Does total hip replacement affect sexual quality of life?

We searched electronically in PubMed, EMBASE and PsycINFO (by EBSCO). We also performed a hand search of reference lists of included articles to identify additional relevant studies. The search strategy was developed in collaboration with a medical database specialist (EPJ). The PubMed search strategy, which can be found in Appendix 1, was adapted for the other databases. The search included articles from January 1970 until February 9th, 2015. The searches included MeSH terms in PubMed, EMtree in EMBASE, thesaurus terms in PsycINFO as well as free text terms. Search terms expressing “total hip” and “osteoarthritis” were used in combination with “sexual quality of life” and “THR” treatment. Search results were imported to a reference manager (Mendeley), and duplicates were removed.

Titles and abstracts were screened using the following eligibility criteria:

Two reviewers (RH and EPJ) independently applied the eligibility criteria to the titles and abstracts. Where there was uncertainty about eligibility, the full text was examined. Titles and abstracts that were identified as potentially eligible were selected for full-article review. The two reviewers independently screened the full-text articles for final study inclusion. Disagreements would have been resolved by a third author (TH), but this did not occur.

There are no other data found in supplementary files. Al data that support our findings are contained within this manuscript.

Two reviewers (RH and IS) extracted data from the included studies, independently and into pre-determined forms, and included the patients’ demographics, such as disease characteristics, study aims and information about study designs (e.g. sample size, response rate, ages, gender, duration of follow-up, and analysis methods). SQoL outcomes were subsequently extracted and categorised into two dimensions of SQoL: physical-functional well-being and psychosocial well-being. We summarised the data into outcomes quantifying SQoL before and after surgery, as a result of surgery (changes in SQoL) and as postoperative (cross-sectional) outcomes.

Two reviewers (RH and IS) independently scored the methodological quality of the included studies. Quality was assessed by using 17 of the 23-items quality checklist previously employed by Schouffoer [17] and Tilbury [18]. This checklist is based on Hayden [19] and Shamlyan [20] and is divided into 3 categories: selection bias (items 1–6), information bias (items 7–14), and statistical analysis bias (items 15–17). This quality checklist can be found in Appendix 2. Items concerning multiple determinants were not included.

Risk of bias was considered to be present if one or more of the items within one category were scored as “unclear”, “negative” or “not described”. When the study represented “high” quality on all items per category, the quality was rated as “0” (absence of risk of bias). When risk of bias was present, or items were not completely or not clearly described, it was rated as “1”. The quality of the study was rated as “high” if there was no risk of selection bias, information bias, or statistical analysis bias. The quality was rated as “moderate” if there was evidence of risk of bias in one of the three categories, and as “low” if there was risk of bias in two or all categories.

We planned to statistically pool data from studies that were clinically and methodologically homogeneous. However, because of the methodological heterogeneity of the studies, further statistical pooling of data was not possible.

Studies were published between 1973 and 2015. Four studies were undertaken in North America and Canada, three of which in the USA [15, 21, 22] and one in Canada [4]. Six were undertaken in Europe: two in the UK [23, 24], one in Denmark [25], one in Sweden [3], one in France [26], and one in the Netherlands [27]; two were undertaken in Asia: one in South Korea [13] and one in China/Japan [14].

Seven of the studies were longitudinally designed [3, 4, 14, 21, 24, 25, 27], and five retrospectively [15, 22, 23, 26, 28], comprising a total of 2099 patients undergoing THR, 60 % of which were males and 40 % females. Seven of the included studies described SQoL before and or after THR multi-dimensionally, and as the main question [13‐15, 22‐24, 26]; the other five described SQoL more indirectly, as one of the outcomes of a broader research question [3, 4, 21, 25, 27]. For example, two studies focused on Quality of Life (QoL) after THR [3, 4], one on function and pain after THR [21]; one translated and validated the Oxford Hip Score questionnaire into Dutch [27], and one focused on alternative outcome measures after THR in young patients [25]. The number of participants in the studies ranged from 22 [21] to 791 [15] and their ages from 20 [22] to 86 [21]. Seventeen RA patients (divided into 5 studies) and five AS patients (indicated in 1 study) are present in this review, as they were part of a group of respondents of which the results were not presented separately. The duration of follow-up ranged between the first post-operative routine visit [22] to a mean of 2.3 (+0,8) years after THR [15]. The characteristics of the included studies that assessed SQoL in patients after THR are presented in Table 1.

