ORIGINAL ARTICLE

Meta-analysis of re-operation, nonunion, and infection after open

reduction and internal fixation of patella fractures

Christopher J. Dy, MD, MSPH, Milton T. M. Little, MD, Marschall B. Berkes, MD, Yan Ma, PhD,
Timothy R. Roberts, MLS, David L. Helfet, MD, and Dean G. Lorich, MD, New York, New York

BACKGROUND: The subcutaneous location of the patella and the demand for early knee motion contribute to the difficulty in treating patients
with patella fractures. The reported rates in the literature for hardware removal after patella open reduction and internal fixation
range from 0% to 60%. The wide variability of these reports leaves the true frequency of re-operation and complications after
patella open reduction and internal fixation in question. Furthermore, gaining a better understanding of the factors that contribute
to re-operation and complications will help to generate hypotheses and research agendas to address these difficult problems.
METHODS: We performed a systematic review to identify publications in which adult patients with patella fractures were surgically treated
with a minimum of 6-month follow-up. The surgical technique (tension band or other), infection rate, nonunion rate, and re-
operation rate (including removal of hardware) were recorded. Meta-regression analysis was used to describe the potential
contributory factors for re-operation, nonunion, and infection while controlling for age, gender, open fracture, surgical
technique, and date of publication. Separate regression models were constructed for each outcome depending on the number
of studies available for inclusion.
RESULTS: The frequency of re-operation was 33.6% in a meta-analysis of 24 studies (737 patella fractures). The frequency of infection
was 3.2% in a meta-analysis of 18 studies (522 patella fractures). The frequency of nonunion was 1.3% in a meta-analysis of
15 studies (464 patella fractures). There were no significant predictors for re-operation, nonunion, or infection in any of the
regression analyses.
CONCLUSION: Although the frequencies of nonunion and infection are relatively low after surgical treatment of patella fractures, the
modern rate of re-operation is substantial. (J Trauma Acute Care Surg. 2012;73: 928Y932. Copyright * 2012 by Lippincott
Williams & Wilkins)
LEVEL OF EVIDENCE: Meta-analysis, level III+.
KEY WORDS: Patella; complications; re-operation; meta-analysis; systematic review.

‘‘
T he incidence of patella fractures has been estimated at
1.2 to 6.1 per 100,000 person-years,1,2 with epidemio-
logic data from Sweden suggesting an increasing incidence
over a recent three decade period.’’3 The potential for symp-
tomatic hardware and knee stiffness after internal fixation of
patella fractures contribute to the difficulty in treating these
fractures. Although hardware failure and nonunion are both
thought to be relatively low following modern techniques of
open reduction and internal fixation (ORIF), symptomatic hard-
ware remains problematic and often requires a second proce-
dure for removal.4
Because of the difficulties associated with conducting
large multicenter studies of surgical patients, there is limited
evidence to guide clinicians about the frequency of re-operation
and complications after patella ORIF. The reported rates in the
literature for hardware removal after patella ORIF range from
0% to 60%.4 The wide variability of these reports leaves the
true frequency of re-operation and complications after patella
ORIF in question. Furthermore, gaining a better understanding
of the factors that contribute to re-operation and complications
will help in generating hypotheses and research agendas to
address this difficult problem.
In this study, we have conducted a systematic review of
the literature and performed a meta-analysis to answer the fol-
lowing research questions:
(1) What are the rates of re-operation, nonunion, and infection
after patella ORIF?
(2) What are the factors that contribute to re-operation, non-
union, and infection after patella ORIF?

