There is a strong tendency for surgical techniques to be improved over time and as new instruments are developed less invasive approaches are possible [1]. Currently the most often used surgical exposures for hip replacement such as the anterolateral and the posterolateral approaches involve splitting muscles with risk of partial denervation and detachment of tendons with a risk for incomplete healing. In many cases this results in weakness of hip abductor muscles and notable limp. Using the direct anterior approach for total hip arthroplasty these problems are largely avoided and several reports in the literature have documented advantages with the technique [2–6]. A short rehabilitation time due to minimal soft tissue trauma is often emphasized. Complications have also been encountered, such as injury to the lateral femoral cutaneous nerve, component malposition, damage to the femoral shaft and delayed wound healing [2–5, 7]. Obviously, the benefits with the new technique are at the expense of a new panorama of complications and intraoperative difficulties. Exceptionally skilled surgeons might be the first to describe the results and when the technique is more widely adopted more problems can become apparent. This report describes the initial experience by a team of four surgeons, a senior surgeon, a recently board certified orthopaedic surgeon and two orthopaedic residents. The aim of the study was to describe early complications so that surgeons who adopt the technique will be informed not only about potential benefits but also of potential risks with the technique.

The material is described in Table 1. Mean values for blood loss, operating time and cup position are shown in Table 2. There was a small but statistically significant influence of BMI on the time of surgery (Table 3). Similarly, obese patients tended to have the cup positioned at a higher degree of deviation (Figure 2). The estimated intra operative blood loss was not significantly correlated to the BMI. It did not change significantly in the whole material during the course of the study. The time of surgery decreased significantly with the experience of the team (Figure 3).

Patients (n = 200)
Women	121
Men	79
Age	29–88 (mean 67,4)
BMI	17–43 (mean 26,7)
Diagnoses	Osteoarthritis	152 (76%)
	Inflammatory arthritis	31 (15,5%)
	Avascular necrosis	9 (4,5%)
	Dysplastic osteoarthritis	8 (4%)

	Mean	SD
Time of surgery (min)	114	28.9
Blood loss (ml)	496	322
Anteversion	18°	7,6°
Deviation	46°	8,5°

	Increasing BMI
Blood loss	n.s.
Operating time	y = 1,1179x + 85,113 R2 = 0,0317 P < 0,05
Anteversion	n.s.
Deviation	y = 0,2816x + 38,473 R2 = 0,0238 P < 0,05

Femur fractures in the early postoperative period occurred in cases 9, 10 and 21 (Figure 4). The cause of the cracks in the proximal femur were likely due to improper broaching. These patients were revised using a standard anterolateral approach; a wire cerclage was used to fix the fractures. In one case the same prosthesis was used, but in case number 10 a long-stem (Wagner™, Zimmer USA) prosthesis was used together with a cemented cup and in case 21, a Meridian™ prosthesis (a cobalt-chrome prosthesis with proximal coating, Stryker USA) was used instead of the Accolade due to its better filling of the metaphysis.

There was no significant correlation between complication and gender, age or BMI.

The introduction of new approaches and instruments can be expected to temporarily be associated with complications, such as neurovascular injury and component malposition. It is important to have information about the risk for various complications and predisposing factors. This study is based on non-selected, consecutive, primary unilateral total hip replacements. The very reason to use this approach is the possibility to minimize surgical soft tissue injury and to maintain normal muscle function and stability of the hip [1, 7, 9–13]. The approach is internervous and does not include the release of muscles or tendons. In the literature there is little evidence for a higher complication rate with the direct anterior approach compared to lateral approaches. It is however of interest to demonstrate and discuss the hardships involved when the technique is adopted by a team of surgeons with mixed experience.

Overall the distribution of complications was fairly even among the 200 cases included in the study. Aside from the first 10 cases, this study does not well delineate a learning curve. An almost linear reduction in intraoperative blood loss was noted among the first 10 cases; in two of them cracks occurred in the proximal femur – this was not seen in cases done later.

Access to the femoral canal can be expected to be difficult especially in patients with a short and varus angulated femoral neck and where the range of motion is restricted due to fibrosis of the joint capsule. Other authors [3, 12, 14, 15] have reported fractures in the proximal femur with this approach. In our series the major predisposing factor seems to have been pronounced osteoporosis. With the currently used offset handle, broaching is safer than with a standard straight handle as in the present series. The risk for fracture of the proximal femur is further minimized by adequate posterolateral capsule release so that an elevator can be placed around the tip of the trochanter while the hip is extended, adducted and externally rotated. However, if the surgeon excessively releases soft tissues from the proximal femur (which is not necessary) instability could result. This was possibly the reason for the dislocations that were seen in cases 118, 189 and 192 in our series. Even if no instability occurs, the release of tendons from the proximal femur counteracts the objective of this approach. Case number 169 was a patient with dysplasia (Figure 5) where the necessary medialization of the femur and not excessive soft tissue release led to postoperative instability. The end result of that operation was excellent.

The direct anterior approach subjects the lateral femoral cutaneous nerve and the femoral nerve to higher risk than lateral approaches. Even though the posterior branch of the lateral femoral cutaneous nerve can be injured [2, 12, 16] the technique described here seems to be effective in avoiding clinically relevant damage to the nerve. This is important since dysesthesia due to injury of this nerve may lead to considerable discomfort in the thigh for several months. Injury to the femoral nerve with the anterior approach has been mentioned by others [3, 17, 18]. We believe that it can be caused by the retractor placed anteriorly so that the iliopsoas muscle and the femoral nerve are compressed against fascial structures. This retraction should be directed cranially, towards the opposite shoulder and not transversely. In addition we have also constructed a retractor which sits around the ilium at the level of the anterior inferior iliac spine. This instrument is useful in difficult cases such as patients with pronounced abdominal obesity. Fortunately, femoral nerve injury seems to have a good prognosis and should resolve completely within a few months.

We decided to use the direct anterior approach even in cases with morbid obesity and in cases with grade 1 and 2 acetabular dysplasia. In obese patients we found that there was a specific problem in positioning the acetabular component. A high BMI was also associated with a longer operative time but not with a significantly higher intraoperative blood loss.

In summary our intention with this report is to discuss the risks involved with the adoption of a MIS technique for total hip replacement. It has the theoretical advantage of minimal soft tissue injury and it has been shown that the expectations of facilitated early rehabilitation can be fulfilled [9, 10]. In our opinion the present results do not obviate the technique as a routine for total hip arthroplasty but it is necessary to have a thorough understanding of the risks beforehand. Although technical difficulties necessitating revision surgery were more obvious among the early cases other complications such as dislocations may appear later. The technique is perhaps more technically demanding than the lateral approaches used today due to the somewhat limited surgical exposure. It should be reserved for specially trained surgeons who have the possibility to treat many patients in order to maintain good skills. It has been our overall impression that morbidly obese or very muscular patients as well as patients with a short femoral neck or acetabular protrusion can represent particular problems.