Original article
Improved detection of hereditary haemochromatosis Catherine Ogilvie,1 Dairena Gaffney,2 Heather Murray,3 Andrew Kerry,4 Caroline Haig,3 Richard Spooner,2 Edward J Fitzsimons1 1
Department of Haematology, West Glasgow Hospitals University NHS Trust, Glasgow, UK 2 Department of Biochemistry, Glasgow Royal Infirmary, Glasgow, UK 3 Robertson Centre for Biostatistics, University of Glasgow, Glasgow, UK 4 Department of Clinical Biochemistry, Royal Alexandra Hospital, Paisley, UK Correspondence to Dr Catherine Ogilvie, Department of Haematology, West Glasgow Hospitals University NHS Trust, 21 Shelley Road, Glasgow G12 0XL, UK; [email protected] Received 13 October 2014 Revised 19 November 2014 Accepted 21 November 2014 Published Online First 24 December 2014
ABSTRACT Aims There is high prevalence of hereditary haemochromatosis (HH) in North European populations, yet the diagnosis is often delayed or missed in primary care. Primary care physicians frequently request serum ferritin (SF) estimation but appear uncertain as how to investigate patients with raised SF values. Our aim was to develop a laboratory algorithm with high predictive value for the diagnosis of HH in patients from primary care with raised SF values. Methods Transferrin saturation (Tsat) was measured on SF samples sent from primary care; 1657 male and 2077 female patients age ≥30 years with SF ≥200 μg/L. HFE genotyping was performed on all 878 male and 867 female patients with Tsat >30%. Results This study identified 402 (206 men; 196 women) C282Y carriers and 132 (58 men; 74 women) C282Y homozygotes. Optimal limits for combined SF and Tsat values for HH recognition were established. The detection rate for homozygous C282Y HH for male patients with both SF ≥300 μg/L and Tsat >50% was 18.8% (52/272) and 16.3% (68/415) for female patients with both SF ≥200 μg/L and Tsat >40%. Conclusions The large number of SF requests received from primary care should be used as a resource to improve the diagnosis of HH in areas of high prevalence.
INTRODUCTION
To cite: Ogilvie C, Gaffney D, Murray H, et al. J Clin Pathol 2015;68: 218–221. 218
Hereditary haemochromatosis (HH) is an autosomalrecessive genetic disorder of iron metabolism. It has been recognised as a clinical disorder for >100 years1 2 and is one of the most common inherited conditions in populations of North European descent.3–5 Clinical complications include liver cirrhosis, cardiomyopathy and diabetes. Treatment is simple and effective, but diagnosis is often either delayed or missed.6 HH is typically associated with the C282Y mutation in the HFE gene.7–9 Large population studies have determined global distribution of the C282Y mutation with highest prevalence (0.5% C282Y homozygotes) in North European populations.10–12 The H63D mutation of HFE is found in 20% of the UK population and can be associated with HH when inherited in combination with C282Y. Clinical iron overload however is less severe and far less frequent.13 Rarer mutations causing HH have been described. These include mutations in hemojuvelin, hepcidin (HAMP) genes, transferrin receptor 2 (TRF2), ferroportin and rare mutations in HFE.14–18 In the absence of C282Y homozygosity or C282Y/H63D compound heterozygosity, further investigation of raised serum ferritin (SF) and transferrin saturation (Tsat) can include screening for these less common mutations. Clinical penetrance as defined by biochemical evidence of iron overload is not complete in
C282Y homozygotes. Estimates vary considerably, but it is found more commonly in male than in female homozygotes. In a study of Danish male C282Y homozygotes, 89% had Tsat ≥50% and 94% SF ≥300 μg/L.10 Other studies suggest that 38–50% of C282Y homozygotes may develop iron overload.19 Estimates of clinical penetrance as defined by iron overload-related morbidity also vary considerably from 10% to 30%19 of men and much less frequently in women.20 In health, SF is a reliable marker of reticuloendothelial iron stores and is highly sensitive to iron overload in HH.21 22 It is the most frequently measured haematinic in the UK, and since its general introduction in the 1980s, SF has replaced the measurements of serum iron and transferrin in primary care.23 Primary care physicians are familiar with the investigation and treatment of low SF values and iron deficiency but are much less comfortable with investigating patients with high SF values even when these values are grossly elevated.24 Transferrin is responsible for iron transport and raised Tsat reflects the potential for parenchymal iron loading and tissue damage.25 The liver, heart, pancreas, pituitary gland, joints and skin are most susceptible to tissue damage. In the liver, iron deposition initially causes a non-specific inflammation that can then progress to cirrhosis, particularly when SF >1000 μg/L.26 Timely diagnosis and treatment with simple venesection prevents organ damage.27–30 HH is however poorly recognised in primary care.31 32 The aim of this study was to develop a hospital laboratory pathway to improve the diagnosis of HH in primary care.