Submitted: October 24, 2011, Revised: February 3, 2012, Accepted: February 21,
2012, Published online: August 17, 2012.
From the Departments of Orthopaedic Surgery (C.J.D., M.T.M.L., M.B.B.), Epi-
demiology and Biostatistics Core (Y.M.), Kim Barrett Memorial Library
(T.R.R.), Hospital for Special Surgery, New York, New York; and Orthopaedic
Trauma Service (D.L.H., D.G.L.), Hospital for Special Surgery, New York, New
York and Weill Cornell Medical College, New York, New York.
Address for reprints: Christopher J. Dy, MD, MSPH, Department of Orthopaedic
Surgery, Hospital for Special Surgery, 535 E 70th Street, New York, NY 10021;
email: dyc@hss.edu.
DOI: 10.1097/TA.0b013e31825168b6
928
METHODS
Identification and Eligibility of Relevant Studies
We conducted a systematic literature review of publica-
tions in English language. We did not impose a limit on pub-
lication year, so we defaulted to the dates of inclusion for each
database. A medical librarian conducted literature searches in
the PubMed, Cochrane Central Register of Controlled Trials,
and EMBASE databases on April 11, 2011, to identify all
studies which discussed complications after patella ORIF based
J Trauma Acute Care Surg
Volume 73, Number 4

Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Trauma Acute Care Surg
Volume 73, Number 4
on the following criteria: (orif OR ‘‘Open Reduction Internal
Fixation’’ OR ‘‘Open Reduction’’ OR ‘‘Internal Fixation’’ OR
‘‘fracture fixation, internal’’ EMeSH Terms^ OR Orthopedic
Fixation DevicesEMeSH Terms^ OR Fractures, Bone/surgery
EMeSH Terms^) AND (patella/injuries OR E‘‘fracture’’ OR
‘‘fractures’’^ AND Patella) AND (Postoperative Complications
OR Treatment Outcome OR Adverse Effects OR reoperation OR
second-look surgery OR Equipment Failure OR ‘‘Hardware
Failure’’ OR posttraumatic arthritis OR posttraumatic arthritis
OR Device Removal OR Follow-Up Studies OR Retrospective
Studies OR retrospective* OR Complications*). We then lim-
ited the results to only studies that included human subjects.
The initial search yielded 454 results for initial screening.
We reviewed the titles of each study as a first pass to exclude
studies that were (1) cadaveric or biomechanical investigations
only, (2) were devoted to treatment of periprosthetic fractures,
osteochondral fractures, sleeve fractures, or patellar malunions
or nonunions, or (3) were case reports, technique descriptions,
or review articles without associated case series. The abstracts
of each of the remaining 145 articles were reviewed to exclude
studies that explicitly stated treatment of patients younger than
16 years, had follow-up outcome less than 1 year, did not in-
clude pertinent outcome data on re-operation, nonunion, or
infection. The abstract review also revealed additional articles
that were either case reports or technique descriptions only
(without associated case series), which also were excluded.
Review of the full text of the remaining 82 articles was per-
formed. An additional 57 studies were excluded because of
inadequate outcomes data, and one study was excluded be-
cause 68% of the patients in the series were treated non-
operatively,5 leaving a total of 24 studies ranging from years
1978 to 2011 that underwent data extraction (Fig. 1).
Data Extraction
A detailed electronic spreadsheet was designed in Micro-
soft Excel (Redmond, WA) to record data for analysis. When the
information was available, three members of the investigative
team extracted the year of publication, number of patients,
Figure 1. Flow diagram depicting study selection for
inclusion in meta-analysis.
* 2012 Lippincott Williams & Wilkins
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Dy et al.
number of patella fractures, mean duration of follow-up, mean
age, closed or open fracture, surgical technique used (tension
band technique or other technique), number of re-operations
(for any reason), number of infections, and number of non-
unions from each article. Missing data were noted in the data
collection sheet, but studies were included if they had con-
tained information on at least one of the three outcomes of in-
terest (re-operation, infection, or nonunion).
Data Analysis
An individual dataset was created for each of the out-
comes of interest (re-operation, infection, or nonunion). This
was done to allow inclusion of the maximum number of studies
for each outcome. Multiple meta-analyses were performed to
pooled proportions of re-operation (defined as number of re-
operations/total number of patella fractures), infections, and
nonunions. In each case, to determine the pooled proportion,
the variances of the raw proportions were stabilized by using a
Freeman-Tukey-type arcsine square root transformation.6 The
pooled proportions were calculated as the back-transform of
the weighted mean of the transformed proportions, using fixed
or random effects models.7 We tested the significance of het-
erogeneity between studies using the Q test.8 Random effects
models were chosen if the Q test was significant. Otherwise,
fixed effects models were applied. Forest plots were used for
presentation of the outcome proportions. Confidence intervals
from individual studies, pooled proportions, and test for ho-
mogeneity were also included.
Meta-regression based on random effects logistic model
was conducted for proportion of re-operation, infection, and
nonunion, respectively, to identify the effects of mean follow-
up, age, closed or open fracture, surgical technique (tension
banding or other technique), and year of publication (during/
before or after year 2000). The year 2000 was chosen because
it allowed a relatively equal division of studies in each of the
datasets. The regression model produced odds ratios with
corresponding 95% confidence intervals for each of the covar-
iates. The threshold for statistical significance of the covariates
was p G 0.05.
The current investigation is reported using the Meta-
analysis of Observational Studies in Epidemiology guidelines
for meta-analyses of observational studies.9
RESULTS
There were 24 studies (with a total of 737 patella frac-
tures) that included information about re-operation. The level
of evidence was Level IV in 20 studies, Level III in 3 studies,
and Level II in 1 study. The frequency of re-operation was
33.6% of 737 patella fractures among 24 studies (Fig. 2). The
infection rate was 3.2% of 522 patella fractures among 18
studies (Fig. 3). The frequency of nonunion was 1.3% of 464
patella fractures among 15 studies (Fig. 4).
None of the examined variables (age, gender, open or
closed fracture, operative technique, or date of publication)
significantly influenced the frequency of re-operation, infec-
tion, or nonunion. Of note, the open or closed fracture variable
was not included in the regression model for nonunion because
929

