Aims. One-stage revision hip arthroplasty for periprosthetic joint infection (PJI) has several advantages; however, resection of the proximal femur might be necessary to achieve higher success rates. We investigated the risk factors for resection and re-revisions, and assessed complications and subsequent re-revisions. Methods. In this single-centre, case-control study, 57 patients who underwent one-stage revision arthroplasty for PJI of the hip and required resection of the proximal femur between 2009 and 2018 were identified. The control group consisted of 57 patients undergoing one-stage revision without bony resection. Logistic regression analysis was performed to identify any correlation with resection and the risk factors for re-revisions. Rates of all-causes re-revision, reinfection, and instability were compared between groups. Results. Patients who required resection of the proximal femur were found to have a higher all-cause re-revision rate (29.8% vs 10.5%; p = 0.018), largely due to reinfection (15.8% vs 0%; p = 0.003), and dislocation (8.8% vs 10.5%; p = 0.762), and showed higher rate of in-hospital wound haematoma requiring aspiration or evacuation (p = 0.013), and wound revision (p = 0.008). The use of of dual mobility components/constrained liner in the resection group was higher than that of controls (94.7% vs 36.8%; p < 0.001). The presence and removal of additional metal hardware (odds ratio (OR) = 7.2), a sinus tract (OR 4), ten years’ time interval between primary implantation and index infection (OR 3.3), and previous hip revision (OR 1.4) increased the risk of proximal femoral resection. A sinus tract (OR 9.2) and postoperative dislocation (OR 281.4) were associated with increased risk of subsequent re-revisions. Conclusion. Proximal femoral resection during one-stage revision hip arthroplasty for PJI may be required to reduce the risk of of recurrent or further infection. Patients with additional metalware needing removal or transcortical sinus tracts and chronic osteomyelitis are particularly at higher risk of needing proximal femoral excision. However, radical resection is associated with higher surgical complications and increased re-revision rates. The use of constrained acetabular liners and dual mobility components maintained an acceptable dislocation rate. These results, including identified risk factors, may aid in preoperative planning, patient consultation and consent, and intraoperative decision-making. Cite this article: Bone Joint J 2021;103-B(11):1678–1685

Aims. There is an increasing demand for hip arthroplasty in China. We aimed to describe trends in in-hospital mortality after this procedure in China and to examine the potential risk factors. Patients and Methods. We included 210 450 patients undergoing primary hip arthroplasty registered in the Hospital Quality Monitoring System in China between 2013 and 2016. In-hospital mortality after hip arthroplasty and its relation to potential risk factors were assessed using multivariable Poisson regression. Results. During the study period, 626 inpatient deaths occurred within 30 days after hip arthroplasty. Mortality decreased from 2.9% in 2013 to 2.6% in 2016 (p for trend = 0.02). Compared with their counterparts, old age, male sex, and divorced or widowed patients had a higher rate of mortality (all p < 0.05). Risk ratio (RR) for mortality after arthroplasty for fracture was two-fold higher (RR 2.0, 95% confidence interval (CI) 1.5 to 2.6) than that for chronic disease. RRs for mortality were 3.3 (95% CI 2.7 to 3.9) and 8.2 (95% CI 6.5 to 10.4) for patients with Charlson Comorbidity Index (CCI) of 1 to 2 and CCI ≥ 3, respectively, compared with patients with CCI of 0. The rate of mortality varied according to geographical region, the lowest being in the East region (1.8%), followed by Beijing (2.1%), the North (2.9%), South-West (3.6%), South-Central (3.8%), North-East (4.1%), and North-West (5.2%) regions. Conclusion. While in-hospital mortality after hip arthroplasty in China appears low and declined during the study period, discrepancies in mortality after this procedure exist according to sociodemographic factors. Healthcare resources should be allocated more to underdeveloped regions to further reduce mortality. Cite this article: Bone Joint J 2019;101-B:1209–1217

