Introduction

Osteoarthritis (OA) is the most common form of arthritis and the leading cause of chronic disability in the United States [1]. The condition affects about eight million people in the United Kingdom [2] and nearly 27 million people in the United States. Osteoarthritic destruction of the knee is the primary reason for total knee arthroplasty (TKA) [3]. The use of surgical drains post-TKA has been standard practice for many years [4]. Surgical drains theoretically reduce the postoperative collection of fluid such as blood in a closed space. It is thought that the reduced formation of haematomas may result in decreased swelling, reduced nidus for infection and improved range of movement [5].

Observational studies on the use of closed-suction drainage systems for wounds following orthopaedic procedures are conflicting, such that the use of drains has been associated with both an increased and a decreased rate of wound infection [6–8]. A number of randomised controlled trials [9–14] have also examined the efficacy of closed suction drains following TKA but no formal review or pooled analysis of the data exists. In 2004, Parker et al. [15] examined the efficacy of drains following total hip and knee arthroplasty and in 2008 the same authors published a similar study on hip fracture surgery [16]. In both studies the authors concluded that there are no benefits to the use of drains after these surgeries. The aim of this systematic review and meta-analysis is to specifically examine the efficacy of closed suction drains on range of movement, swelling, haemoglobin levels and infection rates following TKA.

Methods

Literature search

The following databases were used to identify studies for this review: Pubmed, EMBASE, MEDLINE, Cinahl and the Cochrane Library. The databases were searched using a combination of the following key words: “total knee replacement” OR “total knee arthroplasty” OR “total knee prosthesis” OR “arthroplasty, replacement, knee” [MeSH term] AND “unilateral” AND “bilateral” AND “suction drainage” OR “closed suction drain” OR “suction drainage” [MeSH term]. The search was supplemented by hand searching the references of included articles.

Definitions and inclusion criteria

The PRISMA guidelines for the reporting of systematic reviews and meta-analyses were followed to conduct this review [17]. Types of study design included in this review were randomised controlled trials (RCTs). Participants in the studies included were those who underwent primary TKA. The intervention under investigation was the effect of no surgical drains on outcome and the comparison used was a control group with a drain. In the included studies, the primary outcome measure examined was range of flexion of the knee joint. Other secondary measures included swelling, haemoglobin loss and length of hospital stay.

Study identification and selection

One reviewer (MQ) read the titles of the identified references and eliminated irrelevant studies. Articles that did not meet the inclusion criteria or that included patients for revision arthroplasty, or early clamping of drains were excluded from further analysis. The remaining articles were obtained and, based on the inclusion criteria, one reviewer marked them as relevant, irrelevant or unsure. A second reviewer (RG) independently ranked these trials using the same method. Any disagreements were resolved through discussion between the reviewers.

Methodological quality

Two reviewers (MQ, RG) independently documented the methodological quality of the studies and extracted the relevant data. Studies were assessed in the following domains to eliminate bias: sequence generation; allocation concealment; blinding of participants, personnel and outcome assessors; incomplete outcome data; selective outcome reporting and any other sources of bias. If all of the criteria are met, the study is considered to have a low risk of bias. If one or more criteria are partly met, the study is deemed to have an unclear risk of bias. If one or more of the criteria are not met, then the study is considered to have a high risk of bias. Any disagreements in the quality of the studies were resolved through discussion between the reviewers.

Data extraction

The following data was documented for each study: authors, number of participants in intervention and control groups, results of the study and the outcome measures used. For the purpose of the meta-analysis, scores (mean and standard deviation) on these outcome measures were recorded at first assessment and at all follow-up assessments. Authors were contacted for clarification or to provide any further follow-up.

Quantitative analysis

The extracted data was inputted into Microsoft Office Excel. The meta-analysis was computed using Review Manager 5 from the Cochrane collaboration. The mean difference (MD) between the ‘control’ or ‘drain’ group and the ‘no drain group’ post surgery was used as the mode of analysis of the pooled data for each outcome measure. In studies where the mean value was not available, the median was recorded instead. When the standard deviation was not available, the interquartile range was recorded and if these were not available, the ranges were recorded. The median was used as a proxy for the mean. A multiple of 0.75 times the interquartile range or 0.25 times the range were used as proxies for the standard deviation values. The impact of sample size was addressed by estimating a weighting factor for each study, and assigning larger effect-weights in studies with bigger samples. The random effect model was applied and the I² statistic described study heterogeneity.

Results

Study identification

A flow diagram of the search strategy is presented in Fig. 1. The search was completed in November 2012. Two researchers screened all potential articles. The literature search yielded 254 papers, of which 147 articles were eliminated on the basis of irrelevant title and abstract. Six of the remaining 107 papers met the inclusion criteria and were selected for analysis.

Fig. 1
figure 1

Search strategy

Full size image

Study description

The summary details of RCTs included in the review are contained in Table 1. A total of 495 patients (564 knees) were included in the review.

Table 1 Description of studies included in the review
Full size table

Study quality

The results of the quality assessment are found in Table 2. The overall quality of the studies included is adequate, and only one study [9] demonstrates a high risk of bias. However, the authors identify their own shortcomings and limitations in the discussion.