Dy et al.
Figure 2. Forest plot showing the rate of re-operation.
CI, confidence interval.
Figure 3. Forest plot showing the rate of infection.
CI, confidence interval.
the amount of missing data for this variable did not allow
appropriate convergence of the model.
DISCUSSION
We conducted this study to provide a better understanding
of the frequency of re-operation, infection, and nonunion after
surgical treatment of patella fractures and the potential surgical
930
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Trauma Acute Care Surg
Volume 73, Number 4
Figure 4. Forest plot showing the rate of nonunion.
CI, confidence interval.
contributors to these outcomes. In our analysis, we have esti-
mated the rate of re-operation, infection, and nonunion after
surgical treatment of patella fractures to be 33.6%, 3.2%, and
1.9%, respectively. These rates fall within the ranges described
by a previous nonsystematic review10 and large series,11Y16 but
the meta-analysis methods allow a more accurate estimate of
the true frequency of these events. In addition, our regression
analysis has found that age, gender, operative technique, or
date of publication did not significantly influence the rate of
re-operation, infection, or nonunion. To our knowledge, no
prior studies have attempted to quantify and characterize these
outcomes in a large comparative meta-analysis.
This study has several limitations inherent to any meta-
analysis. First, patient-centered outcomes, such as range of
motion, satisfaction, and return to function, are not included in
the meta-analysis as they are not always reported or are
reported with a nonuniform scale in the original articles. We
originally intended to use clinician-derived measures of out-
come, but there were no uniformly reported outcomes metrics
found in the reviewed studies. We also intended to record
reasons for removal of hardware, but these reasons were not
consistently reported in a manner that would allow inclusion in
the meta-analysis. Additionally, we did not include fracture
classification as a predictor of outcomes because of the het-
erogeneity of classification schemes used in the articles. Fur-
thermore, none of the studies included in all datasets were
nonrandomized studies. Although a meta-analysis of ran-
domized controlled trials is certainly preferred, there is a role
for statistical syntheses of data from observational studies when
high-quality evidence is not available.10,17 This is particularly
relevant in view of the absence of any population-based studies
that could also reflect the frequency of re-operation, infection,
and nonunion after patella fracture treatment. Last, there was
high heterogeneity of the studies included in the meta-regression
analysis for re-operation (I2 = 95.5%; Fig. 2), likely caused by
high variations in clinical sampling (of both fracture pat-
terns and threshold to re-operate) over a broad timeframe. We
* 2012 Lippincott Williams & Wilkins
J Trauma Acute Care Surg
Volume 73, Number 4 Dy et al.

