Abstract
The risk of post-traumatic osteoarthritis following an intra-articular fracture is determined to large extent by the success or failure of osteochondral repair. To measure the efficacy of osteochondral repair in a primate and determine if osteochondral repair differs in the patella (PA) and the medial femoral condyle (FC) and if passive motion treatment affects osteochondral repair, we created 3.2 mm diameter 4.0 mm deep osteochondral defects of the articular surfaces of the PA and FC in both knees of twelve skeletally mature cynomolgus monkeys. Defects were treated with intermittent passive motion (IPM) or castimmobilization (CI) for two weeks, followed by six weeks of ad libitum cage activity. We measured restoration of the articular surface, and the volume, composition, type II collagen concentration and in situ material properties of the repair tissue. The osteochondral repair response restored a mean of 56% of the FC and 34% of the PA articular surfaces and filled a mean of 68% of the chondral and 92% of the osseous defect volumes respectively. FC defect repair produced higher concentrations of hyaline cartilage (FC 83% vs. PA 52% in chondral defects and FC 26% vs. PA 14% in osseous defects) and type II collagen (FC 84% vs. PA 71% in chondral defects and FC 37% vs. PA 9% in osseous defects) than PA repair. IPM did not increase the volume of chondral or osseous repair tissue in PA or FC defects. In both PA and FC defects, IPM stimulated slightly greater expression of type II collagen in chondral repair tissue (IPM 81% vs. CI 74%); and, produced a higher concentration of hyaline repair tissue (IPM 62% vs. CI 42%), but IPM produced poorer restoration of PA articular surfaces (IPM 23% vs. CI 45%). Normal articular cartilage was stiffer, and had a larger Poisson's ratio and less permeability than repair cartilage. Overall CI treated repair tissue was stiffer and less permeable than IPM treated repair tissue. The stiffness, Poisson's ratio and permeability of femoral condyle cast immobilized (FC CI) treated repair tissue most closely approached the normal values. The differences in osteochondral repair between FC and PA articular surfaces suggest that the mechanical environment strongly influences the quality of articular surface repair. Decreasing the risk of posttraumatic osteoarthritis following intra-articular fractures will depend on finding methods of promoting the osteochondral repair response including modifying the intra-articular biological and mechanical environments.
INTRODUCTION
Intra-articular osteochondral repair occurs under unique complex biological and mechanical conditions 4,6,20. Disruption of subchondral and metaphyseal bone creates conditions similar to bone fractures, but the bone repair response extends into the chondral portion of the defect. At the same time the chondral and osseous portions of the defect are exposed to the synovial environment including the synovial fluid and the cellular and vascular reactions in the synovial membrane4. Joint loading and motion causes poorly understood mechanical stresses in the chondral and bony portions of the defect that do not occur with any other type of injury.
The success or failure of acute osteochondral repair is critical in determining the outcome of intra-articular fractures, a clinical problem faced by tens of thousands of patients and surgeons each year20. Yet, basic research and clinical experience give only a vague outline of the osteochondral repair response and the factors that influence this response. Immediately following an acute articular surface fracture, the osseus portion of the defect and a part of the chondral defect fill with a fibrin clot. Over a period of approximately six weeks the fibrin clot is replaced with repair tissue that contains cells with the appearance of chondrocytes in the chondral part of the defect and osteoblasts in the osseous part of the defect2,3,5,6,8,9. In some instances the repair tissue remodels and functions well as a joint surface for decades, but in others the chondral portion of the repair tissue progressively degenerates, a process that frequently leads to the clinical syndrome of joint pain, deformity and dysfunction recognized as post-traumatic osteoarthritis. The mechanical and biological events that influence the formation and remodeling or degeneration of osteochondral repair tissue have received little attention.
Current treatment of intra-articular fractures consists of attempting to restore the alignment and position of the fragments of the fractured articular surface and stabilizing these fragments in the reduced position. Even when surgeons oppose the fracture fragments and restore the alignment and contours of the articular surface, the injured joint frequently undergoes progressive degeneration. The reasons for joint degeneration following anatomic restoration of a fractured articular surface remain poorly understood20, but to a large degree this phenomenon is related to failure of the osteochondral repair response to restore a normal articular surface. Some intra-articular fractures cause such severe osteochondral damage that despite optimal treatment, large gaps or defects are left in the articular surface. In these gaps, the restoration of an articular surface depends on production of substantial quantities of chondral and osseous repair tissue and the remodeling of the repair tissue to form a functional articular surface. For these reasons better understanding of osteochondral healing is critical for the advancement of treatment of intra-articular fractures and decreasing the risk of post-traumatic osteoarthritis.