Aims. Total hip arthroplasty (THA) is one of the most successful surgical procedures. The objectives of this study were to define whether there is a correlation between socioeconomic status (SES) and surgical complications after elective primary unilateral THA, and investigate whether access to elective THA differs within SES groups. Methods. We conducted a retrospective, population-based cohort study involving 202 hospitals in Ontario, Canada, over a 17-year period. Patients were divided into income quintiles based on postal codes as a proxy for personal economic status. Multivariable logistic regression models were then used to primarily assess the relationship between SES and surgical complications within one year of index THA. Results. Of 111,359 patients who underwent elective primary THA, those in the lower SES groups had statistically significantly more comorbidities and statistically significantly more postoperative complications. While there was no increase in readmission rates within 90 days, there was a statistically significant difference in the primary and secondary outcomes including all revisions due (with a subset of deep wound infection and dislocation). Results showed that those in the higher SES groups had proportionally more cases performed than those in lower groups. Compared to the highest SES quintile, the lower groups had 61% of the number of hip arthroplasties performed. Conclusion. Patients in lower socioeconomic groups have more comorbidities, fewer absolute number of cases performed, have their procedures performed in lower-volume centres, and ultimately have higher rates of complications. This lack of access and higher rates of complications is a “double hit” to those in lower SES groups, and indicates that we should be concentrating efforts to improve access to surgeons and hospitals where arthroplasty is routinely performed in high numbers. Even in a universal healthcare system where there are no penalties for complications such as readmission, there seems to be an inequality in the access to THA. Cite this article: Bone Joint J 2022;104-B(5):589–597

Aims. This study describes the variation in the annual volumes of revision hip arthroplasty (RHA) undertaken by consultant surgeons nationally, and the rate of accrual of RHA and corresponding primary hip arthroplasty (PHA) volume for new consultants entering practice. Methods. National Joint Registry (NJR) data for England, Wales, Northern Ireland, and the Isle of Man were received for 84,816 RHAs and 818,979 PHAs recorded between April 2011 and December 2019. RHA data comprised all revision procedures, including first-time revisions of PHA and any subsequent re-revisions recorded in public and private healthcare organizations. Annual procedure volumes undertaken by the responsible consultant surgeon in the 12 months prior to every index procedure were determined. We identified a cohort of ‘new’ HA consultants who commenced practice from 2012 and describe their rate of accrual of PHA and RHA experience. Results. The median annual consultant RHA volume, averaged across all cases, was 21 (interquartile range (IQR) 11 to 34; range 0 to 181). Of 1,695 consultants submitting RHA cases within the study period, the top 20% of surgeons by annual volume performed 74.2% of total RHA case volume. More than half of all consultants who had ever undertaken a RHA maintained an annual volume of just one or fewer RHA, however, collectively contributed less than 3% of the total RHA case volume. Consultant PHA and RHA volumes were positively correlated. Lower-volume surgeons were more likely to undertake RHA for urgent indications (such as infection) as a proportion of their practice, and to do so on weekends and public holidays. Conclusion. The majority of RHAs were undertaken by higher-volume surgeons. There was considerable variation in RHA volumes by indication, day of the week, and between consultants nationally. The rate of accrual of RHA experience by new consultants is low, and has important implications for establishing an experienced RHA consultant workforce. Cite this article: Bone Joint J 2023;105-B(10):1060–1069

Aims. Femoral cement-in-cement revision is a well described technique to reduce morbidity and complications in hip revision surgery. Traditional techniques for septic revision of hip arthroplasty necessitate removal of all bone cement from the femur. In our two centres, we have been using a cement-in-cement technique, leaving the distal femoral bone cement in selected patients for septic hip revision surgery, both for single and the first of two-stage revision procedures. A prerequisite for adoption of this technique is that the surgeon considers the cement mantle to be intimately fixed to bone without an intervening membrane between cement and host bone. We aim to report our experience for this technique. Methods. We have analyzed patients undergoing this cement-in-cement technique for femoral revision in infection, and present a consecutive series of 89 patients. Follow-up was undertaken at a mean of 56.5 months (24.0 to 134.7) for the surviving cases. Results. Seven patients (7.9%) required further revision for infection. Ten patients died of causes unrelated to their infection before their two-year review (mean 5.9 months; 0.9 to 18.6). One patient was lost to follow-up at five months after surgery, and two patients died of causes unrelated to their hip shortly after their two-year review was due without attending. Of the remaining patients, 69 remained infection-free at final review. Radiological review confirms the mechanical success of the procedure as previously described in aseptic revision, and postoperative Oxford Hip Scores suggest satisfactory functional outcomes. Conclusion. In conclusion, we found that retaining a well-fixed femoral cement mantle in the presence of infection and undertaking a cement-in-cement revision was successful in 82 of the patients (92.1%) in our series of 89, both in terms of eradication of infection and component fixation. These results are comparable to other more invasive techniques and offer significant potential benefits to the patient. Cite this article: Bone Joint J 2022;104-B(2):212–220

