Thresholds of handgrip strength for all-cause mortality in patients with chronic kidney disease: a secondary systematic review with dose-response meta-analysis

Abstract

Background

Although handgrip strength is associated with all-cause mortality in patients with chronic kidney disease (CKD), whether this relationship is dose-related is unknown. Therefore, we examined dose-response relationships between handgrip strength and all-cause mortality in CKD patients based on previous studies by meta-analysis.

Methods

Data sources included three electronic databases (PubMed, Web of Science, and Embase) from inception through October 2023. The included cohort was a CKD population not limited to disease stage, and their handgrip strength was objectively measured. Two researchers independently screened studies, extracted data, and assessed the risk of bias. We utilized estimates of handgrip strength categories using robust-error meta-regression (REMR), pooled study-specific estimates, and established dose-response relationships. Outcomes of interest included only all-cause mortality.

Results

A total of 18 studies with 4810 participants (aged 47–71 years) were included. REMR modeling showed a U-shaped trend of association between handgrip strength and all-cause mortality in patients with CKD. Higher handgrip strength values, from 10 kg to approximately 28 kg, were associated with lower mortality risk. After that, the risk of death increased slightly.

Conclusion

A U-shaped association exists between handgrip strength and all-cause mortality risk in CKD patients. Future studies with quantitative measurements for each CKD stage will help to determine precise relative risk estimates between handgrip strength and mortality risk in patients with different stages of CKD.

Keywords:

• Handgrip strength
• chronic kidney disease
• dose-response
• meta-analysis
• mortality

Introduction

Handgrip strength is a reliable proxy for individual upper limb muscle strength [1]. In chronic kidney disease (CKD) patients, handgrip strength has been reported to be lower than in healthy controls and to decline with worsening kidney function [2]. Handgrip strength has also been proposed as a potential marker of frailty and sarcopenia in CKD patients, which are common and severe complications [3]. Recently, there has been increasing evidence that increasing muscle strength may reduce the risk of all-cause mortality in CKD [4,5].

However, some studies differ in sample size, ethnicity, and other characteristics, leading to inconsistencies in their interpretation [6,7]. In addition, there are minimal reports on optimal handgrip strength levels for preventing all-cause mortality in CKD despite the availability of reference handgrip strength values for different populations.

Therefore, we hypothesized that the relationship between handgrip strength and the risk of all-cause mortality in CKD is not purely linear. This study is a further dose-response meta-analysis of our published systematic review [8] to obtain quantitative estimates of the association between handgrip strength and the risk of all-cause mortality in CKD to determine the ‘optimal dose’, aim to explore the potential clinical implications of using handgrip strength as a prognostic and therapeutic tool in CKD management.

Methodology

This secondary systematic review and dose-response meta-­analysis was reported according to the Meta-analysis of Observational Studies in Epidemiology (MOOSE) [9] and Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) checklists (Table S1) [10], and the original systematic review was registered (PROSPERO: CRD42022308810).

Search strategy

Initially, we systematically searched PubMed, Web of Science, and Embase databases from inception to July 2022, updated in October 2023 [8]. The specific search strategy is shown in the Tables S2–S4. Moreover, we reviewed conference abstracts from the International Society of Nephrology (https://www.theisn.org/), European Renal Association (https://www.era-online.org/en/), and American Society of Nephrology (https://www.asn-online.org/) for the past five years to look for other potential studies. We also searched ClinicalTrials.gov for studies registered as completed but not yet published. Finally, we manually investigated a relevant reference of systematic reviews to search for additional potential studies.

Eligibility criteria

The following articles were considered eligible for inclusion: (1) patients: adult patients with a confirmed diagnosis of CKD, including pre-dialysis, peritoneal dialysis, hemodialysis, and kidney transplant recipients; (2) exposures: handgrip strength measured using a dynamometer and reported at least two handgrip strength categories; (3) outcomes: all-cause mortality; (4) study design: cohort studies, both prospective and retrospective; and (5) full-text or non-full-text (e.g., conference abstracts and research letters), published in English with sufficient analyzable data. Studies were excluded if they (1) did not provide sufficient data, if the authors were contacted and did not respond, and (2) were case-control studies, cross-sectional studies, and randomized controlled trials. In the dose-response meta-analysis, homogeneity of measurement units was required; therefore, studies that could not isolate the effect of handgrip strength or studies reporting a mortality risk per 1 unit of elevated handgrip strength were excluded. Reviews, as well as animal studies, were not considered.

