SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Objective

To evaluate the association between radiographic hand osteoarthritis (OA), a disease with marked heredity, and radiographic knee OA in patients treated with meniscectomy.

Methods

We retrospectively identified 170 patients (mean age 54 years [range 33–87 years], 23% women) who had undergone isolated meniscectomy an average of 20 years earlier (range 17–22 years). Patients with cruciate ligament injury were excluded. All subjects were examined by standardized knee and hand radiography. Individual joints were considered to have OA when displaying radiographic features corresponding to a Kellgren/Lawrence (K/L) grade ≥2. Hand OA was considered present if at least 1 of the following criteria was fulfilled: the presence of radiographic OA (K/L grade ≥2) in at least 1 interphalangeal joint in each hand symmetrically, or in at least 2 distal or proximal interphalangeal joints in the same hand in a pattern consistent with primary OA (in the same row or ray), or in the first carpometacarpal joint bilaterally. The association between radiographic hand OA and radiographic knee OA was evaluated using logistic regression.

Results

Radiographic hand OA was present in 57 patients (34%) and radiographic knee OA was identified in 105 patients (62%), within 94 index knees (55%) and 47 contralateral knees (28%). In a multivariate model, radiographic hand OA was associated with an increased likelihood of radiographic OA in the index knee (odds ratio [OR] 3.0, 95% confidence interval [95% CI] 1.2–7.5) and in the nonoperated contralateral knee (OR 3.5, 95% CI 1.0–12.2).

Conclusion

The presence of radiographic hand OA is associated with an increased frequency of radiographic knee OA after meniscectomy. This finding confirms and extends that of a single previous study showing an interaction between hereditary and environmental risk factors for OA, a common and genetically complex disease. Accordingly, the development of OA following a meniscal tear and the resulting meniscal surgery should not be regarded to be of secondary origin only.

Meniscectomy of the knee is associated with a high risk of radiographic knee osteoarthritis (OA) and knee disability at long-term followup (1–5). The OA disease process that commonly follows such injury and intervention is suggested to depend on local biomechanical factors associated with the acute trauma and chronic overload on the joint cartilage generated by the loss of meniscal tissue, and may be further influenced by other risk factors such as obesity. However, recent evidence suggests that the risk of developing symptomatic or radiographic OA following a meniscal tear remains high, even if a limited meniscal resection has been performed (6, 7). These results raise the question as to whether, in addition to knee trauma and loss of meniscal function, an endogenous susceptibility to OA in the individual may contribute to the risk of meniscal tears and OA (8). Importantly, this would provide an example of interaction between heredity and environment for a common and genetically complex disease.

In support of this hypothesis, a single previous study published in 1983 investigated the relationship between radiographic hand OA and radiographic knee OA in individuals who had undergone meniscectomy, and found that the presence of hand OA was associated with a higher frequency of radiographic knee OA (9). Polyarticular hand OA has been suggested as a marker for generalized primary OA and shows marked heritability (10–15). Further studies on the association between hand OA and OA of the knee following meniscectomy might thus provide valuable information.

To evaluate the hypothesis that there is interaction between hereditary and environmental risk factors, we have studied a well-defined and carefully examined cohort of patients who were treated with meniscal surgery but did not have major ligament injuries in either knee; this cohort was examined within a narrow range of time after surgery and the loss to followup was only 20%. The primary objective was to examine the influence of radiographic hand OA on radiographic OA in the operated index knees as well as in the contralateral knees.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Patients

The regional ethics committee approved the study, and informed consent was obtained from all participating subjects. Patients undergoing isolated meniscectomy at Lund University Hospital in 1973 or 1978 were identified retrospectively through the surgical code system (patients in the Lund Health Care district were, at that time, obliged to seek medical treatment at Lund University Hospital). Current addresses were obtained from the National Population records.