Aims. The purpose of this study was to evaluate unexpected positive cultures in total hip arthroplasty (THA) revisions for presumed aseptic loosening, to assess the prevalence of low-grade infection using two definition criteria, and to analyze its impact on implant survival after revision. Methods. A total of 274 THA revisions performed for presumed aseptic loosening from 2012 to 2016 were reviewed. In addition to obtaining intraoperative tissue cultures from all patients, synovial and sonication fluid samples of the removed implant were obtained in 215 cases (79%) and 101 cases (37%), respectively. Histopathological analysis was performed in 250 cases (91%). Patients were classified as having low-grade infections according to institutional criteria and Musculoskeletal Infection Society (MSIS) International Consensus Meeting (ICM) 2013 criteria. Low-grade infections according to institutional criteria were treated with targeted antibiotics for six weeks postoperatively. Implant failure was defined as the need for re-revision resulting from periprosthetic joint infection (PJI) and aseptic reasons. The mean follow-up was 68 months (26 to 95). Results. Unexpected positive intraoperative samples were found in 77 revisions (28%). Low-grade infection was diagnosed in 36 cases (13%) using institutional criteria and in nine cases (3%) using MSIS ICM 2013 criteria. In all, 41 patients (15%) had single specimen growth of a low-virulent pathogen and were deemed contaminated. Coagulase-negative Staphylococcus and anaerobes were the most commonly isolated bacteria. Implant failure for PJI was higher in revisions with presumed contaminants (5/41, 12%) compared to those with low-grade infections (2/36, 6%) and those with negative samples (5/197, 3%) (p = 0.021). The rate of all-cause re-revision was similar in patients diagnosed with low-grade infections (5/36, 14%) and those with presumed contaminants (6/41, 15%) and negative samples (21/197, 11%) (p = 0.699). Conclusion. Our findings suggest that the presumption of culture contamination in aseptic revision hip arthroplasty may increase the detection of PJI. In this cohort, the presence of low-grade infection did not increase the risk of re-revision. Further studies are needed to assess the relevance of single specimen growth and the benefits of specific postoperative antibiotic regimens. Cite this article: Bone Joint J 2021;103-B(6):1070–1077

Aims. We investigated the long-term performance of the Tripolar Trident acetabular component used for recurrent dislocation in revision total hip arthroplasty. We assessed: 1) rate of re-dislocation; 2) incidence of complications requiring re-operation; and 3) Western Ontario and McMaster Universities osteoarthritis index (WOMAC) pain and functional scores. Patients and Methods. We retrospectively identified 111 patients who had 113 revision tripolar constrained liners between 1994 and 2008. All patients had undergone revision hip arthroplasty before the constrained liner was used: 13 after the first revision, 17 after the second, 38 after the third, and 45 after more than three revisions. A total of 75 hips (73 patients) were treated with Tripolar liners due to recurrent instability with abductor deficiency, In addition, six patients had associated cerebral palsy, four had poliomyelitis, two had multiple sclerosis, two had spina bifida, two had spondyloepiphyseal dysplasia, one had previous reversal of an arthrodesis, and 21 had proximal femoral replacements. The mean age of patients at time of Tripolar insertions was 72 years (53 to 89); there were 69 female patients (two bilateral) and 42 male patients. All patients were followed up for a mean of 15 years (10 to 24). Overall, 55 patients (57 hips) died between April 2011 and February 2018, at a mean of 167 months (122 to 217) following their tripolar liner implantation. We extracted demographics, implant data, rate of dislocations, and incidence of other complications. Results. At ten years, the Kaplan–Meier survivorship for dislocation was 95.6% (95% confidence interval (CI) 90 to 98), with 101 patients at risk. At 20 years, the survivorship for dislocation was 90.6% (95% CI 81.0 to 95.5), with one patient at risk. Eight patients (7.2%) had a dislocation of their constrained liners. At ten years, the survival to any event was 89.4% (95% CI 82 to 93.8), with 96 patients at risk. At 20 years, the survival to any event was 82.5% (95% CI 71.9 to 89.3), with one patient at risk. Five hips (4.4%) had deep infection. Two patients (1.8%) developed dissociated constraining rings with pain but without dislocation, which required re-operation. Two patients (1.8%) had periprosthetic femoral fractures, without dislocation, that were treated by revision stems along with exchange of the well-functioning constrained liners. Conclusion. Constrained tripolar liners used at revision hip arthroplasty provided favourable results in the long term for treatment of recurrent dislocation and for patients at high risk of dislocation. Cite this article: Bone Joint J 2019;101-B(6 Supple B):123–126