METHODS All SF requests were initiated and taken in primary care. Serum samples from patients ≥30 years with SF ≥200 μg/L underwent measurement of iron and Tsat. HFE genotyping was carried out on all patients with Tsat >30% using the full blood count (FBC) samples that accompanied the SF samples.
Study population The primary care population is predominantly a ‘well patient’ population, minimising the effect of secondary care disorders on SF values. This population is not entirely unselected as these patients had chosen to consult their general practitioner (GP). The purpose of this study however was to develop a laboratory pathway to assist the diagnosis of HH that might be clinically relevant rather than to screen the general population. Only patients >30 years were included. Iron deficiency due to menstruation or pregnancies may mask HH in young women and clinically relevant HH is unlikely
Ogilvie C, et al. J Clin Pathol 2015;68:218–221. doi:10.1136/jclinpath-2014-202720
Original article to develop before 30 years in men or women due to high iron requirements during growth.33 Large population studies have shown that 90% male and 55% female C282Y homozygotes would have SF values ≥200 μg/L. As such, SF value 50% were seen in 91.9% and 77% of homozygotes, respectively, vs 43.8% and 19.3% of nonhomozygotes ( p300 μg/L. The combination of SF >200 μg/L and Tsat >40% was found in 415 women, which included 92% of C282Y homozygotes.
RESULTS For analysis, 3734 samples with SF values ≥200 μg/L were referred for Tsat measurement. In total, 1745 (46.9%) patients ≥30 years with SF ≥200 μg/L were found to have Tsat >30%. HFE genotype was determined on these patients (table 1). Male patients n=878, mean age 61 years. As shown, 91.9% had SF ≥300 μg/L, 54.2% had Tsat >40% and 32.4% had Tsat >50%. Female patients n=867, mean age 65 years. As shown, 44.1% had SF 200–300 μg/L, 55.9% had SF ≥300 μg/L, 52.1% had
Table 1 Summary of characteristics for all patients genotyped (≥30 years, SF ≥200 μg/L, Tsat >30%) by gender Characteristic
Statistic
Age (years)
30–49 50–69 ≥70 C282Y/C282Y C282Y/+ C282Y/H63D No C282Y allele detected Median (IQR) 200–299 300–499 500–1000 >1000 Median (IQR) >30–40 >40–50 >50
HFE HH C282Y carriers Compound heterozygotes Non-C282Y genotypes SF (mg/L)
Tsat (%)
Men (n=878)
Women (n=867)
212 416 250 58 151 55 614
(24.1%) (47.4%) (28.5%) (6.6%) (17.2%) (6.3%) (69.9%)
165 375 327 74 150 48 595
(19.0%) (43.3%) (37.7%) (8.5%) (17.3%) (5.5%) (68.6%)
456 71 441 256 110 42 402 192 284
(346–639) (8.1%) (50.2%) (29.2%) (12.5%) (35–56) (45.8%) (21.9%) (32.4%)
322 382 257 158 70 39 452 205 210
(240–518) (44.1%) (29.6%) (18.2%) (8.1%) (34–50) (52.1%) (23.6%) (24.2%)
HH, hereditary haemochromatosis; SF, serum ferritin; Tsat, transferrin saturation.
