Iodinated contrast media

Uses

For many types of investigation, contrast media have to be given intravascularly. Intravascular iodine-based contrast media are used for angiography and computed tomography, but since they are excreted mainly by the kidneys, they are also used for excretory urography. Although the low-osmolar contrast media are safer than the high-osmolar contrast media, their use has not been universal because of high costs. However, the prices of low-osmolar contrast media are dropping, and in many countries high-osmolar ionic contrast media are no longer used intravascularly [ , ]. The media now used to outline cavities are mainly water-soluble and are preferred to water-insoluble or oily contrast agents.

• Cholecystography requires a contrast medium that is excreted primarily by the liver; iodipamide and ioglycamide have been long preferred, but iodoxamate and iotroxate are more efficiently excreted by this route. However, cholecystography has become obsolete, with the availability of high quality real-time ultrasound scanning.

• In the gastrointestinal tract the insoluble barium sulfate is still most commonly used, but water-soluble iodinated media can be used in special circumstances.

• For retrograde urography, many different water-soluble iodinated media can be used, provided they are diluted sufficiently.

• For arthrography, ductography, and sinography, many different water-soluble iodinated media can be used, provided they are diluted sufficiently.

• For lymphography and ductography, iodinated oil is used. Iodinated water-soluble non-ionic dimeric media have also been used.

• Hysterosalpingography is now mostly performed with low-osmolar water-soluble media.

• Bronchography was performed in the past with an aqueous or oily suspension, but these formulations have been withdrawn; the water-soluble non-ionic dimer iotrolan has been successfully used for fiberoptic bronchography.

• Myelography represents a special challenge, because of the sensitivity of nervous tissues to direct toxic effects; for this purpose the ionic media (which are neurotoxic) and the oily medium iophendylate have been superseded by non-ionic media, including the iso-osmolar non-ionic dimers (iotrolan and iodixanol), which are isotonic with cerebrospinal fluid and well tolerated by nervous tissues. However, myelography is rarely used in modern diagnostic imaging departments and has been widely replaced by magnetic resonance imaging (MRI).

General adverse effects and adverse reactions

The main types of collateral adverse reactions to contrast media are listed in Table 1 .

Low-osmolar contrast media are better tolerated than high-osmolar ionic media in cardiac and coronary angiography. However, it has been suggested that ionic contrast media may be advantageous during percutaneous transluminal coronary angioplasty, as they have some anticoagulant effect, which non-ionic media do not [ ].

The diagnostic accuracy of bronchography with iotrolan is similar to that reported with Dionosil [ ]. No serious complications or significant change in spirometry measurements and oxygen saturation are observed with bronchography using iotrolan. The main adverse effects are nausea and vomiting. Pyrexia, pulmonary infection, and urticaria-like reactions are rare [ ].

Time-course

Most reactions occur within 5–10 minutes of injection, but they can be delayed for up to 7 days; it is advisable for patients to be under close observation for about 20 minutes after injection. In 449 fatal reactions to ionic media, deaths occurred in all age groups, but the incidence peaked in the age group of 50–70 years; 80% of the reactions began in the first 5 minutes after injection, and the other 20% began at 5–15 minutes after injection [ ]. Late but mild reactions can also occur, notably those that are due to “iodism” [ ]. Serious delayed reactions have been reported from Germany and Japan in 26 cases; the features included shock, hypotension, angioedema, and dyspnea [ ].

Delayed reactions to both ionic and non-ionic media have an incidence of 0.4–18%. Most reactions are not serious or life threatening; they include a flu-like illness, sialadenitis, nausea and vomiting, abdominal pain, headache, and allergy-like reactions with skin manifestations. The incidence of delayed skin reactions to iso-osmolar non-ionic dimers is twice as high as the incidence of delayed skin reactions to low-osmolar non-ionic monomers, and twice as high in Japan and USA as in Europe. The pathophysiology of these delayed reactions is unknown. In response to reports of delayed skin reactions to the non-ionic dimer iotrolan, particularly in Japan, the manufacturer (Schering, Berlin) decided to withdraw it from the market as an intravascular agent [ ].

The incidence of delayed adverse reactions has been investigated in 2001 patients who underwent cardiac angiography with either iopamidol 340 (a non-ionic monomer; n = 738, ages 22–88 years, mean dose 1.79 ml/kg), ioxaglate 320 (an ionic dimer; n = 644, ages 28–86 years, mean dose 1.8 ml/kg) or iodixanol 320 (a non-ionic dimer; n = 619, ages 26–85 years, mean dose 1.74 ml/kg) [ ]. The authors considered reactions that occurred within 24 hours as early reactions and those that occurred after 24 hours but within 1 week of the cardiac catheterization as late reactions. The incidences of early reactions were 22% with ioxaglate, 8.8% with iopamidol, and 7.6% with iodixanol. The commonest early reactions were urticaria and nausea and vomiting, the respective incidences being 6.8 and 4.0% with ioxaglate, 0.8 and 1.0% with iopamidol, and 0.5 and 1.1% with iodixanol. A few patients developed sudden hypotension, 0.5% with iopamidol, 1.4% with ioxaglate, and 0.3% with iodixanol. Cardiac arrest occurred in two patients with iopamidol, three with ioxaglate, and one with iodixanol. The frequencies of delayed skin reactions, which were generally benign, were 12% with iodixanol, 4.3% with ioxaglate, and 4.2% with iopamidol. The authors concluded that selection of a contrast agent for diagnostic cardiac catheterization should take account of adverse effects. Iodixanol was the best-tolerated agent in the early phase of the study, but it was associated with a higher incidence of delayed skin reactions. The authors did not consider that the skin reactions represented a contraindication to iodixanol, although patients should be advised of this particular adverse effect. Furthermore, no contrast agent was free from these effects. Ioxaglate was the least well-tolerated agent in the early phase, with a much greater chance of causing nausea/vomiting and allergic reactions.

Incidence

In various earlier surveys of conventional ionic contrast media, the incidence of minor reactions was one in 13–30 cases, the incidence of intermediate reactions one in 57–130 cases, and the incidence of severe reactions one in 1000–4000 cases. The figures for the non-ionic media are much more favorable. In 1990, the Japanese Committee on the Safety of Contrast Media surveyed 169 284 patients who had received ionic media and 168 363 who had received non-ionic contrast media [ ]. In patients with a previous history of reactions to contrast media, the incidence of severe reactions was 0.73% with ionic media and only 0.18% with non-ionic media. Among patients with asthma, severe and very severe reactions occurred in 1.88% with ionic media and 0.23% with non-ionic media. In a Canadian survey of 1992, the overall incidence of adverse reactions to contrast media was 3.9% for ionic media and only 0.9% for non-ionic media, despite the fact that the proportion of patients with heart disease as a pre-existing susceptibility factor was much higher in the non-ionic group [ ].

In a study from France the incidence of adverse drug reactions after 1480 injections of low-osmolar iodinated contrast media was 0.34%; the frequency was higher in patients with a history of allergy (1.5%) [ ]. In a similar study of the prevalence of adverse drug reactions amongst in-patients in a North Indian referral hospital over 3 years, 317 adverse reactions were reported (0.3%) [ ]. Skin reactions (123 cases, 39%) and gastrointestinal disturbances (90 cases, 28%) made up a large proportion of the reported adverse reactions. Of all the adverse reactions, 15% (48 cases) were due to iodinated contrast media (details of the types of contrast media used were not provided in the report) and the common reactions were nausea/vomiting (24 cases, 7.5%) and rashes (16 cases, 5%); however, there was serious life-threatening anaphylaxis in three cases (0.9%).

The incidence of adverse reactions has been investigated in Taiwan in 28 364 patients after intravenous injection of contrast agents during intravenous urography and CT scanning [ ]. There were adverse reactions in 495 patients (1.75%), including 467 patients (2.03%) of 20 260 examined with ionic contrast agents and 28 (0.34%) of 8076 patients examined with non-ionic contrast agents. The authors concluded that the risk of adverse reactions to non-ionic contrast agents is significantly lower than with ionic agents. Rashes, such as urticaria, were the most common adverse reactions, followed by nausea and vomiting. Shock requiring cardiopulmonary resuscitation was rare and occurred in only six patients (0.02%); there were no deaths in this series.

Adverse reactions to intravascular iodinated agents are usually classified as minor, intermediate, or severe life-threatening. All types of reactions to low-osmolar contrast media are five times less common than reactions to high-osmolar contrast agents [ ], and very severe adverse reactions to contrast media are rare, with a frequency of about 0.04% with high-osmolar agents and 0.004% with low-osmolar agents. However, there are no important differences in the safety profiles of the different low-osmolar non-ionic monomers [ ].

