Iron stores and coagulation parameters in patients with hypoxem ' c polycythemia secondary to chronic obstructive pulmonary disease : The effect of phlebotomies

This study was designed to determine the effects of phlebotomy on iron body contents and coagulation tests of COPO patients with po'ycythemia secondary to hypoxemia. Seventeen patients with COPO and hematocrits higher than 54 percent (mean Hct: 57 :t 0.49 percent ), who had not received anti-inflammatory or antiplatelet aggregation agents recently.Their mean forced expiratory volume at 1 second (FEV1) was 0.92:t 0.11 L. Intervention: Blood work was collected to evaluate the following: serum iron and ferritin leveis, total iron binding capacity, transferrin saturation index, fibrinogen plasma leveis, activated partial thromboplastin time, platelet count, platelet aggregation measurements, and thromboelastography coagulation parameters. The blood samples were obtained before and about 7 days after the hematocrit correction by 300-400 ml phlebotomies done every other day.The mean number of phlebotomies done for each patient was 4.4. Postphlebotomy iron serum leveis decreased from 90.1 :t 14.8 to 59.7:t 9.9 mg/dl and the ferritin serum leveis from 133.8 :t 37.9 to 70.8 :t 32.7 ng/ml (p< 0.05). Regarding the coagulation studies, there were significant increases in the platelet count, from 227,300 :t 13,900 to 312,500 :t 30,200 per mm3, and in the maximum clot amplitude (a) obtained by thromboelastography ( from 53.6 :t 1.4 percent to 60.4 :t 1.1 percent). The coagulation time (k) of the thromboelastography also decreased significantly, from 7.5 :t 0.7mm prephlebotomy to 4.5 :t 0.3mm postphlebotomy. Although the coagulation changes were small amount, the observed significant decrease in iron contents may have important clinicai implications.


INTRODUCTION
B loodlettings, or phlebotomies, are still adyocated in the care of chronic obstructive pulmonary disease (COPD) patients with cor pulmonale and severe secondary polycythemia.1 Several authors have shown that hematocrit (Hct) reductions in hypoxemic patients with COPD and polycythemia are followed by improvement in general symptoms,2-S mental activity2-6 and exercise performance.2 -4 .7 . 8 In a group of patients performing exercise studies in ergometer, the maximum workload and oxygen consumption have increased 31 and 16 percent respectively, when the mean Htc had decreased from 62 to 50 percent,x These effects are not secondary to changes in respiratory mechanics or blood gases. 2 .4.S.7-10 Rather, it seems they result from improvements in hemodynamic conditions related to a better blood viscosity.3.11-13 Although a great deal has been published in this field, no attention has been paid to the potential hematological consequences of phlebotomies.Similarly to what happens with blood donors and during the treatment of hemochromatosis, repeated phlebotomies in COPD patients could lead to reductions in their iron body contents.14  helTIoglobin leveIs, and disturbances in muscular oxidative metaboIism.1ó In the case of reductions in iron body contents actualIy deveIoping in COPD patients folIowing phlebotomies, the former could act as limiting factors for expected exercise improvements.
In addition, polycythemia is a welI known risk factor for thrombosis,17 and lnany studies.suggest that COPD patients have a higher tendency to have thromboembolic episodes.IX -20 SOlne studies of thrombotic risk in COPD have shown an increased tendency towards coagulation with altered values in the throlnboelastogram, hyperfibrinogemia and increased platelet aggregation.1 x'21-23 Therefore, we hypothesized that a reduction in Hct leveIs by phlebotolnies could aIso have a favorable effect on decreasing s'uch thrombotic risk.
The present study investigates the effects of the Hct correction by repeated phlebotomies on paralneters of hemostasis and body iron contents in COPD patients with severe secondary hypoxemic polycythemia.

Patient Selection
Seventeen with COPD and polycythelnia ( Hct 54 percent or more) patients were studied after signing an informed consent formo All patients were in stable respiratory conditions, without clinicaI evidence of right or left ventricular faiIure.They had not been under any antiplatelet or anti-inflammatory agent for at least the prior 10 days, and.during the study period there were no changes in their daily medications.None of the patients showed clinicaI evidence ofblood loss or were using oxygen at that time.
Following an initiallaboratory evaIuation, phlebotomies of 300 to 400 ml were perfonned every other day until a venous Hct value around 45 percent was reached (range: 41-48 percent).A second set of laboratory tests was then performed about 7 days after the last phlebotomy 'session.

Iron Stores Measurements
Venous blood samples with no anticoagulant were collected from fasting patients in the Inorning.Iron seruln leveIs and seruln total iron binding capacity (TIBC) were measured using a commercialIy available kit (Iron Roche@) with readings made with a Colelnan 6/20 spectrophotometer.
Ferritin serum leveIs were also evaluated by a commercially available enzymatic immunoassay (FerrizYlne TM@-Abbot Laboratories).The transferrin saturation index (TSI) was calculated as the percentile ratio between serum iron leveIs and TIBC.

