- Email: [email protected]
Preoperative traction in patients with hip fractures
242
Injury (1992) 23, (4), 242-244
Printed in Great Britain
Preoperative traction in patients with hip fractures V. Finsen, M. Borset, G. E. Buvik and I. Hauke Department of Surgery, Orkda! Sanitetsforenings Hospital, Orkanger, Norway
A series of 80 patients with cervical, trochanteric or subtrochanteric hip fractures were randomized to either treatment without ~raction. skin traction, or skeletal traction during the 1883 h between admission and operation. The institution of skin or skeletal traction was not particularly painful for the patient, but we found no indication that either was of discernible benefit. The number of analgesic medications needed was no higher in patients without traction. We conclude that traction should not be administered routinely to patients awaiting operation for hip fracture.
Introduction It has been customary m many countries to place patients awaiting operation for femoral neck fractures in skeletal or skin traction. This may reduce pain through immobilization. On the other hand, as the extremity is tethered, there might also be additional pain, for instance when having to use a bedpan. Furthermore, traction might partly reduce the fracture, and thus make the operation easier to perform. It seems hard to imagine, however, that this could be accomplished with the 3 kg or so of weights which it is possible to employ when using skin traction. Neither of these putative effects can be obtained by just placing the leg in a comfortable position in the bed, but in this case the patient avoids the pain presumably connected with the additional movement of the extremity when arranging for traction and transfixing the tibia with a Steinman pin. We have found no guidance in the literature concerning these matters and therefore conducted the following randomized investigation.
secured with elastic bandages. The extremity was placed on a Braun splint in the bed and further stabilized with small pillows and more elastic bandages. Traction was applied with a 3 kg weight on a wire which was passed over a pulley at the end of the bed and fastened to the plaster loop under the patient's foot. Patients in the third group were treated as those in the skin traction group, but traction was applied to a Steinman pin which was passed through the proximal tibia under local anaesthesia. Weights equalling approximately 10 per cent of the patient's body weight were used in this group. The 21 patients who on clinical and radiographical grounds initially were considered possibly to have impacted fractures were excluded, as were 17 patients who underwent operation within 6 h of admission, Thus, 80 patients were left for evaluation (Table I). Five patients were not treated according to the protocol. One patient in the no-traction group had to wait 90 h before operation and was after a time placed in skin traction because of pain by a surgeon who was not involved in the study. NO traction was applied to one patient in the skin traction group because the staff did not have time. Three patients who had been randomized to the skeletal traction group were treated with skin traction because of respectively: excessive oedema of the leg, osteosynthesis material in the proximal tibia, and an infected wound below the knee. All protocol deviants were retained in their intended groups for evaluation of the results (Laupacis et al., 1989). Analgesic medication was standardized and recorded. Most patients were treated with hip compression screws. Table I. Characteristics of randomized patients
Patients and methods We studied all 118 patients over the age of 50 years who were admitted to our hospital during a I-year period with recent cervical, trochanteric or subtrochanteric hip fractures. On admission they were allocated with the aid of random numbers to one of three preoperative treatment groups. Patients in the first group had the injured leg placed on a pillow in the bed. Those in the second group had a length of plaster fastened clown each side of the injured extremity and © 1992 Butterworth-Heinemann Ltd 0020-1383/92/040242-03
Traction
Number of patients Mean age (years) Women/men Cerv./troch./su b, Number receiving analgesics Before admission On admission
None
Skin
Skeletal
25 77.2 18/7 15/10/0
26 80.5 19/7 15/8/3
29 78.7 22/7 16/10/3
8 7
8 7
14 6
Finsen et al.: Preoperative traction in hip fractures
243
In seven cervical fractures, an uncemented endoprosthesis was used because this was the method of choice of one surgeon and, in one further case in the no-traction group, a cemented endoprosthesis was resorted to because the fracture could not be reduced. These patients and those with subtrochanteric fractures were excluded in the evaluation of the intraoperative parameters. Two patients died preoperatively, one on the day of admission and one 2 days later. Both had been randomized to the. no-traction group. Statistical evaluation of the results w a s m a d e with non-parametric (Wilcoxon two-sample) tests. P-values < 0.05 were considered significant. The risk of not having detected a clinically significant difference (type 2 error) was tested by the displacement method (Lehmann, 1975; Amfred, 1983; Keiding et al., 1984). Results
The proportion of patients who had received analgesic medication before admission was somewhat higher in the skeletal traction group than in the other two groups (Table I). In other respects the three groups of patients were similar. The mean time between admission and operation was similar in the three groups. The patients reported that the institution of skin or skeletal traction was not particularly painful. The mean number of pain medications was slightly higher in the skeletal traction group (Table !I). The risk of not having detected a real reduction of 33 per cent in analgesic medications in the skin traction group compared with the no-traction group was less than 0.07, while it was less than 0.025 for the skeletal traction group. The risk of not having detected a reduction in analgesic medications of 50 per cent in the skin traction group, and also in the skeletal traction group, when compared with the no-traction group, was less than 0.003.