The patients' medical records were assessed, and exclusion criteria were a report of death, being under 10 years of age at the time of surgery, relocation outside the South Swedish Health-Care region, associated cruciate ligament injury in any knee at the time of surgery or at the time of assessment, a diagnosis of rheumatoid arthritis, psoriatic arthritis, or osteochondritis dissecans, fracture adjacent to the knee, radiographic changes indicating knee OA at the time of surgery, or other knee surgery before the meniscectomy in the index knee. Following these exclusions, a total of 212 patients were invited to the examination, of whom 173 agreed to participate. We were unable to obtain hand radiographs in 3 cases, thus leaving a total of 170 patients in this study and a loss to followup of only 20% (Figure 1). One patient who underwent both medial and lateral meniscectomy at the index surgery was included, while all other index surgeries were performed in 1 knee compartment only. Total meniscectomy was performed on all patients except for 21 (12%) who underwent a limited meniscal resection (Table 1).

thumbnail image

Figure 1. Flow chart detailing the identification and formation of the study cohort.

Download figure to PowerPoint

Table 1. Characteristics of the patient cohort (n = 170)
CharacteristicValue
Demographic data 
 Age at assessment, mean ± SD years54 ± 11
 Sex, no. (%) female39 (23)
 Body mass index, mean ± SD kg/m226 ± 4
 Time since surgery, mean ± SD years20 ± 2
 Occupational workload, medianLight labor
 Spare-time physical activity level, medianModerate
 Medial meniscectomy, no. (%)133 (78)
 Total meniscectomy, no. (%)149 (88)
Type of meniscal tear, no. (%) 
 Degenerative 
  Flap38 (22)
  Horizontal3 (2)
  Tears in a meniscus with degenerative changes15 (9)
 Other 
  Longitudinal77 (45)
  Radial10 (6)
  No visible tear8 (5)
  Nonclassified19 (11)

The classification of the meniscal tears was modified from previous criteria (16). We grouped flap, horizontal, and complex tears and tears in a meniscus with degenerative changes as degenerative meniscal tears, according to findings in epidemiologic (17) and necropsy (18, 19) studies. Questionnaires were used to collect data regarding occupational workload and spare-time physical activity level. Workload was graded as heavy labor, moderate labor, light labor, or clerical work/unemployed/retired. Physical activity was categorized as very high, high, moderate, or no regular sports/physical activity. Examples from each category were given. All estimates were made retrospectively by the subject and divided into 5-year periods from the time of surgery until the time of assessment (Table 1).

Radiographic imaging and OA scoring

Posterioanterior hand images and radiographs of both knees were obtained using a Siemens Basic Radiographic System (Siemens GmbH, Erlangen, Germany). For the tibiofemoral joint, a fluoroscopically positioned anteroposterior radiographic beam was used, with the patient standing with knees in ∼15 degrees of flexion (film-focus distance 1.4 meters at 70 kV and 10 mA). An axial view of the patellofemoral joint was obtained with a vertical beam, with the patient standing with the knee in ∼50 degrees of flexion. One patient did not undergo radiographic examination of the contralateral knee.

All hand and knee radiographs were assessed for joint space narrowing (JSN) and osteophytes on a 4-point scale (range 0–3, with 0 indicating no JSN or no evidence of bony changes), in accordance with the atlas of Osteoarthritis Research Society International (OARSI) (20). We did not score sclerosis, cysts, erosions, or joint malalignment. When the patient had undergone subsequent tibial osteotomy or arthroplasty due to OA, JSN was regarded as grade 3 in the affected knee compartment or hand joint (1 patient had arthroplasty in the base of the thumb). In these cases, the contralateral and patellofemoral compartment of the knees and osteophytes were assessed on preoperative images when available, or when not available, were recorded as missing data. To ensure consistent assessment and avoid drift throughout the reading session, specific radiographs within the study material were selected as reference material and read repeatedly.

Within a period of 3 weeks, 2 trained observers (ME and PTP) separately assessed all distal interphalangeal (DIP) joints, proximal interphalangeal (PIP) joints, the interphalangeal joint of the thumb (IP1), and the first carpometacarpal (CMC1) joint of both hands. The observers were blinded to the surgical details, to each other's assessment, and to the patients' knee radiographs. Interobserver reliability (kappa statistics) was 0.88 for JSN and 0.79 for osteophytes, with complete agreement in 95% of the cases for both features. All discrepancies were reread and data obtained from the consensus reading were used for analysis.