TABLE 1. Studies Included in Meta-Analysis With Associated Demographics and Outcomes

No. of Average Average Men Open Tension No. of No. of No. of
Reference Fractures Age (yr) Follow-Up (mo) (%) fx (%) Band (%) Re-Operation Nonunion Infection
Peeples and Margo18 20 55.8 56.4 28.6 NR 0.0 0 NR NR
Bostman et al.11 64 42 21 NR 10.9 32.8 1 NR NR
Bostman et al.12 93 38 12 73.1 9.7 15.1 2 NR NR
Hung et al.13 68 NR NR 89.7 NR 100.0 0 0 7
Saltzman et al.16 40 11 100.8 38.0 27.5 0.0 1 0 1
Rink and Scott 19 17 54.6 48 47.1 0.0 52.9 14 NR 1
Juutilainen et al.20 9 NR 36 22.2 100.0 0.0 5 1 NR
Catalano et al.21 60 34 21 57.9 100.0 58.3 39 1 0
Torchia and Lewallen15 47 28.3 112.8 51.9 82.5 37.3 15 2 6
Berg22 7 28 31 66.7 28.6 71.4 5 0 0
Wu et al.14 62 39 24 79.4 2.9 100.0 2 0 0
Yang and Byan23 25 37 22 NR NR 20.0 1 0 0
Kastelec and Veselko24 11 55.1 54 NR NR 0.0 8 NR NR
Veselko and Kastelec25 13 50.5 56.4 NR NR 0.0 0 NR 1
Bayar et al26 20 37.75 NR 87.5 10.0 100.0 11 NR 0
Yanmis et al.27 5 32.5 22 NR 0.0 0.0 5 0 0
Singh et al.28 14 35 36 NR NR 0.0 14 NR 0
Uvaraj et al.29 21 43.4 66 81.8 0.0 72.7 2 NR 2
El-Sayed and Ragab30 14 34 26 100.0 0.0 0.0 0 0 0
Li et al.31 60 NR NR NR 0.0 71.7 17 0 0
Haklar et al.32 18 42.1 12.89 75.0 NR 100.0 18 0 NR
Yotsumoto et al.33 13 58.2 18 38.5 NR 100.0 13 0 0
Yang et al.34 21 40 24 57.1 14.3 0.0 9 0 0
Qi et al.35 15 46.2 14 60.0 NR 0.0 0 0 0
NR, not reported.