Table II. Mean time between admission and operation (ADM-OP in hours and mean number of analgesic medications. Ranges in brackets Traction
Adm-op (h) Pain medications Pain medications per 24 h
None
Skin
Skeletal
26 (10-90) 2.1 (0-9) 2.1
24 (10-52) 2.1 (0-9) 2.2
23 (8-68) 2,3 (0-9) 2.4
Table III shows the total analgesic medication administered between admission and operation. In addition, five tablets of 500 mg paracetamo! and one of 0.2 mg buprenorphine were administeredl If, for the sake of illustration, these few protocol-deviant medications are ignored and l i n g of morphine is considered equipotent to 10 mg of pethidine, patients without traction received on average 56 mg of 'pethidine' and 1.6 units of paracetamol/codeine for each 24 h of this period, The corresponding figures were 50 mg and 1.7 units for patients in skin traction and 41 mg and 2.6 units for those in skeletal traction. Preoperative fracture reduction was significantly quicker (TablelV) in the no-traction group (P< 0.001 for the group as a whole). The median duration of the operations was similar in the groups, while intraoperative bleeding was significantly greater (P< 0.01) for trochanteric fractures in the skeletal traction group than in either of the other two groups. Discussion
It is unsatisfying to be forced to measure pain medication merely by the number of medications without taking into account the substance and the quantity administered, but there is no valid way in which centrally and peripherally acting analgesics can be compared. In the present study it nevertheless seems permissible, for the sake of illustration, to omit the few medications which were not according to the protocol and compare the consumption of opiates and paracetamol/codeine. It is not possible to say, however, whether or not the higher medication of paracetamol/ codeine in the skeletal traction group is partly or completely outweighed by the somewhat lower opiate medication. Thus, there was no indication that skeletal traction conferred any benefit compared with no traction. On the contrary, there was a trend for those in skeletal traction to do less well than those in the other groups. They received more pain medications, which would seem to indicate that they complained of pain more often. Furthermore, while the dUration of operations was much the same in each group, the intraoperative bleeding was somewhat higher in the skeletal traction group. Also, skin traction was without demonstrable benefit compared with simply placing the injured leg on a pillow in the bed.
Table IV. Median recorded time for preoperative reduction of fractures (rain), duration of operations (rain), and intraoperative bleeding tml) - excluding subtrochanteric fractures and endoprosthesis operations Traction
Table IlL Analgesic medication administered to each group None
Traction
Adm-op (total hours) Paracet/codeine (units) Pethidine (mg) Morphine (mg)
None
Skin
Skeletal
597 39 1000 40
611 44 1200 7.5
675 74 1000 15
ADM-OP =Total time for whole group between admission and operations. Paracet/codeine (units) =Tablets of suppositories containing 400 mg paracetamol and 30 mg codeine phosphate.
Reduction t~me Cerv, fracture Troch. fracture Operation time Cerv. fracture Troch. fracture Intraoperative bleeding Cerv. fracture Troch. fracture
2.5* 2.5
Skin
Skeletal
5 5
5 4
50 65
45 62
50 65
200 350
200 300
250 500*
(*Significantly different from two other groups: P< 0,01 .)