The limited ability of the natural osteochondral repair response to restore an articular surface has been recognized for more than 250 years15,17,22. Continuous passive motion treatment of one mm diameter osteochondral defects in rabbit knees has been shown to facilitate chondral repair24, and more recently, multiple investigators have reported improved healing of acute experimental articular surface defects with growth factors, cell transplants and artificial matrices2,3,5,6,10,11,13,14,16,18,19,21. However these studies have not examined osteochondral healing in primates, potential differences in osteochondral healing in regions of the joint subjected to different patterns of mechanical forces, and the effects of passive motion on repair of different regions of the joint. The purpose of this study was to answer the following questions. How successfully does the natural osteochondral repair response restore an articular surface in skeletally mature primates? Does repair of patellar and femoral condylar osteochondral defects differ? Do cast and intermittent passive motion treatment of acute osteochondral defects produce different results?
MATERIALS AND METHODS
Experimental Design
We created identical cartilage defects in the central weight-bearing articular surface of the medial femoral condyle and the central articular surface of the patella in both knee joints of twelve skeletally mature cynomolgus male monkeys. All surgery was performed aseptically under inhalation anesthesia at the Ohio State Department of Veterinary Clinical Sciences. A 3.2 mm drill bit and a drill sleeve with an adjustable stop, which ensured that the depth of the hole would be 4 mm, were used to create uniform defects. Animals were given postoperative analgesia. For two weeks postoperatively, the animals were maintained in a seated position in padded restraint chairs so that one limb could be immobilized in a long leg cast in 15 degrees of flexion at the knee and the contralateral limb was subjected to IPM (one cycle/40 seconds) for 16 hours a day, seven days a week. The animals had free arm movement but were controlled from the waist down and the neck up. The casted limb was suspended from the chair with the hip flexed to 90°. The foot of the IPM limb was secured in a cloth boot that was attached to a pedal on a fly wheel. The pedal's height was adjusted to each individual animal so that the limb would go through a full range of motion in the seated position but the limb would not be over-extended or over-flexed. A bicycling motion was thus produced. The motor on the fly wheel was set at a speed that would take the limb through a full range of motion every 40 seconds. The animals were monitored constantly and their skin checked daily for problems associated with the restraint chairs. At the end of one week in the chairs, the monkeys were placed in their cages for 12 hours to allow for chair maintenance and examination of each animal. At the end of two weeks the cast was removed and the animals were allowed cage activity ad libitum for six weeks, before sacrifice. None of the animals developed complications of the surgery or treatment of the knee joints. The eight knees from four randomly selected animals were used for studies of the material properties of the repair tissue and the sixteen knees from the remaining eight animals were used for histologic and biochemical studies. The knees of two normal cynomolgus monkeys were harvested for use as normal controls. Knee joints were harvested at the Ohio State University, frozen and shipped to the University of Iowa for analysis.
Articular Surface Restoration
At the University of Iowa, the articular surfaces of all 48 defects were photographed the images enlarged 40 times with maximum resolution and printed. At this magnification it was possible to identify regions of the repair tissue surface that had the same smooth white appearance as the surrounding articular cartilage and regions that had an irregular appearance. A transparent grid with 100 equally spaced points was superimposed on the images of the repair tissue surface and the number of points falling on repair tissue with the appearance of normal articular cartilage and the total number of points falling on the defect recorded. For each experimental group the percent of repair tissue with a normal appearance was determined. Differences among groups were evaluated using t-tests.