Ogilvie C, et al. J Clin Pathol 2015;68:218–221. doi:10.1136/jclinpath-2014-202720
C282Y carriers and C282Y/H63D compound heterozygotes All patients found to be carriers of C282Y were screened for the H63D mutation. In the male patient group, 206 (23%) C282Y carriers were detected; 55 compound heterozygotes and 151 C282Y carriers only. There was no significant difference in median SF and Tsat between male compound heterozygotes and simple C282Y carriers. In the female patient group, 198 (23%) C282Y carriers were detected with 48 compound heterozygotes and 150 C282Y carriers only. There was no significant difference in median SF between female compound heterozygotes and simple C282Y carriers. Tsat was slightly higher in the compound heterozygote group; 44.7% vs 40.2%; Wilcoxon p value = 0.012.
Statistical analysis and proposed laboratory algorithm Logistic regression analysis confirmed that C282Y homozygosity in men was strongly associated with Tsat >50% compared with Tsat 30–50% (OR (95% CI) 22.0 (9.3 to 51.9), p40% compared with Tsat 30–40% (OR (95% CI) 4.21 (1.54 to 11.6), p40% (women) or >50% (men), samples should be referred for HFE genotyping. The above algorithm was tested using binomial tests of the null hypothesis (table 3). For both men and women, the 219
Original article Table 2 Comparison of patient characteristics; C282Y homozygotes versus non-homozygotes Characteristic
Statistic
Male C282Y/C282Y (n=58)
Male non- C282Y/C282Y (n=820)
Female C282Y/C282Y (n=74)
Female non-C282Y/C282Y (n=793)
Age (years)
Mean (SD)
58.6 (15.0)
61.2 (13.4)
Serum ferritin (mg/L)
30–49 50–69 ≥70 Median (IQR)
18 (31.0%) 25 (43.1%) 15 (25.9%) 724.5 (422–1510)
200–300 300–500 500–1000 >1000
1 (1.7%) 18 (31.0%) 18 (31.0%) 21 (36.2%)
64.9 (15.1) p=0.041* 149 (18.8%) 337 (42.5%) 307 (38.7%) 315.0 (238–500) p=0.0008† 361 (45.5%) 231 (29.2%) 139 (17.5%) 62 (7.8%)
Median (IQR)
85 (65–93)
>30–40 >40–50 >50
4 (6.9%) 2 (3.4%) 52 (89.7%)
61.0 (14.0) p=0.22* 194 (23.7%) 391 (47.7%) 235 (28.7%) 443.0 (344–626) p40% and >50%, respectively. Therefore, only 2.12% (10.6%×20%) of female and 6.02% (17.2%×35%) of male SF requests made from primary care on patients >30 years would generate HFE genotyping with detection rates of 16.4% and 18.8% for C282Y homozygous HH. The clinical impact from implementing this algorithm is less easy to define. It is likely to be influenced by the threshold values of SF and Tsat at which individual clinicians would recommend venesection. These attitudes might also be influenced by factors such as patient age and comorbidities. The European Association for the Study of Liver guidelines however recommend target value of SF 200 μg/L and Tsat >30%. In summary, the laboratory algorithm developed in this study would link a GP-initiated SF request to a sequential reflex laboratory pathway that delivers a high percentage yield of patients with C282Y homozygous HH. This pathway would use samples that are currently in laboratories but that are simply discarded after FBC and SF measurements. The introduction of this algorithm in areas with a high prevalence of HH would help to overcome low awareness of HH in primary care and remove the risk of failure to act on a raised SF result.
