Which are the best treatments for hip fractures in older adults?

Why is this question important?

A hip fracture is a break at the top of the leg bone. We included people with a break just below the ball and socket joint. These types of broken hip are common in older adults whose bones may be fragile because of a condition called osteoporosis. They often happen after a fall from a standing or sitting position. The broken hip can be treated in different ways, and we don't know whether some treatments are better than others.

What are the treatments?

- Replacing the broken hip with an artificial one. This can be done using a hemiarthroplasty (HA), which replaces only the ball part of the joint, and can be unipolar (a single artificial joint) or bipolar which has an additional joint within the HA. A total hip arthroplasty (THA) replaces all of the hip joint, including the socket, and usually has just one artifical joint between the ball and the socket (single articulation) or sometimes two (dual-mobility). All types of artificial joints can be fixed in place with or without bone cement.

- Using metal implants to fix the broken parts of the bones. Pins or screws may be inserted through the two parts of broken bone, or the surgeon may use a 'fixed angle plate' which sits on the outer edge of the broken bone and is attached to the bone with screws or pins. 

- Treatment without an operation, usually requiring a period of rest in bed whilst the leg is held in position using traction with weights.

What did we do?

We searched for studies that compared one or more of these treatments. We wanted to find out the benefits and harms of these different treatments. We combined the findings from studies, and created a 'network' (which is used when researchers perform a 'network meta-analysis' on the results from studies) to see if we could find out if some treatments were better than others.

What did we find?

We found 119 studies, involving 17,653 participants with 17,669 fractures. The average age of study participants ranged from 60 to 87 years; 73% were women, which is usual for people who have this type of hip fracture. We included 75 of these studies in our 'network'.

We found that a modern design of unipolar HA fixed with bone cement, or some of the metal implants (fixed angle plates and pins), seem to have the greatest chance of reducing the number of deaths within 12 months of injury. Compared to people having these treatments, more people who were treated with an uncemented modern bipolar design of HA or with a THA (single articulation) died. 

We didn't find as many studies to include in our 'network' for health-related quality of life, and none of the treatments made a meaningful improvement to people's quality of life. 

We also found that people treated with any of the hip replacements were less likely to need additional surgery on their broken hip than people treated with metal implants or treated without an operation. Amongst all the designs of hip replacements, fewer people needed additional surgery after treatment with a cemented modern unipolar design of HA - but there was not a big difference in the findings for these hip replacement treatments.

So, overall, cemented modern hip replacements tended to produce better outcomes and may be a more successful approach than attempting to fix the broken bone. THA (single articulation) may have increased the risk of death compared with cemented HA, without leading to an important difference in quality of life - but we are not sure about this finding. This type of THA may be an appropriate treatment for some people with these fractures, but we have not studied this in this review.

Are we confident in what we found?

The true effects of these treatments might be very different to what we have found in this review. Many of the studies in this review were published before general reporting standards for research were improved, and so we could not be certain whether or not these studies were well-conducted. Sometimes, the types of fractures were different (particularly between participants treated with metal implants and those treated with hip replacements), and this might have affected the results in the 'network'. We also found that the results included risks of potential benefits and harms, and this is often because there are not enough study participants to find a precise result.

How up to date is this review?

We ran our search in July 2020.     

Authors' conclusions: 

There was considerable variability in the ranking of each treatment such that there was no one outstanding, or subset of outstanding, superior treatments. However, cemented modern arthroplasties tended to more often yield better outcomes than alternative treatments and may be a more successful approach than internal fixation. There is no evidence of a difference between THA (single articulation) and cemented modern unipolar HA in the outcomes measured in this review. THA may be an appropriate treatment for a subset of people with intracapsular fracture but we have not explored this further. 

Read the full abstract...

Hip fractures are a major healthcare problem, presenting a considerable challenge and burden to individuals and healthcare systems. The number of hip fractures globally is rising rapidly. The majority of intracapsular hip fractures are treated surgically.


To assess the relative effects (benefits and harms) of all surgical treatments used in the management of intracapsular hip fractures in older adults, using a network meta-analysis of randomised trials, and to generate a hierarchy of interventions according to their outcomes.

Search strategy: 

We searched CENTRAL, MEDLINE, Embase, Web of Science, and five other databases in July 2020. We also searched clinical trials databases, conference proceedings, reference lists of retrieved articles and conducted backward-citation searches.