Quality assessment

The quality of the included studies was assessed according to the Newcastle-Ottawa Scale (NOS) [11]. Two independent authors (LYH and YB) completed the quality assessment, and a third author (FZ) was consulted if disagreement was encountered.

Literature screening and data extraction

Two independent authors (LYH and YB) first read the title and abstract and, after excluding irrelevant literature, further read the full text to determine inclusion. Data extraction was done independently by two reviewers (LYH and YB). In disagreement, the third author could be consulted (FZ). Characteristics extracted from each study included first author, population country, disease stage, age, number of participants and cases, handgrip strength thresholds, reported effect estimates, and corresponding 95% confidence intervals (95% CI).

For the dose-response meta-analysis, the required data were re-extracted by two authors and reviewed by a third author. For the open-ended highest category of handgrip strength (i.e., >X kg) or the lowest category (i.e., <X kg), we obtained doses by *1.5 or/1.5 [12]. When a study reported a risk estimate relative to a reference category other than the highest category of handgrip strength, we performed centering to make it the lowest category as a reference. Handgrip strength was used to describe effect sizes, and studies that reported a risk ratio was considered to approximate the hazard ratio (HR) [13].

Since most studies used handgrip strength thresholds based on gender classification, we used their mean values as available thresholds. For example, Isoyama et al. [14] reported low muscle strength, defined as handgrip strength below 30 kg for men and 20 kg for women, and then we used 25 as the shared threshold. Where handgrip strength intervals were not reported, they were estimated according to the method of Lauretani et al. [15].

Statistical analysis

A dose-response relationship was established by treating handgrip strength as an independent continuous variable and regressing it against all-cause mortality. Considering the dose-response relationship could likely be a non-linear pattern, we used the restricted cubic spline function to fit the dose-response curve [16]. This was done by setting three fixed knots at the 5th, 50th, and 95th quartiles of handgrip strength distribution (By setting three random knots on the quartiles of handgrip strength distribution) [17]. The Wald test was used to test the non-linearity with the null hypothesis, as the coefficients of the non-linear terms were equal to zero. The robust-error meta-regression (REMR) model synthesized the dose-response relationship across available studies [18]. This is a one-stage procedure that treats each study as a cluster and weighs the effects of each study by its inverse variance while employing the robust variance to address the potential correlation of within-study impacts. In addition, the robust variance has been proven to be a good solution in case of outliers [18]. Results are expressed as an HR and a 95% confidence interval (95% CI). We used funnel plots to assess the risk of publication bias and performed Egger’s tests for the examined. All statistical analyses were performed using Stata 17.0, and p values were two-sided with a significance level of .05.

Quality of evidence

The quality of evidence for the dose-response association of handgrip strength with all-cause mortality was assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) deterministic ratings [19]. Quality of evidence was rated as high, moderate, low, or very low based on risk of bias, inconsistency, indirectness, imprecision, and publication bias.

Results

Based on the screening results of the original published articles [8], we reviewed the newly retrieved literature and read the potential full text. Therefore, 18 articles [6,14,20–35] were included in the analysis (Figure 1). Table 1 lists the main characteristics and ‘dose’ information of the included studies. The methodological assessment of the included articles is shown in Table S5.

Table 1. Characteristics of included studies.