Table 3 Statistical analysis of the proposed algorithm for the detection of C282Y homozygous hereditary haemochromatosis
Women (SF≥200, Tsat >40%) Men (SF ≥300, Tsat >50%)
C282Y/C282Y
Non-C282Y/C282Y
p Value (Binomial test)
68 (16.4%) (95% CI 13.0% to 20.3%) 51 (18.8%) (95% CI 14.3% to 24.9%)
347 (83.6%) 221 (81.3%)
Improved detection of hereditary haemochromatosis Catherine Ogilvie,1 Dairena Gaffney,2 Heather Murray,3 Andrew Kerry,4 Caroline Haig,3 Richard Spooner,2 Edward J Fitzsimons1 1
Department of Haematology, West Glasgow Hospitals University NHS Trust, Glasgow, UK 2 Department of Biochemistry, Glasgow Royal Infirmary, Glasgow, UK 3 Robertson Centre for Biostatistics, University of Glasgow, Glasgow, UK 4 Department of Clinical Biochemistry, Royal Alexandra Hospital, Paisley, UK Correspondence to Dr Catherine Ogilvie, Department of Haematology, West Glasgow Hospitals University NHS Trust, 21 Shelley Road, Glasgow G12 0XL, UK; [email protected] Received 13 October 2014 Revised 19 November 2014 Accepted 21 November 2014 Published Online First 24 December 2014
ABSTRACT Aims There is high prevalence of hereditary haemochromatosis (HH) in North European populations, yet the diagnosis is often delayed or missed in primary care. Primary care physicians frequently request serum ferritin (SF) estimation but appear uncertain as how to investigate patients with raised SF values. Our aim was to develop a laboratory algorithm with high predictive value for the diagnosis of HH in patients from primary care with raised SF values. Methods Transferrin saturation (Tsat) was measured on SF samples sent from primary care; 1657 male and 2077 female patients age ≥30 years with SF ≥200 μg/L. HFE genotyping was performed on all 878 male and 867 female patients with Tsat >30%. Results This study identified 402 (206 men; 196 women) C282Y carriers and 132 (58 men; 74 women) C282Y homozygotes. Optimal limits for combined SF and Tsat values for HH recognition were established. The detection rate for homozygous C282Y HH for male patients with both SF ≥300 μg/L and Tsat >50% was 18.8% (52/272) and 16.3% (68/415) for female patients with both SF ≥200 μg/L and Tsat >40%. Conclusions The large number of SF requests received from primary care should be used as a resource to improve the diagnosis of HH in areas of high prevalence.
INTRODUCTION
To cite: Ogilvie C, Gaffney D, Murray H, et al. J Clin Pathol 2015;68: 218–221. 218
Hereditary haemochromatosis (HH) is an autosomalrecessive genetic disorder of iron metabolism. It has been recognised as a clinical disorder for >100 years1 2 and is one of the most common inherited conditions in populations of North European descent.3–5 Clinical complications include liver cirrhosis, cardiomyopathy and diabetes. Treatment is simple and effective, but diagnosis is often either delayed or missed.6 HH is typically associated with the C282Y mutation in the HFE gene.7–9 Large population studies have determined global distribution of the C282Y mutation with highest prevalence (0.5% C282Y homozygotes) in North European populations.10–12 The H63D mutation of HFE is found in 20% of the UK population and can be associated with HH when inherited in combination with C282Y. Clinical iron overload however is less severe and far less frequent.13 Rarer mutations causing HH have been described. These include mutations in hemojuvelin, hepcidin (HAMP) genes, transferrin receptor 2 (TRF2), ferroportin and rare mutations in HFE.14–18 In the absence of C282Y homozygosity or C282Y/H63D compound heterozygosity, further investigation of raised serum ferritin (SF) and transferrin saturation (Tsat) can include screening for these less common mutations. Clinical penetrance as defined by biochemical evidence of iron overload is not complete in
C282Y homozygotes. Estimates vary considerably, but it is found more commonly in male than in female homozygotes. In a study of Danish male C282Y homozygotes, 89% had Tsat ≥50% and 94% SF ≥300 μg/L.10 Other studies suggest that 38–50% of C282Y homozygotes may develop iron overload.19 Estimates of clinical penetrance as defined by iron overload-related morbidity also vary considerably from 10% to 30%19 of men and much less frequently in women.20 In health, SF is a reliable marker of reticuloendothelial iron stores and is highly sensitive to iron overload in HH.21 22 It is the most frequently measured haematinic in the UK, and since its general introduction in the 1980s, SF has replaced the measurements of serum iron and transferrin in primary care.23 Primary care physicians are familiar with the investigation and treatment of low SF values and iron deficiency but are much less comfortable with investigating patients with high SF values even when these values are grossly elevated.24 Transferrin is responsible for iron transport and raised Tsat reflects the potential for parenchymal iron loading and tissue damage.25 The liver, heart, pancreas, pituitary gland, joints and skin are most susceptible to tissue damage. In the liver, iron deposition initially causes a non-specific inflammation that can then progress to cirrhosis, particularly when SF >1000 μg/L.26 Timely diagnosis and treatment with simple venesection prevents organ damage.27–30 HH is however poorly recognised in primary care.31 32 The aim of this study was to develop a hospital laboratory pathway to improve the diagnosis of HH in primary care.