In one US study of cardiac angiography, the incidence of adverse events was 32% with diatrizoate and 10% with iohexol; it should be noted however that the diatrizoate used in that study (Renografin 76) causes calcium binding, which may have increased the rate of cardiac complications [ ]. The non-ionic agent iodixanol 320, which is virtually iso-osmolar, has been tested in cardiac angiography against iohexol and seemed to be better tolerated [ ].

The safety of the non-ionic contrast agents has also been emphasized in a review from Germany, in which the authors documented a reduced risk of adverse drug reactions with non-ionic contrast agents during interventional cardiology [ ]. According to reports to the FDA, the incidence of lethal complications was 3.9% with ionic monomers, 6.39% with the ionic dimer ioxaglate, and 2.07% with non-ionic monomers. The authors also stated that non-ionic contrast agents do not have thrombogenic potential, a previous concern; currently, most authors would support this view [ , ].

There was no difference in the incidence of adverse reactions between iotrolan 320 and iohexol 350 in cardiac angiography for ischemic heart disease in 120 patients [ ]. There were no serious adverse events. One patient developed mild urticaria after iotrolan. Five patients in each group had mild unspecified delayed reactions.

An analysis of adverse reactions associated with drugs in Poland (1997–98) has shown that 11.5% of these reactions were due to radiographic contrast media. People aged 36–65 were the largest group to have adverse drug reactions [ ].

According to a survey conducted by the Royal Australian College of Radiologists in 1986, the incidence of severe reactions with high-osmolar ionic contrast media was 0.36% in high-risk patients (patients with a strong history of allergy or bronchial asthma or a history of reactions to contrast media) and 0.09% in low-risk patients [ ]. The incidence of these reactions with low-osmolar non-ionic media was 0.03% in high-risk patients, and 0% in low-risk patients.

In another survey, the incidence of contrast media reactions after intravenous administration was evaluated over 14 years [ ]. The incidence of all reactions to contrast media was 6–8% with high-osmolar contrast media and only 0.2% with low-osmolar non-ionic agents. Most of the reactions (over 90%) were allergic-like, and severe reactions were rare (0.05%). One death was reported after the use of a low-osmolar agent. These data are compatible with previous reports, which showed that low-osmolar contrast media have a much better safety profile than high-osmolar media and that there is no significant difference in the incidence of acute adverse reactions between non-ionic dimeric and monomeric contrast media.

In another study there was no significant difference in the incidence of adverse reactions associated with the use of the non-ionic dimer iodixanol or the non-ionic monomer iopromide in femoral angiography [ ].

Delayed reactions

There are conflicting data on the risk of delayed reactions, comparing non-ionic monomers and iso-osmolar non-ionic dimers (iotrolan, iodixanol). European studies [ ] have shown no significant difference (0.62% iotrolan, 0.82% non-ionic monomers), but studies from Japan and the USA have suggested that delayed reactions are 2–5 times more common with iotrolan.

The incidence of delayed adverse reactions to various non-ionic monomers in Japanese patients undergoing contrast-enhanced CT has been compared with the incidence in control patients who did not receive contrast. Delayed reactions occurred in 12% of patients who received contrast compared with 10% in the control group. The authors concluded that the frequency of delayed reactions that can be attributed to contrast media is 2.1%. The most common reactions were itching and limited urticaria [ ]. This report highlights the difficulty of verifying that adverse events that occur many hours after contrast administration are directly caused by the contrast agent.

The incidence of delayed reactions has been assessed in 403 Italian patients who received intravenous iopamidol during CT or urographic examination [ ]. A total of 50 patients (12%) developed delayed reactions. Allergy, previous exposure to a contrast agent, and being female were associated with a significantly higher incidence of delayed reactions. The most frequently reported delayed reactions were nausea and vomiting, drowsiness, rash, itching, and headache. All reported reactions were mild and resolved spontaneously.

In a questionnaire study of 11 121 Japanese patients, 216 (1.4%) developed immediate adverse reactions and 1058 (9.5%) reported having had delayed reactions after intravenous contrast administration during various CT examinations [ ]. Delayed reactions were reported by 18 patients (13%) of the 136 patients with immediate reactions who answered the questionnaire. All the patients, with the exception of 360 who received the non-ionic dimer iotrolan, were given non-ionic monomeric contrast media. The dose was 60–200 ml. Delayed reactions were more frequent in patients with a history of allergy, past adverse reactions to contrast media or with a serum creatinine concentration over 180 μmol/l. Delayed reactions were also more frequent in women and in patients who had not previously received contrast media. There was no significant relation between the occurrence of immediate adverse reactions and the development of delayed reactions. The commonest delayed reactions were itching and skin reactions, which developed in 5.5% and 3.0% of the patients respectively. Skin reactions were twice as common in patients who were given iotrolan than in those who were given monomeric agents. In the iotrolan group, 7.3% of the patients developed a skin reaction, 9.9% reported itching, and 60% of the reactions were severe or moderate. A quarter of the delayed reactions occurred within 6 hours after the examination and more than half occurred within 24 hours. Most of the reactions occurred within the first 3 days.

The incidence of delayed adverse reactions has been investigated in 2001 patients who underwent cardiac angiography with either iopamidol 340 (n = 738, aged 22–88 years, mean dose 1.79 ml/kg), ioxaglate 320 (n = 644, aged 28–86 years, mean dose 1.8 ml/kg) or iodixanol 320 (n = 619, aged 26–85 years, mean dose 1.74 ml/kg) [ ]. The authors considered reactions that occurred within 24 hours as early reactions and those that occurred after 24 hours but within 1 week as late reactions. The incidences of early reactions were 22% with ioxaglate, 8.8% with iopamidol, and 7.6% with iodixanol. The commonest early reactions were urticaria and nausea and vomiting, the respective incidences being 6.8 and 4.0% with ioxaglate, 0.8 and 1.0% with iopamidol, and 0.5 and 1.1% with iodixanol. A few patients developed sudden hypotension, 0.5% with iopamidol, 1.4% with ioxaglate, and 0.3% with iodixanol. Cardiac arrest occurred in two patients with iopamidol, three with ioxaglate, and one with iodixanol. The frequencies of delayed skin reactions, which were generally benign, were 12% with iodixanol, 4.3% with ioxaglate, and 4.2% with iopamidol. The authors concluded that selection of a contrast agent for diagnostic cardiac catheterization should take account of adverse effects. Iodixanol was the best-tolerated agent in the early phase of the study, but it was associated with a higher incidence of delayed skin reactions. The authors did not consider that the skin reactions represented a contraindication to iodixanol, although patients should be advised of this particular adverse effect. Furthermore, no contrast agent was free from these effects. Ioxaglate was the least well-tolerated agent in the early phase, with a much greater chance of causing nausea/vomiting and allergic reactions.

The Federal Institute of Drugs and Medical Devices in Germany has surveyed a total of 1135 adverse drug reactions (acute and late) associated with iotrolan 280 and 1354 reports associated with iodixanol at various iodine concentrations [ ]. There were late adverse reactions (observed later than 1 hour after injection) in 757 cases (67%) associated with iotrolan 280, and in 525 cases (39%) associated with iodixanol. Late reactions were observed mainly in the first 24 hours, with occasional reports at 24–72 hours after injection. Rarely, delayed contrast reactions have been reported more than 72 hours after injection, although they are difficult to substantiate. This study emphasizes the importance of extending the surveillance period for at least 72 hours after contrast administration. Most delayed reactions in this study were non-serious allergic-like reactions, with symptoms including itching, urticaria, erythema, edema, and bronchospasm. In a very few cases there were serious or life-threatening reactions, including Quincke's edema and hypotension.

The incidence of late adverse reactions to the iso-osmolar agent iodixanol has also been investigated in a retrospective comparison with the non-ionic monomer iohexol. Reactions after 3075 injections were reviewed. Patients were sent a written questionnaire to find out about the incidence of any acute or delayed reactions. Those who developed reactions were interviewed by telephone and specific second questionnaires were administered. The incidence of adverse reactions was low (2% for late adverse reactions and 2.3% for acute reactions). There was no significant difference in the incidence of acute or delayed adverse reactions to iohexol and iodixanol and there were no serious reactions [ ].

Delayed reactions are generally benign, but not always. In a Japanese study, the incidence of delayed reactions was investigated in 6764 patients who received the low-osmolar non-ionic contrast medium iohexol intravenously [ ]. Delayed reactions (rash, pruritus, nausea, vomiting, fever, headache, and others) occurred in 192 patients (2.8%). There were no severe delayed reactions. A history of allergy and hay fever were risk factors for delayed adverse reactions.

Delayed adverse reactions to ioxaglate have been documented after coronary angiography with ioxaglate [ ].