Author (year)	Country	Stage of CKD	Mean age at baseline ^†	Death	Total sample size	HGS cutoff	Interval	Dose	HR	Lower 95% CI	Upper 95% CI
Lopes et al. [26] (2012)	Brazil	HD	47.0 ± 14.2	NP	493	<18.5 kg for women and <28.8 kg for men.	>23.65	15.77	2	1.25	3.23
							> =23.65	35.48	1	1	1
Roshanravan et al. [30] (2013)	USA	Predialysis	61 ± 13	39	381	The predicted value is based on age and sex in the general population.	<25	16.67	1.3	0.71	2.37
							> =25	37.5	1	1	1
Isoyama et al. [14] (2014)	Sweden	Dialysis	52.0 ± 26.7	95	330	NP	<25	16.67	1.79	1.09	2.94
							> =25	37.5	1	1	1
Matos et al. [28] (2014)	Brazil	HD	46.61 ± 14.09	102	443	28.3 kg for males.	<28.3	18.87	3.35	1.67	6.73
							> =28.3	42.45	1	1	1
						23.4 kg for females.	<23.4	15.6	1.1	0.49	2.46
							> =23.4	35.1	1	1	1
Chiou et al. [21] (2014)	China	HD	60 ± 11	34	188	26 kg for male and 17 kg for female.	<21.5	14.33	1	1	1
							> =21.5	32.25	0.891	0.838	0.948
Dantas et al. [22] (2014)	Brazil	HD	49.3 ± 15.0	NP	176	18 kg for women <60 y, and 30 kg for men <60 y	<24	16	3.35	1.43	7.41
							> =24	36	1	1	1
						13 kg for women ≥60 y, and 22.7 kg for men ≥60 y.	<24.35	16.23	2.37	0.94	5.99
							> =24.35	36.53	1	1	1
Peng et al. [29] (2015)	China	HD	57.6 ± 13.0	35	108	≤36 kg in males and ≤19 kg in females.	< =27.5	18.33	2.505	1.112	5.642
							>27.5	41.25	1	1	1
Vogt et al. [32] (2016)	Brazil	Dialysis	58 ± 15.2	53	265	22.5 kg for male and 7 kg for female.	< =14.75	9.83	1	1	1
							>14.75	22.13	0.49	0.27	0.89
Kittiskulnam et al. [35] (2017)	USA	HD	56.7 ± 14.5	78	645	< 26 kg for males and < 16 kg for females.	<21	14	1.68	1.01	2.79
							> =21	31.5	1	1	1
Rodrigues et al. [34] (2017)	Brazil	HD	71 ± 7	61	173	<20 kg for women and <30 kg for men.	<25	16.67	1.98	1.12	3.48
							> =25	37.5	1	1	1
Giglio et al. [6] (2018)	Brazil	HD	70.6 ± 7.2	48	170	NP	<25	16.67	1.84	0.92	3.68
							> =25	37.5	1	1	1
Lopes et al. [27] (2018)	Brazil	HD	50.4 ± 15.0	121	413	17.8 kg for women <60 years, 29.5 kg for men <60 years	<23.65	15.77	3.6	2.07	6.55
							> =23.65	35.48	1	1	1
						13.8 kg for women ≥60years, and 21.9 kg for men ≥60years.	<17.85	11.9	2.28	1.25	4.28
							> =17.85	26.78	1	1	1
Chan et al. [20] (2019)	UK	KTR	49.6 ± 15	NP	128	NP	<25	16.67	2.45	1.3	4.64
							> =25	37.5	1	1	1
Kim et al. [24] (2019)	Korea	HD	59.8 ± 13.1	28	142	NP	<25	16.67	5.65	1.99	16.04
							> =25	37.5	1	1	1
Valenzuela et al. [31] (2019)	Spain	Dialysis	71 ± 9	12	30	<24 kg.	<24	16	3	1.01	8.95
							> =24	36	1	1	1
Lee et al. [25] (2020)	Korea	HD	61.7 ± 13.0	26	199	<26 kg for men and <18 kg for women.	<22	14.67	1.15	0.52	2.54
							> =22	33	1	1	1
Xu et al. [36] (2020)	China	PD	56.1 ± 15.0	144	327	<24.5 kg in males and <14 kg in females.	<19.25	12.83	3.37	2.37	4.8
							> =19.25	28.88	1	1	1
Do et al. [33] (2023)	Korea	PD	54.5 ± 15.8	26	199	<28 kg for men and < 18 kg for women (NH-NO versus LH-NO)	<23	15.33	1	1	1
							> =23	34.5	0.12	0.01	0.93
						NH-O versus LH-O	<23	15.33	1	1	1
							> =23	34.5	0.12	0.02	0.98

Abbreviations: HGS: handgrip strength; NP: not report; HD: hemodialysis; PD: peritoneal dialysis; KTR: kidney transplant recipient; UK: the United Kingdom; USA: the United States; NH-NO: patients with normal handgrip strength and without obesity; LH-NO: patients with low handgrip strength and without obesity; NH-O: patients with normal handgrip strength and with obesity; LH-O: patients with low handgrip strength and with obesity.