Repair Tissue Volume and Composition
To evaluate the chondral and osseus repair tissue volume and composition, 32 defect sites from eight animals were briefly decalcified and divided along the sagittal plane into two equal parts. One half of each defect was randomly selected to measure the type II collagen concentration in the chondral and osseous parts of the defect. A dissecting microscope was used to separate the chondral and osseous repair tissue in the specimens used to measure the type II collagen concentration. The osseous repair tissue was decalcified and the type II collagen concentration measured in the chondral and osseous repair tissue as previously described12,23. The other half of each defect was decalcified, fixed and embedded in paraffin. Beginning with the center of the defect, four full thickness sections of the defect and surrounding tissue were collected by taking the first, fifth, tenth and fifteenth section. After staining with Safranin O, images of the sections were projected with a superimposed grid with 200 equally spaced points. The total area of repair tissue relative to the total area of the original defect was determined by point counting for each section as was the area of hyaline cartilage and the area of fibrocartilage or fibrous tissue in the chondral part of the defect and the area of bone matrix, bone marrow, hyaline cartilage, fibrocartilage or fibrous tissue in the osseous part of the defect. The total areas of chondral and bone repair tissue and each repair tissue type were then calculated for each defect.
Repair Tissue Material Properties
The eight knee joints randomly selected for biomechanical testing were shipped frozen from the University of Iowa to Columbia University1. The joints were stored at -20°C and underwent two freeze-thaw cycles before testing. Sixteen repair sites in the patella and the medial femoral condyle were identified. Two of the 16 defects were not tested, because they were devoid of repair cartilage. Normal, cynomolgus monkey knee joints were used as controls for the experimental specimens. Three mechanical properties of repair tissue (aggregate modulus, Poisson's ratio, and permeability), were measured as previously described 1. After mechanical testing, the repair sites were sectioned and the thickness of the chondral repair tissue measured with a digital scale using a stereomicroscope 1.
RESULTS
Articular Surface Restoration
All 48 defect sites could be identified by differences in the appearance of the chondral repair tissue and the surrounding articular cartilage (Figures 1 and 2). The degree of restoration of the articular surface varied among animals (Table 1). The extent of filling of the articular surface defect with chondral repair tissue varied from more than 90% to less than 10% (Figures 1 and 2). The extent of articular surface restoration did not differ between femoral defects treated with CI or IPM, or between femoral and patellar defects treated with CI, but IPM treated patellar defects had significantly less restoration of the articular surface (Table 1 and Figure 3).
Figure 1.
Photographs of the surface of femoral condylar defects eight weeks after creation of the defects showing the variability in the extent of chondral repair tissue among defects.
Top. The poorest repair of a femoral condylar defect.
Middle.
Moderate chondral repair tissue.
Bottom.
The specimen with the most complete repair of the defect. The outlines of the defect are difficult to identify.
Figure 2.
Photographs of the surface of patellar defects eight weeks after creation of the defects showing the variability in the extent of chondral repair tissue among defects.
Top.
The poorest repair of a patellar defect.
Middle.
Some chondral repair tissue is visible.
Bottom.
The specimen with the most complete repair of the defect.
TABLE 1. Articular Surface Restoration: (% of 3.2 mm diameter defect).
Bars connect means that are not significantly different (p<0.01)
Figure 3. Articular Surface Restoration.
Histogram showing that the osteochondral repair response restored more of the articular surface in femoral defects than in patellar defects and that patellar defects treated by cast immobilization had better restoration of the articular surface than patellar defects treated by IPM.
Repair Tissue Volume and Composition
The extent and quality of the repair response varied among animals (Figure 4), but certain characteristics of the repair response occurred consistently. Osseous repair was substantially more effective than chondral repair. The repair response filled 68% of the total volume of the chondral defects and 92% of the volume of the osseous defects (Figure 5). Repair of femoral defects produced significantly more hyaline repair tissue and contained higher concentrations of type II collagen than repair of patellar defects (Table 2 and Figure 6). Intermittent passive motion treatment did not increase the volume of osseous or chondral repair tissue in either patellar or femoral defects. Passive motion treatment did increase the type II collagen concentration in femoral and patellar chondral defects. In patellar defects, passive motion treatment produced a higher concentration of hyaline repair tissue.
Figure 4. Photomicrographs showing the variability in osteochondral repair among animals.
Top.
The osseous portion of the defect is filled with new bone and chondral repair tissue projects above the normal articular surface.
Bottom.
The osseous portion of the defect contains fibrovascular tissue as well as bone and the chondral portion of the defect contains primarily fibrous tissue.
Figure 5. Volume of Repair Tissue.
Histogram showing that repair tissue filled more than 90% of the osseous defect volume but only about 68% of the chondral defect volume.
TABLE 2. Volume and Composition of Repair Tissue.