Selection criteria: 

We included randomised controlled trials (RCTs) and quasi-RCTs comparing different treatments for fragility intracapsular hip fractures in older adults. We included total hip arthroplasties (THAs), hemiarthroplasties (HAs), internal fixation, and non-operative treatments. We excluded studies of people with hip fracture with specific pathologies other than osteoporosis or resulting from high-energy trauma.

Data collection and analysis: 

Two review authors independently assessed studies for inclusion. One review author completed data extraction which was checked by a second review author. We collected data for three outcomes at different time points: mortality and health-related quality of life (HRQoL) - both reported within 4 months, at 12 months, and after 24 months of surgery, and unplanned return to theatre (at end of study follow-up).

We performed a network meta-analysis (NMA) with Stata software, using frequentist methods, and calculated the differences between treatments using risk ratios (RRs) and standardised mean differences (SMDs) and their corresponding 95% confidence intervals (CIs). We also performed direct comparisons using the same codes.

Main results: 

We included 119 studies (102 RCTS, 17 quasi-RCTs) with 17,653 participants with 17,669 intracapsular fractures in the review; 83% of fractures were displaced. The mean participant age ranged from 60 to 87 years and 73% were women. 

After discussion with clinical experts, we selected 12 nodes that represented the best balance between clinical plausibility and efficiency of the networks: cemented modern unipolar HA, dynamic fixed angle plate, uncemented first-generation bipolar HA, uncemented modern bipolar HA, cemented modern bipolar HA, uncemented first-generation unipolar HA, uncemented modern unipolar HA, THA with single articulation, dual-mobility THA, pins, screws, and non-operative treatment. Seventy-five studies (with 11,855 participants) with data for at least two of these treatments contributed to the NMA.

We selected cemented modern unipolar HA as a reference treatment against which other treatments were compared. This was a common treatment in the networks, providing a clinically appropriate comparison. In order to provide a concise summary of the results, we report only network estimates when there was evidence of difference between treatments.

We downgraded the certainty of the evidence for serious and very serious risks of bias and when estimates included possible transitivity, particularly for internal fixation which included more undisplaced fractures. We also downgraded for incoherence, or inconsistency in indirect estimates, although this affected few estimates. Most estimates included the possibility of benefits and harms, and we downgraded the evidence for these treatments for imprecision. 

We found that cemented modern unipolar HA, dynamic fixed angle plate and pins seemed to have the greatest likelihood of reducing mortality at 12 months. Overall, 23.5% of participants who received the reference treatment died within 12 months of surgery. Uncemented modern bipolar HA had higher mortality than the reference treatment (RR 1.37, 95% CI 1.02 to 1.85; derived only from indirect evidence; low-certainty evidence), and THA with single articulation also had higher mortality (network estimate RR 1.62, 95% CI 1.13 to 2.32; derived from direct evidence from 2 studies with 225 participants, and indirect evidence; very low-certainty evidence). In the remaining treatments, the certainty of the evidence ranged from low to very low, and we noted no evidence of any differences in mortality at 12 months. 

We found that THA (single articulation), cemented modern bipolar HA and uncemented modern bipolar HA seemed to have the greatest likelihood of improving HRQoL at 12 months. This network was comparatively sparse compared to other outcomes and the certainty of the evidence of differences between treatments was very low. We noted no evidence of any differences in HRQoL at 12 months, although estimates were imprecise.

We found that arthroplasty treatments seemed to have a greater likelihood of reducing unplanned return to theatre than internal fixation and non-operative treatment. We estimated that 4.3% of participants who received the reference treatment returned to theatre during the study follow-up. Compared to this treatment, we found low-certainty evidence that more participants returned to theatre if they were treated with a dynamic fixed angle plate (network estimate RR 4.63, 95% CI 2.94 to 7.30; from direct evidence from 1 study with 190 participants, and indirect evidence). We found very low-certainty evidence that more participants returned to theatre when treated with pins (RR 4.16, 95% CI 2.53 to 6.84; only from indirect evidence), screws (network estimate RR 5.04, 95% CI 3.25 to 7.82; from direct evidence from 2 studies with 278 participants, and indirect evidence), and non-operative treatment (RR 5.41, 95% CI 1.80 to 16.26; only from indirect evidence). There was very low-certainty evidence of a tendency for an increased risk of unplanned return to theatre for all of the arthroplasty treatments, and in particular for THA, compared with cemented modern unipolar HA, with little evidence to suggest the size of this difference varied strongly between the arthroplasty treatments.