The SI joint is a pain generator in about 15% to 25% of axial low back pain patients [1
]. Many patients with SI joint pain are misdiagnosed or not diagnosed at all, inducing a wrong treatment strategy that will lead to a bad result. It is therefore vital to be aware the kind of pain that originates from the SI joint as well as to diagnose and treat it accordingly. A recent systematic review shows that the evidence for the diagnostic accuracy of SI joint injections is good, the evidence for provocation maneuvers is fair, and evidence for imaging is limited [2
A systematic evaluation of the therapeutic effectiveness of SI joint treatment modalities has shown that evidence was fair for cooled radiofrequency neurotomy and poor for short-term and long-term relief from intra-articular steroid injections, peri-articular injections of steroids or botullinum toxin, pulsed radiofrequency, and conventional radiofrequency neurotomy [3
]. Unfortunately this review did not evaluate surgical treatment. Although in recent years good results have been reported with fusion of the SI joint [4
], no clear-cut evidence has been published showing that surgery is superior to other treatment modalities and further comparative studies are needed [19
Since 1927 when Smith-Petersen described SI joint arthrodesis [26
] several surgical series have been published. The first ones entailed opened approaches with large incisions, significant bone harvesting, lengthy hospital stays and restricted physical activity for several months. Results were acceptable but not too encouraging. Waisbrod et al. [4
] in 1987 reported twenty-one patients treated with excision and package the SI joint with ceramic blocks and iliac crest bone graft, with satisfactory results in eleven of them and a 14% complication rate (2 pseudo-arthrosis and one infection). Keating et al. [27
] in 1995 reported twenty-six patients treated with SI joint debridement, decortication and package with bone graft and secured with 2 lateral compression screws. No complications were seen, the pain scores decreased from 6.1 to 2.9 and there was improvement in work status. Later, in 1997, Moore [28
] reported 77 patients treated with a modified Smith-Petersen technique. Through a 15 cm incision the cartilage of the SI joint was removed and the joint packed with autologous Ilium bone graft
, kept in place with two to three cannulated screws. He reported successful results in 62 out of 77 patients (80.5%) and a 13% complication rate (1 superficial wound infection, 1 post-op radicular pain, 1 sciatic notch fracture and 7 pseudo-arthrosis). Buchowski et al. [5
] reported a similar technique in 20 patients but using instead a plate and screws to stabilize the graft. They achieved a solid fusion in 17 out of 20 patients with a return to work in 8. Complication rate was 30% (3 pseudo-arthrosis, 2 deep wound infections, 1 painful hardware and 3 revision surgeries). Sixty percent of these patients stated that would undergo the surgery again.
Once proved that the surgical treatment can be a solution for those patients with pain refractory to conservative medical treatments, the next step was reduction in surgical aggressiveness with MIS techniques as these are widely established in the treatment of other lesions in other spinal locations [29
]. Advantages are less morbidity, less postoperative pain, shorter hospital stay, better and faster recovery and earlier return to work. This may also result in economic benefits and reduced hospital costs.
Al-khayer et al. [7
] and Wise et al. [8
] published in 2008 their respective series with percutaneous techniques. The first group treated 9 patients with a 10 mm Hollow Modular Anchorage Screw packed with demineralized bone matrix across the SI joint and early mobilization in the post-op. They reported blood loss <50 ml, hospital stay 6.9 days, ODI reduction from 59 to 45, VAS improvement from 8.1 to 4.6, no screw loosening, no screw failure, no nonunion and return to work in 4 out of 9 patients. They saw 1 deep wound infection and a complication rate of 11%. Wise et al. treated 13 patients with 9mm hole drilled through the longitudinal aspect of the SI joint and 2 cages packed with BMP placed across the anterior portion of it. Post-op care entailed limited waist bending and sacral belt for 6 months with full activity at 6 months. Their results were: blood loss < 100 ml per patient, length of stay: 1.7 days, fusion rate: 89%, low back VAS improved 4.9 points and leg VAS 2.4 points. They reported a complication and revision rate of 8% with 1 reoperation (due to nonunion). In 2009 Khurana et al. [9
] using the technique previously reported by Al-khayer et al. [7
] reported 15 patients reducing the hospital stay to 2.7 days and with no complications.