244
Injury: the British Journal of Accident Surgery (1992) Vol. 23/No. 4
The small but statistically significant differences between patients treated with and without traction with respect to intraoperative blood loss and time needed to reduce the fracture may well be spurious findings, and are in any case not of a magnitude to be of any clinical importance. Contrary to expectations, patients did not rate the actual procedure of instituting skin or skeletal traction as particularly painful. The procedures did, however, have the disadvantage of needing the assistance of a nurse in the case of skin traction, and operating theatre facilities in the case of skeletal traction. In some instances an operating theatre nurse had to be called from her home at night to help with the skeletal traction. We were unable to identify a single parameter which indicated even a trend towards a benefit from traction. Although our study only includes 80 patients, the risk of not having detected a clinically important reduction in pain medications due to traction is small. Thus, the risk of having overlooked a reduction in the median number of pain medications from around two to one per 24 h is < 0.3 per cent. We therefore feel it is fair to conclude that neither skin nor skeletal traction preoperatively offer any discernible advantages for patients with hip fractures.
Acknowledgement We are grateful to Mr Terkel Amfred for statistical advice.
References Arnfred T. (I983) A method for calculatingthe risk of type 2 error using non-parametric statistical tests. Ugeskr. L~eger. I45, 2358. Keiding N., Hilden J., Amfred T. et al. (1984) Non-parametric confidence-intervals. Ugeskr. Laeger: 146, 275. LaupacisA., Rorabeck C. H., Bourne R, B. et al. (1989) Randomized trials in orthopaedics: Why, how, and when? J. Bone Joint Surg. 71A, 535. Lehmann E. L. (1975) Nonparametrics. Sta~i#ical Mdhods Based on Ranks. San Francisco:Holden-Day. Paper accepted 9 September 1991.
Requests for reprints should be addressed to: V. Finsen MD, Department of Orthopaedic Surgery, Trondheim University Hospital, N-7006 Trondheim, Norway.
Injury (1992) 23, (4), 242-244
Printed in Great Britain
Preoperative traction in patients with hip fractures V. Finsen, M. Borset, G. E. Buvik and I. Hauke Department of Surgery, Orkda! Sanitetsforenings Hospital, Orkanger, Norway
A series of 80 patients with cervical, trochanteric or subtrochanteric hip fractures were randomized to either treatment without ~raction. skin traction, or skeletal traction during the 1883 h between admission and operation. The institution of skin or skeletal traction was not particularly painful for the patient, but we found no indication that either was of discernible benefit. The number of analgesic medications needed was no higher in patients without traction. We conclude that traction should not be administered routinely to patients awaiting operation for hip fracture.
Introduction It has been customary m many countries to place patients awaiting operation for femoral neck fractures in skeletal or skin traction. This may reduce pain through immobilization. On the other hand, as the extremity is tethered, there might also be additional pain, for instance when having to use a bedpan. Furthermore, traction might partly reduce the fracture, and thus make the operation easier to perform. It seems hard to imagine, however, that this could be accomplished with the 3 kg or so of weights which it is possible to employ when using skin traction. Neither of these putative effects can be obtained by just placing the leg in a comfortable position in the bed, but in this case the patient avoids the pain presumably connected with the additional movement of the extremity when arranging for traction and transfixing the tibia with a Steinman pin. We have found no guidance in the literature concerning these matters and therefore conducted the following randomized investigation.
