Cognitive Impairment and Postoperative Outcomes in Patients Undergoing Primary Total Hip Arthroplasty and Primary Total Knee Arthroplasty: A Systematic Review and Meta-analysis

Abstract

Introduction:

Total hip arthroplasty (THA) and total knee arthroplasty (TKA) are commonly performed procedures in older adults. Preoperative cognitive impairment, including mild cognitive impairment, dementia, and frailty-related cognitive decline, may adversely affect postoperative recovery and outcomes. However, the magnitude of this association in arthroplasty patients remains incompletely defined.

Objective:

To evaluate the association between preoperative cognitive impairment and postoperative outcomes in older adults undergoing primary THA or TKA.

Methods:

A systematic review and meta-analysis was conducted in accordance with PRISMA guidelines. PubMed and Scopus were searched for studies evaluating preoperative cognitive status and postoperative outcomes in patients aged ≥60 years undergoing THA or TKA. Eligible studies assessed cognition using validated tools or clinical diagnoses and reported postoperative neurocognitive or clinical outcomes. A random-effects meta-analysis using the Restricted Maximum Likelihood estimator was performed. Pooled effect estimates were expressed as odds ratios (ORs) with 95% confidence intervals (CIs).

Findings:

Ten studies involving 10,573 patients were included. Preoperative cognitive impairment was consistently associated with adverse postoperative outcomes, including postoperative delirium, postoperative cognitive dysfunction, long-term cognitive decline, nonhome discharge, and mortality. The pooled analysis demonstrated a significant association between preoperative cognitive impairment and adverse postoperative outcomes (OR 2.38; 95% CI 1.78–2.98; P < 0.001). Although substantial heterogeneity was observed (I² = 76.7%), the direction of effect was consistent across studies.

Conclusion:

Preoperative cognitive impairment more than doubles the risk of adverse postoperative outcomes following THA and TKA. Routine cognitive screening should be incorporated into preoperative assessment to identify high-risk patients and support perioperative brain health strategies aimed at improving surgical outcomes.

INTRODUCTION

Total hip arthroplasty (THA) and total knee arthroplasty (TKA) are among the most commonly performed elective procedures in older adults, providing significant improvements in pain relief and functional status. With increasing global life expectancy, a larger proportion of surgical candidates exhibit age-related cognitive vulnerabilities, such as mild cognitive impairment (MCI), dementia, and frailty-associated decline. Cognitive impairment is highly prevalent in elderly populations and may affect perioperative resilience, recovery trajectories, and long-term outcomes.

Multiple studies in orthopedic populations indicate that preoperative cognitive impairment increases the risk of postoperative delirium (POD), prolonged hospitalisation, and discharge to institutional care following TKA.[1] Comparable associations have been observed in THA, where cognitive impairment correlates with higher rates of in-hospital complications, nonroutine discharge, and increased mortality.[2] The broader surgical literature supports these findings. A large systematic review and meta-analysis found that preoperative cognitive impairment significantly elevates the risk of POD, 1-year mortality, readmission, and discharge to assisted-living facilities among older surgical patients.[3] Furthermore, impaired cognition independently predicts postoperative complications, institutionalization, and mortality in elderly individuals undergoing major elective surgery.[4]

Reduced cognitive reserve is considered central to this vulnerability. Preexisting cognitive impairment is associated with increased susceptibility to both POD and postoperative cognitive dysfunction (POCD), with evidence indicating an additive or compounding relationship between these conditions.[5] Postoperative neurocognitive disorders have significant long-term implications. The International Study of POCD (ISPOCD) study reported early cognitive dysfunction in 25.8% of elderly patients and persistent dysfunction at 3 months in nearly 10%.[6] Subsequent research has demonstrated that POCD is associated with increased mortality and long-term socioeconomic consequences,[7] while POD has been linked to the later development of MCI or dementia in individuals who were previously cognitively intact.[8]

Despite the expanding evidence base, significant gaps persist. Much of the existing literature combines heterogeneous surgical populations, which limits the applicability of findings to elective THA and TKA. Many reviews include mixed procedures or focus exclusively on delirium, without considering broader neurocognitive and clinical outcomes. In addition, the magnitude and consistency of the association between preoperative cognitive impairment and adverse postoperative outcomes in total joint arthroplasty are inconsistently reported across studies.[1,2,9] Therefore, a focused quantitative synthesis in this population is warranted.

