Correlation between patient age at total hip replacement surgery and lifeexpectancy

ABSTRACT Total hip arthroplasty (THA) is one of the most cost-effective hip surgeries among orthopedic procedures. We conducted an extensive literature review with 5,394 papers regarding survival rates after THA. We searched PubMed, Embase and the Cochrane library from January 1st, 1970 to July 31th, 2014 looking for all citations about total hip arthroplasty with a long term follow-up (longer than 10 years). The criteria were rigorous: no loss of follow-up, and follow-up more than 10 years. The authors should have known the complete history of patients (whether the patient is alive or dead). Considering the criteria, we found only 15 papers. To evaluate the relationship between follow-up and survival, a linear regression analysis was applied. Considering the papers analysed, and applying the search criteria, we obtained a mean age for the patients above 57.5 years. The chance of survival at 15 years was 57.6%, at 20 years it was 34.6% and at 25 years it is only 11.6%. The relationship between follow-up and survival was significantly linear (p <0.001). Only 11.6% of patients undergoing THA will be still alive 25 years after the surgical procedure. Level of Evidence I, Prognostic Study.


INTRODUCTION
Total hip arthroplasty (THA) is one of the most cost-effective surgeries among orthopaedic procedures. Several authors have suggested that patients undergoing this procedure have a higher survival rate than the general population. [1][2][3][4][5][6][7] However, there are still no publications presenting a clear and unequivocal form of the survival of patients after the performance of total hip arthroplasty. In order to find the answer for the survival of patients after the performance of total hip arthroplasty, we conducted an extensive literature review. After careful review, we selected only papers including a long follow-up (mean follow-up of more than 10 years), known mortality, and in which there was no loss of follow-up regarding the patients.

PATIENTS AND METHODS
A review of the literature was conducted. We searched the computerised databases PubMed, Embase and the Cochrane library from January 1 st , 1970 to July 31 th , 2014 looking for all citations about total hip arthroplasty or hip replacement, and long follow-up (longer than 10 years). All papers with a mean follow-up smaller than 10 years, and papers presenting lost of follow-up were excluded from the study. Most of the initially found papers present significant lack of important information, which compromises the statistical analysis. It was fundamental that the authors know how many patients had died or continued to live at the end of the follow-up. The searching words were: hip arthroplasty or/and hip replacement and follow-up. The search is very sensitive. We found 27,694 papers on the initial search. After a preliminary analysis, 5394 papers were selected. We excluded biomechanical studies, reviews of any kind, surveys, laboratory or radiologic, and experimental animal studies. Then the eligible publications that met the inclusion criteria were evaluated. Initially, we included only papers with minimum follow-up of 10 years, in which the authors cited the number of deaths, and in which there was no lost of follow-up higher than 10%. In the first stage, the authors reviewed all abstracts of articles on THA with a follow-up over 10 years. At this stage, a total of 124 out of 5394 papers were selected.
In the second stage, these 124 articles were read in full and their references were reviewed to identify other papers that might have been lost in the first selection. At the end of the second phase, we found only 44 articles with less than 10% follow-up loss, totalling 13,357 patients. Of these 44 papers, only 15 presented no lost of follow-up, and the authors therefore knew how many patients had died with a minimum follow-up of 10 years. The phrase "no loss of follow-up" should be present somewhere in the text. After this selection, the mean follow-up, patient survival and projections for 15, 20 and 25 years were calculated. Based on this information, an equation that best describes the relationship between these factors was obtained. In performing this calculation, we consider that the boards of living populations of the analysed papers were the same, or very similar, as was the distribution of deaths during follow-up. Thus, it was possible to use the average as a representation of global behaviour. To evaluate the relationship between follow-up and survival, a linear regression analysis was applied, considering the mean age, time and mortality. Table 1 presents the life expectancy of patients in the 15 selected papers [8][9][10][11][12][13][14][15][16][17][18][19][20][21][22] in which the authors have not lost any case to follow-up and knew exactly how many had died during the period. Figure 1 shows the studies through a curve of survival probability versus follow-up of the patients. The chance of pa-tient survival at 15 years is 57.6%, at 20 years is 34.6% and at 25 years is only 11.6%. The relationship between follow-up and survival was significantly linear (p <0.001); with each following year, the probability of survival decreased by 4.6%. When adjusted for age, the association remained significant (p <0.001).