Aims

The duration of systemic antibiotic treatment following first-stage revision surgery for periprosthetic joint infection (PJI) after total hip arthroplasty (THA) is contentious. Our philosophy is to perform an aggressive debridement, and to use a high local concentration of targeted antibiotics in cement beads and systemic prophylactic antibiotics alone. The aim of this study was to assess the success of this philosophy in the management of PJI of the hip using our two-stage protocol.

Two major complications of hip replacement are loosening and infection. Reliable differentiation between these pathological processes is difficult since both may be accompanied by similar symptoms. Our aim was to assess the diagnostic ability of triple-phase bone scanning (TPBS) and positron-emission tomography (PET) to detect and differentiate these complications in patients with a hip arthroplasty. Both TPBS and PET were performed in 63 patients (92 prostheses). The radiotracer for PET imaging was . 18. F-fluorodeoxyglucose (FDG). Image interpretation was performed according to qualitative and quantitative criteria although the final diagnosis was based upon either surgical findings or clinical follow-up. The sensitivity, specificity and accuracy of PET was 0.94, 0.95 and 0.95 respectively, compared with 0.68, 0.76 and 0.74 for TPBS. We found that an image interpretation based exclusively upon quantitative criteria was inappropriate because of its low selectivity. The histological examination indicated that increased periprosthetic uptake of FDG in patients with aseptic loosening was caused by wear-induced polyethylene particles and the subsequent growth of aggressive granulomatous tissue

Juvenile idiopathic arthritis (JIA) is a chronic disease of childhood; it causes joint damage which may require surgical intervention, often in the young adult. The aim of this study was to describe the long-term outcome and survival of hip replacement in a group of adult patients with JIA and to determine predictors of survival for the prosthesis. In this retrospective comparative study patients were identified from the database of a regional specialist adult JIA clinic. This documented a series of 47 hip replacements performed in 25 adult patients with JIA. Surgery was performed at a mean age of 27 years (11 to 47), with a mean follow-up of 19 years (2 to 36). The mean Western Ontario and McMaster Universities osteoarthritis index questionnaire (WOMAC) score at the last follow-up was 53 (19 to 96) and the mean Health Assessment Questionnaire score was 2.25 (0 to 3). The mean pain component of the WOMAC score (60 (20 to 100)) was significantly higher than the mean functional component score (46 (0 to 97)) (p = 0.02). Kaplan-Meier survival analysis revealed a survival probability of 46.6% (95% confidence interval 37.5 to 55.7) at 19 years, with a trend towards enhanced survival with the use of a cemented acetabular component and a cementless femoral component. This was not, however, statistically significant (acetabular component, p = 0.76, femoral component, p = 0.45). Cox’s proportional hazards regression analysis showed an implant survival rate of 54.9% at 19 years at the mean of covariates. Survival of the prosthesis was significantly poorer (p = 0.001) in patients who had been taking long-term corticosteroids and significantly better (p = 0.02) in patients on methotrexate

The in-cement technique for revision hip arthroplasty involves retaining the original cement-bone interface. This has been proven to be a biomechanically stronger method than recementing after complete removal of the original cement mantle. This study reviewed a series of 54 consecutive revision hip arthroplasty procedures, using the in-cement technique, between November 1999 and November 2003. Clinical and radiological follow-up included functional assessment. There were 54 procedures performed in 51 patients, whose mean age at surgery was 70.3 years (45 to 85). A total of 42 were available at a mean follow-up of 29.2 months (6 to 51). There was no radiological evidence of loosening. Functional assessments were available for 40 patients who had a mean Harris hip score of 85.2 (51.9 to 98.5), a mean Oxford hip score of 19.6 (12 to 41), a mean UCLA activity profile score of 5.9 (3 to 8) and a mean SF-36 score of 78.0 (31.6 to 100). The in-cement technique provides consistent, high functional outcomes and should be considered in appropriately selected cases