In the individual hand joints, we considered radiographic OA to be present if any of the following criteria were fulfilled: 1) JSN grade ≥2, or 2) osteophyte grade ≥2, or 3) JSN grade 1 in combination with an osteophyte grade 1. We classified a subject as having radiographic hand OA if at least 1 of the following criteria was fulfilled: the presence of radiographic OA (as defined above) in 1) at least 1 DIP, PIP, or IP1 joint in each hand symmetrically (i.e., corresponding joints), or 2) at least 2 DIP/PIP joints in the same hand in a pattern consistent with primary OA (in the same row or ray), or 3) the CMC1 joint bilaterally.

One trained observer (ME) read the knee radiographs within a period of 5 days and was blinded to the surgical details and hand films. The tibiofemoral images had been read previously (4). Interrater reliability for the present and previous reading was 0.64 for JSN and 0.61 for the maximum osteophyte grade in each compartment (in the first reading, the observers only graded 1 maximum grade for each compartment), with complete agreement in 85% of compartments for JSN and 81% for osteophytes. We considered radiographic knee OA to be present if any of the following criteria were achieved in any of the 2 tibiofemoral compartments or in the patellofemoral compartment: 1) JSN grade ≥2, or 2) sum of the 2 marginal osteophyte grades ≥2, or 3) JSN grade 1 in combination with an osteophyte grade 1 (in the same compartment). The cutoff values for both finger and knee joint OA approximates a Kellgren/Lawrence OA radiographic grade ≥2 (21). To estimate radiographic OA severity in a knee, we calculated an OA score by adding all JSN and osteophyte grades in that knee.

Statistical analysis

Continuous data were analyzed using Student's t-test (parametric) or Mann-Whitney test (nonparametric) as appropriate. Binary data in 2 × 2 tables were evaluated by Fisher's exact test. The association between radiographic hand OA and radiographic knee OA was analyzed by logistic regression, and the other explanatory variables (age [continuous], sex, body mass index [BMI; continuous], occupational workload, spare-time physical activity level, meniscectomy compartment, and type of meniscal tear) were all entered in the model simultaneously. The odds ratio (OR) estimate with 95% confidence interval (95% CI) was based on the model. Crude ORs, and OR estimates after adjusting for age, sex, and BMI only, were also calculated. We considered a P value of less than or equal to 0.05 to be significant, and all tests were 2-tailed. Analyses were performed with SPSS for Windows, release 11.0.0 (SPSS, Chicago, IL).

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Radiographic hand OA.

Radiographic hand OA, as defined in the present study, was found in 57 patients (34%), of whom 48 (84%) had bilateral symmetric hand OA. Subjects with radiographic hand OA were older (P < 0.001) and had a degenerative type of meniscal tear more frequently (P = 0.006) than did patients without hand OA. Medial meniscectomy, rather than lateral meniscectomy, was more common in the subjects with radiographic hand OA (P = 0.048). In other respects, the demographic data of the patients with or without radiographic hand OA were comparable (Table 2).

Table 2. Comparison between subjects with and subjects without radiographic hand osteoarthritis (OA) who underwent meniscectomy
 Radiographic hand OA*P
Present (n = 57)Absent (n = 113)
  • *

    See Patients and Methods for the definition of radiographic hand OA.

  • The sum of all joint space narrowing (JSN) grades and osteophyte grades in the knee.

  • The highest grade of JSN from the knee compartments.

  • §

    The highest sum of the marginal osteophyte grades from the knee compartments.

  • Patients with meniscectomy in the contralateral knee (n = 39) are excluded.

