Introduction

Total hip arthroplasty (THA) is highly effective in alleviating pain, restoring function, and improving quality of life in patients with severe hip arthritis. However, approximately 23% of patients develop hip pain after THA [1,2,3,4]. Some may even experience no improvement or worse pain postoperatively [2]. Known causes of postoperative pain include loosening of implants, infection, periprosthetic fracture, and soft-tissue abnormalities [3]. However, many cases cannot be explained by radiographic or mechanical abnormalities.

Hip pain following THA is often exacerbated by using stairs, walking on uneven surfaces, sitting for extended periods of time, and standing from a seated position. Factors that may influence hip pain involve both modifiable and non-modifiable risk factors. Modifiable risk factors include body mass index (BMI), certain medical and psychological comorbidities, and select socioeconomic variables. Non-modifiable risk factors may cover age, sex, and race. Identification of these modifiable risk factors of increased postoperative pain is critical in guiding preoperative optimization. Understanding non-modifiable risk factors also has added value in determining surgical expectations, individualizing pain management, and guiding the informed decision-making process.

This systematic review aimed to determine preoperative risk factors that are associated with post-THA pain. Hernández et al. reviewed studies prior to 2013 for predictive factors in total knee and total hip replacements [5]. However, THA and TKA are fundamentally distinct surgeries with well-documented differences in outcomes [6]. To our knowledge, there has been no systematic evaluation of risk factors for postoperative pain specific to THA.

Methods

Search strategy

We developed a comprehensive search strategy according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines with the assistance of a trained librarian. A systematic search was conducted in the PubMed (Medline), Scopus, Web of Science, Cumulative Index of Nursing and Allied Health Literature Plus (EBSCO), Embase, and Cochrane databases by using title, abstract, keywords, and medical subject headings (MeSH). MeSH terms included arthroplasty, replacement, total hip, THA, follow-up, risk assessment, risk factors, reinforcing factors, predictors, pain, postoperative, chronic, long-term, and pain measurement. Title, abstract, and keyword search terms were "hip arthroplasty, THA, THR, hip replacement, OR total hip" AND "after, continue, post, recur, ongoing, chronic, persistent, OR long term" AND "pain" AND "risk, predict, factor, associated, correlate, effect, affect, OR influence". The search strategy was tailored to and optimized for each database.

Inclusion and exclusion criteria

Our review screened for prospective controlled trials, prospective and retrospective observational cohort studies, and case–control studies. The inclusion criteria were full-text articles published between January 1995 and August 2020 with postoperative pain outcomes in only adult THA cases for osteoarthritis with respect to preoperative risk factors. Studies of patients undergoing THA in combination with other large-joint replacement surgeries were included if data specific to THA were provided. The literature review was performed independently by three researchers. Search results were reviewed in order of title, abstract, then full text, and excluded if they did not meet all criteria. Additional relevant studies were identified through a manual search of the bibliographies of selected studies. All other studies that did not assess risk factors, pain, risk factors for pain, or specific THA data were excluded from this analysis.

Bias assessment

Criteria for bias are not well described for systematic reviews of observational studies. We performed a bias assessment for each article based on study design and characteristics. The bias assessment was based on (1) loss to follow-up of <20% before 1 year or <30% after 1 year; (2) consecutive patient selection; (3) multicenter recruitment; and (4) use of univariable or multivariable analysis to adjust for confounders. Studies were rated as high risk (meeting zero or one criterion), medium risk (meeting two criteria), or low risk (meeting three or more criteria). If it was unclear whether a study met a criterion, we assumed that it did not.

Data

Data extracted included study design, patient characteristics, preoperative risk factors, study duration, outcome measurements, postoperative pain outcomes, and criteria for bias assessment. Results for each factor were compared across the studies. Medical or psychological comorbidities, BMI, preoperative pain, and certain socioeconomic elements were considered modifiable factors; whereas, age, sex, and race were deemed as non-modifiable. A descriptive analysis was performed as the heterogeneity of outcome measures and study characteristics prohibited a meaningful meta-analysis.

Patient and public involvement

No patients were involved in determining the research question, outcome measures, or study design. There are no plans to involve patients in the dissemination of research findings.

Results

Study characteristics

A total of 11,729 studies were identified, including duplicates. The full-text review was performed for 339 publications, and 54 studies satisfied our inclusion criteria (Fig. 1). Most studies were observational cohort studies. The sample size ranged from 54 to 37,393 patients, and follow-up ranged from 24 h to 12 years. Other than in the studies by Busato et al. [7] and Röder et al. [8], THAs were performed in 1993 or later. Commonly evaluated risk factors for postoperative pain were preoperative pain and function (20 studies), medical and psychological comorbidities (18 studies), BMI (13 studies), socioeconomic status and ethnicity (12 studies), sex (11 studies), and radiographic severity of osteoarthritis (4 studies). Functional outcomes were assessed using the Western Ontario and McMaster Universities Arthritis Index (WOMAC), Oxford Hip Score (OHS), Harris Hip Score (HHS), and Hip Disability and Osteoarthritis Outcome Score (HOOS).