METHODS All SF requests were initiated and taken in primary care. Serum samples from patients ≥30 years with SF ≥200 μg/L underwent measurement of iron and Tsat. HFE genotyping was carried out on all patients with Tsat >30% using the full blood count (FBC) samples that accompanied the SF samples.
Study population The primary care population is predominantly a ‘well patient’ population, minimising the effect of secondary care disorders on SF values. This population is not entirely unselected as these patients had chosen to consult their general practitioner (GP). The purpose of this study however was to develop a laboratory pathway to assist the diagnosis of HH that might be clinically relevant rather than to screen the general population. Only patients >30 years were included. Iron deficiency due to menstruation or pregnancies may mask HH in young women and clinically relevant HH is unlikely
Ogilvie C, et al. J Clin Pathol 2015;68:218–221. doi:10.1136/jclinpath-2014-202720
Original article to develop before 30 years in men or women due to high iron requirements during growth.33 Large population studies have shown that 90% male and 55% female C282Y homozygotes would have SF values ≥200 μg/L. As such, SF value 50% were seen in 91.9% and 77% of homozygotes, respectively, vs 43.8% and 19.3% of nonhomozygotes ( p300 μg/L. The combination of SF >200 μg/L and Tsat >40% was found in 415 women, which included 92% of C282Y homozygotes.
RESULTS For analysis, 3734 samples with SF values ≥200 μg/L were referred for Tsat measurement. In total, 1745 (46.9%) patients ≥30 years with SF ≥200 μg/L were found to have Tsat >30%. HFE genotype was determined on these patients (table 1). Male patients n=878, mean age 61 years. As shown, 91.9% had SF ≥300 μg/L, 54.2% had Tsat >40% and 32.4% had Tsat >50%. Female patients n=867, mean age 65 years. As shown, 44.1% had SF 200–300 μg/L, 55.9% had SF ≥300 μg/L, 52.1% had
Table 1 Summary of characteristics for all patients genotyped (≥30 years, SF ≥200 μg/L, Tsat >30%) by gender Characteristic
Statistic
Age (years)
30–49 50–69 ≥70 C282Y/C282Y C282Y/+ C282Y/H63D No C282Y allele detected Median (IQR) 200–299 300–499 500–1000 >1000 Median (IQR) >30–40 >40–50 >50
HFE HH C282Y carriers Compound heterozygotes Non-C282Y genotypes SF (mg/L)
Tsat (%)
Men (n=878)
Women (n=867)
212 416 250 58 151 55 614
(24.1%) (47.4%) (28.5%) (6.6%) (17.2%) (6.3%) (69.9%)
165 375 327 74 150 48 595
(19.0%) (43.3%) (37.7%) (8.5%) (17.3%) (5.5%) (68.6%)
456 71 441 256 110 42 402 192 284
(346–639) (8.1%) (50.2%) (29.2%) (12.5%) (35–56) (45.8%) (21.9%) (32.4%)
322 382 257 158 70 39 452 205 210
(240–518) (44.1%) (29.6%) (18.2%) (8.1%) (34–50) (52.1%) (23.6%) (24.2%)
HH, hereditary haemochromatosis; SF, serum ferritin; Tsat, transferrin saturation.
