et al., 2020), suggesting that the impact of inflammatory markers on muscle strength can be sex specific. Further studies are needed to elucidate the mechanisms that explain these sex-differences. The mechanism by which CRP is associated with muscle strength is still unclear, but it is suggested that increased CRP levels are related to increased oxidative stress, which seem to affect muscle size, fiber activation, and excitation-contraction (Fulle et al., 2004) that can compromise muscle strength (Khansari et al., 2009;Powers et al., 2011). However, as mentioned before, it is still unclear the reason that the association between CRP and muscle strength was significant only in older men. CRP is normally produced by the liver and regulated by IL-6. However, in obese individuals, adipose tissue-derived CRP production highlighted the body fat as a producer of CRP levels (Farmer et al., 2020). CRP levels seem to be influenced by visceral adiposity in men and subcutaneous adiposity in women (Lindboe and Platou, 1982). Then, as we did not assess the body composition in the present study, we cannot assess the influence of visceral fat on these associations. However, the analyses were adjusted for body weight that reflects both body fat and fat-free masses, which shows that the association between CRP and strength are likely independent of the body composition compartments. NLR is a parameter that allows the assessment of the inflammatory state with a low cost and is related with pro-inflammatory cytokines (Guthrie et al., 2013;Motomura et al., 2013). In addition, NLR has been associated with sarcopenia and muscle strength in some clinical populations, such as cancer patients (Borges et al., 2021;Mainardi et al., 2020;Proctor et al., 2011). For this reason, we hypothesized that NLR could be negatively associated with muscle strength in older adults. However, no association was observed in the present study, which can possibly be explained by the low values of NLR. For healthy adults, NLR has been considered adequate when ranges from >0.78 to <3.58 (Forget et al., 2017). In our study, individuals in the 3rd tertile of NLR presented mean value of NLR = 3.3 indicating that this population did not have high values of NLR, although the values of the 3rd tertile ranged from 2.4 to 21.2. A recent study showed that cancer patients presenting NLR >6.5 were associated with the risk of sarcopenia (Borges et al., 2021), which suggests that NLR can be associated with muscle strength mainly in individuals having high values of NLR.
Study limitations: i) retrospective data regarding 20 years old ago that does not allow to conclude a causal relationship; ii) despite being adjusted for several types of chronic diseases, we did not stratify the sample by the type of diseases; iii) and despite the data presented nonnormal distribution, we performed the analyses using linear regression because even after being transformed into logarithms, non-normality was maintained and there was a worsening in the data distribution. For this reason, we believe that the present analyses are the best form for this study. Conversely, the present study has several strengths: i) a representative population of the United States of America was evaluated, which increases the power to detect small associations; ii) the analyses were adjusted for important variables in order to minimize confounding residues from observational studies.

Conclusion
NLR levels were not associated with strength in older adults independent of the sex. CRP levels were inversely associated with muscle strength in older men, but not in women, suggesting that the association between inflammation and muscle strength in older adults can be sexspecific. Longitudinal studies are needed to confirm whether elevated CRP levels change the muscle strength only in men, as well its effect in different type of population.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

CRediT authorship contribution statement
PCBL, EPO and GDP contributed to study design. PCBL and FMSB  have prepared the statistical analyses. PCBL, EPO and GDP wrote the manuscript. FMSB and CP contributed to discussion of manuscript. All authors read and approved the final version of the manuscript.

Declaration of competing interest
The authors declare that they have no known competing financial  Fig. 2. Linear regression between tertiles of neutrophil-to-lymphocyte ratio and peak force (Newtons) (A, B, C) and C-reactive protein and peak force (Newtons) (D, E, F). (NHANES 1999(NHANES -2002. Adjusted for age, sex, race, marital status, annual family income, education level, body mass index, energy intake (kcal/day), protein intake (g/day), alcohol consumption (g/day), smoking, physical activity, hypertension, diabetes, arthritis, cancer and for women, menopause. Tertile of neutrophilto-lymphocyte ratio ( interests or personal relationships that could have appeared to influence the work reported in this paper.