• A 63-year-old housewife had an acute myocardial infarction. Diagnostic coronary angiography was performed with ioxaglate and repeated 1 week later. On the morning after, she had intense shivering and generalized malaise, a temperature of 39.2 °C, and hypotension. Several hours later she developed a non-pruritic maculopapular rash, starting on the face and extending within the next 48 hours to cover the entire body but sparing the mucous membranes. She had a mild eosinophilia (600 × 10 ⁶ /l) with a raised erythrocyte sedimentation rate (52 mm at 1 hour). Her IgE concentration was raised at 1593 kU/l. Blood cultures were negative. The initial clinical impression was that she had sepsis and she was therefore given vancomycin, netilmicin, and cefotaxime (dosages not stated). She improved rapidly and the fever and the rash resolved completely. She was discharged taking atenolol 100 mg/day and lysine acetylsalicylate 250 mg/day, but 3 days later she returned with a temperature of 39.8 °C and recurrence of the rash. She had a marked eosinophilia (1.4 × 10 ⁹ /l) and raised liver enzymes, lactate dehydrogenase, and aldolase, but creatine kinase activity was in the reference range. She had a significantly raised concentration of circulating immune complexes. Atenolol and lysine acetylsalicylate were withdrawn and she was given acenocoumarol 2 mg/day and amlodipine 5 mg/day. The inflammatory syndrome rapidly disappeared and the enzymes and eosinophil count returned to normal. The total IgE was still raised 2 months later. The leukocyte count and C-reactive protein concentration were within the reference ranges. The serum antinuclear antibody titer was 1/256 and the antimitochondrial antibody titer was 1/128. Tests for antinative DNA, antihistone, and anti-smooth muscle antibodies were negative. Skin prick tests, intradermal tests, and patch tests with lysine acetylsalicylate, atenolol, heparins, vancomycin, netilmicin, cefotaxime, and the contrast agents ioxaglate, sodium ioxitalamate, iopamidol, and iohexol were negative at 15 minutes, but at 48 hours indurated erythematous papules were observed with ioxaglate, ioxitalamate, and iopamidol. A biopsy from the ioxaglate skin reaction showed discrete spongiosis of the epidermis associated with slight lymphocytic exocytosis. The basal layer contained numerous apoptotic keratinocytes. The superficial dermis was edematous and there was a perivascular inflammatory infiltrate composed mainly of T lymphocytes. There were no mast cells or eosinophils. Immunohistological examination showed no staining with anti IgE, IgG, IgM, C1q, C3, or C4. During follow-up she was completely normal at up to 1 year, with a normal eosinophil count and an IgE concentration of 440 kU/l.

This report documents a rare clinical reaction to ioxaglate, with a combination of a maculopapular rash, fever, hepatic and muscle involvement, eosinophilia, and a very high serum IgE concentration. The intradermal tests confirmed a delayed hypersensitivity reaction to ioxaglate. Histological examination of a skin biopsy identified the predominantly T lymphocyte nature of the infiltrate. A contributing role of the beta-blocker atenolol to the seriousness of the clinical syndrome must also be considered.

Mechanisms and susceptibility factors

Although many reactions resemble allergic or anaphylactic events, the evidence does not suggest that they are as a whole induced immunologically, though some certainly are, and there is no doubt that in some cases contrast media act as antigens. Lalli believes that all contrast media reactions are explicable on a neurological basis; this hypothesis is not generally accepted, but most would agree that anxiety can be an important predisposing factor, for example to cardiac dysrhythmias. It has been suggested that contrast media act as histamine liberators, and there is evidence from animal experiments that this is the case, the methylglucamine media being more potent in this respect than the corresponding sodium salts (though ioxaglate may apparently also elicit histaminoid reactions [ ]; the antihistamine diphenhydramine has been recommended both for preventing and treating reactions (see below) and may have some efficacy. A further theory is that bradykinin is an essential mediator in systemic contrast media reactions [ ]. There is also older experimental evidence that contrast media can activate serum complement by the “alternative pathway”, and it is postulated that this may be one of the factors in systemic adverse reactions, possibly by liberation of anaphylatoxin and with a risk of disseminated intravascular coagulation.

Adverse reactions to contrast agents have been linked to an in vivo interaction with enzymes involved in cholinergic activity. However, experimental results on this topic are conflicting. In a study of the in vivo effects of the non-ionic monomer iohexol and the high-osmolar ionic monomer diatrizoate on human plasma acetylcholinesterase and butyrylcholinesterase, both contrast agents significantly inhibited the enzymes [ ]. The effect was more pronounced with iohexol, which has a lower incidence of adverse reactions than diatrizoate. The significance of this observation is not certain, but activation of these enzymes is not likely to be important in mediating adverse reactions to contrast agents. Other enzymes that are inhibited by iodinated contrast agents include hexokinase, glucose-6-phosphate dehydrogenase, and alcohol dehydrogenase [ ].

More minor adverse reactions to the second-generation non-ionic water-soluble contrast media, such as headache, nausea, and vomiting, are more frequent in women than in men, particularly in women aged 26–50 years. Early ambulation increases the incidence of headaches. In one study, there was a higher incidence of delayed headaches after iopamidol in comparison with iohexol. Long-distance travel also appeared to increase the incidence of headaches after myelography [ ]. Dizziness was more frequent in patients over the age of 50 years. In a meta-analysis of 25 published studies, the incidence of symptoms after myelography was significantly higher when needles larger than 22G were used.

The risks (expressed as risk ratios) of adverse reactions to contrast media after intravenous injection related to some susceptibility factors are listed in Table 2 .

Prevention and management

Intravenous pre-testing is unreliable in determining susceptibility; in one survey of 33 400 patients, most of the reactions were not predicted by pretesting, although a positive pre-test did indicate an increased risk of a reaction [ ]. The value of the lymphoblast transformation test is very limited, although it certainly provides interesting evidence that contrast media can sometimes act as antigens [ ].

The value of prophylactic use of glucocorticoids to reduce the risk of contrast reactions is contentious. The Contrast Media Safety Committee (CMSC) of the European Society of Urogenital Radiology (ESUR) considered this issue and produced guidelines on how to reduce the risk of generalized reactions to intravascular use of contrast media ( Table 3 ) [ ].

In high-risk patients, and more particularly those with a history of poor tolerance of contrast media, measures will be necessary to contain risks of hypersensitivity reactions. Apart from using non-ionic and low-osmolar media, and diluting media for certain purposes, one can consider prophylactic premedication. Glucocorticoid premedication alone has often been advocated, particularly in patients with asthma; although it has been condemned as unnecessary with the current generation of contrast media [ , ] it seems to be well founded in experience, reducing both the frequency and severity of reactions, but it certainly does not eliminate risk [ ]. A double regimen based on prednisolone and the antihistamine diphenhydramine has been advocated in various centers [ , ]. To be effective, glucocorticoids should be given sufficiently far in advance; in some cases, an intravenous injection of dexamethasone given immediately before an infusion of contrast media failed to prevent an anaphylactic reaction [ ].

Histamine H ₁ and H ₂ receptor antagonists, aminocaproic acid, and hyposensitization have all been proposed as means of reducing risk prophylactically. Aminocaproic acid was formerly used prophylactically in France, but was associated with a risk of massive intravascular coagulation [ ].

In severe reactions, intravenous glucocorticoids are usually given on an empirical basis, with oxygen as required. Non-cardiogenic hypotensive shock usually responds best to fluid replacement, but vasopressors are occasionally required. Adrenaline is primarily indicated for bronchospasm and other allergic-type reactions, but caution is required to avoid cardiac dysrhythmias. Intravenous antihistamines are useful in angioedema, but can aggravate hypotensive reactions. Chemotoxic convulsions require intravenous diazepam and oxygenation.

Systemic adverse reactions to enteral contrast agents

Oral radiographic contrast media can intravasate during gastrointestinal examination. Septicemia has been reported [ ].

• A 46-year-old patient developed ischemic necrosis of the small bowel complicating mesenteric volvulus. Small bowel resection was carried out and ileal and jejunal stomas were established. Six weeks after the operation Gastrografin was given through the jejunal stoma. A mesenteric vein filled with Gastrografin and this was followed by rapid washout of the intravascular contrast medium. The examination was stopped immediately and 60 minutes later the patient developed chills and a high fever. Hemodynamic instability required the use of vasoactive drugs and infusion of isotonic solutions. Blood cultures grew Escherichia coli , Pseudomonas aeruginosa , and Enterobacter species. Treatment included imipenem and after 48 hours the patient was stable and the vasoactive agents were stopped. Endoscopy through the jejunal stoma showed multiple stenoses, which required surgical treatment, and the jejunum in the area of intravasation was resected.

Gram-negative septicemia possibly related to oral Gastrografin has also been reported in a premature neonate with necrotizing enterocolitis [ ].

• A baby born after 26 weeks of gestation received oral Gastrografin (0.5 ml 8 hourly) for gut stimulation. Three days after the last dose of Gastrografin there was sudden clinical deterioration leading to shock. Blood cultures grew Enterobacter . Progressive deterioration continued to occur, with multisystem failure, leading to death. Autopsy was not performed.