Bars connect means that are not significantly different (p<0.01)
Figure 6.
Photomicrograph showing chondral repair tissue with a hyaline appearance.
Repair Tissue Material Properties
The material properties of repair tissue differed from those of normal articular cartilage (Figure 7)1. The normal articular surface (Table 3) was four times stiffer than the osteochondral repair tissue (p<0.0001). Similarly, the Poisson's ratio of the repair tissue was about half that of the normal articular surface (p<0.01) and the repair tissue was more permeable (p<0.01).
Figure 7. Material Properties of Repair Tissue.
Histogram showing that the osteochondral repair response produced tissue that most closely approached the properties of a normal articular surface in femoral condylar defects treated by cast immobilization.
TABLE 3. Material Properties of Repair Tissue.
| Treatment and Location | Number of Animals | Aggregate Modulus HA (MPa) | Poisson's Ratio vs | Permeability kx1015 (m4/N-s) | Cartilage Thickness at Test Site (mm) | Cartilage Thickness Adjacent to Test Site (mm) |
|---|---|---|---|---|---|---|
| Normal Femoral Condyle |
6 | 0.82±0.18 | 0.24±0.06 | 2.44±1.13 | 0.72±0.09 | |
| IPM Patella |
4 | 0.12±0.04 | 0.09±0.10 | 3.17±1.06 | 2.05±0.81 | 0.66±0.09 |
| Cast Patella |
4 | 0.22±0.06 | 0.10±0.01 | 3.38±0.48 | 2.18±0.73 | 0.87±0.57 |
| IPM Femoral Condyle |
3 | 0.17±0.06 | 0.11±0.03 | 3.86±0.55 | 2.22±0.56 | 0.71±0.14 |
| Cast Femoral Condyle |
3 | 0.27±0.22 | 0.16±0.11 | 2.40±1.45 | 2.19±0.58 | 0.60±0.21 |
Passive motion and cast treated osteochondral repair tissue differed in material properties. The repair tissue had a higher compressive modulus (p<0.10) in joints treated with casts (HA=0.244±0.138 MPa) than with passive motion (HA=0.148±0.053 MPa). The aggregate modulus of patellar repair tissue was significantly higher in the cast treated group (p<0.05). Similarly, repair tissue in the femoral condyle was more permeable in the passive motion treated defects (p<0.10).
Chondral Repair Tissue Thickness
At eight weeks following creation of the osteochondral defects the mean thickness of the chondral repair tissue was 2.16 mm, compared with 0.72 mm for normal femoral articular cartilage and 0.71 mm for the cartilage adjacent to the defects (Table 3). The original depth of the osteochondral defects was 4 mm, a 0.72 mm deep chondral defect and a 3.18 osseous defect. Thus, by eight weeks, bone had not filled the entire original bone defect and the chondral repair tissue was slightly more than three times the thickness of normal articular cartilage.
DISCUSSION
The results of this study confirm that the repair response to an acute osteochondral articular surface injury in a primate fails to restore a normal articular surface by eight weeks. It is possible that with more time the quality of the chondral repair tissue would have improved and the volume would have increased, but given previous studies of articular cartilage repair2,5,9,12 it seems more likely that the chondral repair tissue would degenerate with time. Cynomolgus monkey normal articular cartilage was about four times stiffer than osteochondral repair tissue. The Poisson's ratio of the normal articular surface, which is indicative of the tissue's apparent compressibility, was twice that of the repair cartilage, and the permeability was greater. Because of its inferior material properties, normal joint use would subject the repair tissue to elevated strain fields causing progressive degeneration of the repair tissue.
The difference in the repair tissue in the chondral and osseous portions of the defects eight weeks following injury was striking. Initially the chondral and osseus portions of the osteochondral defect fill with the same clot and then the same initial repair tissue5,7,9. Although hyaline cartilage and synthesis of type II collagen occur in both the chondral and osseous portions of the defect the concentration of hyaline cartilage and type II collagen was greater in the chondral portion. Bone matrix and marrow appeared only in the osseus part of the defect. The explanation for the differentiation of the tissue in the chondral part of the defect toward hyaline cartilage and the repair tissue in the osseus part of the defect toward bone remains unknown. Possible explanations include differences in mechanical environment, and diffusion of biologically active molecules from the synovial fluid, the bone or the cartilage.