DISCUSSION
Although there are numerous papers that have reported survival of THA, few include longer follow-up, without any loss of patients. The total hip arthroplasty decisively modifies the quality of life of patients. However, it is not a goal in the surgical indication to increase the survival of patients. Numerous authors, analysing statistics with long-lasting follow-up show that the survival curves of patients with THA are better than the corresponding analysed populations. [1][2][3][4][5][6][7] Lie et al. 1 observed that after a mean follow-up of 8 years, the mortality of patients with THA was 25% versus 30% in the corresponding Norwegian population. The authors propose several hypotheses to explain this fact and motivate a discussion. Visuri et al. 6 tried to explain this event by the selective process that these patients undergo after the indication for elective surgery. Barrett et al. 4 believe that patients return to having complete mastery over themselves and this determines a favourable change of behaviour. However, the calculation of survival in patients with THA is affected by a very large number of variants and is difficult to assess reliably. However, this expectation of survival can be considered a very important and almost unquestionable factor in planning the surgery. If the patient is aged 90 years or over, has rheumatoid arthritis, pulmonary metastatic cancer, Paget, osteoporosis, etc., that will influence the surgeon about the type of prosthesis, type of fixation, femoral head size, etc. The reliability of the calculation of survival of patients with THA may be questioned if we do not have follow-up of all patients undergoing surgery. This would be essential, but in practice the literature presents losses of 5-52%, depending on the years in which the patients are followed. [23][24][25][26] Toni et al. 26 reported that the loss in his series reached 51.5% after 11 years of follow-up. For this reason, large loss of follow-up can significantly compromise the survival curves of arthroplasty patients, the reviews and own results, as they become virtually unreliable. This does not occur in smaller segments because we have most of the events with patients environmentally con-  trolled. Berstock et al. 27 , in a systematic review of 30-days and 90-days mortality after total hip arthroplasty, concluded that the most commonly identified risk factors for early mortality are increasing age, male gender, and co-morbidity conditions, particularly cardiovascular disease. However, in the literature, most analyses of the results commonly assume that patients with subsequent losses have the same evolution as those who are regularly reviewed. [28][29][30] Few admit that those cases in which the segments have been lost are inferior to those that were regularly analysed. 23,25 In this study, the first selection criterion considered how long patients should be followed in order to offer a reliable statistical curve.
Of course the answer would be the longest possible follow-up. However, there are very few studies with 20, 25 or 30 years of follow-up. We have established that all studies with less than 10 years follow-up should not be included in this review. The second criterion evaluated was the loss to follow-up. First, we analysed papers with loss to follow-up of less than 10%. After reviewing 5394 articles on THA, we found only 44 articles with over 10 years of follow-up and less than 10% loss. However, if the following criteria demand that the authors say that there was no loss of follow-up, we found only 15 articles. After this selection, the index of mortality and survival rates of each paper were calculated according to the average follow-up time, and it was observed that the survival of the population after 25 years was 11.6%. Another interesting topic for discussion is the age of the patients. Nine from 15 of the assessed articles include patients operated at the age of less than 30 years, and this is a bias. The research used the mean age of the selected patients. In four papers, the mean age was about 58 years. 8,15,20,21 Guerritsma--Bleeker et al. 11 with 22 years of follow-up, and a mean age of 70.2 years found a mortality of 84%. Bojescul et al. 8 with 15.6 years of follow-up, and a mean age of 58 years, found a mortality of only 40%. Since, as the mean age of the group becomes higher, as higher we found the mortality ratio.

FINAL REMARKS
Considering the follow-up of more than 10 years, and the no lost of follow-up criteria, and the mean age of patients of more than 57.5 years, only 11.6% of patients undergoing THA will remain alive 25 years after the surgical procedure.