The authors suggested that the high osmolality of Gastrografin may have aggravated the pre-existing mucosal injury in the gastrointestinal tract, leading to complete loss of mucosal integrity and an increase in gut permeability to micro-organisms and toxins. Oral Gastrografin and other hyperosmolar contrast media should not be used in patients, particularly neonates, with compromised bowel integrity, since the high osmolarity may aggravate the bowel injury. Low-osmolar contrast media should always be used in such patients.

Disseminated intravascular coagulation and severe hypotension have been documented after intravenous administration of Gastrografin, which is suitable only for oral or rectal administration [ ].

Cost-benefit ratio

At a time when reductions in the costs of medical care are critical, decisions about which classes of contrast media to use are not determined purely on clinical grounds, but by a consideration of the cost-benefit ratio. Concern about financial implications has been the major factor in preventing a universal conversion to non-ionic contrast agents, which are better tolerated but more expensive than high-osmolar ionic agents. In the USA, the Health Care Finance Administration (HCFA) has recommended that the use of the expensive non-ionic contrast media should be restricted only to patients with severe cardiac disease, a history of asthma, severe allergy, severe debility, sickle cell disease, or a previous severe adverse reaction to contrast media.

The results of a study in 1324 patients who underwent diagnostic arteriography have supported the selective use of non-ionic contrast agents following the HCFA guidelines. A cost saving of $41 per patient was possible without an increase in the incidence of complications [ ].

Observational studies

The incidences of early reactions (< 24 hours) and late reactions (1–7 days) to two non-ionic contrast agents iopamidol 340 and iomeprol 350 (Niopam and Iomeron, Bracco UK Ltd) have been evaluated in a double-blind, randomized trial in 1985 patients undergoing cardiac catheterization [ ]. There was no significant difference between the two groups in the incidence of heat sensation or other early reactions (2.7% of those who received iopamidol and 4% of those who received iomeprol). There were significant electrocardiographic changes in 1.7% and 1.0% respectively; bradycardia was significantly more common after iopamidol (0.8%) than iomeprol (0.1%). Late reactions occurred in 16% and 22% respectively. There was significantly more nausea with iomeprol—39 patients (6.2%) versus 23 (3.7%).

In a study of spontaneous reports of suspected adverse reactions attributed to contrast media sent to the Sicilian Regional Centre for the Spontaneous Reporting of Adverse Drug Reactions during 1996–2006 there were 100 reports, 29 of which involved serious adverse effects, including one death [ ]. Skin, the respiratory system, and the gastrointestinal tract were the most frequently affected sites. Most of the reports described immediate hypersensitivity reactions. Iopromide (53%), iopamidol (14%), and iomeprol (12%) were the most commonly implicated contrast media.

The incidences of immediate and delayed adverse reactions to contrast media have been evaluated in a prospective intensive monitoring questionnaire study in 1514 patients who were given iodinated contrast media in two hospitals in Tuscany [ ]. There were immediate and delayed reactions in 34 (2.2%) and 144 (9.5%) respectively. Delayed allergy of mild severity was more common with the use of monomeric low-osmolar agents (iopromide, iomeprol, iobitridol) and the dimeric iso-osmolar agent iodixanol. Only one immediate reaction was severe. Immediate reactions were more likely with the monomeric agents and delayed reactions with iodixanol.

Comparative studies

The effects of the iso-osmolar non-ionic dimer iodixanol (iodine 320 mg/ml) and the low-osmolar ionic dimer ioxaglate (iodine 320 mg/ml) have been compared in 1411 patients, mean age 62 years, undergoing percutaneous transluminal coronary angioplasty in a multicenter, randomized, double-blind study [ ]. The groups were comparable in relation to the prevalence of cardiac and other medical conditions, including diabetes, obesity, and smoking habits. All the patients received heparin and all but four received an antiplatelet agent (100 mg or more of aspirin and/or ticlopidine). There was no significant difference between the two groups—the incidence of major adverse cardiac events was 4.7% (192 patients) after iodixanol and 3.9% (197 patients) after ioxaglate. However, hypersensitivity reactions and adverse drug reactions were significantly less common with iodixanol (n = 5) than with ioxaglate (n = 18). The reactions to iodixanol were mainly rashes and urticaria-like reactions. The reactions to ioxaglate were anaphylaxis (n = 1), urticaria (n = 12), coughing and throat tightness (n = 12), and in one patient rigors, fever, vomiting, and flushing.

In a large comparative study of iodixanol and iohexol, the incidence of immediate and delayed adverse events was significantly higher with the former [ ]. Iodixanol was used for CT scanning in 15 142 patients and iohexol in 22 044 patients; there were adverse events in 116 patients (0.77%) and 54 patients (0.25%) respectively. There were immediate and delayed adverse events in 76 and 40 patients (0.50% and 0.26% respectively) who received iodixanol and in 52 and 2 patients (0.24% and 0.01% respectively) who received iohexol.

The authors of a large review of the comparative tolerability of contrast media used for coronary intervention recommended the use of ionic or non-ionic dimers over non-ionic low-osmolar monomeric agents, because of the antithrombotic advantage of the dimers [ ]. However, the recommendation to use ionic low-osmolar contrast media in preference to non-ionic contrast media is not widely endorsed, since the latter are less toxic than ionic media. In another study, there was no difference in the incidence of contrast media reactions between the non-ionic monomeric contrast agents iopamidol and iopromide, and both agents were well tolerated following left ventricle and coronary angiography [ ]. The authors concluded that both agents are perfectly acceptable for cardiac angiography.

Iodixanol 320 (n = 307) has been compared with the low-osmolar ionic dimer ioxaglate 320 (n = 311) in patients undergoing percutaneous coronary intervention [ ]. The two groups were well matched for basic demographic data and comparable in relation to cardiovascular state and other medical conditions. Fifteen patients (4.9%) in the iodixanol group and 27 patients (8.7%) in the ioxaglate group had reactions within 24 hours of the procedure; they were likely to have been due to the contrast medium, and included nausea, vomiting, flushing, facial edema, urticaria, and wheezing. There was no significant difference in the outcome of the procedure between the two groups. The authors suggested that in view of the higher incidence of adverse reactions to ioxaglate, its automatic selection for percutaneous coronary intervention should be reviewed.

Cardiovascular

Most severe reactions to contrast media are associated with cardiovascular manifestations, causing hypotensive shock and in some cases ventricular fibrillation and cardiac arrest; these events are reversible in most cases in which prompt treatment is given. In a case of hypotensive collapse reported in 1977, and followed by a small number of others, there was disseminated intravascular coagulation [ ]. In milder cases there is only hypotension, which can be transient and symptomless; in some cases there is bradycardia (due apparently to vagal overactivity) rather than tachycardia.

In more than 90 000 cardiac angiographies performed in US hospitals during 1991, the overall rate of complications with low-osmolar contrast media was 1.5%, including vascular complications in 0.44%, neurological complications in 0.05%, dysrhythmias in 0.31%, and myocardial infarction in 0.06%; the death rate was 0.11%. In percutaneous coronary angioplasty major complications, generally of the same type, occurred in 5% of cases [ ]. In one large series of digital subtraction angiography examinations using iopamidol, the overall incidence of reactions was 2.5%; some occurred with a delay of 1 hour or more [ ].

The safety of iodixanol 320 and iohexol 350 has been investigated in Swedish patients undergoing cardiac angiography for suspected coronary artery disease [ ]. Of 1020 patients, 502 aged 25–83 years received iohexol (median dose 105 ml, range 20–440) and 518 aged 18–85 years received iodixanol (median dose 115 ml, range 30–400). There were 134 patients with unstable angina in the iohexol group and 167 in the iodixanol group. Cardiac adverse events (angina pectoris, dysrhythmias, and dyspnea) within 24 hours of the examination were reported by 9% of the patients who received iohexol and 7% of patients who received iodixanol. There were two cases of ventricular fibrillation, both after iohexol. Cardiac adverse events in patients aged 65 years or more occurred in 11% with iohexol and 7% with iodixanol. The proportions of patients with unstable angina and cardiac adverse events were 18% with iohexol and 12% with iodixanol. The authors concluded that iodixanol could be advantageous in old patients and in those with unstable angina.

Iodixanol (a non-ionic dimer, 320 mg of iodine per ml) and ioxaglate 320 (a low-osmolar ionic dimer, 320 mg of iodine per ml) have been compared in a randomized study in 110 consecutive patients referred for coronary angiography and ventriculography [ ]. The incidence of adverse reactions was significantly higher with ioxaglate (28 versus 3%) but there was no difference in angiographic quality between the two agents. The increase in left ventricular end-diastolic pressure was significantly less with iodixanol than with ioxaglate. The QT interval was significantly prolonged by both agents, but the changes were less marked after iodixanol. The authors concluded that iodixanol and ioxaglate are of comparable diagnostic efficacy in coronary angiography and ventriculography but that iodixanol is better tolerated and has less marked hemodynamic and electrophysiological effects.