An equally interesting result is the increased concentration of hyaline repair tissue and type II collagen in the femoral chondral and osseous defects as compared with the patellar defects. The initial repair tissue in both locations presumably is the same, yet by eight weeks the patellar chondral defects have more fibrous tissue and the patellar osseous defects have more bone matrix than the corresponding femoral defects. Since the defects were made in the same joint at the same time it seems unlikely that the maturing repair tissue would be exposed to important differences in biologically active molecules, cells, free radicals or oxygen tension. The only apparent difference between the two locations is the pattern of mechanical loading with joint use, an observation that suggests that different locations in the same joint and different joints may differ in osteochondral repair depending on the patterns of loading of the repair tissue.
The optimal mechanical environment for osteochondral and chondral repair has not been defined. In this study, two weeks of IPM did not stimulate more effective osteochondral repair than two weeks of cast immobilization. These conclusions do not discount possible beneficial effects that continuous passive motion treatments may have on articular cartilage healing in smaller defects or in other species.
This study shows that the important limitation of the natural acute osteochondral repair response in a primate is the failure to produce a sufficient volume and quality of chondral repair tissue. However, it also shows that the acute osteochondral repair response produces chondral repair tissue that more closely approximates articular cartilage than bone or dense fibrous tissue and that temporary cast treatment of femoral condylar defects followed by active movement produces the best chondral repair. These observations indicate that osteochondral repair can be improved and the risk of post-traumatic osteoarthritis following intra-articular fractures decreased by better understanding of the biological and mechanical conditions that govern acute osteochondral repair.
ACKNOWLEDGMENTS
This work was sponsored by a grant from the OREF (Career Development Award, MPR), NIH grants AR 38733 and P50 AR48939. http://poppy.obrl.uiowa.edu/Specialized Center of Research/SCOR.htm.
References
- 1.Athanasiou KA, Rosenwasser MP, Buckwalter JA, Olmstead M, Mow VC. Biomechanical modeling of repair articular cartilage: effects of passive motion on osteochondral defects in monkey knee joints. Tissue Engineering. 1998;4:185–195. [Google Scholar]
- 2.Buckwalter JA. Articular Cartilage Injuries. Clin Ortho Rel Res. 2002;402:21–37. doi: 10.1097/00003086-200209000-00004. [DOI] [PubMed] [Google Scholar]
- 3.Buckwalter JA. Evaluating methods of restoring cartilagenous articular surfaces. Clin Ortho Rel Res. 1999;367(S):224–238. doi: 10.1097/00003086-199910001-00022. [DOI] [PubMed] [Google Scholar]
- 4.Buckwalter JA, Einhorn TA, Marsh JL. Bone and Joint Healing. In: Rockwood CA, Green DP, Bucholz RW, Heckman JD, editors. Fractures. Philadelphia: Lippincott; 2001. pp. 245–271. [Google Scholar]
- 5.Buckwalter JA, Mankin HJ. Articular cartilage II. Degeneration and osteoarthrosis, repair, regeneration and transplantation. J Bone Joint Surg. 1997;79A(4):612–632. [Google Scholar]
- 6.Buckwalter JA, Martin JA. Promoting Articular Cartilage Repair. In: Tsokos GC, Moreland LW, Kammer GM, Pelletier J-P, Martel-Pelletier J, Gay S, editors. Modern Theraputics in Rheumatic Diseases. Totowa, New Jersey: Humana Press; 2002. pp. 201–214. [Google Scholar]
- 7.Buckwalter JA, Mow VC. Cartilage Repair in Osteoarthritis. In: Moskowitz RW, Howell DS, Goldberg VM, Mankin HJ, editors. Osteoarthritis: Diagnosis and Management. 2nd edition. Philadelphia: Saunders; 1992. pp. 71–107. [Google Scholar]