secured with elastic bandages. The extremity was placed on a Braun splint in the bed and further stabilized with small pillows and more elastic bandages. Traction was applied with a 3 kg weight on a wire which was passed over a pulley at the end of the bed and fastened to the plaster loop under the patient's foot. Patients in the third group were treated as those in the skin traction group, but traction was applied to a Steinman pin which was passed through the proximal tibia under local anaesthesia. Weights equalling approximately 10 per cent of the patient's body weight were used in this group. The 21 patients who on clinical and radiographical grounds initially were considered possibly to have impacted fractures were excluded, as were 17 patients who underwent operation within 6 h of admission, Thus, 80 patients were left for evaluation (Table I). Five patients were not treated according to the protocol. One patient in the no-traction group had to wait 90 h before operation and was after a time placed in skin traction because of pain by a surgeon who was not involved in the study. NO traction was applied to one patient in the skin traction group because the staff did not have time. Three patients who had been randomized to the skeletal traction group were treated with skin traction because of respectively: excessive oedema of the leg, osteosynthesis material in the proximal tibia, and an infected wound below the knee. All protocol deviants were retained in their intended groups for evaluation of the results (Laupacis et al., 1989). Analgesic medication was standardized and recorded. Most patients were treated with hip compression screws. Table I. Characteristics of randomized patients
Patients and methods We studied all 118 patients over the age of 50 years who were admitted to our hospital during a I-year period with recent cervical, trochanteric or subtrochanteric hip fractures. On admission they were allocated with the aid of random numbers to one of three preoperative treatment groups. Patients in the first group had the injured leg placed on a pillow in the bed. Those in the second group had a length of plaster fastened clown each side of the injured extremity and © 1992 Butterworth-Heinemann Ltd 0020-1383/92/040242-03
Traction
Number of patients Mean age (years) Women/men Cerv./troch./su b, Number receiving analgesics Before admission On admission
None
Skin
Skeletal
25 77.2 18/7 15/10/0
26 80.5 19/7 15/8/3
29 78.7 22/7 16/10/3
8 7
8 7
14 6
Finsen et al.: Preoperative traction in hip fractures
243
In seven cervical fractures, an uncemented endoprosthesis was used because this was the method of choice of one surgeon and, in one further case in the no-traction group, a cemented endoprosthesis was resorted to because the fracture could not be reduced. These patients and those with subtrochanteric fractures were excluded in the evaluation of the intraoperative parameters. Two patients died preoperatively, one on the day of admission and one 2 days later. Both had been randomized to the. no-traction group. Statistical evaluation of the results w a s m a d e with non-parametric (Wilcoxon two-sample) tests. P-values < 0.05 were considered significant. The risk of not having detected a clinically significant difference (type 2 error) was tested by the displacement method (Lehmann, 1975; Amfred, 1983; Keiding et al., 1984). Results
The proportion of patients who had received analgesic medication before admission was somewhat higher in the skeletal traction group than in the other two groups (Table I). In other respects the three groups of patients were similar. The mean time between admission and operation was similar in the three groups. The patients reported that the institution of skin or skeletal traction was not particularly painful. The mean number of pain medications was slightly higher in the skeletal traction group (Table !I). The risk of not having detected a real reduction of 33 per cent in analgesic medications in the skin traction group compared with the no-traction group was less than 0.07, while it was less than 0.025 for the skeletal traction group. The risk of not having detected a reduction in analgesic medications of 50 per cent in the skin traction group, and also in the skeletal traction group, when compared with the no-traction group, was less than 0.003.