Accordingly, a systematic review and meta-analysis were conducted to determine whether preoperative cognitive impairment independently increases the risk of adverse postoperative outcomes, including delirium, POCD, neurocognitive decline, nonhome discharge, and mortality, among adults aged 60 years or older undergoing primary THA or TKA.

MATERIALS AND METHODS

Study design

This study was conducted as a systematic review and meta-analysis to evaluate the association between preoperative cognitive impairment and postoperative neurocognitive and adverse outcomes in elderly orthopedic surgical patients. The review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

Eligibility criteria

Studies were considered eligible for inclusion if they enrolled adult patients aged 60 years or older undergoing THA or TKA. Eligible studies were required to assess preoperative cognitive status using validated instruments, including the Montreal Cognitive Assessment (MoCA), Mini-Mental State Examination (MMSE), Mini-Cog, clinical diagnoses of MCI or dementia, or frailty indices incorporating cognitive domains. In addition, studies had to report relevant postoperative outcomes such as POCD, POD, neurocognitive decline, discharge disposition, or mortality. Only studies providing effect estimates (odds ratios (ORs) or relative risks [RRs]) or sufficient data for calculation were included.

Systematic reviews, narrative reviews, editorials, case reports, and studies lacking extractable outcome data were excluded.

Search strategy

A structured literature search was conducted in electronic databases including PubMed and Scopus. The search included combinations of the following keywords:

“preoperative cognitive impairment,” “mild cognitive impairment,” “dementia,” “frailty,” “postoperative cognitive dysfunction,” “postoperative delirium,” “arthroplasty,” “total hip arthroplasty,” and “total knee arthroplasty.”

Reference lists of eligible studies were also screened to identify additional relevant publications. The final selection included 10 studies that satisfied eligibility criteria.

Study selection

The study selection process adhered to PRISMA 2020 guidelines and is presented in Figure 1. A total of 1,264 records were identified through database searching. After removal of 312 duplicate records, 952 unique records remained for title and abstract screening, demonstrating appropriate separation of duplicate removal from the screening phase.

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 flow diagram of study selection process

During title and abstract screening, 863 records were excluded for not meeting the predefined inclusion criteria, including studies unrelated to orthopedic surgery, nonelderly populations, review articles, editorials, and studies lacking assessment of preoperative cognitive status. This resulted in 79 reports being deemed potentially eligible and sought for full-text retrieval. All 79 reports were successfully retrieved and assessed for eligibility, with no reports unavailable for review.

Following full-text evaluation, 69 studies were excluded. The reasons for exclusion were the absence of extractable outcome data (n = 21), lack of preoperative cognitive assessment (n = 18), inappropriate study population (n = 14), nonorthopedic surgical procedures (n = 9), and duplicate cohort reporting (n = 7). Ultimately, 10 studies met all predefined inclusion criteria and were included in both the qualitative synthesis and quantitative meta-analysis.

Data extraction

Data extracted from each study included author, year of publication, study design, sample size, type of surgery, definition and method of assessment of preoperative cognitive impairment, postoperative outcomes assessed, and reported effect estimates with confidence intervals (CIs). When RRs were reported, they were treated as approximations of ORs for pooling purposes.

Outcome measures

The primary outcome was the occurrence of postoperative neurocognitive disorders, including POCD and POD. Secondary outcomes included long-term cognitive decline, discharge disposition, hospital length of stay, readmission rates, and 90-day mortality.

Quality assessment

Given that most included studies were observational in design, methodological quality was assessed with consideration of selection bias, confounding adjustment, outcome measurement validity, and completeness of follow-up. The randomized controlled trial was evaluated for allocation concealment and blinding.

Statistical analysis

Meta-analysis was conducted using a random-effects model with the Restricted Maximum Likelihood (REML) estimator to account for between-study heterogeneity. Effect sizes were log-transformed before pooling. The pooled estimate was expressed as ORs with 95% CIs.