- 8.Buckwalter JA, Rosenberg LA, Hunziker EB. Articular cartilage: Composition, structure, response to injury, and methods of facilitation repair. In: Ewing JW, editor. New York: Raven Press; 1990. pp. 19–56. [Google Scholar]
- 9.Buckwalter JA, Rosenberg LC, Coutts R, Hunziker E, Reddi AH, Mow VC. Articular cartilage: Injury and repair. In: Woo SL, Buckwalter JA, editors. Injury and Repair of the Musculoskeletal Soft Tissues. Park Ridge, IL: American Academy of Orthopaedic Surgeons; 1988. pp. 465–482. [Google Scholar]
- 10.Coutts RD, Healey RM, Ostrander R, Sah RL, Goomer R, Amiel D. Matrices for cartilage repair. Clin Orthop. 2001. pp. S271–S279. [DOI] [PubMed]
- 11.Fortier LA, Mohammed HO, Lust G, Nixon AJ. Insulin-like growth factor-I enhances cell-based repair of articular cartilage. J Bone Joint Surg. (Br) 2002;84(2):276–288. doi: 10.1302/0301-620x.84b2.11167. [DOI] [PubMed] [Google Scholar]
- 12.Furukawa T, Eyre DR, Koide S, Glimcher MJ. Biochemical studies on repair cartilage resurfacing experimental defects in the rabbit knee. J Bone Joint Surg. (Am) 1980;62(1):79–89. [PubMed] [Google Scholar]
- 13.Gao J, Dennis JE, Solchaga LA, Goldberg VM, Caplan AI. Repair of osteochondral defect with tissue-engineered two-phase composite material of injectable calcium phosphate and hyaluronan sponge. Tissue Eng. 2002;8(5):827–837. doi: 10.1089/10763270260424187. [DOI] [PubMed] [Google Scholar]
- 14.Gelse K, Von Der Mark K, Aigner T, Park J, Schneider H. Articular cartilage repair by gene therapy using growth factor-producing mesenchymal cells. Arthritis Rheum. 2003;48(2):430–441. doi: 10.1002/art.10759. [DOI] [PubMed] [Google Scholar]
- 15.Hunter W. On the structure and diseases of articulating cartilage. Philos Trans R Soc London. 1743;42:514–521. [Google Scholar]
- 16.Hunziker EB, Driesang IM, Morris EA. Chondrogenesis in cartilage repair is induced by members of the transforming growth factor-beta superfamily. Clin Orthop. 2001. pp. S171–S181. [DOI] [PubMed]
- 17.Keith A. Menders of the Maimed. London: Oxford University Press; 1919. p. 335. [Google Scholar]
- 18.Lee CR, Grodzinsky AJ, Hsu HP, Spector M. Effects of a cultured autologous chondrocyte-seeded type II collagen scaffold on the healing of a chondral defect in a canine model. J Orthop Res. 2003;21(2):272–281. doi: 10.1016/S0736-0266(02)00153-5. [DOI] [PubMed] [Google Scholar]
- 19.Liu Y, Chen F, Liu W, Cui L, Shang Q, Xia W, Wang J, Cui Y, Yang G, Liu D, Wu J, Xu R, Buonocore SD, Cao Y. Repairing large porcine fullthickness defects of articular cartilage using autologous chondrocyte-engineered cartilage. Tissue Eng. 2002;8(4):709–721. doi: 10.1089/107632702760240616. [DOI] [PubMed] [Google Scholar]
- 20.Marsh JL, Buckwalter JA, Gelberman R, Dirschl D, Olson S, Brown TD, Llinias A. AOA Symposium-Articular Fractures: Does an anatomic reduction really change the result? J Bone Joint Surg. 2002;84A:1259–1271. [PubMed] [Google Scholar]
- 21.Mierisch CM, Cohen SB, Jordan LC, Robertson PG, Balian G, Diduch DR. Transforming growth factor-beta in calcium alginate beads for the treatment of articular cartilage defects in the rabbit. Arthroscopy. 2002;18(8):892–900. doi: 10.1053/jars.2002.36117. [DOI] [PubMed] [Google Scholar]
- 22.Paget J. Healing of cartilage. In: Bick EM, editor. Classics of Orthopaedics. Philadelphia: Lippincott; 1976. pp. 276–277. [Google Scholar]
- 23.Pedrini-Mille A, Pedrini V. Proteoglycans and glycosaminoglycans of human achondroplastic cartilage. J Bone Joint Surg. 1982;64A:39–46. [PubMed] [Google Scholar]
- 24.Salter RB, Simmons DB, Malcolm BW, Rumble EJ, MacMichael D, Clements ND. The biological effect of continuous passive motion on the healing of full-thickness defects in articular cartilage: an experimental investigation in the rabbit. J Bone Joint Surg. 1980;62A:1232–1251. [PubMed] [Google Scholar]