In 2012 Rudolf [11
] reported his results using the minimally invasive technique with triangular, porous plasma spray coated titanium implants. He treated 50 patients with early and sustained statistically significant improvement in pain function identified at all post-operative time points. A clinically significant improvement (>2 point change from baseline) was observed in 7 out of 9 domains of daily living. More than 80% of patients would have the same surgery again. There were, though, 10 peri-operative complications. Three patients with subcuticular skin closure developed superficial cellulitis that resolved after a short course of oral antibiotics. These infections never happened again after changing to Nylon skin sutures. One patient experienced a deep-soft tissue wound infection that resolved after 6 weeks of intravenous antibiotics. Two patients developed a large buttock hematoma, with post-operative pain and difficulty on sitting, resolving spontaneously in 4-6 weeks and leaving no residual symptoms. Implant penetration into the sacral neural foramen with nerve root irritation and radicular pain happened in two patients. Both were brought back to the OR and the misplaced implants retracted to the edge of the neural foramen with complete resolution of symptoms. Similarly, in one patient harboring an unrecognized hemi-sacralized L5 transitional vertebrae
, the first implant was inadvertently placed too cephalad, compressing the L5 nerve. This patient needed a partial retraction of the implant too. Another patient suffered from a non-displaced fracture located at the inferior edge of the ilium adjacent to the sciatic notch at the edge of the lowest implant but the fracture healed spontaneously and no further surgery was required. Finally, a late complication occurred three years after surgery. In a patient presenting persistent and gradually increasing SI joint pain the CT of the pelvis showed findings suggestive of motion surrounding the sacral end of the two most caudal implants. After a CT-guided injection confirming that the SI joint was the pain generator, two additional 7.0 mm implants were placed anteriorly to the loosened implants, resulting in complete symptom resolution. We believe that some of these complications in the Rudolf series are due to the learning curve of the technique and we profited from the advice received from him. We feel very grateful from the support and advice of this great master.
In the last two years more series with the iFuse SI Joint Fusion System have been published. Cummings et al. [12
] have treated 18 patients without intraoperative
complications and one explant at three months post-op due to malposition. All patient-reported outcomes showed both clinically and statistically significant improvement at 12 months. Duhon et al. [13
] have reported the early results of a multicenter prospective single-arm cohort of patients. Mean subject age was 51 years (n=94, safety cohort) and 66% of patients were women. Subjects were highly debilitated at baseline (mean VAS pain score 7.8, mean ODI score 54). There were no severe events device-related. Complete 6-month postoperative follow-up was available in 26 subjects. In the effectiveness cohort, mean SI joint pain improved from a baseline score of 7.8 to a 6-month score of 2.9, mean ODI improved from 55.3 to 38.9 and SF-36 PCS improved from 30.7 to 37.0. Ninety percent of subjects who were ambulatory at baseline regained full ambulation by 6 month post-op and median time to full ambulation was 30 days. Satisfaction with the procedure was high (85%).
Miller et al. [16
] have done an analysis of a post-market complaint database for the iFuse with 5319 patients that were treated between April 2009 and January 2013. Complaints were reported in 204 (3.8%). Pain was the most commonly reported clinical complaint (n=119, 2.2%), with nerve impingement (n=48, 0.9%) and recurrent SI joint pain (n=43, 0.8%). All other clinical complaints were rare (0.2%). They reported ninety-six revision surgeries performed in 94 (1.8%) patients at a median follow-up of four (range 0–30) months. Revisions were typically performed in the early postoperative period for treatment of a symptomatic malpositioned implant (n=46, 0.9%) or to correct an improperly sized implant in an asymptomatic patient (n=10, 0.2%). Revisions in the late postoperative period were performed to treat symptom recurrence (n=34, 0.6%) or for continued pain of undetermined etiology (n=6, 0.1%).
Ledonio et al. [18
] have compared 39 patients, of whom 22 underwent open and 17 underwent minimally invasive SI joint fusion and both resulted in statistically and clinically significant improvement for patients with degenerative SI pain refractory to non-operative management. However, the number of patients reaching the minimally clinically important difference and those showing overall improvement were greater in the minimally invasive surgery
group. The same author recently have published another paper [31
] with 63 patients prospectively studied (open: 36; MIS: 27) who underwent SI joint fusion with minimum 1-year follow-up. Their results have shown that patients in the open group had a higher mean estimated blood loss (681 ml versus 41 ml, p<0.001). Mean surgical time and length of stay were shorter in the MIS group than in the open group (68 minutes versus 128 minutes and 3.3 days versus 2 days, p<0.001 for both). But with the numbers available, mean postoperative ODI scores were not different between groups (47% versus 54%, p=0.272). They concluded that the study size was relatively small and it is possible that the study was underpowered.
We initially used a sharp Steinmann pin all through the operation, but after two cases of temporary post-operative radiculopathic pain, we decided to change to a blunt Steinmann pin. The sharp one is used to start with and set the dilators. Then it is removed and a blunt one is used to guide the drill, the broach and the implant insertion. Since then we have not seen any other case of post-op radiculopathic pain. Thus we strongly recommend it.
The results obtained in our series of patients are consistent with those published in the literature and support the use of percutaneous SI joint arthrodesis in carefully selected patients. This surgical approach offers less morbidity and therefore a better and faster recovery with less postoperative pain and less surgical stay with a faster return to work.
One of the weaknesses and limitations of the study is the small number of patients over a period of more than four years. This makes difficult to do the statistical analysis that could be no demonstrative. It is important to say that there are not long-term complications and patients have remained clinically stable with similar VAS and ODI over time.