accounted for this heterogeneity by a random effects regression AUTHORSHIP

model. Although we recognize these limitations, we think that
this study provides a quantitative description on the current
status of re-operation, infection, and nonunion after surgical
treatment of patella fractures. These quantitative outcomes can
be used by orthopedic surgeons in patient discussions of the
likelihood of re-operation, infection, and nonunion (Table 1).
Although our data show that patella fractures heal with
low nonunion rates, approximately one-third of the patients
undergo re-operation. These findings suggest that further in-
vestigation is needed to use treatment strategies that maintain
these high union rates but decrease the frequency of re-operation.
Patella fractures may be prone to re-operation because the
hardware is often subcutaneous and early motion is required to
maintain function after surgery, which may predispose to hard-
ware loosening. The morbidity, risk of re-fracture, and costs
associated with subsequent surgery cannot be underestimated,
particularly as aging patients represent an increasing portion of
patients with patella fractures.10
Surgical treatment of patella fractures remains challeng-
ing. The results of our descriptive meta-analysis confirm that
re-operation is common, whereas infection and nonunion are
relatively rare. The results of this meta-analysis may stimulate
ideas for patient and population-based research that further
examines the factors that contribute to complications after sur-
gical treatment of patella fractures and an ultimate improvement
in outcomes.
C.J.D., M.T.M.L., M.B.B., D.L.H., and D.G.L. participated in the devel-
opment of the study idea. C.J.D., M.T.M.L., and M.B.B. conducted the
literature search. C.J.D., Y.M., T.R.R., D.L.H., and D.G.L. designed the
study. C.J.D., M.T.M.L., M.B.B., and T.R.R. collected the data, which
Y.M. and T.R.R. analyzed. C.J.D., M.T.M.L., M.B.B., Y.M., D.L.H., and
D.G.L. interpreted the data. C.J.D. drafted the initial manuscript, which
C.J.D., M.T.M.L., Y.M., AND T.R.R. edited. C.J.D., D.L.H., and D.G.L. re-
viewed the initial and final versions of the manuscript.
DISCLOSURE
The authors declare no conflicts of interest.
REFERENCES
1. van Staa TP, Dennison EM, Leufkens HG, Cooper C. Epidemiology of
fractures in England and Wales. Bone. 2001;29:517Y522.
2. Yang NP, Chan CL, Yu IL, Lee CY, Chou P. Estimated prevalence of
orthopaedic fractures in TaiwanVa cross-sectional study based on
nationwide insurance data. Injury. 2010;41:1266Y1272.
3. Bengner U, Johnell O, Redlund-Johnell I. Increasing incidence of tibia
condyle and patella fractures. Acta Orthop Scand. 1986;57:334Y336.
4. Melvin JS, Mehta S. Patellar fractures in adults. J Am Acad Orthop Surg.
2011;19:198Y207.
5. Bostrom A. Fracture of the patella. A study of 422 patellar fractures. Acta
Orthop Scand Suppl. 1972;143:1Y80.
6. Freeman MF, Tukey JW. Transformations related to the angular and the
square root. Ann Math Stat. 1950;21:607Y611.
7. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin
Trials. 1986;7:177Y188.

