All ITPIEs were performed using a medial parapatellar approach and there were no unique aspects of the surgery that could have influenced the outcome. Preoperative computed tomographic scanning was not used for this group of patients. Component fixation was rechecked intraoperatively using manual tools to apply tensile forces and varus-valgus loads to the components. Intraoperative assessment of rotation was based on gross inspection of component position relative to the tibial tubercle and the femoral epicondyles.
Postoperative followup visits were scheduled for patients at 1 month, 3 months, and 12 months postoperatively. Beyond 12 months, followup regimens were variable and reflected individual surgeon preferences, differences in patient compliance, and patient-specific recommendations. Patients were evaluated at all postoperative visits by the operating surgeon, which included clinical examination (pain, range of motion, revision status), radiographs, and patient questionnaires. KS scores were generated based on these visits. Pain at the various followup intervals was determined from an item on the KS questionnaire filled out by the patients, who were asked to rate their pain as nonexistent, mild, moderate, or severe. The most recent followup of each patient was used for the present study. Nine patients unable to follow up in person (nonambulatory or nonlocal) were contacted by telephone by one of us (SEW) to determine their revision status (whether they had additional surgery after the ITPIE). The function portion of the KS was administered over the telephone as well. Families of patients who died were contacted to find out if the patient had a revision of the ITPIE before death. All telephone assessments were made by the first author of this paper.
The ITPIE was defined as a failure if the patients underwent a revision of the ITPIE for any reason (repeat ITPIE or other components revised) or if they had persistent pain of moderate to severe intensity recorded on their pain questionnaire. Implant success was defined as the ITPIE remaining in situ without pain of moderate or severe intensity. The Kaplan-Meier method was used to compute survivorship of the ITPIE with revision for any reason as the end point. Chi square tests were used to compare ITPIE failure rates among the various indications for ITPIE (instability, stiffness, and effusion) and between component design types (cruciate-retaining or posterior-stabilized). An odds ratio was calculated to assess the risk of having a revision of the ITPIE in patients categorized by the time from index TKA to ITPIE (< 3 years or ≥ 3 years). All statistical analyses were conducted using SPSS (Version 13.0 for Windows; SPSS Inc, Chicago, IL).
Twelve of 42 patients (29%) went on to further revision after ITPIE. The mean time to revision was 3 years (range, 2 weeks to 8.7 years). At 11 years, survivorship of the ITPIE with revision as the end point was 58% (95% confidence interval, 42%-72%) (Fig. 1). Of those revised, three patients underwent a repeat ITPIE, eight patients had multiple components exchanged, and one patient had a resection arthroplasty as a result of infection. Thirty of the 42 patients (71%) were unrevised at final followup or death. Of these 30, the mean range of motion before the ITPIE was 103° (range, 55°-125°) and after the ITPIE at last followup was 113° (range, 88°-125°). Mean KS function scores increased from 52.7 preoperatively to 65.4 at followup and mean KS knee scores increased from 58.6 to 85.4. Despite improvements in KS scores and range of motion, moderate to severe pain at final followup was reported by nine of 30 (30%) of the unrevised patients (or 21% of the entire cohort [nine of 42]). The overall failure rate (revision of ITPIE or moderate to severe pain in unrevised patients) was 50%.
Failure rates were not statistically different for patients having ITPIE for instability, stiffness, or effusion (Table 1). Of the 12 patients who had revisions, seven originally underwent IPTIE for instability, one for stiffness, and four for recurrent symptomatic aseptic effusions believed to be the result of polyethylene wear.
Component design did not predict ITPIE revision. Revision of the ITPIE occurred in 31% of cruciate-retaining designs and 27% of posterior stabilized designs (Table 2). Differences in revision rates were not observed among the three diagnostic indications (instability, wear, stiffness) that led to the ITPIE in either the cruciate-retaining or posterior-stabilized cohorts. We identified no differences (p = 0.202) in failure rates among the cruciate-retaining knees that had lipped and nonlipped inserts placed at the time of ITPIE; four of the eight (50%) of the lipped inserts and eight of 33 (24%) of the nonlipped inserts were revised. Patients who underwent ITPIE less than 3 years from their index TKA were 3.8 times more likely (p = 0.059) to later have a revision than patients who underwent ITPIE more than 3 years from their index TKA (Fig. 2). Twenty-three patients underwent ITPIE more than 3 years from their index TKA. This group had an 83% success rate with only four patients in this group having a revision of the ITPIE. One had all components exchanged at 5.6 years after the ITPIE. Two underwent a second ITPIE at 4.7 years and 8.7 years after the initial ITPIE. One additional patient had femoral and tibial components exchanged 1.7 years after ITPIE.