Table 2 Methodological quality of studies included in the review using the Cochrane collaboration’s tool for assessing risk of bias
Full size table

Range of movement

The primary outcome variable considered in this systematic review is range of knee flexion postoperatively. Five studies present results for postoperative range of knee flexion [5–9]. In all studies, range of knee flexion is measured using a standard universal goniometer. Three studies [5–7] report measurements taken between day six and day eight postoperatively. Two studies [8, 9] report measurement of knee flexion six weeks postoperatively. In all studies, it is unreported whether the outcome assessors are blinded to group allocation. In all five RCTs, the authors report no statistical difference between the intervention or ‘no drain’ group, and the control or ‘drain’ group. One RCT [8] is excluded from the meta-analysis as not enough information is available to calculate standard deviations in the control or the intervention group. Three RCTs [5–7] are pooled to ascertain if the presence of a drain after a TKA improves the patients’ postoperative range of knee flexion at six to eight days post TKA. This is illustrated in Fig. 2. A further analysis was also completed including the study by Holter et al. [9] where range of knee flexion is measured at six weeks. This analysis demonstrates no statistical difference in ROM knee flexion [mean difference 0.03° (95 %CI −1.51 to 1.45, p = 0.64, I² = 0 %)].

Fig. 2
figure 2

Range of knee flexion six to eight days post TKA

Full size image

Swelling

Three studies [5–7] report results for postoperative knee swelling. All three RCTs report no statistical difference between the ‘no drain’ group and the ‘drain’ group. Measurement of knee circumference included pre and postoperative measurements of the knee at the mid-patellar point using a manual tape measure. Two RCTs [6, 7] are pooled to ascertain if the presence of a drain after a TKA decreases the patient’s postoperative knee circumference when compared to no drain. The authors assess knee circumference prior to discharge ranging from seven to eight days postoperatively. There is no significant difference in knee circumference between the two groups [mean difference 1.63 cm (95 % CI −1.07 to 4.34 cm, p = 0.34), [I² = 0.12].

Other outcomes

Four studies [4, 5, 7, 9] present results for postoperative haemoglobin (Hb) drop postoperatively. All RCTs report no significant difference between the intervention and the control group. Three RCTs [4, 5, 7] are pooled to ascertain if the presence of a drain after a TKA has an effect on the patients’ Hb levels at two days post surgery. The mean difference in Hb drop between the two groups is 0.25 g/dL [95%CI −0.48 - 0.02, p = 0.59, I² = 0 %]. Figure 3 illustrates these results.

Fig. 3
figure 3

Haemoglobin levels in the groups two days postoperatively

Full size image

It is not possible to pool data on blood transfusion requirements. However, two RCTs [7, 9] report no statistical difference in blood transfusion between the groups, while one RCT [6] reports a statistically significant amount of increased blood transfusion in the ‘drain’ group. Three studies [6, 7, 9] provide results for postoperative length of hospital stay (LOS). All three report no statistical difference between the groups, though not enough data was provided to perform a meta-analysis.

Discussion

Statement of principal findings

This systematic review and meta-analysis examines the available evidence to date in relation to the use of suction drainage in primary TKA. We include six randomised controlled trials which focused solely on TKA. Our findings do not support the routine use of postoperative suction drains in primary total knee arthroplasty in terms of improvements in knee swelling or flexion, length of hospital stay or haemoglobin levels.

Current context and future research directions

In spite of the paucity of evidence to support their use, for many years the majority of major orthopaedic procedures were followed by the use of drains postoperatively to prevent haematoma formation [1]. Methodological issues such as small sample sizes and inconsistent study quality and lack of follow-up data limited the wider application of the findings from these studies. Therefore, we conducted a systematic review of the literature and subsequent meta-analysis of randomised controlled trials. We also evaluated the methodological quality of the studies to examine the internal and external study validity. Our evidence is in keeping with previous research findings, all of which suggest that their use may not be warranted [2, 15, 18–21]. The primary clinical finding that emerges from this systematic review and meta analysis is that the use of drains has no impact on patient outcome post TKA, in terms of improvements in range of motion of the knee, swelling, haemoglobin levels and length of hospital stay. Theoretically it is suggested that range of movement and knee swelling improves with the use of drains post TKA. However, drains may act as a channel for the introduction of infection and may also increase infection by impairing host resistance and allowing pathogens access to a sterile field [13, 14, 22, 23]. Furthermore, drains serve to increase the demands on nursing care and physiotherapy to accommodate the presence of drain.

Larger multi-centre clinical RCTs to assess the impact of drains on length of hospital stay and the cost in terms of nursing costs and adverse events associated with drain use are needed. Furthermore, double-blinded studies where the outcome assessor and patient are blinded to group allocation are needed to improve the quality of the reporting of studies. Furthermore, all outcome measures assessed in this trial are short-term measures. It is important to gauge long-term outcomes in arthroplasty, as this is most pertinent to patients [24]. Therefore, long-term follow up of individuals is also required to assess the long-term impact on recovery post-discharge. In additional to measuring ROM and mobility in the longer term, other outcomes such as deep joint infection and component loosening, which may take a number of years to manifest, should be considered.

Strengths and weaknesses of the study

Due to differences in reporting and data collection, it is not possible to pool all information from each of the studies. None of the six studies used [3–8] report results for all parameters investigated, and some measurements are incompatible for analysis due to temporal differences. Heterogeneity exists among some studies, therefore the findings above need to be considered carefully. Five studies [25–29] were excluded as they are not published in English—while efforts were made to include these studies, it was not possible to include them for the purposes of this review. A study in the use of routine drainage in hip arthroplasty by Zhou et al. [30] showed similar conclusions, including functional recovery and haemoglobin levels. This similarly encountered the problem of not being able to include non-English articles in its review [31, 32].

Conclusion

Contrary to common clinical practice, the randomised controlled trials included in this review suggest that there are no additional benefits, in terms of patient outcome, to using a suction drain following TKA surgery.