INTRODUCTION

TOP ABSTRACT INTRODUCTION REPORT OF CASES DISCUSSION REFERENCES

Both sleep apnea-hypopnea syndrome (obstructive sleep apnea [OSA]) and hypothyroidism are common problems in internal medicine. Each disorder probably occurs in at least 2% (1) of the North American population. Unfortunately, OSA and hypothyroidism can easily be confused. Obesity, fatigue, decreased libido, depressed mood, and impaired concentration are symptoms common to both disorders. Even snoring, which is a prominent presenting complaint with OSA, has been observed universally within at least one group of hypothyroid patients (4). Periorbital and peripheral edema are usually associated with hypothyroidism, but are also frequently seen in OSA (5). Because the two disorders have been estimated to occur together in 1.6 to 11% of sleep clinic patients (6, 7), the sleep-respiratory clinic environment can be an opportunity for misdiagnosis.

Differentiation between the two disorders is made more difficult because hypothyroid patients are also at risk for secondary sleep-disordered breathing. Other respiratory complications that occur in some hypothyroid individuals include upper airway obstruction (secondary to goiter), obesity, respiratory myopathy, and blunted ventilatory chemosensitivity (8). Each of these complications can contribute to sleep-disordered breathing. Obesity, airway obstruction, and altered ventilatory chemosensitivity also occur in primary OSA. Is a "positive" diagnostic polysomnogram which demonstrates abnormal hypopneas or apneas evidence of primary OSA or "secondary sleep apnea," that is, sleep-disordered breathing secondary to underlying hypothyroidism?

This differentiation is not moot. Although an unlikely event in a sleep clinic, misdiagnosing OSA as hypothyroidism could result in inappropriate thyroxine therapy, which has not been shown to be effective treatment for primary OSA. However, the probability of mistakenly declaring OSA when the correct diagnosis is hypothyroidism, is much greater in the sleep clinic, and the consequences could be serious. Inappropriate treatment for OSA, including prescription of continuous positive airway pressure (CPAP), could relieve snoring and might normalize the polysomnogram, as a short-term "success." Meanwhile, the patient's presenting symptoms would persist, while a false diagnosis of OSA would delay or impede proper diagnosis and treatment of hypothyroidism. Supersensitive serum thyroid stimulating hormone (TSH) testing could be used to diagnose hypothyroidism among sleep clinic patients. However, the feasibility and economic justification of this biochemical screening has not been established (7).

A final question remains unanswered for this group of patients with hypothyroidism masquerading as OSA. Even with a correct diagnosis, what is appropriate minimum treatment for "secondary" sleep apnea, i.e., the sleep-disordered breathing of hypothyroidism? Since thyroxine therapy has been used with varying success to treat sleep-disordered breathing in hypothyroidism (4, 9), is CPAP a necessary adjunctive treatment for all such hypothyroid apneic patients?

Therefore, to determine the proportion of undiagnosed hypothyroid patients who were likely to be misdiagnosed as OSA, the efficacy of thyroxine treatment alone for "secondary" apnea of hypothyroidism, and the cost-effectiveness of thyroid testing, we did biochemical thyroid screening prospectively in all sleep clinic patients who underwent polysomnography (PSG) in the investigation of sleep apnea.

	REPORT OF CASES

TOP ABSTRACT INTRODUCTION REPORT OF CASES DISCUSSION REFERENCES

Over 20 mo, 290 sequential new patients were seen in sleep clinic. Based on either abnormal screening nocturnal oximetry, or a history of witnessed apneas, 200 (69%) of these patients underwent PSG for suspected OSA. TSH was measured in all patients undergoing PSG. All sleep studies were reviewed by the authors. Hypopneas were defined as 10- to 60-s periods with reductions in flow  50% accompanied by oximetry desaturations greater than 4%.

Among the 200 patients who underwent PSG, 124 (62%) were judged to have OSA based upon the presence of sleep-disordered breathing characterized by a respiratory disturbance index (RDI)  10/h. Most of these were started on treatment with CPAP. However, before CPAP was prescribed for any individual, the results of the thyroid function tests were reviewed.

Three patients (Cases A, B, and C) who underwent PSG and TSH testing were found to have both sleep-disordered breathing (RDI  10/h) and biochemical hypothyroidism (TSH  20 milliunits/liter [mU/L] with depressed free T₄ index [FT₄I]. These three patients accounted for 1.5% of patients undergoing PSG or 2.4% of those judged to have OSA based on the PSG. These three patients are described in detail subsequently.