Demographic data   
 Age at assessment, mean ± SD years61 ± 950 ± 9<0.001
 Sex, no. (%) male45 (79)86 (76)0.85
 Body mass index, mean ± SD kg/m226 ± 326 ± 40.89
 Time since surgery, mean ± SD years20 ± 220 ± 20.39
 Occupational workload, medianClericalLight labor0.17
 Spare-time physical activity level, medianModerateModerate0.70
 Medial meniscectomy, no. (%)50 (88)83 (73)0.048
 Meniscal tear, no. (%) degenerative27 (47)29 (26)0.006
 Total meniscectomy, no. (%)50 (88)99 (88)1.0
Radiographic outcome in the knee   
 Radiographic OA in either knee, no. (%) subjects46 (81)59 (52)<0.001
 Radiographic OA in index knee, no. (%) subjects41 (72)53 (47)0.002
  OA score, mean ± SD4.0 ± 3.42.6 ± 2.60.004
   JSN, mean ± SD1.1 ± 0.90.7 ± 0.80.004
   Osteophytes, mean ± SD§1.6 ± 1.21.1 ± 1.20.013
 Radiographic OA in nonoperated knee, no. (%) subjects17 (39)9 (10)<0.001
  OA score, mean ± SD2.0 ± 2.30.7 ± 1.2<0.001
   JSN, mean ± SD0.6 ± 0.60.2 ± 0.5<0.001
   Osteophytes, mean ± SD§0.9 ± 0.80.3 ± 0.6<0.001

Radiographic knee OA.

Radiographic knee OA was present in 105 subjects (62%), being localized in 94 index (operated) knees (55%) and 47 contralateral knees (28%). Of the patients without meniscectomy in the contralateral knee, 26 (20%) had radiographic OA in that contralateral knee. Subjects with degenerative meniscal tears in the index knee had radiographic knee OA more frequently than did patients with other tears, both in the index knee (P < 0.001) and in the contralateral knee (P = 0.039). Radiographic patellofemoral OA was found in 30 index knees (18%) and 15 contralateral knees (9%). Isolated patellofemoral radiographic OA (i.e., no tibiofemoral OA) was present in 2 index knees and 3 contralateral knees.

During the time interval between surgery and assessment, 11 subjects underwent further meniscal surgery in the index knee, 1 patient was treated with knee arthroplasty, and 5 patients underwent surgery in the index knee with a high tibial osteotomy for OA; one of these latter subjects and 2 additional patients underwent high tibial osteotomy of the contralateral knee. At the time of assessment, 39 patients (23%) had had meniscectomy performed in the contralateral knee. Information on the index surgery and any subsequent surgery was retrieved from the medical records at Lund University Hospital and by self-reported questionnaire from the subjects.

Association between radiographic hand OA and radiographic knee OA.

In 2 × 2 frequency table analyses, significantly more patients with radiographic hand OA than patients without radiographic hand OA had radiographic knee OA, both in the index knee and in the nonoperated contralateral knee (Table 2). Furthermore, patients with bilateral radiographic knee OA had radiographic hand OA more frequently than did patients with unilateral knee OA (61% versus 35%; P = 0.014).

In the multivariate logistic regression model, the association between radiographic hand OA and radiographic knee OA was statistically significant, both in the operated knees and in the nonoperated knees (Table 3). A similar analysis using an alternate and stricter criterion for radiographic hand OA (using only those with bilateral symmetric disease) also showed an association with radiographic knee OA for the index knees (OR 2.6, 95% CI 1.0–6.4, adjusted for all variables [age, sex, BMI, occupational workload, spare-time physical activity level, meniscectomy compartment, and type of meniscal tear]), but did not reach the level of statistical significance for the nonoperated contralateral knees (OR 2.3, 95% CI 0.7–7.8).

Table 3. The association between radiographic hand osteoarthritis (OA) and radiographic knee OA 20 years after meniscectomy*
Radiographic hand OAIndex (operated) knee (n = 170)Contralateral (nonoperated) knee (n = 130)
Knee OA, no. (%)OR95% CIKnee OA, no. (%)OR95% CI
  • *

    See Patients and Methods for the definition of radiographic hand OA. OR = odds ratio; 95% CI = 95% confidence interval; BMI = body mass index.