The hemodynamic effect of direct intra-arterial injection of contrast agents on capillary perfusion in man has been investigated [ ]. This was achieved through continuous recording of perfusion in the nail-fold capillaries of the right hand before and after a bolus injection of 20 ml of iodixanol 270 (a non-ionic dimer) or iopentol 150 (a non-ionic monomer) into the right axillary artery. The high-viscosity contrast agent iodixanol (5.8 mPa.s) caused a significant reduction in erythrocyte velocity, while iopentol, which has a much lower viscosity (1.7 mPa.s), had no effect. The authors concluded that high-viscosity contrast media can cause reduced organ perfusion. This effect could be significant in patients with atherosclerotic disease, as it might lead to reduced perfusion of the myocardium during coronary angiography.

Respiratory

Transitory changes in lung function can occur after bronchography due to retention of the contrast medium in the bronchi; normal diffusing capacity may not return for 3 days and it is better to avoid thoracic surgery during this period. When bronchography with propyliodone preparations is performed during fiberoptic bronchoscopy, there may be marked arterial oxygen desaturation [ ].

Some contrast medium inevitably reaches the lungs after injection, and tracer studies in the early 1960s showed that some degree of pulmonary oil embolism occurred in every patient [ ]. Pulmonary function studies have shown abnormalities, in particular a reduction in pulmonary diffusing capacity and pulmonary capillary blood volume, even in the absence of clinical signs or symptoms [ ]. Most cases are symptomless, although there can be mild pyrexia, while X-rays show stippling or an arborization pattern. If hypotension, cyanosis, dyspnea, and pleuritic pain occur, infarction should be suspected. The risk that blockage of lung capillaries by oil will endanger the patient is naturally greater if pulmonary function is already compromised, for example by neoplastic or fibrotic disease, or by radiotherapy.

In some patients, this phase of mechanical obstruction is followed after some days or 3–4 weeks by a chemical reaction, presumably as the oil breaks down to irritant fatty acids, which cause exudation and hemorrhage. Interference with the production of lung surfactant can also occur and a marked intravascular cellular reaction has been described [ ]. The chemical irritative phase is marked by fever, cough with sputum (often blood) and a variable degree of respiratory distress; there may even be tachycardia and hypotension.

Non-cardiogenic pulmonary edema has been seen several times after contrast media in patients with a prior history of myocardial infarct [ ]; it can also be a component of anaphylactic shock. Non-cardiogenic pulmonary edema has been reported as a complication of intravenous injection of iomeprol [ ].

• A 68-year-old man with chronic obstructive pulmonary disease underwent CT examination of the abdomen with intravenous infusion of iomeprol for suspected hepatocellular carcinoma and 2 hours later developed severe dyspnea. A chest X-ray showed bilateral diffuse shadowing of the lungs and the heart shadow was not enlarged. A diagnosis of non-cardiogenic pulmonary edema was made and he improved with glucocorticoids.

Contrast media can cause severe bronchospasm, particularly in patients with asthma [ ]. Subclinical bronchospasm can occur after bolus injections of ionic contrast medium; the incidence is lower with non-ionic media [ ].

Life-threatening adult respiratory distress syndrome after intravascular radiographic contrast injection is rare. It has been successfully managed by extracorporeal membrane oxygenation [ ].

• A 62-year-old man developed adult respiratory distress syndrome 5 minutes after receiving a low-osmolar ionic contrast medium (140 ml of ioxaglate 200) during coronary angiography. His respiratory rate was 35/minute, blood pressure 60/40 mmHg, pulse rate 125/minute, PaO ₂ 49 mmHg, main pulmonary artery pressure 21 mmHg, central venous pressure 9 mmHg, and pulmonary capillary pressure 10 mmHg. A chest X-ray showed pulmonary edema. Laboratory investigations, electrocardiography, echocardiography, and coronary angiography were normal. In view of life-threatening hypoxia, extracorporeal membrane oxygenation was given for 50 hours, and 2 weeks later he was discharged without pulmonary symptoms.

In one case adult respiratory distress syndrome was accompanied by disseminated intravascular coagulation [ ].

• Respiratory arrest occurred after aspiration of water-soluble contrast material in a 12-year-old girl who had been injured in a traffic accident. Gastrografin diluted with tap water (540 ml) was given into the stomach via a nasogastric tube as part of CT-enhanced contrast examination of the abdomen. She also received 150 ml of non-ionic contrast media (iopamidol, iodine 300 mg/ml) intravenously, which was followed by vomiting attacks. She became irritable and had an acute fall in oxygen saturation and progressive respiratory distress, which required endotracheal intubation. The CT scan that was performed after she became stable showed contrast material in the lungs.

The authors concluded that aspiration of contrast material can be life-threatening and that administration of oral contrast media after trauma can increase the risk of aspiration of gastric contents [ ].

Accidental inhalation of hypertonic contrast media during oral administration can cause fatal non-cardiogenic pulmonary edema [ ].

In pediatric bronchography, segmental collapse occurs in half the cases, but is particularly common with the aqueous media. Collapse also seems to be more common when halothane and oxygen are used as the anesthetic agents [ ]; this has been attributed to the rapid absorption of the anesthetic gases combined with the partial bronchial block caused by the contrast medium.

Pulmonary fibrosis and microlithiasis has been reported as a late complication of lymphography with iodized oil, but a causative relation was uncertain [ ].

Nervous system

Animal experiments with the ionic media have shown that these damage the blood–brain barrier, thereby creating a degree of risk to brain tissue when they are used in cerebral angiography, although they have often been used safely. Other evidence suggests that sodium increases the neurotoxicity of the contrast medium, possibly by increasing its diffusibility into the brain. Methylglucamine, on the other hand, reduces neurotoxicity. The addition of a small quantity of calcium also appears to reduce toxicity, at least when added to sodium. There are only marginal differences in neurotoxicity as measured in animal experiments between the diatrizoate, iotalamate, and metrizoate anions. Calcium methylglucamine metrizoate probably has the lowest neurotoxicity of the ionic media, followed closely by methylglucamine iotalamate. In clinical use, however, the calcium-containing formulations appear to cause a greater degree of discomfort because they cause a sensation of heat.

Although the non-ionic media produced less injury to the blood–brain barrier in animals, clinical studies have not generally shown them to be better tolerated when used for cerebral angiography, and it has been argued that these expensive media should now only be used where the blood–brain barrier is thought to be defective [ ].

Transient neurological changes after myelography include asterixis, aphasia, and reversible visual defects. In one patient, asterixis and head bobbing were still present 3 months after myelography with metrizamide [ ]. An isolated case of persistent cervical myelopathy has also been reported after lumbar myelography with metrizamide.

A variety of other neurological complications have also been recorded. These include Guillain–Barré syndrome, auditory or visual disturbances, motor aphasia, sixth nerve palsy [ , ], mania, organic brain syndrome [ ], confusional changes, and absence status. Absence status responds rapidly to intravenous diazepam [ ].

Death has occurred following deterioration of neurological status. There has been a higher incidence of such complications in diabetics [ ]. Aspiration of 20–25 ml of cerebrospinal fluid after metrizamide myelography appears to reduce the incidence of neurological adverse effects [ ].

All ionic contrast media are neurotoxic and they should not be injected into the subarachnoid space or intrathecally, and stern warnings against such use have been issued, for example by the American FDA [ ]. When they have accidentally been used, deaths have resulted [ ]. Accidental misuse of diatrizoate for myelography, with fatal consequences, has been a serious problem [ ]. Lavage of the subarachnoid space with saline in such cases has been effective in reducing toxicity [ ].

Accidental injection of diatrizoate into the subarachnoid space has occurred as a result of misplacement of the needle in arteriography, in discography, from inadvertent use in myelography, or during injection of a myelocele. Severe convulsions with extensor spasms occur and can result in death, particularly if the agent comes into contact with the brain; the onset of convulsions can be rapid, but they can also be delayed for several hours. Two patients developed renal insufficiency as an additional complication [ ].

The earliest water-soluble medium used for lumbar myelography or radiculography, sodium methiodal, had an irritant effect and required spinal anesthesia, which sometimes caused hypotension. Methylglucamine iotalamate was less irritant, while methylglucamine iocarmate had still lower neurotoxicity. However, even these two media could cause convulsions or muscle spasms if the contrast agent came into contact with the conus medullaris. For this reason the patient was postured with the head raised for several hours after the examination to prevent the contrast medium passing above the L-l vertebral level. Diazepam might also be given prophylactically. When convulsions did occur they could be severe, even resulting in fractures or dislocation of a hip. A convulsive pattern without actual convulsions has also been seen on the electroencephalogram in some patients.