The methodological quality was rated as “moderate” in two studies [14, 15] and “low” in ten studies [3, 4, 13, 21‐27]. A full description of the methodological quality of all the studies is shown in Table 2, and the individual scores on all items of the methodological quality assessment can be found in Appendix 3. Unclear study participation, sampling and study attrition caused a risk of selection bias in five studies [13, 21‐23, 26]: for example, poor response rates or loss of follow up (more than 30 % was considered inappropriate), unrepresentative cohort study populations (e.g. including only satisfied patients) and unclear presented study details about reasons for loss to follow-up.

Inadequate use of validated outcome tools for the measurement of outcomes (inclusive method and setting), and inadequate or incomplete reporting of confounding variables, caused a risk of information bias in eleven studies [3, 4, 13, 15, 21‐27].

Three studies were only descriptive [23, 24, 27]. Statistical bias was observed in eleven studies due to unclearly described missing values and a lack of proper statistical analysis (e.g. multivariate analysis methods was considered appropriate) [3, 4, 13, 14, 21‐27]. Only one study performed multivariate analysis [15].

Studies used a number of terms to describe SQoL, and these terms were categorised into two main-groups, and several subgroups of SQoL:

Subsequently, we summarised the differences between pre- and postoperative outcomes in Table 3 as “Changes in SQoL after THR”, and the cross-sectional measured postoperative outcomes of SQoL were summarised in Table 4.

Differences between pre- and postoperative outcomes, defined as changes in SQoL after THR (Table 3), were reported in 10 studies [3, 4, 14, 21‐27].

Six studies reported a physical-functional change in SD after surgery [3, 22, 23, 25‐27] which ranged from Δ 25–51 % [3, 22], and by gender between Δ 8–20 % for males and Δ 16–29 % for females [25, 26]. Males preoperatively had less SD than women, women showed greater improvement after THR in three studies [23, 25, 26]. Positive changes between pre-operative and postoperative SA were reported in five studies [4, 14, 21, 24, 25], and ranged widely from Δ 0 to 77 % [14, 24]. Two studies reported a change in hip-pain during SA, respectively Δ 53 % and Δ 56 % [14, 21], while hip-mobility increased in one study (Δ 54 %) [14]. Two studies reported positive changes in SA: one reported an increase of “coital frequency” (Δ 73 %) [22], and one an improvement in “SA” (Δ 44 %) [4]. Two studies reported that the preoperatively sexual active patients had regained SA after THA [24, 25]; both reported a postoperative loss of follow-up (Table 3).

Two studies reported changes in psychosocial outcomes of SQoL [14, 24]. One of those, reported an increase in the patients’ “need for information” after THR of Δ 28 % [24]; the other study (only males) reported reduced “impairment of relationship” (Δ -37 %) (p = 0.026) and a change on “sexual satisfaction of patients” of Δ 40 % (p = 0.018) [14]. One study assessed associations between pre- and postoperative SQoL and clinical and demographic characteristics, and found no correlation (p > 0.05) between these variables and postoperative SQoL [14].

Seven studies reported cross-sectional outcomes [13, 15, 22‐26] (Table 4); mostly because they were designed retrospectively [13, 15, 22, 23, 26]. Physical-functional outcomes of SQoL were extracted from these seven studies [13, 15, 22‐26]. SD was reported in one study [23]: 51 % females and 50 % males had complete to considerable relief of SD after THR. Six studies reported on SF [13, 15, 22, 24‐26] including two that reported on increased “coital frequency” [15, 26] in 18.5 and 43.5 % of the patients, respectively; coital frequency stayed unchanged in 70 and 52 %, respectively, and it decreased in 11.5 and 4.5 %, respectively. Four studies reported on the “time to resume SA” [13, 22, 25, 26]: the majority of the patients (>50 %) resumed SA within 2 months [22, 26]; one Eastern study reported a mean of 6.9 months (3 weeks - 48 months) [13]. Females (87 days) resumed later than males (54 days) (p = 0.0005) [26]. “Decreased SA” was mostly affected by pain, mobility and apprehension [15]. In one study, three patients never resumed SA again [26], and one study reported that 3 males (4 %) under 60 years experienced erectile dysfunction after THA [25].