Ogilvie C, et al. J Clin Pathol 2015;68:218–221. doi:10.1136/jclinpath-2014-202720
C282Y carriers and C282Y/H63D compound heterozygotes All patients found to be carriers of C282Y were screened for the H63D mutation. In the male patient group, 206 (23%) C282Y carriers were detected; 55 compound heterozygotes and 151 C282Y carriers only. There was no significant difference in median SF and Tsat between male compound heterozygotes and simple C282Y carriers. In the female patient group, 198 (23%) C282Y carriers were detected with 48 compound heterozygotes and 150 C282Y carriers only. There was no significant difference in median SF between female compound heterozygotes and simple C282Y carriers. Tsat was slightly higher in the compound heterozygote group; 44.7% vs 40.2%; Wilcoxon p value = 0.012.
Statistical analysis and proposed laboratory algorithm Logistic regression analysis confirmed that C282Y homozygosity in men was strongly associated with Tsat >50% compared with Tsat 30–50% (OR (95% CI) 22.0 (9.3 to 51.9), p40% compared with Tsat 30–40% (OR (95% CI) 4.21 (1.54 to 11.6), p40% (women) or >50% (men), samples should be referred for HFE genotyping. The above algorithm was tested using binomial tests of the null hypothesis (table 3). For both men and women, the 219
Original article Table 2 Comparison of patient characteristics; C282Y homozygotes versus non-homozygotes Characteristic
Statistic
Male C282Y/C282Y (n=58)
Male non- C282Y/C282Y (n=820)
Female C282Y/C282Y (n=74)
Female non-C282Y/C282Y (n=793)
Age (years)
Mean (SD)
58.6 (15.0)
61.2 (13.4)
Serum ferritin (mg/L)
30–49 50–69 ≥70 Median (IQR)
18 (31.0%) 25 (43.1%) 15 (25.9%) 724.5 (422–1510)
200–300 300–500 500–1000 >1000
1 (1.7%) 18 (31.0%) 18 (31.0%) 21 (36.2%)
64.9 (15.1) p=0.041* 149 (18.8%) 337 (42.5%) 307 (38.7%) 315.0 (238–500) p=0.0008† 361 (45.5%) 231 (29.2%) 139 (17.5%) 62 (7.8%)
Median (IQR)
85 (65–93)
>30–40 >40–50 >50
4 (6.9%) 2 (3.4%) 52 (89.7%)
61.0 (14.0) p=0.22* 194 (23.7%) 391 (47.7%) 235 (28.7%) 443.0 (344–626) p40% and >50%, respectively. Therefore, only 2.12% (10.6%×20%) of female and 6.02% (17.2%×35%) of male SF requests made from primary care on patients >30 years would generate HFE genotyping with detection rates of 16.4% and 18.8% for C282Y homozygous HH. The clinical impact from implementing this algorithm is less easy to define. It is likely to be influenced by the threshold values of SF and Tsat at which individual clinicians would recommend venesection. These attitudes might also be influenced by factors such as patient age and comorbidities. The European Association for the Study of Liver guidelines however recommend target value of SF 200 μg/L and Tsat >30%. In summary, the laboratory algorithm developed in this study would link a GP-initiated SF request to a sequential reflex laboratory pathway that delivers a high percentage yield of patients with C282Y homozygous HH. This pathway would use samples that are currently in laboratories but that are simply discarded after FBC and SF measurements. The introduction of this algorithm in areas with a high prevalence of HH would help to overcome low awareness of HH in primary care and remove the risk of failure to act on a raised SF result.
Table 3 Statistical analysis of the proposed algorithm for the detection of C282Y homozygous hereditary haemochromatosis
Women (SF≥200, Tsat >40%) Men (SF ≥300, Tsat >50%)
C282Y/C282Y
Non-C282Y/C282Y
p Value (Binomial test)
68 (16.4%) (95% CI 13.0% to 20.3%) 51 (18.8%) (95% CI 14.3% to 24.9%)
347 (83.6%) 221 (81.3%)