Modular design of TKA allows the surgeon to exchange the polyethylene without revising well-fixed components. Although an excellent concept, previous reports have failed to demonstrate predictable positive long-term results. We measured survival of ITPIE in patients who underwent the procedure for instability, stiffness, or effusions believed to be the result of polyethylene wear. We also assessed whether the reason for ITPIE influenced the likelihood of revision and if component design or time in situ of the original arthroplasty were associated with the likelihood of success after ITPIE.
We note several limitations of the study. First, our study consisted of a small, heterogeneous patient population. The results of this study might therefore not directly apply to a larger heterogeneous population. Second, the operative reports that were reviewed did not always provide detail as to surgical findings such as ligamentous integrity and foci of osteolysis, and the clinical records did not always detail the magnitude and direction of instability patterns. Other surgeons might define ligamentous stability differently because this is a subjective evaluation. Third, the sterilization method of the polyethylene substantially affects clinical outcome, especially during the era of gamma sterilization in air. Unfortunately, we were unable to determine the sterilization technique for the polyethylene in all of the included patients. With new sterilization techniques, results in the future with this procedure may improve. Lastly, the cases were gathered over a 14-year interval, which included multiple surgeons with multiple implants. Prosthetic designs have evolved over this era as has our understanding of the potential pitfalls of ITPIE. These design changes and the expertise of the various surgeons may have affected the outcomes positively or negatively. Despite these limitations, the high failure rate seen across diagnostic groups leaves a clear message that this procedure should be undertaken with caution.
We observed a 29% revision rate in our cohort of patients undergoing ITPIE. Although we wanted to see if reoperation rates were different depending on the diagnostic indication for ITPIE, we ultimately found the revision rate in our series was similar for knees that had ITPIE for instability, stiffness, and effusion believed to be the result of polyethylene wear (30%, 25%, and 27%, respectively). The discouraging results we observed were similar to most other authors who have reported on ITPIE [1, 2, 4-6, 8] (Table 3).
We found 30% of ITPIEs performed for instability went on to further revision. Our findings are in agreement with another published study  of 27 knees that underwent ITPIE for instability. In that study, failure was defined as revision or severe pain. At a mean of 5.8 years, 55.6% (15 of 27) had survived. Our results were similar, with 57% of the ITPIE performed for instability ultimately judged to be failures. In another small series of patients  who had ITPIE for knee instability, four of 14 patients had persistent instability at a mean of 56 months postoperatively.
The second indication we investigated was stiffness. We only had four patients in our series who underwent ITPIE for stiffness. Nevertheless, one of these patients went on to further revision surgery, and two of the others had persistent pain. In all, three of the four patients in this group did poorly with ITPIE in our hands. Our findings are similar to those reported in another small series of seven knees that underwent ITPIE for stiffness . Two knees in that study were revised at 2.5 years after ITPIE. Of the five remaining patients, four reported severe pain and one patient had moderate or occasional pain.
The third indication we investigated was effusion believed to be the result of polyethylene wear. Our finding that 27% of knees that had ITPIE for effusion went on to further surgery is similar to a prior report  of 48 knees with ITPIEs for wear. In that study, 17% (eight of 48) of the overall cohort failed at an average of 4.5 years after the exchange. However, a subgroup of 22 patients in that study had ITPIE for severe polyethylene wear, and six patients (27%) within that subgroup failed at less than 5 years. Despite the poor results observed for ITPIEs we performed for wear and the previously referenced study , at least one report  has described better results. In that study, 68 knees from four centers were followed for an average of 3.6 years after ITPIE for wear and osteolysis. Those authors reported 84% of the patients retained their components and lacked progression of osteolysis after the ITPIE. Although more encouraging than our results, the failure rate of 16% reported by those authors is still worrisome, and their followup (3.6 years) was shorter than ours (5.6 years).