Fig. 1
figure 1

Flow diagram of a systematic review of articles published from January 1995 to May 2018 about adults undergoing total hip arthroplasty. The databases searched were PubMed (Medline), Scopus, Web of Science, Cumulative Index of Nursing and Allied Health Literature Plus (EBSCO), Embase, and Cochrane

Full size image

According to the bias criteria, 10 studies had a high risk of bias, 31 had a medium risk, and 13 had a low risk (Table 1). Most studies adjusted for confounders. However, because of the retrospective design of many studies, selection methods and follow-up rates often could not be assessed.

Table 1 Risk of bias among 53 studies of adult THA
Full size table

Preoperative BMI

We found that high BMI was associated with worse pain outcomes or increased opioid use after THA in 9 of 13 studies, although conclusions varied regarding its clinical importance (Table 2). Of the studies, 5 had a low risk of bias, 7 had a medium risk, and 1 had a high risk.

Table 2 Associations of BMI with pain/opioid use after THA
Full size table

Preoperative pain and hip function

We found that preoperative pain and function were significant predictors of pain or persistent opioid use after THA in 16 of 20 studies (Table 3). Of the studies, 2 had a low risk of bias, 14 had a medium risk, and 4 had a high risk.

Table 3 Associations of preoperative pain medication or opioid use and function with pain medication or opioid use after THA
Full size table

Age

We found an association between patient age and pain or opioid use after THA in 9 of 12 studies, but these results were controversial (Table 4). Of these studies, 3 showed older age to be predictive of persistent pain, whereas 5 found young age to be predictive. One study concluded that two opposite age ranges studied were associated with worse pain, and three studies found no association. Notably, 4 studies had a high risk of bias, 3 had a medium risk, and 5 had a low risk.

Table 4 Associations of patient age with pain/opioid use after THA
Full size table

Sex

Female gender was a predictor of worse pain or opioid use after THA in 9 of 11 studies (Table 5). There were no studies with a high risk of bias, and three had low risk.

Table 5 Associations of female sex with pain/opioid use after THA
Full size table

Radiographic severity of arthritis

Of the studies investigated, four examined the association between preoperative radiographic severity of arthritis and postoperative pain. We noted that 3 of these studies reported better pain outcomes in patients with severe arthritis, while one study found no predictive value (Table 6). Regarding risk of bias, 2 studies had a low risk and 2 had a medium risk.

Table 6 Associations of preoperative radiographic severity of arthritis with pain after THA
Full size table

Socioeconomic status and race/ethnicity

Associations of various socioeconomic parameters and race/ethnicity with pain or opioid use after THA were assessed in 12 studies (Table 7). Three studies had a high risk of bias, 6 had a medium risk, and 3 had a low risk. Of these studies, 4 found worse pain in African Americans at 2–3.5 years postoperatively [10, 19, 38, 42]. Reports on the educational level were mixed. Three studies reported that low socioeconomic status was a risk factor for poor pain outcomes after THA [31, 38, 40].

Table 7 Associations of race/ethnicity and SES with pain/opioid use after THA
Full size table

Preoperative comorbidities

Seventeen of 18 studies found a negative association between medical or psychological comorbidities and pain after THA or postoperative opioid use (Table 8). Of these, 2 articles had a high risk of bias, 11 had a medium risk, and 5 had a low risk. Psychological comorbidities were another frequently studied risk factor.

Table 8 Associations of preoperative medical/psychological comorbidities with pain/opioid use after THA
Full size table

Discussion

The most consistent association were found between poor pain outcomes and the female sex, high preoperative pain or low function, and various medical or psychiatric comorbidities. Females not only had worse pain at both short- and long-term follow-ups but also had higher odds of severe acute postoperative pain and long-term opioid use. Although this should not affect the patient selection, an effort should be made to optimize multimodal pain management in women to achieve better short-term pain control, decrease chronic pain, and minimize opioid dependence [59]. Postoperatively, multidisciplinary pain therapy has been shown to provide substantial pain relief and may be a valuable referral [60].

Preoperative pain and loss of function are two primary criteria for performing THA; therefore, they cannot be treated like other modifiable factors. A difference should be noted for patients using chronic pain medication, who are at higher risk of postoperative opioid dependence and may benefit from referral to a pain specialist preoperatively for intervention and a weaning protocol at the cost of delaying surgery [11, 16].