Where HGS intervals were not reported, it was estimated according to the method of Lauretani et al. [15].

A total of 4810 patients with CKD were included in the dose-response meta-analysis. One, two, 11, and one studies had pre-dialysis, peritoneal dialysis, hemodialysis, and kidney transplant recipients, respectively (Table 1). Two other studies recruited dialysis patients. The mean age of the participants was between 46.61 and 71.

There was a clear U-shaped relationship between handgrip strength and all-cause mortality risk in CKD with a significant non-linear association (Figure 2). Higher handgrip strength values were associated with a lower risk of death from 10 kg to approximately 28 kg. After that, the risk of death increased slightly. In addition, we noted that handgrip strength between 18 and 40 kg was associated with a low risk of all-cause mortality. Exceeding 40 kg appears to be associated with an increased mortality risk. The HR and 95% CI for each ‘dose’ are available in Table S6.

Regarding the assessment of publication bias, severe asymmetry was found in our funnel plot (Figure S1). Egger’s test also confirmed a significant risk of publication bias (Figure S2). Therefore, the quality of evidence was rated as very low due to inconsistency and publication bias (Table S7).

Discussion

Our dose-response meta-analysis allowed the estimation of handgrip strength values associated with a reduced risk of all-cause mortality in CKD. This finding is particularly interesting as it provides a range of reference values for preventing all-cause mortality in patients with CKD, which may be particularly useful in clinical settings and inform future guidelines and public health recommendations.

Low handgrip strength is a reliable marker for morbidity and mortality outcomes in chronic diseases, including CKD. Multiple findings report a strong association between lower handgrip strength and premature all-cause and cardiovascular mortality [36–39]. Similarly, our prior meta-analysis observed a consistent inverse relationship between higher upper extremity muscle strength and all-cause mortality [8]. A recent cohort study evaluating the association between handgrip strength index and mortality in kidney transplant recipients yielded similar results [40]. Despite the availability of handgrip strength reference values for different populations, there are few reports on the optimal handgrip strength for preventing all-cause mortality in CKD. Most published studies assume a linear relationship between handgrip strength and mortality outcomes, although some evidence suggests otherwise.

The evidence for dose-response relationship between handgrip strength and all-cause mortality risk in CKD was analyzed with low certainty due to high heterogeneity and lack of precision. Rücker et al. [41] proposed three sources of heterogeneity in meta-analyses, including clinical heterogeneity (e.g., differences between sample characteristics), statistical heterogeneity, and other sources of heterogeneity (e.g., design-related heterogeneity). Testing individual handgrip strength using different assessment tools was one of the sources that led to the observed heterogeneity. On the other hand, there were differences in disease stage of participants recruited into the included studies.

Age, baseline nutritional status, and physical activity level are key variables that may influence handgrip strength [42–44]. In addition, previous studies have shown that dialysis treatment leads to muscle atrophy, which reduces handgrip strength [45], whereas transplantation may have beneficial effects on muscle strength in patients with end-stage renal disease [46]. These characteristics may have various implications for heterogeneous samples, thus making the external validity of this meta-analysis challenging for generalizability.

The main novelty of this study is estimating the range of handgrip strength associated with a low risk of all-cause mortality in CKD. Notably, we found that the relationship between handgrip strength and all-cause mortality in CKD showed a U-shaped dose-response relationship rather than a higher handgrip strength. Importantly, handgrip strength cannot be assumed to be a proxy for overall muscle strength [47], and an excessively high handgrip strength may suggest that patients have other health problems, such as obesity, which may explain this observation.

It is well known that because handgrip strength decreases with declining renal function in pre-dialysis CKD patients, patients with end-stage renal disease gain some restoration of physical function, including handgrip strength, after receiving a kidney transplant. However, a significant decrease in muscle strength due to rapid loss of muscle protein in dialysis-dependent CKD patients may be sufficient to reduce their risk of mortality even if it is increased modestly or maintained at acceptable levels [48]. Thus, treatment at different CKD stages, lifestyle-related factors, and how it affects muscle strength may explain a broader range of handgrip strength thresholds. Future research is needed to elucidate the optimal levels of handgrip strength values for different periods of CKD to reduce the mortality risk associated with different CKD stages.