Metrizamide was the first non-ionic contrast medium and became widely used for myelography and ventriculography. It was much less neurotoxic than iocarmate, although electroencephalographic changes lasting up to 3 days were found in some 16% of cases in early studies. Today, metrizamide is no longer used and has been succeeded by new generations of non-ionic monomers (iopamidol, iohexol) and dimers (iotrolan, iodixanol) which are well tolerated with low neurotoxicity [ ]. The need for intrathecal injection of contrast media in the developed world has dramatically fallen with the wide availability of MRI.

Visual evoked responses 20 hours after myelography can be delayed, with a correlation with the severity of headache after myelography; the delay in visual-evoked response is less marked with iopamidol than with metrizamide. It has been suggested that this technique may be useful in the assessment of myelographic contrast media toxicity [ ].

In 439 myelographic procedures with metrizamide the most frequent adverse reaction was headache, which could be differentiated into early-onset headache (related to hydrodynamic modifications in the spinal fluid following lumbar puncture) and late-onset headache (reflecting a specific metrizamide effect) [ ]. The frequency of late-onset headache was at least 27%, but altogether 46% of the patients had headache at one time or another. In this series, meningeal irritation was seen in 5%, sometimes in a severe form, mimicking a septic complication. There was spinal irritation in two cases and epileptic fits in one. An acute psychotic organic syndrome was common after cervical myelography with high doses of metrizamide. There was a severe anaphylactic reaction in one patient. The results of this study have generally been confirmed by others; seizures are clearly very rare with these non-ionic media, though they have occurred, even with iohexol [ ], as have involuntary movements and facial twitching [ ] and nystagmus [ ]. Meningeal irritation and paraplegia have both been seen with iohexol [ ] and a case of aseptic meningitis with iotrolan [ ].

In comparative trials, iopamidol and iohexol produced fewer adverse reactions than metrizamide and they do not so far appear to have given rise to the psychosyndrome, but slight electroencephalographic changes can occur even with the newer media. Seizures and clonic jerks have been reported when relatively large doses of iopamidol have been used in myelography [ , ].

Four cases of cauda equina syndrome have been described with methylglucamine iocarmate used for radiculography, although in two cases faulty technique was perhaps contributory.

Loss of consciousness can occur if there is hypotensive collapse; rarely there can be prolonged coma.

Convulsions, seen as part of an idiosyncratic reaction, tend to occur in patients with an existing tendency to epilepsy, or to occur as consequences of hypotensive collapse, cardiac arrest, or overdose.

Intraventricular metrizamide can cause perivascular mononuclear infiltration in the walls of the ventricles. This histological appearance may be mistaken for encephalitis if previous exposure to metrizamide is not considered [ ].

A few cases of severe purulent meningeal reactions to iopamidol have been briefly described [ ]. In one exceptional report, a hemorrhagic meningeal reaction and thrombosis of the superior longitudinal sinus followed sacroradiculography with iopamidol [ ]. A case of transitory abducent nerve palsy has been reported after myelography with iopamidol [ ].

Sensory systems

Psychiatric

To reduce the incidence of generalized reactions to contrast media in high-risk patients some have advocated the prophylactic administration of glucocorticoids (prednisolone 30 mg orally or methylprednisolone 32 mg orally, 12 and 2 hours before contrast injections). In one case an acute psychosis complicated glucocorticoid premedication to reduce the risk of contrast reactions [ ].

• A 13-year-old girl with bipolar disorder and a history of adverse reactions to contrast media was given methylprednisolone (32 mg/day) and ranitidine (300 mg/day) before a CT scan of the head with intravenous contrast enhancement. One day after, she developed psychiatric symptoms, which were more severe than her initial symptoms, including extreme agitation and mental confusion. All medications were withdrawn and her symptoms resolved within 2 weeks.

The authors suggested that the recurrence of the manic symptoms could have been due to premedication with prednisolone. Exacerbation of manic symptoms after the use of glucocorticoids has been documented before, but never in a case of short-term premedication before contrast-enhanced radiographic examination. This report shows that even a short course of glucocorticoids can cause significant adverse reactions in patients with a history of mood disorders.

In 2500 cases of cervical myelography with metrizamide, there were transient mental reactions in 25, including 13 cases of confusion or disorientation, four of depression, two of hallucinations, two of psychosis, and one each of anxiety, drowsiness, dysphasia, and nightmares [ ].

Of 18 German patients undergoing lumbar myelography with metrizamide, six had an organic psychosis, characterized by impaired memory and depression, but it was demonstrable only by psychometric tests and disappeared within 5 days [ ]. In four of the 18 patients there was hyporeflexia or areflexia and in three there were electroencephalographic changes; there was no correlation between these various types of effect.

Visual hallucinations are very rare adverse reactions to contrast media, with isolated reports after vertebral angiography or myelography. The mechanism of this adverse reaction could be similar to that reported in transient cortical blindness after infusion of contrast agents. However, other possibilities include a toxic effect of contrast media on the optic nerve, transient impairment of cerebral blood flow, which could be mediated through the release of the potent vasoconstrictor endothelin, or the formation of microclots. Two cases of visual hallucinations after coronary angiography have been reported [ ].

• A 70-year-old woman with a history of mastectomy developed syncope which lasted a few seconds. She had taken tamoxifen 10 mg bd for 10 years and had no history of allergic reactions. Doppler ultrasound showed aortic stenosis and coronary angiography was performed using 150 ml of iopromide (a non-ionic contrast medium, iodine 370 mg/ml). She had visual hallucinations (spiders on the wall, moving curtains) 30 minutes after the injection of iopromide. The symptoms resolved 72 hours later without any specific treatment. Neurological and psychiatric examinations were normal, as were brain MRI and Doppler ultrasound of the carotid and vertebral arteries.

• A 64-year-old man with a history of ischemic heart disease underwent coronary angiography with 150 ml of iopromide (iodine 370 mg/ml). One hour later he had visual hallucinations (moving objects, pictures of familiar persons), which resolved about 40 hours later without any treatment. He had taken the following drugs for a year: nifedipine 10 mg tds, metoprolol 50 mg bd, and aspirin 325 mg/day. His serum creatinine concentration was in the reference range and there was no history of allergies or previous exposure to contrast media.

Myelography with either iopamidol or metrizamide can cause transitory deterioration in memory as determined by psychological tests, but the effect is less with iopamidol [ ].

Many of the psychiatric complications of cerebral angiography may be due to arterial trauma rather than to the toxic effect of the contrast agent. If the investigation is undertaken under general anesthesia, the use of a volatile anesthetic can in itself cause an increase in intracranial pressure and thereby constitute an aggravating factor [ ]. Focal electroencephalographic changes can occur on the side of the injection, and if these are prolonged they can be followed by evidence of neurological involvement. Transient global amnesia and confusional states have been reported after cerebral angiography, even with non-ionic media [ ].

Endocrine

Contrast agents contain very large amounts of iodine, although it is in a bound form. Liberation of iodine [ ] from these agents can produce some inhibition of thyroid function in healthy subjects for up to 3 months, but can also increase hormonal synthesis in a thyroid adenoma, and cases of frank thyrotoxicosis have been attributed to these agents, the effect starting within a few days [ ]. Hypothyroidism has also been reported, particularly in neonates. Contrast medium-induced hyperthyroidism is rare and usually occurs in patients with autonomous thyroid function. Treatment is exclusively symptomatic. Prophylaxis with sodium perchlorate should be considered in cardiac patients with a goiter and a subnormal concentration of thyroid stimulating hormone (TSH) [ ]. In premature babies and neonates thyroid complications can develop after intravascular administration of iodinated contrast media and great caution should be exercised during radiological examinations in infants [ ].

Thyroid metabolism has been prospectively investigated in 102 patients undergoing diagnostic coronary angiography [ ]. Thyroid function tests (T3, rT3, T4, free T4, and TSH) and urinary iodine excretion were measured before and 3 weeks after diagnostic intra-arterial administration of iodinated contrast agents. Only euthyroid patients were included, in order to determine whether the administration of non-ionic iodine-containing contrast agents causes significant thyroid function changes in euthyroid patients and whether thyroid morphology is a prognostic factor for the risk of hyperthyroidism. Serum concentrations of thyroid autoantibodies (TPO-Ab, Tg-Ab, TSH-receptor-Ab) were measured. Thyroid ultrasound showed that 37 patients had normal thyroid glands. The gland was of normal size but nodular in 16 patients, there was a diffuse goiter in 15 patients, and a nodular goiter in 34 patients. In 25 patients Tg-Ab was positive and in 13 patients TPO-Ab was positive; TSH-receptor-Ab was not detected in any patient. T3 concentrations did not change significantly after the administration of iodine. T4 and free T4 concentrations underwent significantly different changes in the four groups. The amount of iodine given did not affect the changes in the serum concentrations of TSH, T3, T4, free T4, or rT3. Raised concentrations of urinary iodine correlated with the amount of contrast medium given. There were no cases of hyperthyroidism. The study showed that thyroid function was significantly altered after coronary angiography, independent of antibody status and the amount of contrast agent given, but dependent on thyroid morphology.