Between-study heterogeneity was assessed using the Cochran Q test and quantified using the I² statistic. I² values >50% were considered indicative of substantial heterogeneity.

Publication bias was assessed using funnel plot inspection, Kendall’s rank correlation test, Egger’s regression test, and fail-safe N analysis.

Equivalence testing using two one-sided tests (TOST) was also performed to determine whether the pooled effect was statistically different from predefined equivalence margins.

Ethical considerations

As this study was a systematic review and meta-analysis of previously published data, ethical approval and patient consent were not required.

RESULTS

Study characteristics

Ten studies were included in the quantitative synthesis, comprising approximately 10,573 patients undergoing predominantly TKA and THA. Study designs included prospective cohort studies, retrospective cohort studies, one population-based cohort, and one randomized controlled trial. The majority of studies evaluated elderly surgical populations aged 65 years and older.

Preoperative cognitive vulnerability was assessed using validated instruments including the MoCA, MMSE, Mini-Cog, formal clinical diagnoses of MCI or dementia, and frailty indices such as the Reported Edmonton Frail Scale. Outcomes evaluated across studies included POCD, POD, neurocognitive decline at short- and mid-term follow-up, discharge disposition, and 90-day mortality.

Overall, definitions of exposure and outcome varied across studies; however, all directly examined the prognostic significance of baseline cognitive status in relation to postoperative complications. The included studies are described in Table 1.

Table 1.

Characteristics of studies included (n=10)

Author (year)	Study design	Surgery type	Sample size (n)	Preoperative cognitive assessment	Primary postoperative outcome	Reported effect estimate
Ren (2022)[10]	Retrospective cohort	TKA	105	MoCA	POCD	OR 2.85
Silbert (2015)[11]	Prospective cohort	THA	300	Neuropsychological battery	POCD (7 days, 3 months)	RR 2.10
Tang (2024)[12]	Population-based cohort	THA/TKA	1,250	Cognitive trajectory analysis	Long-term cognitive decline	RR 1.15
Culley (2017)[13]	Prospective cohort	THA/TKA	211	Mini-cog	Postoperative delirium	OR 4.52
Silverstein (2007)[14]	Prospective cohort	Mixed orthopedic	1,218	Neuropsychological testing	Early POCD	RR 1.90
Blair (2021)[15]	Retrospective cohort	Mixed surgery (elderly)	6,590	Clinical MCI/dementia diagnosis	90-day mortality	RR 1.62
Takazawa (2022)[16]	Randomized controlled trial	TKA	185	Baseline cognitive assessment	POCD (intervention study)	OR 2.22
Krishnan (2021)[17]	Retrospective matched cohort	TKA	208	Cognitive impairment status	Postoperative delirium	OR 3.40
Evered (2020)[18]	Cohort study	THA	300	Frailty index (REFS)	Neurocognitive decline	OR 2.00
Wang (2022)[19]	Prospective cohort	TKA	206	MCI (Winblad criteria)	Postoperative delirium	RR 2.62

REFS=Reported Edmonton Frail Scale, MCI=Mild cognitive impairment, TKA=Total knee arthroplasty, THA=Total hip arthroplasty, POCD=Postoperative cognitive dysfunction, MoCA=Montreal cognitive assessment, RR=Relative risk, OR=Odds ratio

INDIVIDUAL STUDY FINDINGS

Ren et al. evaluated 105 elderly patients undergoing TKA and reported that low preoperative MoCA scores were independently associated with POCD. Multivariate analysis identified low MoCA as a significant predictor of POCD severity, with an OR of 2.85. Perioperative hemoglobin drop, elevated C-reactive protein, and severe postoperative pain were also associated with cognitive decline; however, baseline cognitive impairment remained the strongest independent predictor.[10]

Silbert et al., in a prospective cohort of 300 patients undergoing THA, reported that 32% had preoperative cognitive impairment. POCD occurred in 17% at 7 days and 10% at 3 months postoperatively. Patients with preexisting cognitive impairment demonstrated significantly higher rates of POCD at both early and intermediate time points and were more likely to experience persistent cognitive decline at 12 months.[11]

