Patients with hip fractures are a frail group with a recognised perioperative and short-term mortality rate. In Norway, the 30-day mortality of hip fractures is currently at 8%, which is fairly consistent with the UK, but the influence of the actual time to treating the hip fracture on both short- and long-term mortality is not yet clear.

In this interview, Andrew Duckworth is joined by Dr Sunniva Leer-Salvesen and Dr Jan-Erik Gjertsen from Heartland Hospital in Bergen, Norway to discuss their study entitled ‘Does time from fracture surgery affect mortality and interoperative medical complications for hip fracture patients: an observational study of 73,557 patient reported to the Norwegian Hip Fracture Register’, which was published in the September 2019 issue of The Bone & Joint Journal.

Can you give us a brief introduction to the paper and what we currently know about the influence of the actual time to treating hip fractures on both the short- and long-term mortality?

Our paper is a register-based study combining data from the Norwegian Hip Fracture Register, and the Norwegian Patient Registry. We aimed to investigate mortality and risk of interoperative medical complications, depending on time to surgery for hip fracture patients.

Several studies have investigated the effect of surgical delay for our refracture patients. Some studies find a higher in-hospital mortality and a higher 30-day mortality when surgical delay exceeds either 24 hours or 48 hours from admission to surgery. On the other hand, other studies have found no association between time to surgery and the patient! For mortality, surgical delay has been connected to longer lengths of stay and morbidity, as well as things like urinary tract infections and bedsores.

Can you give us a brief overview of the current guidelines in Norway in relation to time to surgery for hip fracture patients?

There were no official guidelines in Norway regarding time to surgery until recently, however, there is a note from the national knowledge centre of the health service from 2015, recommending treatment within 24 hours, if possible, or at least within 48 hours. The national quality indicator for the operation is surgery within 48 hours but in 2018, new multidisciplinary guidelines on treatment of hip fractures were introduced, and they recommend the treatment within 24 hours.

This is a big data study utilising data from the Norwegian Hip Fracture Register and the Norwegian Patient Registry. Can you give us a brief overview of these databases and what they contain and collect routinely?

The Norwegian Hip Fracture Register was established in 2005 to collect nationwide information on the treatment of hip fractures as a basis for improvement of treatment. It now includes data from almost 120,000 patients and is owned by the Norwegian Orthopedic Association and is co-located with the Norwegian Arthroplasty Register in Bergen. The register collects data from each surgery by use of paper forms, which are filled in by surgeons immediately after each operation. The form includes information on the patients such as ASA classification and cognitive function, age and sex, as well as time of fracture, the type of fracture and surgery performed and several details about the surgery, including interoperative complications. Both primary operations and reoperations are included and linked together.

In this study we also used the Norwegian Patient Registry. This is one of the central health registries in Norway; it receives administrative data on patients from all hospitals, including the exact time of admission which we used in our study and using the patient’s identification number, it was possible to link data from bother registers.

What were the primary and secondary outcomes or goals you were looking at for the study itself?

The primary outcome in our study was mortality following hip fracture surgery, and the secondary outcome was risk of reported intraoperative complications for our patients. An important factor is that all existing research to our knowledge targets hospital delay, meaning the time from admission to surgery. We chose to study both the pre-hospital hours and hospital hours, then targeted the total delay for any fracture patients.

You performed two analyses, with different numbers of hip fracture patients in each; can you explain why that is?

We used two different patient groups when investigating total delay and hospital delay. As mentioned, we used data from the Norwegian Hip Fracture Register and the Norwegian Patient Registry; time of fracture is gathered from the Norwegian Hip Fracture Register by forms filled out by the independent surgeons. In about 47% of the cases, we do not know the exact time of fracture or the exact time of fracture isn't reported. So, we ended up with about 39,000 patients when investigating total delay.

When we investigated the hospital delay, we gathered information on time of admission from the Norwegian Patient Registry. This is administratively reported and constitutes the grounds for activity-based financing for hospitals. Consequently, these data are nearly complete, and we were able to study about 73,000 patients when investigating hospital delay.