The effects of iopromide on thyroid function have been investigated in 20 preterm infants with very low birth weights and 26 matched premature infants who did not receive contrast medium [ ]. The dose of iopromide (iodine 300 mg/ml) was 0.3–1.0 ml. Iopromide did not affect the concentrations of free thyroxine and thyroid stimulating hormone. This was attributed to the small amount of free iodide that iopromide contains (0.6 microgram/ml) compared with other contrast media, in which the free iodide concentration ranges from 1.8 micrograms/ml (iohexol) to 4 micrograms/ml (ioxaglate). Furthermore, hypothyroidism has previously been described after the injection of less than 1 ml of ioxaglate 320 in 13 premature infants of less than 34 weeks gestational age and in other children after the injection of iopamidol. The authors concluded that iopromide may be superior to other contrast media in protecting infants of very low birth weight from thyroid dysfunction. It is advisable to monitor thyroid function when contrast media are given to such infants.

• A 54-year-old man developed Graves' disease and hypoadrenalism secondary to adrenocorticotropin deficiency soon after a cranial CT scan with an iodine-containing contrast agent [ ].

It was presumed that the iodine load (about 30 g) had precipitated thyrotoxicosis in this patient, who had antibodies to the thyrotropin receptor, which in turn precipitated collapse due to adrenal insufficiency.

In three women (aged 63, 72, and 75 years) with subclinical goiters, hyperthyroidism developed after the intravenous administration of iodinated contrast medium [ ]. There was a marked rise in the concentration of free T4. The hyperthyroidism improved spontaneously in all three.

In 51 sick neonates given two different non-ionic, iodine-containing contrast agents, metrizamide and iohexol, urinary iodine excretion was increased on day 5 after iodine exposure [ ]. In 17 term neonates given Amipaque, the median TSH concentration was normal after 5 days and 2 weeks, and there was only one case of transient hypothyrotropinemia; median concentrations of T3 and T4 were in the lower reference ranges. However, in 15 neonates given Omnipaque the median TSH was raised and T3 and T4 concentrations were very low. There was hypothyroidism in six of the eight preterm and one of the seven term neonates.

• Mild hypothyroidism with a goiter developed in a 15-year-old boy 6 weeks after lymphangiography with Lipiodol ultrafluid; the goiter disappeared after 3 months treatment with levothyroxine [ ].

Metabolism

Iopanoic acid is as potent a uricosuric agent as probenecid and this effect might explain some renal complications; aspirin reduces the uricosuric effect but can also impair X-ray visualization because of competition at plasma protein-binding sites. Fluctuations of serum urate after oral cholecystography can interfere with diagnostic tests and even precipitate an attack of gout [ ].

Electrolyte balance

Iodinated contrast media can cause increased release of potassium from blood cells and vascular endothelial cells, as has been investigated in vitro using blood, collected from 52 patients, mixed with iodinated contrast media for 30 minutes [ ]. The following contrast media were used: iopamidol (iodine 370 mg/ml), ioxaglate (iodine 320 mg/ml) and diatrizoate (iodine 370 mg/ml). Potassium release increased after exposure to the contrast media and the high-osmolar diatrizoate caused the greatest release, followed by iopamidol and then ioxaglate. The osmolality of contrast media may play an important role in the mechanism responsible for the release of potassium from blood cells. Chemotoxicity may also play a role. There are no data to suggest that the release of potassium is due to hemolysis, and it is most likely due to increased membrane permeability.

Mineral balance

Contrast agents can lower serum calcium and magnesium concentrations [ ], which may be relevant to the occasional occurrence of tetany [ ].

Hematologic

A single case of severe but reversible hypoplastic anemia has been attributed to sodium diatrizoate [ ]. Ionic contrast media have a disaggregating effect on erythrocytes, and hyperosmolar agents reduce their elasticity [ ]. When blood is diluted with 90% sodium diatrizoate in vitro, there is initially a reduction in erythrocyte diameter, due to the hypertonic environment, but as more contrast medium is added, the erythrocytes increase in diameter because of damage to the cell wall; this tallies with the fact that cases of hemolysis and hemoglobinuria have been reported with amidotrizoate [ ].

Iodinated water-soluble contrast agents have traditionally been contraindicated in patients with sickle cell disease, because of possible shrinkage of erythrocytes secondary to the high osmolality of these agents, which can lead to impaired blood flow through the microcirculation and can precipitate or exacerbate a sickle cell crisis. The hematological and rheological effects in vitro of four contrast agents of different osmolalities (iodixanol 290 mmol/kg, ioxaglate 600 mmol/kg, iohexol 844 mmol/kg, and diatrizoate 1940 mmol/kg) have been compared [ ]. Blood was tested from 10 healthy and 10 sickle cell donors at drug concentrations of 0, 1, 10, and 30% w/v in an attempt to approximate the circulating concentrations of contrast medium that might occur during bolus injection. There were significant hematological effects in the blood of both the healthy and sickle cell donors: there was a concentration-related reduction in hematocrit and MCV and an increase in MCHC, all of which varied directly with the osmolality of the contrast medium (amidotrizoate > iohexol > ioxaglate > iodixanol). Only with amidotrizoate at concentrations of 10–30% was there marked echinocytosis. There was no significant increase in the number of irreversibly sickled cells in donors with hemoglobin S. The filterability of erythrocyte suspensions through capillary-sized pores was impaired in both healthy and sickle cell samples in direct proportion to the osmolality of the contrast medium. Filterability effects were greater with sickle cells than healthy erythrocytes. Iodixanol, which is iso-osmolar with blood, had little effect on erythrocyte volume and had no significant effect on the filterability of healthy or sickle cells. These results suggest that microcirculatory impairment after infusion of contrast agents can occur in sickle cell disease because of the unusual rheological sensitivity of HbSS erythrocytes and may be avoided by using low-osmolar or iso-osmolar contrast agents.

Acute thrombocytopenic purpura has been reported in three patients given contrast agents [ , ]. Severe thrombocytopenia has rarely been reported after iopanoic acid [ ], iocetamic acid [ ], and sodium iopodate; in the last case there was evidence that the patient had developed platelet antibodies of the type associated with other drug-induced thrombocytopenias [ ].

• Thrombocytopenia occurred 24 hours after 100 ml of iopamidol was given intravenously during cranial CT scanning to investigate a 9-month history of headache [ ]. The patient reported purpuric lesions on her legs, abdomen, and gingival bleeding within 24 hours of the scan, and examination of the peripheral blood at 48 hours confirmed severe thrombocytopenia. A bone marrow smear showed a prominent increase in megakaryocytes and dysmegakaryopoiesis. The bleeding time was longer than 15 minutes. Other laboratory values were within normal limits. Within 10 days, all the lesions disappeared spontaneously and the platelet count improved gradually and returned to normal within 2 months.

The pathogenesis of this complication is not understood but was most likely an immunological response to the contrast agent.

Clotting time is longer shortly after administration of ionic contrast agents [ ]. Non-ionic contrast agents do not have a similar anticoagulant action. In contrast, if blood is allowed to mix with a non-ionic medium in a syringe or catheter, thrombus formation can occur and this could be a cause of thromboembolism [ ]. The non-ionic agents are not so much thrombogenic as less anticoagulant than their ionic predecessors [ ]. The risk of thromboembolism with ionic agents has been described in some contested studies as being 4–10 times higher [ ]. The non-ionic agents produce profound degranulation of platelets in vitro, but this is unrelated to thrombin generation [ ]. However, some prefer to heparinize when giving the non-ionic media. This issue is contentious, and although non-ionic agents are viewed as being less anticoagulant than ionic agents, they are not considered to be procoagulant [ ]. Adherence to a high standard of angiographic technique, with regular flushing of the catheter with isotonic saline is crucial to avoid thromboembolic complications during angiography.

The effects of contrast agents on leukocytes, platelets, and endothelium have been investigated in 19 subjects (mean age 63 years) undergoing angiography with the non-ionic contrast medium iohexol 350 (median volume 40–160 ml) for leg ischemia [ ]. Blood was obtained from the external iliac vein before and at several intervals after the injection of the contrast agent into the ipsilateral femoral artery. Markers of endothelial cell injury (von Willebrand factor), platelet activation (soluble P selectin), and leukocyte activation (neutrophil elastase and soluble L selectin) were measured in citrated plasma. Soluble intracellular adhesion molecule-1 and thromboxane B ₂ , which are non-specific markers of inflammation, were also measured. Compared with the sample before angiography, concentrations of soluble L selectin and soluble intracellular adhesion molecule-1 (ICAM-1) were reduced immediately after passage of the last bolus of contrast medium; 15 minutes later the concentrations returned to normal, but the concentration of von Willebrand factor had increased. After 30 minutes, only thromboxane B ₂ concentrations were increased. On the next day both von Willebrand factor and soluble P selectin were increased. These data point to both early and late effects of contrast agents on markers of endothelial, platelet, and leukocyte function. The authors suggested that these adverse changes may increase the risk of coagulopathy and thrombosis after contrast examination and increase the risk of re-stenosis after angioplasty.