Three additional patients were found to have both OSA and hypothyroidism, but are not included in this report. Two combined OSA-hypothyroidism patients who were excluded had a prior history of hypothyroidism, and were already on thyroxine prior to their PSG. The third patient had persistent elevations in TSH and OSA, but had serially normal FT₄I values. He was treated with CPAP.

Following informed consent, the three patients (Cases A, B, and C) who were found to be hypothyroid at the same time as their PSG had confirmed sleep-disordered breathing, were started on thyroxine therapy alone as treatment for both disorders. Other treatments for OSA, including CPAP and dental appliances, were not prescribed.

Case A

A 52-yr-old white male with a remote history of angina pectoris, depression, and hypercholesterolemia presented with a history of daytime somnolence, fatigue, witnessed nocturnal apneas, dysphonia, and loud snoring. He had no prior or family history of thyroid disease. He was on diltiazem and fluvoxamine. On physical examination he had alopecia of the lateral eyebrows, dysphonia, a smooth firm goiter, and normal tendon reflexes.

His initial PSG showed an RDI of 30/h with 53% of his total sleep time (TST) spent with oxygen saturation  85% while breathing room air. His initial TSH was 189 mU/L; creatine kinase, 8,060 U/L; lactate dehydrogenase, 554 U/L; and random cholesterol, 8.2 mmol/L. A flow- volume loop showed no evidence of upper airway obstruction. He was placed on 25 µg of thyroxine daily, which was increased slowly to 150 µg daily given his history of angina. Within 2 wk his symptoms had resolved. Over 20 mo on thyroxine his TSH decreased to 8.4 mU/L (Figure 1A). Serial sleep studies showed decreased respiratory disturbances during sleep (minimum RDI, 1.7/h) and resolution of the nocturnal hypoxia (minimum 9% of TST with Sa_O2  85%) (Figure 1A). Over the same 20 mo his body-mass index (BMI) remained unchanged: 30.7 to 29.2 kg/m².

View larger version (21K): [in this window] [in a new window]   Figure 1.   The effect of thyroxine therapy on serum TSH, RDI, and nocturnal hypoxia in three patients with hypothyroidism and sleep-disordered breathing.

Case B

A 44-yr-old white woman with hypercholesterolemia and prior left mastectomy for localized adenocarcinoma presented with a 2-yr history of loud snoring, witnessed nocturnal apneas, dyspneic awakenings, fatigue, and poor concentration. She had no prior or family history of thyroid disease. She was on lovastatin and cholestyramine resin. On physical examination she had a normal thyroid gland, a mastectomy scar, no palpable breast masses, no palpable adenopathy, and normal tendon reflexes.

Her initial PSG showed an RDI of 14/h with 7% of TST spent with Sa_O2  85% while breathing room air. Her initial TSH was 204.6 mU/L; lactate dehydrogenase, 384 U/L; antithyroglobulin antibody titer, 1:6,400; and antimicrosomal antibody titer, 1:25,600. Her cholestyramine was discontinued, and she was started on 150 µg of thyroxine daily. Within 4 wk her symptoms had resolved. Over 12 mo on thyroxine her TSH decreased to 4.09 mU/L (Figure 1B). Serial sleep studies showed decreased respiratory disturbances during sleep (minimum RDI, 1.0/h) and resolution of her nocturnal hypoxia (minimum 1% of TST with Sa_O2  85%) (Figure 1B). Over the same 12 mo her BMI was unchanged: 21.9 to 21.7 kg/m².

Case C

A 47-yr-old white man with a history of hyperlipidemia, gastroesophageal reflux, myocardial infarction, and paraphilia presented with daytime somnolence, loud snoring, fatigue, and headaches on waking. He had no prior or family history of thyroid disease. He was on omeprazole, acetylsalicylic acid (ASA), and metoprolol. On physical examination he was obese with a normal thyroid gland and normal tendon reflexes.

His initial PSG showed an RDI of 24/h with 23% of his TST spent with Sa_O2  85% while breathing room air. His initial TSH was 190.4 mU/L. He was placed on thyroxine 100 µg daily, which was subsequently increased to 175 µg daily. Within 2 mo his symptoms had improved substantially although he remained in a chronic psychiatric hospital for his paraphilia. Over 5 mo on thyroxine his TSH decreased to 4.4 mU/L (Figure 1C). Serial sleep studies showed decreased respiratory disturbances during sleep (minimum RDI, 16/h) and nocturnal hypoxia (minimum 0.2% of TST with Sa_O2  85%) (Figure 1C). Over the same 5 mo his BMI remained stable: 33.8 to 34.7 kg/m².