Patients with existing comorbidities are more reluctant to undergo elective surgeries [61]. In THA, they also experience increased complications requiring revisions [62]. Notably, patients with diabetes had eight times higher odds of having persistent pain than those without. Other medical comorbidities had a similar effect. Increased odds of acute and chronic pain, as well as opioid dependence, were repeatedly observed in patients with anxiety or depression. Our data suggest that preoperative medical optimization may be beneficial to pain outcomes. Additionally, treating those with active psychiatric conditions may also improve perceived pain and satisfaction [63].

Less consistent association was seen between poor pain outcomes and high BMI, low radiographic grade, and low socioeconomic status. The difference in pain outcomes among BMI groups was often small when compared with the overall improvement. Although most data show worse postoperative pain scores in patients with high BMI, this may be due to greater preoperative pain rather than a less surgical benefit. Nevertheless, weight loss may be beneficial in decreasing baseline pain, ultimately improving postoperative pain [64].

Advanced preoperative radiographic severity of osteoarthritis was mostly found to be associated with worse pain outcomes. Although the data were not robust, they support the current guidelines of attempting non-operative modalities for those with low-grade radiographic arthritis [65].

With respect to race and socioeconomic factors, most studies found worse pain outcomes in African Americans, immigrants, patients with low educational levels, and patients with low socioeconomic status. Additionally, African Americans and patients of lower socioeconomic status had worse pain on presentation. A lack of access to resources, along with health care disparities affecting these populations, likely contribute to a delayed presentation with greater pain from advanced disease or improper non-operative management [66]. Although it is encouraging that these patients achieved similar improvement from surgery as their counterparts, they may still benefit from attention to pain management and preoperative education.

The relationship between age and pain outcomes is less clear. Studies showed that older age could be associated with better, worse, or no difference in postoperative pain. Possible explanations for worse outcomes in younger patients include an increased level of activity and expectations [67]. Conversely, confounding comorbidities and poor recovery may contribute to persistent pain in older patients. Notably, two studies reported worse outcomes in older patients at 3- and 6-month follow-up, whereas, most studies at later time points showed no effect or the opposite. This may be attributed to decreased healing and rehabilitation potential in older patients who are less healthy and less active. Although results conflict, data exist showing both ends of the age spectrum having worse pain outcomes. We may consider holding off surgery in young patients due to worse pain outcomes, in addition to other complications such as early implant failure [68]. Older patients, if they meet surgical criteria, should not delay the operation, or the recovery and rehabilitation potential is diminished.

This is the first systematic review, to our knowledge, that assessed common preoperative risk factors for pain after THA. A strength of this study is that a large number of studies were included, most of which had a low or medium risk of bias. However, there are several limitations to consider. The majority of outcome-based studies are observational cohorts with varying quality and risk of bias. In retrospective studies, appropriate patient selection and loss of follow-up are difficult to assess. Additionally, not all studies adjusted for confounding factors. Most studies reported follow-up of < 2 years, and only two studies continued beyond 5 years. The effects, if any, that these risk factors may have beyond this time frame are impossible to evaluate. Additionally, it is apparent that there is no consensus on an outcome measure for postoperative pain. The differences in the various questionnaires may also be a source of bias. Although some studies reported the effect of risk factors to be small relative to overall improvement from surgery, the heterogeneity of study design and outcome measures prohibited a meaningful meta-analysis to determine the magnitude of effect for each predictive factor. Future outcomes research will benefit from standardized design and outcome measures that allow for meta-analysis and the production of stronger evidence.

Persistent or severe postoperative pain is often difficult to explain and remains a major detriment to overall patient satisfaction and recovery after THA. Identification and management of preoperative risk factors is crucial. Although age, sex, and certain socioeconomic elements cannot be altered, they provide value in the discussion of surgical benefits and patient expectations. Additionally, recognizing patients at higher risk of worse pain outcomes allows the provider to appropriately manage their pain. Over-prescription of opioids is becoming a dangerous epidemic, and more THA patients are relying on them by the year [11]. Referring patients at high risk of postoperative pain to specialists will provide safer and more reliable pain regimens. Finally, these risk factors hold value in surgical decision-making. Modifiable characteristics such as obesity, mental health, and medical comorbidities present the opportunity to improve pain outcomes with preoperative optimization.

Conclusion

We have identified preoperative risk factors that were consistently predictive of greater postoperative pain after THA, despite the varying quality of studies that prohibit reaching concrete conclusions. Modifiable factors should be optimized preoperatively, whereas non-modifiable factors may be valuable to patient education, shared decision-making, and individualized pain management.