  • Age, sex, BMI, occupational workload, spare-time physical activity level, meniscectomy compartment, and type of meniscal tear.

Absent53/113 (47)1.0Reference9/86 (10)1.0Reference
Present      
 Crude estimate41/57 (72)2.91.5–5.817/44 (39)5.42.1–13.5
 Adjusted for age, sex, and BMI 2.91.3–6.4 2.60.9–7.5
 Adjusted for all variables 3.01.2–7.5 3.51.0–12.2

We also repeated the multivariate analysis after excluding patients with radiographic OA in the contralateral knee, because the radiographic changes in the operated knee might be considered more likely, being exclusively secondary to the meniscectomy. The estimated OR still suggested a tendency toward an increased frequency of knee OA if radiographic hand OA was present (OR 2.6, 95% CI 0.8–8.1, adjusted for all variables).

Association between radiographic hand OA and radiographic severity of the knee OA.

In order to avoid any influence of disease prevalence when comparing the knee OA severity between patients with and without radiographic hand OA, we confined our analyses to only patients with radiographic features of OA (i.e., OA score ≥1) in the knee investigated. In age-categorized analyses, we found that patients below 50 years of age who had radiographic hand OA had more severe radiographic changes in the index knee than did patients without hand OA in the same age category (Table 4). There were no significant differences in the severity of the radiographic changes in the index knee in the 2 older age categories. Similar to the evaluation of disease prevalence, we also repeated the analyses of radiographic severity in the index knee when subjects with radiographic OA in the contralateral knee were excluded. The OA severity score remained significantly higher for subjects with radiographic hand OA than for those without hand OA in the younger age category (P = 0.05).

Table 4. Comparison of the radiographic severity of knee osteoarthritis (OA) between subjects with and without radiographic hand OA who underwent meniscectomy*
Age, yearsIndex (operated) knee (n = 146)Contralateral nonoperated knee (n = 71)
Hand OA presentHand OA absentPHand OA presentHand OA absentP
  • *

    Except where indicated otherwise, values are mean ± SD OA scores. The OA score was calculated as the sum of all joint space narrowing grades and osteophyte grades in the knee. To avoid influence of the disease prevalence, only patients with radiographic features of OA in the knee (OA score ≥1) are included. See Patients and Methods for the definition of radiographic hand OA.

<506.8 ± 4.4 (n = 6)2.7 ± 1.9 (n = 49)0.0131.5 ± 0.7 (n = 2)1.5 ± 0.8 (n = 15)0.94
50–593.9 ± 3.2 (n = 17)3.3 ± 3.3 (n = 28)0.222.4 ± 1.9 (n = 11)1.1 ± 0.5 (n = 14)0.021
≥604.3 ± 3.0 (n = 28)3.7 ± 3.0 (n = 18)0.422.6 ± 2.5 (n = 22)3.1 ± 2.3 (n = 7)0.67

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

We found that the presence of radiographic hand OA is associated with an increased frequency of radiographic knee OA after meniscectomy. Patients with hand OA at an age younger than 50 years already have more advanced radiographic knee changes than do patients in the same age category whose hands are healthy. These findings support previous evidence of the influence of endogenous factors on the risk of OA after meniscal lesions and meniscectomy.

Several epidemiologic studies have shown an association between OA of the hand and OA of the knee in community-based study populations, suggesting that hand OA is often part of a more generalized OA (22–24). Hand OA has a significant heredity, indicating the presence of a genetic risk factor (13, 15). The only previous study investigating an association between hand OA and knee OA in individuals who had undergone meniscectomy, as a human model of isolated joint damage, was in 1983 (9). That study was the first to suggest an interaction between local joint injury and systemic factors in OA. Thus, the classic view of secondary OA may be incorrect and any distinction from primary OA may not be as clear as previously thought. Some limitations associated with the previous study provided a rationale for the present investigation. The previous study involved a heterogeneous study group with regard to injury type and extent, a wide range of time since surgery, and a high dropout rate, and evaluation of radiographic severity was confounded by disease prevalence. Furthermore, the type of meniscal tear was not evaluated and the radiographic criteria used to consider a joint diseased were difficult to interpret.