Can you give us a concise overview of the analyses you performed?

For our primary outcome (mortality) we used Cox multiple regression models to compare the relative risks of postoperative death among patients divided into groups based on the preoperative delay. We adjusted our analyses for the possible influences of age, sex and comorbidity using ASA score and types of surgery, as well as type of fracture. Further on we divided our patients into groups based on 12-hour intervals either from the time of fracture or the time of admission to surgery.

Can you give us the detailed key findings in relation to the total delay and mortality?

First of all, we found no effect of total delay on surgery as long as operations were performed within 48 hours of the fracture. However, total delay more than 48 hours was associated with increased mortality after three days postoperatively and one-year postoperatively for any fracture patients.

Beyond this, we went on to study comorbidity and stratify the patients into the healthcare ASA 1 to 2 group, and then more comorbid patients with an ASA class of 3 to 5. For these patients, we found that total delay of more than 48 hours was associated with an increased three-day mortality for the comorbid patients while the healthier patient groups did not have an altered mortality.

We went on to study patients based on the type of surgery performed. In patients receiving osteosynthesis, time to surgery did not seem to influence the patient mortality at any point of observation.

What did you find in terms of the hospital delay and the sort of medical complications that can potentially develop?

We went on to study the hospital delay and, on average, our patients waited 22 hours from admission to surgery in our study. We found out patients that waited more than 24 hours for surgery had an increased risk of reported intraoperative medical complications. In conclusion, the limit for unwanted outcomes was even lower when studying complications compared with mortality.

These are interesting findings and very robust in such a large group of patients. What do you feel are the key findings of the work?

The key finding is that hip fracture should be operated on within 48 hours of fracture, not within 48 hours of admission. Further, the number of reported medical complications were higher when hospital delay exceeded 24 hours. But there are some limitations, even if we adjust for age, sex and ASA class, there may be confounders that we did not adjust for. For example, it is difficult to conclude whether the medical complications occurred due to the long waiting time to surgery, or whether the patients waited longer to surgery because they have comorbidities that had to be stabilised before surgery and that they, therefore, had an increased risk of incorporative medical complications that we were not able to adjust for.

When you look at the results from previous studies, particularly the details of the fast track, hip fracture, how do you feel that these are related to your study?

I think our results compare well both with the new Norwegian multiple disciplinary guidelines recommending surgery within 24 hours, and also with, for example, the NICE guidelines in the UK which recommend treatment within 36 hours.

Our results also compare well to a large meta-analysis from Shiga which found increased one-year mortality and 30-day mortality for patients waiting longer than 48 hours.

Do you feel that surgery was delayed because of a patient’s increased risk of unwanted outcomes, or do the unwanted outcomes occur due to the delayed surgery?

It's a hard question to answer because we know that both in earlier studies and our own study, a high level of comorbidity has been reported among patients with a long hospital delay compared with the patients with earlier surgical interventions. Therefore, we are potentially facing a question of confounding by indication in our study. We think that to fully investigate the dilemma, we need other studies to understand our observational register study, and randomised controlled trials, where can target the consequences of comorbidity and preoperative management in a deeper way.

What do you feel are the potential implications of this study moving forward, back in Norway, but also worldwide?

I think it's important to focus on the time from facture to surgery, particularly in Norway where there is a long journey to a hospital for many patients. Most other studies and guidelines focus on the time from admission to surgery, and although waiting time to surgery is a very vulnerable time for the patients, the most vulnerable period is probably the time between fracture and admission because in that period patients are in a lot of pain. They are immobilised, perhaps on the floor, maybe for several hours if they live alone, before they get help. A patient who has had a long pre-hospital waiting period must be prioritised for surgery after admission.

Our results also show that we don't have to rush too much to surgery. The results do not suggest that hip fractures must be treated faster than 24 hours; there’s probably no need to operate on hip fractures at night.

If you’d like to read the full paper you can do so here. You can listen to the podcast version of this interview here.