Coagulation Studies
Coagulation tests were performed on 3.8 percent sodium citrate venous blood samples collected simuItaneously with the iron blood work.The following determinations were made: -Activated paltial thromboplastin time (APTT) using human cephalin.24 The results were expressed as the ratio to the control values for the Iaboratory on that day.
-Platelet aggregation determinations by the Born lnethod in a Chronolog aggregometer.2ó The aggregation agents ADP in concentrations of 1 mM and 3 mM, epinephrine at 1:1000 dilution, and hUlnan colIagen were used in salnples froln each patient.The percentage of aggregated pIatelets was caIculated at a curve point corresponding to an interval of 5 minutes from the initial.stimulus for each aggregation agent.
-Throlnboelastography (TEG), using a Hartert throlnboelastograph lnade by HelIige.This test exalnines the entire bloocl coaguIation process and the interaction of alI involvecl cOlnponents.27 The reaction tilne r, the coagulation time k, and the lnaxil11um cIot alnpIitude a , were lneasured using a special ruIer on the developed filmo

Statistical Methods
All data are expressed as the mean :t standard error.The WiIcoxon test was used in the statistical analysis for the nonindependent variables.A. leveI of 5 percent was fixed as the litnit for rejection of the null hypothesis.

RESULTS
Thirteen male and four female patients were studiecl.Most of theln had severe or 1110derate COPD with arterial hypoxemia (Table 1).The mean Hct fell frol11 57 percent before phebotolnies to 45 percent at the time of the final

