Our findings show that hip and knee arthroplasty in our DH had a low complication rate, not only comparable to our TH, but also to other centres from high-income countries (HIC). 32–34,41
A recent 10-week multicentred prospective observational study has also shown that arthroplasty at district hospitals and tertiary hospitals in South Africa had comparable peri-operative morbidity results as seen in our study. 35
The average length of stay at the TH (7.6 ± 7.1 days) was longer than at the DH (5.2 ± 2.0 days) with a p value of 0.0001. This is likely attributed to the fact that the TH catered for a higher proportion of patients with multiple co-morbidities and higher ASA grades, requiring multidisciplinary co-management and additional medical support.
In our study the mortality rate did not differ significantly between the two hospitals (DH 0.4% vs TH 0.6%; p 0.76) despite the differences in patients ASA grades in the 2 cohorts. The mortality rates are also comparable to international reports from HIC. The one-year mortality rates in the National Joint Registry for England and Wales (NJR) were 10.8 and 8.9 per 1,000 patient-years after hip and knee arthroplasty, respectively. 32–34 The cause of death at the district hospital was attributed to a thromboembolic event more than 2 weeks after surgery and to perioperative cardiorespiratory complications in patients with ASA grade III and IV at the TH. Although previous studies indicated the risk of mortality following surgery in patients across Africa is twice as high as the global average, this was not reflected in our study. 36
Our study showed an average blood loss of 1.41g/dL and 1.66 g/dL for DH and TH respectively. This is consistent with the reported average decline in haemoglobin of 2.35 g/dL ± 1.14 and 2.29 g/dL ± 1.16 for THA and TKA.37 In our TH, 26 (3.27%) patients received blood transfusion compared to 3 (1.31%) at the DH. This finding again may be a reflection of more complex cases (e.g.; inflammatory arthropathies) being done at TH level with longer surgical time and possibly higher transfusion requirements.
The readmission rate of the TH was 4.40%, which was less than for the DH (1.75%), while still being within a range which is acceptable and reported in other centres38. Reasons for re-admission varied from infection, knee stiffness, wound-related problems, or cardiovascular complications. There were 3 TKR (1.04%) revision surgeries from district level. Reasons for revisions included aseptic loosening, early PJI and a polyethylene insert dislodgement.
At tertiary level 33 (4.1%) patients required revision surgery. The most common indication for revision surgery was PJI (see Fig. 5). The higher revision rate from the TH could be the result of more complex pathology as well as higher ASA grade of patients treated at the TH.
There were 11 (1.38%) peri-prosthetic fractures recorded in the tertiary group, 4 (0.50%) occurred intra-operatively, all during THR (2 femur and 2 acetabular fractures). The district cohort had 2 (0.87%) peri-prosthetic fractures, both intra-operative tibial fractures during TKR. Our cohort incidence of intraoperative periprosthetic fractures is comparable to that reported in the analysis of International Registry data (0.8%) by Pivec et al.39 There were no dislocations recorded at district level, 5 (0.62%) at tertiary level. This falls within the acceptable range of 2–3%.40 The rates of thromboembolic events recorded at TH and DH (1.01% and 1.75% respectively) were similar to those reported in the literature.38,39 (see Table 4) One thromboembolic event in each hospital resulted in death.
Leg length discrepancy after a THA is one of the major causes of patient dissatisfaction as demonstrated by Fujimaki et al.41 Both hospitals showed analogous percentages of LLD; 3.65% (29) vs 4.38% (10) in tertiary and district level respectively. (see Fig. 3) However, a limitation of this study is that the change in leg length was not measured quantitatively but was rather a qualitative observation by the treating surgeons.
Overall tertiary and district level surgery had similar rates of physiological or mechanical concerns (11 vs 10%) (see Table 5), although many of these variables share the same concern as for LLD and were not measured comparisons.
The strength of this study is that it documents that an experienced surgeon using well accepted standards and techniques can perform TJA safely in a DH given careful patient selection (as per ASA classification and theatre protocols as discussed above). However, we must acknowledge its shortcomings, in that the study was conducted in a single district center with a relatively small sample size, and therefore results may not be generalisable or extrapolated to other district centers.
Only two state-funded facilities with a large discrepancy in the cohort sizes were assessed. Limited by the date of commencement of arthroplasty at the DH; the TH cohort had more than three times the number of patients. An important dispersion is generated, which makes it difficult to obtain statistically significant values. There was also no matching of cohorts to control for potential cofounders such as the discrepancy in comorbidity profile of patients. Nevertheless, we can see that low risk patients can safely have THR and TKR in the district setting provided the surgical methodology is up to the same standards as the TH. Another limitation is a relatively high loss to follow up; in low-income setting follow-up is often challenging and highlights the distances travelled by patients to access arthroplasty services.22