	DISCUSSION

TOP ABSTRACT INTRODUCTION REPORT OF CASES DISCUSSION REFERENCES

Cases A, B, and C illustrate the potential for misdiagnosis and therapeutic misadventure if hypothyroidism is not definitively excluded among individuals who are diagnosed with OSA. These patients all presented in sleep clinic with features that were strongly suggestive of OSA and underwent a PSG. We found these three among 200 similar patients for a 1.5% point prevalence of undiagnosed hypothyroidism among patients undergoing PSG or a 2.4% prevalence of undiagnosed hypothyroidism among 124 individuals judged to have OSA based on their sleep study. Typically, in a sleep-respiratory clinic, such patients with a significantly elevated RDI on PSG, would be judged to have "primary" OSA, and would be prescribed CPAP or other treatment for OSA, without further investigation. However, Cases A, B, and C actually had "secondary sleep apnea" caused by previously undetected hypothyroidism; only deliberate TSH screening unearthed the correct primary diagnosis. For these individuals, their "OSA" was more precisely sleep-disordered breathing secondary to undiagnosed hypothyroidism. With a proper diagnosis, their "sleep apnea" then responded to thyroxine treatment alone, with relief of symptoms and resolution of abnormal nocturnal respiration without CPAP.

Consider the probable outcome if CPAP or some other treatment for OSA had been prescribed instead, in error, for these individuals. Treatment for OSA in each of these patients would have provided a temporary "improvement" in sleep-disordered breathing, with inevitable, long-term failure. The "improvement" would have significantly delayed correct diagnosis and treatment of hypothyroidism. These cases demonstrate the importance of careful examination of the OSA patients for important, coexisting conditions such as hypothyroidism. To avoid inappropriate treatment for OSA, we may consider routine screening for hypothyroidism among sleep apnea patients. However, to make such a recommendation, these cases must be representative of the expected prevalence and coprevalence/concurrence of these two disorders, and further, correct diagnosis of hypothyroidism among sleep apnea patients must be reasonably cost-effective. We will consider each of these conditions in turn.

Prevalence of Hypothyroidism with OSA

Several other studies have investigated the prevalence/coprevalence and management of hypothyroidism with OSA. Among a comparable group of patients referred to sleep clinic, Winkelman and coworkers (7) found 1.6% of sleep clinic patients had hypothyroidism, and 2.9% of patients with confirmed OSA also had hypothyroidism. In a recent abstract report, the prevalence of coexistent hypothyroidism with OSA was even greater (6); of 95 patients suspected of having OSA 72 underwent PSG, 53 had OSA, and six of 53 (11%) were found additionally to be hypothyroid. Conversely, among 65 Taiwanese sleep clinic patients, Lin and coworkers (4) found only one case of hypothyroidism and one case of postpartum panhypopituitarism. Thus, exact values may vary between reports, but our observation that 2.4% of patients judged to have OSA based on their polysomnogram actually had "secondary sleep apnea" arising from previously undiagnosed hypothyroidism appears to be a reasonable estimate of coprevalence within sleep-respiratory clinic patients.

Clearly, these individuals with sleep-disordered breathing secondary to hypothyroidism responded to thyroxine replacement therapy alone as sole treatment for their "sleep apnea." However, the exact proportion of such patients who can be fully treated with thyroxine alone is uncertain. Several investigators (10) have reported on individual patients with hypothyroidism and sleep apnea for whom sleep respiratory abnormalities resolved with thyroxine replacement. In one series (4), all five patients with hypothyroid-related respiratory abnormalities were effectively treated with thyroxine replacement, whereas in another report (9), only five of 10 newly diagnosed, combined hypothyroid-apneic patients had alleviation of their apnea on thyroxine treatment. Here, sleep disordered breathing in all three of our cases was resolved or significantly improved with thyroxine. Rather than debate the proportion of patients with sleep-disordered breathing secondary to hypothyroidism who might be fully treated with thyroxine alone, we prefer another perspective. Certainly, no patient with sleep-disordered breathing secondary to hypothyroidism will be treated safely or effectively using CPAP or other treatments prescribed in error for primary OSA.