Tang et al. conducted a population-based observational study of 1,250 patients undergoing THA or TKA and examined long-term cognitive trajectories. Baseline cognitive decline was associated with modest worsening in long-term postoperative cognitive outcomes (RR 1.15), suggesting that preoperative vulnerability may influence cognitive aging trajectories following surgery.[12]

Culley et al. studied 211 older orthopedic patients and found that 24% had probable cognitive impairment based on Mini-Cog screening. Cognitive impairment was strongly associated with POD (21% vs. 7%; OR 4.52, 95% CI 1.30–15.68), increased likelihood of discharge to nonhome facilities (67% vs. 34%; OR 3.88, 95% CI 1.58–9.55), and prolonged hospitalization.[13]

Silverstein et al. analyzed 1,218 surgical patients and identified 74 individuals with baseline cognitive impairment. At 1 week postoperatively, deterioration occurred predominantly in attention, executive function, and processing speed domains. Patients with preoperative impairment were particularly vulnerable to early POCD, especially in executive and psychomotor domains.[14]

Blair et al. (2021)[15] evaluated 6,590 Medicare beneficiaries aged ≥65 years undergoing surgery. Among them, 20.1% had MCI and 9.9% had dementia. While MCI was not associated with increased short-term mortality, dementia was associated with significantly higher 90-day mortality (8.4% vs. 5.2%, P = 0.002) and readmission rates (17.3% vs. 13.9%, P = 0.038), indicating a severity-dependent risk gradient.[15]

Takazawa et al. (2022)[16] conducted a randomized, double-blind trial involving 185 TKA patients to evaluate minocycline for POCD prevention. At 1 week, POCD incidence was 8.9% in the minocycline group and 4.2% in the placebo group (OR 2.22; 95% CI 0.64–7.65), with no statistically significant difference. At 3 months, incidence remained comparable between groups, suggesting that short-term pharmacologic intervention did not mitigate cognitive vulnerability-related risk.[16]

Krishnan et al. analyzed 208 TKA patients in a matched retrospective cohort and reported that preexisting cognitive impairment significantly increased POD risk and worsened functional status. Cognitively impaired patients were more likely to require nonhome discharge, although short-term mortality rates were low.[17]

Evered et al. examined 300 THA patients and assessed frailty as a marker of cognitive vulnerability. Baseline frailty was associated with cognitive decline at 3 months (OR 1.51, 95% CI 1.02–2.23) and 12 months (OR 2.00, 95% CI 1.26–3.17), supporting frailty as a predictor of postoperative neurocognitive disorders.[18]

Wang et al. evaluated 206 TKA patients and found that MCI significantly increased POD incidence (32.5% vs. 12.4% in non-MCI patients; overall POD 20.2%). MCI and elevated cerebrospinal fluid tau biomarkers were identified as significant risk factors, supporting a biological substrate underlying POD susceptibility.[19]

Across studies, preoperative cognitive impairment consistently demonstrated an adverse association with postoperative outcomes. The strongest and most consistent associations were observed for early POD and POCD, with several studies reporting two-to fourfold increase in risk. Long-term cognitive outcomes showed smaller but persistent associations. Dementia-level impairment was associated with increased mortality and readmission, suggesting a severity-dependent gradient of risk.

Collectively, these findings support the role of baseline cognitive vulnerability as a significant perioperative risk amplifier in elderly orthopedic surgical populations.

OVERALL POOLED EFFECT

A random-effects meta-analysis using the Restricted Maximum Likelihood (REML) estimator was performed on 10 studies incorporating approximately 10,573 patients. Effect estimates were standardized to ORs to ensure comparability across studies reporting ORs and RRs. The pooled analysis demonstrated a statistically significant association between preoperative cognitive impairment and adverse postoperative outcomes, with a pooled OR of 2.38 (95% CI 1.78–2.98). The corresponding Z statistics was 7.78, indicating a highly significant overall effect (P < 0.001).