Table II. Mean time between admission and operation (ADM-OP in hours and mean number of analgesic medications. Ranges in brackets Traction
Adm-op (h) Pain medications Pain medications per 24 h
None
Skin
Skeletal
26 (10-90) 2.1 (0-9) 2.1
24 (10-52) 2.1 (0-9) 2.2
23 (8-68) 2,3 (0-9) 2.4
Table III shows the total analgesic medication administered between admission and operation. In addition, five tablets of 500 mg paracetamo! and one of 0.2 mg buprenorphine were administeredl If, for the sake of illustration, these few protocol-deviant medications are ignored and l i n g of morphine is considered equipotent to 10 mg of pethidine, patients without traction received on average 56 mg of 'pethidine' and 1.6 units of paracetamol/codeine for each 24 h of this period, The corresponding figures were 50 mg and 1.7 units for patients in skin traction and 41 mg and 2.6 units for those in skeletal traction. Preoperative fracture reduction was significantly quicker (TablelV) in the no-traction group (P< 0.001 for the group as a whole). The median duration of the operations was similar in the groups, while intraoperative bleeding was significantly greater (P< 0.01) for trochanteric fractures in the skeletal traction group than in either of the other two groups. Discussion
It is unsatisfying to be forced to measure pain medication merely by the number of medications without taking into account the substance and the quantity administered, but there is no valid way in which centrally and peripherally acting analgesics can be compared. In the present study it nevertheless seems permissible, for the sake of illustration, to omit the few medications which were not according to the protocol and compare the consumption of opiates and paracetamol/codeine. It is not possible to say, however, whether or not the higher medication of paracetamol/ codeine in the skeletal traction group is partly or completely outweighed by the somewhat lower opiate medication. Thus, there was no indication that skeletal traction conferred any benefit compared with no traction. On the contrary, there was a trend for those in skeletal traction to do less well than those in the other groups. They received more pain medications, which would seem to indicate that they complained of pain more often. Furthermore, while the dUration of operations was much the same in each group, the intraoperative bleeding was somewhat higher in the skeletal traction group. Also, skin traction was without demonstrable benefit compared with simply placing the injured leg on a pillow in the bed.
Table IV. Median recorded time for preoperative reduction of fractures (rain), duration of operations (rain), and intraoperative bleeding tml) - excluding subtrochanteric fractures and endoprosthesis operations Traction
Table IlL Analgesic medication administered to each group None
Traction
Adm-op (total hours) Paracet/codeine (units) Pethidine (mg) Morphine (mg)
None
Skin
Skeletal
597 39 1000 40
611 44 1200 7.5
675 74 1000 15
ADM-OP =Total time for whole group between admission and operations. Paracet/codeine (units) =Tablets of suppositories containing 400 mg paracetamol and 30 mg codeine phosphate.
Reduction t~me Cerv, fracture Troch. fracture Operation time Cerv. fracture Troch. fracture Intraoperative bleeding Cerv. fracture Troch. fracture
2.5* 2.5
Skin
Skeletal
5 5
5 4
50 65
45 62
50 65
200 350
200 300
250 500*
(*Significantly different from two other groups: P< 0,01 .)
244
Injury: the British Journal of Accident Surgery (1992) Vol. 23/No. 4
The small but statistically significant differences between patients treated with and without traction with respect to intraoperative blood loss and time needed to reduce the fracture may well be spurious findings, and are in any case not of a magnitude to be of any clinical importance. Contrary to expectations, patients did not rate the actual procedure of instituting skin or skeletal traction as particularly painful. The procedures did, however, have the disadvantage of needing the assistance of a nurse in the case of skin traction, and operating theatre facilities in the case of skeletal traction. In some instances an operating theatre nurse had to be called from her home at night to help with the skeletal traction. We were unable to identify a single parameter which indicated even a trend towards a benefit from traction. Although our study only includes 80 patients, the risk of not having detected a clinically important reduction in pain medications due to traction is small. Thus, the risk of having overlooked a reduction in the median number of pain medications from around two to one per 24 h is < 0.3 per cent. We therefore feel it is fair to conclude that neither skin nor skeletal traction preoperatively offer any discernible advantages for patients with hip fractures.
Acknowledgement We are grateful to Mr Terkel Amfred for statistical advice.
References Arnfred T. (I983) A method for calculatingthe risk of type 2 error using non-parametric statistical tests. Ugeskr. L~eger. I45, 2358. Keiding N., Hilden J., Amfred T. et al. (1984) Non-parametric confidence-intervals. Ugeskr. Laeger: 146, 275. LaupacisA., Rorabeck C. H., Bourne R, B. et al. (1989) Randomized trials in orthopaedics: Why, how, and when? J. Bone Joint Surg. 71A, 535. Lehmann E. L. (1975) Nonparametrics. Sta~i#ical Mdhods Based on Ranks. San Francisco:Holden-Day. Paper accepted 9 September 1991.
Requests for reprints should be addressed to: V. Finsen MD, Department of Orthopaedic Surgery, Trondheim University Hospital, N-7006 Trondheim, Norway.