In the present study, we confirmed the association between radiographic hand OA and radiographic knee OA after meniscal injury, for both the operated knee and the nonoperated knee. These findings provide additional support for an interaction between genetic and environmental factors in OA. We also confirmed a strong association between degenerative meniscal tears and radiographic OA in the operated knee. We have previously reported that degenerative meniscal tears are associated with a worse long-term patient-relevant knee outcome in individuals after limited meniscal resection (7), and this type of tear has also been associated with radiographic OA and with combined radiographic and symptomatic OA, i.e., clinical OA (5). Meniscal degenerative pathology has been correlated with degenerative cartilage changes (25), and meniscal tears have been found to be highly prevalent in OA of the knee but lack a clear effect on the patient's symptoms and functional status (26). We therefore hypothesize that the meniscal tear itself may, in some individuals, indicate the presence of incipient OA, a disease involving changes not only in joint cartilage, but also in the whole joint (27). This type of tear could be caused by the degenerated tissue's decreased ability to withstand loads and force transmissions during knee joint movement, and may develop spontaneously or in conjunction with minor knee trauma. Thus, patients with these types of meniscal symptoms and who undergo surgery due to degenerative tears may provide a subpopulation enriched in individuals with an inherited tendency to develop OA.

It was suggested that, in addition to an increased frequency of knee OA, the severity of the radiographic changes was greater if hand OA was present (9). This previous suggestion was based on data from all knees, regardless of whether or not they were considered osteoarthritic. In order to compare the disease severity, patients who are not classified as having the disease, or, alternatively, patients without any radiographic sign of the disease, should be excluded from analysis; otherwise, the result would be influenced by the disease prevalence. Due to the small number of patients in both the previous and present study, an accurate estimation of the association between disease severity and hand OA is difficult. In the present study, we could not confirm a generally increased disease severity if radiographic hand OA was present, when taking into account the age difference between the groups. However, our results do indicate that patients with radiographic hand OA at an early age may have both more frequent and more severe radiographic knee changes than that seen in patients with no hand OA in the corresponding age category, which further supports the possibility of an influence by a genetic risk factor.

The present study has the limitations inherent in any retrospective cohort study. The surgical records were not standardized, but were usually detailed and easily interpreted by the investigators. The radiologist's written statement from the preoperative examination was used to exclude patients with knee OA at the time of surgery. In 16 cases (9%), the preoperative radiographic statement or films were not available, and therefore individual patients with radiographic knee OA present already at baseline may have been included. We excluded patients with cruciate ligament injury in any knee, thereby creating a homogeneous study cohort with regard to the environmental risk factor. Exclusion of subjects with this injury may have increased the proportion of degenerative meniscal tears in the study population. The OA classification was based on separate readings of JSN and osteophytes, in accordance with the atlas from OARSI scoring method (20), and the cutoff value for individual knee OA and hand joint OA approximated a Kellgren/Lawrence grade ≥2 (21). Thus, we did not classify a single grade 1 osteophyte or grade 1 JSN alone as radiographic OA, but we did recognize that a grade 1 osteophyte may represent emergent disease (28, 29). We based our definition of radiographic hand OA on polyarticular unilateral involvement in a typical pattern, or bilateral symmetric hand OA, thus eliminating patients with possible secondary hand OA. The small number of patients limits the statistical power of some of the analyses, especially for the nonoperated contralateral knees.

In conclusion, we report that the presence of radiographic hand OA is associated with an increased frequency of radiographic knee OA after meniscectomy. It would appear from present and previously reported results that there is an interaction between hereditary and environmental factors, and not only knee trauma, that leads to meniscal tears and meniscal surgery. A degenerative tear may be regarded as an early signal of susceptibility to osteoarthritic disease. Consequently, we consider the widely accepted classification of knee OA after meniscectomy as secondary OA to be misleading. The relative importance of genetics, knee trauma, and surgery in the initiation and progression of OA after meniscal tears remains to be further explored.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES