Chingford phenotype list

The Chingford Study population was established in 1989 as a retrospective case-control study to determine prevalence rates of osteoarthritis (OA) and osteoporosis (OP) in middle-aged women in the general population, and to assess a number of known risk factors and their associations with these two diseases. It has since become a prospective population-based longitudinal cohort of women seen annually and described in detail (Hart DJ & Spector TD). It is listed by the NIH as an important epidemiological resource and one of the few such cohorts with wide-ranging musculoskeletal data (http://www.niams.nih.gov/ne/oi/oaepipappen_b.htm).

The original response rate of the sample was 78%. The cohort consisted originally of 1003 middle-aged women aged 45-64 from a general practice in Chingford, North-East London. Since 1989 the women have been assessed annually with a number of investigations. At the beginning of the 15th year of the study 776 women are still attending the clinic, 77% of the original cohort. Eighty-four women have died, 62 women have moved away and 81 women have left the study. The women in the study are flagged and their deaths and any cancers are being tracked and reported to us through Office of National Statistics. The adjusted response rate (minus those who can no longer participate) is 85%. Reasons for attaining such a good response rate include close location of the clinic, continuity of well-prepared staff, care and preparation for each visit, and close GP liaison. The women are confident of an efficient service enabling the annual visit to be very time and cost effective.

The primary focus to date has been the study of the natural history of osteoarthritis and osteoporosis. A number of clinical, anthropometric, psychosocial, radiological and metabolic variables have been collected at two or more time points. Investigations have been carried out on all subjects and DNA samples on 900 individuals have been collected at least twice between 1994 and 2000. Over 70 publications to date (in high quality rheumatology and bone journals) have been based on data from this cohort. A summary of the Chingford study descriptive statistics is shown in Table 1.

The strength of the Chingford Study in the last 5 years has been its use for collaborative research due to the detail of the longitudinal data. We have been able to determine not only common risk factors for musculo-skeletal disease within the subjects, but to investigate associated genetic and biologic phenotypes in the cohort and explore new methodology in imaging longitudinal changes in these common diseases.

Table 1 Chingford cohort: descriptive statistics

Trait mean (SD) (%)	Year 1 (n=1003)	Year 9/10 follow-up (n=830)
Age (years)	54.05 (6.4)	64.5 (6.4)
BMI (kg/m2)	25.3 (4.3)	26.9 (6.5)
Waist/hip ratio	0.77 (0.06)	N/A
Ever smoker	45.8%	44.7%
Ever HRT	45.8%	48.1%
Current HRT	45.8%	20.8%
Hypertension I	16.4%	27.9%
Hypertension II-IV	5.7%	12.0%
% diabetes mellitus	0.9%	1.4%
Fasting glucose mmol/L	4.92 (1.15)	5.26 (0.97)
c-reactive protein mg/L	2.28 (3.49)	3.55 (4.54)
% radiographic knee OA	21.0%	37.9%
% radiographic hip OA	12.7%	33.1%
% radiographic spine OA	55.5%	76.1%
% radiographic hand OA	22.3%	52.1%
BMD hip (g/cm2)	0.77 (0.13)	0.75 (0.12)
BMD spine (g/cm2)	0.98 (0.17)	0.95 (0.15)
Total body fat (kg)	N/A	33.29 (18.03)
% body fat	N/A	41.8% (9.7%)
Colles #'s	69 (7%)	96 (12%)
Vertebral #'s	13 (1.2%)	52 (6.3%)
Total #'s	233 (23.2%)	407 (49%)
HAQ1	0.18 (0.36)	N/A
SF362	45.86 (25.89)	48.94 (28.40)
GHQ3	18.4 (9.72)	19.0 (10.79)

BMI= body mass index. HRT= hormone replacement use, self-reported. BMD= bone mineral density, HAQ=Health Assessment Questionnaire, GHQ=General Health Questionnaire, The psychological tests were collected at different time intervals ¹ Year 4 & Year 9; ² Year 3 & Year 9; ³ Year 5

Radiographic assessment and progression of disease.

Grading assessments of knee (baseline/5 & 10 years), spine (baseline and 9 years), hand (baseline & 11 years), hip (year 1 & year 8) and feet (year 6) are available. Radiographs were read by investigators, using an atlas of radiographic features with good reproducibility (Spector 1993, Hart 1994, Hassett 2003). A summary or global degeneration score of all affected joints is possible from the acquired data. We have used this method for calculating a total hand score (Spector TD, 1996). In addition, we have recently analysed data comparing progression at the hand, spine and knee showing the sites are closely correlated. Lumbar spine OA progression predicted 67% of progression of knee OA. Progression of JSN in knee OA was predicted by progression in lumbar spine disc space narrowing (OR 2.9, 95% CI 1.2 –7.5) and hip OA (JSN) (OR 2.0, 95% CI 1.0 – 4.2) predicting 41 and 49% of knee progression respectively. This study demonstrated that progression of knee OA can be predicted by studying progression of lumbar spine and hip OA. This is of particular importance as it suggests that global scores of total joint damage may be more informative for longitudinal studies. The progression data from this cohort is very unique. We have reproducibility of over 70% Kappas with less than 3% improvers.

Current OA Status of Cohort

In 1989 200 cases of radiographic OA were identified in the cohort. In the knee, 25% developed progressive disease, and 95 new cases of OA were identified after 4 years. There are now 150 progressors with knee OA, and 100 non-progressors. There are also 115 incident knee OA cases, with an incident rate of new knee OA of approximately 3% per year.

Joint Group		Non-progressors	Progressors
Knee	OP	11%	15%
	JSN	7%	8%
Hand	OP	8%	20%
	JSN	13%	24%
Hip	OP	15%	15%
	JSN	2%	2%
Spine	OP	58%	31%
	DSN	46%	19%

Bone mineral density and fracture data

At the baseline examination, as well as at year 10 follow-up, BMD was measured at the lumbar spine (L1–L4) and femoral neck region (gm/cm2) using dual x-ray absorptiometry (QDR 1000; Hologic, Waltham, MA cross-calibrated with a QDR 2000/w at follow-up). Reproducibility was tested on 20 paired subjects 1 week apart and found to be 1-2% for spine and 2-3% for the hip (Keen RW, 1998). Detailed data were sought on all fractures and falls occurring during the first 48 months of the study. Information was obtained between 1989 and 1993 by interview at each annual visit and by phone calls every 6 months to reduce error due to poor recall. A total of 145 validated fractures were observed during the first 48 months of the study, the majority were of the distal forearm (40, or 27.6%) and vertebrae (41, or 28.3%). The remaining fractures were of the hip (6, or 4.1%), lower limb and foot (30, or 20.7%), upper limb and hand (20, or 13.8%), and sternum (8, or 5.5%).

The women have a fracture/falls questionnaire administered by phone every 6 months, and at year 15 (present) 407 women have experienced a fracture. Not all fractures have yet been validated with hospital and GP records. The women now have average age of 65, there have been 96 colles fractures, and up to year 10, 52 vertebral fractures.

Bone loss at the spine and hip over 9 years of study

Major achievements

In 15 years, the Chingford Study has generated over 70 peer reviewed scientific publications on cross-sectional and longitudinal data (see scientific references). Of these publications 20 have been in journals with high impact ratings; A&R: 11, JBMR: 2, BMJ: 5, Lancet: 2

Examples of our major achievements in the last ten years include:

• Classification of OA and progression of disease. This study was the first to develop atlases of standard radiographs of individual features of knee, hip, spine, hand and foot used in epidemiological studies of OA worldwide (Burnett 1994).
• Aids design of trials. The longitudinal obesity paper from this study led to the design of the protocol for the multicentre NIH Doxycycline trial (Brandt et al A&R, 2005). Chingford natural history data is often used to help design phase II and III pharma studies.
• Role of inflammation in OA. In 1997 this was the first study to suggest that low-grade inflammation may be a significant aspect of early OA (Spector TD A&R, 1997)
• Links of OA and OP. Two major population studies clarified the relationship between these diseases. BMD was higher in OA but those with previous fractures were protected from OA, independent of BMD, suggesting a possible common role of bone turnover and repair in the early manifestations of OA (Hart (A&R 2002) (ARD 1994))
• Role of genes in OA progression. This study identified 9 novel genes whose genotype correlates with knee OA susceptibility and/or progression and could have potential diagnostic or therapeutic value by pointing to possible new disease mechanisms (Valdes A & R, 2004)
• Role of obesity in OA progression. An important study showing one third of middle aged women with unilateral disease will progress to bilateral knee OA within two years and a fifth will progress in the index joint (Spector ARD 1994)
• Family history of Colles fracture. This study showed a positive family history of wrist fracture was associated with a 4- fold risk of wrist fracture, independent of BMD (Keen Ost Int, 1999)
• Role of the vitamin D receptor gene in OA and OP. Study found a polymorphism of the VDR gene associated with an increased risk of knee OA. This study was the first genetic locus that was shown to influence the risk of early knee OA within the general population (Keen A & R 1997)
• Hypermobility unrelated to OA. Contrary to popular belief- this study showed joint hypermobility may be a marker for fitness, manifested by reduced knee OA and increased hip BMD (Dolan J Rheumatol, 2003)
• Importance of spine disc progression. This was the first population-based longitudinal study to assess progression of the individual radiographic features (4% per annum) and show risk factors for lumbar spine disc degeneration.(Hassett A & R, 2003)
• Importance of anxiety in falls. This study showed Fear of falling index was related to early reduction of mobility function rather than psychological factors. It allows identification of individuals at risk of subsequent functional decline (Martin, Age & ageing, 2005)
• Role of bone turnover in OA progression. This study demonstrated for the first time that bone resorption is increased in patients with progressive knee OA compared with those with nonprogressive knee OA (Bettica A & R, 2002)
• Classification of vertebral fractures for trials. Chingford normative data was used to develop a rapid semi-automated technique for assessing vertebral deformities on lateral spine radiographs. The method has become a world-wide standard in assessing vertebral deformity both in population studies and in clinical trials (McCloskey Ost Int, 1993)