Aims. The purpose of this study was to validate the diagnosis of periprosthetic joint infection (PJI) in the Danish Hip Arthroplasty Register (DHR). Patients and Methods. We identified a cohort of patients from the DHR who had undergone primary total hip arthroplasty (THA) since 1 January 2005 and followed them until first-time revision, death, emigration or until 31 December 2012. Revision for PJI, as registered in the DHR, was validated against a benchmark which included information from microbiology databases, prescription registers, clinical biochemistry registers and clinical records. We estimated the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for PJI in the DHR alone and in the DHR when combined with microbiology databases. Results. In total, 1382 of the 37 826 primary THAs in the DHR were registered as having been revised for any cause once 26 patients with errors in registration had been excluded: 232 of these were for PJI. For this group, the sensitivity was 67%, specificity 95%, PPV 77%, and NPV 92%. Combining the data from the DHR with those from microbiology databases increased the sensitivity to 90% and also improved specificity (100%), PPV (98%) and NPV (98%). Conclusion. Only two thirds of revisions for PJI were captured in the DHR and only 77% of the PJI reported to the DHR could be confirmed to be infected. . Take home message: combining the data from the DHR with those from microbiology databases substantially improved the validity of the diagnosis of PJI and should enable future register-based studies. Cite this article: Bone Joint J 2016;98-B:320–5

Aims

To determine mortality risk after first revision total hip arthroplasty (THA) for periprosthetic femoral fracture (PFF), and to compare this to mortality risk after primary and first revision THA for other common indications.

We implanted 300 uncoated cementless PM prostheses into 271 patients and followed 251 (92.6%) of them for four to seven years. By then 37 had already been revised for aseptic and three for septic loosening. The survival rate with implant failure as the endpoint was 88.8% for the cup and 85.3% for the stem after six years. There was a higher risk of implant loosening in congenital dysplasia, unilateral hip arthroplasty and obesity. The results of 225 unrevised hip replacements were assessed by questionnaire. Only 27.4% of the patients were completely free from pain and 17.9% had pain on walking any distance or at all times. The walking distance was for less than 30 minutes in 40%. Because of the poor results in comparison with other prostheses we do not recommend further use of the uncoated PM prosthesis

We retrospectively reviewed 45 hip arthroplasties which were performed over a period of 20 years in 38 patients with cirrhosis of the liver. There was a high perioperative 30-day complication rate (26.7%). Advanced cirrhosis was associated with a higher risk of complications (p = 0.004) as also was increased age, a high level of creatinine, a low level of albumin, a low platelet count, ascites, encephalopathy and an increased operative blood loss. The survival of the prosthesis at five years was 77.8% and infection was a major cause of failure. In view of the high rate of early complications and the limited longevity of the prosthesis, surgeons who perform hip arthroplasty on such patients should counsel them appropriately preoperatively

We undertook a randomised controlled trial to compare the piriformis-sparing approach with the standard posterior approach used for total hip replacement (THR). We recruited 100 patients awaiting THR and randomly allocated them to either the piriformis-sparing approach or the standard posterior approach. Pre- and post-operative care programmes and rehabilitation regimes were identical for both groups. Observers were blinded to the allocation throughout; patients were blinded until the two-week assessment. Follow-up was at six weeks, three months, one year and two years. In all 11 patients died or were lost to follow-up.

There was no significant difference between groups for any of the functional outcomes. However, for patients in the piriformis-sparing group there was a trend towards a better six-minute walk test at two weeks and greater patient satisfaction at six weeks. The acetabular components were less anteverted (p = 0.005) and had a lower mean inclination angle (p = 0.02) in the piriformis-sparing group. However, in both groups the mean component positions were within Lewinnek’s safe zone. Surgeons perceived the piriformis-sparing approach to be significantly more difficult than the standard approach (p = 0.03), particularly in obese patients.

In conclusion, performing THR through a shorter incision involving sparing piriformis is more difficult and only provides short-term benefits compared with the standard posterior approach.