We found time in situ of the index arthroplasty predicted success of the ITPIE; TKAs that had been in situ for 3 years or longer had a 3.8-fold lower revision rate than TKAs that had been in situ less than 3 years at the time of ITPIE. In total, 83% of TKAs in our series with a service life of over 3 years at the time of ITPIE were free of further revision at final followup. In one study with the highest success rate after ITPIE , a group of patients with ITPIE was compared for reoperation and severe knee pain at a mean of 3.3 years with a group of patients with full TKA revision. Survivorship in the two groups was similar. The patients in that series were operated on at an average of 9 years after the index TKA and, as such, we believe the generally favorable survival these authors report is in agreement with our findings that knees with longevity before ITPIE are likely to do well after ITPIE. This appears intuitive, because knees that have functioned well over a long period of time are unlikely to have substantial abnormalities of alignment, rotation, or soft tissue balance.
In summary, we studied a heterogeneous group of patients undergoing ITPIE for instability, stiffness, and effusions believed to be the result of wear. We observed a high rate of failure across all three diagnoses with 29% (12 of 42) undergoing reoperation. We also identified moderate to severe pain at final followup in 30% (nine of 30) of the unrevised patients. Including the patients with moderate to severe pain as failures increased the overall clinical failure rate to 50% (21 of 42 patients). The patients who fared best in our series were those who had the longest successful in situ service of the original arthroplasty. We believe ITPIE is a procedure that should be undertaken with caution. For patients who are less than 3 years from their index TKA, ITPIE is associated with an unpredictable outcome. For patients who have a well-functioning TKA and begin to experience symptoms related to instability or effusions believed to be the result of wear, more than 3 years from their index procedure, ITPIE is a reasonable treatment option.
1. Babis, GC., Trousdale, RT. and Morrey, BF. The effectiveness of isolated tibial insert exchange in revision total knee arthroplasty. J Bone Joint Surg Am.
2002; 84: 64-68.
2. Babis, GC., Trousdale, RT., Pagnano, MW. and Morrey, BF. Poor outcomes of isolated tibial insert exchange and arthrolysis for the management of stiffness following total knee arthroplasty. J Bone Joint Surg Am.
2001; 83: 1534-1536.
3. Bert, JM., Reuben, J., Kelly, F., Gross, M. and Elting, J. The incidence of modular tibial polyethylene insert exchange in total knee arthroplasty when polyethylene failure occurs. J Arthroplasty.
1998; 13: 609-614. 10.1016/S0883-5403(98)80001-6
4. Brooks, DH., Fehring, TK., Griffin, WL., Mason, JB. and McCoy, TH. Polyethylene exchange only for prosthetic knee instability. Clin Orthop Relat Res.
2002; 405: 182-188. 10.1097/00003086-200212000-00023
5. Engh, GA., Koralewicz, LM. and Pereles, TR. Clinical results of modular polyethylene insert exchange with retention of total knee arthroplasty components. J Bone Joint Surg Am.
2000; 82: 516-523.
6. Griffin, WL., Scott, RD., Dalury, DF., Mahoney, OM., Chiavetta, JB. and Odum, SM. Modular insert exchange in knee arthroplasty for treatment of wear and osteolysis. Clin Orthop Relat Res.
2007; 464: 132-137.
7. Insall, JN., Dorr, LD., Scott, RD. and Scott, WN. Rationale of the Knee Society clinical rating system. Clin Orthop Relat Res.
1989; 248: 13-14.
8. Jensen, CL., Petersen, MM., Jensen, KE., Therbo, M. and Schroder, HM. Outcome of isolated tibial polyethylene insert exchange after uncemented total knee arthroplasty: 27 patients followed for 8-71 months. Acta Orthop.
2006; 77: 917-920. 10.1080/17453670610013222
© 2010 Lippincott Williams & Wilkins, Inc.
9. Whiteside, LA. and Katerberg, B. Revision of the polyethylene component for wear in TKA. Clin Orthop Relat Res.
2006; 452: 193-199. 10.1097/01.blo.0000238802.90090.82