A recent study of the beneficial association between handgrip strength and cancer mortality did not find a minimum threshold, i.e., a low level of handgrip strength (i.e., a threshold of 16 kg for women and 22 kg for men), to be beneficial in reducing mortality risk in cancer patients [38]. In contrast, our findings on all-cause mortality showed a U-shaped association, and because kidney disease accelerates the loss of muscle strength, implementing targeted strategies to increase handgrip strength or maintain it at acceptable levels in patients with CKD may be able to reduce their risk of death. However, as noted above, the relationship between handgrip strength and all-cause mortality was not the same for participants with different baseline characteristics, and large, long-follow-up cohorts will be needed in the future to clarify the optimal level of handgrip strength.

Strength and limitations

In this review, we found 18 articles (22 studies) that collected available evidence on the dose-response association between handgrip strength and the risk of all-cause mortality in CKD. Limitations of this study include, first, that the handgrip strength thresholds determined for all-cause mortality in CKD are limited by the estimates obtained from the included studies, and individuals with higher handgrip strength than observed in this study may also benefit from the risk of all-cause mortality. A related limitation is that at the higher end of the exposure spectrum, the rise in the dose-response curve may represent a lack of data rather than an actual lack of association. The inversion of the right part of the dose-response curve in this study may reflect a scarcity of data/events rather than a genuine lack of beneficial association at higher levels of handgrip strength. Second, low handgrip strength thresholds range widely between men and women, and direct averaging may introduce some degree of measurement bias into the results. Also, due to the few studies that reported effects for gender separately, subgroup analyses could not be performed in this study to clarify handgrip strength thresholds between men and women. Nonetheless, the credibility of our analysis is moderate due to the inclusion of studies with a high number of participants and high quality. Third, because the participants in this study were predominantly hemodialysis-dependent patients, it is impossible to generalize it to full-spectrum CKD patients, as noted above. Fourth, although we used estimates from the most adjusted models in each cohort, this does not necessarily imply that choosing covariates was appropriate [1]. Sixth, a certain degree of measurement bias may exist due to the significant heterogeneity in the follow-up period and different covariate-adjusted lists across studies. On the other hand, we reduced the quality of evidence by potential publication bias. Therefore, the certainty of evidence from existing studies is very low and has limited ability to inform clinical care or policy, and future studies may provide different results.

Despite these limitations, our study provides some guidance for clinical practice. Diminished handgrip strength is a significant predictor of mortality risk in patients with CKD. Health caregivers can assess muscle strength in patients with CKD by the simple method of measuring handgrip strength and take appropriate measures, such as recommending proper exercise and nutritional support, so that this population’s handgrip strength can be improved or maintained within a specific range to improve overall health and reduce the mortality risk.

Conclusion

In patients with CKD, handgrip strength is associated with all-cause mortality in a U-shaped curve. Until further studies determine the optimal handgrip strength ‘dose’, handgrip strength management in CKD patients should be enhanced and maintained at least above 18 kg and possibly limited to less than 40 kg. This study provides two important directions for future research, including (1) determining the optimal range of handgrip strength for different CKD stages and (2) reporting the effect of handgrip strength and mortality risk between genders with the data obtained.

Author contributions

Study conception and design, data acquisition: FZ and YFZ; data interpretation and statistical analysis: FZ, YB, and LYH; manuscript writing: HC, FZ and YB; revision of the manuscript: HC, YL, and YFZ.

Acknowledgments

We thank Prof. Chang Xu for selflessly disclosing the Stata code of REMR.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Figure 1. Flowchart of study selection for the analysis. AKI acute kidney injury.

Figure 2. Dose-response plots of handgrip strength and risk of all-cause mortality.

Additional information

Funding

This study was supported by Shanghai Hospital Development Center [SHDC2022CRD003], Shanghai University of Traditional Chinese Medicine [602059D], Shanghai Xuhui District Health Care Committee [XHLHGG202105], National Natural Science Foundation of China [82374589], Demonstration-oriented Research Ward Construction Project of Shanghai Hospital Development Center [SHDC2022CRW006], and Shanghai Municipal Health Commission [20224Y0166].