Rationale for Thyroid Screening among Sleep Clinic Patients

Based on our case prevalence and the other reports cited previously, we can reasonably conclude that without thyroid screening of sleep clinic patients, approximately 2 to 3% of patients diagnosed with primary sleep apnea will be misdiagnosed and inappropriately treated because of undetected hypothyroidism. In addition, the same 2 to 3% of OSA patients will be "treatment failures" on CPAP or other standard sleep apnea therapy, and the correct diagnosis of hypothyroidism will be delayed.

We know that there is a personal and clinical cost for each case of undiagnosed and untreated hypothyroidism. However, these penalties are hard to quantify. Hypothyroid patients face increased risks of hypercholesterolemia (13), an optic neuropathy mimicking glaucoma (14), and organic mental disorders (15), as well as the sleep-disordered breathing described here. We cannot estimate the risk of other complications and overall mortality for these misdiagnosed patients. Obviously, a proper diagnosis based on biochemical thyroid testing of sleep clinic patients would be preferred, but is such thyroid screening cost-effective?

Cost-effectiveness of Thyroid Screening

One previous report concluded that wholesale thyroid screening among patients "either suspected of, or even with confirmed, OSA" was not justified financially (7). That decision was based upon a projected saving of only US$1,200, approximately equivalent to the cost of one CPAP device, balanced against costs of $50 per thyroid test for a total expenditure of over US$5,000 to exclude hypothyroidism in a group of patients slightly smaller than we reported on here. We surveyed Medicare billing practices of several public and private health care organizations in Washington State as they were the closest centers of similar size to ours. The differences between TSH screening costs and CPAP prescription for our provincial health plan and modal Medicare billings is shown in Table 1. The perceived "cost-effectiveness" of thyroid screening in patients with sleep apnea confirmed by polysomnogram depends on the method of billing and local costs for laboratory testing (TSH), polysomnogram, and CPAP purchase. Based on our costs, in this region there was a projected net saving of $494 if all these patients had thyroid screening tests. On the other hand, in a neighboring region of the United States, routine thyroid screening tests for this entire group would have cost $584 extra. However, we are convinced that these regional calculations underestimate the dollar savings, because the calculations do not consider the unavoidable additional costs of leaving three misdiagnosed patients with undetected, symptomatic hypothyroidism in the community. There must be additional expense before a delayed diagnosis of hypothyroidism is finally made.

In summary, we believe that there is a financial justification for biochemical thyroid testing among patients diagnosed with sleep apnea in a sleep clinic. However, the personal and clinical "costs" for individual patients who are misdiagnosed may be even greater.

Finally, we propose a different perspective on the cost- effectiveness of thyroid screening in this patient group, derived from the notion of "secondary" sleep apnea. That is, used alone in this group, the specificity of the polysomnogram was in error by at least 2.4%, because that proportion of patients were declared incorrectly to have primary OSA. Our local cost of a PSG is approximately US$550. If we consider the thyroid screening test to be an adjunct to the PSG, then the local sleep study cost increases to $550 plus the cost of a TSH, which is $15.37, for a total of $565.37, an increase of about 2.8%. In return, the specificity of the sleep investigation would increase by 2.4% because the hypothyroid misdiagnosis would be avoided. Thus, the diagnostic accuracy of the sleep study, as expressed by specificity, would be improved by 1% for each additional 1% of testing expense. That is an extraordinarily cost-effective change in diagnostic procedures for a serious disease.

There is a strong analogy with the exclusion of uncommon aggravating or causal factors in the investigation of other serious chronic diseases. In the investigation of presumed hypertension, a complete blood cell count, blood glucose, potassium, calcium, creatinine, uric acid, cholesterol level, triglyceride level, and electrocardiogram are recommended as routine initial investigations (16), before a diagnosis of primary hypertension is entertained. Although the yield of such investigations is very low, available evidence and expert opinion concur that the benefit of identification and treatment of secondary causes of hypertension and the prevention of complications warrant these screening costs. We believe that the same argument applies to testing and identification of hypothyroidism as a secondary cause of sleep-disordered breathing.

Follow-up

We have identified patients with sleep apnea and hypothyroidism, where thyroxine replacement reversed both secondary sleep-disordered breathing and other symptoms, that would not have been controlled with CPAP. However, it must be emphasized that careful follow-up of both the respiratory disturbance and thyroid status is required in such patients. In particular, close observation is required to ensure that the sleep apnea is substantially cured once euthyroid status is achieved.