This figure presents the forest plot of the 10 included studies evaluating the association between preoperative cognitive impairment and adverse postoperative outcomes. Each square represents the individual study effect estimate (OR or RR converted to OR), with the size of the square reflecting the study weight. Horizontal lines denote 95% CIs. The vertical dotted line indicates the line of no effect (OR = 1). The diamond at the bottom represents the pooled effect estimate from the Random Effect Model (REM), Restricted Maximum Likelihood (REML), showing a combined OR of 2.38 (95% CI 1.78–2.98). Most studies demonstrate effect estimates greater than unity, indicating increased postoperative risk among cognitively impaired patients [Figure 2].

Figure 2.

Forest plot of the association between preoperative cognitive impairment and adverse postoperative outcomes

The lower bound of the CI remained well above unity, confirming that the association is unlikely to be due to random variation. Clinically, this pooled estimate indicates that patients with baseline cognitive vulnerability had more than twice the odds of experiencing postoperative neurocognitive disorders, delirium, cognitive decline, or mortality compared with cognitively intact individuals.

BETWEEN-STUDY HETEROGENEITY

Substantial heterogeneity was observed across the included studies. The between-study variance (Tau²) was 0.693, and the standard deviation of true effects (Tau) was 0.832. The inconsistency statistic (I²) was 76.72%, indicating that approximately 77% of the variability in observed effect sizes was attributable to real differences between studies rather than sampling error. The Cochran Q statistic was 37.04 with 9 degrees of freedom, and this was statistically significant (P < 0.001), confirming heterogeneity beyond chance.

The magnitude of heterogeneity likely reflects the differences in definitions of cognitive impairment, outcome measurement methods, surgical populations, follow-up duration, and study design. Despite this variability, the direction of effect was consistently positive across the majority of included studies, suggesting that cognitive vulnerability uniformly confers increased postoperative risk even if the magnitude varies.

PRECISION AND STATISTICAL STRENGTH

The standard error of the pooled estimate was 0.305, reflecting reasonable precision given the large cumulative sample size. The 95% CI (1.78–2.98) spans a range that remains clinically meaningful throughout, with even the lower bound suggesting a 78% increase in odds of adverse postoperative outcomes. This degree of precision, coupled with the large Z value, supports the robustness of the overall association.

PUBLICATION BIAS ASSESSMENT

Visual inspection of the funnel plot [Figure 3] demonstrated asymmetry in the distribution of included studies. While several studies were distributed within the expected triangular region around the pooled effect size, a number of smaller studies appeared outside the pseudo 95% confidence limits, particularly on the right side of the plot. This pattern suggests the presence of small-study effects, with smaller studies tending to report larger effect sizes.

Figure 3.

Funnel plot assessing publication bias in the meta-analysis (n = 10)

Statistical testing supported this visual impression. Kendall’s rank correlation test indicated significant funnel plot asymmetry (P = 0.007), and Egger’s regression test was also statistically significant (P = 0.004). Despite this asymmetry, the fail-safe N was 890, indicating that a large number of unpublished null studies would be required to render the pooled effect non-significant.

Overall, while Figure 3 suggests possible publication bias or small-study effects, the magnitude and statistical robustness of the pooled association remain unlikely to be fully explained by publication bias alone.

This figure shows the funnel plot of the 10 included studies assessing the association between preoperative cognitive impairment and adverse postoperative outcomes. The vertical dashed line represents the pooled effect estimate from the random-effects model. The triangular region represents the pseudo 95% confidence limits. Each dot corresponds to an individual study plotted according to its effect size and standard error. Asymmetry in the plot suggests potential small-study effects or publication bias.

EQUIVALENCE TESTING

Equivalence testing was performed using predefined equivalence bounds of −0.5 to +0.5 to determine whether the pooled effect could be considered clinically trivial. As shown in Figure 4, the pooled effect size was 2.38, with a 95% CI of 1.78-2.98. The 90% CI used for the TOST procedure ranged from 1.87-2.88.

Figure 4.

Equivalence testing plot of the pooled effect estimate

The 95% CI did not cross zero, and the null hypothesis significance test was statistically significant (P < 0.001), confirming that the pooled effect differed significantly from no effect.