The hematological effects of ioxaglate (a low-osmolar ionic dimer) 105 (range 95–114) ml (n = 15) and iopromide (a low-osmolar non-ionic monomer) 102 (range 90–108) ml (n = 16) have been investigated in patients undergoing abdominal and femoral angiography [ ]. The aim was to investigate in vivo whether non-ionic contrast media are less anticoagulant or more prothrombotic than ionic agents. Activation of coagulation and platelets were found in almost 50% of patients before any contrast medium was given. Both iopromide and ioxaglate caused further increases in thrombin-antithrombin complex, prothrombin fragments 1 + 2, and beta-thromboglobulin; the degree of activation was similar with both agents. In contrast to the findings in in vitro studies, there were no significant differences between the effects of the non-ionic agent iopromide and the ionic agent ioxaglate. The results supported the notion that the catheterization procedure per se may represent a source of hemostatic activation and that ionic contrast agents have insufficient anticoagulant effect to prevent clotting activation being induced by the procedure. The study also yielded no support for the concept that non-ionic contrast media are less anticoagulant or more prothrombotic than ionic agents.

However, it has been suggested that anticoagulant effects are high with ionic media and low with non-ionic media [ ]. The authors also suggested that there is no direct activation of platelets with low-osmolar ionic and non-ionic dimeric contrast agents but a high degree of activation with non-ionic monomeric contrast media. They concluded that although the interaction between contrast media and coagulation has been widely studied in vitro and in vivo, this issue is contentious and further studies are required for better understanding.

Within a few hours after lymphography, the lymph nodes show dilatation of the marginal and intermediate sinuses and a giant-cell reaction, with diffuse reticulocytosis and sinus histiocytosis. Diffuse plasmacytosis and an increase in the eosinophil count can also occur. The response is maximal after 10–14 days, but the changes in the lymph nodes may not be eliminated for as long as 15 months [ ].

The thrombotic complications after angioplasty have been investigated in patients with unstable angina [ ]. There was no significant difference in the incidence of thrombotic complications between patients who received the low-osmolar ionic dimer ioxaglate (n = 103) and those who received the non-ionic agent iopamidol (n = 102), although there was a non-significant trend towards more thrombus formation in the non-ionic group (21 of 129 patients) compared with the ionic group (15 of 141 patients). The two groups were well matched with respect to age, sex, class of unstable angina, and susceptibility factors. There was no significant difference between the two groups with respect to clinical outcome in the first 24 hours after percutaneous coronary angioplasty.

• Disseminated intravascular coagulation has been described in a 63-year-old man who received 50 ml of the non-ionic monomer iobitridol (iodine 300 mg/ml) for arteriography [ ].

Salivary glands

Iodine is highly concentrated by the salivary glands, and radioiodine can cause direct damage to the glands [ ] (see the monograph on radioactive iodine).

Sialadenitis associated with the administration of non-radioactive iodinated compounds (“iodide mumps”) is an uncommon adverse reaction that was first reported in 1956 by Sussman and Miller in relation to an intravenous high-osmolar ionic contrast medium (sodium diatrizoate + meglumine diatrizoate) [ ]. Since then sporadic reports have appeared.

Iodide mumps has been reported to affect the parotid glands [ ], the submaxillary glands [ ], and the submandibular glands [ ]. In one case asymptomatic parotid gland swelling after injection of iotalamate salts was accompanied by pancreatic swelling, to which the authors gave the name “pancreatic mumps” [ ]. In another case submandibular gland swelling was accompanied by thyroid swelling in a patient with renal insufficiency and a very high serum iodine concentration after administration of iodixanol; ultrasonography showed that both glands were edematous [ ]. The authors recommended that patients on renal replacement therapy should undergo dialysis immediately after the use of iodine-containing contrast media.

In a systematic review of 79 published case reports of iodide mumps, of which 68 were due to iodinated contrast media, the ratio of men to women was 3:2; their respective mean ages were 61 (range 27–83; median 62) and 55 (range 8–80; median 63) years [ ]. The glands affected and whether there was associated pain are shown in Table 4 ; pain was a common feature.

The time to onset of symptoms was stated in 76 of the cases, in whom it ranged from “immediate” to 5 days (median 1 day). The times were bimodally distributed. Renal function was reported in 58 patients, of whom 26 were reported as having impaired renal function. In those with renal impairment the median time to onset was 24 hours (range 5 minutes to 48 hours) and in those without renal impairment it was 12 hours (range 2–120). Significantly more patients with a time to onset beyond 24 hours had impaired renal function. Taken together, these results suggest that there may be two different types of reaction: immediate hypersensitivity and either delayed hypersensitivity or a toxic effect. The EIDOS and DoTS descriptions are shown in Figure 1 .

Iodide mumps is rare. In one study of adverse reactions to intravenous contrast media in 337 647 patients there were no reports, which suggests that the maximum likely incidence is about 9 per million [ ]. However, it can occur on a second occasion after re-exposure [ ], suggesting that some individuals may be particularly susceptible.

In most cases resolution is spontaneous. Some patients have been treated with glucocorticoids.

Gastrointestinal

Nausea and vomiting can occur in reaction to contrast agents. Diarrhea is less common but has been repeatedly reported, sometimes with angioedema of the bowel [ ].

Infarction of the bowel was formerly a very occasional complication of abdominal aortography and was due to injection of contrast medium into the mesenteric arteries. Most of these cases were due to the older media, such as acetrizoate; however, small bowel injury has occurred after injection of a concentrated bolus of sodium iotalamate [ ]. Ileus has been reported after mesenteric angiography in a patient with renal insufficiency [ ].

Intra-arterial injection of lipiodol (iodized oil) has been used to enhance the accuracy of computed tomography in hepatic tumors. This can cause transient bowel ischemia with nausea, vomiting and diarrhea.

The use of Gastrografin for preoperative mechanical bowel preparation has been investigated in 58 patients (aged 45–80 years) listed for elective colorectal operations [ ]. One group (30 patients, mean age 67 years) was given oral Gastrografin 200 ml and 3 liters of water for 2 days before the operation. The rest (28 patients, mean age 65 years) were given Ringer’s solution 5–20 liters, warmed to body temperature, through a nasogastric tube; preparation was considered complete when the patient excreted clear fluid. All were given metronidazole and cefuroxime as antibiotic prophylaxis during the peri-operative period. There were no significant adverse effects in the Gastrografin group, apart from nausea in six patients. In the Ringer’s solution group there was nausea in 15 patients and vomiting in eight. A clean colon was found at operation in 93% of patients in both groups. The authors concluded that Gastrografin is well tolerated and can be used successfully for mechanical bowel preparation before elective colorectal surgery.

The rectal administration of Gastrografin 400–1000 ml in patients with suspected diverticular disease before CT examination of the abdomen has been investigated in 308 patients (aged 19–97 years) [ ]. None of the CT scans showed extravasated contrast material in the peritoneal cavity as a sign of bowel perforation. No patient had sudden clinical deterioration after the examination. All tolerated the contrast medium well and there were no allergic reactions. High-quality diagnostic examinations were obtained in all patients.

Contrast agents caused release of vasoactive intestinal polypeptide in a patient with a vipoma of the pancreas and hepatic secondaries [ ].

The high-osmolar agents amidotrizoate and iotalamate were formerly recommended for oral use in preference to barium where there was a risk of perforation, but the osmotic effect of such material can lead to diarrhea and fluid loss, which can be dangerous in weak patients. Amidotrizoate, despite its cathartic action, can occasionally cause ileus when given postoperatively [ ]; it can also be precipitated as a solid mass in the stomach when gastric acidity is high, but also in an achlorhydric gastric stump following partial gastrectomy if there is stomal obstruction. On one occasion, when diatrizoate was used to fill the gastric balloon of a Sengstaken–Blakemore tube, hydrogen ions from the gastric content apparently penetrated the balloon, precipitated the contrast medium, and thus prevented the balloon’s removal.

Low-osmolar media such as iohexol are now preferred for oral use; they can still cause some diarrhea, for example in 18 out of 40 cases in one series [ ]. Absorption can be increased if there is mucosal damage in the bowel, such as in Crohn’s disease, resulting in delayed excretion but not apparently involving risk [ ].

Amidotrizoate enemas are still sometimes used to treat meconium ileus or constipation, and it is important to give intravenous fluids so as to avoid dehydration. Hypomagnesemia can also occur [ ]. Osmotic effects lower in the gastrointestinal tract have even led to distention and cecal perforation [ ]. Stasis of amidotrizoate in dilated loops of bowel can cause inflammatory changes or necrosis [ , ].