Database contents

In these years, we have developed an extensive database covering both baseline variables, and 15 years of repeat assessments for a number of variables and additional assessments have been recorded.

1. Xrays
   

   Spine: Lateral lumbar and thoracic spine radiographs from baseline, year 3 and year 9 scored for vertebral fractures using the algorithm developed by Eugene McCloskey, Sheffield University. Assessment of osteopenia, and spinal osteophytes, disc space narrowing and aortic calcification on all women also recorded at both timepoints.
   

   Knees: Weightbearing AP and skyline views graded for osteoarthritis using Kellgren & Lawrence and individual features of osteophytes and joint space narrowing, and continuous measurement of joint space by magnified reticule. Available at baseline, year 3, year 5, year 10 and currently being collected at year 15.
   

   Pelvis: Films graded for hip joint osteophytes, joint space, and other features including singh index at baseline, year 4 and year 9.
   

   Hands: Graded for Kellgren & Lawrence, individual features of OA at baseline and year 3, and year 11.
   

   Feet: Films taken at year 6 and graded for MTP osteophytes and narrowing and angle degree of hallux valgus.

2. DXA
   

   Annual DXA scan of lumbar spine and hip using Hologic QDR 1000/2000 and calculated bone loss on all women, years 1 -10, whole body DXA at year 8 and year 9. Central fat analysis from DXA available on all women at year 9 & year 15.

3. MRI
   

   Sagittal, coronal T1, FOV 15, Long TE STIR & axial MRI sequences of the knee in 100 knee OA progressors and 100 non-progressors.

4. Fractures/Falls
   

   Detailed fracture history on prevalent and incident fractures. Six month fracture questionnaire completed for years 8-15 by phone. All prevalent and incident fractures validated with hospital and GP records. Detailed questionnaire on falls and non-accidental fracture years 6-15.
   

   Computerised measurement of postural instability using a Balance Performance Monitor.

5. Examination
   

   Annual anthropomorphic measures (weight, height, blood pressure, pulse, quads measurements, body circumference measures using standardised techniques).Clinical examination of signs of osteoarthritis of the hands and knees, and detailed symptom questionnaires.

6. Medical History
   

   Annual detailed medical and social history, gynaecological variables, drug, surgical, major illnesses, and family history.

7. Biological Samples
   

   Annual stored serum and urine on each subject up to 2.5ml in 5ml containers stored at -20° C (years 1,2,3,4,5,6,7,8,9,10, 11 & 15)
   

   DNA extracted on all women and stored at 3 timepoints (>200µg)
   

   Blood glucose, serum cholesterol, HDL, uric acid, serum calcium and routine biochemistry at years 3, 6, 9. C-reactive protein, E2,SHBG,DHEAS, IGG's, rheumatoid factor at baseline
   

   Collagen cross-links (CTX-I & CTX-II) (markers of bone and cartilage)
   

   Information (SNPs and microsatellites) on over 30 OA & OP candidate genes

8. Other data
   

   Standardised physical activity assessment from 20s and middle-age.
   

   Life employment patterns and levels of activity related to occupation.
   

   Repeat measures of standardised assessment of health status and disability (GHQ,SF36,HAQ).
   

   Pain assessment
   

   Footwear assessment
   

   Questionnaire on alternative remedies
   

   Soft tissue injury questionnaires & assessment.
   

   Social class
   

   Mobility
   

   GP records available for prescriptions, attendances and co-morbidity etc.
   

If you are interested in using the Chingford cohort in collaborative work please contact Victoria Vazquez, PA to Prof. Tim Spector.