In contrast, the TOST procedure was nonsignificant for equivalence, as the 90% CI lay entirely outside the predefined equivalence region (−0.5 to + 0.5). As illustrated in Figure 4, the pooled estimate and its CIs are positioned well beyond the equivalence margins, indicating that the observed effect is not small enough to be considered clinically negligible.

Taken together, these findings demonstrate that the association between preoperative cognitive impairment and adverse postoperative outcomes is both statistically significant and clinically meaningful, rather than trivial in magnitude.

This figure presents the equivalence testing results for the pooled meta-analytic effect. The square represents the overall effect size (2.38), and the horizontal line indicates the 95% CI (1.78–2.98). The vertical dashed lines denote the predefined equivalence bounds (−0.5 and +0.5). The 90% CI used for the TOST (1.87–2.88) lies entirely outside the equivalence region, indicating nonequivalence. Simultaneously, the 95% CI does not cross zero, confirming statistical significance. These findings demonstrate that the pooled association is both statistically significant and clinically meaningful rather than trivial.

INFLUENCE AND MODEL STABILITY

Diagnostic analyses, including influence measures and leave-one-out assessments, did not indicate that any single study disproportionately influenced the pooled estimate. Although heterogeneity remained substantial, removal of individual studies did not eliminate statistical significance, suggesting that the association between preoperative cognitive impairment and adverse postoperative outcomes is stable across the included body of evidence.

The quantitative synthesis demonstrates a strong and statistically robust association between preoperative cognitive impairment and adverse postoperative outcomes. While substantial heterogeneity and evidence of small-study effects were present, the magnitude of association remained consistently elevated across studies. The pooled effect estimate of 2.38 represents a clinically meaningful increase in postoperative risk and supports the prognostic importance of baseline cognitive vulnerability in elderly surgical populations.

DISCUSSION

This review and meta-analysis combine the data from 10 studies, involving about 10,573 elderly patients who mostly underwent total knee or hip replacements. The combined results show that having cognitive impairment before surgery more than doubles the risk of negative outcomes after surgery (pooled OR 2.38, 95% CI 1.78–2.98, P < 0.001). Although there was considerable variation between studies (I² = 76.7%), all showed a similar trend. These results suggest that preoperative cognitive problems are a major risk factor for older patients undergoing orthopaedic procedures.

Our results match other research showing that many older patients have cognitive impairment before surgery. One cross-sectional study found that cognitive problems were common in elderly orthopedic patients and were linked to older age, less education, sleep issues, atherosclerosis, and high cholesterol, pointing to existing health vulnerabilities.[20] Another study found that 23.5% of older surgical patients had abnormal cognitive screening results, which were connected to less independence and more health problems.[21]

A systematic review and meta-analysis further demonstrated pooled prevalence rates of unrecognized cognitive impairment of 37% in elective noncardiac surgery and 50% in emergency procedures.[22] When over one-third of elderly surgical candidates harbour cognitive impairment, a twofold increase in postoperative risk carries substantial population-level implications.

The link we found between cognitive problems and delirium after surgery matches other large studies showing that cognitive impairment before surgery raises the risk of delirium and longer hospital stays in older patients.[23] Reviews focused on hip and knee replacements report that about 14.8% of patients develop delirium and 19.3% experience cognitive problems after surgery, with preexisting cognitive decline being a key risk factor.[24]

Long-term studies add weight to these findings. One large study showed that delirium after major surgery speeds up cognitive decline by about 40% over 6 years, suggesting that delirium can have lasting effects, not just short-term ones.[25] Earlier research from the ISPOCD1 study found that 25.8% of elderly patients had cognitive problems 1 week after major noncardiac surgery, and 9.9% still had issues after 3 months.[6] These results support the idea that the link we found is both biologically likely and consistent over time.

Our results about death rates and long-term outcomes also agree with earlier studies. One meta-analysis found that older patients with cognitive impairment before surgery had higher risks of delirium, death within a year, being sent to assisted care, being readmitted, and other complications. Dementia made the risk of death even higher.[3] Other long-term studies showed that having cognitive problems 3 months after surgery is linked to higher death rates and social impacts, like leaving work early and needing more help.[7] These findings support the idea that more severe cognitive problems are associated with worse outcomes.