Psychosocial outcomes of well-being were reported in six studies [13, 15, 22‐24, 26]. Two studies reported on the need for more advice, ranging from 57 to 89 % [22, 23]; two reported that 80 % of the patients were not able to obtain information [13, 26]; and two studies mentioned SD as an argument to undergo THR [22, 26]. Two studies reported on the terminology subgroups “effect on relationship”, “effect on satisfaction”, “sex quality” and “concerns” [13, 15]. One study examined “effect on relationship” as the lack of understanding from the spouse (in 4.7 %, only males) [13]; in this study, satisfaction increased in 23.4 %; stayed the same in 68.8 % and decreased in 7.9 % [13]. One study found “sex quality” was experienced as better after surgery in 69.9 %, and experienced as worse in 2.2 % [15].”Concerns” were quantified in three studies [13, 15, 24]: in one study, 51.6 % of patients worried about fear of dislocation [13]; in another, the spouses worried about hurting the partner [24]; a third study reported that 3.1 % of the patients felt their hip was slipping out during SA after surgery [15].

Our systematic review is, to our knowledge, the first that summarises SQoL after THR. However, we are aware of one recent narrative review that reports the same beneficial but heterogenic effects of THR on SQoL [38], and we found a literature review of RA and sexuality, published in 1999, which reviewed 19 eligible papers with predominantly the same results [39]. Nonetheless, studies into SQoL and surgical hip treatment in the orthopaedic literature are few and far between: The first study was published in 1970, which corresponds to the time that the THR technique was being developed and gradually became safer [1]. However, since then, scant studies have acknowledged SQoL, and of the twelve studies included in this review, five were published more than 20 years ago. Although the methodology of the newer studies (after 2005) is more advanced, and recent studies show more depth of analyses of the topic, the overall results of the studies do not differ essentially between older and more recent studies.

We rated ten of the twelve included studies methodologically as “low” because they had numerous sources of risk of bias; and eleven studies lacked multivariate analyses methods. Five of the twelve studies were retrospectively designed, however, in our opinion, these studies were useful for focussing on the study question of SQoL more thoroughly, identifying detailed information and “feasibility issues for future longitudinal research” about SQoL after THR [40]. Generally speaking, we suppose investigating SQoL is complex because it is a sensitive issue. However, Fenton et al. (2001) suggest that sexual behaviour research is as difficult as other areas of self-reported behaviour, including diet, smoking, and alcohol consumption” [41], and they conclude that “continued methodological research is needed to better identify the sources of measurement error.” [41].

Additionally, the included studies paid little attention to comorbidities and other potential confounders. For example, a review concerning SQoL in psoriasis patients found that diabetes, hypertension or depression could play an important causal role towards erectile dysfunction [42], and another review suggests that the use of beta-blockers and diuretics may also have negative effects on SF [43]. Given the average age of THR patients, these confounders will likely be present in a considerable part of the hip population. The available studies further paid little attention to gender-specific complaints and outcomes; in addition, whereas females generally outnumber males in THR treatment [7], our review includes 60 % males, which indicates selection bias, and this is only partly explained by the fact that one great cohort-study included only men [14].

Our review was characterized by a number of different − prospectively and retrospectively measured − heterogeneously defined factors of SQoL. This caused marked heterogeneity; consequently that data synthesis was not possible. Moreover, we may have missed potentially eligible studies in other languages as well as studies on QoL that mention SQoL in the full text only. Although we intended to exclude studies about SQoL in RA and or AS patients (because of the systemic illness and the multiple joint involvements interfering with SQoL), we decided to include 5 studies in which the majority of the patients had OA, but a minority were RA or AS patients (between 2 and 11 % of the population). In these studies data were not presented separately for the different diagnosis and therefore the total samples were included in this review. This may have slightly biased the results of our review.

SQoL in patients with HA, before and after THR, is gaining importance as the total number of patients increases and the age range of patients broadens. Given the rising number of patients worldwide, we feel that SQoL should be better quantified routinely, for example by using Patient Reported Outcome Measures (PROMs) that are validated for this particular purpose. Longitudinal representative cohort studies would be helpful to accurately understand SD, beginning at the early stages, through to end-stage HA and postoperatively after THR.

Ethics approval and consent to participate was not required for this study, because this is a systematic review of previously published studies.

Not applicable.

Al data that support our findings are contained within the manuscript.