Aims. The pelvis rotates in the sagittal plane during daily activities. These rotations have a direct effect on the functional orientation of the acetabulum. The aim of this study was to quantify changes in pelvic tilt between different functional positions. Patients and Methods. Pre-operatively, pelvic tilt was measured in 1517 patients undergoing total hip arthroplasty (THA) in three functional positions – supine, standing and flexed seated (the moment when patients initiate rising from a seated position). Supine pelvic tilt was measured from CT scans, standing and flexed seated pelvic tilts were measured from standardised lateral radiographs. Anterior pelvic tilt was assigned a positive value. Results. The mean pelvic tilt was 4.2° (-20.5° to 24.5°), -1.3° (-30.2° to 27.9°) and 0.6° (-42.0° to 41.3°) in the three positions, respectively. The mean sagittal pelvic rotation from supine to standing was -5.5° (-21.8° to 8.4°), from supine to flexed seated was -3.7° (-48.3° to 38.6°) and from standing to flexed seated was 1.8° (-51.8° to 39.5°). In 259 patients (17%), the extent of sagittal pelvic rotation could lead to functional malorientation of the acetabular component. Factoring in an intra-operative delivery error of ± 5° extends this risk to 51% of patients. Conclusion. Planning and measurement of the intended position of the acetabular component in the supine position may fail to predict clinically significant changes in its orientation during functional activities, as a consequence of individual pelvic kinematics. Optimal orientation is patient-specific and requires an evaluation of functional pelvic tilt pre-operatively. Cite this article: Bone Joint J 2017;99-B:184–91

We present the development and results of a nationwide, prospective, observational follow-up programme including patient-reported outcome measures (PROMs) for the Swedish Hip Arthroplasty Register. The programme started in 2002 and has gradually expanded to include all units performing total hip replacement in Sweden. The self-administered PROMs protocol comprises the EQ-5D instrument, the Charnley class categorisation and visual analogue scales for pain and satisfaction. These current analyses include 34 960 total hip replacements with complete pre- and one-year post-operative questionnaires. Patients eligible for total hip replacement generally report low health-related quality of life and suffer from pain. One year post-operatively the mean EQ-5D index increased to above the level of an age- and gender-matched population, with a considerable reduction of pain (p < 0.001). Females, younger patients and those with Charnley category C reported a lower EQ-5D index pre-operatively than males, older patients and Charnley category A or B, respectively (all p < 0.001). In a multivariable regression analysis Charnley category C, male gender and higher age were associated with less improvement in health-related quality of life (p < 0.001). Nationwide implementation of a PROMs programme requires a structured organisation and effective data capture. Patients’ response rates to the Registry are good. The continuous collection of PROMs permits local and national improvement work and allows for further health-economic evaluation

Aims

The rate of dislocation when traditional single bearing implants are used in revision total hip arthroplasty (THA) has been reported to be between 8% and 10%. The use of dual mobility bearings can reduce this risk to between 0.5% and 2%. Dual mobility bearings are more expensive, and it is not clear if the additional clinical benefits constitute value for money for the payers. We aimed to estimate the cost-effectiveness of dual mobility compared with single bearings for patients undergoing revision THA.

We have used the Oxford hip score to monitor the progress of 1908 primary and 279 revision hip replacements undertaken since the start of 1995. Our review programme began in early 1999 and has generated 3900 assessments. The mean pre-operative scores for primary and revision cases were 40.95 and 40.11, respectively. The mean annual score for primary replacement at between 12 and 84 months ranged between 20.60 and 22.57. A comparison of cross-sectional and longitudinal data showed no significant differences. All post-operative reviews showed a significant improvement (p ≤ 0.0001). The 50- to 60-year-old group scored significantly better than the patients over 80 years of age up to 48 months (p < 0.01). A subgroup of 826 National Health Service (NHS) and 397 private patients, treated by the senior author (2292 Oxford assessments), had a higher (i.e. worse) mean pre-operative score for the NHS patients (p ≤ 0.001). The private patients scored better than the NHS group up to 84 months (p < 0.05). Patients treated by a surgeon performing more than 100 replacements each year had a significantly better outcome up to five years than those operated on by surgeons performing fewer than 20 replacements each year. The age of the patients at the time of operation, and their pre-operative level of disability, have both been identified as affecting the long-term outcome. Awareness of the influence of these factors should assist surgeons to provide balanced advice