Recent research on frailty adds to these findings. Studies using the modified frailty index show that patients with higher frailty are more likely to develop delirium and have slower cognitive recovery after planned joint replacement surgery.[26] Frailty seems to reflect lower physical and mental reserves, which helps explain why these patients are more vulnerable after surgery.

Expert consensus statements now frame perioperative brain health as a patient safety imperative and recommend experts now see protecting brain health around the time of surgery as a key patient safety issue. They recommend regular cognitive screening and team-based approaches to lower risks.[27] However, reviews show that these recommendations are not always implemented, highlighting the need for better brain health programs.[28] Our results give strong evidence for adding structured cognitive screening to standard care for joint replacement patients. Humeidan et al (2021) demonstrated that cognitive prehabilitation reduced POD incidence from 23.0% to 14.4% among older adults undergoing major noncardiac surgery.[29] Although not limited to arthroplasty populations, this finding supports early identification and targeted preventive strategies.

However, reviews of quick screening tools show that none have been fully validated to reliably detect MCI before surgery.[30] Hence, while our results support regular cognitive checks, better methods are still needed to make screening more accurate and practical.

Even though the studies included were quite different, removing any one study did not change the overall results, showing that the findings are stable. While the funnel plot suggested that smaller studies might have some effect, the large fail-safe N means that publication bias is unlikely to explain the association we found.

In summary, this meta-analysis shows that having cognitive impairment before surgery more than doubles the risk of problems after hip and knee replacements in older patients. Given how common unrecognized cognitive issues are, the known links to delirium and long-term decline, and new findings about frailty, these results strongly support adding structured cognitive checks to routine pre-surgery evaluations. Cognitive vulnerability should be seen as a key part of assessing and improving brain health around surgery, not just as a risk marker.

Limitations

This study has some limitations. Most of the included studies were observational, which means there could be confounding factors, and it is harder to draw firm conclusions about cause and effect. The definitions of cognitive impairment and postoperative outcomes were inconsistent across studies, resulting in substantial heterogeneity (I² = 76.7%). Differences in how cognition were measured, how long patients were followed, and the types of surgeries included may have affected the results. While tests for publication bias suggested that smaller studies might have influenced the findings, the large fail-safe N helps reduce this concern, although it does not eliminate it. Furthermore, RRs and ORs were combined for analysis, which could introduce some minor bias. Most studies were conducted in high-income countries, so the results may not apply to all healthcare settings.

FUTURE SCOPE

Future research should use consistent definitions of preoperative cognitive impairment and postoperative neurocognitive outcomes to facilitate easier comparisons across studies. Large, multicenter studies that use the same cognitive tests and follow patients for longer periods are needed to better understand the long-term outcomes. Randomized controlled trials should examine whether structured cognitive screening and targeted brain health programs help patients undergoing joint replacement surgery. To make this research more actionable, future studies could include specific cognitive prehabilitation elements such as memory and attention exercises, sleep optimization protocols, or medication review to reduce potentially inappropriate drugs. In addition, interventions like structured orientation sessions, caregiver involvement, or tailored physical activity programs could be tested for their impact on perioperative cognitive health. Adding biomarkers, frailty scores, and brain imaging could improve our understanding of the risks and help identify patients who need extra support. Studies should also look at the cost-effectiveness of routine cognitive screening and interventions. Including research from more countries will make the findings more relevant worldwide.

CONCLUSION

This review and meta-analysis show that older patients with cognitive impairment before surgery are much more likely to have problems after total hip or knee replacement. The combined results suggest their risk is more than twice as high, highlighting that cognitive issues are an important risk factor. Even though the studies varied, the overall trend was consistent, which supports the strength of this link. Since many patients have cognitive impairment that goes unnoticed, regular cognitive checks should be the part of presurgery evaluations. Using structured brain health strategies could help identify risks, guide care during surgery, and possibly improve recovery. Routine cognitive screening should now be considered standard pre-arthroplasty care. Clinicians are encouraged to screen, stratify, and intervene to protect brain health and optimize the outcomes for older adults undergoing joint replacement.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.