Coagulation
studies obtained before and after phlebotomies are listed in Table 3. Ali the initial Inean values were within the norlnal limits' of the authors' laboratory except for the r paralneter of the TEG, vvhich was slightly increased.No significant decreases were seen in the Inean APTT and fibrinogen leveIs after phlebotolnies.The platelet count significantly increased fronl 227,3001Inln 3 to 312,000/mm 3 after the procedure.Fifteen patients (88.2).Serum ferritin is a good indicator of iron body stores, and the drop in its leveIs denotes that the tissue iron contents 28 ,29  were also affected.Iron deficiency states may have effects on red blood cells and solid organ tissues.Although it would be expected that iron deficiency could have an inhibitory effect on COPD hypoxic hemopoietic response, this is not always the case.Secondary polycythemia associated to iron deficiency in COPD patients has been described. 30,31  Some authors report that iro,n depletion may not be an impeding factor to increases in red cell mass, and that in such cases polycythemia is associated with hypochromic erythrocytes. 32 An iron deficiency state may also metabolism, interfering with the patient's work performance.16 The tissue haem iron compounds include the cytoçhromes, myoglobin, catalase and peroxidase.In additiQD, muscles harbor non-haem iron-containing enzymes such as NADH dehydrogenase, succinic dehydrogenase and xanthine oxidase.Finally, some enzymes such as aconitase and tryptophan pyrrolase require iron or haem as a cofactor.
Experimental studies on rats have found that mild and severe iron deficiency are associated with deficits 9f cytochrome c and mioglobin in skeletal muscle. 33FINCH et a1. 34completed studies of work performance on irondeficient rats, keeping hemoglobin leveIs constant through exchange transfusions.It was seen that iron-deficient rats had a significant shorter running time than çontrols with similar hemoglobin leveIs.After 4 days of iron administration, the iron-deficient group achieved normal running times.
Biochemical analysis done at that time suggested that an increase in the rate of oxidative phosphorylation with a-glycerophosphate as a substrate played a key role in the improvement of work performance.Non-published data from our laboratory have shown small but significant increases in the maximal work load and oxygen consumption reached by 15 iron-deficient, anemic wom' 4 days after administration ofIV iron-dextran, ev~n though their hemoglobin leveIs had not changed.
Therefore, we can hypothesize that COPD patients who show ameIioration of exercise performance after phIebotomies lnight have this effect attenuated if simuItaneous iron deficiency states deveIop.The DISCUSSION platelet count, 4 of them with final measurements beyond 400,OOO/mm. 3Studies of platelet aggregation made with different stimulant agents did not show any significant postphlebotomies change.Adequate thromboelastography studies before and after phIebotomies could not be obtained in 4 patients due to technical problems.
In the 13 studied subjects, a significant decrease was noticed in k during the final studies.Although r also experienced a decrease, such change did not reach statistical significance.Finally, a values showed a significant postphlebotomies increase.
Phlebotomy may be used in the care of COPD patients in an attempt to reduce detrimental cardiovascular effects of blood hyperviscosity due to polycythemia.The basis for this procedure is the effect of phIebotomy in improving hemodynamic conditions and exercise performance in such patients. 2-4 ,7,8,11-13 However, the effects of this kind of therapy on hemostasis and iron reserves of COPD patients had not been thoroughIy examined before.
Our present resuIts showed that the normalization of hematocrits by phlebotomies led to an important reduction in iron body contents.This became apparent by the significant decreases in serum iron, ferritin and TSI (Table São Paulo   development of hypochromic red ceIls in such situations could impair oxygen transport to the tissues, even with the presence of a high hematocrit.Furthermore, iron depletion at the skeletal muscle leveI could lead to poor enzymatic function and oxygen utilization. In order to evaluate the effects of Het correction by phlebotomies on the hemostasis of COPD polycythemic patients, tests were done to assess the intrinsic pathway, platelet and fibrinogen concentrations, platelet aggregation and TEG.A renewed interest in TEG monitoring has been seen recently because its tracing represents the shear elasticity of a blood clot as it forms, matures, retracts and eventuaIIy lyses.This test is extremely sensitive to identify hypercoagulabi lity.27 The initial evaluation prephlebotomy did not show any evidence of hypercoagulability, since most of the test resuIts were within the normal range for our coagulation laboratory (Table 3).The only exception was the reaction time r, which represents the time required for initiaI fibrin fornlation, which was slightly increased.This last finding indeed is lnore compatible with a hypocoagulability condition.
The mean platelet count increased significantly after phlebotomy from 227,300 to 312,500 mm 3 (Table 3).Increases in platelet number have been reported after chronic and acute blood 10ss.35.36The lnechanisms involved in such rises may be related to iron depletion, since an inhibitory effect of this metal in throbopoiesis has been described.35 Other studies suggest the presence of an humoral factor, the thrombopoietin, in plasma from animaIs with posthemorrhagic thrombocytosis.37 In opposition to previous reports on COPD patients, we did not observe an increase in platelet aggregation after phlebotomy.'XJX-40Thosereports attempted to relate their findings to increases in circulating cathecolamines, especial1y serotonin, secondary to the bleeding.The differences between present and former results lnay be explained by the fact that our blood samples were collected later, around one week after the Iast phlebotomy, when normalization of cathecolamine leveIs probably had already occurred.
Other significant postphlebotomy changes were an extension of a and shortening of k in the TEG tracings (Table 3).The paralneter a represents the lnaximuln clot strength.The parameter k represents the ~peed of clot forlnation, fibrin crosslinking and platelet-fibrin interaction.Both parameters may be decreased in cases of throlnbocytopenia or hypofibrinogenemia.Although with no statistical significance, the r value also decreased postphlebotomy.
These last results suggest that improvements in thrombin fonnation have occurred postphlebotomy.This could be a non-specific reaction to bleeding, as increases in coagulation activity measured by TGE following blood losses during surgery have been described. 41Such changes were documented even 2 hours after surgery and could have accounted for cathecolamine releases related to stress, and the entry into circulation of coagulation factors froln the hepatic interstitial fluido However, our final coagulation tests were performed about 7 days after the last phlebotomy, again making this interpretation less probable.Another possibility is that decreased helnatocrits led to falls in viscosity with better liver perfusion and oxygenation, promoting improvelnent its production of coagulation factors.
The increase in a Inay also be related to the en)argement' of the platelet counting.Although the fibrinogen leveIs have not changed, the growth in a postphlebotomy may also be in part secondary to a higher availability of fibrinogen for the fibrin network making.In high Hct samples, the total alnount of fibrinogen may not be enough to make a fibrin network capable of strongly containing an increased nUlnber of red cells.Low Hct samples may have proportionally Inore fibrinogen to be used in the fibrin network, producing clots with greater texture.
In summary, the coagulation tests made after phlebotolny have shown slnall changes towards an augmented coagulability.Such changes could be the result of an organic reaction to bleeding, especially the increase in platelet counting.Although the precise clinicaI meaning of these findings are unknown, the literature contains a reference to a patient with pulmonary artery thrombosis following phlebotomies. 42However, alnong the different tests done, only k reached a final value out of our normal laboratory range.Therefore, we speculate that the detected changes do not represent a harmful thrombotic risk related to the phlebotomies Based in the present results, we do oot think the prophylactic adlninistration of anticoagulant agents after phlebotomies in COPD patients is necessary.In addition, we recolnmend oral iron supplementation after phlebotolnies for polycythelnic COPD patients in order to avoid iron deficiency.

Table 1
There were significant decreases in the serunl leveIs of iron, ferritin and TSI postphlebotolny (Table2).Twelve studies.The second set of blood tests were collected 8.7 :t 1.3 days after the Iast phlebotomy.The time between blood collection and actuaI sample analysis was no longer than two hours in every case.