* 2012 Lippincott Williams & Wilkins 931

Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Dy et al.
8. Cochrane WG. The combination of estimates from different experiments.
Biometrics. 1954;10:101Y129.
9. Stroup DF, Berlin JA, Moron SC, et al. Meta-analysis of observational
studies in epidemiology: a proposal for reporting. JAMA. 2000;283:
2008Y2012.
10. Warriner AH, Patkar NM, Curtis JR, et al. Which fractures are most
attributable to osteoporosis? J Clin Epidemiol. 2011;64:46Y53.
11. Bostman O, Kiviluoto O, Nirhamo J. Comminuted displaced fractures of
the patella. Injury. 1981;13:196Y202.
12. Bostman O, Kiviluoto O, Santavirta S, Nirhamo J, Wilppula E. Fractures
of the patella treated by operation. Arch Orthop Trauma Surg. 1983;102:
78Y81.
13. Hung LK, Chan KM, Chow YN, Leung PC. Fractured patella: operative
treatment using the tension band principle. Injury. 1985;16:343Y347.
14. Wu CC, Tai CL, Chen WJ. Patellar tension band wiring: a revised technique.
Arch Orthop Trauma Surg. 2001;121:12Y16.
15. Torchia ME, Lewallen DG. Open fractures of the patella. J Orthop Trauma.
1996;10:403Y409.
16. Saltzman CL, Goulet JA, McClellan RT, Schneider LA, Matthews LS.
Results of treatment of displaced patellar fractures by partial patellectomy.
J Bone Joint Surg Am. 1990;72:1279Y1285.
17. Chesney A, Chauhan A, Kattan A, Farrokhyar F, Thoma A. Systematic
review of flexor tendon rehabilitation protocols in zone II of the hand.
Plast Reconstr Surg. 2011;127:1583Y1592.
18. Peeples RE, Margo MK. Function after patellectomy. Clin Orthop Relat
Res. 1978;180Y186.
19. Rink PC, Scott F. The operative repair of displaced patellar fractures.
Orthop Rev. 1991;20:157Y165.
20. Juutilainen T, Patiala H, Rokkanen P, Tormala P. Biodegradable wire
fixation in olecranon and patella fractures combined with biodegradable
screws or plugs and compared with metallic fixation. Arch Orthop
Trauma Surg. 1995;114:319Y323.
21. Catalano JB, Iannacone WM, Marczyk S, et al. Open fractures of the
patella: long-term functional outcome. J Trauma. 1995;39:439Y444.
22. Berg EE. Extensile exposure of comminuted patella fractures using a tibial
tubercle osteotomy: results of a new technique. J Orthop Trauma. 1998;
12:351Y355.
932
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Trauma Acute Care Surg
Volume 73, Number 4
23. Yang KH, Byun YS. Separate vertical wiring for the fixation of comminuted
fractures of the inferior pole of the patella. J Bone Joint Surg Br. 2003;85:
1155Y1160.
24. Kastelec M, Veselko M. Inferior patellar pole avulsion fractures:
osteosynthesis compared with pole resection. J Bone Joint Surg Am. 2004;
86-A:696Y701.
25. Veselko M, Kastelec M. Inferior patellar pole avulsion fractures:
Osteosynthesis compared with pole resection. Surgical technique. J Bone
Joint Surg Am. 2005;87(suppl 1):113Y121.
26. Bayar A, Sener E, Keser S, Meray J, Simsek A, Senkoylu A. What leads to
unfavourable cybex test results for quadriceps power after modified tension
band osteosynthesis of patellar fractures? Injury. 2006;37:520Y524.
27. Yanmis I, Oguz E, Atesalp AS, et al. Application of circular external
fixator under arthroscopic control in comminuted patella fractures:
technique and early results. J Trauma. 2006;60:659Y663.
28. Singh RP, Shah RK, Srivastava MP. Treatment of inferior patellar pole
avulsion fractures with pole resection and patellotibial cerclage wire.
Nepal Med Coll J. 2007;9:93Y95.
29. Uvaraj NR, Mayil Vahanan N, Sivaseelam A, Mohd Sameer M, Basha IM.
Surgical management of neglected fractures of the patella. Injury. 2007;38:
979Y983.
30. El-Sayed AM, Ragab RK. Arthroscopic-assisted reduction and stabilization
of transverse fractures of the patella. Knee. 2009;16:54Y57.
31. Li Y, Ouyang YP, Gou SH, Yuan Q, Liu Y, Sun W. Comparison of curative
effects among three internal fixation implants for comminuted patellar
fracture. J Clin Rehab Tis Eng. 2008;760.
32. Haklar U, Kocaoglu B, Gereli A, Nalbantoglu U, Guven O. Arthroscopic
inspection after the surgical treatment of patella fractures. Int Orthop.
2009;33:665Y670.
33. Yotsumoto T, Nishikawa U, Ryoke K, Nozaki K, Uchio Y. Tension band
fixation for treatment of patellar fracture: Novel technique using a
braided polyblend sutures and ring pins. Injury. 2009;40:713Y717.
34. Yang L, Yueping O, Wen Y. Management of displaced comminuted patellar
fracture with titanium cable cerclage. Knee. 2010;17:283Y286.
35. Qi L, Chang C, Xin T, et al. Double fixation of displaced patella fractures
using bioabsorbable cannulated lag screws and braided polyester suture
tension bands. Injury. 2011;42:1116Y1120.
* 2012 Lippincott Williams & Wilkins